Angiogenesis

Models, Modulators, and Clinical

Applications
NATO ASI Series
Advanced Science Institutes Series
A series presenting the results of activities sponsored by the NA TO Science Committee,
which aims at the dissemination of advanced scientific and technological knowledge,
with a view to strengthening links between scientific communities.
The series is published by an international board of publishers in conjunction with the
NATO Scientific Affairs Division

A Life Sciences Plenum Publishing Corporation

B Physics New York and London

C Mathematical Kluwer Academic Publishers

and Physical Sciences Dordrecht, Boston, and London
D Behavioral and Social Sciences
E Applied Sciences

F Computer and Systems Sciences Springer-Verlag

G Ecological Sciences Berlin, Heidelberg, New York, London,
H Cell Biology Paris, Tokyo, Hong Kong, and Barcelona
I Global Environmental Change

PARTNERSHIP SUB-SERIES
1. Disarmament Technologies Kluwer Academic Publishers
2. Environment Springer-Verlag
3. High Technology Kluwer Academic Publishers
4. Science and Technology Policy Kluwer Academic Publishers
5. Computer Networking Kluwer Academic Publishers

The Partnership Sub-Series incorporates activities undertaken in collaboration with

NA TO's Cooperation Partners, the countries of the CIS and Central and Eastern
Europe, in Priority Areas of concern to those countries.

Recent Volumes in this Series:

Volume 295 - Prions and Brain Diseases in Animals and Humans
edited by Douglas R. O. Morrison

Volume 296 - Free Radicals, Oxidative Stress, and Antioxidants: Pathological and
PhYSiological Significance
edited by Tomris Ozben

Volume 297 - Acute Respiratory Distress Syndrome: Cellular and Molecular

Mechanisms and Clinical Management
edited by Sadis Matalon and Jacob lasha Sznajder

Volume 298 - Angiogenesis: Models, Modulators, and Clinical Applications

edited by Michael E. Maragoudakis

Series A: Life Sciences

Angiogenesis
Models, Modulators, and Clinical
Applications
Edited by

Michael E. Maragoudakis
University of Patras Medical School
Patras, Greece

Springer Science+Business Media, LLC

Proceedings of a NATO Advanced Study Institute on
Angiogenesis: Models, Modulators, and Clinical Applications,
held June 20-30,1997,
in Rhodes, Greece

NATO-PCO-DATA BASE

The electronic index to the NATO ASI Series provides full bibliographical references (with
keywords and/or abstracts) to about 50,000 contributions from international scientists
published in all sections of the NATO ASI Series. Access to the NATO-PCO-DATA BASE is
possible via a CD-ROM "NATO Science and Technology Disk" with user-friendly retrieval
software in English, French, and German (©WTV GmbH and DATAWARE Technologies, Inc.
1989). The CD-ROM contains the AGARD Aerospace Database.

The CD-ROM can be ordered through any member of the Board of Publishers or through
NATO-PCO, Overijse, Belgium.

Library of Congress Cataloging-in-Publication Data

On file

ISBN 978-1-4757-9187-7 ISBN 978-1-4757-9185-3 (eBook)

DOI 10.1007/978-1-4757-9185-3

© 1998 Springer Science+Business Media New York

Originally published by Plenum Press, New York in 1998.
Softcover reprint of the hardcover 1st edition 1998
http://www.plenum.com
10987654321
All rights reserved
No part of this book may be reproduced, stored in a retrieval system, or transmitted in anl
form or by any means, electronic, mechanical, photocopying, microfilming, recording, or
otherwise, without written permission from the Publisher
PREFACE

Interest in angiogenesis research remains strong in recent years and exciting new
discoveries, about modulators of angiogenesis, their receptors, the transduction
mechanisms and the angiogenic genes involved, have contributed to our present day
understanding of this complex process. This knowledge has provided the basis and
broadened the scope of angiogenesis - based therapy in oncology and many other clinical
conditions.

This monograph contains the contributions to the NATO Advanced Study

Institute on "Angiogenesis: Models, Modulators and Clinical Applications", which was
held in Rhodes, Greece, from June 20-30, 1997. This was the fourth of a series of NATO
supported international meetings on Angiogenesis aiming to bring together basic
scientists with clinicians to exchange ideas, disseminate new knowledge and discuss the
present status and potential new directions in this fast moving area of biomedical
research.

The International Organising Committee that included Drs. E. Dejana, C

Haudenschild, M. Hackel, H. Kleinman, P. Lelkes, M. Presta, P. Polverini, D. Thompson,
has provided invaluable help with their insightful suggestions in the formulation of the
scientific program for which I am grateful. I wish to thank all the participants for their
enthusiastic participation and their complimentary comments on the success of the
conference.

Special thanks are due to the Scientific Affairs Division of NATO for providing
the major portion of the grants for the organization of the meeting and also for
publication of this book. The contributions of the following organizations: Carbomed
(USA), Janssen (Greece), Schering Plough (Greece), SmithKline Beecham (USA) and
Viannex (Greece), which was used to support the support of young scientists is
gratefully acknowledged.

I am particularly grateful to Mrs. Anna Marmara and Mrs. Georgia Tziora for their
enthusiastic and dedicated work throughout the organization of the meeting and the
editing of this monograph.

Michael E. Maragoudakis (Greece)

v
CONTENTS

Introductory Comments
M. E. Maragoudakis

MODELS AND METHODS FOR ASSESSING ANGIOGENESIS

Microvascular Endothelial Cells from Adrenal Medulla- A Model for in Vilro

Angiogenesis ...................... . .... .. . . .... ... ...... .. ........ . 7
D. K. Banerjee and J. A. Martinez

Angiogenesis in the Heart and skeletal Muscle- Models for Capillary Growth 19
O. Hudlicka, S. Egginton, and M. D. Brown

Basement Membrane Biosynthesis as a Biochemical Index of Angiogenesis in

Chick Chorioallantoic Membrane ...................................... 35
N. Tsopanoglou, and M. E. Maragoudakis

The Rat and Rabbit Cornea Assays ... .. .. ..... .. . ..... . ..... . . . .... . . . ... .. 39
L. Morbidellia, and M. Ziche

The Rat aorta Model of Angiogenesis: Methodology and Applications. . . . . . . . . . . .. 47

R. F. Nicosia

Assessment of Angiogenesis by MRI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 55

M. Neeman

Measuring Intratumoral Microvessel Density .............. . . ....... ..... . .... 61

N. Weidner

Patterns of Physiological Angiogeneis in Adult Mesentery. . . . . . . . . . . . . . . . . . . . . .. 75

F. Hansen-Smith, and L. Morris

Angiogenesis---Critical Assessement of in Vitro Assays and in Vivo Models. . . . . . . .. 85

D. BenEzra

ANGIOGENIC FACTORS AND THEIR RECEPTORS

The Several Roles of Oxygen in Wound Healing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 93

J. Gibson, T. Hunt, 1. Feng, M. Rollins, A. Sheikh, and Z. Hussain

VII
 Autocrine Role of Basic Fibroblast Growth Factor (B-FGF) in Angiogenesis and
Angioproliferative Diseases ........................................... 99
A. Gualandris, P. Dell'Era, M. Rusnati, R. Giuliani, E. Tanghetti.
M. P. Molinari-Tosatti, M. Ziche, D. Ribatti. and M. Presta

Inflammatory Angiogenic Factors in a Model of Chronic Inflammation ....... .... . 113

J. D. Winkler. M. P. Seed, and 1. R. Jackson

Angiogenic Mediators in Wound Healing 121

L. DiPietro, and N. N. Nissen

Modulation of VEGF Angiogenic Activity by ADP-Ribosylation ................. 129

J . J. Feng, Q. P. Ghani, G. Ledger. R. Barkhordar, T. Hunt, and Z. Hussain

Role of Fibroblast Growth Factor-2 and Endothelial Cell Stimulating Angiogenic

Factor (ESAF) in Capillary Growth in skeletal Muscles Exposed to Long-Term
High Activity ....................................................... 135
M. D. Brown, H. Walter, O. Hudlicka, F. M. Hansen-Smith, and 1. B. Weiss

Angiogenesis in Atherosclerosis: Possible Roles for ESAF, VEGF, VWF and

Soluble E-Selectin ......... . .. . .... . . . .. ... .......... . ......... . ..... 149
J. Weiss, A. Blann, J. L. Li, C. N. McCollum, and A. Bate

Thymosin Beta 4 Promotes Endothelial Cell Migration and Angiogenesis. . . . . . . . . .. 157

K. Malinda, A. L. Goldstein, D. S. Grant, and H. Kleinman

Structural Studies on Angiogenin, a Protein Implicated in Neovascularisation during

Tumour Growth. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 165
K. R. Acharya, D. D. Leonidas, A. C. Papageorgiou, N. Russo, and R. Shapiro

Lesional Levels of ESAF and VEGF Are Elevated in Psoriasis 179

B. McLaughlin, M. Bushan, C. M.Griffiths

ROLE OF CELL ADHESION MOLECULES AND EXTRACELLULAR

MATRIX IN ANGIOGENESIS

Structure and Functional Role of Endothelial Cell-to Cell Junctions ........ ... .... 187
P. Navarro, M. G. Lampugnani, and E. Dejana

The Interaction of Human Neutrophils with Type IV Collagen Involves an

Inhibitory Signal Transduction Pathway ................. . ............... 203
1. C. Monboisse, G. Bellon, R. Garnotel, A. Fawzi, N. Ohno, N. A. Kefalides, and
J. P. Borel

Angiogenic Polypeptides in Breast Cancer. Expression of mRNAs in Primary

Human Tumours, MCF -7 Cell Transfection and Xenograft Models ............ 213
H. T. Zhang, R. Choudhuri, P. A.E. Scott, L. Zhang, M. Ziche, L. Morbidelli,
S. Donnini, R. T. Jaggar, H. Y. Chan, K. Smith, S. Peak, M. c.P. Rees,
A. L. Harris, and R. Bicknell

Vlll
 ROLE OF THROMBOSIS AND FIBRINOLYSIS IN ANGIOGENESIS

The Role of Thrombin and Its Receptors in Angiogenesis: Physiological and

Pathological Implications . . .... . .. . .. . . . . .... ............. . ....... .. .. 225
M. E. Maragoudakis, N. Tsopanoglou, and E. Pipili-Synetos

Fibrin Degradative Pathways and Angiogenesis in Healing, Atherosc1erosis and

Tumor Invasion ........ . : .. ......... . .... .. ........ . ........... .... . 233
D. Thompson, C. M. Stirk, A. 1. Keating, A. Reid, E. B. Smith, and W. T. Melvin

Proteases and Angiogenesis: Regulation of Plasminogen Activators and Matrix

Metalloproteases by Endothelial Cells ..... . .. . ........ . . . ........... . ... 241
P. Koolwijk, R. Hanemaaijer, and V. W.M. van Hinsbergh

Cellular and Molecular Effects of Thrombin in the Vascular System ...... .... ..... 263
C. Kanthou, V. V. Kakkar, and O. Benzakour

REGULATION OF ANGIOGENESIS

Endogenous Regulation of Angiogenesis in Vitro ... . .................... . . ... . 285

R. F. Nicosia

Nitric Oxide and Angiogenesis .. . .. . .... .. ... . . .. .. . .. . ... . . . . . ...... . ..... 297
M. Ziche

Nitric Oxide: A Promoter or Suppressor of Angiogenesis ........ . .......... . .... 307

E. Pipili-Synetos, and M. E. Maragoudakis

Hypoxia/Reoxygenation Enhances Tube Formation of Cultured Human

Micro-Vascul~r Endothelial Cells: The Role of Reactive Oxygen Species ... . ... 321
P. I. Lelkes, K. A. Hahn, S. Karmiol, and D. H. Schmidt

Tumor Angiogenesis and Metastasis: The Regulatory Role of Hyaluronan and Its
Degradation Produces ... ......... .. ..... . .... . .......... .... . ...... .. 337
D. West, D. M. Shaw, and M. Joyce

Cytokines and Growth Factors.. Early Gene Expression during Angiogenic Stimuli ... 349
D. BenEzra, S. Yarkoni, G. Maftzir, and H. Loberboum-Galski

The Role of Laminin Peptide SIKVAV in the Revascularization ofIschemic Tissue . .. 355
R. W. Rose, R. C. Morrison, M. G. Magno, J. Mannion, H. K. Kleinman, and
D. S. Grant

Oxidized Lipoproteins Enhance the in Vitro Tube Formation by Endothelial Cells

Cultured on Matrigel .. . .. . .. . . .. . . ..... . . . . . ....... .. .. .. .. ....... ... 367
B. Sundstrom, K. Venkitesnaran, P. Selvaraj, and D. S. Sgoutas

A Peptide from the NCl Domain of the Alpha-3 Chain of Type IV Collagen
Prevents Damage to Basement Membranes by PMN . . ... ... . ... . . . .. ..... . . 377
Z. Ziaie, J.-c. Monboisse, A. Fawzi , G. Bellon, 1. P. Borel, and N. A. Kefalides

ix
 HUMAN PATHOLOGY AND CLINICAL DEVELOPMENTS

Tumor Vascularity: What Does It Tell Us about the Growth and Spread of Cancer? ... 389
N. Weidner

Tumor Vascularity, Hypoxia and Malignant Progression in Solid Neoplasms ..... ... 407
M. Hockel, K. Schlenger, B. Aral, U. Schaffer, and W. Weikel

Scatter Factor as a Mediator of Tumor Angiogenesis: Recent Clinical and

Experimental Studies . .... .. ................................. . ... .... 415
E. Rosen, K. Lamszus, 1. Laterra, P. Polverini, 1. S. Rubin, and I. D. Goldberg

CLINICAL APPLICATIONS

The Vascularization of Experimental and Human Primary Tumors: Comparative

Morphometric and Morphologic Studies ................................. 429
M. Konerding, E. Fait, A. Gaumann, Ch. Dimitropoulou, and W. Malkush

The Evolution of an Anti-Tumor Agent from a Neonatal Pathogen: Preclinical and

Clinical Evidence for Targeting Tumor Neovasculature . .............. ...... 449
C. G. Hellerqvist

Mapping Neovascularization and Antineovascularization Therapy Using MRI ....... 459

M. Neeman, G. Meir, C. Tempel, Y. Schiffenbauer, and R. Abramovitch

Clinical Trials of Angiogenesis-Based Therapies: Overview and New Guiding

Principles . ..... ...................................... .. ............ 475
W. W. Li, V. W. Li, R. Casey, D. Tsakayannis, E. A. Kruger, A. Lee, Y-L Sun,
C. A. Bonar and S. Cornelius

Abstracts of Posters ..................................................... 495

Participants' Photo ...................................... . ............... 558

List of Participants ...................................................... 560

Index ...................... . .. . ...... . ............... " ............... 569

x
INTRODUCTORY COMMENTS

Michael E. Maragoudakis

Department of Pharmacology, Medical School, University of Patras

261 10 Rio Patras, Greece

When we first started this series of NATO/AS I meetings on angiogenesis in 1991,

we were wondering whether enough new data will be available for discussion in two
years time to justify another meeting. Time has shown that we were unjustifiably
pessimistic about the future progress in the field of angiogenesis.
The exponential increase in the number of publications in peer reviewed journals
both in basic and clinical science from the hundred of laboratories around the world that
are involved in studies of various aspects of angiogenesis has increased our understanding
of this important biological process. The complexity of the cascade of angiogenesis has
challenged scientists and attracted wide-spread interest in basic science such as
embryology, physiology, biochemistry, cell and molecular biology, anatomy,
pharmacology etc. This multidisciplinary approach and the use of the powerful new
techniques that have been developed has made the progress in the field impressive
indeed.
The explosion of clinical research in angiogenesis is a result of the realisation that in
many disease states a common underlying pathology is a derangement in angiogenesis.
Practically every medical speciality deals with disease states where angiogenesis is
involved. Particularly interested are oncologists, ophthalmologists, rheumatologists and
pediatricians. The concept of angiogenesis - based therapy is well founded and calls for
the development of both suppressors and inhibitors of angiogenesis. Suppressors of
angiogenesis have potential clinical applications in situations where abnormal proliferation
of blood vessels is related to disease progression e.g. solid tumors, hemangiomas, diabetic
retinopathy, inflammatory disease etc. On the other hand promoters of angiogenesis are
clinically useful in conditions resulting from insufficient blood supply e.g. wound
healing, coronary artery disease, stroke, organ grafts, diabetic leg ulcers, some types of
hair loss, certain types of infertility etc.
There are visionaries who believe that angiogenesis-based therapy may become for
the next century what antibiotics-based therapy has been for the 20th century. As many
infections diseases, seemingly unrelated, were successfully treated with antibiotics. The
same way angiogenesis inhibitors may be effective in a variety of unrelated diseases,
which have common pathology. For example, significant contributions have been made
the past 3-4 years in early clinical development of suppressors of angiogenesis as
antitumor agents. In addition natural angiogenesis inhibitors and synthetic drugs with

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, J998
 different mechanisms of action have been discovered in recent years, which suppress
specific steps in tumor angiogenesis. These new data provide direct evidence that
tumors area indeed angiogenesis dependent.
The enormous economic potential in angiogenesis-based therapeutics justifies the
keen interest of many pharmaceutical and biotechnology firms, which are represented in
this meeting. The officers of the Angiogenesis Foundation will bring us up to date with
the results of clinical trials in progress. At least 10 such clinical studies are in Phase I, II
& III.
There is a strong interest in the use of vascular density of various solid tumors,
such as breast, lung, prostate etc., as an independent factor for prognosis and guide to
therapy. Dr. Weidner and others will cover this important topic and attempt to answer
important questions: Is the vascular density always a good prognostic indicator? Is it
of value for all solid tumors? What are the methodological problems?
The initial difficulties of the 70's in characterising angiogenic factors were rapidly
resolved with the discovery and molecular sequencing of endothelial angiogenic factors
and endogenous angiogenesis inhibitors as well as other molecules of the extracellular
matrix, which are involved in the modulation of angiogenesis. Now we are faced with the
problem of deciding the relative importance of the various factors in the angiogenesis,
which is taking place in different tissues and under different conditions. This is an
important question that I am hoping will be discussed. Most investigators in the field
believe that there is a reduntance of promoters and suppressors of angiogenesis which
are in balance under physiological conditions. This balance can be tipped off at any
moment and in any tissue by overproduction of an endogenous promoter or
underproduction of an endogenous inhibitors leading to activation of the normally
quiescent angiogenic process.
This meeting will cover various aspects of modulators of angiogenesis and their
receptors, the role of extracellular matrix and adhesion molecules and other cell types
involved in angiogenesis. Dr. Polverini, in the keynote lecture, will give us an overview
of the orchestration of soluble mediators, matrix components and various cell types In
physiological and pathological angiogenesis.
There are fundamental differences in the control, duration and the extent of
angiogenesis under physiological rs pathological conditions. Understanding these
differences will be very important for implementation of this knowledge in the
development of therapeutic agents. A major issue to be resolved is the molecular
mechanisms, which lead to a tight control of the initiation, the progression and especially
the cessation of the angiogenic cascade under physiological conditions. This is evident in
wound healing, the establishment of placenta, the changes in the endometrium etc.
In the past few years angiogenesis research is rapidly moving from descriptive to
mechanistic studies. Our basic knowledge on the mechanism of action of angiogenic
peptides or natural inhibitors is progressing steadily. Studies on the angiogenic genes
and the signal transduction mechanisms have contributed to our understanding of many
aspects of experimental oncology. The previously prevailing view that tumor growth
depends on angiogenesis as the only regulatory mechanism has to be revised in the face
of new data. Today it is recognised that in tumor progression from the premalignant to
advanced tumor stage tumor angiogenesis is controlled by genetic and epigenetic
alterations, where many oncogenes are involved. Angiogenesis can be promoted by loss
of a tumor suppression gene or by overexpression of one or more angiogenic peptides.
There is a progression from a mutual inhibition of endothelial cells, parenchymal cells
and tumor cells, in the dormant tumor stage, to a reciprocal stimulation of angiogenesis
and tumor growth in the advanced tumor. In the pre-malignant stage endothelial cells
secrete cytokines (IL.6 , TGFb, IL. 1), which inhibit the growth of tumor cells. Tumor
cells also secrete anti-angiogenic factors (TSPI, 6D-AIF). At the early and intermediate
tumor stages endothelial cells may be activated by growth factors such as VEGF and

2
TGFa released by the tumor. At later stages angiogenic factors which are produced by
tumor cells, tumor parenchyma and endothelial cells, favor both angiogenesis and tumor
cell growth. Thus a mutual interaction of tumor cells and stroma cells is emerging as an
important mechanism of promotion of angiogenesis and tumor growth. These
interactions through paracrine mechanisms involving production of growth factors,
cytokines, activation of endothelial cells, macrophages and fibroblasts lead to stimulation
of tumor growth.
The discovery of endogenous circulating peptide inhibitors (angiostatin and
endostatin) produced by tumors implies that tumors may remain in a dormant state by
these inhibitors. Both angiostatin and endostatin are capable of inducing tumor
dormancy in animal tumor models. Similar results are obtained with non-peptide
angiogenesis inhibitors. What is exciting in these experiments is that angiosuppression
enhances apoptosis without effect on proliferation of tumor cells. Thus the remodelling
of tumor vasculature in connection with inhibition of angiogenesis can explain the
observed shrinkage of solid tumors after treatment with angiogenesis inhibitors. With
angiostatin experimental tumors shrink to 5mm in size. When treatment is interrupted the
tumor growth resumes to be shrank again on new treatments up to 15 times. These
remarkable results are impossible with conventional chemotherapy or radiotherapy,
because drug resistance develops very quickly.
The role of extracellular matrix and the involvement of the coagulation cascade in
relation to angiogenesis has been the subject of intense investigations in recent years.
There are several presentations in this meeting on the role of thrombin, fibrin peptides
and integrins a v ,) and a v ,5 in angiogenesis. VEGF activates ~,3 while bFGF activates av ,5,
which leads to activation of the U-PAlP AI-I system. Integrin av ,3 also activates
metalloproteinases, which release sequestered b-FGF and angiogenic peptides from the
extracellular matrix.
This new knowledge has led to anti-angiogenic therapies with inhibitors of
metalloproteinases and antibodies to a v ,) integrin.
In our first NATO/AS I in 1991 Dr. Denekamp was making a strong case in favor
of anti-vascular rather anti-angiogenic therapy of cancer. Dr. Thorpe and his group
reported recently a very exciting experiment in this connection. They succeeded in
directing with a specific antibody a truncated form of tissue factor to tumor endothelium.
The resulting infarction in tumor vessels caused complete tumor regression in 38% of
the animals. This validates the concept of anti-vascular therapy of cancer and as expected
the death of tumor endothelium causes an avalanche of death in tumor cells in a ratio of
1-100 or 1-1000.
Up to now we are used to think of angiogenesis in relation to solid tumors only.
More recent evidence from Dr. Folkman's group shows that in bone marrow biopsies
taken from children with acute lymphoblastic leukaemia before therapy, an intense
neovascularization is evident. Microvessel counts were 7-10 times more than in non-
leukaemic specimens. This raises the question if angiogenesis is also involved III
leukaemia.
Another new and exciting development in the field is the cloning of TIE] and TIE2
receptors of endothelial cells and the isolation and cloning of the angiopoietins, the
legends of these receptors. With this we begin to understand the functional role of TIE-
angiopoietin system in the maturation and stabilisation of blood vessels.
The distinction between vasculogenesis and angiogenesis was thought to be clear.
However, in a recent paper it was shown that peripheral blood contains progenitor
endothelial cells that can differentiate into endothelial cells. These cells can be
incorporated into sites of active angiogenesis such as ischemic tissues. These in vivo
results suggest that circulating progenitor endothelial cells may contribute to angiogenesis
in adult species. This is consistent with vasculogenesis, which is thought to be restricted
only to embryogenesis.

3
 From the aforementioned comments the complexity and diversity of angiogenesis
under different conditions is evident by the number of factors, oncogenes, extracellular
matrix and the contribution of many activated cell types (e.g. mesenchymal cells releasing
angiopoietin-l). This meeting presents an opportunity for in depth discussion of many
of these complex aspects of angiogenesis.
I am hoping that as a result of these discussions many collaborations will be
initiated, information, materials and techniques will be exchanged, and differences in
experimental findings can be resolved. Hopefully, at the end we leave with a better
understanding of the present status of angiogenesis and possible future directions.

4
Models and Methods for
Assessing Angiogenesis
MICROVASCULAR ENDOTIlELIAL CELLS FROM ADRENAL MEDULlA - A
MODEL FOR IN VITRO ANGIOGENESIS

Dipak K. Banerjee and Juan A. Martinez

Department of Biochemistry, School of Medicine
University of Puerto Rico, San Juan, PR 00936-5067, USA

INTRODUCTION

Angiogenesis is the development of new and small blood vessels by budding and
sprouting from larger, extant vessels (Beck Jr, and D'Amore, 1997; Bussolino, Montavani,
and Persico, 1997). In adult tissues endothelial cells are quiescent but rapid proliferation
occurs for a limited period of time during menstruation, ovulation, reproduction,
implantation, mammary gland changes during lactation, and wound healing (Cockerill,
Gamble, and Vadas, 1995; Folkman, and Shing, 1992). Abnormal or uncontrolled
angiogenesis has been seen in diabetic retinopathy, arthritis, hemangiomas, psoriasis as well
as for growth and maintenance of many types of benign and malignant tumors (Cockerill et
ai, 1995; Folkman, Watson, Ingber, and Hanahan, 1989; Liotta, Stug, and StetIen-
Stevenson, 1992; Saclarides, Speziale, Drab, Szeluga, and Rubin, 1994; Folkman, 1992).
Inducers of angiogenesis can act directly on endothelial cells, or indirectly, via accessory
cells (monocytes, mastocytes, T cells). Vascular growth factor A (VEGF-A), VEGF-B,
VEGF-C and placental growth factor (PIGF) are angiogenic glycoproteins and display high
amino acid similarity in the platelet-derived growth factor (PDGF) and tumor necrosis
factor a (TNF a) and requires interaction with integrins a)33 ; the other is via VEGF-A and
is integrins avBs - dependent (Friedlander, Brook, Shaffer, Kincaid, Verner, and Cheresh,
1995). It is also becoming evident that there are different classes of endogenous inhibitors
of endothelial cell growth and motility that work in concert with inducer molecules to
control angiogenesis.
Significant progress has been made in our understanding that angiogenesis occurs in
stages: (i) an initiation phase, characterized by increased permeability; (ii) progression,
constituted by the production of proteolytic enzymes that degrade the extracellular matrix
and promote endothelial cell migration, and the entry of cells into either a cell-cycle or an
apoptotic response; (iii) differentiation into new vessels; and (iv) the stabilization and
maturation of vessels by mediator molecules that recruit mesenchymal cells to vessel walls.
But very little is known about the molecular events that trigger the withdrawal of the

7
Figure 1: Light microscopy of the endothelial cells in culture. The cells were cultured in
two-chamber slides. At desired time the cells were washed and mounted with a
coverslip. The pictures were taken with Nomarski interference-contrast on a
Zeiss Ultraphot II inverted microscope. Sequentially the pictures are I-day-old
(a), 2-day-old (b), and 8-day-old (c) taken at a magnification ofx40.

8
Figure 2: The cells were cultured in a plastic flask for 3 weeks. The microscopic picture
was taken on a Zeiss inverted microscope with Hoffman optics.

endothelial cell from the cell cycle, that subsequently regulate their differentiation to form
new vessels, or that finally switch off the process.
To understand the relative balance of inducers and inhibitors that activate the
"angiogenic switch" and the stage specific expression of genes during initiation, progression,
differentiation, and stabilization and maturation of the vessel like structure, a microvascular
endothelial cell line from the vascular bed of bovine adrenal medulla has been developed and
characterized.

CAPILLARY-LIKE STRUCTURE FORMATION IN VITRO

Dissociated cells when cultured in EMEM with Earle's salt supplemented with
glutamine (2 mM), streptomycin (50 ~glml), penicillin (50 units/ml) and 10% fetal bovine
serum (heat-inactivated) at 37°C in 5% CO 2 were frequently seen to be attached to one
another in an end-to-end and side-to-side fashion. In succeeding days they became more
flattened and elongated and exhibited more contact. By eighth day extensive and ordered
networks of cells could be observed which eventually generated macroscopic capillary-like
structures (Fig. 1 & 2; Banerjee, Omberg, Youdim, Heldman, and Pollard, 1985).
Ultrastructural studies indicated that the thin processes extended for up to 50 ~m
and the cells were flattened except at the nucleus. The cytoplasm was filled with rough
endoplasmic reticulum containing a dense matrix, and it became more specialized in the
flattened processes where bundles of actin stress fibers and micotubules predominated.
Numerous surface cisterns of smooth, coated vesicles and small vesicles exhibiting

9
Figure 3: Ultrastructure of 8-day old endothelial cell cultures. (a) Low-magnification
view of putative precapillary lumen (*). The lumens are free of extracellular
matrix (x841O); (b) Extracellular matrix being exocytotically released and
presumably transforming from a nebulous web (lower arrowhead) into parallel
filamentous strands (upper arrowhead) (x54520); (c) Cross-sectional view of
mitochondria typical of these endothelial cells (x9860); (d) Vesicular shuttling
or transcytosis (arrowhead) between a putative precapillary lumen (*) and a
thin intracellular space (x4 11 80); (e) Gap junction (x98020); (f) Intercellular
filaments that predominate at points where cells juxtapose, perhaps
forming precapillary lumens (*) (x2001O).

exocytotic and endocytotic phenomena were also seen to underlie and stud the
plasmalemma. Images consistent with vesicular shuttling ("transcytosis"), the sine qua non
of capillary endothelia (Bruns, and Palade, 1968) were observed in 8-day and older
cultures. Exocytosed matrix component formed a nebulous web that rearranged into the
more fibrillar backbone typical of capillary basal lamina adhering to the plasmalemma.
Dispersed mitochondria of the long tubular type parallel to the cellular axis were frequently
seen. Both tight and gap junctions, intercellular filaments and intercellular spaces
reminiscent of capillary lumen (*) were present. Tight and gap junctions were common in
places where neighboring fine-stranded processes attached the cells to one another. Fine-
stranded filaments (10 nm in diameter), quite distinct and physically separated from
extracellular matrix and basal lamina, fonned a second kind of interaction between these
cells. These fine filaments passed between juxtaposed plasma membranes over distances of
tens of nanometers from separate membrane insertion point (Fig. 3).
Immunocytochemical analysis indicated that a well-developed cytoskeletal network
existed in these cells. The cells were tested positive for a number of cytoskeletal proteins
such as actin, tubulin, myosin, and keratin (Fig. 4).

10
Figure 4: Immunofluorescence of cytoskeletal proteins in endothelial cells. The cells
were grown in two-chamber glass slides. At the end the cells were washed,
permeabilized with digitonin and specific staining were performed for (a) actin,
(b) tubulin, (c) myosine, and (d) keratin.

EXPRESSION OF FACTOR VIII:C GENE IN CAPILLARY ENDOTHELIAL CELLS

Factor VIlI:C is a cofactor for Factor IX. dependent activation of Factor X in the
blood coagulation cascade and is distinctly different from the von Willebrand Factor. In
these endothelial cells Factor VIII:C was expressed as a Mr 200,000 dalton heavy chain and
Mr 46,000 dalton light chains but upon secretion, the heavy chain became M r 215,000
dalton and the two chains were held together by a disulfide bridge to give a 270,000 dalton
asparagine-linked (N-linked) glycoprotein. The carbohydrate content of the light chain was
approximately 18% (BaneIjee, Tavarez, and Oliveira, 1994). De ~ synthesis of Factor
VIlI:C in endothelial cells was due to expression of a transcribable gene because
actinomycin D treatment blocked its expression. The cellular as well as the secretory
Factor VIII:C was biologically active as evidenced by the proteolytic cleavage of a
synthetic peptide (S-2222) in COATEST assay. Immunocytochemical analysis with a
monoclonal antibody (IgG1k) to human Factor VIII:C indicated that Factor VIII:C was
located in the perinuclear region of the cell (Fig. 5).

11
 The exact reason for the presence of Factor VIll:C in capillary endothelial cells is
currently unknown. It is proposed here that during wound healing released Factor VIII:C
would accelerate the blood clotting while angiogenesis is in progress. It is also proposed
that Factor VIll:C would activate a protease essential for the endothelial cells to invade
and penetrate the extracellular matrix during angiogenesis.

PRESENCE OF MEMBRANE RECEPTORS AND TRANSPORTERS ON THE

ENDOTHELIAL CELL SURFACE

Evidence obtained from the isolated plasma membranes or the use of intact cells
strongly supported the presence of 13-adrenoreceptors (Das, Mukherjee, and Bane~ee,
1994), insulin receptors and IGF -1 receptors (Brush, Tavarez-Pagan, and Bane~ ee, 1991)
as well as glucose and catecholamine transporters (unpublished) on the surface of these
capillary endothelial cells.
eH]Dihydroalprenol (eH]DHA) binding to the isolated plasma membranes
demonstrated the presence of 13-adrenoreceptors with two different affinities with an
approximate distributions of 7% and 93%. The binding was specific, saturable and
reversible. The dissociation constants (Kd) and the corresponding binding capacities (Bmax)
for these receptors, however differed by one order of magnitude. Inhibition of eH]DHA
binding by isoproterenol, epinephrine, and norepinephrine suggested that the majority of
these receptors were of (3) subtype or its subclass. When the inhibition of e]DHA binding
was examined in the presence of a series of adrenergic agonists at concentrations ranging
from 10-8 to 1O-3M, the Ki values were found to be isoproterenol (0.56 ± 0.19 x 1O- 9M) <

Figure 5: Immunofluorescence of Factor VIll:C in the endothelial cells. The cells were
cultured in two-chamber glass slides. At the end the cells were washed, fixed
and stained for Factor VIII:C with a mouse monoclonal antibody to human
Factor VIll:C.

12
epinephrine (0.77 ± 0.26 x 10- 9M) > norepinephrine (0.71 ± 0.24 x 1O- 9M) for the high
affinity site, an order expected for a B,-subtype receptor. The corresponding values for the
low affinity site were 4.62 ± 0.64 x 1O-9M, 6.21 ± 0.86 x 1O- 9M and 5.90 ± 0.82 x 1O- 9M,
respectively for the same agonists. This in conjunction with the Scatchard analysis
suggested that the majority of these receptors are indeed of the B,-subtype. The B,-
selective antagonist atenolol was 3 times more effective than the Brantagonist ICI 118,551
in competing with eH]DHA for binding [IC50cor (= K j): atenolol (0.08 ± 0.03 x 10-'2M) <
ICI 118,551 (0.25 ± 0.08 x 1O- 12M) for high affinity sites and inhibited approximately 35%
of total eH]DHA binding at 1O-1~ concentration. It was believed that atenolol interacted
exclusively with the putative B,-adrenoreceptor but from our experiments the possibility of
minor interaction with the Bradrenoreceptor can not be ruled out. It is worth noting that
eH]DHA is a least selective radioligand (Neve, Mcgoningl, and Molinoff, 1986) which
labeled both B, and B2-adrenoreceptors with equal facility (Limberd, 1986). Nevertheless,
these results along with the total ligand binding data supported the presence of a relatively
high concentration of B,-adrenoreceptor and a low concentration of Bradrenorecptor on
these cells. Coexistence of B-adrenoreceptor subtypes had also been described in brain
capillary endothelial cells (Durieu-Trautmnn, Foignant, Strosberg, and Couraud, 1991).
Initial physiologic responses to B-adrenergic receptor stimulation is usually
followed by an increase in adenylte cyclase activity, resulting an increase in adenylate
cyclase concentration (Levitzki, 1986). Our study (Das et ai, 1994) indicated that
isoproterenol at 1O- 7M increased intracellular cAMP concentrations by 38% when
administered for only 30 min and as measured by the radiobinding assay (cAMP assay kit
of Amersham). This clearly established that these capillary endothelial cells contain B-
adrenoreceptors which are coupled to the adenylate cyclase system of the cell.
Furthermore, in analogous experiments, 114-236% increases in intracellular cAMP was also
observed in cells treated with cholera toxin (B-subunit; 100 ng/ml) and forskolin (1 x 10-
6M).
' 25 I-insulin (10 fmols) binding plus intranalization (BI) to these capillary endothelial
cells reached a steady state after 20 min. Acid-washed fraction accounted for nearly half
of the total specifically-bound hormone. Dissociation constants (Kd) for insulin-surface
receptor in acid-extractable fraction were 0.04 nM (high affinity) and 4.7 nM (low affinity)
with a total number of 210,000 high affinity receptors per cell. The time sequence of
binding plus internalization for insulin indicated that 64% of insulin remained specifically
bound to its receptor even after 1 hr. Out of the 36% internalized insulin 92% was found
to be specifically bound and is consistent with the idea that the internalization was
accomplaished only via its receptor. The estimated affinity constant of insulin was within
the range reported for other cell types but the receptor number (i.e., 210,000 per cell) was
more than four-fold greater than that estimated for the rat adipocytes (Gammeltoft, and
Gliemann, 1973). The rapidity with which insulin reaches equilibrium in being bound and
internalized, and the presence of a large number of surface receptors are consistent with the
hypothesis (Carson, Peterson, Moynahan, and Shepro, 1983) that the endothelial cells
perform a through transfer function for insulin with little retention intracellularly. When
125I-labeled IGF-l (15 fmols) was incubated, a transient plateau of binding and
internalization was reached between 6 and 15 min and then continued to increase for 2 hr
(Brush et ai, 1991). This would be consistent with its transfer through the cell as for
utilization in growth processes (Bar, Siddle, Dolash, Bose, and Dake, 1988). This study
also indicated that there was cross-reactivity of insulin with the IGF-l receptor similar to
that reported by others ( King, Goodman, Buzney, Mosses, and Kahn, 1985; Bar, Hoak,

l3
and Pecock, 1978; Bar et ai, 1988). It is probable, however, that there are some differences
in their respective dispositions after combination of each hormone with its own receptor.
The physiologic significance of insulin and IGF-l receptors in these particular
endothelial cells is not yet fully understood. However, glucose transport studies described
with these cells by Brush et al (1991) indicated that the effect of insulin and IGF-l are
complex. Inability ofIGF-l to stimulate glucose uptake in these cells as opposed to that
observed with insulin clearly established the presence of homologous receptors in these
cells. Furthermore, the differential glucose transport capability by insulin and IGF -1 in
these endothelial cells made these capillary endothelial cells physiologically unique from
other microvessel endothelial cells such as one from the bovine adipose tissue where
glucose uptake was stimulated by both hormones (Bar et ai, 1988).

METABOLIC AND ENVIRONMENTAL REGULATIONS OF ENDOTHELIAL CELL

GROWTH, PROLIFERATION AND DIFFERENTIATION

Endothelial cells described here are dependent on serum for their growth and
proliferation. Cells are normally maintained at 10% fetal bovine serum (heat-inactvated)
but 2% serum would be enough. These cells when cultured in EMEM containing 10% fetal
bovine serum in the presence of 5% CO 2 exhibited a cell doubling time of 68 hours and
fusiform appearence upon attachment. Coincident with this doubling in cell number, an
e
increase in 4C]thymidine and [14C]leucine consistent with the cell division and protein
synthesis needed for growth were also observed (BaneIjee et ai, 1985). A dramatic shift in
the morphology however, was observed when a parallely dispersed cells were cultured in
the absence of CO2 . The cells failed to attach the culture dish, did not proliferate and died
within 24 hours. Addition of 10 mM HEPES-NaHC0 3, pH 7.4 improved the cells
attachment to a some degree but the growth and morphology remained impaired. In
HEPES-NaHC03 supplemented media, cells proliferated slowly and exhibited a minimum
cell-to-cell contact (BaneIjee, 1988). The exact role played by the environmental CO 2 in
culturing eukaryotic cells is unclear but it is believed that a subtle decline in pH at the
contact point between the cell surface and the substratum facilitates the cells attachment to
the culture dish and increases its survivality.
A series of agents, such as heparin, thrombin, thyroxine, glucagon, insulin and
phorbol myristate acetate (PMA) were then tested for there effects on endothelial cell
growth and proliferation. Many of these are constituents of normal plasma as well as the
vital hormones of metabolism. Serum concentration of 2% was maintained in these studies.
Only heparin (5 units), thrombin (0.07 units), glucagon (0.10 f,lglml) and thyroxine (0.01
f,lglml) reduced the cell doubling time by 15-22 hours with no morphological abnormality
(Oliveira, and BaneIjee, 1990) satisfying their role as growth factors (Maciag, Hoover,
Stermermann, and Wei stein, 1981). In addition, these reults were consistent with the fact
that heparin reduced the requirement for endothelial cell growth factor (Thronton, Mueller,
and Levine, 1983); thrombin, a serine protease with high specificity for arginyl bonds
stimulated angiogenesis (Maragoudakis, 1996); thyroxine a potentitor of human
development (Baxter, and MacLeod, 1980). Insulin (0.01 f,lgiml) and PMA (20 nglml) on
the other hand, increased the cell doubling time by 17-24 hours (Oliveira, and BaneIjee,
1990) even though it was shown earlier that insulin a vital hormone during mammalian
embryogenesis (Eriksson, Lewis, and Freikel, 1984) and PMA increased secretion of
collagen, plasminogen activator in endothelial cells to make them invade underlying collagen
matrix to form capillary network in bovine microvascular endothelial cells (Montesano, and

14
Orci, 1985). Analysis of the cell cycle indicated that insulin induced S phase but held the
cells in the G, phase for a longer period of time, thereby, reducing the total mitotic activity.
PMA arrested the cells in the G, phase during the entire period of treatment.
13-Agonist isoproterenol as well as cholera toxin (13-subunit), prostaglandin E"
forskolin all enhanced intracelluar cAMP to a significant level and also the proliferation of
capillary endothelial cells to varying degrees. When the functional implication(s) of the
cAMP generating system were studied in these cells, it was observed that the overall
morphology of the cells remained the same but cAMP-related stimuli accelerated the lumen
formation in vitro (Das et ai, 1994).

INTERRELATIONSIDP BElWEEN THE DOLICHOL-CASCADE PATHWAY OF

PROTEIN N- GLYCOSYLATION AND CAPILLARY ENDOTHELIAL CELL
PROLIFERATION

Many inducers of angiogenesis such as VEGF, bFGF, PIGF, vascular cell adhesion
molecule-I, soluble E-selectin and many of their receptors, e.g., integrins, sialyl lewis X as
well as the endothelila cell marker Factor VIII:C are glycoproteins and carry N-linked
glycan chains as a part of their structures. A number of cell adhesive molecules such as
fibronectin, laminin, and thrombospondin were expressed by these endothelial cells
(Banerjee, personal communication). In addition, the expression of Factor VIII:C paralleled
the cell proliferation. Building of Glc3Man9GlcNAc2 oligosaccharide chain on the dolichol
backbone through a pyrophosphate bridge in the endoplasmic reticulum is a prerequisite
for the asparagine residues to be N-glycosylated when present in the consensus sequence
Asn-X-Ser/Thr. We then examined the various steps of the dolichol-cascade pathway after
exposing the endothelial cells to a variety of experimental conditions.
Culturing of cells in an environment deprived of CO 2 resulted in a 32%
(approximately) reduction in protein synthesis but the ratio of [3H]mannose to ['4C]leucine
was increased to 4.3. Analysis of Dol- P-Man synthase, a key glycosyltransferase
(Kornfeld, and Kornfeld, 1985) in the Dol-P pathway suggested that the Km for GDP-
mannose in CO 2 deprived cells was reduced by 32% (approximately) without a significant
change in the Vmax (Banerjee, 1988). Amphomycin, a lipopeptide which binds Dol- P in a
Ca2+-dependent manner and inhibited Dol-P-Man, Dol-P-Glc and Dol-PP-GlcNAc
synthesis (Banerjee, 1989) also inhibited capillary endothelial cell proliferation. At 0.5 g,
amphomycin increased the cell doubling time by 52 hours (Banerjee,and Vendrell-Ramos,
1993). A similar result was also obtained with tunicamycin (Martinez, Torress, and
Banerjee, 1997) a specific inhibitor of GlcNAc-lP transferase (Elbain, 1987). Other
laboratories used N-glycan processing inhibitors I-deoxymannojirimycin and
castanospermine and reached the conclusion that inhibition of hybrid and complex type
oligosaccharides blocked capillary tube formation and tumor growth (Nguyen, Folkman,
and Bischoff, 1992; Pili, Chang, Partis, Muller, Chrest, and Passaniti, 1995).

CONCLUSION

Increased cell proliferation by a cAMP-related stimuli, enhanced protein

glycosylation (120-220%) by a 13-agonist isoproterenol and reduction of cell proliferation
by a glycosylation inhibitor suggested not only the protein N-glycosylation and endothelial
cell proliferation are interlinked but also cAMP played a major role in linking these two

15
processes. Isoproterenol manifested its effect through 13-adrenoreceptors because atenolol
(a 13 1-antagonist), ICI 118,551 (a 132-antagonist) as well as propranolol (an antagonist which
binds 13 1- and 132 -receptors with equal potencies) all inhibited isoproterenol mediated
protein glycosylation. Earlier we had demonstrated that 13-adrenoreceptor stimultion
enhanced Glc3Man9GlcNAcrPP-Dol biosynthesis and turnover and subsequently the
protein N-glycosylation in eukaryotic cells through intracellular cAMP following protein
phosphorylation of Dol-P-Man synthase by a cAMP-dependent protein kinase
(Kousvelari, Grant, and Baum, 1983; Baneljee, Kousvelari, and Baum, 1987). It may
therfeore be concluded that cAMP plays an important role during angiogenesis. A cell
synchronization protocol has now been developed which will allow us to address the
geneti c basis of the "angi ogeni c switch".

ACKNOWLEDGEMENT

The work described here supported in part by the United States Public Health
Service Grants ROI HL35011 and S06RR08224 as well as the by CIDIC funds from the
Medical Sciences Campus.

REFERENCES

Banerjee DK. Amophomycin inhibits mannosylphosphoryldolichol synthesis by forming

a complex with dolichylmonophosphate. J Bioi Chern 264:2024-2028, 1989.

Baneljee DK. Microenvironment of endothelial cell growth and regulation of protein N-

glycosylation. Indian J Biochem Biophys 25:8-13, 1988.

Baneljee DK, Kousvelari EE, Baum B1. cAMP-mediated protein phosphorylation of

microsomal membranes increases mannosylphosphodolichol synthase activity. Proc
Nat! Acad Sci (USA) 84:6389-6393, 1989.

Baneljee DK, Ornberg RL, Youdim MBH, Heldman E and Pollard HB. Endothelial cells
from bovine adrenal medulla develop capillary-like growth patterns in culture. Proc Nat!
Acad Sci (USA) 82:4703-4706, 1985.

Banerjee DK, Tavarez n, Oliveira CM. Expression of blood clotting Factor VIII:C gene in
capillary endothelial cells. FEBS Letts 306:33-37, 1992.

Baneljee DK, Vendrell-Ramos M. Is asparagine-linked protein glycosylation an obligatory

requirement for angiogenesis? Indian J Biochem Biophys 30:389-394, 1993.

Bar RS, Hoak JC, Peacock ML. Insulin receptors in human endothelial cells:Identification
and characterization. J Clin Endocrinol Metab 47:699-702, 1978.

Bar RS, Siddle K, Dolash S, Boes M, Dake B. Actions of insulin and insulin like growth
factors I and II in cultured microvessel endothelial cells from bovine adipose tissue.
Metabolism 37:714-720,1988.

Baxter JD, Macloed KM. In Metabolic Control and Disease. PK Bondy and IN
Rosenberg, eds. WB Saunders, Philadelphia, pp 140, 1980.

16
Beck Jr L, D'Amore P A. Vascular development Cellular and molecular regulation. F ASEB
J 11:365-373 1997.

Bruns RR, Palade GE. Studies on blood capillaries. II.Transport of ferritin molecules
across the wall of muscle capillaries. J Cell Bioi 37:277-299, 1968.

Brush JS, Tavarez-Pagan JJ, BaneIjee DK. Insulin and IGF-1 manifest differential effects
in a clonal capillary endothelial cell line. Biochem Intl25:537-545, 1991.

Bussolino F, Montavani A, Persico G. Molecular mechanisms of blood vessel formation.

TIBS 22:251- 256,1997.

Carson MP, Peterson SW, Moynahan ME, Shepro D. In Vitro 19:833-840,1983.

Cockerill GW, Gamble JR, Vadas MA. Angiogenesis:Model and modulation. Int Rev
Cytol 159:113-159, 1995.

Das SK, Mukherjee S, BaneIjee DK. u-Adrenoreceptors of multiple affinities in a clonal

capillary endothelial cell line and its functional implication. Molec Cellular Biochem
140:49-54, 1994.

Durieu-Trautmann 0, Foignant N, Strosberg AD, Couraud 00. Coexpression of Ul- and

uradrenergic receptors on bovine brain capillary endothelial cells in culture. J Neurochem
56:775-781,1991.

Elbain AD. Inhibitors of the biosynthesis and processing of N-linked oligosaccharide

chains. Annu Rev Biochem 56:497-534, 1987.

Eriksson 0, Lewis N, FreikeI N. Growth reterdation during early organogenesis III

embryos of experimentally diabetic rats. Diabetes 33:281-284, 1984.

Folkman 1. The role of angiogenesis in tumor growth. Seminar in Cancer Bioi 3:65-71,
1992.

Folkman J, Shing Y. Angiogenesis. J Bioi Chern 267: 10931-10934, 1992.

Folkman J, Watson K, Ingber D, Hanahan D. Induction of angiogenesis during the

transition from hyperplasia to neoplasia. Nature 339:58-61, 1989.

Friedlander M, Brook PC, Shaffer RW, Kincaid CM, Verner JA, Cheresh DA. Definition
of two angiogenic pathways by distant v integrins. Science 270:1500-1502,1995.

GammeItoft S, Gliemann 1. Binding and degradation of 125I_labelled insulin by isolated rat

fat cells. Biochim Biophys Acta 320: 16-32, 1973.

King GK, Goodman AD, Buzney S, Mosses A, Kahn CR. Receptors and growth
promoting effects of insulin and insulin like growth factors on cells from bovine retinal
capillaries and aorta. J Clin Invest 75:1028-1036, 1985.

Kornfeld R, Kornfeld S. Assembly of asparagine-linked oligosaccharides. Annu Rev

Biochem 54:631-664, 1985.

17
Kousvelari EE, Grant SR,Baum BJ. -Adrenergic receptor regulation of N-linked protein
glycosylation in rat parotid acinar cells. Proc Nat! Acad Sci (USA) 80:7146-7150, 1983.

Levitzki A. a-Adrenergic receptors and their mode of coupling to adeny1ate cyclase.

Physiol Rev 66: 819-854, 1986.

Limbard LE. Cell surface receptors:A short course on theory and methods. Martinus
Publishing, Boston, pp 75-115, 1986.

Liotta LA, Stug PS, Stet!en-Stevenson WG. Cancer metastasis and angiogenesis: An
imbalance of positive and negative regulation. Cell 64:327-336, 1992.

Maciag T, Hoover GA, Stermerman MB, Weistein R. Serial propagation of human

endothelial cells in vitro. J Cell Bioi 91:420-426, 1981.

Maragoudakis ME. The role of thrombin in angiogenesis. In Molecular, Cellular, and

Clinical Aspects of Angiogenesis (ed. Michael E. Maragoudakis), pp 95-103, 1996.
Plenum Press, New York.

Martinez JA, Torres IN, Banerjee DK. Expression of Glc3Man9GlcNAc2-PP-Dol is a pre-

requisite for endothelial cell growth and proliferation. FASEB J 11(9):A1385, 1997.

Montesano R, Orci L. Tumor promoting phorbol esters induce angiogenesis in vitro. Cell
42:469-477,1985.

Neve KA, Mcgonigl P, Molinoff PB. Quantitation analysis of the selectivity of

radioligands for subtypes of a-adrenergic receptors. J Pharmacol Exp Ther 238:46-53,
1986.

Nguyen M, Folkman J, Bischoff J. I-Deoxymannojirimycin inhibits capillary tube

formation in vitro. J Bioi Chern 267:26157-26165, 1992.

Oliveira CM, Banerjee DK. Role of extracellular signaling on endothelial cell proliferation
and protein N-glycosylation. J Cellular Physiol 144:467-472, 1990.

Pili R, Chang J, Partis RA, Muller RA, Chrest FJ, Passaniti A. The a-glucosidase I
inhibitor castanospermine alters endothelial cell glycosylation prevents angiogenesis, and
inhibits tumor growth. Cancer Res 55:2920-2926, 1995.

Thronton SC, Mueller SN, Levine EM. Human endothelial cells:Use of heparin in cloning
and long-term serial cultivation. Science 222:623-625, 1983.

Saclarides TJ, Speziale NJ, Drab E, Szeluga DJ, Rubin DB. Tumor angiogenesis and rectal
carcinoma. Dis Colon Rectum 37:921-926, 1994.

18
ANGIOGENESIS IN THE HEART AND SKELETAL MUSCLE - MODELS FOR
CAPILLARY GROWTH

O. Hudlicka 1, S. Egginton 1 and M.D. Brown2

I Department of Physiology, University of Birmingham, Birmingham

B15 2TT, UK.
2 Department of Sport and Exercise Sciences, University of Birmingham,
Birmingham B15 2TT, UK.

INTRODUCTION

Growth of vessels under physiological circumstances in adults appears only in female

reproductive organs, but can be elicited by endurance training in skeletal and sometimes in
cardiac muscle (for review see Hudlicka et 01, 1992; Hud1icka and Brown, 1996).
However, to achieve an increase in capillary supply of 20% in skeletal muscles takes 5-8
weeks of very intensive endurance training in man (Andersen and Henriksson, 1977), and
usually about 12 weeks in animals such as rat (see Gute et 01., 1994) Endurance training is
obviously linked with a long-term increase in activity of skeletal muscles and hence
overload resulting sometimes but not always, in muscle hypertrophy. It also results in an
increase in the capacity of the whole vascular bed (Snell et aI., 1987) due to high blood
flow in contracting muscles.
In the heart, one of the accompanying signs of endurance training is a decrease in resting
heart rate (bradycardia) and, of course, higher coronary blood flow during exertion as well
as increased stroke volume and force of contraction. As with skeletal muscle, these changes
may take weeks or even months to develop.
We have used a model of pronounced increase in activity to a limited group of skeletal
muscles, originally designed to alter their contractile properties (Salmons and Vrbova,
1969), to stimulate capillary growth. To induce growth of capillaries in the heart we
utilised the findings of Wachtlova et al. (1965, 1967) on the high capillarization in hearts of
'athletic' animals, such as hare or wild rat, in comparison with related sedentary species
such as rabbit and laboratory rat, and the knowledge that the heart rate in athletic animals is
considerably lower than in the sedentary types.
Several indices were used to assess capillary growth. The total number of capillary profiles
relative to skeletal or cardiac muscle cross sectional area was identified either in low-power
electron micrographs or in sections where capillary endothelium was stained for alkaline
phosphatase activity. Using the indoxyl tetrazolium method, capillaries were depicted as
dark reaction product while endothelium in arterioles or venules was not stained (Ziada et

Angiogenesis: Models. Modulators. and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 19
 Alkaline phosphatase

mean !SE.
FREQUENCY % 2332 :67
30 n : 31

20

10

1000 2000 3000

2
CAPS/nvn
Toluidine Blue

mean! S.E.

FREQUENCY % 2239 ! 42
n : 31

30

20

10

1000 2000 3000

CAPS/mm 2
, : 0.794 y : l.268x • - 506.8

Fig. 1
Comparison of the estimation of capillary density in rabbit papillary muscle using
histochemical staining for alkaline phosphatase in one half, and low power electron
microscopy in toluidine blue stained semithin (O.SJlm) sections in the other half, gave very
similar results with both methods.

ai., 1984). As this staining does not work in all tissues or species (e.g. not in human or pig
skeletal muscles or hearts, or in rabbit brain tissue) other histochemical methods, such as
staining for the lectin Grif.fonia simplicifolia I have also been used. Since EM or
histochemical stains depict all capillaries present in the tissue, whether they are perfused or
not, any increase in capillary density or capillary/myocyte ratio is a clear indication of
capillary growth. This has been confirmed by comparisons of EM with alkaline
phosphatase staining in rabbit papillary muscle (Brown and Egginton, 1988) (Fig 1).
Capillary growth can also be demonstrated by increased incorporation of either 3H-
thymidine or bromodeoxyuridine (BrdU) into capillary endothelial cell nuclei. Capillary
endothelial cells in normal skeletal or cardiac muscle have a very low turnover (Denekamp
et ai., 1984) and an increased labelling index is a clear indication of capillary growth.
However, proliferation is not a condition sine qua non: Sholley et ai. (1984) demonstrated
capillary sprouts in the absence of endothelial cell proliferation. Quite clearly, the presence
of capillary sprouts visualized by vascular casts (Dawson and Hudlicka, 1989), electron
microscopy (Schoefl, 1963; Cliff, 1963) or intravital observations (Clark, 1918; Myrhage
and Hudlicka, 1978) offers a definitive proof of capillary growth, although recently Burri
and Tarek (1990) demonstrated capillary growth without sprouting during development
called 'intussusceptive growth' where capillaries are divided by ingrowing columns of
connective tissue (Fig 2).

20
In assessment of capillary growth in skeletal muscles we used all the methods referred to
above and in the heart assessment was predominantly by the total number of capillaries
related to the tissue area.
The models used to induce angiogenesis in adult skeletal and cardiac muscle are shown in
Table 1.

1. SKELETAL MUSCLE

1.1. Long-term increase in muscle activity

Skeletal muscles involved in the maintenance of posture, such as lower limb postural
muscles or deep back muscles in man, or quadriceps or soleus muscles in most mammals,
are slowly contracting and composed of highly oxidative and highly vascularised muscle
fibres. In contrast, muscles used for rapid movements, such as those involved in the flexion
of the ankles, are fast contracting with a relatively high proportion of muscle fibres with
high glycolytic metabolism and a more sparse capillary supply (Romanul, 1965). Nerves
supplying the postural muscles have a constant relatively low frequency of action
potentials (about 10 Hz) most of the time, i.e. whether the animal is sitting or moving,
while those supplying fast contracting muscles show bursts of high frequency (60-100 Hz)
appearing only during movement (Adrian and Bronk, 1929) (Fig 3B). Stimulation of fast
contracting muscles at a frequency appearing in nerves supplying postural muscles induced
capillary growth in rabbits within 2-4 days (Brown et aI., 1976; Hudlicka et aI., 1982), and
in rats within 4 days, with an increase of -40-50% within 7 days. The stimulation was
performed by implantation of the electrodes in the vicinity of the nerve supplying ankle

A.

0 ) }} ~.
,/" ~\
,, ''
,,, '\
, ~~
~)
\" ...... _ - - /
/~ ) //- ---~

1. 2. 3. 4. 5. 6.

=w
B.

®
~ :JXC
1. 2.

A · SPROUTING ··········BM, * PROLIFERATING CELLS

* 3. 4.

B . INTUSSUSCEPTIVE
Fig. 2
The two most common types of capillary growth.
A an angiogenic stimulus to quiescent endothelial cells (1) leads to breakage of the basement
membrane (2), allowing the endothelial cells to migrate outside the parent vessel (3 and 4),
with proliferation of the cells closest to the capillary (5) and formation of a sprout (6).
B intussusceptive growth: extracellular mesenchymal tissue partially (2), and later fully (3),
occludes the existing capillary which will be eventually divided into two separate vessels
(4).

21
 stimulator

peroneal nerve

10 Hz

TA. EOl - SOLEUS

Fig. 3
A action potentials recorded in nerves supplying fast (TA and EDL) and slow (soleus)
muscles.
B schematic drawing of the stimulation regime for fast (T A and EDL) contracting muscles.

flexors - tibialis anterior (TA), extensor digitorum longus (EDL) and extensor hallucis
proprius (EHP) (Fig 3A) at 10 Hz, short pulse duration (O.3msec) and voltage resulting in
palpable strong muscle contraction, usually for 8 hours/day, a regime which does not cause
the animals any discomfort. Stimulation with other frequency patterns, such as
intermittent stimulation at 40 Hz, also elicited capillary growth, provided the total number
of stimuli was similar to that at 10Hz although the duration required was longer (Tyler and
Hudlicka, 1984). The low frequency (10 Hz) stimulation resulted not only in an increase in
the total number of capillaries estimated by histochemical methods (Brown et aI., 1976;
Myrhage and Hudlicka, 1978) or electron microscopy (Hudlicka et al., 1987), but also by
increased labelling of capillary linked nuclei with BrdU (pearce et aI., 1995; Fig 4), the
presence of capillary sprouts demonstrated by intravital observations (Myrhage and
Hudlicka, 1978), electron microscopy (Hansen-Smith et aI., 1996; Zhou et al., 1997) and
vascular corrosion casts (Dawson and Hudlicka, 1989). It also resulted in the growth of the

1400 100
l 1200
0
:;-
::; 80 n
!: 1000 (j)
Q) 60 II>
v.
II) 800 C!>
"'t;

~
Q)
600 40 5'
0
.E :.;,
•
400 20
;200 ~

2d 7d 14d 28d 3mo

Length of muscte stimutatlon

Fig. 4
Changes in labelling index (LI) for bromodeoxyuridine and capillary/fibre ratio in
chronically stimulated muscles as a function of time.

22
 20 Saline injected
elF ~ 1·20±0·04
15 (mean±S.E.M.)
n ~ 87
10

15
""c
0

10
0
:;
.D 5
";:
<f)

"0
> Xanthine injected
u
c 10 elF ~ 1·69±0·04
Q)
::J n ~ 100
0- 5
~
LL

25
Prazosin drinking
20 elF ~ 1· 78±0·03

15

10

1·0 1·5 2·0 2·5 3·0

elF ratio

Fig. 5
Changes in capillary/fibre ratio in rat skeletal muscles, in animals treated with saline
(control) and three different vasodilators: dipyridamole (persanthine), a xanthine derivative
(propentofylline), and prazosin.

whole vascular bed demonstrated by cast weight (Dawson and Hudlicka, 1989), increased
number of arterioles (Hansen-Smith el af., 1995; 1997) and increased maximal conductance
(i.e. maximal blood flow divided by mean arterial pressure which is a good estimate of a
total capacity of the vascular bed) (Brown and Hudlicka, 1995).
The obvious question to ask is what mechanism is responsible for capillary growth induced
in chronically stimulated muscles. One possibility is local hypoxia since the
predominantly glycolytic fibres with a low content of mitochondria are exposed to activity
with a high energy demand. Chronic stimulation increased significantly the mitochondrial
content (Hoppeler et aI., 1987) but this might again be due to hypoxia. However, the
direct measurement of oxygen tension in chronically stimulated muscles did not show
values lower than those in controls (Hudlicka el al., 1984). In addition, when blood supply
to stimulated muscles was limited by ligation of the common iliac artery thereby making
muscles hypoxic, capillary growth was attenuated (Hudlicka, 1991). Thus, increased blood
flow during muscle contractions seemed to be an important stimulus for capillary growth,
and we tested this hypothesis by increasing blood flow without an increase in muscle
activity.

23
1.2. Long-term increase in blood flow

Tornling et al. (1980) demonstrated increased labelling of capillaries with 3H thymidine in

skeletal muscles of animals treated with dipyridamole - a drug which inhibits the removal of
adenosine, a well known dilator in skeletal and cardiac muscle, by red blood cells. As
dipyridamole does not induce proliferation of endothelial cells in tissue cultures (Jakob et
aI., 1982), the assumption is that its effects would be solely due to an increased blood
flow. If blood flow is an important stimulus for capillary growth, other vasodilators
should be similarly effective. Fig 5 demonstrates a comparison of changes in capillary
supply, expressed as capillary/muscle fibre ratio in animals which received for 5 weeks
either saline, dipyridamole, torbafylline - a xanthine derivative which increased blood flow
in skeletal and cardiac muscle (Hudlicka et al.,1981) or prazosin - an a.l blocker which is
used for treatment of hypertension since it produces peripheral dilatation and increases
capillary flow (Dawson and Hudlicka, 1989). Since prazosin was the most effective drug,
we used it in further experiments to establish whether a shorter administration would also
induce capillary growth. Two and even one week of oral administration was enough to
stimulate capillary growth demonstrated either by histochemical staining for all
anatomically present capillaries (Brown et al., 1996) or by electron microscopy (Zhou et
aI., 1996). However, in this case capillary growth appeared without endothelial cell
proliferation (Brown et aI., 1996) or abluminal sprouting (Zhou et aI., 1996). Prazosin also
stimulated growth of arterioles, presumably by arteriolarisation of capillaries as described
by Price and Skalak (1996).
We therefore assumed that the common denominator for the induction of capillary growth
in chronically stimulated and prazosin treated muscles are haemodynamic forces acting
predominantly on the luminal side of capillaries. One of the important forces is shear
stress which stimulates endothelial cell proliferation in tissue cultures (Ando et al., 1987).
Measurement of velocity of red blood cells (Vrbc) in capillaries and their diameter enabled
us to calculate shear stress in stimulated and prazosin treated muscles and to show that
shear stress is indeed considerably higher than in control muscles (Dawson and Hudlicka,
1993). Since the diameters of capillaries in stimulated muscles were marginally greater
while those in prazosin treated animals somewhat smaller, and Vrbc was higher in the
former than in the latter, shear stress seems to be a more important factor in prazosin
treated animals. By similar reasoning for the opposite changes in diameters, the calculated
capillary wall tension seemed to be more important in chronically stimulated muscles. In
addition, stimulated muscles are repeatedly contracting and relaxing, thus changing the
configuration of capillaries which are tethered by the extracellular matrix to muscle fibres.
The effect of muscle fibre length thus could exert an influence on capillaries acting from the
abluminal side. If so, it should be possible to elicit capillary growth by modification of the
muscle fibre length even without a concomitant increase in blood flow and shear stress.

1.3. Long-term muscle stretch

Muscle overload resulting from stretch can easily be induced by removal of muscles with a
similar function - muscle agonist. For instance in the case of EDL removal of TA would
result in muscle hypertrophy and stretch (Frischknecht and Vrbova, 1991). We used this
model to investigate changes in capillary supply in relation to blood flow (Egginton and
Hudlicka, 1992). After two weeks, capillary/fibre was increased to a similar extent as in
one week chronically stimulated muscles while blood flow was not appreciable higher than
in control muscles (Fig 6). Sarcomere length was increased by about 20% and capillary
growth was also demonstrated by the presence of abluminal sprouts using electron
microscopy (Zhou and Egginton, 1997).

24
 --
ClF
• • D control
2 prazOSin

~ stImulation

stretch

• p < 0.05 vs control

o lwk 2wk

Fig6.
Capillary/fibre ratio (CIF) in three different models of angiogenesis in skeletal muscle.

2. THE HEART

In the heart, capillary growth was induced by long-tenn decrease in heart rate, chronic
administration of drugs that increase coronary blood flow or long-tenn administration of
dobutamine at a dose with a relatively little effect on heart rate or blood flow but which
increases cardiac perfonnance on long-tenn administration (measured by cardiac work or
development of tension with time (dP/dt index).

2.1. Long-term bradycardia

Bradycardia has been induced by alinidine, a drug affecting directly the sino-atrial node, or
by electrical bradycardial pacing. Alinidine was administered by intraperitoneal injection
twice daily (6 mglkglday) over a period of 5 weeks. Heart rate was lowered by ~30% and
capillary supply, expressed as capillary/myocyte ratio, increased from 1.31 to 1.54
(p<O.05).
In rabbits (Wright and Hudlicka, 1981) and pigs (Brown et ai., 1994) heart rate was
decreased by electrical bradycardial pacing. This was done by eliminating every second
heart beat in rabbits (Fig 7) and in pigs by decreasing heart rate by ~35% using dual
chamber pacemakers with two electrodes sensing the atrial rhythm and adjusting the atrio-
ventricular delay to achieve prolongation of the thus decrease in the interval between
ventricular systoles. The increase in capillary supply varied directly with the magnitude of
the decrease in heart rate (Fig 8).
Since bradycardia prolongs the duration of diastole more than the duration of systole, and
since coronary blood flow is very limited during systole because of the high pressure
developed in the left ventricle, we assumed that one of the factors which might stimulate
capillary growth would be higher flow. Blood flow per beat was indeed increased in
bradycardia (Hudlicka et aI., 1989) and we thus wanted to establish whether increased
blood flow per se could stimulate capillary growth.

25
 100

mmHg a[ LVP

1 s

Fig. 7
Bradycardial pacing in rabbit. Pacing electrodes (connected to a portable stimulator which
is not shown) are implanted into the right atrium . ECG record and record of the left
ventricular pressure (L VP) demonstrate how bradycardia is achieved by imposing an
electrical impulse (arrows) into the heart to eliminate every second heart beat.

2.2. Long-term increase in blood flow

When two different vasodilators, adenosine and a xanthine derivative propentofylline, were
administered by intravenous infusion using portable infusion pumps over a period of 3-5
weeks in rabbits, capillary density was increased by -30 and -20% respectively compared
to animals infused with saline (Ziada et aI., 1984). Infusion of prazosin, which did not
increase coronary blood flow, did not lead to capillary growth, whereas infusion of
adenosine did (Fig 9). Capillary growth was also described in rat hearts after chronic
administration of dipyridamole (Tornling, 1982) and ethanol (Mall et al., 1982). Thus
increased blood blow can stimulate capillary growth, possibly by the same mechanism as in
skeletal muscle, i.e. by increased shear stress. Using the data on the Vrbc and capillary

26
 RABBIT

Degr.e 0' bradycardia

Hearl rate
decreased by 28% 34% 45%

Fig. 8
Relative increase in capillary density (compared to control animals with a similar heart
weight) and percentage decrease in heart rate in different species.

diameters obtain by Tillmans et af. (182) in rats treated with dipyridamole, the calculated
shear stress was indeed higher. However, the same calculation in bradycardial paced hearts
did not indicate any substantial elevation of shear stress,because the diameters of capillaries
are substantially greater during diastole (Tillmans et al., 1974) and the increase in the
velocity of flow is not high enough to compensate for it. However, the wider capillaries
accommodating greater blood flow are exposed to higher pressure, and this increases
capillary wall tension (Hudlicka, 1994) and might lead to changes in the basement
membrane and extracellular matrix thus activating growth factors. In agreement with this
assumption is the good correlation between capillary density and levels of endothelial cells
stimulating angiogenic factors (ESAF) in control and bradycardially paced rabbit and pig
hearts (Hudlicka et al., 1995).

% change from control capilia r lesJmm2

14

12

80

60

40

20

o control
!21 prazosin . * p<O.OOl v.
control
~ adenosine

Fig. 9
Acute changes in coronary blood flow and in capillary density after 5 weeks of
administration of a vasodilator (adenosine), or a drug with very little effect on coronary
blood flow (prazosin).

27
 capillarles/mm 2 mls.mln ·l .g. l J.g.l.mln· l .xl0· 4

12 2.5
*
*
3000

0 control
2.0
~ dobutamlne
2000 8 treated
1.5
6
1.0
1000 4

0.5
2

CAPILLARY CORONARY CARDIAC

SUPPLY BLOOD FLOW MINUTE WORK

Fig. 10
Administration of dobutamine for 2 weeks caused a substantial increase in cardiac work (by
a positive isotropic action) and increased capillary supply without affecting coronary
blood flow.

Prolongation of diastole during long-term bradycardia also means that the filling of the
ventricles is increased and thus the muscle fibers are stretched. This in tum induces a
greater force of contraction and increased stroke volume and stroke work, which actually
precedes capillary growth (Wright and Hudlicka, 1981). Increased force of contraction,a
positive inotropic effect, inevitably influences capillaries tethered to myocytes. Hence,
mechanical distortion of capillaries may be yet another stimulus for growth.

2.3. Long-term increase in cardiac inotropism

Beta receptor agonists are a family of drugs which increases the force of contraction in the
heart, and we selected dobutamine in a dose which would not effect blood pressure (and
thus the load on the heart by changes in the peripheral resistance) to induce capillary
growth (Brown and Hudlicka, 1991). Administration of the drug by intravenous infusion
over a period of two weeks resulted in significantly higher cardiac minute work and
capillary density in the absence of a significant change in coronary blood flow (Fig 10).
Thus changes in the configuration of capillaries due to forces acting from the abluminal side
also represent a stimulus for capillary growth.

3. CAPILLARY GROWTH IN SKELETAL AND CARDIAC MUSCLE WITH

LIMITED BLOOD SUPPLY

Whatever the mechanism of capillary growth, the ultimate aim of these studies was to
induce it in tissues where limited capillary perfusion would impair performance. Cardiac
performance is limited either by pressure or volume overload hypertrophy where capillary

28
 Table 1 Models of angiogenesis.

SKELETAL MUSCLE

Long-tenn increase in activity: chronic electrical stimulation

Long-tenn vasodilatation: administration of adenosine, dipyridamole or prazosin

Long-tenn stretch: overload due to removal of synergistic muscle

HEART

Long-term bradycardia: chronic electrical pacing

Long-tenn vasodilatation: adenosine, dipyridamole

Long-tenn increase in heart work: b agonist dobutamine

growth does not accompany the increasing size of myocytes and the diffusion distances for
oxygen and other nutrients are greater eventually resulting in heart failure. Another
example is, of course, myocardial infarction where some collateral circulation develops but
it is extremely difficult to achieve growth of capillaries. Moreover, the non-infarcted tissue
hypertrophies and the diffusion distances are increased in the same way as in other types
of hypertrophy. Thus stimulation of capillary growth in either case would be extremely
beneficial, and thus far has been achieved only to a very limited extent in pressure overload
heart hypertrophy by long-tenn administration of the Ca2+ channel blocker (and
vasodilator) nifedipine (Turek et aI., 1987).
Using either the vasodilator prazosin (Fulgenzi and Hudlicka, 1994) or a specific regime of
chronic electrical stimulation (Hudlicka et aI., 1994) as a means to increase capillary
growth, it was possible to increase capillary supply in skeletal muscles where blood flow
had been limited by ligation of the common iliac artery. An increase in capillary/muscle
fibre ratio nonnalised the maximal muscle blood flow which had been lowered by ligation to
less than one third of values in nonnal muscles, and improved muscle performance. The
tension developed in contracting muscles with limited blood supply decreased over a
period of five minutes to about 35% compared with a decrease to 70% in nonnal muscles,
and administration ofprazosin or chronic stimulation over the whole period (two weeks)
of the limitation of blood supply prevented this decrease.
In the heart, alinidine increased capillary supply in hypertrophy induced by increased
blood pressure in rats (Hudlicka and Brown, 1991) and bradycardial pacing had a similar
effect in volume overload hypertrophy in rabbits where it was linked with a significant
improvement in perfonnance measured as stroke work (Wright et al., 1989). Bradycardial
pacing in pigs with myocardial infarction also resulted in capillary growth in the border
zone to infarction as well as in the intact tissue and again improved stroke work (Brown et
al., 1995).

CONCLUSIONS

Models used to elicit capillary growth in nonnal adult skeletal and cardiac muscle include
long-tenn administration ofvasodilating drugs. In skeletal muscle a very powerful stimulus

29
is long-tenn increase in actlVlty achieved very effectively by chronic stimulation of a
specific muscle group, or long-tenn muscle overload achieved by stretch due to removal of
a synergistic (agonist) muscle. In the heart, the most powerful stimulus is a drastic
reduction of heart rate achieved either by electrical bradycardial pacing or drugs with a
relatively specific negative chronotropic effect, but substantial capillary growth can be
achieved by long-tenn administration of drugs with a positive inotropic effect. The
mechanism of capillary growth is linked with various mechanical factors such as shear
stress or stretch of the capillary wall. Capillary growth in tissues with limited blood
supply has a beneficial effect on blood flow and perfonnance.

REFERENCES

Adrian, E.D., and Bronk,D.w., 1929, The discharge of impulses in motor nerve fibres, 1.
Physiol. ,Lond 67: 119-151.

Andersen, P., and Henriksson, I., 1977, Capillary supply of the quadriceps femoris muscle
of man: adaptive response to exercise, 1. Physiol., Lond 270:677-690.

Ando, I., Nomura, H., and Kamiya, A., 1987, The effect of fluid shear stress on the
migration and proliferation of cultured endothelial cells, Microvasc. Res. 33:62-70.

Brown, M.D., Cotter, M.A., Hudlicka, 0., and Vrbova, G., 1976, The effects of different
patterns of muscle activity on capillary density, mechanical properties and structure of
slow and fast rabbit muscles, Pflugers Arch. 361:241-250.

Brown, M.D., and Egginton, S., 1988, Capillary density and ultra-structure III rabbit
papillary muscle after a high dose of norepinephrine, Microvasc. Res. 36: 1-12.

Brown, M.D., and Hudlicka, 0., 1991, Capillary supply and cardiac perfonnance in the
rabbit after chronic dobutamine treatment, Cardiovasc. Res. 25:909-915.

Brown, M.D., and Hudlicka, 0., 1995, Vascular conductance and capillary supply in heart
and skeletal muscle in response to altered activity, Microcirculation 2:98.

Brown, M.D., Hudlicka, 0., Weiss, J.B., Bate, A., and Silgram, R., 1996, Prazosin-induced
capillary growth in rat skeletal muscle; link with endothelial cell-stimulating angiogenic
factor, Int. 1. Microcirc. 16:207.

Brown, M.D., Davies, M.K., Hudlicka, 0., and Townsend, P., 1994, Long-tenn
bradycardia by electrical pacing: a new method for studying heart rate reduction,
Cardiovasc. Res. 28: 1774-1779.

Burri, P.R., and Tarek, M.R., 1990, A novel mechanism of capillary growth in the rat
pulmonary microcirculation, Anat. Rec. 2281:35-45.

Clark, E.R., 1918, Studies on the growth of blood vessels in the tail of the frog, Am. 1. Anat.
23:37-38.

Cliff, W.J., 1963, Observations on healing tissues: a combined light and

electronmicroscopic investigation, Phi/os. Trans. R. Soc. Lond B BioI. Sci. 246:305-325.

Dawson, J.M., and Hudlicka, 0., 1989a, The effect of long-tenn activity on the
microvasculature ofrat glycolytic skeletal muscle, Int. 1. Microcirc. 8:53-69.

30
Dawson, I.M., and Hudlicka, 0., 1989b, The effects of long-tenn administration of
prazosin on the microcirculation in skeletal muscles, Cardiovasc. Res. 23:913-920.

Dawson, J.M., and Hudlicka, 0., 1993, Can changes in microcirculation explain capillary
growth in skeletal muscle?, Int. 1. Exp. Path. 74:65-71.

Denekamp, l, 1984, Vasculature as a target for tumour therapy, Progr. Appl. Microcirc.
4:28-38.

Egginton, S., and Hudlicka, 0., 1992, Effect of long-tenn muscle overload on capillary
supply, blood flow and performance in rat fast muscle, 1. Physiol., Lond. 452:9P.

Frischknecht, R., and Vrbovit, G., 1991, Adaptation of rat extensor digitorum longus to
overload and increased activity, Pflugers Arch. 419:319-326.

Fulgenzi, G-L., and Hudlicka, 0., 1994, The effect of alpha 1 blocker prazosin on
capillarization, blood flow and perfonnance in ischaemic skeletal muscles, Int. 1. Microcirc.
14: Suppl 1,37.

Gute, D., Laughlin, M.H., and Amann, F.F., 1994, Regional changes in capillary supply in
skeletal muscle of interval-sprint and low-intensity, endurance trained rats,
Microcirculation 1: 183-193.

Hansen-Smith, F., Egginton, S., and Hudlicka, 0., 1997, Growth of arterioles in chronically
stimulated adult rat skeletal muscle, Microcirculation 5: In press.

Hansen-Smith, F., Egginton, S., Owens, G.K., and Hudlicka, 0., 1995, Growth of arterioles
in chronically stimulated rat fast muscles, 1. Physiol., Lond. 489P: 16P.

Hudlicka, 0., 1991, What makes blood vessels grow?,J. Physiol., Lond. 444:1-24.

Hudlicka, 0., and Brown, M.D., 1991, Capillary growth in normal and hypertrophic heart,
in: Cardiac Electrophysiology, Activation, Circulation and Transport. (Review) (S.
Sideman, R. Beyar, and A.G. Kleber, eds.), pp. 307-316, Kluwer Academic Publishers,
Boston.

Hudlicka, 0., and Brown, M.D., 1996, Postnatal growth of the heart and its blood vessels,
1. Vasco Res. 33:266-287.

Hudlicka, 0., Brown, M.D., Egginton, S., and Dawson, J.M., 1994, Effect of long-tenn
electrical stimulation on vascular supply and fatigue in chronically ischemic muscles, 1.
Appl. Physiol. 77:1317-1324.

Hudlicka, 0., Brown, M.D., Walter, H., Weiss, lB., and Bate, A., 1995, Factors involved
in capillary growth in the heart, Mol. & Cell. Biochem. 147:57-68.

Hudlicka, 0., Brown, M., and Egginton, S., 1992, Angiogenesis in skeletal and cardiac
muscle, Physiol. Rev. 72(2):369-417.

Hudlicka, 0., and Tyler, K.R., 1984, The effect oflong-tenn high frequency stimulation on
capillary density and fibre types in rabbit fast muscle, 1. Physiol., Lond. 353:435-445.

Hudlicka, 0., Tyler, K.R., Wright, AJ.A., and Ziada, A.M.A.R., 1984, Growth of
capillaries in skeletal muscles, Progress in applied microcirculation, 5, pp44-64.

31
Hudlicka, 0., West, D., Kumar, S., EI Khelly, F., and Wright, A1.A, 1989, Can growth of
capillaries in the heart and skeletal muscle be explained by the presence of an angiogenic
factor? Br. J. Exp. Pathol. 70: 237-246.

Jakob, W., Zipper, 1., Savoly, S.B., Siems, W.-E., and Jentzsch, KD., 1982, Is
dipyridamole an angiogenic agent? Exp. Patho/. 22:217-224.

Mall, G., Mattfeldt, T., Reiger, P., Volk, B., and Frolov, VA, 1982, Morphometric
analysis of the rabbit myocardium after chronic ethanol feeding - early capillary changes,
Basic Res. Cardio/. 77:57-67.

Myrhage, R, and Hudlicka, 0., 1978, Capillary growth in chronically stimulated adult
skeletal muscle as studied by intravital microscopy and histological methods in rabbits and
rats, Microvasc. Res. 16:73-90.

Pearce, S., Hudlicka, 0., and Egginton, S., 1995, Early stages in activity-induced
angiogenesis in rat skeletal muscles: incorporation of bromodeoxyuridine into cells of the
interstitium, J. Physio/., Lond. 483: 146P.

Price, R1., and Skalak, I.C., 1996, Chronic alpha-I adrenergic blockade stimulates terminal
and arcade arteriole formation in a network model, Am. J. Physiol. 40:H752-H759.

Romanul, F.CA, 1965, Capillary supply and metabolism of muscle fibers, Arch. Neurol.
12:497-509.

Salmons, S., and Vrbova, G., 1969, The influence of activity on some contractile
characteristics of mammalian fast and slow muscles, J. Physiol., Lond. 210:535-549.

Schoefl, G.I., 1963, Studies on inflammation. III. Growing capillaries: their structure and
permeability, Virchows Arch. A Pathol. Anat. Histol. 337:97-141.

Sholley, M.M., Ferguson, G.P., Seibel, HR, Montour, J.L., and Wilson, J.D., 1984,
Mechanisms of microvascularization and vascular sprouting can occur without proliferation
of endothelial cells, Lab. Invest. 51:624-637.

Snell, P.G., Martin, W.H, Buckey, J.C., and Blomqvist, e.G., 1987, Maximal vascular leg
conductance in trained and untrained men, J. App/. Physio/' 62:606-610.

Till mans, I.H, Ikeda, S., Hansen, H., Sarma, 1.S., Fauvel, J.H., and Bing, R.J., 1974,
Microcirculation in the ventricle of the dog and turtle, Orc. Res. 34:561-569.

Tillmans, H., Steinhausen, M., Leinberger, H, Thederan, H, and Kubler, W., 1982, The
effect of coronary vasodilators on the microcirculation of the ventricular myocardium, in:
Microcirculation oj the Heart, (H.TilImans, W. Kubler, H. Zebes, eds.), pp. 305-312,
Springer Verlag, Berlin.

Tornling, G., 1982, Capillary neoformation in the heart of dipyridamole treated rats, Acta
Pathol. Microbiol. Scand. Sect. A Pathol. 90:269-271.

Tomling, G., Unge, G., Adolfsson, 1., Ljungqvist, A, and Carlsson, S., 1980, Proliferative
activity of capillary wall cells in skeletal muscle in rats during long-term treatment with
dipyridamole, Arzneim. Forsch. 30:622-623.

Turek, Z., Kubat, K, Kazda, S., Hoofd, L., and Rakusan, K, 1987, Improved myocardial

32
capillarisation in spontaneously hypertensive rats treated with nifedipine, Cardiovase.
Res. 21:725-729.

Wachtlova, M., Rakusan, K, and Poupa, 0., 1965, The coronary tenninal vascular bed in
the heart of the hare (Lepidus europeus) and the rabbit (Oryctolagus domestieus), Physiol.
Bohemoslov. 14:328-331.

Wachtlova, M., Rakusan, K, Roth, Z., and Poupa, 0., 1967, The tenninal bed of the
myocardium in the wild rat (Rattus norvegieus) and the laboratory rat (Rattus norvegieus
Lab.), Physiol. Bohemoslov. 16:548-554.

Wright, AlA, and Hudlicka, 0., 1981, Capillary growth and changes in heart performance
induced by chronic bradycardial pacing in the rabbit, Cire. Res. 49:469-478.

Wright, AJ.A., Hudlicka, 0 ., and Brown, M.D., 1989, Beneficial effect of chronic
bradycardial pacing on capillary growth and heart perfonnance in volume overload heart
hypertrophy, Circ. Res. 64:1205-1212.

Zhou, A-L., and Egginton, S., 1997, Capillary growth in overloaded rat skeletal muscle: an
ultra-structural study, J. Physioi., Lond. 499P:39P.

Zhou, A.-L., Egginton, S., and Hudlicka, 0., 1996, Ultrastructural evidence for a novel
mechanism of capillary growth in rat skeletal muscle, J. Physiol., Lond. 491:28P.

Ziada, AMAR., Hudlicka, 0., Tyler, KR., and Wright, AlA, 1984, The effect of long-
term vasodilatation on capillary growth and perfonnance in rabbit heart and skeletal
muscle, Cardiovasc. Res. 18:724-732.

33
BASEMENT MEMBRANE BIOSYNTHESIS AS A BIOCHEMICAL INDEX OF
ANGIOGENESIS IN CHICK CHORIOALLANTOIC MEMBRANE

Nikos E. Tsopanoglou and Michael E. Maragoudakis

Department of Pharmacology, Medical School

University ofPatras, Patras, Greece.

The chorioallantoic membrane (CAM) of the chick is formed by fusion of the somatic
mesoderm of the chorion with the splachnic mesoderm of the allantois during the 4th to 5th
day of embryonic development. After 6 days of incubation, the CAM is covering a surface
area of approximately 6 cm 2 . By day 10, and extending to day 14, mean surface area of the
CAM is approximately 65 cm 2. This expansion of CAM is accompanied by an increased
complexity of the patterns of capillary microvessels. The number of capillary vessels per
square centimetre is increased about 150% from day 7 to day 12 and the intercapillary
distances have been reduced about 40% during the same period (1). This highly vascularized
membrane serves as the first respiratory system of the avian.
Because of its extensive microvascular network, the CAM is a commonly used system
for studying angiogenesis. It is relatively simple and cheap and it permits screening of a large
number of samples. Test substances are applied on the CAM surface placed on tissue culture
pellets, on gelatine sponges, on filter paper etc., and the evaluation of modulating-angiogenesis
is usually carried out by morphological and histological technics.
Basement membranes (BMs) are heterogeneous, highly specialised structures of
extracellular matrix proteins. They provide anchorage of adjacent cells as well as stimuli for cell
differentiation, migration and cell phenotype. The most abundant component of BMs is
collagen type IV, which is the collagen characteristically found in BMs. A three dimensional
network of collagen type IV appears to form the basic structure upon which the other
components are attached.
All blood vessels have a layer of endothelial cells which adhere to a thin layer of BM.
This BM is an essential structure element and provides the support and the adhesive surface
for the anchorage of endothelium. Angiogenic process involves the local dissolution of the BM,
proliferation, migration, differentiation of endothelial cells, the formation of a lumen and the

AngiogeneSis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 35
deposition of a new BM which is synthesised by stimulating endothelial cells (2). Since at the
early stages of their formation blood vessels consist only of endothelial cells and the
underlying BM, a quantitative relationship must exist between the rate of new BM synthesis
and angiogenesis. Background activity ofBM synthesis in the absence of angiogenesis is very
low, because both BM synthesis and angiogenesis are extremely slow processes under normal
physiological conditions.
Considering the above rational, an assay system in which we monitor the rate of BM
collagen biosynthesis as an index of angiogenesis has been described (3). Briefly, fresh
fertilised eggs were placed horizontally in an incubator. On day 4 of incubation, a square
window was opened in the egg shell after removal of 3 ml of albumin so as to detach the
developing CAM from the cell. Test materials with the radiolabelled proline (O.S !-lCi/disc)
were placed on sterile round plastic discs and were allowed to dry under sterile conditions. The
loaded discs were inverted and placed on the surface of CAM on day 9. Control discs
containing equal amounts of radiolabelled proline were placed on the CAM about 1 cm away
from test discs. Cortisone acetate was added to all discs to prevent inflammatory responses.
On day 11, the CAM under the discs was cut off. The round pieces of CAM were transferred
into centrifuge tubes containing cycloheximide and dipyridyl to stop protein synthesis and
hydroxylation of proline and lysine. The samples were boiled and subsequently were washed
extensively from non-protein bound radioactivity. Samples were then suspended in buffer
containing 7.S IU bacterial collagenase vn and SOmM calcium chloride. The mixture was
incubated at 37°C for 4 hours and the collagenase digestion products were separated by
centrifugation. Radioactivity in the supernatant corresponds to radiolebelled tripeptides from
BM collagen and other collagenous materials synthesised by the CAM. The collagenous
proteins synthesised by CAM under these conditions is mostly BM collagen type IV. As
shown by column chromatography, the collagenous protein synthesised by CAM had the
same molecular size as type IV collagen from bovine lenses (4). Also, a specific collagenase
type IV purified from Walker 256 rat carcinosarcoma degraded the collagenous protein
biosynthesised by CAM to the same extent as bacterial collagenase VII (S).
Maximum rate of angiogenesis in the CAM system occurs between days 8 and 10 of
chick embryo development as shown by the increase in vascular density (3). After day 12 the
vascular density reaches a plateau. Between days 8-12 the vascular density increases about
three-fold. At that period the rate of collagenous protein was maximum. At day 10 rate of
collagenous protein synthesis was three-fold higher than that observed on day 7 and II-fold
higher than that observed on day IS when vascular density reached a plateau. This shows that
maximum rate of collagenous protein biosynthesis coincides with the stage of maximum
angiogenesis as measured by morphological evaluation of the vascular density of CAM (3).
Collagenous protein synthesis as an index for angiogenesis was validated using a
commonly employed method for the morphological evaluation of angiogenesis in the CAM,
developed by Harris-Hooker et al. (6).1t involves the measurement of all vessels intersecting
three concentric rings of 4,S and 6mm in diameter. Recently, image analysis has been used to
study neovascularization after irradiation in vivo in the rat corneal model (7), as well as
microvascular proliferation in the disc angiogenesis system (8), and the effect of hypoxia on
the vascularity of the CAM (9).
We have used five compounds that have been shown to modulate angiogenesis via
different mechanisms. The tumour promoter and protein kinase C activator PMA and
thrombin have been shown to promote angiogenesis. On the other hand Ro 318220, a selective

36
inhibitor of protein kinase C, the antitumor agent D609 and GPA 1734 which inhibits prolyl
and lysyl hydroxylase have been shown to inhibit angiogenesis (10). The angiogenic or
antiangiogenic responses are evaluated biochemically using the collagenous protein
biosynthesis method or morphologically using Harris-Hooker et al. method and computer-
assisted image analysis method. It has shown that the results obtained using the three methods
are comparable.

CONCLUSION

Collagenous protein synthesis offers a reliable, unbiased, and reproducible index of

angiogenesis which provides comparable results to established morphological methods. The
advantages of using this index are based on the fact that it involves biochemical evaluation of an
essential vessel component and takes into account all sizes of newly formed vessels, induding
capillaries.

REFERENCES

1. Defouw, O.D., Rizzo, V.J., Steinfeld, R. and Feinberg, R.N. (1989). Mapping of the
microcirculation in the chick chorioallantoic membrane during normal angiogenesis.
Microvasc. Res., 38: 136-147.
2. D' Amore, P.A and Thompson, R.N. (1987). Mechanism of angiogenesis. Ann. Rev.
Physiol., 49: 453-464.
3. Maragoudakis, M.E., Panoutsakopoulou, M. and Sarmonica, M. (1988). Rate of
basement membrane biosynthesis as an index of angiogenesis. Tissue and Cell, 20: 531-
539.
4. Maragoudakis, M.E., Missirlis, E., Sarmonika, M., Panoutsakopoulou, M. and
Karakioulakis, G. (1989). Basement membrane biosynthesis as a target to tumour
therapy. J. Pharmacol. Exp. Ther., 252: 753-757.
5. Karakioulakis, G., Missirlis, E., Aletras, A. and Maragoudakis, M.E. (1988).
Degradation of intact basement membranes by human and murine tumour enzymes.
Biochim. Biophys. Acta. 967: 163-175.
6. Harris-Hooker, SA, Gajdusek, C.M., Wight, T.N. and Schwartz, SM. (1983).
Neovascular responses induced by cultured aortic endothelial cells. J. Cell. Physio!., 114:
302-310.
7. Scroggs, AW., Proia, AD., Smith, C.F., Halperin, E.c. and Klintworth, G.K. (1991).
The effect of total body irradiation on cornea neovascularization in the fischer 344 rat
after chemical cauterisation. Invest. Ophthalmol. Visual. Sci., 32: 2105-2111.
8. Kowalski, J., Khwan, H.H., Prionas, S.D., Allison, AC. and Fajardo, L.F. (1992).
Characterisation and application of the disc angiogenesis system. Exp. Mol., 56: 1-19.
9. Strick, D.M., Waycarter, R.L., Montani, J.P., Gay, W.J. and Adair, T.H. (1991).
Morphometric measurements on chorioallantoic membrane vascularity: Effects of
hypoxia and hyperoxia. Am. J. Physiol., 260:H1385-H1389.

37
10. Maragoudakis, M.E., Haralabopoulos, G.C., Tsopanoglou, N.E. and Pipili-Synetos, E.
(1995). Validation of collagenous protein synthesis as an index for angiogenesis with the
use of morphological methods. Microvasc. Res., 50:215-222.

38
THE RAT AND THE RABBIT CORNEA ASSA Y

Lucia Morbidelli and Marina Ziche

Dept. Pharmacology, University of Florence, Viale Morgagni 65, 50134

Florence, Italy.

In order to develop angiogenic and antiangiogenic strategies, there are concerted efforts to
provide animal model for more quantitative analysis of in vivo angiogenesis. In vivo
techniques consist in the rat and rabbit cornea assay, the sponge implant assay, the fibrin
clots, sodium alginate beads and matrigel plugs and the chick embryo chorioallantoic
membrane. In this chapter we will discuss on the avascular cornea assay and the advantages
and divadvantages in the rat and the rabbits, the two animal species most used.
The cornea assay consists in the placement of an angiogenesis inducer (tumor tissue, cell
suspension, growth factors) into a corneal pocket in order to evoke vascular outgrowth
from the peripherally located limbal vasculature. This assay, respect to the other in vivo
assays, has the advantage of measuring only new blood vessels, since the cornea is initially
avascular.

EXPE~NTALPROCEDURES

Rabbit cornea assay

The corneal assay performed in New Zealand white rabbits was firstly described by
Gimbrone et al. (1974). We have been using this technique extensively during the years
and have substantially modified it to fulfill different experimental requirements. The
material under test can be in the form of slow-release pellets incorporating recombinant
growth factors, cell suspensions, or tissue samples (Ziche et aI., 1982; Ziche and Gullino,
1982).
Surgical procedure: After being anaesthetised with sodium pentothal (30 mglkg, i.v.), a
micro pocket (1.5 x 3 mm) is surgically produced under aseptic conditions using a pliable
iris spatula 1.5 mm wide in the lower half of the cornea. A small amount of the aqueous
humor can be drained from the anterior chamber when reduced corneal tension is required.
The implant is positioned at 2.5-3 mm from the limbus. Implants sequestering the test
materials and the controls are coded and implanted in a double masked manner.
Slow release preparations: Recombinant growth factors are prepared as slow-release
pellets by incorporating the protein under test into an ethylen-vinyl-acetate copolymer

Angiogenesis: Models, Modulators. and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 39
(Elvax-40) (DuPont de Nemours, Wilmington, DE). In order to avoid non specific reactions
the Elvax-40 for implantation has to be carefully prepared as follows: after extensive
washings of the Elvax-40 beads in absolute alcohol for 100 fold at 37°C, a 10% casting
stock solution is prepared in methyl en-chloride and tested for its biocompatibility (Langer
and Folkman, 1976). The casting solution is eligible for use ifno of the implants performed
with this preparation induces the slightest or histological reaction in the rabbit cornea. For
testing, a pre-determined volume ofElvax-40 casting solution is mixed with a given amount
of the compound to be tested and the polymer is allowed to dry under a laminar flow hood.
After drying the film sequestering the compound is cut into 1 x 1 x 0.5 mm pieces. Empty
pellets of Elvax-40 are used as controls.
Cell and tissue implants: Cell suspensions are obtained by trypsinization of confluent cell
monolayers. Five Jll containing 2x105 cells in medium supplemented with 10% serum are
introduced in the cornea micropocket. If the overexpression of growth factors by stable
transfection of specific cDNA is studied, one eye is implanted with transfected cells and
the other with the wild type cell line. When tissue samples are tested, samples of 2-3 mg
are obtained by cutting the original fragments under steril conditions. The angiogenic
activity of tumor samples is compared with macroscopically healthy tissue.
Quantification: Subsequent daily observation of the implants is made with a slit lamp
stereomicroscope without anaesthesia. Angiogenesis, edema and cellular infiltrate are
recorded daily by an independent operator from the surgeon. An angiogenic response is
scored positive when budding of vessels from the limbal plexus occurs after 3-4 days and
capillaries progress to reach the implanted pellet according to the scheme previously
reported (Ziche et aI., 1989). The number of positive implants over the total implants
performed is scored during each observation, as shown in Table 1. The potency of
angiogenic activity is evaluated on the basis of the number and growth rate of newly
formed capillaries, and an angiogenic score is calculated [vessel density x distance from
limbus] (Ziche et aI., 1994, 1997c). A density value of 1 corresponds to 0 to 25 vessels per
cornea, 2 from 25 to 50,3 from 50 to 75, 4 from 75 to 100 and 5 for more than 100 vessels.
The distance from the limbus is graded with the aid of an ocular grid. An example of the
angiogenic score parameters is reported in Fig. 1.
Histological examination: Corneas are removed at the end of the experiment as well as at
defined intervals after surgery and/or treatment and fixed in formalin for histological
examination. Newly formed vessels and the presence of inflammatory cells are detected by
hematoxylin/eosin staining or specific immunohistochemical procedure (i.e. anti-rabbit
macrophages (RAMI 1), anti-CD-31 for endothelium) (Ziche et aI., 1997c). Double staining
(i.e. anti-CD-31 for vascular endothelium and specific markers for tumor cells) could be
useful to label newly formed vessels of the host and proliferating tumor cells implanted in
the cornea.

Rat cornea assay

Due to its size, the rat eye is essentially suitable for studing the angiogenic effect of
purified growth factors by the use of slow release preparations. Test substances are
combined 1: 1 with hydroxy-ethylmethacrylate (HydronR) (Interferon Science, New
Brunswick, NJ) as described by Polverini and Leibovich (1984). Hydron is used as a
casting solution of 12% (w/v) prepared dissolving the polymer in absolute alcohol at 37°C
(Langer and Folkman, 1976). Pellets are implanted 1-1.5 mm from the limbus of the cornea
of anaesthetised rats (sodium pentobarbital, 30 mglkg, i. p.). Neovascularization is assessed
at fixed days (usually 3, 5 and 7 days). To photograph responses, animals are perfused
with colloidal carbon solution to label vessels, eyes are enucleated and fixed in 10% neutral
buffered formalin overnight. The following day, corneas are excised, flattened and
photographed. A positive neovascularization response is recorded only if sustained
directional ingrowth of capillary sprouts and hairpin loops toward the implant is observed.

40
 5
A

..
Q;'
0
v
4

.c
0;;;
3

=
..
OJ
~
2

'"'"OJ
>

0
0 5 10 15 20

5 -0-- VEGF121
B

e 4 • VEGF165

~
e
..c: 3
0iJ
=
..'"
~

'"OJ
2

>

04------r----~----~r_--~
o 5 10 15 20
Time (days)
Figure 1
Effect of VEGF isoforms (VEGF 121 and VEGF 165) on capillary density and progression.
Slow release pellets containing the purified growth factors (200 ng/pellet) were implanted
in the cornea stroma. The two parameters of the angiogenic score are reported separately
during time. Vessel density (graded as score) is shown in panel A, while the vessel length
(distance in mm from the limbus) is reported in panel B. Data are the means of 4 implants.

Negative responses are recorded when either no growth is observed or when only an
occasional sprout or hairpin loop showing no evidence of sustained growth is detected.

ADVANTAGES AND DISADVANTAGES OF TIlE RABBIT CORNEA ASSA Y

Advantages:
I) Multiple observations. The use of slit lamp stereomicroscope and of not anaestesized
animals allows the observation of newly formed vessels during time with long time
monitoring, even for 1-2 months.
2) Size. Rabbits have large cornea which has been found avascular in all strains examined so
far. Rabbits are more docile and amenable to handling and experimentation. The rabbit size
(2-3 kg) lets an easy manipulation of both the whole animal and the eye to be easly
extrused from its location and to be surgically manipulated. General anaesthesia is required
only for surgery while daily examinations occasionally need local anaesthesia.

41
3) Systemic treatment. Systemic drug administration is possible by different routes, i.e.
parenteral and oral (Ziche et aI., 1994, 1997b, 1997c).
4) Easy monitoring of inflammation. Inflammatory reactions are easly detectable by
stereomicroscopic examination as corneal opacity.
5) Different experimental procedures. In the rabbit eye, due to its wide area, stimuli in
different forms can be placed. In particular the activity of specific growth factors can be
studied in the form of slow-release pellets (Ziche et aI., 1982, 1990, 1997a; Bussolino et
aI., 1990, 1991; Albini et aI., 1996) and of tumor or non-tumor cell lines stably transfected
for the over-expression of angiogenic factors (Zhang et aI., 1995; Coltrini et aI., 1995;
Gualandris et aI., 1996; Ziche et aI., 1997c; Chouduri et aI., 1997). The modulation of the
angiogenic responses by different stimuli can be assessed in the rabbit cornea assay through
the implant and/or removal of multiple pellets placed in parallel micropockets produced in
the same cornea (Ziche et aI., 1989, 1994). The implant of tumor samples from different
locations can be performed both in corneal micropockets and in the anterior chamber of the
eye to monitor angiogenesis produced by hormone-dependent tissues or tumors (i.e human
breast or ovary carcinoma in female rabbits) and it allows the detection of both the iris and
the corneal neovascular growth.
Disadvantages
1) Species. Reagents of rabbit origin or against rabbit antigens are not available.
2) Size. Large amounts of drugs are required for systemic treatment.

ADVANTAGES AND DISADVANTAGES OF THE RAT CORNEA ASSAY

Advantages:
l) Size: The size of the animal allows the use of small amounts of growth factors or
substances to be tested.
2) Species: The species allows the use of reagents such as antibody, synthesized or
purified from these animals and the use of cell lines, expecially tumoral, explanted form
mice, without giving antigenic reactions.

Table 1. Comparison of the angiogenic activity elicited by PIGF-l, VEGF165 and

FGF-2 on the rabbit cornea assay

Time (days)
ComEound 5 7 10
Control 0110 0110 0110

PlGF-1 4/6 6/6 6/6

VEGF165 4/6 4/6 5/6

FGF-2 2/6 3/6 5/6

The angiogenic activity induced by purified placenta growth factor-1 (PlGF-l), vascular
endothelial growth factor)65 (VEGF 165) and fibroblast growth factor-2 (FGF -2) was
assessed in the rabbit cornea assay as slow release pellets containing 200 ng of each
peptide. Data are expressed as positive implants exhibiting neovascularization over total
implants during time (days). An angiogenic response was scored positive when budding of
vessels from the limbal plexus occurred after 3-4 days and capillaries progressed to reach
the pellet containing the angiogenic factors.

42
3) Systemic treatment: The systemic treatment of mice with drugs or the implantation of
tumor cells by different routes is possible.
Disadvantages
I) Size: The relatively small size of the eyes in these animals and the presence of preexiting
vessels within the cornea in some of the strains are serious disadvantages. Surgical
procedures and manipulation are difficult because of the small size and the need of general

B
.......

~'
.\

''C
'.
(
'-

Figure 2
Image analysis of a positive angiogenic response. A, Starting image in black and white, and
8, elaboration of the digitalized image to remove the background and to put in evidence the
newly formed vessels.

anaesthesia of these animals during each step is a drawback for their routine use in large-
scale experiments.
2) Time-point results: the evolution of the angiogenic response in the same animal is not
recommended because each time the cornea is observed the animal has to be anaestesized
since it is not easy to keep it quiet. Thus experiments are made with a large number of rats
and vessel growth during time is visulized by perfusion with colloidal carbon solution in
individual animals.
3) Spontaneous inflammatory reactions can easly develop because of manipulation
procedures.

43
MORPHOLOGICAL ANALYSIS AND QUANTIFICATION OF
NEOVASCULARIZATION

In the past angiogenesis has been described qualitatively (strong, weak, absent) or by
measuring the length of the most advanced vascular sprout at daily intervals. Moreover,
flat mount preparations of cornea were used to obtain more quantitative assessment using
imaging techniques and vector analysis to determine the directional properties of newly
penetrating blood vessels (Proia et al ., 1988). The use of fluorescence to provide better
imaging properties has led to more precise measurements of neovascularization. The
availability of sensitive video cameras and image recording now permit sequential recording
of neovascular process in individual animals without the need to resort to the dissection
and fixation processes previously needed for obtaining imaging data from the flat mount
preparations (Conrad et al.,.1994).
Inherent with the in vivo approach is the capacity to establish, quantify and characterize
the spatial and temporal pattern of corneal angiogenesis elicited by distinct angiogenesis
effectors in each animal over a long period of time. By the use of an accurate method of
measurement by image analysis (Fig. 2), the effect of intervention on the process can be
shown and quantified.

CONCLUDING REMARKS

Continuous monitoring of angiogenesis in vivo is required for the development and

evaluation of drugs acting as suppressors or stimulators of angiogenesis. In this respect the
avascular cornea of New Zealand albino rabbits offers a unique model, since progression of
neovascularization can be monitored for extended periods of time with a non invasive
approach, and the comparison in the same animal of distinct effectors is possible (Ziche et
aI., 1982; Ziche et aI., 1989). Measurements of corneal angiogenesis is useful for
quantitating the effects of angiogenic stimuli and for evaluating the efficacy of potential
inhibitors of neovascularization. Because accurate methods suitable for recording the entire
pattern of corneal neovascularization over time and for obtaining a quantitative evaluation
of the process in individual living animals do not exist, we are developing a non invasive
method to achieve this goal by the use of a computerized image analysis system.

ACKNOWLEDGEMENT

This work was supported by funds from the European Communities BIOMED-2
(PL950669): "Angiogenesis and Cancer", the Italian Association for Cancer Research (AIRC,
Special Project "Angiogenesis") and the National Research Council of Italy (Project #:
96.03745 .14).

REFERENCES

Albini, A., R Benelli, M. Presta, M. Rusnati, M. Ziche, A. Rubartelli, G. Paglialunga, F.

Bussolino and D. Noonan. HIV-tat protein is a heparin-binding angiogenic growth
factor. Oncogene, 12: 289-297, 1996.
Bussolino, F., M. Ziche, 1. M. Wang, D. Alessi, L. Morbidelli, O. Cremona, A. Bosia, P.
Marchisio and A. Mantovani. In vitro and in vivo activation of endothelial cells by
colony-stimulating factors. 1. Clin. Invest., 87: 986-995, 1991.
Bussolino, F., M .F. Di Renzo, M. Ziche, E. Bocchietto, M. Oliviero, L. Naldini, G.
Gaudino, L. Tamagnone, A. Coffer, P.c. Marchisio and P.M. Comoglio. Hepatocyte

44
 growth factor is a potent angiogenic factor which stimulates endothelial cell motility
and growth. 1. Cell BioI., 119(3): 629-641,1992.
Choudhuri, R., H.-T. lhang., S. Donnini, M. liche and R. Bicknell. An angiogenic role for
the neurotrophins midkine and pleiotrophin in tumorigenesis. Cancer Res., 57: 1814-
1819, 1997.
Coltrini, D., A Gualandris, E.M. Nelli, S. Parolini, M.P. Molinari-Tosatti, N. Quarto, M.
liche, R. Giavazzi and M. Presta. Growth advantage and vascularization induced by
basic fibroblast growth factor overexpression in endometrial HEC-I-B cells: an export-
dependent mechanism of action. Cancer Res. 55: 4729-4738, 1995.
Conrad, T.J., D.B. Chandler, J.M. Corless and G.K. Klintworth. In vivo measurement of
corneal angiogenesis with video data acquisition and computerized image analysis. Lab.
Invest., 70: 426-434, 1994.
Gimbrone, M. Jr., R. Cotran, S.B. Leapman, and 1. Folkman. Tumor growth and
neovascularization: An experimental model using the rabbit cornea. 1. Natl. Cancer Inst.
52: 413-427,1974.
Gualandris, A, M. Rusnati, M. Belleri, E.M. Nelli, M. Bastaki, M. P. Molinari-Tosatti, F.
Bonardi, S. Parolini, A. Albini, L. Morbidelli, M. liche, A. Corallini, L. Possati, A.
Vacca, D. Ribatti, and M. Presta. Basic fibroblast growth factor overexpression in
mouse endothelial cells: an autocrine model of angiogenesis and angioproliferative
diseases. Cell Growth Differentiation 7: 147-160,1996.
Langer, R., and 1. Folkman. Polymers for the sustained release of proteins and other
macromolecules. Nature 363: 797-800, 1976.
Polverini, P.J., and S.J. Leibovich. Induction ofneovascularization in vivo and endothelial
cell proliferation in vitro by tumor associated macrophages. Lab. Invest., 51: 635-642,
1984
Proia, AD., D.B. Chadler, w.L. Haynes, C.F. Smith, C. Suvarenamani, F.R Erkel and
G.K. Klintworth. Quantification of corneal neovascularization using computerized
image analysis. Lab. Invest., 58: 473-479, 1988.
lhang RT., P. Kraft, P.A.E. Scott, M. liche, H.A Weich, AL. Harris and R. Bicknell.
Enhancement of tumour growth and vascular density by transfection of vascular
endothelial cell growth factor (VEGF) into MCF7 human breast carcinoma cells. J.
Natl. Cancer Inst., 87 (3): 213-219,1995.
liche, M. and P.M. Gullino. Angiogenesis and neoplastic progression in vitro. J. Natl.
Cancer Inst., 69: 483-487, 1982.
liche, M., D. Maglione, D. Ribatti, L. Morbidelli, c.T. Lago, M. Battisti, 1. Paoletti, A.
Barra, M. Tucci, G. Parise, V. Vincenti, H.J. Granger, G Viglietto, and G.M. Persico.
Placenta growth factor-l (PIGF -1) is chemotactic, mitogenic and angiogenic. Lab.
Invest., 76: 517-531, 1997 (a).
liche, M., G. Alessandri, and P.M. Gullino. Gangliosides promote the angiogenic
response. Lab. Invest., 61: 629-634, 1989.
liche, M., G. Gasparini, R. Choudhuri, R. Bicknell, L. Morbidelli, A. Parenti and F.
Ledda. Antiangiogenic effect of Linomide on breast cancer VEGF transfectants.
Oncology Report, 4: 253-256, 1997 (b).
liche, M., 1. Jones, and P.M. Gullino. Role of prostaglandinE 1 and copper in angiogenesis.
1. Natl. Cancer Inst., 69: 475-482, 1982.
liche, M., L. Morbidelli, E. Masini, S. Amerini, H.J. Granger, C.A Maggi, P. Geppetti
and F. Ledda. Nitric oxide mediates angiogenesis in vivo and endothelial cell growth and
migration in vitro promoted by Substance P. 1. Clin. Invest., 94: 2036-2044, 1994.
liche, M., L. Morbidelli, M. Pacini, P. Geppetti, G. Alessandri and C.A Maggi.
Substance P stimulates neovascularization in vivo and proliferation of cultured
endothelial cells. Microvasc. Res., 40: 264-278, 1990.

45
Ziche, M., L. Morbidelli, R. Choudhuri, H.-T. Zhang, S. Donnini, H.1. Granger and R.
Bicknell. Nitric oxide-synthase lies downstream of Vascular Endothelial Growth Factor
but not basic Fibroblast Growth Factor induced angiogenesis. 1. Clin. Invest., 99: 2625-
2634, 1997 (c).

46
THE RAT AORTA MODEL OF ANGIOGENESIS: METHODOLOGY AND
APPLICATIONS

Roberto F. Nicosia

Department of Pathology, Allegheny University of the Health Sciences,

Philadelphia PA 19102, USA

1. ABSTRACT

Angiogenesis can be studied in vitro by culturing rings of rat aorta in collagen gels under
serum-free conditions. Ring-shaped explants are obtained by cross sectioning the rat aorta
under a dissecting microscope. The aortic rings are washed with serum-free MCDB 131
growth medium and embedded in collagen gels prepared in agarose culture wells. After
gelation has occurred, each collagen gel is freed of the surrounding agarose support and
transferred to a plastic culture well where it floats in 0.5 ml of serum-free medium. The
cultures are kept in a humidified CO 2 incubator at 3S.S o C. Aortic rings grown under these
conditions generate branching microvessels in the absence of serum or exogenous growth
factors. The angiogenic response can be modulated by adding soluble angiogenic agonists or
antagonists to the culture medium. The solid phase in which the aortic rings are embedded
can be modified by incorporating additional extracellular matrix molecules to the collagen
solution before gelation. The neovessels arise primarily from the luminal cut edges of the
explants and are composed of endothelial cells surrounded by pericytes. The cultures
contain fibroblasts which are the first cells to migrate out of the aortic wall. The angiogenic
response starts at day 3-4 and ends at day 9-10. After it has formed, the vascular
outgrowth undergoes a process of regression and remodeling with retraction of the small
endothelial branches into the main stem of the microvessels. During this phase, pericytes
increase in number by migrating and proliferating at the abluminal surface of the
endothelium. Angiogenesis can be quantitated by direct visual counts of microvessels and
by computer-assisted image analysis. The rat aorta model can be used to evaluate various
aspects of the angiogenic process including autocrine/paracrine regulation of angiogenesis
by cells of the vessel wall, modulation of angiogenesis by the extracellular matrix,
stimulation of microvessel formation by angiogenic factors, and inhibition of angiogenesis

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 47
2. INTRODUCTION

In 1982 we reported that explants of rat aorta cultured in plasma clot in presence of fetal
bovine serum generated neovascular outgrowths (Nicosia, Tchao and Leighton, 1982). We
later found that plasma clot and serum were not required for the angiogenic response of the
rat aorta which occurred also in gels of purified fibrin or collagen under serum-free
conditions (Nicosia and Ottinetti, 1990a). Similar observations were made concurrently by
Kawasaki, Mori and Awai (1989). Serum-free cultures of rat aorta proved to be a reliable
and reproducible assay for studying angiogenesis and its mechanisms This chapter
describes the methodology currently used in our laboratory to prepare and analyze these
cultures and briefly summarizes the applications of this model.

3. METHODOLOGY

3. 1 Excision of the Rat Aorta

The thoracic aorta is obtained under sterile conditions from a 2-3-month-old Fischer 344
male rat sacrificed by a lethal intraperitoneal injection of sodium pentobarbital or by CO2
asphyxiation. After exposing the posterior mediastinum, the aorta is ligated with silk
sutures (3-0) both proximally, below the aortic arch, and distally, above the diaphragm.
Fine curved microdissection forceps (Fine Science Tools Inc., Foster City, CA) are used for
this procedure. The aorta is then cut below the distal future, dissected away from the
vertebral column, excised below the proximal suture and transferred to a compartmentalized
Felsen dish containing serum-free MCDB 131 growth medium (Knedler and Ham, 1987;
Endothelial Basal Medium, Clonetics, San Diego, CA). Transfer is carried out by cutting
the aorta below the distal suture directly into the dish. During excision, attention is paid
not to stretch or traumatize the aorta and not to cut the surrounding veins, to avoid
unnecessary bleeding.

3.1.1. Preparation of Collagen Gel Cultures of Rat Aorta

The main steps in the preparation of serum-free collagen gel cultures of rat aorta are
summarized in Figure 1 (Villaschi and Nicosia, 1993; Nicosia, 1997). A similar approach
can be used to obtain fibrin gel cultures (Nicosia and Ottinetti, 1990a). Fibroadipose tissue
and intraluminal blood clots are removed from the aorta using Noyes scissors and curved
microdissection forceps (Fine Science Tools Inc, Foster City, CA). This procedure is
performed under a dissecting microscope in a tissue culture room with REP A-filtered air.
Care is taken not to stretch, cut or crush the aortic wall. While it is cleaned, the aorta is
transferred to successive compartments of the Felsen dish. After four serial washes, the
aorta is cross sectioned into 1-2 mm long rings with Noyes scissors. The proximal and
distal 1 mm-long segments of the aorta, which are used to hold the explant during
dissection, are discarded. The rings are then subject to twelve consecutive washes, each
containing 4 ml of serum-free medium (three Felsen dishes). Microdissection forceps are
used to transfer the rings, avoiding mechanical damage. This procedure and the procedure
for preparing the collagen gels (see below) are performed in a black wooden hood closed on
top by a transparent glass which protects the cultures from dust contamination and allows
direct view from above. This arrangement facilitates the handling of the rat aorta which is
otherwise cumbersome in a laminar flow hood. The same procedures can be performed
comfortably in an open tissue culture hood with positive air flow.

48
The aortic rings are embedded in collagen gel using cylindrical agarose wells. A sterile 1.5%
solution of agarose (type VIA, Sigma Chemical Company, St Louis, MO) is poured into
100 X 150 culture dishes (35 ml/dish; Falcon, Lincoln Park, NJ) and allowed to gel in a
laminar flow hood. Dishes with solidified agarose can be stored up to 4 weeks in an air-
tight container at 4 0c. Agarose rings are prepared before each experiment by punching
two concentric circles in the agarose plate with nylon punchers having diameters of 10 and
17 mm respectively. The agarose inside and outside the rings is removed with a bent glass
pipette. The agarose rings are then transferred with a bent spatula to clean 100 X 150
dishes and gently tapped with the spatula to insure adherence to the bottom of the dish.
This method produces a cylindrical well with a plastic bottom and an agarose wall. Four
drops of collagen solution are placed with a transfer pipette on the bottom of each agarose
well . Collagen is purified from the rat tail as reported (Elsdale and Bard, 1972). The
collagen solution is prepared by mixing on ice eight volumes of 1 mg/ml collagen in 1110
Minimum Essential Medium (MEM), pH 4.0, with a freshly made solution containing one
volume of lOX MEM and one volume of23.4 mg/ml Na HC0 3 . After the bottom collagen
has gelled for 5 min at 35.5-37 °C, each aortic ring is placed on the agarose surface at the
edge of the culture well. The well is then filled with collagen solution and the aortic ring is
tipped into it with a transfer pipette (Fisher, Malvern, PA). After it sinks to the bottom,
the aortic ring is positioned in the collagen solution with fine microdissection forceps so
that its cut edges are clearly visible. The cultures are then incubated for 30 min at 35.5-37
0c. Once the collagen has gelled, the agarose is removed using a scalpel blade and a bent
spatula. Each collagen gel is then transferred with the spatula to an 18 mm well where it
floats in 0.5 ml of serum-free medium (4-well NUNC dish, Interlab, Thousand Oaks, CA).
Each experimental group includes 3 to 4 cultures. The dishes are transferred to a
humidified CO2 incubator at 35.5 0c. The growth medium is changed three times a week
starting from day 3 of culture.

(()) )
I
RAT AORTA

PLASTIC CULlURE WELL

I

SERUM-FREE COLLAGEN
MEOIUM GEL

Figure 1. Steps for the preparation of serum -free collagen gel cultures ofrat aorta.

49
3.1.1.1. Quantitation of Angiogenesis
Aortic rings in serum-free collagen gel culture generate branching microvessels surrounded
by fibroblasts (Figure 2). The angiogenic response starts at days 3-4 and ends at days 9-
10. The neovessels arise primarily from the luminal cut edges of the explants and are
composed of endothelial cells surrounded by pericytes. The cultures contain fibroblasts
which are the first cells to migrate out of the aortic wall. Mter it has formed, the vascular
outgrowth undergoes a process of regression and remodeling with retraction of the small
endothelial branches into the main stem of the microvessels. During this phase, pericytes
increase in number by migrating and proliferating at the abluminal surface of the
endothelium.
The angiogenic response of the rat aorta can be quantitated by visual counts (Nicosia and
Ottinetti, 1990a, Nicosia and Bonanno, 1991) or by computer-assisted image analysis
(Nicosia, Bonanno and Smith, 1993; Nissanov et aI., 1995). For visual quantitation,
curves of microvascular growth are generated by counting the number of microvessels daily
or every other day. Cultures are examined and scored under bright-field microscopy using
an inverted Leitz microscope equipped with 2X and 4X objectives. Optimal contrast and
depth offield are obtained by closing the iris diaphragm of the condenser. Angiogenesis is
scored according to the following criteria. (a) Microvessels are distinguished from
fibroblasts based on their thickness and cohesive pattern of growth; (b) the branching of
one microvessel generates two new microvessels, which are added to the count; (c) each
loop is counted as two microvessels because it originates from two converging
microvessels. The length of microvessels can be measured directly from living cultures by
digitizing morphometry (Nicosia et aI., 1993). Data obtained by manually tracing
individual microvessels with a computer mouse are recorded and evaluated with Bioquant
IV image analysis software. The aortic outgrowth can also be measured automatically using
an image processing algorithm which segments the neovessels from gray scale images

Figure 2. Photomicrograph of living collagen gel culture of rat aorta. The aortic explant
has generated an outgrowth of microvessels (arrows) and fibroblasts (arrowheads).
Magnification:
X30.

50
(Nissanov et a!., 1995). A digital filter separates the images into vascular and nonvascular
(fibroblasts) compartments based on object size and shape. Quantification relies on the
identification of vessels intersecting a closed transect set at a fixed distance from the aortic
explant. Correlation between computer-assisted quantification and visual microvessel
count is high. This automated method can process approximately 30 images/hour and is
particularly useful for pharmacologic screening which requires quantitative analysis of a
large number of cultures.

4. APPLICATIONS OF THE RAT AORTA MODEL

The angiogenic response of the aortic explants can be either stimulated or inhibited by
modifying the soluble phase (growth medium) or the solid phase (collagen gel). Soluble
angiogenesis agonists (Nicosia, Nicosia and Smith, 1994) or antagonists (Nicosia and
Ottinetti, 1990a) can be added to the growth medium. The effect of extracellular matrix
molecules can be tested by incorporating these molecules into the collagen or fibrinogen
solutions before gelation (Nicosia et al ., 1993; Nicosia, Bonanno and Yurchenco, 1994;
Nicosia and Tuszynski, 1994). Time-lapse videomicroscopy can be used to study the
motile and proliferative behavior of the endothelium and pericytes (Nicosia and Villaschi,
1995). Collagen gel cultures can be fixed in formalin for light microscopic studies or in
glutaraldehyde for ultrastructural analysis (Nicosia et a!., 1982; Nicosia and Ottinetti,
1990a). For paraffin embedding, the collagen gels are processed manually with
conventional methods for dehydration, paraffin infiltration and embedding. Formalin-fixed
or unfixed collagen gels can also be embedded in Optimal Cutting Temperature (OCT)
compound (Baxter, McGaw Park, IL), and snap frozen in cold isopentane for
cryosectioning (Nicosia and Villaschi, 1995). Soluble factors secreted by the aortic rings
and outgrowths in the culture medium can be studied by Western blot, immunoblot and
enzyme-linked immunosorbent assay (Villaschi and Nicosia, 1993; Nicosia, Lin, Hazelton,
Quian,1996) For molecular biology studies, the limited number of cells growing out of the
aortic explants under serum-free conditions requires the use of sensitive techniques such as
reverse transcriptase polymerase chain reaction (Nicosia et a!., 1996). There are no reports
about in situ hybridization studies which should, however, be feasible in this system.

5. ADVANTAGES AND LIMITATIONS OF THE RAT AORTA MODEL

The rat aorta model is essentially an ex vivo assay which combines advantages of both in
vivo and in vitro models of angiogenesis. A major advantage is that the aortic cells have not
been modified by repeated passages in culture and generate vascular outgrowths which
strongly resemble and behave like those formed during angiogenesis in vivo. In-viva-like
features ofrat aortic angiogenesis include the formation of microvessels composed of both
endothelial cells and pericytes, the self-limited nature of the angiogenic response, as well as
the regression and remodeling of the neovasculature. Angiogenic factors can be added to the
growth medium while extracellular matrix molecules can be incorporated into the collagen
gel. The effect of angiogenesis agonists and antagonists can be evaluated in the abscence of
serum molecules which may otherwise bind, inactivate or simulate the action of the
substances being tested Testing is not complicated by inflammatory reactions, which may
affect the interpretation of in vivo assays . Angiogenesis can be quantitated reproducibly by
either visual counts or image analysis. Finally, the use of animals is minimized since many
assays can be prepared from one aorta.

51
Like any other model of angiogenesis, the rat aorta model has some limitations. Direct
quantitation of angiogenesis from the living cultures becomes difficult when the outgrowth
contains more than 250 microvessels. This occurs, for example, when aortic explants are
cultured in plasma clot in the presence of fetal bovine serum. The margin of error for the
observer becomes too high due to the three-dimensional complexity of the vascular network
and the masking effect of the surrounding fibroblasts. Quantitation in these cases is
accomplished by processing the cultures for histologic studies and by counting the number
of microvessels in sections taken from the central portion of the gels (Nicosia and Ottinetti,
1990b). The lack of antibodies against rat proteins represents a limiting factor for certain
studies. As more private laboratories and biotechnology companies enter the field of
angiogenesis, this limitation may be overcome by the availability of a greater selection of
rat-specific reagents.

6. ACKNOWLEDGEMENTS

I gratefully acknowledge the many colleagues and co-workers who have contributed to the
studies reviewed in this chapter. This work was supported by NllI Grant ~52585.

7. REFERENCES

Elsdale, T., and Bard, 1., 1972, Collagen substrata for studies on cell behavior, J. Cell BioI.
54:626-637.

Kawasaki, S., Mori, M., and Awai, M., 1989, Capillary growth of rat aortic segments
cultured in collagen without serum. Acta Pathol. Japonica 39:712-718.

Knedler, A, and Ham, R., 1987, Optimized medium for clonal growth of human
microvascular endothelial cells with minimal serum, In Vitro Cell. Dev. Bioi., 23: 481-491.

Nicosia, R.F., 1997, The rat aorta model of angiogenesis and its applications, in Vascular
Morphogenesis in Vivo, in Vitro and in Sapio, (Little, H. Sage and V. Mironov eds.),
Birkhauser, Cambrige, MA (in press).

Nicosia, RF.and Bonanno, E.,1991, Inhibition of Angiogenesis in Vitro by Arg-Gly-Asp-

Containing Synthetic Peptide, Am. J. Path. 138:829-833,.

Nicosia, R.F. and Ottinetti, A, 1990a, Growth of microvessels in serum-free matrix culture
of rat aorta: a quantitative assay of angiogenesis in vitro. Lab. Invest. 63: 115-122 ..

Nicosia, RF.and Ottinetti, A, 1990b, Modulation of microvascular growth and

morphogenesis by reconstituted basement membrane-like gel in three-dimensional culture
of rat aorta: a comparative study of angiogenesis in matrigel, collagen, fibrin, and plasma
clot. In Vitro Cel(Dev. BioI. 26:119-128.

Nicosia, RF. and Tuszynski, G., 1994, Matrix-bound thrombospondin promotes

angiogenesis in vitro. J. Cell Bioi. 124:183-193.

52
Nicosia, R.F., Villaschi, S., 1995, Rat aortic smooth muscle cells become pericytes during
angiogenesis in vitro, Lab. Invest. 73:658-666

Nicosia, R.F., Tchao, R., and Leighton, J., 1982, Histotypic angiogenesIs III vitro: light
microscopic, ultrastructural, and radioautographic studies, In Vitro 18: 538-549.

Nicosia, RF., Bonanno, E., and Smith, M., 1993, Fibronectin Promotes the Elongation of
Microvessels During Angiogenesis in Vitro. J. Cell. Physiol. 154:654-661,.

Villaschi, S. and Nicosia, RF., 1993, Angiogenic role of basic fibroblast growth factor
released by rat aorta after injury, Am. J. Pathol. 143:182-190,.

Nicosia, R. F., Nicosia S. V. and Smith, M, 1994a, Vascular endothelial growth factor,
platelet derived growth factor, and insulin-like growth factor-1 promote rat aortic
angiogenesis in vitro, Am. J. Pathol. 145:1023-1029

Nicosia, R.F., Bonanno, E., and Yurchenco, P., 1994, Modulation of angiogenesis in vitro
by laminin-entactin complex, Dev. BioI. 164:197-206.

Nicosia, R.F., Lin YJ., Hazelton D, and Quian. x., 1996, Role of vascular endothelial
growth factor in the rat aorta model of angiogenesis, J. Vase. Res. 33 (S 1): 73

Nissanov, J., Tuman, RW., Gruver, L.M., and Fortunato, J.M, 1995, Automatic vessel
segmentation and quantification of the rat aortic ring assay of angiogenesis. Lab. Invest.
73:734-739.

53
ASSESSMENT OF ANGIOGENESIS BY MRI

Michal Neeman

Department of Biological Regulation

The Weizmann Institute of Science
Rehovot, 76100 Israel

1. INTRODUCTION- THE BASIS FOR ANALYSIS OF ANGIOGENESIS BY MRI

Analysis of vascular remodeling in vivo is a major challenge both for the study of the
regulatory mechanisms of its initiation and inhibition, and for clinical evaluation of
pathological processes. Magnetic resonance imaging provides an attractive approach for
non invasive analysis of angiogenesis. The aim of this review will be to introduce a
number of specific MRI experiments that can provide information on angiogenesis, and
these will be demonstrated through a few examples. The scope of this manuscript does
not allow a thorough introduction of the principles of magnetic resonance imaging.
Discussion will be further limited to the MRI signal that arises from the nuclear spin of the
water hydrogens, as used in conventional clinical MRI. It is important to mention that
while the MRI images are two dimensional, MRI is inherently a volumetric technique, and
three dimensional data sets can be reconstructed either from a set of two dimensional slices
or from a true isotropic 3D measurement. The non invasive nature ofMRI allows for time
course kinetic studies. The spatial resolution of MRI is limited primarily by the inherent
insensitivity of nuclear magnetic resonance, and in vivo also by motion. Resolution of I
mm is easily achieved in the clinic and 0.1 mm in research systems, in small laboratory
animals. This resolution is not sufficient for resolving individual capillaries. However, the
presence and characteristics of the vasculature can be inferred from the signal intensity
(image contrast) in specific types of MRI experiments. The chapter will be divided into
two major parts. Section 2 will review the use of exogenous MRI contrast agents that are
injected systemically, usually intra venously. Section 3 will introduce a number of
intrinsic contrast mechanisms that do not require the injection of any tracer.

2. EXOGENOUS CONTRAST ENHANCEMENT:

The use of contrast agents is widely spread in medical imaging including MRI.
Gadolinium-DTPA is the most commonly used drug for clinical MRI contrast
enhancement, but a large number of additional agents are being developed. The interaction

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 55
between the nuclear spin of water hydrogens and the local magnetic fields in the vicinity of
the contrast agent, affects the longitudinal (T 1) and transverse (T2) relaxation of water.
Follow-up studies show the kinetics of uptake and clearance of the agent and can be
analyzed to derive the perfusion parameters and the permeability of the contrast agent into
the extra vascular interstitial space.

2.1. Gd-DTPA.
The accumulation of Gd-DTPA in tumors, resulting in significant MRI signal
enhancement, led to its wide spread clinical application. The increased permeability of
tumor neovasculature, the relatively large interstitial space (low cellularity) as well as the
presence of angiogenic 'hotspots' with extremely high vessel density, all contribute to
specific patterns of signal enhancement in tumors. Detailed kinetic analysis using a
pharmacological model of contrast agent uptake and clearance was applied for example in
the study of angiogenic activity induced in the periphery of necrotic regions in MCF7
breast cancer tumors implanted in nude mice (Furman-Haran et aI., 1996). Using this
analysis, the permeability of the contrast agent and the fraction of leakage space (i.e. the
interstitial space) were mapped in the rapidly enhancing regions within the tumor, .
While it is clear now that Gd-DTPA enhancement in clinical MRI reflects angiogenic
activity, an important reservation for the use of Gd-DTPA as a research tool for the
assessment of angiogenesis is the fact that this tracer is neither freely diffusible nor
impermeable thus complicating the analysis. On one hand Gd- DTPA diffuses through
normal as well as tumor vasculature wall and therefore is not a good marker of increased
permeability, while on the other hand, it does not completely equilibrate with the
interstitial space as rapidly as water, and thus does not measure total perfusion accurately.

2.2. Macromolecular contrast agents- albumin Gd-DTPA.

A better evaluation of vascular permeability and plasma volume can be achieved using
contrast agents that have a molecular weight similar to serum proteins (van Dijke et aI.,
1996). Analysis of rats with a subcutaneously implanted R3230 mammary carcinoma
showed highly significant correlation (r2 == 0.85) between MRI derived tumor plasma
volume and permeability and histologic capillary density (van Dijke et aI., 1996).
Permeability was found to correlate with angiogenesis, and treatment of tumors with anti
VEGF neutralizing antibodies resulted in reduced permeability (Brasch et al., 1997).

2.3 Differentiation between tumor and normal capillaries based on their relative
diameter.
Tumor microvessels have a diameter 2 fold larger than normal (11.3 and 6.3 mm
respectively). The relaxation of water is affected by the presence of blood vessels that
contain paramagnetic nuclei, either intrinsic, such as deoxy haemoglobin, or exogenous
such as Gd-DTP A or other contrast agents. Spin echo sequences sensitive to T2 relaxation
are maximally sensitive to normal capillaries, and are less sensitive to larger vessels while
GE sequences (T2*) are equally sensitive to all vessels (Boxerman et aI., 1995). Therefore
the changes in vascular morphology associated with tumor angiogenesis can be determined
from the differences in the two relaxation mechanisms (Dennie et aI., 1997). Intravascular
equilibrium iron oxide contrast agent (MION) was injected to a glioma model in Fisher
rats. T2 and T2* relaxation were measured at low spatial resolution, at different doses of
the contrast agent. Significant differentiation was found between the relaxation rates in the
tumor and in the contralateral normal gray matter in the brain, which were consistent with
a two fold increase in the mean capillary diameter.

56
2.4. Targeted contrast agents for avb3.
The methods outlined above as well as in the following sections rely on the functional
characteristics of neovasculature for vessel detection. Molecular targeting of contrast for
detection of angiogenesis was demonstrated in the study of Sipkins et al (Sipkins et aI.,
1997). This study employed intravascular paramagnetic polymerized liposomes (PPLs) as
stable recirculating contrast agents (Storrs et aI., 1995). These liposomes were conjugated
with biotinylated antibodies against the endothelial integrin receptor avb3 and were
administered to a rabbit model of squamous cell carcinoma (Sipkins et al., 1997). Images
acquired 24 h after administration of the contrast agent showed specific labeling of tumor
neovasculature including enhancement of angiogenic 'hot spots'.

2.5. Advantages and disadvantages of exogenous contrast enhancement

The inj ection of a vascular tracer provides the advantage of internal precontrast
background images, that can be subtracted from the post contrast images to improve
sensitivity. The temporal pharmaco kinetics of the contrast agent is a source of important
physiological information on the structure and properties of the blood vessels. The maj or
disadvantage of this methodology is the need for tracer injection. Obviously this limits the
possibility of repetitive scans in clinical follow-up. Approval for clinical use is not yet
available for many of the newly developed contrast agents, thus limiting their application.
For basic research, where kinetic mapping of vascular remodeling occurs over a period of a
few days, the need for intravenous injections and relatively long scan times can become a
severe experimental limitation on the number of animals and the frequency at which they
can be studied.

3. INTRINSIC MRI CONTRAST OF THE VASCULAR SYSTEM.

Intrinsic contrast techniques for following the vascular system by MRI rely on the physical
properties of blood, such as the paramagnetic properties of deoxyhaemoglobin, as well as
on the dynamics of translational water motion due to blood flow.

3.1 Detection of vessel density through the effect of deoxyhaemoglobin.

The first contrast mechanism that will be discussed here relies on the paramagnetic
properties of deoxyhaemoglobin. The presence of vessels containing deoxygenated blood
reduces the signal intensity of water in the periphery of the vessels in gradient echo MRI
due to the shortened T2* relaxation time (Ogawa et aI., 1990a; Ogawa et aI., 1990b). The
main application of this methodology has so far been for the study of changes in blood
oxygenation associated with cognitive brain function (,functional MRI'). The extent of
signal loss due to the presence of blood vessels is however also dependent on the partial
volume fraction of blood, and thus can be used to follow changes in vascular density. We
have developed this method for the study of the vascularization of multicellular tumors
spheroids implanted in nude mice as a model of the 'angiogenic switch', namely the
transition of tumors from the avascular to the vascular phase (Abramovitch et aI., 1995).
The main advantage of this MRI experiment is the very short measurement time, which
implies minimal manipulation of the mice. This method provides detailed kinetics of
tumor growth and vascularization. Similar MRI experiments can also be used for the study
of the kinetics of angiogenesis and vessel regression in wound healing and for evaluation
of the activity of antiangiogenic agents.

3.2. Tissue oxygenation and angiogenesis.

Hypoxia is a potent activator of angiogenesis, by inducing the expression of VEGF as well
as by inducing the infiltration of macrophages that can trigger an angiogenic response.

57
The angiogenic process itself increases perfusion thereby restoring nonnoxia. The content
of deoxyhaemoglobin in blood is a product of oxygen supply, that is dependent on blood
flow and thus on vascular density, and on oxygen extraction. Modulation in the oxygen
inhaled results in modulations in signal intensity, and provides information of the
functional properties of the neovasculature. Vasoconstriction due to 100% oxygen, and
vasodilation due to C02 can be different in tumor angiogenesis relative to normal
vasculature due to the high fraction of capillaries that are devoid of pericytes and smooth
muscle cells (Karczmar et aI., 1994; Robinson et aI., 1995).

3.3 Perfusion imaging by steady state saturation transfer (SSST).

This methodology is an example of a family of experiments which rely on magnetic
tagging of either the flowing or the stationary water, by affecting the longitudinal
magnetization in a particular slice. For example, magnetization of spins crossing a main
feeding artery can be saturated using a combination of a magnetic field gradient and a
frequency selective radio-frequency pulse. The influx of saturated spins is then monitored
in the irrigated region. This method provides a completely non invasive measure of tissue
perfusion (Detre et ai., 1994).

3.4. Diffusion and microcirculation: analysis of incoherent water motion.

Saturation transfer experiments require a large coherent motion of blood across the
imaging slice. Microcirculation however tends to be relatively incoherent, and randomly
oriented capillaries will create water motion that can be quite similar to rapid diffusion.
Measurement of water diffusion by MRl can be achieved by tagging the water hydrogen
spins with a position dependent shift of phase. After an experimentally determined
diffusion time D, the mean displacement is measured as a reduction in the signal intensity.
In this experiment, bulk coherent flow creates a signal phase shift rather than loss of signal
intensity, thereby allowing differentiation between blood flow in large vessels and through
the capillary bed. A rapidly diffusing component of water observed only in the presence of
blood flow has been attributed to incoherent microcirculation (Le Bihan, 1995). The more
slowly diffusing components can be differentiated into intra cellular, extracellular and
necrotic or acellular regions in the images (Neeman et al., 1991; Tempel et a!. , 1995).

3.5 Advantages and disadvantages of intrinsic contrast.

The major advantage of the intrinsic contrast methods is the fact that they can be repeated
essentially as frequently as the patient (or experimental animal) can be placed in the MRI
scanner. The disadvantage in some cases is the requirement of specific hardware
modifications, that may not be available yet on all clinical MRI systems, and the low
contrast to noise in some of the experiments, which will imply large errors in the
measurement.

4. SUMMARY

A number of approaches for mapping angiogenesis by MRI were described here. These
different MRI experiments are not mutually exclusive and can be combined in a single
study to provide complementary infonnation on different features of the neovasculature.
As a research tool MRI provides the whole set of kinetic data from each animal thereby
revealing the connection and interdependence of the different processes involved in
vascular modeling, in addition to significantly reducing the number of experimental
animals used in the study. The 3 dimensional non invasive MRI images allow design of
experiments that are less prone to false positive errors and interference by tissue damage as
found in the window chamber preparations. However it is important to note here that even

58
the highest resolution MR microscopy systems do not have the spatial and temporal
resolution obtained by light microscopy in the window chambers. A very important aspect
promoting the application ofMRI for evaluation of angiogenesis is obviously the ability to
apply almost any newly developed method as a clinical diagnostic or prognostic tool.
Here, MRI may provide a non invasive whole volume alternative to histological counting
of microvessel density.

REFERENCES

Abramovitch, R., Meir, G. and Neeman, M., 1995, Neovascularization induced growth of
implanted C6 glioma multicellular spheroids: magnetic resonance microimaging. Cancer-
Res. 55: 1956-1962.

Boxerman, J. L., Hamberg, L. M., Rosen, B. R. and Weisskoff, R. M., 1995, MR contrast
due to intravascular magnetic susceptibility perturbations. Magn-Reson-Med. 34: 555-566.

Brasch, R., Pham, c., Shames, D., Roberts, T., van Dijke, K., van Bruggen, N., Mann, 1.,
Ostrowitzki, S. and Melnyk, 0., 1997, Assessing tumor angiogenesis using
macromolecular MR imaging contrast media. JMRI. 7: 68-74.

Dennie, J., Mandeville, 1. B., Weisskoff, R. M . and Rosen, B. R ., 1997, Quantitative

imaging of tumor angiogenesis, in: ISMRM, Vol. 1. pp. 489: Vancouver.

Detre, 1. A., Zhang, W., Roberts, D. A., Silva, A. c., Williams, D. S., Grandis, D. J.,
Koretsky, A. P. and Leigh, 1. S., 1994, Tissue specific perfusion imaging using arterial
spin labeling. NMR-Biomed. 7: 75-82.

Furman-Haran, E., Margalit, R., Grobgeld, D. and Degani, H., 1996, Dynamic contrast-
enhanced magnetic resonance imaging reveals stress-induced angiogenesis in MCF7
human breast tumors. Proc-Natl-Acad-Sci-U-S-A. 93: 6247-6251.

Karczmar, G. S., River, 1. N., Li, 1., Vijayakumar, S., Goldman, Z. and Lewis, M. Z ., 1994,
Effects of hyperoxia on T2* and resonance frequency weighted magnetic resonance
images of rodent tumours. NMR-Biomed. 7: 3-11.

Le Bihan, D., 1995, Molecular diffusion, tissue microdynamics and microstructure. NMR-
Biomed. 8: 375-386.

Neeman, M., Jarrett, K. A., Sillerud, L. O. and Freyer, 1. P., 1991, Self-diffusion of water
in multicellular spheroids measured by magnetic resonance microimaging. Cancer Res. 51:
4072-4079.

Ogawa, S., Lee, T. M., Kay, A . R. and Tank, D. W., 1990a, Brain magnetic resonance
imaging with contrast dependent on blood oxygenation. Proc-Natl-Acad-Sci-U-S-A. 87:
9868-9872.

Ogawa, S., Lee, T. M., Nayak, A. S. and Glynn, P., 1990b, Oxygenation-sensitive contrast
in magnetic resonance image of rodent brain at high magnetic fields . Magn-Reson-Med.
14: 68-78.

Robinson, S. P., Howe, F. A. and Griffiths, 1. R., 1995, Noninvasive monitoring of

carbogen-induced changes in tumor blood flow and oxygenation by functional magnetic
resonance imaging. Int-J-Radiat-Oncol-Biol-Phys. 33: 855-859.

59
Sipkins, D. A., Kazemi, M., Cheresh, D., Bednarski, M. D. and Li, K. C. P ., 1997,
Detection of tumor angiogenesis in vivo by avb3-targeted magnetic resonance imaging, in:
ISMRM, Vol. l. pp. 334: Vancouver.
Storrs, R. W., Tropper, F. D., Li, H . Y., Song, C. K., Sipkins, D. A., Kuniyoshi, J. K.,
Bednarski, M. D., Strauss, H . W. and Li, K. C., 1995, Paramagnetic polymerized
liposomes as new recirculating MR contrast agents. J-Magn-Reson-Imaging. 5: 719-724.

Tempel, C., Schiffenbauer, Y. S., Meir, G. and Neeman, M., 1995, Modulation of water
diffusion during gonadotropin-induced ovulation: NMR microscopy of the ovarian follicle.
MagnResonMed. 34: 2l3-218.

van Dijke, C. F., Brasch, R. C., Roberts, T. P., Weidner, N., Mathur, A., Shames, D. M.,
Mann, J. S., Demsar, F., Lang, P . and Schwickert, H. C., 1996, Mammary carcinoma
model: correlation of macromolecular contrast-enhanced MR imaging characterizations of
tumor microvasculature and histologic capillary density. Radiology. 198: 813-818.

60
MEASURING INTRA TUMORAL MICROVESSEL DENSITY

Noel Weidner, MD

Department of Pathology; University of California, Box 0102; San

Francisco CA 94143-0102, USA

1. MICROVESSEL DENSITY AND TUMOR AGGRESSIVENESS

Although Brem et al. (Brem, Cotran, Folkman, 1972) were among the first to suggest that
the intensity of intratumoral angiogenesis may correlate with tumor grade and
aggressiveness, the first clear-cut evidence that tumor angiogenesis in a human solid tumor
could predict the probability of metastasis was reported for cutaneous melanoma.
Srivastava et aI. (Srivastava, Laidler, Davies, Horgan, Hughes, 1988) studied the vascularity
of 20 intermediate thickness skin melanomas. Vessels were highlighted and the stained
histologic sections analyzed with an image analysis system. The 10 cases that developed
metastases had a vascular area at the tumor base that was more than twice that seen in the
10 cases without metastases (p=0.025).
In 1990, my colleagues and I asked if the extent of angiogenesis (as measured by a spectrum
of intratumoral microvessel densities) in human breast carcinoma correlated with metastasis
(Weidner, Semple, Welch, Folkman, 1991). Such information might prove valuable in
selecting subsets of breast carcinoma patients for aggressive adjuvant therapies. When the
microvessel counts in a number of invasive breast carcinomas are sorted in ascending order
on a log scale, the spectrum of low to high microvessel densities becomes apparent. The
densities are an evenly distributed continuum, extending from about 10 to 200 microvessels
per 0.74 mm' (200x) field. My colleagues and I examined primary tumor specimens from
randomly selected patients with invasive breast carcinoma (Weidner et aI., 1991).
Hematoxylin and eosin (H&E)-stained sections of the breast tumor were used to choose
one paraffin-embedded tissue block clearly representative of a generous cross-section of the
invasive carcinoma, and one 5-micron-thick section from this block was immunostained for
factor VIII-related antigen/von Willebrand's factor (F8RNvWF) to highlight the endothelial
cells lining the blood vessels. Intratumoral microvessel density was assessed by light
microscopic analysis for areas of the tumor that contained the most capillaries and small
venules (microvessels). Finding these neovascular "hot spots" is critical to accurately
assess a particular tumor's angiogenic potential. This is to be expected since there is

Angiogenesis: Models. Modulators. and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 61
considerable evidence that, like tumor proliferation rate, tumor angiogenesis is
heterogeneous within tumors (Folkman, 1994; Folkman, Watson, Ingber, Hanahan, 1989;
Kandel, Bossy-Wetzel, Radvani, Klagsbum, Folkman, Hanahan, 1991; Weidner et al.,
1991). The technique for identifying neovascular "hot spots" is very similar to that for
finding mitotic "hot spots" for assessing mitotic figure content and is subject to the same
kinds of inter- and intraobserver variability. In our study, sclerotic, hypocellular areas
within tumors and immediately adjacent to benign breast tissue were not considered in
intratumoral microvessel density determinations. Only tumors that produced a high quality
and distinct microvessel immunoperoxidase staining pattern with low background staining
were included in this or subsequent studies. This is very important, because the quality of
immunoperoxidase staining can vary considerably between laboratories, and before
measuring intratumoral microvessel density, high quality immunoperoxidase staining must
be consistently achieved.
Areas of highest neovascularization were found by scanning the tumor sections at low
power (4Ox and IOOx total magnification) and selecting those areas of invasive carcinoma
with the greatest density of distinct F8RA1vWF-staining microvessels. These highly
neovascular areas could occur anywhere within the tumor, but most frequently appeared at
the tumor margins. After the area of greatest neovascularization was identified, individual
microvessels within a 200x field (2Ox objective and lOx ocular, Olympus BH-2 microscope,
0.74 mm 2 per field with the field size measured with an ocular micrometer) were counted.
Any highlighted endothelial cell or endothelial-cell cluster clearly separate from adjacent
microvessels, tumor cells, and other connective-tissue elements was considered a single,
countable microvessel. Even those distinct clusters of brown-staining endothelial cells,
which might be from the same microvessel"snaking" its way in and out of the section, were
considered distinct and countable as separate microvessels. Vessel lumens, although usually
present, were not necessary for a structure to be defined as a microvessel, and red cells
were not used to define a vessel lumen. Results were expressed as the highest number of
microvessels in any single 200x field. An average of multiple fields was not performed.
Invasive breast carcinomas from patients with metastases (either lymph nodal or distant
site) had a mean microvessel count of 101 per 200x field. For those carcinomas from
patients without metastases the corresponding value was 45 per 200x field (p=0.003). We
plotted the percent of patients with metastatic disease in whom a vessel count was carried
out within progressive 33 vessel increments. The plot showed that the incidence of
metastatic disease increased with the number of microvessels, reaching 100% for patients
having invasive carcinomas with> 100 microvessels per 200x field.
To further define the relationship of intratumoral microvessel density to overall and
relapse-free survival and to other reported prognostic indicators in breast carcinoma, a
blinded study of 165 consecutive carcinoma patients was performed using identical
techniques to measure intratumoral microvessel density (Weidner, Folkman, Pozza,
Bevilacqua, Allred, Moore, Meli, Gasparini, 1992). The other prognostic indicators
evaluated were metastasis to axillary lymph nodes, patient age, menopausal status, tumor
size, histologic grade (i.e., Scarff-Bloom-Richardson criteria), peritumoral lymphatic-
vascular invasion (PLVI), flow DNA ploidy analysis, flow S-phase fraction, growth
fraction by Ki-67 binding, c-erbB2 oncoprotein expression, pro-cathepsin-D content,
estrogen-receptor content, progesterone-receptor content, and EGFR expression. We found
a highly significant association of intratumoral microvessel density with overall survival
and relapse-free survival in all patients, including node-negative and node-positive subsets.
All patients with breast carcinomas with more than 100 microvessels per 200x field

62
experienced tumor recurrence within 33 months of diagnosis, compared to less than 5% of
patients who had fewer than 33 microvessels per 200x field. Moreover, intratumoral
microvessel density was the only significant predictor of overall and relapse-free survival
among node-negative women.
Other studies performed on different patient databases by different investigators at
different medical centers have observed this same association of increasing intratumoral
vascularity with various measures of tumor aggressiveness, such as higher stage at
presentation, greater incidence of metastases, and/or decreased patient survival. This has
been shown in studies of patients not only with carcinoma of the breast (Barbareschi,
Gasparini, Weidner, Morelli, Forti, Eccher, Fina, Leonardi, Mauri, Bevilacqua, Dalla Palma,
1995; Bevilacqua, Barbareschi, Verderio, Boracchi, Caffo, Dalla Palma, Meli, Weidner,
Gasparini, 1995; Bosari, Lee, DeLellis, Wiley, Heatley, Silverman, 1992; Charpin,
Devictor, Bergeret, Andrac, Boulat, Horschowski, Lavaut, Piana, 1995; Fox, Leek, Smith,
Hollyer, Greenall, Harris, 1994; Fox, Turner, Leek, Whitehouse, Gatter, Harris, 1995;
Gasparini, Barbareschi, Boracchi, Bevilacqua, Verderio, Dalla Palma, Menard, 1995a;
Gasparini, Barbareschi, Boracchi, Verderio, Caffo, Meli, Palma, Marubini, Bevilacqua,
1995b; Gasparini, Bevilacqua, Boracchi, Maluta, Pozza, Barbareschi, Dalla Palma,
Mezzetti, Harris, 1994a; Gasparini, Weidner, Bevilacqua, Maluta, Dalla Palma, Caffo,
Barbareschi, Boracchi, Marubini, Pozza, 1994; Horak, Leek, Klenk, Lejeune, Smith, Stuart,
Greenall, Stepniewska, Harris, 1992; Inada, Toi, Hoshina, Hayashi, Tominaga, 1995;
Lipponen, Ji, Aaltomaa, Syrjanen, 1994; Obermair, Czerwenka, Kurz, Buxbaum,
Schemper, Seve1a, 1994a; Obermair, Czerwenka, Kurz, Kaider, Seve1da, 1994b; Obermair,
Kurz, Czervenka, Thoma, Kaider, Wagner, Gitsch, Seve1da, 1995; Ogawa, Chung, Nakata,
Takatsuka, Maeda, Sawada, Kato, Yoshikawa, Sakurai, Sowa, 1995; Scopinaro, Schillaci,
Scarpini, Mingazzini, diMacio, Banci, Danieli, Zerilli, Limiti, Centi Colella, 1994; Toi,
Hoshina, Yamamoto, Ishii, Hayashi, Tominaga, 1994; Toi, Inada, Suzuki, Tominaga, 1995;
Toi, Kashitani, Tominaga, 1993; Visscher, Smilanetz, Drozdowicz, Wykes, 1993; Weidner
et aI., 1991; Weidner, et aI., 1992), but in those with prostate carcinoma (Brawer, Deering,
Brown, Preston, Bigler, 1994; Fregene, Khanuja, Noto, Gehani, Van Egmont, Luz, Pienta,
1993; Hall, Troncoso, Pollack, Zhau, Zagars, Chung, von Eschenbach, 1994; Vesalainen,
Lipponen, Talja, Alhava, Syrjanen, 1994; Wakui, Furusato, Itoh, Sasaki, Akiyama,
Kinoshita, Asano, Tokuda, Aizawa, Ushigome, 1992; Weidner, Carroll, Flax, Blumenfeld,
Folkman, 1993), head-and-neck squamous carcinoma (Albo, Granick, Jhala, Atkinson,
Solomon, 1994; Alcalde, Shintani, Yoshihama, Matsumura, 1995; Gasparini, Weidner,
Bevilacqua, Maluta, Boracchi, Testolin, Pozza, Folkman, 1993; Mikami, Tsukuda,
Mochimatsu, Kokatsu, Yago, Sawaki, 1991; Williams, Carlson, Cohen, Derose, Hunter,
Jurkiewicz, 1994), non-small-cell lung carcinoma (Angeletti, Lucchi, Fontanini, Mussi,
Chell a, Ribechini, Vignati, Bevilacqua, 1996; Fontanini, Bigini, Vignati, Basolo, Mussi,
Lucchi, Chine, Angeletti, Bevilacqua, 1995; Macchiarini, Fontanini, Hardin, Hardin,
Squartini, Angeletti, 1992; Macchiarini, Fontanini, Dulmet, de Montpreville, Chapelier,
Cerrin, Le Roy Ladurie, Dartevelle, 1994; Yamazaki, Abe, Takekawa, Sukoh, Watanabe,
Ogura, Nakajima, Isobe, Inoue, Kawakami, 1994; Yuan, Yang, Yu, Yao, Lee, Kuo, Luh,
1995), malignant melanoma (Barnhill, Levy, 1993; Fallowfield, Cook, 1991; Graham,
Rivers, Kerbel, Stankiewicz, White, 1994; Smolle, Soyer, Hofmann-Wellenhof, SmoIle-
Juettner, Kerl, 1989; Srivastava, Laidler, Davies, Horgan, Hughes, 1988; Srivastava, Laidler,
Hughes, Woodcock, Shedden, 1986; Vacca, Ribatti, Roncali, Lospalluti, Serio, Carrel,
Dammacco, 1993), gastrointestinal carcinoma (Saclarides, Speziale, Drab, Szeluga, Rubin,
1994; Maeda, Chung, Takatsuka, Ogawa, Sawada, Yamashito, Onoda, Kato, Nitta,

63
Arimoto, Kondo, Sowa, 1995a; Maeda, Chung, Takatsuka, Ogawa, Onoda, Sawada, Kato,
Nitta, Arimoto, Kondo, Sowa, 1995b; Takebayashi, Akiyama, Yamada, Akiba, Aikou,
1996), testicular genn-cell malignancies (Olivarez, Ulbright, DeRiese, Foster, Reister,
Einhorn, Sledge, 1994), multiple myeloma (Vacca, Ribatti, Roncali, Ranieri, Serio,
Silvestris, Dammacco, 1994), central nervous system tumors (Leon, Folkerth, Black, 1996;
Li, Folkerth, Watanabe, Yu, Rupnick, Barnes, Scott, Black, Sallan, Folkman, 1994), ovarian
carcinoma (Hollingsworth, Kohn, Steinberg, Rothenberg, Merino, 1995; Volm, Koomagi,
Kaufmann, Mattern, Stammler, 1996), cervical squamous carcinoma (Bremer, Tiebosch,
van der Putten, SChouten, deHaan, Arends, 1996; Schlenger, Hockel, Mitze, Schaffer,
Weikel, Knapstein, Lambert, 1995; Wiggins, Granai, Steinhoff, Calabresi, 1995),
endometrial carcinoma (Abulafia, Triest, Sherer, Fansen, Ghezzi, 1995), and transitional-
cell carcinoma of the bladder (Bochner, Cote, Weidner, Groshen, Chen, Skinner, Nichols,
1995; Dickinson, Fox, Persad, Hollyer, Sibley, Harris, 1994; Jaeger, Weidner, Chew,
Moore, Kerschmann, Waldman, Carroll, 1995; Kohno, Iwanari, Kitao, 1993). In many of
these studies, intratumoral vascularity was found to have independent prognostic
significance when compared to traditional prognostic markers by multivariate analysis.
Thus far, tumors originating in the liver, biliary system, or pancreas have not been
extensively studied by these techniques. Nonetheless, Shimoyama et al. (Shimoyama,
Gansauge, Gaunsage, Negri, Oohara, Beger, 1996) found that the expression of angiogenin
in pancreatic carcinoma is related to cancer aggressiveness, Egawa et al. (Egawa, Tsutsumi,
Konishi, Kobari, Matsuni, Nagasaki, Futami, Yamaguchi, 1995) discovered that tumor
angiogenesis plays an important role in growth of a hamster pancreatic cell line, and
multiple groups have reported an association between tumor angiogenesis and a propensity
of gastrointestinal carcinomas to metastasize to the liver (Konno, Tanaka, Matsuda, Kanai,
Maruo, Nishino, Nakamura, Baba, 1995; Maeda, Chung, Ogawa, Takatsuka, Kang, Ogawa,
Sawada, Sowa, 1996; Tomisaki, Ohno, Ichiyoshi, Kuwano, Maehara, Sugimachi, 1996).
Additional work needs to be done.
Although assaying intratumoral microvessel density has been the most common means of
deterining tumor vascularity, some investigators found similar associations with tumor
aggressiveness using image analysis to measure vascular area or doppler ultrasound to
measure blood flow. The optimum technique for assaying intratumoral vascularity has not
been completely defined. Finally, it is important that some studies have found that high
intratumoral microvessel density can be significantly associated with a favorable outcome,
apparently when specific forms of therapy are given wherein therapeutic effectiveness
depends directly on the extent of blood flow (Schlenger, et aI., 1995; Zatterstrom, Brun,
Willen, Kjellen, 1995). Kohno et al. (1993) found that cervical carcinomas treated with
hypertensive intra-arterial chemotherapy are more responsive when highly vascular, and
Zatterstrom et al. (1995) found that highly vascular squamous carcinomas of the head-and-
neck are more responsive to radiation therapy when highly vascular.

2. MEASURING INTRATUMORAL MICROVESSEL DENSITY

When highlighting microvessels, it is important that previously published protocols for

measuring intratumoral microvessel density be followed carefully. Considerable experience
is needed at the senior staff pathologist level for assessing intratumoral microvessel
density. It is necessary not only for supervising the immunostaining of endothelial cells,
but also for selecting generous sections of representative invasive tumor and for localizing

64
the neovascular "hot spot." Counting microvessels has been shown to be reproducible
(Weidner, 1992), especially following a period of training. Brawer et. al. (1994) compared
manual intratumoral microvessel determinations to those determined by automated
counting (i .e., Optimas bnage Analysis) and found a very tight correlation (r=0.98,
p<O.OOI). Proper technique must be supplemented with unbiased case selection and
proper statistical analysis of data. Finally, accurate staging and adequate patient follow-up
are needed to determine those patients who have metastases or will experience tumor
recurrences. These reasons may explain why some investigators have not found the
association of intratumoral microvessel density with prognosis in some solid tumors
(Axelsson, Ljung, Moore, Thor, Chew, Edgerton, Smith, Mayall, 1995; Barnhill, Busam,
Berwick, Blessing, Cochran, Elder, Fandrey, Daraoli, White, 1994; Carnochan, Briggs,
Westbury, Davies, 1991; Costello, McCann, Carney, Dervan, 1995; Dray, Hardin,
Sofferman, 1995; Goulding, Rashid, Robertson, Bell, Elston, Blarney, Ellis, 1995; Hall,
Fish, Hunt, Goldin, Guillou, Monson, 1992; Kainz, Speiser, Wanner, Obermair, Tempfer,
Sliutz, Reinthaller, Breitenecker, 1995; Leedy, Trune, Kronz, Weidner, Cohen, 1994;
MacLennan, Bostwick, 1995; Miliaras, Kamas, Kalekou, 1995; Rutger, Mattox, Vargas,
1995; Siitonen, Haapasalo, Rantala, Helin, Isola, 1995; Tahan, Stein, 1995; Van Hoef,
Knox, Dhesi, Howell, Schor, 1993; van Diest, Zevering, Zevering, Baak, 1995). Why these
reports are contradictory is not clear; however, work by Leedy et al. (1994) may provide a
clue. They implied that tumor growth and spread might be facilitated by preexisting vessels
in highly vascular organs such as the tongue, liver, skin, kidney, or gastrointestinal tract. If
such were the case, intratumoral microvessel density would not be so useful in predicting
outcome. Clearly, it should not be of value in solid tumor systems in which there is no
spectrum from low to high intratumoral microvessel density (MacLennan and Bostwick,
1995). Paradoxically, Kainz et al. (1995) found that patients with cervical cancers showing
relatively low microvessel density had a significantly poorer recurrence-free survival than
those with higher densities.

3. HIGHLIGHTING MICROVESSELS FOR MICROVESSEL COUNTING

Although no endothelial marker is perfect, when applied properly, anti-F8RA/vWF

remains the most specific, providing very good contrast between microvessels and other
tissue components. Unfortunately, anti-F8RAJvWF may not highlight all intratumoral
microvessels. Although apparently more sensitive, CD31 strongly cross reacts with plasma
cells (DeYoung, Wick, Fitzgibbon, Sirgi, Swanson, 1993; Longacre TA, Rouse RV, 1994).
This complication can markedly obscure the microvessels in those tumors with a
prominent plasma cellular inflammatory background. CD34 is an acceptable alternative and
the most reproducible endothelial-cell highlighter in many laboratories, but CD34 will
highlight perivascular stromal cells and has been noted to stain a wide variety of stromal
neoplasms (Traweek, Kandalaft, Mehta, Batti fora, 1991; van de Rijn, Rouse, 1994). Like
antibodies to F8RAJvWF, antibodies to CD31, CD34, and PAL-E also do not immunostain
all intratumoral microvessels (Schlingemann, Rietveld, Kwaspen, van de Kerkhof, de Waal,
Ruiter, 1991). Ulex europeus lectin will stain many tumor cells, seriously decreasing its
specificity; it is not recommended. Wang et al. (Wang, Kumar, Pye, van Agthoven,
Krupinski, Hunter, 1993; Wang, Kumar, Pye, Haboubi, AI-Nakib, 1994) have developed a
monoclonal antibody (Mab E9), which was raised against proliferating or "activated"
endothelial cells of human umbilical-vein origin and grown in tissue culture. Mab E9

65
strongly reacted with endothelial cells of all tumors, fetal organs, and in regenerating and/or
inflamed tissues, but it only rarely and weakly immunostained endothelial cells of normal
tissues. Unfortunately, in their study, Mab E9 immunoreacted only in frozen tissue
sections; however, they did not mention if microwave antigen-retrieval techniques applied
to formalin-fixed, paraffin-embedded tissues were tried. Antibodies such as Mab E9 may
provide the most sensitive staining of intratumoral microvessels and preferentially
immunostain activated or proliferating endothelial cells such that the overall staining
intensity may correlate best with the intensity of tumor angiogenesis and, hence, tumor
aggressiveness. Staining for the endothelial activation marker _v_3 may also yield very
useful data for clinicopathologic correlations (Brooks, Montgomery, Rosenfeld, Reisfeld,
Hu, Ilier, Cheresh, 1994). Automated ("machine") immunostaining and application of
computer-aided image analysis may help to standardize microvessel counts and help
eliminate interobserver and even intraobserver variables, such as inexperience and "hot
spot" selection biases (Barbareschi et aI., 1995). The latter approach may make
determination of intratumoral microvessel density a simple, reliable, and reproducible
prognostic factor in a variety of solid tumors, not just in breast carcinoma.
Other methods may prove even more reliable and reproducible. These include measuring
levels of angiogenic molecules in serum or urine, or directly measuring angiogenic molecules
or inhibitors from tumor extracts (i.e., in a manner similar to hormone receptor assays).
Indeed, using an immunoassay, Watanabe et al. (Watanabe, Nguyen, Schizer, 1992) and
Nguyen et al. (Nguyen, Watanabe, Budson, 1994) reported elevated levels of bFGF in the
serum and urine of patients with a wide variety of solid tumors, including breast carcinoma.
Higher levels were found in patients with metastatic disease than in those with localized
disease. Moreover, Li et al. (Li, Folkerth, Watanabe, Yu, Rupnick, Barnes, Scott, Black,
Sallan, Folkman, 1994) measured bFGF in the cerebral spinal fluids of children with various
brain tumors and correlated increasing fluid levels with greater intratumoral microvessel
density and increased likelihood of recurrence. As a final note in this section, it is important
to keep in mind that immunohistochemistry of any given angiogenic factor (ie., bFGF or
VPFNEGF) in a tumor would not, by itself, be expected to correlate with microvessel
count (or with patient outcome), because the microvessel count is the sum total of positive
and negative angiogenic factors, and most likely more than one of each. Finally, magnetic
resonance imaging (cMRI) enhanced with special contrast agents will allow assessment of
angiogenesis in vivo; a technique that will allow monitoring the effects of therapy
(Esserman, Hylton, George, Yassa, Weidner, submitted).

REFERENCES

Abulafia, 0., Triest, W. E., Sherer, D. M., Hansen, C. C., Ghezzi, F., 1995, Angiogenesis
in endometrial hyperplasia and stage I endometrial carcinoma, Obstet. Gynecol. 86:479-485.

Albo, D., Granick, M. S., Ihala, N., Atkinson, B., Solomon, M. P., 1994, The relationship
of angiogenesis to biological activity in human squamous cell carcinomas of the head and
neck, Ann. Plast. Surg. 32:588-594.

Alcalde, R. E., Shintani, S., Yoshihama, Y., Matsumura, T., 1995, Cell proliferation and
tumor angiogenesis in oral squamous cell carcinomas, Anticancer Res. 15:1417-1422.

66
Angeletti, C. A, Lucchi, M., Fontanini, G., Mussi, A, Chell a, A., Ribechini, A., Vignati,
S., Bevilacqua, G., 1996, Prognostic significance of tumoral angiogenesis in completely
resected late stage lung carcinoma (stage IIIA-N2). Impact of adjuvant therapies in a
subset of patients at high risk of recurrence, Cancer 78:409-15.

Axelsson, K., Ljung, B. M. E., Moore, D . H., Thor, A D., Chew, K. L., Edgerton, S. M.,
Smith, H. S., Mayall, B. H., 1995, Tumor angiogenesis as a prognostic assay for invasive
ductal carcinoma, J. Natl. Cancer Inst. 87:997-1008.

Barbareschi, M ., Gasparini, G., Weidner, N., Morelli, L., Forti, S., Eccher, c., Fina, P.,
Leonardi, E., Mauri, F. , Bevilacqua, P., Dalla Palma, P., 1995, Microvessel density
quantification in breast carcinomas: assessment by manual vs. a computer-assisted image
analysis system, Appl. Immunohistochem. 3:75-84 .

Barnhill, R. L., Levy, M. A, 1993, Regressing thin cutaneous malignant melanomas «1.0
mm) are associated with angiogenesis, Am. J. Pathol. 143:99-104.

Barnhill. R. L. , Busam, K. 1, Berwick, M., Blessing, K., Cochran, A 1, Elder, D. E.,

Fandrey, K., Daraoli, T., White, W.L., 1994, Tumor vascularity is not a prognostic factor
for cutaneous melanoma [letter], Lancet 344: 1237-1238.

Bevilacqua, P., Barbareschi, M., Verderio, P., Boracchi, P., CaITo, 0., Dalla Palma, P.,
Meli s, Weidner, N., Gasparini, G., 1995, Prognostic value of intratumoral microvessel
density, a measure of tumor angiogenesis, in node-negative breast carcinoma - results of a
multiparametric study, Breast Cancer Res. Treat. 36:205-217.

Bochner, B. H., Cote, R. 1, Weidner, N., Groshen, S., Chen, S-c., Skinner, D. G., Nichols,
P. W., 1995, Angiogenesis in bladder cancer: relationship between microvessel density and
tumor prognosis, J. Natl.Cancer Inst. 87: 1603-1612.

Bosari, S., Lee, A K., DeLellis, R. A, Wiley, B. D., Heatley, G.1, Silverman, M. L. ,1992,
Microvessel quantitation and prognosis in invasive breast carcinoma, Hum. Pathol. 23:755-
761.

Brawer, M. K., Deering, R. E., Brown, M., Preston, S. D., Bigler, S. A. 1994, Predictors of
pathologic stage in prostate carcinoma, Cancer 73:678-687.

Brem, S., Cotran, R., Folkman, 1, 1972, Tumor angiogenesis: a quantitative method for
histologic grading, J. Natl. Cancer Inst. 48:347-356.

Bremer, G. L., Tiebosch, A T. M. G., van der Putten, H. W. H. M., Schouten, H. 1 A., de
Haan, 1, Arends, J-W., 1996, Tumor angiogenesis: an independent prognostic parameter in
cervical cancer, Am. J. Obstet. Gynecol. 174(1 Pt 1):126-31.

Brooks, P. C., Montgomery, A M. P, Rosenfeld, M ., Reisfeld, R. A, Hu, T., llier, G.,

Cheresh, D. A, 1994,. Integrin av~3 antagonists promote tumor regression by inducing
apoptosis of angiogenic blood vessels, Cell 79: 1157-1164.

67
Carnochan, P., Briggs, J. C., Westbury, G., Davies, A J., 1991, The vascularity of
cutaneous melanoma: a quantitative histologic study oflesions 0.85-1.25 mm in thickness,
Brit. J. Cancer 64: 102-107.

Charpin, C., Devictor, B., Bergeret, D., Andrac, L., Boulat, J., Horschowski, N., Lavaut,
M. N., Piana, L., 1995, CD31 quantitative immunocytochemical assays in breast
carcinomas. Correlation with current prognostic factors, Am. J. Clin. Pathol. 103:443-448.

Costello, P., McCann, A, Carney, D . N., Dervan, P. A, 1995, Prognostic significance of

microvessel density in lymph node negative breast carcinoma, Hum. Pathol. 26: 1181-1184.

DeYoung, B. R., Wick, M. R., Fitzgibbon,J. F., Sirgi, K E., Swanson, P. E., 1993, CD3l:
An immunospecific marker for endothelial differentiation in human neoplasms, Appl.
Immunohistochem. 1:97-100.

Dickinson, A J., Fox, S. B., Persad, R. A, Hollyer, J., Sibley, G. N., Harris, A L., 1994,
Quantification of angiogenesis as an independent predictor of prognosis in invasive bladder
carcinomas,Brit. J. Urol. 74:762-766.

Dray, T. G., Hardin, N . J., Sofferman, R. A, 1995, Angiogenesis as a prognostic marker in

early head and neck cancer, Ann. Otol. Rhino!. Laryngol. 104:724-729.

Egawa, S., Tsutsumi, M., Konishi, Y, Kobari, M., Matsuni, S., Nagasaki, K, Futami, H.,
Yamaguchi, K, 1995, The role of angiogenesis in the tumor growth of syrian hamster
pancreatic cancer cell line HPD-NR,Gastroenterology 108:1536-1533.

Esserman, L., Hylton, N., George, T., Yassa, L., Weidner, N., (submitted), Constrast-
enhanced magnetic resonance imaging (cMRI) provides a window to visualize anatomic
extent and tumor angiogenesis in breast carcinoma, Cancer Res.

Fallowfield, M. E., Cook, M. G., 1991, The vascularity of primary cutaneous melanoma, J.
Pathol. 164:241- 244.

Folkman, J., Watson, K, Ingber, D., Hanahan, D., 1989, Induction of angiogenesis during
the transition from hyperplasia to neoplasia, Nature 339:58-61 .

Folkman,J., 1994, Angiogenesis and breast cancer, J. Clin. Oncol. 12:441-443.

Fontanini, G., Bigini, D ., Vignati, S., Basolo, F., Mussi, A., Lucchi, M., Chine, S.,
Angeletti, C. A, Bevilacqua, G., 1995, Recurrence and death in non small cell lung
carcinomas: a prognostic model using pathological parameters, microvessel count, and gene
protein products, J. Pathol. 177:57-63.

Fox, S. B., Leek R. D ., Smith, K, Hollyer, J., Greenall, M., Harris, A. L., 1994, Tumor
angiogenesis in node-negative breast carcinomas: relationship with epidermal growth factor
receptor, estrogen receptor, and survival, Breast Cancer Res. Treat. 29: 109-116.

Fox, S. B., Stuart, N., Smith, K, Brunner, N., Harris, A. L., 1993, High levels of uP A and
PA-l are associated with highly angiogenic breast carcinomas, J. Pathol. 170 (Suppl):388a.

68
Fregene, T A, Khanuja,P. S., Noto, A c., Gehani, S. K, Van Egmont, E. M., Luz, D. A,
Pienta, K 1.,1993, Tumor- associated angiogenesis in prostate cancer, Anticancer Res.
13:2377-2381.

Gasparini, G., Weidner, N., Bevilacqua, P., Maluta, S., Boracchi, P., Testolin, A., Pozza,
F., Folkman, 1.,1993, Intratumoral microvessel density and p53 protein: correlation with
metastasis in head-and-neck squamous-cell carcinoma, Int. J. Cancer 55:739-744.

Gasparini, G., Bevilacqua, P., Boracchi, P., Mal uta, S., Pozza, F., Barbareschi, M., Dalla
Palma, P., Mezzett" M., Harris, A L, 1994a, Prognostic value of p53 expression in early-
stage breast carcinoma compared with tumour angiogenesis, epidermal growth factor
receptor, c-erbB2, cathepsin D, DNA ploidy, parameters of cell kinetics and conventional
features, Int. J. Oncol. 4:155-162.

Gasparini, G., Weidner, N, Bevilacqua, P., Maluta, S., Dalla Palma, P., CaITo, 0,
Barbareschi, M., Boracchi, P., Marubini, E., Pozza, F., 1994b, Tumor microvessel density,
p53 expression, tumor size, and peritumoral lymphatic vessel invasion are relevant
prognostic markers in node-negative breast carcinoma, J. Clin. Oncol. 12:454-466.

Gasparini., G., Barbareschi, M., Boracchi, P., Bevilacqua, P., Verderio, P., Dalla Palma, P.,
Menard, S., 1995a, 67-kDa laminin-receptor expression adds prognostic information to
intra-tumoral microvessel density in node-negative breast cancer, Int. J. Cancer 60:604-610.

Gasparini, G., Barbareschi, M., Boracchi, P., Verderio, P., CaITo, 0., Meli, S., Palma, P. D.,
Marubini, E., Bevilacqua, P., 1995b, Tumor angiogenesis may predict clinical outcome of
node-positive breast cancer patients treated either with adjuvant hormone therapy or
chemotherapy, Cancer J. Sci. Am. 1: 131-141.

Goulding, H., Rashid, N. F. N. A, Robertson, J. F., Bell, 1. A, Elston, C. W., Blarney, R.

W., Ellis, I. 0., 1995, Assessment of angiogenesis in breast carcinoma: An important factor
in prognosis? Hum. Pathol. 26: 1196-1200.

Graham, C. H., Rivers, J, Kerbel, R S., Stankiewicz, K S., White, W. L., 1994, Extent of
vascularization as a prognostic indicator in thin «0.76 mm) malignant melanomas, Am. J.
Pathol.145:510-514

Hall, N. R, Fish, D. E., Hunt, N., Goldin, R. D., Guillou, P. J., Monson, 1. R. T, 1992, Is
the relationship between angiogenesis and metastasis in breast cancer real? Surg Oncol
1:223-229.

Hall, M. c., Troncoso, P., Pollack, A., Zhau , H. Y, Zagars, G. K, Chung, L. W., von
Eschenbach, A c., 1994, Significance of tumor angiogenesis in clinically localized prostate
carcinoma treated with external beam radiotherapy, Urology 44:869-875.

Hollingsworth, H. c., Kohn, E. C., Steinberg, S. M., Rothenberg, M. L., Merino, M. 1.,
1995, Tumor angiogenesis in advanced stage ovarian carcinoma, Am. J. Pathol. 147:33-41.

69
Horak, E., Leek, R, Klenk, N., Lejeune, S., Smith, K., Stuart, N., Greenall, M.,
Stepniewska, K., Harris, A L.,1992, Angiogenesis, assessed by platelet/endothelial cell
adhesion molecule antibodies, as indicator of node metastases and survival in breast cancer,
Lancet 340:1120-1124.

Inada, K., Toi, M., Hoshina, S., Hayashi, K., Tominaga, T., 1995, Significance of tumor
angiogenesis as an independent prognostic factor in axillary node-negative breast cancer,
Gan. To. Kagaku Jap. J. Cancer Chemo. 22 (suppl 1):59-65.

Jaeger, T. M., Weidner, N., Chew, K., Moore, D. H., Kerschmann, R. L., Waldman, F. M.,
Carroll, P. R, 1995, Tumor angiogenesis correlates with lymph node metastases in
invasive bladder cancer. J. Urol. 154:59-71.

Kainz, c., Speiser, P., Wanner, c., Obermair, A, Tempfer, c., Sliutz, G., Reinthaller, A,
Breitenecker, G., 1995, Prognostic value of tumor microvessel density in cancer of the
uterine cervix stage IB to lIB, Anticancer Res. 15: 1549-1551.

Kandel, 1., Bossy-Wetzel, E., Radvani, F., Klagsburn, M., Folkman, 1., Hanahan, D., 1991,
Neovascularization is associated with a switch to the export of bFGF in the multi-step
development of fibrosarcoma, Cell 66: 1095-1104.

Kohno, Y, Iwanari, 0., Kitao, M., 1993, Prognostic importance of histologic vascular
density in cervical cancer treated with hypertensive intraarterial chemotherapy, Cancer
72:2394-2400.

Konno, H., Tanaka, T., Matsuda, I., Kanai, T., Maruo, Y, Nishino, N., Nakamura, S.,
Baba, S.,1995, Comparison of the inhibitory effect of the angiogenesis inhibitor TNP-470
and mitomycin C on the growth and liver metastasis of human colon cancer, Int. J. Cancer
61:268-271,

Leedy, D. A, Trune, D. R., Kronz, 1. D., Weidner, N., Cohen, 1. I., 1994, Tumor
angiogenesis, the p53 antigen, and cervical metastasis in squamous carcinoma, Otolaryngol.
Head Neck Surg 111:417-422.

Leon, S. P., Folkerth, R D., Black, P. M., 1996, Microvessel density IS a prognostic
indicator for patients with astroglial brain tumors, Cancer 77(2):362-72.

Li, V. W., Folkerth, R. D., Watanabe, H., Yu, c., Rupnick, M., Barnes, P., Scott, R. M.,
Black, P. M., Sallan, S.E., Folkman, 1., 1994, Microvessel count and cerebrospinal fluid
basic fibroblast growth factor in children with brain tumors, Lancet 334:82-86.

Lipponen, P., Ji, H., Aaltomaa, S., Syrjanen, K., 1994, Tumor vascularity and basement
membrane structure in breast cancer as related to tumor histology and prognosis, J. Cancer
Res. Clin. Oneol. 120:645-650.

Longacre, T. A, Rouse, R. v., 1994, CD31: A new marker for vascular neoplasia, Adv.
Anat. Pathol. 1:16-20.

70
Macchiarini, P, Fontanini, G., Hardin, M. J., Hardin, M. J., Squartini, F., Angeletti, C. A.,
1992, Relation of neovasculature to metastasis of non-small-cell lung cancer, Lancet
340: 145-146.

Macchiarini, P, Fontanini, G., Dulmet, E., de Montpreville, V., Chapelier, AR., Cerrin, J,
Le Roy Ladurie, F., Dartevelle, P.G., 1994, Angiogenesis: an indicator of metastasis in
non-small-cell lung cancer invading the thoracic inlet. Ann Thorac Surg 57: 1534-1539.

MacLennan, G. T., Bostwick, D . G.1995, Microvessel density in renal cell carcinoma: lack
of prognostic significance, Urology 46:27-30.

Maeda, K., Chung, Y.-S., Takatsuka, S., Ogawa, Y., Onoda, N., Sawada, T., Kato, Y.,
Nitta, A., Arimoto, Y., Kondo, Y., Sowa, M., 1995a, Tumor angiogenesis and tumor cell
proliferation as prognostic indicators in gastric carcinoma, Br. J Cancer 72:319-323.

Maeda, K., Chung, Y-S., Takatsuka, S., Ogawa, Y., Sawada, T. , Yamashito, Y., Onoda, N.,
Kato, Y., Nitta, A, Arimoto, Y., Kondo, Y., Sowa, M., 1995b, Tumor angiogenesis as a
predictor of recurrence in gastric carcinoma, J Clin. Oncol. 13:477-481.

Maeda, K., Chung ,y'-S., Ogawa, Y., Takatsuka, S., Kang, S, Ogawa, M ., Sawada, T.,
Sowa, M., 1996, Prognostic value of vascular endothelial growth factor expression in
gastric carcinoma, Cancer 77:858-863.

Mikami, Y., Tsukuda, M., Mochimatsu, I., Kokatsu, T., Yago, T. , Sawaki, S., 1991,
Angiogenesis in head and neck tumors, Nip. Jib. Gak. Kai. 96:645-50.

Miliaras, D., Kamas, A., Kalekou, H., 1995, Angiogenesis in invasive breast carcinoma: is it
associated with parameters of prognostic significance? Histopathology 26: 165-169.

Nguyen, M., Watanabe, H., Budson, A. E., Richie, J. P., Folkman, J, 1993, Elevated levels
of the angiogenic peptide basic fibroblast growth factor in urine of bladder cancer patients,
J Natl. Cancer Instil. 85:241-242.

Obermair, A, Czerwenka, K., Kurz, c., Buxbaum, P., Schemper, M., Sevela, P., 1994a,
Influence of tumoral microvessel density on the recurrence-free survival in human breast
cancer: preliminary results, Onkologie 17:44-49.

Obermair, A, Czerwenka, K., Kurz, c., Kaider, A., Sevelda, P., 1994b, Tumor vascular
density in breast tumors and their effect on recurrence free survival, Chirurg. 65:611-615.

Obermair, A, Kurz, c., Czervenka, K., Thoma, M., Kaider, A, Wagner, T., Gitsch , G ,
Sevelda, P, 1996,. Microvessel density and vessel invasion inlymph node negative breast
cancer: effect on recurrence-free survival, Int. J Cancer 62:126-130.

Ogawa, Y., Chung, Y-S ., Nakata, E, Takatsuka, S., Maeda, K., Sawada ,T., Kato ,Y.,
Yoshikawa, K., Sakurai, M., Sowa, M. , 1995, Microvessel quantitation in invasive beast
cancer by staining for factor VIII-related antigen, Brit. J Cancer 71: 1297-1301.

71
Olivarez, D., Ulbright, T., DeRiese, w.,Foster, R., Reister, T., Einhorn, L., Sledge, G.,
1994, Neovascularization in clinical stage A testicular geIm cell tumor: prediction of
metastatic disease, Cancer Res 54 :2800-2802.

Rutger, 1. L., Mattox, T. F., Vargas, M. P., 1995, Angiogenesis In uterine cervical
squamous cell carcinoma,/nt. J. Gynecol. Pathol. 14:114-118.

Saclarides, T. 1., Speziale,N. J., Drab, E., Szeluga, D. J., Rubin, D. B., 1994, Tumor
angiogenesis and rectal carcinoma, Dis. Colon Rectum 37:921 -926.

Schlenger, K., Hockel, M., Mitze, M., Schaffer, u., Weikel, W ., Knapstein, P . G., Lambert,
A., 1995, Tumor vascularity -- A novel prognostic factor in advanced cervical carcinoma,
Gynecol. Oncol. 59 :57-66.

Schlingemann, R. 0 ., Rietveld, F. J. R., Kwaspen, F., van de Kerkhof, P. C. M ., de Waal,

R. M . W ., Ruiter, D. 1., 1991, Differential expression of markers for endothelial cells,
pericytes, and basal lamina in the microvasculature of tumors and granulation tissue, Am. J.
Pathol. 138: 1335-1347.

Scopinaro, F., Schillaci, 0., Scarpini, M ., Mingazzini, P. L., diMacio, L., Banci, M.,
Danieli, R., Zerilli, M., Limiti, M. R., CentiColella, A., 1994, Technetium-99m sestamibi:
an indicator of breast cancer invasiveness, Eur. J. Nucl. Med. 21 :984-987.

Shimoyama, S., Gansauge, F., Gansauge, S., Negri, G., Oohara, T. , Beger, H. G ., 1996,
Increased angiogenin expression in pancreatic cancer is related to cancer aggressiveness,
Cancer Res. 56: 2703-2706.

Siitonen, S. M., Haapasalo, H. K., Rantala, I. S., Helin, H. 1., Isola, 1. 1., 1995, Comparison
of different immunohistochemical methods in the assessment of angiogenesis: lack of
prognostic value in a group of 77 selected node-negative breast carcinomas, Mod. Pathol.
8:745-752.

Smolle, 1., Soyer, H-P., Hofmann-Wellenhof, R ., Smolle-Juettner, F. M., Kerl, H., 1989,
Vascular architecture of melanocytic skin tumors. A quantitative immunohistochemical
study using automated image analysis, Pathol. Res. Pract. 185:740-745.

Srivastava, A., Laidler, P., Hughes, L. E., Woodcock, 1., Shedden, E. J., 1986,
Neovascularization in human cutaneous melanoma: a quantitative morphological and
Doppler ultrasound study, Eur. J. Cancer Clin. Oncol. 22: 1205-1209.

Srivastava, A., Laidler, P., Davies, R., Horgan, K., Hughes, L. E., 1988, The prognostic
significance of tumor vascularity in intermediate-thickness (0.76-4.0 mm thick) skin
melanoma, Am. J. Pathol. 133:419-23 .

Tahan, S. R ., Stein, A. L., 1995, Angiogenesis in invasive squamous cell carcinoma of the
lip: tumor vascularity is not an indicator of metastatic risk, J. Cutan. Pathol. 22:236-240.

Takebayashi, Y., Akiyama, S.-I., Yamada, K., Akiba, S., Aikou, T., 1996, Angiogenesis as
an unfavorable prognostic factor in human colorectal carcinoma, Cancer 78:226-31.

72
Toi, M., Kashitani, J., Tominaga ,T., 1993, Tumor angiogenesis is an independent
prognostic indicator of primary breast carcinoma, Int. 1. Cancer 55:371-374.

Toi, M., Hoshina, S., Yamamoto, Y., Ishii, T., Hayashi, K., Tominaga, T.,1994, Tumor
angiogenesis in breast carcinoma: significance of vessel density as a prognostic indicator,
Gan. To. Kagaku Ryoho Japanese 1. Cancer Chemo. 21 (Suppl 2): 178-182.

Toi, M., Inada, K., Suzuki, H., Tominaga, T., 1995, Tumor angiogenesis in breast cancer:
its importance as a prognostic indicator and the association with vascular endothelial
growth factor expression, Breast Cancer Res. Treat. 36: 193-204.

Tomisaki, S., Ohno, S., Ichiyoshi, Y., Kuwano, H., Maehara, Y., Sugimachi, K., 1996,
Microvessel quantification and its possible relation with liver metastasis in colorectal
cancer, Cancer (Suppl), 77: 1722-1728.

Traweek, S. T., Kandalaft, P. L., Mehta, P., Battifora, H., 1991, The human hematopoietic
progenitor cell antigen (CD34) in vascular neoplasia, Am. 1. Clin. Pathol. 96:25-31.

Vacca, A., Ribatti, D., Roncali, L., Lospalluti, M., Serio, G., Carrel, S., Dammacco, F.,
1993, Melanocyte tumor progression is associated with changes in angiogenesis and
expression of the 67-kilodalton laminin receptor, Cancer 72:455-461.

Vacca, A., Ribatti, D., Roncali, L., Ranieri, G., Serio, G., Silvestris, F., Dammacco, F.,
1994, Bone marrow angiogenesis and progression in multiple myeloma, Brit. .J. Haematol.
87:503-508.

Van Hoef, M. E. H. M., Knox, W. F., Dhesi, S. S., Howell, A., Schor, A. M., 1993,
Assessment of tumor vascularity as a prognostic factor in lymph node negative invasive
breast cancer, Eur. 1. Cancer 29A:1141-1145.

van de Rijn, M., Rouse, R.Y., 1994, CD34: A review, Appl. Immunohistochem. 2:71-80.

van Diest, P. J., Zevering, J. P., Zevering, L. C., Baak, 1. P. A., 1995, Prognostic value of
microvessel quantitation in cisplatin treated Figo 3 and 4 ovarian cancer patients, Pathol.
Res. Pract. 191 :25-30.

Vesalainen, S., Lipponen, P., Talja, M., Alhava, E., Sytjanen, K., 1994, Tumor vascularity
and basement membrane structure as prognostic factors in TI-2MO prostatic
adenocarcinoma, Anticancer Res. 14:709-714.

Visscher, D. W., Smilanetz, S., Drozdowicz, S., Wykes,S. M., 1993, Prognostic
significance of image morphometric microvessel enumeration in breast carcinoma, Anal.
Quant. Cytol. Histo!. 15:88-92.

Volm, M., Koomagi, R., Kaufmann, M., Mattern, J., Stammler, G., 1996, Microvessel
density, expression of proto- oncogenes, and resistance-related proteins and incidence of
metastases in primary ovarian carcinomas,Clin. Exp. Met. 14:209-14 ..

73
Wakui, S., Furusato, M., !toh, T., Sasaki, H., Akiyama, A., Kinoshita, I, Asano, K.,
Tokuda, T., Aizawa S., Ushigome, S., 1992, Tumor angiogenesis in prostatic carcinoma
with and without bone marrow metastasis: a morphometric study, J. Patho!. 168:257-262.

Wang, 1. M., Kumar, S., Pye, D., Haboubi, N., Al-Nakib, L., 1994, Breast carcinoma:
comparative study of tumor vasculature using two endothelial-cell markers, J. Natl. Cancer
Inst. 86:386-388.

Wang, 1. M., Kumar, S., Pye, D., van Agthoven, A. 1., Krupinski, 1., Hunter, R. D., 1993,
A monoclonal antibody detects heterogeneity in vascular endothelium of tumors and
normal tissues, Int. J. Cancer 54:363-370.

Watanabe, I. I., Nguyen, M., Schizer, M., 1992, Basic fibroblast growth factor in human
serum -- a prognostic test for breast cancer, Mol. BioI. Cell. 3:324a.

Weidner, N., 1992, The relationship of tumor angiogenesis and metastasis with emphasis
on invasive breast carcinoma, in: Advances in Pathology and Laboratory Medicine, Volume
5 (R.S. Weinstein RS, ed), pp. 101-121, Mosby Year Book, Chicago.

Weidner, N., Semple, J. P., Welch, W. R., Folkman, J., 1991, Tumor angiogenesis and
metastasis - correlation in invasive breast carcinoma, N. Eng!. J. Med 324: 1-8.

Weidner, N., Folkman, J., Pozza, F., Bevilacqua, P, Allred, E. N, Moore, D. H., Meli S.,
Gasparini G., 1992, Tumor angiogenesis: a new significant and independent prognostic
indicator in early-stage breast carcinoma, J. Natl.Cancer Inst. 84: 1875-1887.

Weidner, N., Carroll, P. R., Flax, 1., Blumenfeld, w., Folkman, J., 1993, Tumor
angiogenesis correlates with metastasis in invasive prostate carcinoma, Am. J.
Pathol.143:40 1-409.

Wiggins, D. L., Granai, e.O., Steinhoff, M. M., Ca1abresi, P., 1995, Tumor angiogenesis as
a prognostic factor in cervical carcinoma, Gynecol. Oncol. 56:353-356.

Williams, 1. K, Carlson, G. W., Cohen, C., Derose, P. B., Hunter, S., Jurkiewicz, M. J.,
1994, Tumor angiogenesis as a prognostic factor in oral cavity tumors, Am. J. Surg.
168:373-380.

Yamazaki, K., Abe, S., Takekawa, H., Sukoh, N., Watanabe, N., Ogura, S., Nakajima, I.,
Isobe, H., Inoue, K, Kawakami, Y., 1994, Tumor angiogenesis in human lung
adenocarcinoma, Cancer 74:2245-2250.

Yuan, A., Yang, p-e., Yu, C-1., Yao, Y-T., Lee, Y-C., Kuo, S-H., Luh, K-T., 1995, Tumor
angiogenesis correlates with histologic type and nodal metastasis in non-small cell lung
carcinoma, Am. J. Resp. Critical Care Med Dec. 152(6 Pt 1): 2157-62.

Zatterstrom, U. K., Brun, E., Willen, R., Kjellen, E., Wennerberg, J., 1995, Tumor
angiogenesis and prognosis in squamous cell carcinoma of the head and neck, Head & Neck
17:312-318.

74
PATTERNS OF PHYSIOLOGICAL ANGIOGENESIS IN ADULT MESENTERY

F Hansen-Smith and Laura Morris

Department of Biological Sciences

Oakland University
Rochester, Michigan, 48309-4401 USA

The structural process by which new microvessels form from pre-existing ones has been
documented in classical studies of wound healing. (Folkman and Shing, 1992; Hudlicka and
Tyler, 1986). However, the process by which new microvessels originate de novo and
continue to expand under physiological conditions is less well understood. The
physiological formation and subsequent growth of a new vascular bed is normally
considered to be something that occurs predominantly during embryonic development. In
mammalian systems this puts certain restrictions on the types of studies which can be
done on a developing vascular bed to understand how the growth is regulated. Most
tissues in the adult have little ongoing angiogenesis, with the exception of the female
reproductive system. However, it has been found that the rat mesentery undergoes a
spontaneous angiogenesis, leading to the formation of a two-dimensional microvascular bed
which seems to be well suited for studies of angiogenesis under either physiological or
pathological conditions (Hansen-Smith, Joswiak, and Baustert, 1994; Norrby, Jakobsson,
and Sorbo, 1990; Rhodin and Fujita, 1989). Since this growth occurs in adults, many types
of physiological studies, as well as structural and biochemical studies, are feasible to help
understand not only how the initiation of angiogenesis and angiostasis are regulated, but
also how microvascular patterns arise and are remodelled during the enlargement of
microvascular networks.

The majority of in vivo and in vitro studies of angiogenesis have used a simple method of
quantitation ie, vascular density, in vivo; cell migration or proliferation, in vitro( Hudlicka
and Tyler, 1986, D' Amore, 1992; D' Amore and Smith, 1993). However, in order to form
a coherent physiological pattern within a microvascular network, there must be an order to
the growth ofvessels--ie, if all cells migrated or proliferated at the same rate, the outcome
would presumably be a sheet of solid cells. Yet, in vivo, sprouting, branching and
anastamosis must occur, as must differentiation of vascular segments and selective growth
within regions needing new or more blood vessels, whether pathological or physiological

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 75
 5 -r------ ---------~

*
4

Q;
.0
E 3
::>
Z
Q;
<J)
<J)
(1)
>c
III
2
(1)

--,
:2

- - - - --- - - - - -
o -+------,---~---~

o 2 4 6 8 10
Window Number

Figure 1: Mean density of capillaries per window in mesenteries offemale rats ranging in
age from 6-8 weeks (squares) to 18-24 weeks (circles). Capillary internodes
segments were counted from nine different samples sites per window.
Samples were taken from the first, fifth, and ninth, or last, loop of bowel.
Rhodamine GSI lectin used to visualize the microvessels. * = p<O.OS relative
to the first window.

(Colville-Nash and Willoughby, 1997; Folkman and Shing, 1992). An understanding of the
regulation of angiogenesis--beyond whether changes in vascular density occur-- may be
extremely useful in understanding the basic process, as well as in designing treatments to
promote or inhibit the process. We used the mesentery as a model of physiological
angiogenesis to evaluate what types of patterning during angiogenesis can feasibly be
studied in the intact tissue. Here we report several types of growth patterns we have
identified which may be of significant value in further understanding questions about
angiogenesis at the inter-cellular and tissue level and which are difficult, if not impossible,
to study using either in vitro angiognesis models or three-dimensional tissues.

METHODS

Female Sprague-Dawley rats in two age ranges (6-8 weeks) and (18-24 weeks) were used
for this study, with n=6 per group. Mesenteric windows were isolated from the small
intestine after euthanasia of the rats with CO2. Two windows each were selected from the
jejunum and the proximal and distal regions of the ilium and lightly fixed in a flattened
sheet, using 4% formalin. Mter rinsing with physiological saline, the windows were
trimmed of most of the adipose tissue and immersed overnight in the tetramethyl
rhodamine derivative of the Griffonia simplieiJolia (GSI) lectin (Vector Laboratories,
Burlingame, CA), as previously described to delineate microvessels (Hansen-Smith,
Watson, Lu, and Goldstein, 1988). The windows were mounted on microscope slides after
extensive rinsing with physiological saline.

76
 4
1- .- .._ _._. - - - -- 1
. *
3
Q;
.c
E
OJ
Z

3l
(J)
2
Q)
>
c
m
Q)
::;;

o
Intestinal Portal

Figure 2: Density of capillaries on the intestinal side vs. portal side of mesenteric
windows, using rhodamine GSI lectin to visualize microvessels. Solid bars: 6-8
wks; open bars: 18-24 weeks. * = p<0.05 relative to all other groups.

The density and patterns of capillary angiogenesis in the windows were evaluated by
epifluorescence (Nikon Optiphot). To assess microvascular density, the total number of
microvessels was determined in 800um 2 rectangular fields at pre-determined sites. Three
equidistant sites were selected along each side of the triangular borders of the window,
giving a total of 9 sampling sites per window. Statistical analysis was conducted using the
Student-Newman-Keuls a posteriori test, with statistical significance at p < 0.05

RESULTS

The density of capillaries per sample site per window, as assessed by the number of
microvessel segments, was significantly greater in the older than in the younger rats (Fig.
1). In addition, in the adult rats there appeared to be an assymetry to the distribution of
the capillaries, with more capillaries along the intestinal side, vs. the two portal sides.
Consequently the data were sub-divided and re-analyzed according to regions, ie, intestinal
side (3 sites) vs. portal side (6 sites). There were approximately 3-4 times more vessels
per site on the intestinal side relative to the portal side in both ages of animals. (Fig. 2) On
the intestinal side the density of vessels in the adults was significantly greater than that of
the juveniles, whereas the differences seen between adult and juvenile on the portal side did
not reach statistical significance. There were qualitative differences in the organization of
the microvessels at the two different sites as well. Even in the mature rats, the capillary
network on the portal sides was much simpler than that along the intestinal side.
As indicated in Fig. 1, there was a significant difference in the extent of angiogenesis found
in the younger vs. the more mature rats. Microvessels growing into the windows of the
younger rats formed simpler networks, often with only 2-3 branches(Fig. 3). Most of the

77
Figure 3: GSI-Iabelled capillaries on the intestinal side of a mesenteric window from an 8-
week rat. Sprouts (arrows) growing from the adipose (A) tissue sheath are
simple, only occasionally forming closed loops.

vessels ended in sprout tips rather than forming a closed loop which would provide return
circulation from this network. Thus, the microvessels in the younger rats were at early
stages of angiogenesis. In contrast, the microvascular networks in the mature rats were
very complex, often extending more centripitally than the field used for quantitation. In
some cases the microvessels appeared to "meander" in a yarn-like manner (Fig. 4). Many

Figure 4: GSI-Iabelled capillaries (arrow) and lymphatic vessels (double arrows) in the
central region of the mesenteries from an I8-week rat. In this region both the
capillaries and the more faintly staining lymphatic vessels are "meandering".
Capillary sprouts are pointed, whereas lymphatics are club-shaped.

78
Figure 5: GSI-labeJled capiJlaries forming an arterio-venous arcade (A,V) in the mid-
region of a mesenteric window from a 22-week rat. Both closed capiJlary loops
and open sprouts are seen. Note that capillary sprouts often appear to be
progressing toward closure of a loop (arrows).

such microvessels ended as free-ending sprouts, but in many other cases, the microvessels
formed a closed loop (Fig. 5). Some, but not all, of the microvascular networks examined
had progressed to the formation of clearly recognizable arcades which contained arterioles,
capillaries, and venules. Among these arcades, however, free-ending sprouts were still
commonly observed. Sprouts in these location were often in a curved shape, as if
progressing toward closure of a loop (Fig. 5).

Focusing through the tissue by standard fluorescence techniques established that most of
the networks within the clear area of the windows arose from a small number of sprouts
from the deeper feed vessels rather than the closer and seemingly more obvious source, ie,
the capillaries of the perivascular adipose tissue. Sprouts of lymphatic vessels were also
seen in some windows (Fig. 4). These were "club-like", and irregular in their branching
pattern. While some windows contained only capiJlaries, others contained only
lymphatics, and stiJl others contained both.

A second source of sprouting capillaries was seen in some windows which were
undergoing division by the formation of a dividing artery, vein, and, often, a nerve.
Branches of the intestinal arteriole and portal venule crossed the clear area, forming a
"bridge". This A-V unit was invariably accompanied by an anastamosing group of
capillaries, from which new sprouts sometimes formed (Fig. 6) The latter type of vascular
unit w~s normally found on the portal sides of the window and was much more common in
the older rats. Those "bridges" which were more established were partially populated by
adipose tissue, like the surrounding walls of the established window.

Two cell types in the mesenteric windows were GSI-positive in addition to the
microvessels: a larger cell (> 15 urn) was granular in appearance and has previously been

79
Figure 6: Arterio-venous "bridge" accompanied by capillaries has grown across a
mesenteric window from a 24-week rat and forms the basis of two new
"daughter" windows (A,B). Capillary sprouts may also originate from these
regions (arrows).

positively identified by us as a mast cell (Fig. 7, Hansen-Smith, unpublished). Smaller oval

cells (6-8 urn) have been tentatively identifed as immature mast cells, based on mast cell
immunostaining. Neither fibroblasts nor macrophages are GSI-positive in the mesentery.
Both types of mast cells were found throughout the mesentery, whether angiogenesis was
present or not. In angiogenic areas, in contrast to non-angiogenic areas, the mast cells were
often clustered in groups of 4-6. Partially degranulated mast cells were often found near the
tips of sprouts or at branch points in the developing capillary network (Fig. 7).

DISCUSSION

In our approach to employing the mesentery as a model of angiogenesis we rely on the GSI
lectin to delineate the microvessels, usually using a fluorescent conjugate, although other
conjugates may also be used. One of the great strengths of the GSI lectin method
introduced by our laboratory (Hansen-Smith, et ai, 1988), is that the all of the
microvessels, including even lymphatics, are visualized in their natural pattern and
capillary sprouts are easily detected and followed to their tips. The GSI lectin binds to a-
galctosyl pyrranosyl moities, and competition with the hapten sugar, a-D-galactose, blocks
the GSI binding to capillaries. An additional attribute of this method is that since the GSI
lectin binding sites are associated primarily with the ablumenal membrane and/or basement
membranes, perfusion of the marker is unnecessary. Hence, the method reveals numerous
capillary sprouts which might have been missed by a perfused marker, and, additionally,
lymphatics which would not have been detected at all. GSI binding to arterioles is stronger
than that to venules, so their identification is possible without intravital studies of flow
direction (Hansen-Smith, et aI, 1988).

80
Figure 7: High magnification field of capillary sprout tips shows two peri-capillary cell
types: the granular GSI-positive mast cells (*) and smaller, intensely GSI-
positive cell (arrows) tentatively identified as immature mast cells.

The two-dimensional pattern in which capillaries sprout into the clear mesenteric windows
makes this microvascular bed an attractive candidate for studying the mechanisms
regulating the initiation of physiological angiogenesis and also the factors regulating the
coordinated growth of an integrated microvascular network. It is also appealing for use as a
bioassay system somewhat analagous to the way the chick choriaoallantoic membrane has
been used for angiogenesis studies. This vascular bed has already been used extensively in
this manner by Norrby, et al (1989, 90,92,94), although the predominant application has
been in the context of a mast-cell induced angiogenesis (using a degranulating agent), which
may be more pathological than physiological. Early studies employed an assay system
which samples a relatively small number of microvessels in cross sections of tissues
embedded in plastic, giving limited information on the growth of the vessels (Norrby, et ai,
1986). Norrby and colleagues have extended their vessel identification to include image
analysis of toluidine-blue-stained or ink-perfused networks (Norrby, Jakobsson, and
Sorbo, 1990; Norrby, 1995). This method enables an estimate of vascularized area of the
mesenteric windows, a measurement which is difficult using the fluorescent technique used
here.

Our laboratory is interested in the determinants of patterning during angiogenesis.

Although the literature relating to angiogenesis is extensive, there are actually relatively
few studies focusing on the issue of how angiogenesis leads to the establishment of either
normal or abnormal microvascular patterns. Pathological angiogenesis almost invariably has
an abnormal pattern for the tissue in which it occurs. Intraperitoneal injections of different
types of cancer cells induce angiogenesis in the mesentery, but the pattern and extent of
angiogenesis varies with the type of tumor cell (Nagy, et al, 1995; Yanagi and Oshima,
1996). Leakiness and bleeding of mesenteric vessels are common under these
circumstances, whereas the physiological angiogenesis does not appear to give rise to leaky

81
vessels. In addition, certain test treatments appear to overcome the intrinsic patterns of
angiogenesis found here (Heuser, Taylor, and Folkman, 1984; Nagy, Morgan, Herzberg, et
al, 1995; Yanagi and Oshima, 1996) The mesentery model of spontaneous angiogenesis
may be useful in identifying factors which are critical to the "correct" patterning of
networks during angiogenesis. This study has identified several types of patterns: 1.) age-
related increase in angiogenesis: 2) proximal-distal increase in angiogenesis along the length
of the intestinal tube; 3.) locus-dependent sprouting--ie, portal vs. intestinal sides; 4.)
types of vessels from which sprouting occurs--ie, "deep" arterioles or venules, or
lymphatics, rather than the more extensive perivascular capillary plexus within the
adipose tissue sheath; 5.) tendency to form closed capillary loops in more angiogenically
mature regions. Another pattern of angiogenesis which has been identifed in the mesentery
is a predictable temporal pattern (in response to mast cell secretion), as reported by
lakobsson, 1994.

Our finding of age-related and proximal-distal differences in angiogenesis raises obvious

questions regarding whether these phenomena are a function of increased angiogenic stimuli
or are instead a function of a release from angiostatic factors. Further studies probing for
established autocrine and paracrine growth factors and inhibitors will be required to clarify
these concerns. Both the proximal-distal differences and the portal vs intestinal side
differences could be indicative of metabolic or hormonal influences within the digestive
system which could promote or inhibit angiogenesis in the different regions. Furthermore,
factors such as mechanical stretch (Hudlicka and Tyler, 1986) on the intestinal side may
favor angiogenesis under the same conditions as the less angiogenic portal regions, in which
stretch is not a factor. Hemodynamic influences on angiogenesis (Hudlicka and Tyler,
1986) may well play a role in the patterning associated with the locus-dependent
sprouting from the portal vs. intestinal side. Hemodynamic factors may also function in
determining the sites from which sprouting occurs along the arterioles and venules, vs. the
more abundant capillary plexus around them (Zweifach, 1973). While the hemodynamics
and other physiological parameters of the mesenteric arteries and portal vessels have been
extensively studied for physiological and pharmacological reasons, those associated with
the intestinal side remain poorly characterized (Unthank, Lash, and Bohlen, 1990).

One of the least appreciated aspects of the mesentery model is that unique questions may
be explored pertaining to the perivascular millieu around sprouts themselves, as well as
areas in which anasamoses occur. This level of interaction has been difficult to study using
other models of in vivo angiogenesis due to the difficulties in detecting the actual regions of
sprouting. Sprouts are readily detected in the mesenteric windows using the GSI lectin.
Thus, one of the most important strengths of this model is that the spatial relationship
between capillary sprouts and the greater microvascular network can be appreciated. The
model can be used in double-labelling studies to identify interstitial cells, such as
fibroblasts, or capillary pericytes associated with tips of capillary sprouts (Nehls and
Drenckhaahn, 1991). Rhodin and Fujita (1989 ) used intravital microscopy and electron
microscopy to study the cellular components of sprout tips.

Macrophages and mast cells are present throughout the mesentery as well as fibroblasts.
Our studies using the GSI lectin, in combination with antibodies to mast cells have shown
an intimate relationship between mast cells and sprouting areas and branch points. How the
mast cells are actually interacting with the growing microvessels remains to be completely
determined. However, using different techniques, Norrby, et al (1992, 1994), have

82
identifed histamine and heparin as likely mast-cell derived factors in regulating
angiogenesis. The presence of mast cells does not in itself ordain angiogenesis to occur.
Indeed, there are mast cells distributed throughout the mesenteric window, yet only certain
regions sprout. The question of how mast cells promote angiogenesis remains unsolved,
but the mesenteric model is certainly a likely candidate for further study in this regard. We
postulate that mast cells, possibly along with the fibroblasts and macrophages in the
perivascular environment are key players in the regulation of sprouting, formation of
anastamoses, as well as in directing the patterning of microvessel growth during
angiogenesis. Further study using the mesentery model will elucidate how these
interactions occur under physiological conditions.

REFERENCES

Colville-Nash, P.R. and Wiloughby, D.A., 1997, Growth factors in angiogeneis: current
interest and therapeutic potential. Mol. Med. Today, 14-23.

D' Amore, P., 1992, Mechanisms of endothelial growth control, Am . J. Resp. Cell. Mol.
BioI. 6: 1-8.

D'Amore, P., and Smith, S.R, 1993, Growth factor effects on cells of the vascular wall: a
survey. Growth Factors 8: 61-75, 1993.

Folkman, J. and Shing, Y, 1992, Angiogenesis. 1. BioI. Chern. 267: 10931-34.

Hansen-Smith, F.M., 1988, Fluorescent delineation of microvessels in thin muscle

preparations and mesentery. FASEB J. 2, A1189

Hansen-Smith, F.M. , Watson, L., Du, D., and Goldstein, I., 1988, Griffonia simplicijolia I:
Fluorescent tracer for microcirculatory vessels in non-perfused thin muscles and sectioned
muscle. Microvasc. Res. 36: 199-215.

Heuser, L., Taylor, S., Folkman, J. , 1984, Prevention of carcinomatosis and blood
malignant ascites in the rat by an inhibitor of angiogenesis. 1. Surg. Res. 36: 244-50.

Hudlicka, 0., and Tyler, KR. Angiogenesis. New York, Academic Press, 1986.

lakobsson, A., 1994, Angiogenesis induced by mast cell secretion in rat peritoneal
connective tissue is a process of three phases. Microvasc. Res . 47: 252-69.

Nagy, J., Morgan, E. , Herzberg, K, Manseau, E., Dvorak, A., and Dvorak, H., 1995,
Pathogenesis of ascites tumor growth: angiogenesis, vascular remodeling, and stroma
formation in the peritoneal lining. Cancer Res. 55: 376-85.

Nehls, V. and and Drenckhahn, D., 1991, Heterogeneity of microvascular pericytes for
smooth muscle type alpha-actin. J. Cell Bioi 113: 147-54.

Norrby, K and Sorbo, J., 1992, Heparin enhances angiogenesis by a systemic mode of
action. Int. 1. Exp. Pathol. 73: 147-155.

83
 Norrby, K., Jakobsson, A., and Sorbo, J., 1986, Mast-cell mediated angiogenesis: a novel
experimental model using the rat mesetery. Virchow's Arch. B. Cell Pathol. 52: 195-206.

Norrby, K., Jakobsson, A., and Sorbo, J., 1989, Mast-cell secretion and angiogenesis: a
quantitative study in rats and mice. Virchows. Arch. 57: 251-56.

Norrby, K., Jakobsson, A., and Sorbo, J., 1990, Quantitative angiogenesis in spreads of
intact rat mesenteric windows. Microvasc. Res. 39: 341-348.

Rhodin, J.A. G. and Fujita, H., 1989, Capillary growth in the mesentery of normal young
rats: Intravital video and electron microscope analyses. J. Submicrosc. Cytol. Pathol. 21:
1-34.

Sorbo, J., Jakobsson, A., and Norrby, K, 1994, Mast-cell histamine is angiogenic through
receptors for histamine! and histamin~. Int. J. Exp. Path. 75: 43-50.

Unthank, J., Lash, J., and Bohlen, H., 1990, Maturation of the rat intestinal
microvasculature from juvenile to early adult life. Am. J. Physiol. 259: G282-289.

Yanagi, K. and Oshima, N., 1996, Angiogenic vascular growth in the rat peritoneal
disseminated tumor model. Microvasc. Res. 51: 15-28.

Zweifach, B., 1973, The microcirculation in the intestinal mesentery. Microvasc. Res. 5:
363-367.

84
ANGIOGENESIS - CRITICAL ASSESSMENT OF IN-VITRO ASSAYS AND IN -
VIVO MODELS

David BenEzra

Dept. of Ophthalmology
Hadassah University Hospital
Jerusalem, Israel

Angiogenesis or neovascularization is a complex in-vivo phenomenon. Originally

postulated more than 50 years ago by Michaelson, the growth of new blood vessels (either
from stem cells during embryogenesis or from pre-existing mature vessels during wound
healing processes) is most probably stimulated by released chemical factors I Michaelson
alluded to the existence of an x-factor of neovascularization and studied its expression in
experimental models and clinical conditions 1,2 Attempts to isolate the Michaelson x-factor
have led to the discovery of a myriad of "angiogenic factors"3-6..

Based on extensive experimental observations (in vitro and in vivo) as well as on clinical
observations of embryological healing and aging processes within the eye, I had suggested
in 1978 that in-vivo angiogenesis is most probably a step-ladder complex of phenomena4
(figure 1). This unique in-vivo process could be triggered via various pathways involving
many factors. The production of these factors lead to a cascade of events ending in the
sprouting of new capillaries and the fomlation of a new (and patent) vessel network.
Following the accumulation of additional knowledge regarding angiogenesis, an amended
scheme for the cascade of events is suggested (figure2).

The essential biological events associated with the sprouting of new blood vessels during
healing and/or aging processes (angiogenesis) in vivo are:
1. "Loosening" of intercellular adhesions between the vascular endothelial cells and
degradation of basement membrane elements. This process is carried out by the
activation of pre-existing proteases and/or their de novo synthesis. These activated
proteases degrade the "adult" vessel basement membrane and the tight intercellular
adhesions. At the same time, the natural contact inhibition is overcome and the
vascular endothelial cells are "ready to respond" to exogenous stimuli.

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenwn Press, New York, 1998 85
 Specific
cells?

(NAF)

Figure 1. Illustrates the original concept regarding the possible pathways and cascade of
events during angiogenic stimuli leading to the growth of new blood vessels. (D.
BenEzra. Neovasculogenic ability of prostaglandins, growth factors and
synthetic chemoattractants. Am. J. Ophthalmol. 86:455-461, 1978.)

Integrins

IFNs, ECM

TGF p, bFGF,
Il-1 , IL-6, IL-8

Figure 2. Scheme of angiogenic pathways and cascade of events amended from Figure 1.
Within the various pathways, factors involved in the cascade of events have been
introduced based on our present knowledge of the processes.

86
2. Directed taxis of vascular endothelial cells towards the site of highest concentration of
chemotactic factors released by the cells of the affected tissue.
3. Mitogenesis of vascular endothelial cells and/or dedifferentiation of other "accessory
cells".
4. Lumen formation with re-establishment of the intercellular adhesions. In many
cases however, the lumen of these new vessels is not tightly "sealed". Therefore,
these vessels are more "leaky" and prone to more rapid extravasation of fluids and to
hemorrhages.
5. Circulation of blood within the new vessels with the formation of appropriate
anastomoses and gradients between the arterial and venous channels.

The above processes are tightly regulated in vivo. Probably, the thresholds of stimulation
are specific for each tissue. Once a threshold of stimulation is reached and the process
initiated, its regulation is canied out by characteristic feedback mechanisms typical to the
tissue involved and to the type of triggering stimulus. Therefore, it is conceivable to
assume that not all factors able to induce basement membrane degradation or chemotaxis
and/or mitogenesis of vascular endothelial cells are necessarily "angiogneic factors". For
example, a specific substance with chemotactic and/or mitogenic activity on isolated
vascular endothelial cells in vitro may not be able to induce these activities in vivo if an
earlier process of basement membrane degradation had not been initiated earlier. Moreover,
even if a degradation of the basement membrane and "loosening" of the vascular endothelial
cells had been initiated earlier, these tested factors may not overcome the regulatory and
inhibitory processes taking place in vivo. Thus, these substances, although demonstrating
chemotactic and mitogenic activities on vascular endothelial cells in vitro may not be active
angiogenic factors in vivo. The peptide f-Meth-Leu-Phe is a strong chemoattractant for
polymorphonuclear cells in vitro but has no angiogenic activity in viv04 On the other
hand, the bacterial lipopolysaccharide (LPS) is not mitogenic to endothelial cells in vitro
but is a strong angiogenic stimulant in vivo'- More importantly, the fact that substances
have a potential inhibitory effect on chemotactic and/or mitogenic activity (or on tube
formation, or matrigel invasion) in vitro does not necessarily indicate that they are
"inhibitors of angiogenesis" in vivo. Excellent examples for these are the transforming
growth factor beta (TGFI3) and the interferons alpha and gamma (IFNa and IFNy). Both
have an inhibitory effect on proliferation (and chemotaxis) of endothelial cells in vitro, but
promote angiogenesis in viv06 Also, ref,'Ulatory feedback mechanisms with possible
synergistic and antagonist activities occurring in a different milieu offactors in vivo have to
be considered7

Furthermore, it is not inconceivable to assume that substances with no evident chemotactic

and/or no mitogenic activities in vitro, may playa most crucial role in angiogenesis in vivo.
This role can be carried out (for example) by their ability to counteract the regulatory (and
inhibitory) mechanisms which prevent the activation of crucial pathways leading to the
formation of new blood vessels in vivo. Alternatively, these types of substances may
directly activate crucial processes unassociated with mitogenic or chemotactic activities but
are necessary for the triggering of the cascade of events leading to angiogenesis in vivo.

Therefore, when evaluating in vitro systems we have to take in consideration the following:
1. Observations and findings have to be kept in their proper context of isolated artificial
systems.

87
2. Extrapolation to in-vivo situation should be avoided and if alluded to, it has to be
made with great care and ample reservations.
3. The use of the term angiogenesis (or neovascularization) for in-vitro systems should
be avoided as this is a strict in-vivo complex cascade of events.

Because of the complexity of angiogenesis as an in-vivo phenomenon, only reliable and

reproducible in vivo models can be relevant for the study of angiogenesis. However, even
with the most adequate in vivo models we have to be aware of the following potential
pitfalls:
1. Exogenous materials (factors) ectopically placed may express a function which my be
different from their real role in vivo.
2. New functions can be induced in different tissues in vivo by changing the
environment and the regulation of gene expression.

Taking all the above reservations in consideration, we can ask: "What are the essential
characteristics and minimal prerequisites for valuable and practical in vivo models?" The
answer to this should include, on the one hand, conditions intrinsic to the model and on the
other hand, properties associated with the performance of the experiments.

Essential conditions intrinsic to the in vivo model are:

1. Reproducibility of the results under the same conditions.
2. High reliability regarding the positive and negative controls. These should be strictly
consistent even under different experimental designs.
3. Interpretation and measurement of the observed results has to be easy and
unequivocal with minimal interobserver variations.
4. Intraexperimental repeats and interexperimental variabilities should be minimal and in
accordance with the fact that we are dealing with an in vivo system.
5. The model should mimic as closely as possible similar clinical conditions observed in
humans.

Essential conditions intrinsic to the performance of the experiments:

1. An experimenter should be able to accurately use the model after a short period of
training experience.
2. Neutralization of possible external influence on the experiment outcome should be
straightforward.
3. Cost should be reasonable in order to allow for many repeats to increase accuracy.
4. The experiment has to be finalised within a reasonable time framework. Handling a
living tissue for a long period of time may induce secondary and uncontrollable
changes.

Considering all above prerequisites for a valuable and practical model for the study of
angiogenesis, there is no doubt that the eye and its various tissues are excellent candidates
for this purpose. Within the eye, the cornea is a near ideal model for the study of the
angiogenic process in vivo. The anatomy of the cornea with a large acellular stroma allows
for the easy implantation of tested materials. The cornea is clear and easily accessible and
can be directly visualised under maginification. The cornea is also avascular and clearly
delineated by pre existing blood vessels at the limbus. Because of these characteristics, the
cornea has become an indispensable organ for the study of angiogenesis.

88
Finally, the observations derived from any in vivo model have to be evaluated in a double
masked (double blind) manner. Recording ofresuIts derived from in vivo models in an open
manner can be (and are!) plagued with inaccuracies due to unavoidable biases innate to
biological systems. These have to be overcome in order to be able to derive sound and
accurate conclusions from the collected data. Unbiased recording of the data in a double
masked manner is one of the most essential conditions to be implemented when analysing
in vivo systems. In angiogenesis, due to the complexity of the process and the involved
cascade of events, only experiments carried out in a double masked manner can provide
data of scientific value.

REFERENCES

1. Michaelson IC: The mode of development of the vascular system of the retina
with some observations of its significance for certain retinal diseases.
Trans. OphthalmoL Soc. UK. 1948; 68: 137-180.

2. Michaelson IC: "Retinal Circulation in Man and Animals", Charles C. Thomas,

Publisher, Springfield, Illinois (1954).

3. Folkman J., Merler E., Abernathy c., Williams G: Isolation of a tumor factor
responsible for angiogenesis.
Exp. 1. Med. 1987; 235: 442-447.

4. BenEzra,D: Neovasculogenic ability of prostaglandins, growth factors and

synthetic chemoattractants.
Am. 1. OphthalmoL 86:455-461, 1978.

5. BenEzra, D: Neovasculogenesis. Triggering factors and possible mechanisms.

Survey OphthalmoL 24:167-176,1979.

6. Maftzir G., Aharonov,O. and BenEzra,D: Influence of interferons on corneal

angiogenesis induced by basic fibroblast growth factor and lipopolysaccharide.
Ocular Immunology and Inflammation 1:143-150,1993.

7. Aharonov 0., Maftzir, G. and BenEzra, D: The role of cytokines in angiogenesis

Ocular Immunology and Inflammation 1:135-141, 1993.

89
 Angiogenic Factors
and Their Receptors
THE SEVERAL ROLES FOR OXYGEN IN WOUND ANGIOGENESIS

Jeffrey 1. Gibson, MD., Thomas K. Hunt, MD., John 1. Feng, MD.,

Mark D. Rollins, MD., Ahmad Y. Sheikh, M. Zamirul Hussain, PhD.

University of California, Parnassus Avenue, San Francisco, CA 94143-

0758

Healing wounds provide a unique opportunity to study angiogenesis. Ever since the
development of multicellular organisms necessitated the development of a circulatory
system, injury has been a common feature of life. In all likelihood repair and replacement of
injured vessels occured simultaneously with the original processes by which blood vessels
developed in evolution. It is difficult even to imagine that reparative angiogenesis differs in
any material way from that which occurs in normal growth and development; or, for that
matter, from angiogenesis in many tumors. Wound angiogenesis is also extremely rapid and
is unencumbered by the irrelevant, often distracting properties of tumors. Furthermore,
wound healing presents a unique and liberating methodological opportunity because the
artifact of the measurement, the wound, is the object of the exercise. Finally, the wound
environment is easily measured and duplicated and intercellular messengers are easily
sampled. With respect to oxygen, for instance, wounds are hypoxic and hyperlactated by
nature l .2 This was found simply by aspirating and measuring the fluid which ordinarily
collects in wounds 2 Furthermore, oxygen tension, lactate levels and numerous other
characteristics are easily influenced and measured experimentally.

Surgeons have long recognized that oxygen enhances formation of granulation tissue,
i.e. angiogenesis in wounds3 Note that the relationship is direct, not inverse. While basic
scientists use hypoxia to elicit vascular endothelial growth factor from a variety of cells,
clinicians use hyperoxia to aid vessel growth. Both clinical and experimental experience in
wounds of intact animals leaves no doubt that in wounds hypoxia retards both angiogenesis
and deposition of its co-dependent structural protein, collagen, while hyperoxia accelerates
both 4 Therefore, any supposition that hypoxia stimulates wound angiogenesis in vivo
must be carefully questioned.

Nevertheless, a large body of evidence, including our own, proves that hypoxia does,
in fact, stimulate VEGF release from many cells, including macrophages which are generally

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 93
considered the main sources of angiogenic signals in wounds 5. Since both the cell biologic
and clinical facts are clearly correct, there must be a mechanism which reconciles them. We
undertook it to find this mechanism.

In a paper presented at the last NATO meeting we demonstrated that hypoxia or

15mM lactate or both increase VEGF mRNA and VEGF protein. VEGF was increased by
lactate in the presence of normal tissue oxygen tensions. VEGF accounted for about 90% of
the angiogenic substance(s) released by lactated or hypoxic macrophages!.

In most metabolic circumstances, the relationship between oxygen and lactate is

reciprocal--as oxygen supply diminishes, lactate appears, and as oxygen supply increases,
lactate falls. However, in wounds, the relationship is the opposite. As oxygen supply in
wounds rises, lactate does not fall because literally all of the cells which appear in
significant numbers in wounds are aerobically glycolytic and release lactate even when
oxygen is plentiful 6 . In other words, the finding that lactate stimulates VEGF release,
means that as long as the lactate is high, the stimulus to angiogenesis remains in wounds!,
regardless of oxygen status.

This finding then changes the question to how can both hypoxia and lactate stimulate
VEGF? As a start, we noted that the only effect of both oxygen and lactate which is not
reciprocal is that both reduce NAD+ to NADH. NADH has only one function, the release
of hydrogens to other substances7 . NAD+, (actually NADPR), however, is not merely a
means of transferring hydrogens. It has another important function which is to rid itself of
N, so as to produce ADPR (adenosine diphosphoribose, NAD without the N). ADPR is
then used in a number of "ADPribosylations". At least three enzymes do this. The one
that we are most interested in for this work is poly ADPR synthetase, a nuclear enzyme
which removes the N from NADH and polymerases the resultant ADPR. This product,
pADPR performs a number of functions including reparing single DNA strand breaks8 . It
also represses collagen and VEGF gene transcription, thus, its absence increases the
prevalence ofVEGF mRNA.

ADPribosylations originate from NAD+, not NADH. Therefore, the size of the
NAD+ pool, which regulates ADPribosylations, is influenced by the local redox state7 ,9.
The amount of NAD+ which is available for ribosylation is dependent upon the
availablility of oxygen and the concentration of lactate. In hypoxia or excess lactate, the
NAD+ concentration diminishes and ADPR ribosylations diminish as well lO .

Explaining these statements is perhaps best done in terms of their angiogenic

development. Our first step was to recognize that angiogenesis and collagen deposition are
co-dependent phenomina 4 . This is particularly clear in wounds where collagen cannot be
synthesized without angiogenesis, and where, concommitantly, new collagen fibers must
surround each budding vessel in order to prevent vessel rupture when the force of blood
pressure is inevitably imposed. The capillary fragility found in scurvy (in which collagen
synthesis is retarded) is compelling evidence of this.

While investigating the controlling factors in collagen synthesis, we found that collagen
synthesis and deposition have a curious property which also bears on angiogenesis and
begins to explain why oxygen enhances angiogenesis. Hypoxia and/or lactate, each alone or
together, stimulate the assembly of collagen synthetic and depository mechanisms

94
including increasing its mRNA prevalence and its post translational modifying enzymes 1,7,9
However, oxygen is specifically required at relatively high concentrations because the post-
translation hydroxylation of prolines and lysines in the collagen peptide is an oxygenase
reaction (i.e. requires oxygen as its substrate). Hydroxylation is necessary for the
exportation and cross binding of collagen molecules l l While oxygen is required for
hydroxylation, lactate is not. In other words, though a lack of oxygen causes the synthetic
apparatus to be assembled, oxygen is also required to run it to completion. When lactate is
added to fibroblast cell cultures, the rate of collagen export becomes proportional to the
local oxygen concentration l2 Clearly, the presence of lactate in wounds is critical to
collagen deposition. It is accepted that lactate increases collagen synthesis (but not export).

Hussain has reported that ADPribosylation reversably inhibits prolyl hydroxylase,

the enzyme which adds molecular oxygen to collagen peptide thus allowing it to be
exported and cross bound7 Adding NAD+, and thus increasing ADPribosylation inhibits
this enzyme. Tests, not yet published by Suski, in wounds confirm that lactate affects in
vivo wound healing as predicted.

But, if oxygen is present how can lactate accumulate? As mentioned earlier in this
paper, aerobic glycolysis is a feature of many cells including leukocytes and fibroblasts
both of which are plentiful in wounds. It is also a feature of almost all tumor cells.
Leukocytes and most malignant cells secrete lactate regardless of oxygen supply, and thus
raising oxygen supply to wounds has little effect on local lactate levels 6

In summary, macrophages and fibroblasts exist in the same environment in wounds.

We reasoned that each of these cells has its own response to that environment (high lactate,
low oxygen) and these responses would be cooperative. On testing this hypothesis, lactate
and hypoxia increased both VEGF and collagen mRNA and activated post translational
modifications of collagen. Data by Hussain on post translational modification of VEGF will
be disscussed later in this book (Hussain et. al.). The molecular means by which oxygen
stimulates collagen synthesis and angiogenesis is still unknown. We undertook to delineate
the role of oxygen in angiogenesis in the following experiments.

We now report, using the in-vivo matrigel model, that exposing mice to hyperbaric
oxygen stimulates angiogenesis over and above that seen in both hypoxic and normoxic
animals Matrigel implants were injected into mice and explanted, sectioned, and examined
microscopically. The results were semi-quantified on a angiogenesis grading scale as 0, .5
and 1 by the number of vessels seen on microscopic examination.

Animals were exposed to room air at sea level pressure, to 100% oxygen at sea level
pressure, and at 100% oxygen at 1, 2, 2.5 and 3 atmospheres pressure by means of a
hyperbaric chamber. Each "hyperoxic group" was exposed to the hyperoxic atmosphere for
90 minutes twice daily for seven days (Figure 1).

The results of these experiments were as follows: When no other substances were
added and when animals breathed room air at ambient pressure, no angiogenesis was seen.
However, angiogenesis increased in every case in which the animals breathed hyperoxic
atmospheres. This was statistically significantly in all of the groups when compared to
controls. The effect was greatest in the groups exposed to oxygen at 2.5 atm. Higher
pressures of 3.0 atm seemed to produced toxicity with a plateau of the angiogenesis

95
 1.1i

1.4

Angiogenesis
Score

1.2

T
·p<0.05 I LE.s.pZlL~JI4§1_lJJIp§L~E4JIiL_J2JI$EL
1.0 ATM 21% 1.0ATM 100% 2.0ATM 100% 2.5ATM 100% 3.0 ATM 100%

Tissue PO 2 65mmHg 120mmHg 190mmHg 270mmHg 450mmHg

Figure 1. Angiogenesis score with non-supplemented matrigel under increasing

oxygen tensions. Asterisk indicates P<O.05 compared to control at 1.0
ATM, 21 % oxygen,Fishers LSD. Mean ± SD.

stimulating effect. When VEGF (lOOng/per I ml implant) was added, an appropriate

angiogenesis score of.4 was seen in the room air group. The oxygen effect was seen in both
supplemented (VEGF) and unsupplemented hyperoxygenated matrigel plugs but was
greater in the VEGF-added groups.

In other groups of mice, oxygen probes were placed in the subcutaneous tissue at the
same level as the matrigel plugs, and the local oxygen tensions were measured and averaged.
The results were 65mrnHg at room air, ambient pressure, 120mmHg at 100% oxygen at
ambient pressure, 190mmHg at 100% oxygen at 2 atmosphers pressure, and 270mrnHg at
100% oxygen at 2.5 atm (see bottom of Figure I).

To make sense of this finding, it is necessary to review one more fact of the ADPR
system. Exposing cells to oxidants produces DNA strand breaks8,13,14. This activates
polyADPR synthetase which is a repair mechanism for strand breaks. In this process,
NAD+ is consummed and its prevalence is markedly diminished. This effect has been
demonstrated by others as a mechanism in apoptosis. If the point of cell death is not
reached, however, the recovering cell finds itself acutely short of NAD+. In fibroblasts,
these conditions lead to stimulation of collagen synthesis. There is, therefore, one
commonality in three seemingly individual stimulants of collagen and VEGF synthesis;
lactate, hypoxia, and oxidant exposure. All three of the aforementioned conditions reduce
NAD+ levels 15 . With the reduction ofNAD+, the inhibitory effect of ADPribosylation on
both VEGF and collagen synthesis is released. As is the clinical experience with hyperbaric
oxygen, angiogenesis is increased by all three.

96
 Feeling confidant that the effects of oxygen that we were measuring represented an
endothelial response, we added anti-VEGF antibody to the matrigel in a further test of the
same design. This eliminated the oxygen stimulating effect in all groups. IgG was added to
matrigel as a control for the anti-VEGF protein and led to the same results as the
unsupplemented groups.

Recent studies in an animal model suggest that high oxygen tension at 2.5 atm does not
increase VEGF concentration in wound cylinders when compared to controls at room air.
This leads us to suspect that the increased angiogenic response observed in high oxygen
tensions is predominantly due to the endothelial cell response to VEGF at elevated P02.
Another possible explanation of elevated P02' s effect may involve the regulation of VEGF
bioactivity by post translational modification. However, these possibilities remain
untested.

Clearly, a result in which hypoxia and hyperoxia yield the same finding must be the
result of two mechanisms. However, our current hypothesis is that both sets of data are
correct, and that these seemingly contradictory results can be explained through an
adenosine diphosphoribosylation mechanism. We postulate that ADPribosylation is the
effector mechanism of hypoxia and lactate excess as we previously noted. However, the
NAD+ pool is equally sensitive to excess oxidant concentration via activation of P ADPR
synthetase by DNA strand breaks. Both circumstances are well known to cause a marked
reduction of the NAD+ pool and this, in turn, leads to the synthesis of both collagen and
VEGF i ,7

REFERENCES

1. Constant,1. MD., Suh, D. MD., Hussain, M. PhD., Hunt, T. MD. Wound Healing
Angiogenesis: The Metabolic Basis of Repair. In: Mol, Cell, Clin Aspects of
Angiogenesis. 1996. New York. Plenum Press;151-159.

2. Hunt TK, Conolly WB, Aronson SB, Goldstien P. Anaerobic metabolism and wound
healing: An Hypothesis for the initiation and cessation of collagen synthesis in
wounds. Am J Surg. 1978, 135:328-332.

3. Hunt TK. "The Physiology of Wound Healing." Ann of Emerg. Med. 1988; 17: 1265-
1273.

4. Hunt TK, Pai MP. The effect of varying ambient oxygen tensions on wound
metabolism and collagen synthesis. Surg Gynecol Obstet. 1972;135:561-567.

5. Farrara N, Davis-Smyth T. The biology of vascular endothelial growth factor.

Endocrine Reviews. 1997; 18(1):4-25.

6. Newsholme P., Costa Rosa LFP, Newsholme E., Curi R. The importance offuel
metabolism to macrophage function. Cell Biochem and Function 1996; 14: 1-10.

7. Hussain MZ, Ghani QP, Hunt TK.. Inhibition of prolyl hydorxylase by poly [ADP-
ribose] and phosphoribosyl-AMP. Possible role of ADP ribosylation in intracellular
prolyl hydroxylase regulation. J Bioi Chern. 1989;264; 14:7850-5.

97
8. Satoh MS, Lindahl T. Role of Poly [ADP-ribose] fonnation in DNA repair. Nat.
1992;356:356-358.

9. Ghani QP, Hussain MZ, Zhang J, Hunt TK. Control of pro collagen gene transcription
and prolyl hydorxylase activity by poly[ADP-ribosel G Poirier and A Moreaer (eds).
In: ADP-Ribosylation Reactions. 1992. New York. Springer Verlag 111-117.

10. Koch AE; Cho M; Burrows JC; Polverini PJ; Leibovich SJ. Inhibition of production
of monocyte/macrophage-derived angiogenic activity by oxygen free-radical
scavengers. Cell Biology International Reports, 1992 May, 16(5):415-25.

11. Eyre, D.R., Paz, M.A., Gallop, P.M. Cross-linking in collagen and elastin. Annu.
Rev. Biochem . 1984. 53:717-748.

12. Jonsson K, Jensen JA, Goodson WH, Scheuenstuhl H, West J, HopfH, Hunt TK.
Tissue oxygenation, anemia and perfusion in relation to wound healing in surgical
patients. Annals ofSurg, 1991 214:605-613.

13 . Berger NA. Poly[ADP-ribose] in the cellular response to DNA damage.Rad. Res.

1985; 101 :4-15.

14. Satoh MS, Poirier GG, Lindahl T. NAD+-dependent repair of damaged DNA by
human cell extracts. J. Bioi Chern. 1993;268:5480-5487.

15. Winkler J; Cochran F. Apoptosis: insight into its role in inflammation. Inflammation
research, 1997 Jan, 46(1):3

98
AUTOCRINE ROLE OF BASIC FIBROBLAST GROWTH FACTOR (bFGF) IN
ANGIOGENESIS AND ANGIOPROLIFERATIVE DISEASES

Anna Gualandris, Patrizia Dell'Era, Marco Rusnati, Roberta Giuliani,

Elena Tanghetti, *Maria Pia Molinari-Tosatti, **Marina Ziche,
#Domenico Ribatti, and Marco Presta

Unit of General Pathology and Immunology and *Unit of Histology,

Department of Biomedical Sciences and Biotechnology, University of
Brescia, 25123 Brescia, Italy
**Department of Pharmacology, University of Florence, 50134 Florence,
Italy
#Institute of Anatomy, Histology and Embryology, University of Bari,
70124 Bari, Italy

1. AUTOCRINE ROLE OF bFGF IN ANGIOGENESIS

bFGF belongs to the family of the heparin-binding growth factors (Basilico and Moscatelli,
1992). The single copy human bFGF gene encodes multiple bFGF isoforms with molecular
weights ranging from 24 kD to 18 kD. High molecular weight isoforms (HMW-bFGFs) are
colinear NH2-terminal extensions of the better characterized 18 kD protein (Florkiewicz
and Sommer, 1989). Both low and high molecular weight bFGFs exert angiogenic activity in
vivo and induce cell proliferation, protease production, and chemotaxis in endothelial cells
in vitro (Gualandris, Urbinati, Rusnati, Ziche, and Presta, 1994). Also, bFGF has been
shown to stimulate endothelial cells to form capillary-like structures in collagen gels
(Montesano, Vassalli, Baird, Guillemin, and Orci, 1989) and to invade the amniotic
membrane in vitro (Mignatti, Tauboi, Robbins, and Rifkin, 1989). Moreover, the
phenotype induced in vitro by bFGF in endothelial cells includes modulation of integrin
expression (Klein, Giancotti, Presta, Albelda, Buck, and Rifkin, 1993), gap-junctional
intercellular communication (Pepper and Meda, 1992) and urokinase receptor upregulation
(Mignatti, Mazzieri, and Rifkin, 1991a). Studies with neutralizing anti-bFGF antibodies
have implicated endogenous bFGF in wound repair (Broadly, Aquino, Woodward, Bickley-
Sturrock, Sat, Rifkin, and Davidson, 1989), vascularization of the chorioallantoic membrane
during chick embryo development (Ribatti, Urbinati, Nico, Rusnati, Roncali, and Presta,
1995), and tumor growth under defined experimental conditions (Baird, Mormede, and

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 99
Bohlen, 1986; Gross, Herblin, Dusak, Czerniak, Diamond, Sun, Eidsvoog, Dexter, and
Yoayon, 1993).
Several cell types, including tumor cells of different origin (Moscatelli, Presta, Joseph-
Silverstein, and Rifkin, 1986; Ohtani, Nakamura, Watanabe, Mizoi, Saku, and Nagura,
1993; Presta, Moscatelli, Joseph-Silverstein, and Rifkin, 1986; Schulze-Osthoff, Risau,
Vollmer, and Sorg, 1990; Takahashi, Mori, Fukumoto, Igarashi, Jaye, Oda, Kikuchi, and
Hatanaka, 1990; Yamanaka, Friess, Buchler, Beger, Uchida, Onda, and Kobrin, 1993),
macrophages (Baird, Mormede, and Bohlen, 1985) and T lymphocytes (Blotnik, Peoples,
Freeman, Eberlein, and Klagsbrun, 1994), express bFGF in vitro and in vivo. bFGF lacks a
classic signal peptide for secretion (Abraham, Mergia, Whang, Tumolo, Friedman, Hjerrild,
Gospodarowicz, and Fiddes, 1986). However, cell damage may cause the release of bFGF
from producing cells (Gajdusek and Carbon, 1989; McNeil, Muthukrishnan, Warder, and
D' Amore, 1989; Witte, Fuka, Haimovitz, Vlodavski, Goodman, and Edor, 1989). Also, an
alternative mechanism of exocytosis of bFGF, independent of the endoplasmic
reticulum/Golgi pathway, has been proposed (Mignatti, Morimoto, and Rifkin, 1991b).
Accordingly, bFGF has been found associated with the extracellular matrix (ECM) of cell
cultures in vitro (Rogelj, Klagsbrun, Atzmon, Kurokawa, Haimovitz, Fuks, and Vlodavski,
1989; Vlodavski, Folkman, Sullivan, Friedman, Ishai-Michaell, Sasse, and Klagsbrun,
1987a) and located in the basement membranes of blood vessels in vivo (DiMario,
Buffinger, Yamada, and Strohman, 1989; Folkman, Klagsbrun, Sasse, Wadzinski, Ingber,
and Vlodavski, 1988). On this basis, bFGF is thought to exert its effects on endothelial
cells via a paracrine mode consequent to its release by other cells and/or mobilization from
proteoglycans ofECM (Figure 1).
Besides experimental evidence for paracrine mode of action for bFGF, some observations
raise the hypothesis that bFGF may also play an autocrine role in endothelial cells. In
vitro, it has been shown that different endothelial cells produce bFGF (presta, Majer,
Rusnati, and Ragnotti, 1989; Schweigerer, Neufeld, Friedman, Abraham, Fiddes, and
Gospodarowicz, 1987; Vlodavski, Friedman, Sullivan, Sasse, and Klagsbrun, 1987b) and
that endogenous bFGF modulates cell proliferation and migration, as well as the production

paracrine mode autocrine mode

J:Jq) bFGF
G~
• / .

ondotheliOl coli

Figure 1. Paracrine and autocrine activity of bFGF on endothelium. Tumor cells and
inflammatory cells (M<l» release bFGF which acts on endothelial cells in a paracrine mode
of action. Alternatively, endogenous bFGF is upregulated in endothelial cells causing an
autocrine loop of stimulation.

100
Figure 2. bFGF immunostaining of blood vessel endothelium in human adenocarcinoma of
the cervix.

of proteinases and their receptors (Itoh, Mukoyama, Pratt, and Dzau, 1992; Pepper,
Sappino, Stocklin, Montesano, Orci, and Vassalli, 1993; Sato and Rifkin, 1988). In vivo, it
has been shown that bFGF expression occurs in the endothelium adjacent to neoplastic
cells in several human tumor types. These neoplasms include neuroblastoma, astrocytoma,
glioblastoma and meningioma (Schulze-Osthoff et aI., 1990; Takahashi et al., 1990; Zagzag,
Miller, Sato, Rifkin, and Burstein, 1990), pheochromocytoma (Statuto, Ennas, Zamboni,
Boneti, Pea, Bemardello, Pozzi, Rusnati, Gua1andris, and Presta, 1993), melanoma
(Schulze-Osthoff et al., 1990), carcinomas of the stomach and colon (Ohtani et al., 1993;
Schulze-Osthoff et aI., 1990), and adenocarcinomas of the larynx, endometrium, and cervix
(Figure 2). Thus, bFGF expression is a common feature of vascular endothelium during
tumor angiogenesis.
These observations strongly support the hypothesis that neovascularization may be
triggered by molecule(s) released by tumor cells and/or infiltrating inflammatory cells that
induce bFGF upregulation in the quiescent endothelium. In keeping with this hypothesis is
the observation that tumor cells of different origin release molecule(s) able to interact with
endothelium and to upregulate the expression of bFGF which, in tum, stimulates the
fibrinolytic potential of the endothelial cell in an autocrine manner (peverali, Mandriota,
Ciana, Marelli, Quax, Rifkin, Della Valle, and Mignatti, 1994). In addition, bFGF itself,
thrombin, and interIeukin-2 stimulate bFGF production in endothelial cells (Cozzolino,
Torcia, Lucibello, Morbidelli, Ziche, Platt, Fabiani, Brett, and Stern, 1993; Weich, Oberg,
Klagsbrun, and Folkman, 1991).
bFGF has been detected in cardiac myocytes (Speir, Tanner, Gonzales, Farris, Baird, and
Casscells, 1992) and cells of the coronary vasculature (Hawker and Granger, 1993). Also,
cultured coronary endothelium exhibits the FGF receptor on its surface and expresses
bFGF mRNA (Hawker and Granger, 1993), suggesting that bFGF might induce coronary
angiogenesis by an autocrine/paracrine mechanism. Indeed, we have shown (Ziche, Parenti,
Ledda, Dell' Era, Granger, Maggi, and Presta, 1997) that nitric oxide (NO) induces an
angiogenic phenotype (including cell proliferation and urokinase-type plasminogen
activator upregulation) in coronary venular endothelial cells by inducing endogenous bFGF
and that this pathway mediates the angiogenetic response to the vasoactive neuropeptide
substance P (Figure 3).

101
 ~
\ ..!L _ anti-bFGF Ab
. 3.\ 1' => '0

- 'If
~
f uPA
ub lance I' ~ :\0 ~ "'!F> proliferatioll

Figure 3. Endogenous bFGF mediates NO-induced angiogenic phenotype in coronary

venular endothelial cells. Substance P and the NO donor nitroprusside (NaNP) induce
bFGF upregulation which stimulates endothelial cell (EC) proliferation and urokinase-type
plasminogen activator (uPA) upregulation. Neutralizing anti-bFGF antibody prevents EC
response to NO.

2. AUTOCRINE ROLE OF bFGF IN VASCULAR TUMORS

Blood vessels may represent the site of origin for neoplasms, hamartomas, and vessel
malformations. Neoplasms include benign tumors and tumor-like lesions (hemangioma),
tumors of intermediate malignancy (hemangioendothelioma), and malignant tumors
(angiosarcoma) (Enzinger, 1995). The pathogenesis of vascular tumors is at present
unknown, even though the local, uncontrolled release of growth factors and/or lytic
enzymes has been hypothesized to facilitate endothelial cell proliferation and the formation
of vascular lacunae (Enzinger, 1995).
A close relationship exists between the formation of vascular tumors and angiogenesis. This
relationship is also apparent in Kaposi's sarcoma (KS). Classic KS is a relatively benign,
highly vascularized neoplasm. A clinically aggressive form of KS develops in a significant
percentage of acquired immune deficiency syndrome (AIDS) patients (Levine, 1993).
Histologically, KS is characterized by the presence of spindle-shaped cells, inflammatory
cells and newly formed blood vessels (Enzinger, 1995). KS lesions express various markers
for vascular endothelial cells, suggesting that KS spindle cells are of endothelial cell lineage
(Sturzl, Brandstetter, and Roth, 1992).
Several experimental evidences implicate bFGF in the pathogenesis of vascular lesions,
including KS and hemangiomas. III vitro, AIDS-KS cells derived from different patients
express high levels ofbFGF which is released in the extracellular media (Albini, Fontanini,
Masiello, Tacchetti, Bigini, Luzzi, Noonan, and Stetler-Stevenson, 1994). Antisense
oligonucleotides directed against bFGF mRNA inhibit both the growth of AIDS-KS cells
and the angiogenic activity associated with these cells, including the induction of KS-like
lesions in nude mice (Ensoli, Markham, Kao, Barillari, Fiorelli, Gendelman, Raffeld, Zon,
and Gallo, 1994b). bFGF immunoreactivity is detected both in classic and AIDS-associated
KS lesions in humans (Ensoli, Gendelman, Markham, Fiorelli, Colombini, Raffeld, Cafaro,
Chang, Brady, and Gallo, 1994a) and recombinant bFGF synergizes with HIY-I-Tat
protein in inducing the formation of vascular lesions closely resembling early KS into nude
mice (Ensoli et aI., 1994). Interestingly, cytokines from activated T cells induce bFGF
upregulation and the acquisition of a AIDS-KS spindle cell-like phenotype in normal
endothelial cells (Barillari, Buonaguro, Fiorelli, Hoffman, Michaels, Gallo, and Ensoli, 1992;
Fiorelli, Gendelman, Samaniego, Markham, and Ensoli, 1995). Finally, coexpression of
bFGF and endothelial phenotypic markers CD31 and von Willebrand factor has been found
in the proliferating phase of human hemangioma but not in vascular malformations

102
(Takahashi, Mulliken, Kozakewich, Rogers, Folkman, and Ezekowitz, 1994) Taken
together, the data suggest that bFGF produced by cells of the endothelial lineage may play
important autocrine and paracrine roles in the pathogenesis of vascular tumors.

3. bFGF OVEREXPRESSION IN MOUSE ENDOTHELIAL CELLS

To investigate the biological consequences of endothelial cell activation by endogenous

bFGF, immortalized Balb/c mouse aortic endothelial cells (MAE cells) and brain
microvascular cells (MBE cells) were transfected with a retroviral expression vector
harboring a human bFGF cDNA (Gualandris, Rusnati , Bell eri , Nelli, Bastaki, Molinari-
Tosatti, Bonardi, Parolini, Albini, Morbidelli, Ziche, Corallini, Possati, Vacca, Ribatti , and
Presta, 1996a; Gualandris, Rusnati, Belleri, Molinari-Tosatti, Bonardi, Parolini, Albini,
Ziche, and Presta, 1996b). bFGF transfectants (Figure 4) express all bFGF isoforms and
are characterized by a transformed morphology and an increased saturation density. bFGF
transfectants show invasive and morphogenetic behavior in three-dimensional gels which is
prevented by anti-bFGF antibody, revealing the autocrine modality of the process
(Gualandris et aI., 1996a).
The biological consequences of this autocrine activation were investigated in vivo. bFGF-
transfected MAE cells induce the rapid growth of highly vascularized, non infiltrating
tumors. Lesions were observed also when cells were injected in x-irradiated syngeneic mice
but grew poorly in immunocompetent syngeneic animals, indicating that the growth of
these lesions is dependent on the immunological status of the host. Histologically, the
tumors have the appearance of hemangioendothelioma with spindled areas resembling KS
(Gualandris et al ., 1996a) and with numerous CD31-positive blood vessels and lacunae
(Figure SA). Southern blot analysis revealed that less than 10% of the cells in the tumor
mass were transplanted bFGF-transfected MAE cells. Accordingly, disaggregation of the

Figure 4. AcLDL uptake (A) and bFGF immunostaining (B) of bFGF-transfected MAE
cells.

103
 lesion and in vitro cell culture demonstrate that less than 10% of total cell population retain
bFGF overexpression and neomycin-resistance. These data indicate that bFGF-
overexpressing endothelial cells cause vascular lesions in the immunocompromised host
that are sustained to a large extent by recruitment of host cells, including endothelial cells.
In agreement with these observations, bFGF-transfected MAE cells induce an angiogenic
response when implanted in the avascular rabbit cornea (Gualandris et al., 1996a). Also,
they cause an increase in vascular density (Gualandris et al., 1996a) and formation of
hemangiomas in the chorioallantoic membrane when injected into the allantoic sac of the
chick embryo (Figure 5B).

4. CONCLUDING REMARKS

The capacity to cause opportunistic vascular lesions is not limited to bFGF-overexpressing

MAE cells. Previous observations had indicated in fact that human KS spindle cells cause

Figure 5. bFGF-transfected MAE cells induce vascular tumors with CD31-positive blood
vessels and lacunae when implanted s.c. in nude mice (A) and cause chorioallantoic
membrane hemangioma when injected into the chick embryo allantoic sac (B).

104
vascular tumors by host cell recruitment when injected into nude mice (Salahuddin,
Nakamura, Biberfeld, Kaplan, Markham, Larsson, and Gallo, 1988). It is interesting to note
that also KS spindle cells produce and secrete significant amounts of bFGF (Albini et aI.,
1994) and that antisense oligonucleotides directed against bFGF mRNA inhibit their
capacity to form vascular lesions in vivo (Ensoli et al., 1994b). Conversely, recombinant
bFGF synergizes with HIV-1-Tat protein in inducing the in vivo formation of vascular
lesions closely resembling early KS (Ensoli et aI., 1994a). These data point to a role for
bFGF in the pathogenesis of opportunistic vascular tumors. It must be pointed out that
this hypothesis does not rule out the possibility that other cytokines and growth factors
may also be of importance in the genesis and growth of these tumors, as demonstrated by
the capacity of PmT-transformed mouse endothelial cells to form cavernous hemangiomas
by host cell recruitment into nude mice (Garlanda, Parravicini, Sironi, De Rossi, Wain stock
de Calmanovici, Carozzi, Bussolino, Colotta, Mantovani, and Vecchi, 1994). PmT-
transformed endothelial cells produce various cytokines such as interleukin-6, chemokines,
and a 40 kD factor that is chemotactic for endothelial cells and distinct from bFGF
(Bussolino, De Rossi, Sica, Colotta, Wang, Bocchietto, Martin-Padura, Bosia, Dejana, and
Mantovani, 1991; Mantovani, Bussolino, and Dejana, 1992; Taraboletti, Belotti, Dejana,
Mantovani, and Giavazzi, 1993).
One important question raised by the in vivo behavior of bFGF-transfected MAE cells, as
well by that of KS-derived spindle cells or PmT-transformed endothelial cells, is how a
limited number of cells can sustain the growth of lesions formed essentially by recruited
host elements. As stated above, both KS spindle cells and PmT-transformed endothelial
cells produce a wide array of cytokines and many of them have mitogenic and chemotactic
effects for normal hosts cells, including vascular endothelial cells, smooth muscle cells, and
fibroblasts. Also, normal endothelial cells undergo phenotypic conversion of their
immunological and cytokine profiles when exposed to the conditioned medium of KS
spindle cells or of activated T cells (Barillari et aI., 1992; Fiorelli et aI., 1995). Thus,
autocrine and paracrine effects consequent to a cytokine network imbalance triggered by a
limited number of "activated" cells might lead to the progressive, self-sustained
amplification of the lesion. It seems possible to hypothesize that bFGF produced and
secreted by transfected MAE cells (Gualandris et aI., 1996a) may trigger those events
required to initiate the formation of hemangioendotheliomas, including endothelial cell
recruitment and angiogenesis.
Besides bFGF, other constitutively and/or bFGF-induced factors produced by MAE cells
may play a relevant role in the formation of the vascular lesions. Indeed, also parental
MAE cells induce the growth of tumors in nude mice even though with a very long delay of
appearance when compared to bFGF-transfected cells (Bastki, Nelli, Dell'Era, Rusnati,
Molinari-Tosatti, Parolini, Auerbach, Ruco, Possati, and Presta, 1997). Preliminary
observations have shown that both parental and bFGF-transfected MAE cells secrete the
chemokine MCP-1 and their conditioned media exert a chemotactic activity on
macrophages and endothelial cells. On the other hand, only the conditioned medium of
bFGF-transfected MAE cells induce an invasive and morphogenetic phenotype in normal
endothelial cells which is not prevented by anti-bFGF antibodies (Gualandris and Presta,
unpublished observations). In conclusion, bFGF acting directly on target cells or indirectly
through the induction of other factors that remain to be identified may facilitate vascular
tumor growth. Recently, it has been demonstrated that a minority of tumor cells
overexpressing acidic FGF induces a community effect on an entire population of non-
productive cells and that as few as 0.001% of acidic FGF-producing cells can cause a
significant reduction in the delay of tumor appearance (Jouanneau, Moens, Bourgeois,

105
 bFGF overexpression
in ECce11s

Autocrine effects: Paracrine effects:

• EC proliferation
• EC invasion • EC recruitment
• EC sprouting
• EC morphogenesis

• angiogenesis
• angioproliferative diseases

Figure 6. Biological consequences of bFGF overexpression in endothelial cells.

Poupon, and Thiery, 1994). This may explain the short delay of appearance of
pZipbFGF2-MAE cell-induced lesions when compared to those caused by parental MAE
cells. Other factors constitutively produced by MAE cells may act in a synergetic manner
with bFGF and contribute to the growth of the lesion by host cell recruitment.
In conclusion, our data indicate that bFGF-overexpressing endothelial cells acquire an
angiogenic phenotype and reclute quiescent endothelium originating angioproliferative
lesions in vivo. These findings demonstrate that bFGF overexpression exerts an autocrine
role for endothelial cells and support the notion that tumor neovascularization and
angioproliferative diseases can be triggered by stimuli that induce vascular endothelium to
produce its own autocrine factor(s) (Figure 6). bFGF-transfected MAE cell-induced lesions
represent a novel murine model for opportunistic vascular tumors and for the study of
novel positive and negative regulators of angiogenesis.

ACKNOWLEDGEMENTS

This work was supported by grants from Associazione Italiana per la Ricerca sui Cancro
(Special Project Angiogenesis), from Istituto Superiore di Sanita (IX AIDS Project), and
from European Communities (Human Capital Mobility Project "Mechanisms for the
Regulation of Angiogenesis") to M.P.

106
REFERENCES

Abraham, lA., Mergia, A, Whang, 1.L., Tumolo, A, Friedman, l, Hjerrild, J.,

Gospodarowicz, D., and Fiddes, 1., 1986, Nucleotide sequence of a bovine clone encoding
the angiogenic protein, basic fibroblast growth factor, Science 233 :545-548.

Albini, A, Fontanini, G., Masiello, L., Tacchetti, c., Bigini, D., Luzzi, P., Noonan, D .M.,
and Stetler-Stevenson, W.G., 1994, Angiogenic potential in vivo by Kaposi sarcoma cell-
free supernatants and H1Vl -tat product: inhibition of KS-like lesions by TIMP-2. AIDS
8:1237-1244.

Baird, A, Mormede, P, and Bohlen, P., 1985, Immunoreactive fibroblast growth factor in
cells of peritoneal exudate suggests its identity with macrophage-derived growth factor,
Biochem. Biophys. Res. Commun. 126:358-364.

Baird, A., Mormcde, P., and Bohlen, P., 1986, Immunoreactive fibroblast growth factor
(FGF) in a transplantable chondrosarcoma: inhibition of tumor growth by antibodies to
FGF, J. Cell. Biochem. 30:79-85.

Barillari, G., Buonaguro, L., Fiorelli, v., Hoffman, 1., Michaels, F., Gallo, R.C., and Ensoli,
B., 1992, Effects of cytokines from activated immune cells on vascular cell growth and
mY-I gene expression, J. Immuno!. 149:3727-3734.

Basilico, c., and Moscatelli, D., 1992, The FGF family of growth factors and oncogenes,
Adv. Cancer Res. 59:115-165.

Bastaki, M., Nelli, EE, Dell'Era, P., Rusnati, M., Molinari Tosatti, MP, Parolini, S,
Aurbach, R., Ruco, L.P., Possati, L., and Presta, M, 1997, Basic firboblast growth factor-
induced angiogenic phenotype in mouse endothelium A study of aortic and microvascular
endothelial cell lines, Arterioscler. Thromb. VascoBio!. 17:454-464.

Blotnick, S., Peoples, G.E., Freeman, M.R., Eberlein, T.J., and Klagsbrun, M, 1994,. T
lymphocytes synthesize and export heparin-binding epidermal growth factor-like growth
factor and basic fibroblast growth factor, mitogens for vascular cells and fibroblasts:
differential production and release by CD4+ and CD8+ T cells, Proc. Natl. Acad. Sci. U S.A .
91 :2890-2894.

Broadly, K.N., Aquino, A.M., Woodward, Sc., Buckley-Sturrock, A., Sato, Y, Rifkin,
D.B. , and Davidson, lM, 1989, Monospecific antibodies implicate basic fibroblast growth
factor in normal wound repair, Lab. Invest. 61:571-575 .

Bussolino, F., De Rossi , M., Sica, A, Colotta, F., Wang, 1.M., Bocchietto, E , Martin-
Padura, I., Bosia, A, Dejana, E, and Mantovani, A, 1991, Murine endothelioma cell lines
transformed by polyoma middle T oncogene as target for and producers of cytokines, J.
Immuno!. 147:2122-2129.

Cozzolino, F., Torcia, M, Lucibello, M. , Morbidelli, L., Ziche, M., Platt, l, Fabiani, S.,
Brett, 1., and Stern, D., 1993, Cytokine-mediated control of endothelial cell growth
interferon-a and interleukin-2 synergistically enhance basic fibroblast growth factor

107
synthesis and induce release promoting cell growth in vitro and in vivo, J. Clin. Invest.
91:2504-2512.

DiMario, J., Buffinger, N., Yamada, S., and Strohman, R.C., 1989, Fibroblast growth factor
in the extracellular matrix of dystrophic (mdx) mouse muscle, Science 244:688-690.

Ensoli, B., Gendelman, R., Markham, P., Fiorelli, V., Colombini, S., Raffeld, M., Cafaro,
A, Chang, H.K., Brady, J.N., and Gallo, RC., 1994a, Synergy between basic fibroblast
growth factor and mV-1 Tat protein in induction of Kaposi's sarcoma, Nature 371:674-
680.

Ensoli, B., Markham, P., Kao, v., Barillari, G., Fiorelli, V., Gendelman, R, Raffeld, M.,
Zon, G., and Gallo, R.C., 1994b, Block of AIDS-Kaposi's sarcoma (KS) cell growth,
angiogenesis, and lesion formation in nude mice by antisense oligonucleotide targeting basic
fibroblast growth factor. A novel strategy for the therapy of KS, J. Clin. Invest. 94: 1736-
1746.

Enzinger, F.M., and Weiss S.w., 1995, Soft tissue tumors, pp. 579-677, Mosby-Year
Book Inc., St. Louis.

Fiorelli, v., Gendelman, R, Samaniego, F., Markham, P.D., and Ensoli, B., 1995,
Cytokines from activated T cells induce normal endothelial cells to acquire the phenotypic
and functional features of AIDS-Kaposi' s sarcoma spindle cells, J. Clin. Invest. 95: 1723-
1734.

Florkiewicz, RZ., and Sommer, A, 1989, Human basic fibroblast growth factor gene
encodes four polypeptides: three initiate translation from non-AUG codons, Proc. Natl.
Acad Sci. U.S.A. 86:3978-3981.

Folkman, J., Klagsbrun, M., Sasse, J., Wadzinski, M., Ingber, D., and Vlodavski, I., 1988.
A heparin-binding angiogenic protein -basic fibroblast growth factor- is stored within
basement membran,. Am. J. Pathol. 130:393-400.

Gajdusek, C.M., and Carbon, S., 1989, Injury-induced release of basic fibroblast growth
factor from bovine aortic endothelium, J. Cell. Physiol. 139:570-579.

Garlanda, c., Parravicini, c., Sironi, M., De Rossi, M., Wainstock de Calmanovici, R,
Carozzi, F., Bussolino, F., Colotta, F., Mantovani A., and Vecchi, A., 1994, Progressive
growth in immunodeficient mice and host cell recruitment by mouse endothelial cells
transformed by polyoma middle-sized T antigen: implications for the pathogenesis of
opportunistic vascular tumors, Proc. Natl. Acad Sci. U.S.A. 91:7291-7295.

Gross, J.L., Herblin, W.F., Dusak, B.A, Czerniak, P., Diamond, M.D., Sun, T., Eidsvoog,
K., Dexter, D.L., and Yayon, A, 1993, Effects of modulation of basic fibroblast growth
factor on tumor growth in vivo, J. Natl. Cancer Inst. 85: 121-131.

Gualandris, A., Urbinati, c., Rusnati, M., Ziche, M., and Presta, M., 1994, Interaction of
high molecular weight basic fibroblast growth factor (bFGF) with endothelium: biological
activity and intracellular fate of human recombinant Mr 24,000 bFGF, J. Cell. Physioi.
161:149-159.

108
Gualandris, A, Rusnati, M., Belleri, M., Nelli, E.E., Bastaki, M., Molinari-Tosatti, M.P.,
Bonardi, F., Parolini, S., Albini, A, Morbidelli, L., Ziche, M., Corallini, A, Possati, L.,
Vacca, A, Ribatti, D., and Presta, M., 1996a, Basic fibroblast growth factor overexpression
in endothelial cells: an autocrine mechanism for angiogenesis and angioproliferative diseases,
Cell Growth & Differ. 7:147-160.

Gualandris, A, Rusnati, M., Belleri, M., Molinari-Tosatti, M.P., Bonardi, F., Parolini, S.,
Albini, A., Ziche, M., and Presta, M., 1966b, Angiogenic phenotype induced by basic
fibroblast growth factor transfection in brain microvascular endothelial cells: an in vitro
autocrine model of angiogenesis in brain tumors, Int. 1. Oncol. 8:567-573.

Hawker, 1.R.H., and Granger, 1., 1993, Tyrosine kinase inhibitors impair fibroblast growth
factor signaling in coronary endothelial cells, Am 1. Physiol. 266:HI07-HI20.

Itoh, H., Mukoyama, M., Pratt, R.E., and Dzau, V.J., 1992, Specific blockade of basic
fibroblast growth factor gene expression in endothelial cells by antisense oligonucleotide,
Biochem. Biophys. Res. Commlill. 188: 1205-1213.

Jouanneau, 1., Moens, G., Bourgeois, Y., Poupon, M.F., and Thiery, 1.P, 1994, A
minority of carcinoma cells producing acidic fibroblast growth factor induces a community
effect for tumor progression, Proc. Nat!. Acad Sci. US.A. 91:286-290.

Klein, S., Giancotti, F.G., Presta, M., Albelda, S.M., Buck, C.A, and Rifkin, DR, 1993,
Basic fibroblast growth factor modulates integrin expression in microvascular endothelial
cells, Mol. BioI. Cell. 4:973-982.

Levine, AM., 1993, AIDS-related malignancies: the emerging epidemic, 1. Natl. Cancer
Inst.85:1382-1387.

Mantovani, A, Bussolino, F., and Dejana, E., 1992, Cytokine regulation of endothelial cell
function, FASEB 1. 6:2591-2599.

McNeil, P.L., Muthukrishnan, L., Warder, E., and D'Amore, P., 1989, Growth factors are
released by mechanically wounded endothelial cells, 1. Cell Bioi. 109:811-822.

Mignatti, P., Tauboi, R., Robbins, E., and Rifkin, D.E., 1989, In vitro angiogenesis on the
human amniotic membrane: requirement for basic fibroblast growth factor, 1. Cell Bio!.
108:671-682.

Mignatti, P., Mazzieri, R., and Rifkin, D.B., 1991a, Expression of the urokinase receptor in
vascular endothelial cells is stimulated by basic fibroblast growth factor, 1. Cell Bioi.
113:1193-1201.

Mignatti, P., Morimoto, T., and Rifkin, D.B., 1991b, Basic fibroblast growth factor
released by single, isolated cells stimulates their migration in an autocrine manner, Proc.
Nat!. Acad. Sci. US.A. 88:11007-11011.

Montesano, R., Vassalli, 1.D., Baird, A., Guillemin, R., and Orci, L., 1986, Basic fibroblast
growth factor induces angiogenesis in vitro, Proc. Nat!. Acad Sci. US.A. 83:7297-7301.

Moscatelli, D., Presta, M., Joseph-Silverstein, J, and Rifkin, D.B., 1986, Both normal and
tumor cells produce basic fibroblast growth factor,1. Cell. Physiol. 129:273-276.

109
Ohtani, H., Nakamura, S., Watanabe, Y., Mizoi, T., Saku, T., and Nagura, H., 1993,
Immunocytochemical localization of basic fibroblast growth factor in carcinomas and
inflammatory lesions of the human digestive tract, Lab. Invest. 68 :520-527.

Pepper, M.S., and Meda, P., 1992, Basic fibroblast growth factor increases junctional
communication and connexin 43 expression in microvascular endothelial cells, J. Cell.
Physiol. 153: 196-205.

Pepper, M .S., Sappino, A.P. , Stocklin, R., Montesano, R., Orci, L., and Vassalli, J.D.,
1993, Upregulation of urokinase receptor expression on migrating endothelial cells, J. Cell
Bioi. 122:673-684.

Peverali, F.A., Mandriota, SJ., Ciana, P., Marelli, R., Quax, P., Rifkin,D.B., Della Valle,
G., and Mignatti, P., 1994, Tumor cells secrete an angiogenic factor that stimulates basic
fibroblast growth factor and urokinase expression in vascular endothelial cells, J. Cell.
Physiol. 161:1-14.

Presta, M., Moscatelli, D., Joseph-Silverstein, J., and Rifkin, D.B., 1986, Purification from
a human hepatoma cell line of a basic fibroblast growth factor-like molecule that stimulates
capillary endothelial cell plasminogen activator production, DNA synthesis, and migration,
Mol. Cell. Bioi. 6:4060-4066.

Presta, M., Maier, J.A.M., Rusnati, M., and Ragnotti, G., 1989, Basic fibroblast growth
factor: production, mitogenic response, and post-receptor signal transduction in cultured
normal and transformed fetal bovine aortic endothelial cells, J. Cell. Physiol. 141 :517-526.

Ribatti, D., Urbinati, C, Nico, B., Rusnati, M., Roncali, L., and Presta, M., 1995,
Endogenous basic fibroblast growth factor in the vascularization of the chick embryo
chorioallantoic membrane, Dev. BioI. 170:39-49.

Rogelj, S., Klagsbrun, M., Atzmon, R., Kurokawa, M., Haimovitz, A., Fuks, Z., and
Vlodavski, I., 1989, Basic fibroblast growth factor is an extracellular matrix component
required for supporting the proliferation of vascular endothelial cells and the differentiation
ofPC12 cells, J. Cell Bioi. 109:823-831.

Sato, Y., and Rifkin, D.B., 1988, Autocrine activities of basic fibroblast growth factor:
regulation of endothelial cell movement, plasminogen activator synthesis, and DNA
synthesis,J. Cell Bioi. 107:1199-1205.

Salahuddin, S.Z ., Nakamura, S., Biberfeld, P., Kaplan, M.H., Markham, P.D., Larsson, L.,
and Gallo, R.C, 1988, Angiogenic properties of Kaposi's sarcoma-derived cells after long-
term culture in vitro, Science 242:430-433 .

Schulze-Osthoff, K., Risau, W., Vollmer, E., and Sorg, C., 1990, In situ detection of basic
fibroblast growth factor by highly specific antibodies, Am. J. Pathol. 137:85-92.

Schweigerer, L. Neufeld, G. , Friedman, J., Abraham, J.A., Fiddes, J.C, and

Gospodarowicz, D., 1987, Capillary endothelial cells express basic fibroblast growth
factor, a mitogen that promotes their own growth, Nature 325:257-259.

Speir, E., Tanner, V., Gonzales, A.M., Farris, J., Baird, A., and Casscells, W., 1992, Acid

110
and basic fibroblast growth factors in adult rat heart myocytes: localization, regulatio in
culture, and effects on DNA synthesis, Circ. Res. 71:251-259.

Statuto, M., Ennas, M.G., Zamboni, G., Bonetti, F., Pea, M., Bemardello, F., Pozzi, A,
Rusnati, M., Gualandris, A, and Presta, M., 1993, Basic fibroblast growth factor in human
pheochromocytoma: a biochemical and immunohistochemical study, Int. J Cancer. 53:5-
10.

Sturzl, M., Brandstetter, H, and Roth, W.K., 1992, Kaposi's sarcoma: a review of gene
expression and ultrastructure of KS spindle cells in vivo, AIDS Res. Human Retrov.
8:1753-1763.

Taraboletti, G., Belotti, D., Dejana, E., Mantovani, A, and Giavazzi, F., 1993, Endothelial
cell migration and invasiveness are induced by a soluble factor produced by murine
endothelioma cells transformed by polyoma virus middle T oncogene, Cancer Res.
53:3812-3816.

Takahashi, 1.A, Mori, H, Fukumoto, M., Igarashi, K., Jaye, M., Oda, Y., Kikuchi, H., and
Hatanaka, M., 1990, Gene expression of fibroblast growth factors in human gliomas and
meningiomas: demonstration of cellular source of basic fibroblast growth factor mRNA and
peptide in tumor tissues, Froc. Nat!. Acad Sci. U.S.A. 87:5710-5714.

Takahashi, K., Mulliken, J.B., Kozakewich, H.P.w., Rogers, R.A., Folkman, 1., and
Ezekowitz, RAB., 1994, Cellular markers that distinguish the phases of hemangioma
during infancy and childhood, J Clin. Invest. 93:2357-2364.

Vlodavski, I., Folkman, 1., Sullivan, R, Friedman, R, Ishai-Michaell, R., Sasse, 1., and
Klagsbrun, M., I 987a, Endothelial cell-derived basic fibroblast growth factor: synthesis and
deposition into subendothelial extracellular matrix, Froc. Natl. Acad. Sci. U.S.A. 84:2292-
2296.

Vlodavski, I., Friedman, R., Sullivan, R., Sasse, J., and Klagsbrun, M., 1987b, Aortic
endothelial cells synthesize basic fibroblast growth factor which remains cell associated and
platelet-derived growth factor-like protein which is secreted, J. Cell. Physiol. 131 :402-408.

Weich, H, Iberg, N., Klagsbrun, M., and Folkman, 1., 1991, Transcriptional regulation of
basic fibroblast growth factor gene expression in capillary endothelial cells, J Cell.
Biochem.47:158-194.

Witte, L., Fuka, Z., Haimovitz, F.A., Vlodavski, I., Goodman, D.S., and Eldor, A., 1989,
Effects of irradiation on the release of growth factors from cultured bovine, porcine, and
human endothelial cells, Cancer Res. 49:5066-5072.

Yamanaka, Y., Friess, H., Buchler, M., Beger, HG., Uchida, E., Onda, M., and Kobrin,
M.S., 1993, Overexpression of acidic and basic fibroblast growth factors in human
pancreatic cancer correlates with advanced tumor stage, Cancer Res. 53:5289-5296.

Zagzag, D., Miller, D.C., Sato, Y., Rifkin, D.B., and Burstein, DE, 1990,
Immunohistochemical localization of basic fibroblast growth factor in astrocytomas,
Cancer Res. 50:7393-7398.

III
Ziche, M ., Parenti, A., Ledda, F., Dell'Era, P., Granger, H.J., Maggi, c.A., and Presta, M.,
1997, Nitric oxide promotes proliferation and plasminogen activator production by
coronary venular endothelium through endogenous bFGF, Cire. Res. in press.

112
INFLAMMATORY ANGIOGENIC FACTORS IN A MODEL OF CHRONIC
INFLAMMA nON.

James D. Winkler, Michael P. Seed* and Jeffrey R. Jackson

Department of Immunopharmacology, SmithKline Beecham

Pharmaceuticals, King of Prussia, PA. 19406 USA
*Department of Experimental Pathology, The William Harvey Research
Institute, St. Bartholomew's Hospital Medical College, London, United
Kingdom

1. SUMMARY

In many diseases, there is an apparent co-dependence of inflammation and angiogenesis.

There's a very strong inflammatory component in rheumatoid arthritis, psoriasis,
inflammatory bowel disease, during tumor growth and in some angiogenic ocular disorders.
Many of the mediators that we think about in inflammation can promote angiogenesis, such
as some interleukins, TNFa, prostaglandins, arachidonic acid metabolites and platelet-
activating factor. Some of these appear to act indirectly by changing the balance between
angiogenesis stimulation and inhibition, promoting the production of some of the more
directly acting angiogenesis inducers, such as VEGF.
We have been using an animal model to study angiogenesis in an inflammatory setting, an
air pouch granuloma model, in which there is an inflammatory response and the formation
of granulomatous tissue. We have assessed the characteristics of this granuloma over time,
it's weight, histological structure, vascular volume, cellularity and mediators produced.
Many of these characteristics change during time after the start of granuloma growth.
We have characterized a number of different classes of compounds in this model.
Protamine and the angiostatic steroid medroxyprogesterone each blocked about fifty
percent of the angiogenesis produced by adjuvant and croton oil injection. SB 203347, an
inhibitor of class IT, fourteen kDa PLA 2 , an enzyme thought to be responsible for the
production of leukotrienes and platelet-activating factor, produced a striking reduction in
the angiogenic response in mice. Inhibitors of p38 MAP kinase were also found to cause a
dose-related inhibition of the angiogenic response. These compounds block the production
of cytokines, such as TNFa and ILlb.

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 113
Thus, in this model of inflammatory angiogenesis, inhibitors of inflammatory mediator
production block angiogenesis, possibly by tipping the balance against the production of
pro-angiogenic mediators.

2. LINK BE1WEEN INFLAMMATION AND ANGIOGENESIS

There is growing evidence that there exists a link between the process of angiogenesis and
of inflammation. This link has been recently reviewed (Folkman and Brem, 1992; Jackson
et al., 1997e) and will be briefly outlined in the next two sections.

2.1 Angiogenesis supports inflammation

Chronic inflammatory disease is by its nature a proliferative process, similar in many ways
to cancer. In the case of inflammation there is both an influx of inflammatory cells from
other sites in the body as well as an in situ proliferation of both inflammatory and other
resident cells. One example for illustration is rheumatoid arthritis, in which there is a
proliferation of inflammatory tissue called pannus. This pannus tissue is composed of
numerous inflammatory cells as well as proliferating synovial cells. As this tissue expands,
it has an increased need for blood supply, to both provide nutrients and remove waste. In
addition, the neovasculature facilitates the access of invading inflammatory cells to the
inflammatory site. The increased vascular surface area resulting from the angiogenesis can
enhance the activation of inflammatory cells and the ability to produce pro-inflammatory
mediators.
Indeed, the support that angiogenesis provides to the proliferating inflammatory tissue
appears to be critical in many ways for the function of that tissue. The therapeutic
hypothesis has been made that blockade of angiogenesis will be of benefit in chronic
inflammatory conditions, in a fashion analogous to cancer (Colville-Nash and Seed, 1993;
Folkman, 1995; Seed et al., 1997). The therapeutic restriction of angiogenesis in chronic
inflammation should not be seen purely as a molecular tourniquet but as a disease
modifying treatment targeting the microvasculature.

2.2 Inflammation supports angiogenesis

The signals controlling angiogenesis, while directed at the endothelial cells, come from cells
in the nearby tissues. Many cells are capable of producing angiogenic factors when given
the correct environmental signals. Inflammatory cells, especially macrophages, are
exceptional producers of pro-angiogenic molecules. This inflammatory
monocyte/macrophage cell type can be found at most sites where angiogenesis is occurring
and almost every growth factor and cytokine known to regulate angiogenesis can be
produced by macrophages (Sunderkotter et al., 1994). Thus the presence of inflammatory
cells can promote angiogenesis.
In addition, many factors that are produced to promote the inflammatory response can also
promote angiogenesis. Some of these factors are indirect, that is those which act via
intermediary mechanisms, and are often found to induce angiogenesis only in vivo (Jackson
etal., 1997e). ProstaglandinsEI and E2 (PGEI, PGE2), tumor necrosis factor a (TNF-a),
interleukins I, 6, and 8 (IL-l, IL-6, IL-8) have all been shown to induce angiogenesis in
vivo. The E prostaglandins induce angiogenesis in both the cornea and chicken
chorioallantoic membrane assays (Jackson et al., 1997e) and may mediate the angiogenic
actions of bFGF (Spisni et al., 1992). IL-8 is both chemotactic and mitogenic for
endothelial cells in addition to causing angiogenesis in vivo in the cornea assay (Jackson et

114
al., I 997e). Aberrant increases in IL-8 produced by psoriatic keratinocytes have been said
to be central to their ability to produce an angiogenic response (Nickoloff et aI., 1994). IL-
1b has been reported to be significantly more potent than VEGF in inducing corneal
angiogenesis, requiring as little as I ng (BenEzra et al., 1990). TNF-a appears to have dose-
dependent pleiotropic effects on angiogenesis: in vivo, it is a potent inducer at low doses
in corneal implants and an inhibitor at higher doses (Fajardo et aI., 1992). Finally, platelet-
activating factor, a potent pro-inflammatory lipid, has been shown to directly stimulate
endothelial cell proliferation and to have angiogenic effects in vivo (Camussi et aI., 1995).
These agents can induce many facets of inflammation and their angiogenic effects could be
linked to the recruitment of inflammatory cells or increased vascular permeability.
However, recent data suggest that many of these inflammatory mediators can induce the
expression of more directly acting angiogenic factors, such as VEGF. PGE2, 1L6 and IL-I
have been shown to induce VEGF mRNA and protein expression (Cohen et al., 1996; Ben-
Av et al., 1995; Li et aI., 1995; Jackson et aI., 1997d). VEGF also appears to be regulated
by the oxygen concentration that cells are exposed to, with hypoxia inducing its expression
(Shweiki et aI., 1995; Shima et al., 1995; Jackson et al., 1997d). Thus, during growth of
inflammatory tissue, the hypoxic conditions, coupled with high concentrations of
inflammatory mediators, can cause an increase of VEGF production. Such increases have
been seen in rheumatoid arthritis (Nagashima et al., 1995).

3. MOUSE AIR POUCH GRANUWMA MODEL OF INFLAMMATORY

ANGIOGENESIS

In order to study the interactions between inflammation and angiogenesis, we have

employed a murine air pouch granuloma model, first described by Kimura et al. (1985;
1986) and modified and characterized by Dr. Willoughby's laboratory (Colville-Nash et al.,
1995). This model involves the creation of an air pouch that in many ways resembles a
normal synovial space. The response is started by the injection of Freund's complete
adjuvant and croton oil. A key feature of this model is that the granuloma growth is not
dependent on the angiogenesis; removing the croton oil results in the growth of an avascular
granuloma (Kimura et aI., 1985). Thus, treatments that selectively block angiogenesis can
be observed to have different effects from those that block inflammation. For example, the
anti-inflammatory steroid dexamethasone will decrease the size of the granuloma, whereas
the anti-angiogenic steroids medroxyprogesterone and tetrahydrocortisol will block block
vessel growth without affecting granuloma size (Colville-Nash et al., 1995; Jackson et al.,
I 997c).

4. TEMPORAL CHANGES IN THE INTERACTION BElWEEN

INFLAMMA TION AND ANGIOGENESIS

The granuloma within the air pouch grows over time, typically peaking around 7 days.
Over that time, there are a number of changes that occur, at both a structural and a
molecular level. The temporal changes are important to consider as they may reflect
changes in the interaction between inflammation and angiogenesis. They may also point to
differences in therapeutic strategies that should be considered. The next sections detail
some of these temporal changes, summarized in Table 1.

115
 Table 1. Granuloma Characteristics Change Over Time

Time Period
Earl Middle Late

Cells Neutrophils Monocytes Macrophages

Lymphocytes Fibroblasts

Organization Diffuse Cellular Fibrous

Matrix

Mediators PDGF, EGF ILl bFGF

VEGF TNF TGFb

Angiogenesis Large Branches Fine Networking Reorganizati on

Regression

4.1 Granuloma growth

As mentioned above, the granuloma grows over time. It obviously starts at zero size, then
quickly expands over 7 days, then slowly diminishes over the subsequent 2-4 weeks
(Colville-Nash et aI., 1995). In contrast to the steady increase then decrease in the size of
the granuloma, the percentage of granuloma vasculature appears to fluctuate. This vascular
index is measured by vascular casting of the vessels, using carmine dye in gelatin, followed
by extraction and optical density measurement. The vascular index increases over the first 5
days. In careful studies there is observed a decrease in the vascular index between days 5-7,
followed by an increase (Colville-Nash et al., 1995; Jackson et al., 1997c). Thus, the
vascular index, or angiogenesis of the tissue, is not constant over time, implying that
different stimuli and/or amounts of stimuli may drive it at different time.

4.2 Histological changes

One of the most striking changes that occurs over time is in the histological organization of
the granuloma tissue. The early granuloma is extremely cellular, with little organization
observed. By 3 days, there is clear structure emerging, with developed blood vessels easily
seen. This structural aspect is heightened by 12-14 days, when there emerges a clear
fibrotic nature to the granuloma. In fact, a 3-dimensional organization is apparent, with
many blood vessels running perpendicular to tracts of fibroblast-like cells and extracellular
matrix. Thus, the structural complexity of the granuloma increases over time.

4.3 Cellular changes

As in many chronic inflammatory situations, there are waves of different inflammatory
cells that come to the site (Kimura et al., 1985; Appleton et al., 1993). The first wave
observed in the granuloma is of polymorphonuclear cells, and this wave peaks 3 days after
the initiation of the lesion. This wave is soon followed by a wave of mononuclear cells,
including monocytes and lymphocytes. This wave occurs over a longer time course,
peaking from 5-7 days after insult. The final wave is the appearance of more "residential"
inflammatory cells, such as tissue macrophages and mast cells, in addition to fibroblasts,
indicating a transition from acute to chronic inflammation. Because these different cells

116
respond to inflammatory stimuli in different ways, and also respond to different stimuli, it
is not surprising that the presence of different inflammatory cells can affect the character of
the granuloma is several ways, including the mediators produced and the structure of the
granuloma itself. The structure was discussed above and the mediators will be discussed in
the following section.

4.4 Mediator changes

Because of the dramatic changes that occur over time in the structure of the granuloma and
in its cellular composition, it is reasonable to hypothesize that there may be changes in the
amounts and type of mediators that are produced. Recent evidence is beginning to support
this hypothesis. For example, the report by Appleton et al. (1996) documented that there
is a rapid yet transient expression of VEGF in the granuloma. The peak of VEGF
expression appears around I day after insult and may account for the earliest
vascularization of the granuloma. Treatment with antibody against VEGF was shown to
block early vascular responses (Appleton et aI., 1996). Other early mediators include
PDGF and EGF (Appleton et al., 1993).
In contrast to the results with VEGF, several inflammatory mediators appear to peak at
later times. For example, there are dramatic increases in the content of !LIb and TNFa in
the granuloma at 5-7 days after inflammatory insult. This peak of mediator production
coincides with the influx of mononuclear cells into the granuloma and the change from acute
to chronic inflammatory character. It is possible that a second wave of angiogenesis occurs
as the result of this increase in indirect angiogenesis mediators, and preliminary evidence
with selective inhibitors supports this notion.
Increases in other mediators or growth factors may occur at later times, such as during the
fibrotic phase of the granuloma (days 10-21). However, these later times are much less
studied. There are indications that bFGF and TGFb are increased at later times (Appleton
et al., 1993). Certainly, as the complexity of the tissue increases, it is possible that
numerous mediators are interacting with the vasculature to support angiogenesis and it may
become correspondingly more difficult to disrupt angiogenesis at later times.

4.5 Pharmacological manipUlation of inflammatory mediators

As mentioned above, there are times during the angiogenic response in which selective
inhibitors of inflammatory mediators are effective at blocking neovascularization. For
example, we have recently observed that inhibition of cytokine production, using inhibitors
of CSBP/p38 MAP kinase, reduced angiogenesis in the air pouch granuloma model
(Jackson etal., 1997a). This reduction of angiogenesis occurred during days 5-7, the time of
peak cytokine production and was accompanied by dramatic reductions in TNFa and !LIb.
Additional supportive evidence is seen with inhibitors of 14 kDa PLA2. Inhibition of this
enzyme blocks leukotriene and platelet-activating factor production (Jackson et aI., 1997c)
and reduces angiogenesis in the granuloma. Non-steroidal antiinflammatory drugs have also
been affective anti angiogenic agents in this model (Colville-Nash et al., 1992).
Taken together, these results suggest that, during the vascularization of the granuloma,
many inflammatory mediators and cytokines contribute to the pro-angiogenic envimment.
Drugs that can decrease these mediators can affect this environment, shifting the balance
away from neovascularization. A brief summary of some of the compounds that have been
shown to inhibit angiogenesis in this model is shown in Table 2.

117
Table 2. Summary of the Effects on Various Inhibitors on Angiogenesis in the Mouse Air
Pouch Granuloma Model

Compound Molecular Effect on Reference

Target Granuloma
An io enesis

Protamine Growth factor 55% (Jackson et al., 1997c)

binding?

Hyaluronan + Steroid receptor? 52% (Colville-Nash et al.,

Cortisone 1995)

Medroxyprogesterone Steroid receptor? 44% (Jackson etal., 1997b)

Tetrahydrocortisol Steroid receptor? 55% (Colville-Nash et al.,

1995)
SB 203347 PLA2 34% (Jackson et al., 1997b)

SB 220025 p38 kinase 42% (Jackson etal., 1997a)

Ro 24-4736 P AF Receptor 38% (Jackson et aI., 1997b)

Antagonist

Ibuprofen COX 50% (Colville-Nash et al.,

1992)
Methotrexate DNA Synthesis 72% (Colville-Nash et aI.,
1992)
Anti-VEGF Ab VEGF 47% (Appleton et al., 1996)

5. CONCLUSIONS

The evidence is growing that there exist multiple links between angiogenesis and
inflammation, with both processes supporting each other. Because of this mutual support,
it is possible that inhibition of the production of inflammatory mediators will inhibit the
angiogenic process. In a model of inflammatory angiogenesis, the murine granuloma model,
the angiogenic process appears to be multifactorial and dynamic over time, with different
mediators stimulating angiogenesis at different times. Inhibitors of inflammatory mediator
production were shown to specifically block angiogenesis. Such evidence supports both the
link between angiogenesis and inflammation and its temporal nature.

6. REFERENCES

Appleton, I., Tomlinson, A., Colville-Nash, P.R. and Willoughby, D.A., 1993, Temporal
and spatial immunolocalization of cytokines in murine chronic granulomatous tissue, Lab.
Invest. 69:405-414.

118
Appleton, I., Brown, N.J., Willis, D., Colville-Nash, P.R., Alam, e., Brown, J.R. and
Willoughby, D.A., 1996, The role of vascular endothelial growth factor in a murine chronic
granulomatous tissue air pouch model of angiogenesis, 1. Pathol. 180:90-94.

Ben-Av, P., Crofford, L.J., Wilder, RL. and Hla, T., 1995, Induction of vascular
endothelial growth factor expression in synovial fibroblasts by prostaglandin E and
interleukin-I: a potential mechanism for inflammatory angigenesis, FEBS Lett 372:83-87.

BenEzra, D., Hemo, I. and Maftzir, G., 1990, In vivo angiogeneic activity of interleukins,
Arch Opthamoll08:573-576.

Camussi, G., Montrucchio, G., Lupia, E., De Martino, A., Perona, L., Arese, M.,
Vercellone, A., Toniolo, A. and Bussolino, F., 1995, Platelet-activating factor directly
stimulates in vitro migration of endothelial cells and promotes in vivo angiogenesis by a
heparin-dependent mechanism, 1. Immunol. 154:6492-6501.

Cohen, T., Nahari, D, Cerem, L.W., Neufeld, G. and Levi, BL, 1996, Interleukin-6
induces the expression of vascular endothelial growth-factor, J Bioi Chem 271 :736-741.

Colville-Nash, P.R., Seed, M.P. and Willoughby, D.A., 1992, Anti-rheumatic drugs and the
development of vasculature in murine chronic granulomatous air pouches, Br J Pharmacol
107:421P.

Colville-Nash, P.R. and Seed, M.P., 1993, The current state of angiostatic therapy, with
special reference to rheumatoid arthritis, Curr Opin Invest Drugs 2:763-813.

Colville-Nash, P.R., Alam, C.A.S., Appleton, I., Browne, J.R., Seed, M.P. and Willoughby,
D.A., 1995, The pharmacological modulation of angiogenesis in chronic granulomatous
inflammation, J Pharmacol Exp Ther 2741463-1472.

Fajardo, L.F., Kwan, H.H., Kowalski, J., Prionas, S.D. and Allison, A.e., 1992, Dual role
of tumor necrosis factor-a in angiogenesis, Am J PathoI140:539-544.

Folkman, J., 1995, Angiogenesis in cancer, vascular, rheumatoid and other disease, Nat.
Med. 1:27-31.

Folkman, J. and Brem, H., Angiogenesis and inflammation. In: J.1. Gallin, I.M. Goldstein
and R Snyderman (Eds.), Inflammation: Basic Principles and Clinical Correlates, Second
Edition, Vol. 2, Raven Press, Ltd, New York, 1992, pp. 821-839.

Jackson, J.R, Bolognese, B, Hillegass, L., Adams, J., Griswold, D.E. and Winkler, J.D.,
1997a, The role ofCSBP, p38, a stress response MAP kinase, in angiogenesis and chronic
inflammatory disease models, submitted.

Jackson, J.R, Bolognese, B., Hubbard, W.e., Marshall, L.A. and Winkler, J.D., 1997b,
Platelet-activating factor derived from 14 kDa phospholipase A2 contributes to
inflammatory angiogenesis, Submitted

119
Jackson, lR., Bolognese, B., Kircher, C.R., Marshall, L.A. and Winkler, lD., I 997c,
Modulation of angiogenesis in a model of chronic inflammation, Inflammation Research in
press.

Jackson, J.R., Minton, lA.L., Ho, M.L., Wei, N. and Winkler, lD., 1997d, Expression of
VEGF in synovial fibroblasts is induced by hypoxia and interleukin-lb, J. Rheum. in
press.

Jackson, lR., Seed, M.P., Kircher, C.R., Willoughby, D.A. and Winkler, lD., 1997e, The
co-dependence of angiogenesis and chronic inflammation, FASEB J. 11:457-465.

Kimura, M., Suzuki, J. and Amemiya, K, 1985, Mouse granuloma pouch induced by
freund's complete adjuvant with croton oil, J Pharmacobio-Dyn 8:393-400.

Kimura, M., Amemiya, K, Yamada, T. and Suzuki, l, 1986, Quantitative method for
measuring adjuvant-induced granuloma angiogenesis in insulin-treated diabetic mice, J
Pharmacobio-Dyn 9:442-446.

Li, l, Perrella, M.A., Tsai, l-C., Yet, S.-F., Hsieh, C.-M., Yoshizumi, M., Patterson, c.,
Endege, W.O., Zhou, F. and Lee, M.-E., 1995, Induction fo vascular endothelial growth
factor gene expression by interleukin-lb in rat aortic smooth muscle cells, J Bioi Chem
270:308-312.

Nagashima, M., Yoshino, S., Ishiwata, T. and Asano, G., 1995, Role of vascular endothelial
growth factor in angiogenesis of rheumatoid arthritis, J Rheum 22:1624-1630.

Nickoloff, B.l, Mitra, R.S., Varani, l, Dixit, V.M. and Polverini, P.l, 1994, Aberrant
production of interleukin-8 and thrombospondin-l by psoriatic keratinocytes mediates
angiogenesis, Am J PathoI144:820-828.

Seed, M.P., Colville-Nash, P.R., Jackson, J.R. and Winkler, lD., Angiogenesis in
inflammation. In: T.-P. Fan and R. Auerbach (Eds.), The new angiotherapy, Humana Press,
1997, pp. In press.

Shima, D.T., Adamis, A.P., Ferrara, N., Yeo, KT., Yeo, T.K, Allende, R., Folkman, J. and
Damore, P.A., 1995, Hypoxic induction of endothelial-cell growth-factors in retinal cells -
identification and characterization of vascular endothelial growth-factor (vegf) as the
mitogen, Molecular Medicine 1: 182-193.

Shweiki, D., Neeman, M., Itin, A. and Keshet, E., 1995, Induction of vascular endothelial
growth-factor expression by hypoxia and by glucose deficiency in multicell spheroids -
implications for tumor angiogenesis, Proc Natl Acad Sci USA 92:768-772.

Spisni, E., Manica, F. and Tomasi, V, 1992, Involvement of prostanoids in the regulation
of angiogenesis by peptide growth factors, Pros. Leuko. Essen. Fatty Acids 47: 111-115.

Sunderkotter, C., Steinbrink, K, Goebeler, M., Bhardwaj, R. and Sorg, c., 1994,
Macrophages and angiogenesis, J Leuk Bioi 55:410-422.

120
ANGIOGENIC MEDIATORS IN WOUND HEALING

Luisa A. DiPietro and Nicholas N. Nissen

Bum and Shock Trauma Institute, Department of Surgery, Loyola

University Medical Center, Maywood, IL 60153, USA.

1. INTRODUCTION

Angiogenesis is an essential component of normal wound repair. Neovascularization

provides nutrient support to healing tissue, promotes granulation tissue formation, and
assists in the clearance of debris. Despite the prominent role of angiogenesis in the repair
process, the primary mediators of wound angiogenesis have been difficult to define. An
abundance of potential angiogenic agonists and antagonists have been identified in healing
wounds (Table I), suggesting that the net angiogenic stimulus relies upon a balance of
positive and negative mediators. The identification of this rather plentiful number of
candidate mediators seems to indicate that angiogenesis in wounds is regulated by a most
redundant and complex mechanism.
Neovascularization within wounds, as in all tissues, depends upon many factors, including
cell to cell interaction, cell to extracellular matrix interaction, and the balance between pro-
angiogenic and anti-angiogenic soluble factors. Our efforts have been directed at identifying
the soluble pro-angiogenic factors that are critical to normal wound healing. As part of this
undertaking, we began to characterize the angiogenic profile of fluid collected from human
surgical wound drains placed at the time of various operations (Nissen, Polverini, Gamelli,
and DiPietro, 1996). We reasoned that wound fluid would be generally representative of
the growth environment of the wound, as it would be expected to contain soluble pro-
angiogenic growth factors at levels similar to that of the wound bed itself. While the
proliferative capacity of wound fluid has been widely reported, studies of the net
angiogenic stimulus in this fluid, and of the specific mediators responsible for this stimulus,
are few. In a rabbit model, wound fluid derived from subcutaneous wound chambers at 21
days after implantation simulated angiogenesis in vivo (Banda, Knighton, Hunt, and Werb,
1982) and endothelial cell proliferation in vitro (Greenburg and Hunt, 1978). This activity
is believed to be reflective of macrophage-derived angiogenic factors, as the macrophage is
the predominate cell type in the chamber at this time. Wound macrophages themselves are
angiogenic and produce a variety of cytokines and angiogenic factors (Hunt, Knighton,
Thakral, Goodson, and Andrews, 1984; Rappolee, Mark, Banda, and Werb, 1988).

Angiogenesis: Models. Modulators. and Clinical Applications

Edited by Maragoudakjs, Plemll11 Press, New York, 1998 121
 Table I. Angiogenic mediators that are present in healing wounds.
Promoters Inhibitors

bFGF . Thrombospondin
EGF ~ TIMP-l, TIMP-2
TGF-alpha ~ nitric oxide
TGF-beta 1p53
TNF-alpha
VEGF
thrombin
Prostaglandin E2
Leukotrienes

Although the angiogenic environment of the early wound has not been as well described as
the later wound, early human wound fluid has been shown to stimulate endothelial
proliferation (Katz, Alvarez, Kirsner, Eaglstein, and Falanga, 1991). In addition, early
human wound fluid contains a variety of potential angiogenic mediators, including tumor
necrosis factor-alpha, epidermal growth factor, platelet derived growth factor, and basic
fibroblast growth factor (bFGF) (Dvonch, Murphey, Matsuoka, and Grotendorst, 1992;
Grayson, Hansborough, Zapata-Sirvent, Dore, Morgan, and Nicolson, 1993).

2. ANGIOGENIC ACTIVITY IN EARLY WOUND FLUID

To discern the angiogenic activity of human wound fluid, wound fluids were tested for the
ability to 1) induce endothelial cell proliferation, 2) stimulate endothelial cell chemotaxis
and 3) induce angiogenesis in vivo in the corneal pocket assay (Nissen et al ., 1996).
Surprisingly, surgical drain fluid collected immediately after injury showed significant
angiogenic activity. This early wound fluid stimulated endothelial cell proliferation, was
chemotactic, and caused new vessel formation in the in vivo rat corneal assay. The rapid
generation of a positive angiogenic stimulus in wound fluid was remarkable, as the in situ
synthesis of growth factors is unlikely to reach significant levels within such a short
period. Thus, the finding that wound fluid collected immediately after surgery was
angiogenic strongly suggested that this early activity was provided by a preformed growth
factor. Several previous studies, both in vitro and in vivo, suggested that a likely mediator
of this early angiogenic activity was basic fibroblast growth factor (bFGF), or FGF-2.
bFGF can be sequestered within cells, and in vitro, these intracellular stores of bFGF can
be released during cell injury, lysis, and death (Gajdusek and Carbon, 1989;
Muthukrishnan, Warder, and McNeil, 1991). bFGF may also be stored in the extracellular
matrix, where it can be released by the action of several proteases found in early wounds
(Bashkin, Doctrow, Klagsbrun, Svahn, Folkman, and Vlodavsky, 1989; Villaschi and
Nicosia, 1993). Such findings have lead to speculation that bFGF could act as an early
stimulus for repair in injured tissue.
To assess the likelihood that bFGF plays a role in the immediate angiogenic stimulus found
in wound fluid, the levels of bFGF in human wound fluids were determined by ELISA
analysis. The bFGF content of wound fluid was maximal immediately after surgery

122
Table II. Effect of neutralizing anti-bFGF antibody on angiogenic activity of early wound
fluids. *
Assay % of control response*

Endothelial cell ~
proliferation
1 33 %

Endothelial cell
Il
chemotaxis
1 56%

In vivo angiogenesis ~ 42%

(corneal assay) !
*Neutralizing antibody to human bFGF was added to wound fluids prior to assay
tControl response = proliferative, chemotactic, or corneal response of untreated wound
fluid

(824+/-505 pglml), falling by 60% at day 1, by 80% at day 2, and to near serum levels (16
+/- 17 pglml) thereafter. The role of bFGF as an angiogenic mediator in early human
wound fluid was further defined by neutralization experiments. The addition of neutralizing
antibody to bFGF decreased the proliferative activity of early wound fluid to near that of
control, and substantially decreased both the chemotactic and the in vivo angiogenic
activity (Table II). Taken together, these findings support the hypothesis that preformed
stores of bFGF are released shortly after injury, and that angiogenic activity of early
wound fluids is largely due to bFGF.
The source and synthesis of bFGF within tissues has been a subject of many studies.
More specifically, several previous studies have provided evidence for a role for bFGF in
healing wounds. bFGF mRNA is temporally regulated in animal wound models
(Antoniades, Galanopoulos, Neville-Golden, Kiritsy, and Lynch, 1994; Werner, Peters,
Longaker, Fuller-Pace, Banda, and Williams, 1992; Werner, Breeden, Hubner, Greenhalgh,
and Longaker, 1994). More significantly, antibody to bFGF decreases granulation tissue
formation in subcutaneous implants (Broadley, Aquino, Woodward, Buckley-Sturrock,
Sato, Rifkin, and Davidson, 1989).
Our results demonstrate a temporal pattern of bFGF release and activity that is highest in
early surgical wounds. It is possible that after the clearance of soluble bFGF, the active
synthesis of this growth factor continues within the wound. This possibility is supported
by reports that bFGF mRNA can be detected in healing wounds for many days after injury
(Antoniades et aI, 1994; Werner et aI., 1992). However, because bFGF lacks a signal
peptide, secretion of this molecule from cells may be delayed (Abraham, Mergia, Whang,
Tumolo, Friedman, Hjerrild, Gospodarowicz, and Fiddes,1986). Therefore, although active
bFGF synthesis may occur within wounds, a net increase in bioactive extracellular bFGF
may not follow.
Our studies regarding the angiogenic activity of early wound fluid are in agreement with
previous reports that early wound fluid supports endothelial cell proliferation ( Katz et aI,

123
1991). Our studies extend these observations by providing evidence that early wound fluid
also contains chemotactic activity and stimulates in vivo angiogenesis. Further, our results
begin to define a specific role for bFGF in the early wound environment. Interestingly, our
finding that early wound fluid is angiogenic in the in vivo corneal assay is in contrast to the
previous findings of Arnold, West, Schofield, and Kumar (1987). Using the chick
chorioallantoic membrane assay, these investigators showed that early wound fluid was not
angiogenic, primarily due to a high level of angiogenic inhibitors. This discrepancy may
relate to methodological differences in the assay systems.

3. ANGIOGENIC ACTIVITY IN LATER WOUND FLUID

°
An examination of wound fluids from the later time points of days 2 through 6
demonstrated that, similar to day and day 1, these fluids were also potently angiogenic.
Nearly all fluids tested, be they from day 0, 1, 2, 3, 4, 5 or 6, exhibited the capacity to
induce endothelial cell chemotaxis, and were angiogenic in the corneal assay. This finding
was certainly not surprising. In rodent models of excisional wound healing, capillary
density peaks at 7 to 10 days after injury (Figure 1). Although soluble bFGF levels
rapidly decline in wounds, a sustained angiogenic stimulus would be needed to promote the
angiogenic process. This stimulus most likely arises from the synthesis of additional
positive angiogenic factors within the wound itself.

15
~

..
~
:; 10
u
•
> 5
~

0
s 3 5 7 10 14 21 28

Post-operltlve DIY

Figure 1. Rate of neovascularization in murine full-thickness dermal wounds. Capillary

density was measured by image analysis of sections that had been immunostained with
endothelial specific antibodies.

Many cell types within the wound are capable of synthesizing angiogenic factors, yet
several lines of evidence point to the tissue macrophage as a key contributor. Studies in
guinea pigs by Leibovich and Ross (1975) have shown that macrophages are essential to
optimal wound repair. Further, activated macrophages, including wound macrophages are
potently angiogenic (Thakral, Goodson, and Hunt, 1979). In vitro, macrophages make a
plethora of angiogenic chemokines, including IL-8, TGF-beta, and vascular endothelial cell
growth factor (VEGF).
While the prolonged angiogenic stimulus in wounds may include several different angiogenic
mediators, VEGF seems particularly likely to be involved. VEGF has direct angiogenic
effects similar to bFGF, and is capable of stimulating endothelial cell migration and
activation in vitro and angiogenesis in vivo (Dvorak, Brown, Detma, and Dvorak, 1995;
Ferrara, Houck, Jakeman, and Leung, 1992). VEGF has been implicated in wound repair,
as VEGF is produced by keratinocytes in skin wounds in mouse, rat, and guinea pig models

124
 \ '[ F,
TGF-~,
other? apillary
growth

---
.... Proliferation ..... Migration ..... Capillary Formation ... Resolution

Figure 2. A model for the production of pro-angiogenic growth factors in wounds.

(Brown, Kiang-Teck, Berse, Yeo, Senger, Dvorak, and Van De Water, 1992; Frank, Hubner,
Breier, Longaker, Greenhalgh, and Werner, 1995). Further, VEGF is produced by many
cells in response to hypoxia in vitro systems (Schweiki, Itin, Soffer, and Keshet, 1992),
suggesting that the hypoxic environment of the wound would favor VEGF production. In
accordance with this prediction, high levels of VEGF are found human wound fluids at
days 3 through 6 post-operative (5.3 +/- 3.1 nglml at day 6) (LAD and NNN, unpublished
observations). Finally recent studies in a pig model of omental repair have suggested that
VEGF is a critical mediator in this system (Gupta, McNeil, Riegner, and Howdieshell,
1996). To date, the role of VEGF in human surgical wounds, which may differ
substantially from animal skin injury models, has not been clearly defined.

4. A MODEL OF WOUND ANGIOGENESIS

Our studies suggest the following model for the regulation of wound angiogenesis by
soluble pro-angiogenic growth factors (Figure 2). Tissue injury causes the immediate
release ofbFGF, providing a nearly instantaneous angiogenic stimulus. This stimulus can
initiate cell proliferation, and may serve as an anti-apoptotic signal to endothelial cells.
Within a day or two of injury, cells within the wound, such as macrophages, begin to
synthesize new angiogenic mediators such as TGF-beta and VEGF. The stimulus for this
production may be hypoxia, cytokines, or other inflammatory mediators within the wound
(Knighton, Hunt, Scheuenstuhl, Halliday, Werb, and Banda, 1983). This second wave of
positive angiogenic factors would cause endothelial cell chemotaxis into the wound bed, and
would promote capillary formation. As wound healing proceeds, the production of these
mediators would slow as granulation tissue formation occurs and tissue hypoxia
diminishes.

5. CONCLUSIONS

There remain many unanswered questions regarding the generation of soluble positive
angiogenic factors in wounds. First, the source of bFGF in the early wound is not yet
clear. We and others have suggested the possibility that bFGF is released from damaged

125
and dying cells within the injured tissue. However, the total bFGF within normal uninjured
tissue is substantially less that than found in a similar amount of injured tissue. One
possibility is that the injured area derives bFGF from a large surrounding area of normal
tissue. Another possibility is that platelets contribute bFGF at the wound site. Recent
reports suggest that platelet degranulation may account for some of the early bFGF in
wound fluid. However, the low levels of bFGF reported in pure platelet preparations
suggest that additional sources are needed (Brunner, Nguyen, Gabrilove, Rifkin, and
Wilson, 1993). A final potential source is serum itself. Although serum contains little
bFGF (Nissen et al., 1996), it is conceivable that the wound acts as a trap for bFGF as
serous exudate flows though the wound bed.
Another question that arises from our studies is the relative importance of soluble versus
bound bFGF in regulating wound angiogenesis. While the amount of bFGF in the fluid
phase of the wound drops precipitously after the first 24 hours, the level of bFGF than
remains bound to the extracellular matrix or to cell surfaces within the wound bed is not yet
known. Healing wounds that are subjected to immunostaining for bFGF show that
substantial bFGF remains present in the wound bed long after the observed drop in soluble
bFGF (Gibran, Isik, Heimbach, and Gordon, 1994; Kurita, Tsuboi, Ucki, Rifkin, and
Ogawa, 1992; Whitby and Ferguson, 1991). The function of this bound bFGF, and the
relative amounts of soluble and bound bFGF within the wound, have not yet been
established.
A final controversy involves the characterization of the primary angiogenic factors in the
later phases of wound healing. While experimental evidence suggests that VEGF may play
a role in mediating wound angiogenesis, several other mediators may also be involved. In
summary, our knowledge of the predominate pro-angiogenic mediators in wounds, and the
time course of their production, has been greatly expanded by recent studies. Many
intriguing questions remain to be answered before the complex process of wound
angiogenesis can be fully understood.

ACKNOWLEDGMENTS

This work was supported by the Dr. Ralph and Marion C. Falk Foundation and the
National Institutes of Health (GM50875).

REFERENCES

Abraham lA., Mergia, A., Whang, l, Tumolo, A., Friedman, J., Hjerrild, K.A.,
Gospodarowicz, D., and Fiddes, le., 1986, Nucleotide sequence of a bovine clone
encoding the angiogenic protein basic fibroblast growth factor, Science 233:545-548.

Antoniades, H.N., Galanopoulos, T., Neville-Golden, l, Kiritsy, c.P., and Lynch, S.P.,
1994, Expression of growth factor and receptor mRNAs in skin epithelial cells following
acute cutaneous injury, Am. J. Pathol. 142:1099-1110.

Arnold, F., West, D.C., Schofield, P.F., and Kumar, S., 1987, Angiogenic activity in human
wound fluid, Int. J. Microcirc. Clin. Exp. 5:381-386.

126
Banda, MJ., Knighton, D.R., Hunt, TK., and Werb, Z., 1982, Isolation of a nonmitogenic
angiogenesis factor from wound fluid, Froc. Natl. Acad. Sci. USA 79:7773-7777.

Bashkin, P., Doctrow, S., Klagsbrun, M., Svahn, C.M., Folkman, 1, and Vlodavsky, I.,
1989, Basic fibroblast growth factor binds to subendothelial extracellular matrix and is
released by heparitinase and heparin-like molecules, Biochemistry 28: 1737-1743.

Broadley, K, Aquino, A, Woodward, S., Buckley-Sturrock, A, Sato, Y., Rifkin, DR, and
Davidson, 1M., 1989, Monospecific antibodies implicate basic fibroblast growth factor in
normal wound repair, Lab. Invest. 61 :571-575.

Brown, LF., Kiang-Teck, Y., Berse, B, Yeo, T., Senger, D.R., Dvorak, H.F., and Van De
Water, L., 1992, Expression of vascular permeability factor (vascular endothelial growth
factor) by epidermal keratinocytes during wound healing, 1. Exp. Med. 176: 1375-1379.

Brunner, G., Nguyen, H., Gabrilove, J., Rifkin, D.B., and Wilson, E.L, 1993, Basic
fibroblast growth factor expression in human bone marrow and peripheral blood cells,
Blood 81:631-638.

Dvonch, V.M., Murphey, R.I., Matsuoka, J., and Grotendorst, G., 1992, Changes in
growth factor levels in human wound fluid, Surgery 112:18-23.

Dvorak, H.F., Brown, LF., Detmar, M., and Dvorak, AM., 1995, Vascular permeability
factor/vascular endothelial growth factor, microvascular hyperpermeability, and
angiogenesis, Am. 1. Path. 146: 1029-1 039.

Ferrara, N., Houck, K, Jakeman, L., and Leung, D., 1992, Molecular and biological
properties of the vascular endothelial growth factor family of proteins, Endocrine Rev.
13: 18-32.

Frank, S., Hubner, G., Breier, G., Longaker, M.R, Greenhalgh, D.G., and Werner, S., 1995,
Regulation of vascular endothelial growth factor expression in cultured keratinocytes, 1.
Bioi. Chem. 270:12607-12613.

Gajdusek, CM., and Carbon, S, 1989, Injury-induced release of basic fibroblast growth
factor from bovine aortic endothelium, 1. Cell. Physiol.139: 570-579.

Gibran, N.S., Isik, F.F., Heimbach, D.M., and Gordon, D., 1994, bFGF in the early human
burn wound. 1. Surg. Res. 56:226-234.

Greenburg, G.B., and Hunt, T.K , 1978, The proliferative response in vitro of vascular
endothelial and smooth muscle ce\1s exposed to wound fluids and macrophages, 1. Cell.
Physiol. 97:353-360.

Grayson L.S., Hansborough, 1F., Zapata-Sirvent, RL, Dore, CA, Morgan, 1L and
Nicolson, M.A, 1993, Quantitation of cytokine levels in skin graft donor site wound fluid,
Burns 19:401-405.

Gupta, V.K., McNeil, P.L, Riegner, C, and Howdieshell, T.R., 1996, Vascular endothelial
growth factor is a key mediator of omental angiogenesis, Surg. Forum 47:746-749.

127
Hunt, TK, Knighton, D.R., Thakral, K.K., Goodson, W.H. III, and Andrews, W.S., 1984,
Studies on inflammation and wound healing: Angiogenesis and collagen synthesis
stimulated in vivo by resident and activated wound macrophages, Surgery 96:48-54.

Katz, M.H, Alvarez, A.F., Kirsner, R.S., Eaglstein ,W.H, and Falanga, V., 1991, Human
wound fluid from acute wounds stimulates fibroblast and endothelial cell growth, J. Amer.
Acad Dermatof. 25:1054-1058.

Knighton, DR, Hunt, TK, Scheuenstuhl, H, Halliday, B.I., Werb, Z., and Banda, M.J.,
1983, Oxygen tension regulates the expression of angiogenesis factor by macrophages,
Science 221:1283-1285.

Kurita, Y., Tsuboi, R., Ucki, R., Rifkin, D.B., and Ogawa, H., 1992, Immunohistochemical
localization of basic fibroblast growth factor in wound healing sites of mouse skin, Arch.
Dermatol. Res. 284: 193 -197.

Leibovich, S.J., and Ross, R., 1975, The role of the macrophage in wound repair. A study
with hydrocortisone and anti macrophage serum, Am. J. Pathol. 78 :71-91.

Muthukrishnan, L., Warder, E., and McNeil, P., 1991, Basic fibroblast growth factor is
efficiently released from a cytosolic storage site through plasma membrane disruptions of
endothelial cells, J. Cell. Physiol. 148: 1-16.

Nissen, N.N., R.L. Gamelli, PJ. Polverini, and DiPietro, L.A., 1996, Basic fibroblast
growth factor mediates angiogenic activity in early surgical wounds, Surgery 119:457-465.

Rappolee, D.A., Mark, D., Banda, MJ., and Werb, Z., 1988, Wound macrophages express
TGF-a and other growth factors in vivo: analysis by mRNA phenotyping, Science
241 :708-712.

Schweiki, D., Itin, A., Soffer, D., and Keshet, E., 1992, Vascular endothelial growth factor
induced by hypoxia may mediate hypoxia-initiated angiogenesis, Nature 359:843-845 .

Thakral, KK, Goodson, W.H III, and Hunt, TK, 1979, Stimulation of wound blood
vessel growth by wound macrophages, J. Surg. Res. 26:430-436.

Villaschi, S., and Nicosia, R.F., 1993, Angiogenic role of endogenous basic fibroblast
growth factor released by rat aorta after injury, Am. J. Pathol. 143: 181-190.

Werner, S., Peters, K.G., Longaker, M.T., Fuller-Pace, F., Banda, MJ., and Williams, L.T.,
1992, Large induction of keratinocyte growth factor expression in the dermis during wound
healing, Proc. Natl. Acad Sci. USA 89:6896-6900.

Werner, S., Breeden, M., HUbner, G., Greenhalgh, D.G., and Longaker, M.T., 1994,
Induction of keratinocyte growth factor expression is reduced and delayed during wound
healing in the genetically diabetic mouse, J. Invest. Dermato!' 103:469-473.

Whitby, DJ., and Ferguson, M.W., 1991, Immunohistochemical localization of growth

factors in fetal wound healing. Developmental Bioi. 147:207-215.

128
MODULATION OF VASCULAR ENDOTHELIAL GROWTH FACTOR
ANGIOGENIC ACTIVITY BY ADP-RIBOSYLA TION

John J. Feng, * Q. Perveen Ghani, Gabriel Ledger, Rahmat Barkhordar and

Thomas K. Hunt, * M. Zamirul Hussain.

Departments of Restorative Dentistry and Surgery, * University of

California, San Francisco, California, USA 94143-0758.

INTRODUCTION

Angiogenesis is a complex series of interdependent events which basically involves

chemotaxis, proteolytic digestion of extracellular matrix, cell proliferation, vascular tube
formation, anastomoses with other vascular sprouts, and cell-cell, and cell-matrix adhesion.
Studies on angiogenesis have revealed that these processes are regulated by a balance of
positive and negative inducers of which cytokines and growth factors are the primary
stimulators (1-3). Research over the last few years have indicated the pivotal role of
vascular endothelial growth factor (VEGF) in the regulation of normal and abnormal
angiogenesis (3). Because of the specificity of VEGF for endothelial cell migration,
proliferation and tube formation, VEGF is considered to be a major stimulus for wound
angiogenesis.

Little is known about how the angiogenic activity of VEGF is regulated. Generally
the angiogenic potential is assessed by measuring VEGF mRNA and VEGF synthesis.
This assumes that once synthesized, all VEGF molecules are biologically active.
Observations indicating inactivity or angiogenically less potent VEGF in some conditions
are beginning to emerge. This is true of VEGF produced by macrophages maintained in
normoxic environment (4). Furthermore, in some tumor models, VEGF found during the
premalignant stage to the intermediate stage, is inactive, which in the malignant stage,
becomes highly angiogenic (5). The biochemical reason for this behavior is not known.
While several putative explanations particularly involving endothelial cell response can be
conceived, no unifying postulate has yet emerged.

We considered the possibility of post-translational modification of VEGF

polypeptide by ADP-ribosylation after we found that conditioned media derived from
macrophage cultures exposed to 15 mM lactate enhanced angiogenesis in animal models
even when de novo VEGF synthesis was blocked by cyc10hexamide A number of
cytoplasmic protein functions are known to be regulated by mono-ADP-ribosylation (6).

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenwn Press, New York, 1998 129
We hypothesized that VEGF is reversibly ADP-ribosylated in the cell and that ADP-
ribosylated VEGF is angiogenically less potent. Our results demonstrate that recombinant
human VEGF 165 as well as macrophage VEGF are avid acceptors of ADP-ribose in the
presence of enzymes which transfer ADP-ribosyl moiety of NAD+ to the arginyl groups
ofVEGF. Furthermore, the angiogenic activity of ADP-ribosylated VEGF was found to be
significantly reduced.

METHODS FOR MEASUREMENT OF ANGIOGENESIS

A number of in vitro and in vivo methods are currently used to determine

angiogenesis. Each procedure has its advantages and weaknesses. Wound healing
investigators have generally relied on in vivo assays for angiogenesis research. We used
two in vivo methods to evaluate angiogenic activity ofVEGF and modified VEGF.

Rabbit Corneal Assay

The rabbit corneal assay is widely used in wound angiogenesis research. It provides
a reliable, reproducible and semi-quantifiable measure of newly formed blood vessels (7).

Test samples or serum-free conditioned media from cultured cells, is dialyzed,

lyophilized, reconstituted in phosphate buffered saline, mixed with Hydron and vacuum
dried into discs each containing approximately 10 ng of protein in 10 /11 (8). Discs are
implanted into a rabbit corneal stroma micropocket 3 mm from the limbus under anesthesia.
Corneas are examined for maximal peak angiogenic response, which occurs around 9 days,
and scored (0-5) utilizing an anatomic grading scale: 0 = No vessels; 1 = 1-5 vessels, no
broad response; 2 = 114 distance to pellet, broad response; 3 = 112 distance, broad
response; 4= up to edge of pellet, broad response; 5 = encompassing pellet, broad
response.

Murine Matrigel Assay

Although Matrigel assay was originally developed as an in vitro test for endothelial
cell tube formation, murine Matrigel system is now being utilized as an in vivo assay for
angiogenesis (9). Matrigel appears to be a suitable substrate for endothelial cell migration
and response because of its similarity with the biologic basement membrane and matrix.
This assay measures the early infiltration of endothelial cells (evidenced by Factor VIII
staining), their organization into tube-like structures and appearance of intravasculature
erythrocytes (10, 11).

Matrigel (Collaborative Biomedical, Bedford, MA), a reconstituted basement

membrane complex containing primarily laminin and type IV collagen, is isolated from the
Engelbreth-Holm-Swarm (EHS) murine tumor (10, 12). At 4°C, Matrigel is liquid; but
when injected subcutaneously into a mouse it reconstitutes as a gel. Test samples are
dialyzed and lyophilized as for the corneal assay, mixed with the Matrigel (10% v/v
reconstituted in 1 m1 Matrigel) on ice and then injected subcutaneously into a mouse
dorsum (5 month old, female Swiss-Webster). After one week, the mouse is sacrificed and
the gel plugs (two per mouse) harvested. Samples are fixed in buffered formalin, paraffin
embedded, sectioned (4/1m), and H&E stained (11). Angiogenic activity is expressed as
mean score ± standard error. Matrigel scoring system: 0 = No tube-like structure, 0.5 =
Incomplete tube formation, 1 = Well-formed tubes throughout Matrigel plug.

130
 Autoradlograph Western Blot
12 r--=,=~-­

••211

II

ABCDEF
G H 1 J K

A; 2119 VEGF, 37"C G: No VEGF (D)

B: 1119 VEGF, 37"C H: 1119 VEGF, 4°C (C)
C: 1119 VEGF, 4°C I; 1119 VEGF, 37"C (B)
D: NoVEGF J: 2119 VEGF, 37°C (A)
E: 2119 VEGF, 37"C } K; 1119 VEGF, 37°C + NH20H (F)
+ NH 2 0H
F; 1119 VEGF, 37"C

Figure 1. In vitro ADP-ribosylation of VEGF 165 by cholera toxin.

Incorporation of radiolabeled ADP-ribose from NAD+ is time and
concentration dependent. Protamine was used as the positive control.

RESULTS

Our experimental approach consisted of a combination of in vitro and in vivo

experiments: (i) Recombinant human VEGF and macrophage derived VEGF were tested
for the ability to accept ADP-ribose from NAD+ by ADP-ribosyl transferase. (ii) Two in
vivo angiogenesis assays were used to determine the effect of ADP-ribosylation on VEGF
angiogenic activity. (iii) The ADP-ribosylated form ofVEGF in cultured macrophages was
analyzed by the incorporation of radiolabeled precursor of ADP-ribose into VEGF bound
ADP-ribose.

120
400 1 mM 32p-NAD+
0.75 119 YEGF, 20 119 cholera toxin
1 mM 32p-NAD+ 30 minutes
20 11 9 cholera toxin
300 80
(32PI-YEGF
(pmole)
200
40
100

0
0 0.1 0.5 0.75 1.0
0 30 60 90
VEGF 1119)
Time (min)
61 ± 7 30 minutes
Protamine 10 119
114 ± 13
20 j.l9

Figure 2. SDS-Page of ADP-ribosylated VEGF and corresponding Western blot.

Hydroxylamine (NHzOH) treatment removed ADP-ribose from VEGF
Western blot analysis demonstrated intact VEGF before and after NHzOH
treatment.

131
 120
VEGF. 0.25 119. 90 min
100 Cholera toxin = 106
pmoles
80
FP}VEGF
(pmole) 60

40

20

Macrophage Homogenate (}tIl

Figure 3. In vitro ADP-ribosylation ofVEGF 165 by macrophage enzyme is concentration

dependent.

1
* P < 0.05
Matri gel
Angiogenesi s
Score 0.5
(mean .. SEM)

o
VEGF A OP· A~ AOP-A_ ADP·A,_ A-"'
(n • 11) VEGF CIuW<I CIHwd ConVOl
(n. '5) en. "~I • on VEGF In. I)
anuboc!y
(n. '5)

Figure 4. Angiogenic activity ofVEGF and modified VEGF.

ADP-ribosylation decreases VEGF angiogenic activity. Removal of ADP-
ribose restores activity that is inhibited by anti-VEGF antibody.

t Lactate~ pyruvate
LDH

{ NADH
_ _.. ~.A.DP.rlbose I
INAO+ ~ ~
, ADPR.T Active VeGF 'II Inactive VEGF

Figure 5. Hypothetical scheme for regulation of VEGF angiogenic activity by ADP-

ribosylation.
Increase in lactate lowers NAD+ stores and lowers ADP-ribosylation ofVEGF.
Decrease in ADP-ribosylation releases inhibition of angiogenic activity. LDH
= lactate dehydrogenase; ADP-RT = ADP-ribosyl transferase.

132
Table 1. In vivo ADP-ribosylation ofVEGF in macrophages.
A significant amount ofVEGF is ADP-ribosylated. Lactate treatment reduces
ADP-ribosylation of VEGF. Oxamate blocks this effect.

Treatment Radioactivity in VEGF

(dpm x 10')
None 3.2 ±0.3
Lactate, 15 mM 1.7 ± 0.2
Lactate, 15 mM + 2.9 ±0.3
Oxamate, 20 mM

• Macrophage. (6 x 10') were labeled wi1h 1~Ci/mL of "C·

adenosine for 16 h with the above conditions.
• VEGF was isolated from conditioned media using antl-VEGF
antibody affinity column.
• Radlolabel in VEGF fraction was characterized as ADPR and
total radioactivity measured.

(i) Cholera toxin is often used to catalyze the mono-ADP-ribosylation reaction onto
arginine residues of target proteins (6). Figure 1 demonstrates the time and dose dependent
manner by which cholera toxin catalyzes ADP-ribosylation of recombinant human
VEGF 165 in the presence of [32 p ]_NAD+ The results indicate that VEGF can increasingly
accept ADP-ribose in a time-dependent fashion. In addition, increasing the amount of
VEGF also increases the amount of radiolabeled, ADP-ribosylated VEGF. It was
estimated that 12 out of 20 total arginyl residues of VEGF homodimer were modified by
ADP-ribose. The modification increased the molecular weight and the corresponding
electrophoretic mobility ofVEGF (Figure 2). The ADP-ribose on VEGF was removed by
the action of snake venom phosphodiesterase and hydroxylamine without any adverse
effect on VEGF activity. This is consistent with the ADP-ribosyl group attached to the
arginyl moiety on other proteins (6, 13).

Figure 3 shows that macrophages contain the enzyme that can ADP-ribosylate
recombinant VEGF ,65 . In subsequent experiments, we found that macrophage-derived
VEGF can also accept ADP-ribose in the presence of cholera toxin or its own endogenous
ADP-ribosyl transferase. ADP-ribosylated VEGF was hydrolyzed by macrophage
homogenate to yield ADP-R indicating the possibility of a reversible VEGF modification.

(ii) Next we evaluated the effect of ADP-ribosylation on VEGF 165 and macrophage-
derived VEGF activity. Figure 4 illustrate the angiogenic activity of unmodified VEGF and
ADP-ribosylated-VEGF in the mouse Matrigel angiogenesis assay. It is clear that ADP-
ribosylation significantly decreases the original angiogenic activity of recombinant and
macrophage VEGF. In addition, removal of ADP-ribose from ADP-ribosylated-VEGF
returns angiogenic activity to control levels. Similar results were observed in the rabbit
corneal assay (data not shown).

(iii) To detect ADP-ribosylated VEGF in macrophages, quiescent cultures were

labeled with [14C]-adenosine for 16 hand [14C]ADP-ribosylated- VEGF was recovered
from conditioned media by affinity chromatography using anti-VEGF monoclonal
antibody. As shown in Table 1, an appreciable amount of released VEGF contained
covalently bound ADP-ribose which corresponded to the total ADP-ribosylation activity
of the cell. Earlier, we reported that incubation of macrophages with 15 mM lactate
decreases cellular ADP-ribosylation by depleting NAD+ (14). As shown in the table,
ADP-ribose on VEGF was significantly decreased when macrophage cultures were
maintained in the presence of 15 mM lactate. However, the lactate effect was reversed

133
upon simultaneous addition of oxamate, which competes against lactate for the enzyme
lactate dehydrogenase. In this case, the level of ADP-ribose on VEGF returns to that of
control value.

DISCUSSION

Results of the present study demonstrate that VEGF can undergo enzymatic mono-
ADP-ribosylation and that the reversible post-translational ADP-ribosylation modulates
VEGF angiogenic activity. These findings provide the basis of possible existence of a novel
mechanism to regulate VEGF mediated angiogenesis. The basic idea for this line of VEGF
research is not new. Glycosylation was conceived to influence VEGF biological activity .
However, when tested, glycosylation did not exhibit any regulatory effect (I5, 16). Other
bioactive polypeptides such as ACTHI-24 (corticotropin), TSH (thyroid stimulating
hormone), FSH (follicle stimulating hormone), LH (leutinizing hormone), hCG (human
chorionic gonadotropin) have been reported to act as acceptors for cholera toxin-catalyzed
ADP-ribosylation (17). However, the functional significance of ADP-ribosylation of these
pituitary hormones remains to be elucidated. It appears that the conformational change of
VEGF polypeptide caused by ADP-ribose linkage is sufficient to alter its angiogenic
activity. We assume that loss of angiogenic potential of ADP-ribosylated VEGF is related
to an inadequate endothelial cell migration, receptor binding and/or proliferation. Studies to
evaluate which of these processes are adversely affected by ADP-ribosylation are
underway . It is also important to determine which arginines in the VEGF sequence are
ADP-ribosylated .

Figure 5 depicts the role of reversible mono-ADP-ribosylation to regulate VEGF

angiogenic activity. The activity ofVEGF is down-regulated by ADP-R which modifies the
conformation of active VEGF polypeptide. This follows that normally VEGF molecules
exist as a mixture offree (angiogenic) and ADP-ribosylated (poorly angiogenic) forms. The
ratio of these forms which normally determines the angiogenic potential, is sensitive to
metabolic alterations involving a change in ADP-R synthesis. We suggest this ratio to be
different in different physiologic and pathologic conditions. As for example, it is
anticipated that highly angiogenic tissues have greater proportion of the free form while
VEGF produced in normoxic conditions may constitute higher proportion of ADP-
ribosylated form.

In conclusion, a reversible modification of VEGF by mono-ADP-ribosylation

affects angiogenic activity . This structural alteration has the potential to regulate
angiogenesis because ADP-ribose synthesis is intimately linked to the metabolic status of
growing tissue. Thus this mechanism provides a novel means to modulate VEGF mediated
angiogenesis through metabolic means. Furthermore, our findings provide a new approach
to study the regulation of the activity of other biologically active polypeptides.

REFERENCES

I. Battegay EJ. Angiogenesis: mechanistic insights, neovascular diseases, and

therapeutic prospects. J Mol Med 1995; 73, 333-346.

2. Folkman J. Angiogenesis. J Bioi Chem 1992: 267, 10931-10934.

3. Ferrara N, Davis-Smyth T. The biology of vascular endothelial growth factor.

Endocrine Reviews 1997;18(1):4-25.

134
4. Leibovich S1. Presented at 1997 Gordon Conf on Angiogenesis.

5. Shweiki D, Neeman M, Itin A, Keshet E. Induction of vascular endothelial growth

factor expression by hypoxia and by glucose deficiency in multi cell spheroids:
implications for tumor angiogenesis. Proc Natl A cad Sci USA 1995;92(3):768-72.

6. Okazaki IJ, Moss 1. Mono-ADP-ribosylation: A reversible post-translational

modification of proteins. Adv Pharm 1996;35:247-280

7. Gimbrone M, Cotran R, Leapman S. and Folkman J. Tumor growth and

neovascularization: An experimental model using the rabbit cornea. J Natl Cancer
Inst 1974; 52: 413-416.

8. Knighton DR, Hunt TK, Scheuentuhl H, Halliday BJ, Werb Z, Banda MJ. Oxygen
tension regulates the expression of angiogenesis factor by macrophages. Science
1983; 221: 1283-1285.

9. Constant J, Suh DY, Hussain MZ, Hunt TK. Wound healing angiogenesis: The
metabolic basis of repair. In: Maragoudakis ME, editor. Molecular, Cellular, and
Clinical Aspects ofAngiogenesis. New York: Plenum Press, 1996: 151-159.

10. Kleinman HK, McGarvey ML, Hassel JR, Star VL, Cannon FB, Lauri JW, Martin
GR. Basement complexes with biological activity. Biochemistry 1986; 25 :312-318.

11. Passaniti A, Taylor RM, Pili R, Guo Y, Long PV, Haney JA, Pauly RR, Grant DS,
Martin GR. A simple quantitative method for assessing angiogenesis and
antiangiogenic agents using reconstituted basement membrane, heparin, and fibroblast
growth factor. Lab Invest 1992; 67: 519-528.

12. Grant DS, Lelkes PI, Fukuda K, Kleinman HK. Intracellular mechanisms involved in
basement membrane induced blood vessel differentiation in vitro. In Vitro Cell Dev
Bioi 1991; 27a: 327-336.

13. Althaus FR, and Richter C. Mono-ADP-ribosylation Reactions: The bond, in ADP-
Ribosylation of Proteins, Enzymology and Biological Significance, (Althaus FR and
Richter C, eds), Springer-Verlag, New York, 1987; pp 216-220.

14. Zabel DD, Feng JJ, Scheuenstuhl H, Hunt TK, Hussain MZ. Lactate stimulation of
macrophage-derived angiogenic activity is associated with inhibition of poly(ADP-
ribose) synthesis. Lab Invest 1996;74:644-649.

15. Walter DH, Hink U, Asahara T, Van Belle E, Horowitz J, Tsurumi Y, et al. The in
vivo bioactivity of vascular endothelial growth factor/vascular permeability factor is
independent ofN-linked glycosylation. Lab Invest 1996;74(2)546-56.

16. Peretz D, Gitay GH, Safran M, Kimmel N, Gospodarowicz D, Neufeld G.

Glycosylation of vascular endothelial growth factor is not required for its mitogenic
activity. Biochem Biophys Res Commun 1992; 182(3): 1340-7.

17. Trepel JB, Chuang DM, and Neff NH. Polypeptide hormones and chromatin-
associated proteins act as acceptors for cholera toxin-catalyzed ADP-ribosylation. J
Neurochem 1981; 36: 538-543.

135
ROLE OF FIBROBLAST GROWTH FACTOR -2 AND ENDOTHELIAL CEIL
STIMULATING ANGIOGENIC FACTOR (ESAF) IN CAPILLARY GROWTH IN
SKELETAL MUSCLES EXPOSED TO LONG-TERM IDGH ACTIVITY

M.D. Brownl, H. Walter, O. Hudlicka2, F.M. Hansen-Smith3 and lB.

Weiss4

1School of Sport and Exercise Sciences and 2Dept. of Physiology,

University of Birmingham, UK, 3Dept. of Biological Sciences, Oakland
University, Rochester, MI, USA, and 4Wolfson Angiogenesus Unit,
Department of Rheumatology, Hope Hospital, Salford, UK.

1. INTRODUCTION

Angiogenesis, the development of new blood vessels, is a process controlled by

many different mediators, of which different growth factors have been considered to be key
regulators. In particular, fibroblast growth factors (FGFs) are known to be involved in
angiogenesis in many pathological situations e.g. tumours (Folkman and Klagsbrun, 1987),
wound healing (Broadley, Aquino, Woodward, Buckley-Sturrock, Sato, Rifkin and
Davidson, 1989), inflammatory conditions (D'Amore, 1992), growth of collateral vessels in
ischaemic heart (Schaper, Sharma, Quinkler, Markert, Wunsch and Schaper, 1990) and
skeletal muscle (Yang, Deschenes, Ogilvie and Te~ung, 1996). However, FGFs do not
appear to cause proliferation of endothelial cells in uninjured tissue (D'Amore, 1990) and it
has therefore not been established whether they play a role in angiogenesis under normal
physiological conditions when the microvascular bed is undamaged. During development,
FGFs may be involved in vasculogenesis in the heart (Tomanek, Haung, Suvama, O'Brien,
Rataj ska and Sandra, 1996), but in skeletal muscle, angiogenesis during postnatal growth
does not seem to be associated with basic fibroblast growth factor FGF-2 (Hansen-Smith,
Morris and Joswiak, 1992).
In normal adult skeletal muscle, angiogenesis can be induced as a physiological
adaptation to increased contractile activity. Morrow, Kraus, Moore, Williams and Swain
(1990) stimulated fast muscles of the rabbit for 24 hours per day and did not find any
change in the expression ofFGFs after three days of stimulation but observed elevation of
FGF mRNA after three weeks. It is known from our previous work that capillarization in
stimulated muscles is increased at this latter time (Brown, Cotter, Hudlicka and Vrbova,

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York , 1998 137
 1976) but the initiation of angiogenesis would have taken place much earlier. Morrow et al
(1990) considered that satellite cells were more likely to be the source for FGFs. FGF-2
has been localised by immunohistochemistry to the extracellular matrix of skeletal muscle
fibres where it is up-regulated in situations of hypertrophy and regeneration (Yamada,
Buffinger, DiMario and Strohman, 1989). Since continuous muscle stimulation for an
extended period such as that employed by Morrow et al (1990) causes muscle damage
(Pette, Muller, Leisner and Vrbova, 1976; Lexell, Jarvis, Downham and Salmons, 1992),
this could account for the presence of FGF-2 at this time. However, a possible link
between the presence of FGF-2 in skeletal muscle and activity-induced angiogenesis
remains to be established.
An important primary step in angiogenesis is considered to be breakage of the
capiIIary basement membrane which we have observed in chronically stimulated muscles
(Hansen-Smith, Hudlicka and Egginton, 1996). In addition, there must be degradation of the
extracellular matrix in order for a capillary to form a sprout, and matrix metalloproteinases
(MMPs) collagenase, stromelysin-I and gelatinase A which accomplish these actions are
thus very important for the angiogenic process. Another growth factor, endothelial-cell-
stimulating angiogenesis factor (ESAF) is a non-protein, non-enzymic but naturally
occurring low molecular mass (-600Da) factor which has been identified and extracted from
a variety of tissue sources - bovine pineal gland, epiphysial growth plate, and the serum of
patients with active neovascularization e.g. proliferative retinopathy and fracture healing
(Odedra and Weiss, 1991). ESAF is a mitogen specific for microvascular endothelial cells,
alone and synergistically with FGF-2 (Odedra and Weiss, 1991). ESAF is not only capable
of activating the pro-forms of MMPs that degrade extracellular matrix components, it is
unique in that it is the only physiological molecule capable of reactivating the complexes
that MMPs form with their tissue inhibitors (TIMPs) (McLaughlin and Weiss, 1996). Its
ability to activate enzymes capable of dissolution of the capillary basement membrane, a
prerequisite for initiation of the angiogenic process, is clearly a key feature of its function
as an angiogenic factor. We have previously found that levels of ESAF correlated with
capillary density in normal rabbit and pig hearts where capillary growth was induced by
long-term bradycardial pacing (Hudlicka, Brown, Walter, Weiss and Bate, 1995), and it is
possible that it is also involved in activity-induced angiogenesis in skeletal muscle.
The time course of capillary growth in rat fast muscles subjected to chronic
electrical stimulation has been shown to be comparatively rapid. Stimulation for 8
hours/day causes proliferation of capillary endothelial cells within two days (pearce,
Hudlicka and Egginton, 1995) and capillary supply is increased by 30-40% within a week
(Myrhage and Hudlicka, 1978; Brown, Hudlicka, Makki and Weiss, 1995). The aims of the
present study were therefore to investigate whether FGF-2 has a role in this capillary
growth by examining its relationship to the expression of mRNA for FGF-2, and the
amount and distribution of the peptide and its receptors, and furthermore, to study
whether levels of ESAF are linked with the increase in capillary supply (Brown et aI.
1995).

2. METHODS

2.1 Chronic electrical stimulation procedure and muscle sampling

Experiments were performed on male Sprague Dawley rats. Extensor digitorum

longus (EDL) and tibialis anterior (TA) muscles were stimulated in one hind limb by

138
implanted electrodes as described previously (Dawson and Hudlicka,1989). Muscle
contractions were evoked for 8 hours per day for periods of two, four or seven days by
stimulation at 10Hz, pulse width O.3msec, which does not activate either sympathetic or
unmyelinated nerve fibres, and voltage sufficient to produce palpable muscle contractions
(3-5 V). Stimulated and contralateral EDL and T A muscles were removed from animals
killed by cervical dislocation at least 16 h after the end of the last stimulation. A mid-
muscle steak approximately 6mm thick was frozen in isopentane pre-cooled in liquid
nitrogen for histology while the rest of the muscle was rapidly weighed and frozen in liquid
nitrogen for either subsequent RNA extraction or estimation ofESAF levels.

2.2 Estimation of capillary supply

Capillaries were identified by histochemical stammg for endothelial alkaline

phosphatase in l2J..lm thick cryo-sections using an indoxyl tetrazolium method which
depicts all anatomically present capillaries (Ziada, Hudlicka, Tyler and Wright, 1984).
Capillaries and muscle fibres were counted in fields covering a total of OA-O.8mm2 per
muscle section and capillary supply of the EDL and TA was evaluated as capillary-per-
fibre ratio (CIF). For T A, capillary counts were also made separately in the glycolytic
surface cortex and the oxidative deep core.

2.3 Immunolocalization ofFGF-2 and FGF receptors

Peptide FGF-2 was localised in 6 or 8J..lm cryosections of EDL and TA using

several different antibodies. Poly clonal antibodies to a synthetic peptide fragment of FGF-
2 or to human recombinant FGF-2 (a gift of Dr. Andrew Baird, The Whittier Institute, La
Jolla, CA, USA) were used on sections adjacent to those from which capillary/fibre counts
were obtained, and visualised by the avidin-biotin complex technique using
diaminobenzidine (Gonzalez, Berry, Maher, Logan and Baird, 1995). In addition,
monoclonal antibodies against either purified bovine brain FGF-2 (Upstate Biotechnology
Inc.) or against purified18kDa human FGF-2 (Santa Cruz Biotechnology) were used to
localise FGF-2 peptide by an immunofluorescence method using an FITC-labelled
secondary antibody. Similarly, FGF receptors were identified fluorescently by a polyclonal
antibody to synthetically produced chicken flg receptor FGFR-l (Upstate Biotechnology
Inc.) and a monoclonal antibody recognizing the human flg receptor and FGFR-2 bek
receptors (Santa Cruz Biotechnology Inc.). Immuno-fluorescent detection ofFGF-2 and its
receptors was performed concurrently with labelling of microvessels with the rhodamine
labelled Griffonia simplicifolia I (GSI) lectin to visualize capillaries, arterioles and venules
and enable colocalisation of growth factors in relation to vessels, as described previously
(Hansen-Smith, Watson, Lu and Goldstein, 1988).

2.4 RNA extraction and FGF-2 mRNA analysis

Total cellular RNA was extracted from all muscles either by the one-step method
(Chomcyznski and Sacchi, 1987) or a modification of that method using RNAzol B method
(Biogenesis). The integrity, concentration and purity of the RNA samples was assessed as
described by Sambrook, Fritsch and Maniatis (1989). 20ug or total cellular RNA from each
muscle was analysed by Northern hybridisation (Maniatis, Fritsch and Sambrook, 1982).
Since the Northern blot analysis yielded only weak positive results, the Ribonuclease
Protection Assay (RPAII, Ambion) with a 32P-labelled FGF-2 cDNA probe (O.5kb of the

139
 TA

C/F

2.

1.5

A T A oxidative core

1.0 • TA glycolytic cortex

0.5

EDL

C/F
2.0

1.5

1.0

0.5

0
0 2 4 7

days of stimulation

Figure 1. Changes in capillary per fibre ratio in rat T A and EDL muscles after chronic
electrical stimulation at 10Hz, 8 hours/day for 2 (n=4), 4 (n=4) and 7 (n=ll) days. Values
are plotted as means ± s.e.m and those at day 0 are for control unstimulated muscles
(n=12). For TA, data are shown separately for the deep oxidative core and the superficial
glycolytic cortex. *P<O.OS v. control.

coding region of rat FGF -2 cDNA, clone RobFGF 103, Shimasaki, Emoto, Koba, Mercado,
Shibata, Cooksey, Baird and Ling, 1988) was used to detect specific hybridisation.

2.5 ESAF estimation

ESAF was assayed in control muscles and those stimulated for 7 days by its ability
to activate latent collagenase as described by Weiss, Hil1, Davis, McLaughlin, Sedowofia
and Brown (1983). Results are expressed as units of ESAF where 1 unit is the percent
activation of the enzyme per hour per miligram of protein in the supernatant.

140
3. RESULTS

Capillary/fibre ratio in control muscles was 1.44±O.06 in EDL and 1.62 ±O.05 in
T A (all values are given as means ± SEM). There was no significant increase in CIF in
muscles that had been stimulated for 2 days. Muscles stimulated for 4 days showed a
slightly higher CIF (1.65±O.04 EDL, 1.89±O.08 TA) and the increase was significantly
greater in those stimulated for 7 days (2.09±O.08 and 2.37±O.06 respectively, p<O.05 vs
controls, ANOV A). In TA, the increased CIF occurred preferentially in the glycolytic
cortex of the muscle while in the core, CIF did not substantially change (Figure 1). Muscles
contralateral to those stimulated did not show any significant difference in CIF ratio from
control values (data not shown).

Figure 2. Photomicrographs of control rat tibialis anterior muscle A) and B) stained by

antibody to human recombinant FGF-2 and C) and D) adjacent section stained for alkaline
phosphatase to depict capillaries which appear as black dots. FGF-2 immunoreactivity is
seen associated with an arteriole in A) and C) at arrows. B) shows that there is no antibody
stain which corresponds to the site of capillaries, the only immunoreactivity present (at
arrow) being that associated with nerves alongside the larger blood vessels, probably
venular, apparent in D) at arrow. M = identification of the same muscle fibre in adjacent
sections. Scale bar represents 50llm.
141
 RPA demonstrated expression of mRNA for FGF-2 which was variable among
muscles, with no obvious relationship to capillary supply. All control muscles, EDL as
well as TA, showed some expression of FGF-2. It was absent in 4 out of 7 muscles
stimulated for 2 days, and in 3 out of 4 and 3 out of3 muscles stimulated for 4 and 7 days
respectively. In contrast, all muscles contralateral to stimulated showed FGF-2 expression,
which was slightly enhanced in some after 2 and 4 but not 7 days.
There was no indication with any of the antibodies used that FGF-2 peptide could
be localised to capillaries in either control, contralateral, or stimulated muscles (Figure 2).
However, the peptide was detected in nerves and in the walls of some arterioles and larger
blood vessels. Both FGF receptors, flg and bek, were distributed throughout the muscles
particularly in the smooth muscle layer of arterioles, in myonuclei, adjacent interstitial cell
types and nerves. Very occasionally, immunostaining for FGF receptors was colocalised to
capiIlaries identified by lectin fluorescence but there was no obvious quantitative difference
in their distribution among control, contralateral or stimulated muscles.

TA
C/F • ESAF

D
•
control

~ sham operated
2
stimulated

EDl
C/ F
• ESAF
5

• 2

o
Figure 3. Capillary per fibre ratio is shown on the left and levels of ESAF units on the
right for rat TA and EDL muscles (n=ll) after stimulation for 7 days at 10Hz, 8 hours/day
in comparison with values for control unoperated (n=20) and sham-operated (n=5)
muscles. All values are shown as means ± s.e.m. *P<O.OS v. control.

142
 The levels of ESAF were assayed in whole TA and EDL muscles. Figure 3 shows
that ESAF levels were significantly increased after 7 days stimulation in both muscles in
association with the increase in capillary supply. For comparison, data is shown for sham-
operated muscles, i.e. electrodes implanted but no stimulation, in which neither CIF nor
ESAF levels were different from controls.

4. DISCUSSION

The present findings show an absence of increased numbers of capillaries in skeletal

muscles subjected to increased contractile activity by stimulation for 2 or 4 days, but a
significant increase after 7 days. However, the labelling index for nuclei associated with
capillaries, based on incorporation of the thymidine analogue bromodeoxyuridine, was
increased after only 2 days of stimulation (pearce et al. 1995) and ultrastructural signs of
capillary endothelial cell activation were observed in muscles stimulated for 4 days
(Hansen-Smith et al. 1996). If FGF-2 is implicated in capillary growth in this situation, it
could be expected that at least some of the criteria for its involvement such as upregulation
of message, its expression and the presence of its receptors would be met by 4 days. The
expression of mRA for FGF-2 in small amounts in normal skeletal muscles confirms
previous findings, and indicates that the radioprotection assay method was sensitive
enough to detect possible changes as a result of stimulation. However, the differences in
FGF-2 mRNA observed were not related to capillary growth. If anything, mRNA
expression tended to be decreased in 2 and 4 day stimulated muscles and enhanced in
contralateral muscles, a finding which would not support its participation in the early
stages of capillary proliferation. Variations in mRNA levels between muscles may relate to
the differing proportions of the structures in muscles to which FGF-2 has been localised,
such as nerves, blood vessels, basement membranes and interstitial cells (Yamada et al.
1989). It is also possible that the presence of FGF -2 mRNA in contralateral muscles may
be linked with alterations of muscle fibres in relation to increased weight-bearing activity,
as animals favour their unoperated limb initially and increased expression of FGF-2 has
been reported in skeletal muscles undergoing hypertrophy (Yamada et al. 1989). In
addition, mRNA signal for FGF-2 was found to be increased in muscles exposed to
intermittent tetanic contractions (Hudlicka and Egginton, unpublished results) and
increased expression of FGFs observed in those stimulated continuously at 10Hz for 3
weeks, associated with satellite cell proliferation (Morrow et al. 1990), suggesting that its
role may be related more to skeletal muscle rather than capillary growth.
Immunological localisation ofFGF-2 peptide in relation to capillaries in stimulated
muscles would be another indicator of its involvement in angiogenesis but none of the
antibodies used against four different forms of FGF-2 - synthetic, recombinant human,
purified bovine and human- detected any significant amounts of peptide in association with
capillaries in any of the muscles examined, despite that fact that they showed
immunoreactivity to FGF-2 in rat brain endothelial cells (synthetic, Gonzalez et al. 1995),
to cultured bovine aortic endothelial cells (recombinant human, Speir, Sasse, Shrivastav and
Casscells, 1991) and in postnatal skeletal muscle (purified bovine and human, Hansen-
Smith et al. 1992). Likewise, the fact that FGF-2 receptors were only occasionally co-

143
 % change from
resting diameter
50 ____ 3rd order arterioles

** ___ 4th order arterioles

40
**
**

0.0001 0 .001 0.01 0. 1 10 100

FGF-2 concentration ng.ml·'

Figure 4. Concentration-response curves for the dilator effect, expressed as % change from
resting diameter, of FGF-2 when applied topically to 3rd order (sqaure symbols) and 4th
order (circles) arterioles of the hamster cheek pouch. Each point represents the mean ±
s.e.m response from 8-10 vessels. *p<O.OS, **P<O.OI v. resting diameter.

localised immunologically to capillaries does not support a strong involvement of this

factor in stimulation-induced angiogenesis. However, the fact that FGF-2 and its receptors
were found in arterioles suggested that it may be vasoactive. We therefore tested its effect
on arterioles in hamster cheek pouch, a tissue where FGF-2 has so far not been found.
Topical application of FGF-2 either by micropipettes onto the adventitial surface of
arterioles resulted in dilatation of 3rd order (resting diameter 22.S±O.S mm) and fourth
order (resting diameter 14A±OA mm) arterioles (Figure 4), resulting in almost 80% of
maximal dilatation (Brown, Hudlicka, Damon and Duling, 1996).
It is well known that some vasodilators can induce angiogenesis in skeletal muscle
(Hudliclci, Brown and Egginton, 1992) and thus the presence of FGF-2 in relation to
arterioles may indicate that it contributes to vasodilatation of arterioles in contracting
muscles and possibly to the increase in arteriolar diameters seen during the first few days
of stimulation (pearce and Hudlicka, 1994). Such a role for FGF-2 and vasodilatation in
angiogenesis has been proposed by Folkman and Shing (1992).
Although we could not obtain support for the involvement of FGF-2 in capillary
growth in stimulated skeletal muscles, we did observe a significant increase in levels of
ESAF in muscles stimulated for 7 days, when capillary growth is evident. This is in
agreement with previous findings that the presence ofESAF is linked with capillary growth
in the heart (Hudlicka et aI. 1995). While we do not have data from previous time points
during the course of capillary growth during stimulation, one animal showed increased
ESAF after stimulation for only 3 days (unpublished observation) indicating that this
factor may be a corequisite for the angiogenic process, whereby its dual role as an
endothelial cell mitogen and in extracellular matrix degradation (Odedra and Weiss, 1991)
would be important.
In conclusion, capillary growth in normal skeletal muscles subjected to increased
activity by chronic stimulation does not appear to involve FGF -2, whereas it is linked with

144
ESAF. However, the presence of FGF-2 and its receptors in arteriolar blood vessels in all
muscles may suggest a role in vasodilatation.

REFERENCES

Broadley, K.N., Aquino, A.M., Woodward, S.c., Buckley-Sturrock, A., Sato, Y., Rifkin,
D.B. and Davidson, lM., 1989, Monospecific antibodies implicate basic fibroblast growth
factor in normal wound repair, Lab. Invest. 61: 571-575.

Brown, M.D., Cotter, M.A., Hudlicka, O. and Vrbova, G., 1976, The effect of different
patterns of muscle activity on capillary density, mechanical properties and structure of
slow and fast rabbit muscles, Pfliigers Arch. 361: 315-323.

Brown, M.D., Hudlicka, 0., Makki, R.F. and Weiss, lB., 1995, Low-molecular-mass-
endothelial cell-stimulating angiogenic factor in relation to capillary growth induced in rat
skeletal muscle by low-frequency electrical stimulation, Int. l Microcirc. 15: 111-116.

Brown, M.D., Hudlicka, 0., Damon, D. and Duling, B.R., 1996, Vasoactive effects of basic
and acidic fibroblast growth factors in hamster cheek pouch arterioles, Int. l Microcirc. 16:
308-312.

Chomcyznski, P. and Sacchi, N., 1987, Single-step method of RNA isolation by

guanidinium thiocyanate -phenol-chloroform extraction, Anal. Biochem. 162: 156-159.

D'Amore, P.A., 1990, Modes ofFGF release in vivo and in vitro, Cancer and Metast. Rev.
9: 227-238.

D'Amore, P.A., 1992, Mechanism of endothelial growth control, Am. l Resp. Cell. Mol.
BioI. 6: 1-8

Dawson, lM. and Hudlicka, 0., 1989, The effect of long-term activity on the
microvasculature of rat glycolytic skeletal muscle, Int. l Mi croci rc. 8: 53-69.

Folkman, land Klagsbrun, M., 1989, Angiogenic factors, Science 235: 442-446.

Folkman, land Shing, Y., 1992, Angiogenesis, l BioI. Chern. 267: 10931-10934.

Gonzales, A.M., Berry, M., Maher, P.A., Logan, A. and Baird, A., 1995, A comprehensive
analysis of the distribution of FGF -2 and FGFR 1 in the rat brain, Brain Res. 701: 201-226.

Hansen-Smith, F.M., Watson, L. , Lu, D.Y. and Goldstein, I. , 1988, Griffonia simplicifolia
I: fluorescent tracer for microcirculatory vessels in nonperfused thin muscles and sectioned
muscle, Microvasc. Res. 36: 199-215.

Hansen-Smith, F.M., Morris, L. and Joswiak, G.R., 1992 Postnatal proliferation of

microvessels and the distribution of basic fibroblast growth factor (bFGF) in rat
sternomastoid muscle, FASEB J. 6: A1600

Hansen-Smith, F.M., Hudlicka, O. and Egginton, S., 1996, In vivo angiogenesis in adult rat
skeletal muscle: early changes in capillary network architecture and ultrastructure, Cell
Tissue Res. 286: 123-136.

145
Hudlicka, 0., Brown, M.D. and Egginton, S., 1992, Angiogenesis in skeletal and cardiac
muscle, Physio!. Rev. 72: 369-417.

Hudlicka, 0., Brown, M.D., Walter, H. , Weiss, 1.B. and Bate, A. , 1995, Factors involved
in capillary growth in the heart, Mol.and Cell. Biochem. 147: 57-68.

Lexell, T., Jarvis, 1., Downham, D. and Salmons, S., 1992, Quantitative morphology of
stimulation-induced damage in rabbit fast twitch skeletal muscles. Cell Tiss. Res. 269: 195-
204.

Maniatis, T., Fritsch, E.F. and Sambrook, 1., 1982, Molecular cloning: a laboratory manual,
Cold Spring Harbour Laboratory, New York.

McLaughlin, B . and Weiss, 1.B., 1996, Endothelial-cell-stimulating angiogenesis factor

(ESAF) activates progelatinase A (72kDa type IV collagenase), prostromelysin 1 and
procollagenase and reactivates their complexes with tissue inhibitors of metalloproteinases:
a role ofESAF in non-inflammatory angiogenesis, Biochem. 1. 317: 739-745.

Morrow, N .G., Kraus, W .E., Moore, 1.W., Williams, R.S . and Swain, 1.L., 1990, Increased
expression of fibroblast growth factor in a rabbit skeletal muscle model of exercise
conditioning, 1. Clin. Invest. 85: 1816-1820.

Myrhage, R. and Hudlicka, 0., 1978, Capillary growth in chronically stimulated adult
skeletal muscles as studied by intravital microscopy and histological methods in rabbits
and rats, Microvasc. Res . 16: 73 -90.

Odedra, R . and Weiss, J.B ., 1991, Low molecular weight angiogenesis factors, Pharmacol.
Ther. 49 : 111-124.

Pearce, S. and Hudlicka, 0., 1994, Are prostaglandins involved in capilary growth In
chronically stimulated muscles? Int. 1. Microcirc. 14: 243.

Pearce, S., Hudlicka, O. and Egginton, S., 1995, Early stages In activity induced
angiogenesis in rat skeletal muscle: incorporation of bromodeoxyuridine into cells of the
interstitium, 1. Physiol. 483 : 143 P

Pette, D., Muller, W., Leisner, E. and Vrbova, G., 1976, Time dependent effects on
contractile properties, fibre population, myosin light chains and enzymes of energy
metabolism in intermittently and continuously stimulated fast twitch muscles of the rabbit,
PflOgers Archiv. 364: 103-112.

Sambrook, 1., Fritsch, E.F. and Maniatis, T., 1989, Molecular cloning: a laboratory manual,
2nd ed., Cold Spring Harbor, New York.

Schaper, W., Sharma, H.S., Quinkler, W., Markert, T., Wunsch, M . and Schaper, 1., 1990,
Molecular biologic concepts on coronary anastamoses, 1. Am. Coli . Cardiol. 15: 513-518.

Shimasaki, S., Emoto, N., Koba, A., Mercado, M.. Shibata, F .. Cooksey, K.. Baird, A. and
Ling N., 1988, Complementary DNA cloning and sequencing of rat ovarian basic fibroblast
growth factor and tissue distribution study of its mRNA, 1988, Biochem. Biophys. Res.
Commun. 157: 256-263.

146
Speir, E., Zilou, Y.F., Lee, M., Shrivastav, S. and Casscells, W., 1989, Fibroblast growth
factors are present in adult cardiac myocytes in vivo, Biochem. Biophys. Res. Commun.
159: 1336-1340.

Tomanek, R.J., Haung, L., Suvama, P.R., O'Brien, L.c., Ratajska, A and Sandra, A, 1996,
Coronary vascularization during development in the rat and its relationship to basic
fibroblast growth factor, Cardiovasc. Res. 31: E116-E126

Weiss, JB., Hill, C.R., Davis, R.J., McLaughlin, B., Sedowofia, K.A. and Brown, R.A.,
1983, Activation of a procollagenase by low molecular weight angiogenesis factor. Biosci.
Rep. 3: 171-177.

Yamada, S., Buffinger, N., DiMario, J. and Strohman, R.C., 1989, Fibroblast growth factor
is stored in fiber extracellular matrix and plays a role in regulating muscle hypertrophy,
Med. Sc. Sports Exerc. 21: SI73-S180.

Yang, H.T., Deschenes, M.R., Ogilvie, R.W. and TeIjung, R.L., 1996, Basic fibroblast
growth factor increases collateral blood flow in rats with femoral arterial ligation, Circ. Res.
79: 62-69.

Ziada, AM.A.R., Hudlicka, 0., Tyler, K.R. and Wright, A.J.A., 1984, The effect of long-
term vasodilatation on capillary growth and performance in rabbit heart and skeletal
muscle, Cardiovasc. Res. 18: 724-732.

147
ANGIOGENESIS IN ATHEROSCLEROSIS: POSSmLE ROLES FOR VASCULAR
ENDOTHELIAL CELL GROWTH FACTOR, ENDOTHELIAL CELL
STIMULATING ANGIOGENESIS FACTOR AND SOLUBLE E-SELECTIN.

JB Weiss', A Blann", lun-Ling Li"', CN McCollom" and A Bate'.

'Wolfson Angiogenesis Unit, Rheumatic Diseases Centre, University of

Manchester, Hope Hospital, Manchester M6 8HD,UK.
"Department of Surgery, South Manchester University Hospital,
Manchester M20 8LR. UK.
"'Thrombosis, Haemostsis and Vascular Biology Unit, University
Department of Medicine, The City Hospital,Birmingham, B18 7QH,UK.

1. INTRODUCTION

Endothelial cell injury is believed to be an important early step in the pathogenesis of

atherosclerosis, and there is evidence of endothelial cell dysfunction in the risk factors for
this disease[1-4]. It follows that if there is generalsied endothelial cell destruction in
atherosclerosis then there must also be a proportional increase in repair if the intima is to
avoid being de-endothelialised. A further vascular consequence of atherosclerosis is
angiogenesis of the vasa vasorum within the arterial media and adventicia, in response to
the ischaemia following arteriole occlusion. Such a consequence may be dangerous as
angiogenesis in neovascularising blood vessels affected by atherosclerosis can cause
atheroma to rupture or thrombose, or cause vascular spasm.
Both of these clinical events may lead to cerebral, peripheral or myocardial infarction [5-8].
Numerous microvessels are present in both the wall surrounding the atheroma and within
the plaque itself, and the degree of angiogenesis correlates with the extent of the atheroma
[9,10].
The current study was designed to determine the levels of circulating factors such as ESAF,
VEGF and soluble E-Selectin and vWF, which might reflect endothelial cell growth,
regeneration and damage, in patients with peripheral vascular disease and patients with
ischaemic heart disease. Any meaningful relationship between the circulating levels of these
factors and any relation to the risk factors for atherosclerosis might contribute to the
understanding of angiogenesis.

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 149
2. SUBJECTS

Twenty-four subjects (seven women) with peripheral vascular disease were recruited from
a dedicated Vascular Disease clinic at the South Manchester University Hospital, UK.
Doppler or angiography proven atherosclerosis (stenosis > 70% or occlusion) was
confirmed in the patients by Doppler/ultrasound scanning of the appropriate symptomatic
areas of the carotids and the arteries of the lower abdomen and leg ( iliac, femoral and
popliteal). Peripheral vascular disease subjects were asymptomatic for ischaemic heart
disease. Twenty patients (five women) with ishaemic heart disease were recruited whilst in
hospital following myocardial infarction as proven by raised levels of creatinine kinase
«150 u/ml), typical history of retrosternal chest pain and changes on electrocardiogram.
They were asymptomatic for peripheral vascular disease. Patients were given an
appointment to return for blood sampling at least six weeks after their discharge from
hospital. This was in order to minimise the effects of the acute phase response on von
Willebrand factor, and post - infarction changes in the lipid profile. These patients were
asymptomatic for peripheral vascular disease.
The twenty-seven controls were asymptomatic attenders at hospital clinics for endoscopy,
varicose veins, hernia or for minor operations, or were apparently healthy hospital staff.
They were age and sex matched to the patients with atherosclerosis. Exclusion criteria for
all subjects were acute or chronic renal or liver disease, connective tissue disease, increased
erythrocyte sedimentation rate, (Westergren, > 20 mrnlhr), excessive alcohol consumption,
diabetes, neoplasia or treatment with vasopressin, or cytotoxic or steroid drugs. There were
three smokers in the control group, four in the ischaemic heart disease group, and eight in
the peripheral vascular disease group.

3. METHODS

3.1. ESAF
Originally described by its ability to stimulate the proliferation of endothelial cells in vitro
and its angiogenic activity in vivo on the chick chorioallantoic membrane, has also been
shown to be capable of activating the neutral promatrix metalloproteinases [11,12,13]. This
latter ability enables ESAF to be assayed rapidly by its ability to activate procollagenase in
a validated functional assay. The technique is fully described elsewhere [11,13].

3.2. VEGF
VEGF was measured by an ELISA using commercial antisera and standardsised with
recombinant human VEGF(R&D Systems UK).[I4]

3.3. vWF
von Willebrand Factor was measured by an established and fully validated 96 well
microtitre plate ELISA using commercial polyclonal rabbit antisera at a dilution of 1/500 as
fully described elsewhere [15,16].

3.4. Soluble E-Selectin

Soluble E-Selectin was measured using a commercial ELISA kit (R&D Systems, Abingdon,
UK)

150
 60

ESAF
(Ulml)

50

40

30

20 o
000

00
o
00
o
o
°ao o
o
10 000

00
00

IHC Control PVC

Figure 1. ESAF Levels in Ischaemic Heart Disease(IHD), Peripheral Vascular

Disease(pVD) and Normal Controls

Table 1.
Levels of von Willebrand factor, soluble E-selectin, ESAF, and VEGF in Patients and
Controls.

Healthy Controls IHD Patients PVD Patients

vWF(IU/dLj 102±17 133±30 * 140±35 *
E-Selectin (ng/ml) 5l±17 57±20 ** 53±22
VEGF (ng/ml) 1.5 8.0 * 5.0 *
ESAF(U/ml) 9.4 19.6 * 17.1 *

Table I: Data are mean and standard deviation or median and range. IHD = Ischaemic heart
disease. PVD = Peripheral vascular disease. * = raised (p<O.OI) relative to healthy controls
but not between groups. ** = raised (p<0.05) relative to levels in both healthy controls and
patients with peripheral vascular disease.

151
 700
• •
VEGF •
(nglml)

•
100 •

• • •
•• •••
.~

I. I
I
10
•
---¥---
Y
•
•
• • -...
t-
••
••
• •
•••• •
•••
•

<1 •• 11111 .....

A. A.
IHD Control PVD

Figure 2. VEGF Levels in Ischaemic Heart Disease(IHD), Peripheral Vascular

Disease(pVD) and Normal Controls

4. STATISTICS

Numerical data was analysed by analysis of variance with Tukeys post hoc test (ESAF
and VEGF data were non-parametrically distributed: therefore a logarithmic transformation
was applied).
Categorical data (sex and smoking) was analysed by the chi-squared test. Data is presented
as mean and standard deviation, median and range, or as percentage incidence. Results were
correlated by Spearmans rank method and a multivariate stepwise regression analysis with
von Willebrand factor, VEGF, soluble E-Selectin and ESAF as independent variables versus
the risk factors for atherosclerosis as dependent variables was performed on a Minitab 8
system .

5. RESULTS

5.1. Endothelial Cell Related Indices

Table I shows levels of ESAF (fig I), VEGF (fig 2) and soluble E-Selectin . Levels of
ESAf, vWF and VEGF were all significantly raised (p<O.Ol) in both groups of patients

152
with no differences between the groups of patients. Soluble E-Selectin was weakly raised in
patients with ischaemic heart disease (p<0.05)
Figures 1 and 2 show comparisons of data for circulating levels of VEGF, ESAF and vWF.
The distribution pattern of VEGF and ESAF are notably similar.

6. CORRELATIONS

There was a positive Spearmans correlation between vWF and ESAF (r=0.25, p=0.025).
Despite the similarities in the patterns of ESAF and VEGF they failed to correlate
significantly.

7. DISCUSSION

Increased levels ofESAF and VEGF were observed in patients with ischaemic heart disease
and in patients with peripheral vascular disease. Raised soluble E-Selectin was found only
in patients with ischaemic heart disease, raised vWF was found in both groups of patients,
and both these findings support previous reports.
Although there is histological, physiological and metabolic evidence of endothelial cell
damage in atherosclerosis and its major risk factors, there is also evidence of active ongoing
endothelial cell proliferation and renewal, probably in response to the disease process.
Evidence of angiogenesis in atherosclerosis has focused on the histological demonstration of
neovascularisation, and the animal and in vitro experiments [5-10,17]. Knowledge of the
role of growth factors in this process is improving but is frequently focused on histological,
animal and tissue culture work, to the detriment of an understanding of the endothelium in
humans [11,17-19,20-22].
Raised circulating ESAF levels in patients with two of the major symptoms of
atherosclerosis, peripheral vascular disease and ischaemic heart disease were observed. High
levels of ESAF correlated with levels of vWF, an established marker of endothelial cell
damage[23] .
Lack of correlation between circulating VEGF and vWF suggests that VEGF is unrelated to
vascular damage, and conflicts with the report that VEGF will induce vWF release from
endothelial cells in vitro [24].
Levels of soluble E-Selectin appear to be independent of endothelial cell damage, and
increased levels of ischaemic heart disease may be due to another aspect of the
pathophysiology of this manifestation of atherosclerosis.
Increased levels ofESAF and VEGF in patients with long standing ischaemic heart disease
and peripheral vascular disease, indicate ongoing angiogenesis. This may occur separately in
the endothelium of large arteries, or in the microvessels of the vasa vasorum, or in the
capillary beds and perhaps in more than one site. It has been suggested that hypoxia of
myocardial iscaemia promotes angiogenesis, providing the stimulus for the production of
these growth factors, particularly VEGF [19,21] Furthermore hypoxia of the myocardium
induced in vivo and in vitro will upregulate VEGF mRNA, and immunoblotting has shown
that smooth muscle cells can also secrete VEGF, leading to the hypothesis that they have a
role in the neovascularisation of the atherosclerotic plaque [25,26].

These preliminary studies warrant further investigation of the role of VEGF and ESAF in
atherosclerosis. They may represent different aspects of the angiogenic process in

153
microvessels, or of endothelial repair in large vessels. In particular, it would be instructive
to learn via serial studies, if increased angiogenesis preceded or followed endothelial cell
damage.

8. REFERENCES

1. Ross R., 1993, The pathogenesis of atherosclerosis - A perspective for the 1990's.
Nature 362: 801
2. Davies MJ.,Woolf N., 1993, Atherosclerosis: what is it and why does it occur.
Br.Heart J. 69:(Suppl) :s3
3. Badimon L.,Badimon J.J.,Chesebro lH.,Fuster V., 1993, von Willebrand factor and
cardiovascular disease. Thromb.Haemosts. 70: 111
4. Luscher T.F.,Yang Z.,Diederich D.,Buhler F.R., 1989, Endothelium derived
vasoactive substances: Potential role in hypertension, atherosclerosis and vascular
occlusion. J Cardiovasc.PhharmacoI14: (Suppl 6) s63
5. Fryer lA.,Myers P.C.,Appleberg M., 1987, Carotid intraplaque haemorrhage: The
significance ofneovascularity. J.Vasc. Surg., 6: 341
6. Falk E., 1983, Plaque rupture with severe preexisting stenosis precipitating coronary
thrombosis. Br.Heart J. 50:127
7. Lusby RJ.,Farrell L.D.,Ehrenfeld W.K.,Stoney RJ.,Wylie EJ., 1982, Carotid plaque
haemorrhage. Arch.Surg. 117:1479.
8. Barger A.c., Beeuwkes R, (ill) Lainey L.L., Silverman KJ., 1984,
Hypothesis:Vasovasorum and neovascularisation of human coronary arteries. New
Eng.J.Med, 310:175
9. Zamir M.,Silver M.D., 1985, Vasculature in the walls of human coronary arteries.
Arch.Pathol.Lab.Med 109:659
10. Kamat B.R., Galli SJ., Barger A.C., Lainey L.L., Silverman KJ., 1987,
Neovascularistion and coronary atheromatous plaque: cinematographic localisation
and quantative histological analysis. Hum.Pathol. 18: 1036
11. Keegan A., Hill C., Kumar S., Phillips P., Schor A., Weiss lB., 1982, Purified
tumour angiogenesis factor enhance proliferation of capillary but not large vessel
endothelial cells in vitro. J.Ce/l Sci. 55:261
12. Taylor C.M.,Weiss lB.,1989, Raised endothelial cell stimulating angiogenesis factor
in diabetic retinopathy. Lancet ii,1329
13. Weiss J.B., Hill C.R., Davis R.I., McLaughlin B., Sedowofia K.A., Brown R.A.,
1983, Activation of procollagenase by a low molecular weight angiogenesis factor.
Biosci.Reps. 3: 171
14. Leung DW, Cachianes G., Kuang WJ., Goeddel D.V.,Ferrara N.,1989, Vascular
endothelial growth factor is a secreted angiogenic mitogen. Science 246:1306
15. Blann A.D.,McCollum C.N., 1994, Circulating endothelial celllleukocyte adhesion
molecules in atherosclerosis. Thromb.Haemost. 72: 151
16. Blann A.D., Tse W., Maxwell SRJ., Waite M.A., 1994, Increased levels of the
soluble adhesion molecule E-selectin in essential hypertension. J.Hyperten. 12:925
17. Kahlon R., Shapero J., Gotlieb A.I., 1992, Angiogenesis in atheroscerosis.
Can.J.Cardiol. 8:60
18. Weiss lB.,Brown RA.,Kumar S.,Phillips P., 1979, An angiogenic factor isolated from
tumours: a potent low molecular weight compound Br.J.Cancer 40:493

154
19. Dvorak H.F.,Brown L.F.,Detmat M.,Dvorak AM.,1995, Vascular permeability
factor/vascular endothelial growth factor, microvascular hyperpermeability and
angiogenesis. Am.JPathoI146:1029
20 . Folkman J.,Shing Y, 1992, Angiogenesis JBiol. Chem. 267: 10931
21. Battegay E.1.,1995, Angiogenesis:mechanistic insights, neovascular diseases, and
therapeutic prospects. JMol.Med. 73:333
22. Eisenstein R., 1991, Angiogenesis in arteries, Review. Pharma. Ther. 49: 1
23. Blann AD.,1993 von Willebrand factor and the endothelium in vascular disease.
Br.JBiomedSci. 50: 125
24. Brock T.A.,Dvorak HF.,Senger D.R., 1991, Tumour secreted vascular permeability
factor increases Ca2+ and von Willebran factor release in human endothelial cells.
Am.JPathol138:213
25 . Banai S.,Shweiki D.,Pinson A.,Chandra M. ,Lazarovici G.,Keshet E., 1994,
Upregulation of vascular endothelial growth factor expression induced by myocardial
ischaemia: implications for coronary angiogenesis. Cardiovasc.Res. 28: 1176
26. Kuzuya M.,Satake S.,Esaki T.,Yamada K.,Hayashi T.,Naito M.,Asai K.,Iguchi A,
1995, Induction of angiogenesis by smooth muscle cell derived factor: Possible role in
neovascularisation in atherosclerotic plaque. JCell.Physiol. 164:658.

155
THYMOSIN BETA 4 PROMOTES ENDOTHELIAL CELL MIGRATION AND
ANGIOGENESIS

Katherine M. Malinda, Allan L. Goldstein, Derrick S. Grant and Hynda

K. Kleinman

National Institute of Dental Research, NIH, Bethesda, MD,

Department of Biochemistry, George Washington University Medical
Center, Washington, DC,
Cardeza Foundation for Hemaotological Research, Thomas Jefferson
University College of Medicine, Philadelphia, PA

INTRODUCTION

The basement membrane associated with endothelial cells is important in regulating the
passage of macromolecules and cells in and out of the circulation, forming a barrier to the
underlying stroma and maintaining the differentiated phenotype of the endothelial cells
(Grant et aI., 1990). During the formation of new blood vessels, the basement membrane is
first degraded and then the endothelial cells migrate away from the vessel, proliferate at the
site of migration initiation and subsequently a new basement membrane is synthesized
when the vessel is formed. Normally the vasculature is fairly stable with minimal turnover.
During tissue formation, in wound repair and in certain diseases, considerable increases in
vessel formation occur (Folkman, 1992). Using in vitro and in vivo assays, a number of
factors have been described that regulate this process. Many of these factors are either
stored in the basement membrane or produced by cells in response to the basement
membrane.
When endothelial cells are plated in vitro on a layer of basement membrane (commercially
available as Matrigel), the cells attach, migrate and form capillary-like structures with a
lumen (Kubota et aI., 1988; Grant et aI., 1989). The cells are usually attached by two
hours and have begun to migrate towards each other to form tubes by 4 hours. Nearly all
of the cells form the tubes and the process is generally complete within 18 hours (Figure 1).
This differentiation is faster than other in vitro assays employing collagen substrates or
growth factor deprivation. This morphological differentiation on basement membrane
mimics many of the steps in angiogenesis and has been used as an assay to screen for
factors which may be angiogenic or antiangiogenic. It is accepted in the field that

Angiogenesis: Models. Modulators. and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 157
Figure 1 Appearance of human umbilical vein endothelial cells on plastic and
onMatrigel. Cells were plated and photographed on plastic at 20 hours and on
Matrigel as 2, 4 and 20 hours as indicated.

angiogenesis is generally confirmed by more than one assay so additional assays are
generally performed. Many of the known angiogenic factors, such as bFGF, HGF, TGF
beta and phorbol esters, promote tube formation in this assay (Table 1). In addition,
known antiangiogenic factors, such as the protease inhibitors TIMP-1 and TIMP-2 and
PAl-I, block tube formation (Schnaper et aI., 1995). Inhibitors of collagen synthesis block
tube formation while stimulators and recombinant 72 KDa gelatinase promote tube
formation as expected due to the role of collagen breakdown and synthesis in angiogenesis
(Schnaper et al., 1992). Likewise a number of unexpected factors regulating angiogensis
have been identified using the tube assay as an initial screen. For example, haptoglobin
was identified as the factor in sera from vasculitis patients that promoted tube formation
(Cid et aI., 1993). Haptoglobin activity was confirmed in a number of angiogenesis assays
and found to be increased in certain cancers and in endometriosis. Estrogen was also found

158
 Table I FACTORS ACTIVE IN IN VITRO TUBE ASSAY

FACTOR ACTIVITY
laminin peptide SIKVAV stimulates
laminin peptide YIGSR or RGD inhibits
thymosin beta 4 inhibits
bFGF stimulates
TGF beta stimulates
alpha interferon stimulates
gamma interferon inhibits
HGF/SF stimulates
haptoglobin stimulates
phorbol esters stimulates
estrogen stimulates
inhibitors of collagen synthesis inhibits
stimulators of collagen synthesis stimulates
recombinant 72 Kda gelatinase stimulates
TIMP-l or TIMP-2 inhibits
PAI-l inhibits
amino terminal fragment of uPA stimulates
thrombin stimulates

to promote tube formation in vitro on basement membrane and angiogenesis in vivo

(Morales et aI., 1995). Furthermore, overectomized mice were found to have a reduced
angiogenic response to bFGF relative to overectomized animals which received slow release
implants of estrogen. These data demonstrate that the in vitro tube forming assay on
basement membrane Matrigel mimics some of the steps in angiogenesis and can be used
reliably to predict the activity of test compounds.

THYMOSIN BETA 4 IS ELEVATED IN TUBE FORMING ENDOTHELIAL CELLS

We used subtractive cDNA cloning to identify factors produced by endothelial cells early
in tube formation on basement membrane. After four hours on Matrigel, several novel and
known genes were found to be increased (Grant et al., 1995). One gene coded for a small
polypeptide of 4.9 KDa, termed thymosin beta 4, that was first isolated as a product of
calf thymus (Badamchian et aI., 1988). It contain 43 amino acids and is important in
regulating and inducing T-cell differentiation. It also binds to actin and can sequester actin
(Cassimeris et aI., 1992). It is produced by many cells types and could have other
functions (Low and Goldstein, 1984; Condon and hall, 1992; Lin and Morrison-Bogorad,
1990). A variant form, thymosin beta 15, has been found to be important in prostate
tumor cell migration (Bao et aI., 1996). Four hours after plating on Matrigel, thymosin beta
4 mRNA is increased 4 to 6 fold (Grant et aI., 1995). At this time, the cells have attached
to the Matrigel and begin to migrate toward each other to form the tubes. Thymosin beta 4
appears to be important in tube formation since cells transfected with this gene form tubes
more quickly than mock transfected controls. In addition, an antisense oligo to thymosin
beta 4 blocks tube formation on Matrigel. Thymosin beta 4 is located in vessels and has
been found in both growing and mature vessels. Thymosin beta 4 has many effects on

159
Table 2 EVIDENCE THAT THYMOSIN BETA 4 IS IMPORTANT IN
ENDOTHELIAL CELL BEHAVIOR

Transfected cells form tubes more quickly on Matrigel

Transfected cells attach more quickly to substrates
Transfected cells are more spread
Antisense treated cells show decreased tube formation on Matrigel
Exogenous thymosin beta 4 increases cell migration in the Boyden chamber and
scratch assay
Exogenous thymosin beta 4 increases invasion and angiogenesis in subcutaneous
Matrigel plugs

endothelial cells related to angiogenic activity besides the ability to regulate tube formation
(Table 2). Thymosin beta 4 transfected cells adhere more quickly than the mock
transfected cells to various substrates (Figure 2). The transfected cells also spread more
quickly having an increased cell spreading within 30 minutes after plating (Figure 3). These
data demonstrate that endogenously produced thymosin beta 4 has powerful biological
effects on endothelial cells and likely is important in tube formation and in angiogenesis.

THYMOSIN BETA 4 PROMOTES CELL MIGRATION

Because thymosin beta 4 is most highly expressed by endothelial cells on Matrigel at a time
of high migration, it was not surprising to find that it was very active in promoting cell

18

16 • Control
Thymosin B4
14
~
"iii 12
c
Q.I
0 10
Q)
<.> 8
c
RI
Q.I 6
~
4

0
Plastic Lamlnln Collagen IV

Substrate

Figure 2. Attachment of human umbilical vein endothelial cells to various substrates.

Cells either transfected with thymosin beta 4 or mock transfected (control)
were allowed to attach in the absence of serum for I hour to either plastic,
laminin or collagen IV (10 ugl16 mm diameter culture dish). The unattached
cells were removed by washing and the attached cells were counted after
trypsininzation in a Coulter counter.

160
Figure 3. Appearance of nonnal and thymosin beta 4 transfected human umbilical vein
endothelial cells. Upper panels show endothelial cells immunostained with an
antibody to thymosin beta 4 following transfection. The lower panels show
normal and transfected cells 30 minutes after plating onto tissue culutre plastic.
The transfected cells are more spread than the non transfected cells.

migration in vitro and in vivo (Malinda et aI., 1997). Using a Boyden chamber assay,
thymosin beta 4 was found to stimulate endothelial cell migration in a dose-dependent
manner with half maximal activity at 10 ng/mi. Since it is found at 1 x 10 -5 to 5 x 10 -4
molar in tissues, this is a reasonable physiological level of activity. Thymosin beta 4
stimulates directed migration in the Boyden chamber checkerboard assay and is cell type
specific. Human umbilical vein endothelial cells and coronary artery endothelial cells
migrated towards thymosin beta 4 whereas smooth muscle cells, neutrophils and HT1080
fibrosarcoma cells did not. Thymosin beta 4 was also active in another in vitro migration
assay using endothelial cells. Here a monolayer of endothelial cells was "scratch wounded"
and then the rate of migration of the cells into the wounded area was detennined. In the
presence ofthymosin beta 4, the rate of migration was greatly increased at all time points
tested in a dose-dependent manner (Figure 4). These data confinn that thymosin beta 4 can
promote cell migration when added exogenously to cells either in a gradient fonn (the
Boyden chamber assay) or as a constant amount (the scratch wound assay).

Thymosin beta 4 was also tested in vivo for its ability to promote migration into a
subcutaneously injected basement membrane Matrigel plug. This assay has been used to
assess angiogenesis. Thymosin beta 4 was found to increase endothelial cell migration and
angiogenesis with maximal activity at 5 ug/mi. The maximal activity was comparable to
that observed with FGF. Thus, thymosin beta 4 is active in vivo in promoting
angiogenesis.

161
 80.-----------------------------,

.......
OJ
60 ~ :/S. ..............().
J ' J' .....

:;
<Jl
0
G
"0
I

•J (.'
..
" .....
....
••••

40
,.' . ....
c:
::J
0
~
0
,/ ...../
0 ~ ...-

20 J/ . . . . --0- ECGS (+ control)

,. Ii ______ 1000 ngl ml
AI.'.:. . . ·. ·. . . .+.. 100 ngl ml
..... ...t:J..... 0 ngl ml
o .."""""~----r_--.....----.--=<.---.--___l
o 2 4 6 8 10 12

Time (hr)

Figure 4. Migration of human umbilical vein endothelial cells in the "scratch wound"
assay. A monolayer of confluent endothelial cells was scratched with a blue tip
from PGC. The distance the cells had migrated at various times in the presence
ofthymosin beta 4 was determined.

SUMMARY
The morphological differentiation of endothelial cells on basement membrane Matrigel is a
model system to study some of the steps in angiogenesis. Nearly all of the factors found
to either stimulate or impair tube formation on Matrigel have similar activty in other in
vitro assays and in vivo. Thymosin beta 4 mRNA was found by subtractive cloning of
endothelial cells on Matrigel to be increased some 4 to 6 fold at four hours after plating on
Matrigel. Thymosin beta 4 regulates many of the activities involved in angiogenesis
including cell adhesion, migration, protease activity and tube formation . Our data suggest
that thymosin beta 4 could be an important angiogenic factor in vivo. Its small size ensures
that it is highly diffusable and accessible in areas undergoing angiogenesis.

REFERENCES

Badamchian, M ., Strickler, M.P., Stone, MJ., and Goldstein, A.L., 1988, Rapid isolation
of thymosin beta 4 from thymosin fraction 5 by preparative high-performance liquiq
chromatography, 1. Chromatography 459:291-300.

Bao, L., Loda, M., Janmey, P.A., Stewart, R., Anand-Apte, B., and Zetter, B.R. , 1996,
Thymosin beta 15: A novel regulator of tumor cells motility upregulated in metastatic
prostate cancer. Nature Medicine 2: 1322-1328.

Cassimeris, L., Safer, D., Nachmias, V.T., and Zigmond, S.H., 1992, Thymosin beta 4
sequesters the majority ofG-actin in resting human polymorphonuclear leukocytes, J. Cell
BioI. 119: 1261-1270.

162
Cid, M. e., Grant, D.S., Hoffman, G.S., Auerbach, R., Fauci, AS., and Kleinman, H.K.,
1993, Identification of haptoglobin as an angiogenic factor in sera from patients with
systemic vasculitis, I Clin. Invest. 91:977-985.

Condon, M.R., and Hall, A.K, 1992, Expression of thymosin beta 4 and related genes in
developing human brain, I Mol. Neurosci. 3:1650170.

Grant, D.S., Tashiro, KJ., Sequi-Real, 8., Yamada, Y, Martin; G.R., and Kleinman, H.K,
1989, Two different laminin domains mediate the differentiation of human endothelial cells
into capillary-like structures in vitro, Cell 58: 933-943.

Grant, D.S., Kleinman, H.K, and Martin, G.R., 1990, The role of the basement membrane
in vascular development, Ann.N.Y Acad. Sci 588:61-72.

Grant, D.S., Kinsella, IL., Kibbey, M.e., Laflamme, S., Burbelo, P.O., Goldstein, AL.,
and Kleinman, H.K, 1995, Matrigel induces thymosin beta 4 gene in differentiating
endothelial cells, I Cell Sci. 108:3685-3694.

Folkman, I, 1992, The role of angiogenesis in human growth, Semin. Cancer Bio 3: 65-71.

Kubota, Y, Kleinman, H.K, Martin, G.R., and Lawley, TJ., 1988, Role of laminin and
basement membrane in the morphological differentiation of human endothelial cells into
caplillary-like structures, I Cell BioI. 107: 1584-1598.

Lin, S., and Morrison-Bogorad, M., 1990, Developmental expression ofmRNAs endcoding
thymosin beta 4 and beta 10 in rat brain and other tissues, I Mol. Neurosci. 2: 35-44.

Low, T., and Goldstein, A, 1984, Thymosins:structure, function and therapeutic

application, Thymus 6: 27-42.

Malinda, KM., Goldstein, AL., and Kleinman, H.K, 1997, Thymosin beta 4 stimulates
directional migration of human umbilical vein endothelial cells, FASEB I, in press.

Morales, D.E., Grant, D.S., Maheshwari, S., Bhartiya, D., Cid, M.e., Kleinman, H.K , and
Schnaper, H.W., 1995, Estrogens promote angiogenic activity in human umbilical vein
endothelial cells in vitro and in a murine model, Circulation 91: 755-763.

Schnaper, H.W., Barnathan, E.S., Mazar, A., Maheshwari, S., Ellis, S., and Kleinman,
H.K, 1995, Plasminogen activators augment endothelial cell organization in vitro by two
distinct pathways, I Cellul. Physiol. 165107-118.

Schnaper, H.W., Grant, D.S., Stetler-Stevenson, W.G., Fridman, R., O'Orazi, G., Bird,
B.E., Hoythya, M., Fuerst, T.R., French, D.L., Quigley, IP., and Kleinman, HK, 1992,
Type IV collagenase activity promotes endothelial cell formation into capillary-like
structures on basement membrane in vitro, I Cellul. Phsyiol. 156: 235-246.

163
STRUCTURAL STUDIES ON ANGIOGENIN, A PROTEIN IMPLICATED IN
NEOVASCULARIZATION DURING TUMOUR GROWTH

K. Ravi Acharya l , Demetres D. Leonidas l , Anastassios C.

Papageorgiou l , Nello Russ02 and Robert Shapiro2,3

1 Department of Biology and Biochemistry, University of Bath,

Claverton Down, Bath BA2 7AY, UK

2 Center for Biochemical and Biophysical Sciences and Medicine,
Harvard Medical School, Boston, Massachusetts 02115, USA
3 Department of Pathology, Harvard Medical School, Boston,
Massachusetts 02115, USA

1. INTRODUCTION

Angiogenesis, the formation of new blood vessels, is an essential part of normal

physiological processes such as embryonic growth, wound healing, and the cyclical
development of the uterine endometrium. It also occurs in a variety of pathological
conditions including arthritis, diabetic retinopathy, and tumour growth (Folkman and
Cotran, 1976). Early observers had noted a proliferation of blood vessels in the vicinity of
such tumours (see Vallee e/ al., 1985), and it was later proposed by Folkman (1971) that
these tumours are totally dependent on angiogenesis for growth beyond a diameter of 1-2
mm. Angiogenesis is also thought to be a prerequisite for the development of metastases
since it provides the means whereby cells disseminate from the original primary tumour.

A central feature of this 'tumour angiogenesis' model has been the idea that the process is
mediated by a messenger derived from the tumour cells. Consequently, over the past two
decades considerable effort has been invested in the search for such angiogenic factors,
resulting in the isolation and molecular characterisation of at least 10 proteins, and the
identification of several smaller factors, all with angiogenic activity (Folkman and Shing,
1992). The precise molecular events triggered by angiogenic molecules still remains unclear.
However, a single, common mechanism seems unlikely since the biological properties of
these agents vary considerably. For example, acidic and basic fibroblast growth factors
(FGFs), angiogenin (Ang), vascular endothelial growth factor (VEGF), and hepatocyte
growth factor (HGF) are mitogenic for endothelial cells in culture, whereas platelet derived

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 165
endothelial growth factor (pD-ECGF) has been found to have no direct effect on
proliferation, and transforming growth factor-~ and tumour necrosis factor-a are in fact
inhibitory. The FGFs, VEGF and Ang bind tightly to heparin, whereas the others do not.
At least two of the proteins, Ang and PD-ECGF, are enzymes, and in the former case this
activity clearly is an essential mechanistic component. Other properties of the various
proteins that may be associated with angiogenesis include their capacities to induce
proteases, increase vascular premeability, attract monocytes, and support adhesion of
endothelial cells.

Human angiogenin, a single-chain polypeptide (M r 14,124), is a potent inducer of

neovascularization on the chicken embryo chorioallantoic membrane (Fett et aI., 1985) and
the rabbit knee meniscus (King and Vallee, 1991). Although originally isolated from tumour
cell conditioned medium (Fett et al., 1985), it is also a component of normal serum
(Shapiro et aI., 1987). Among the angiogenic molecules, Ang is unique in that it is a
ribonucleolytic enzyme (Shapiro et aI., 1986), with a sequence that is 33 % identical to that
of bovine pancreatic RNase A (Strydom et al., 1985). Its enzymatic activity toward
conventional RNase substrates is several orders of magnitude lower than that of RNase A.

Ang has recently been shown to be a direct mitogen for vascular endothelial cells in sparse
cultures and to bind specifically to a 170-kDa endothelial cell-surface protein that may be
its functional receptor (Hu et al., 1997). Subsequent to binding, Ang is internalized and
transported to the nucleolus (Moroianu and Riordan, 1994). This translocation is critical
for the angiogenic response and may bring Ang into contact with its natural RNA

Figure 1: Polypeptide fold for human Ang [drawn with the program MOLSCRIPT
(Kraulis,1991)].

166
substrate. Ang also stimulates cell-associated proteolytic activity and invasiveness (Hu et
aI., 1994) and is capable of serving as an effective substratum for endothelial cell adhesion
and spreading (Soncin, 1992).

Biochemical studies involving mutagenesis, proteolysis, and chemical modification of Ang

have identified several amino acid residues that play important roles in catalysis or
substrate binding and have at least partially accounted for the enzymatic differences
between Ang and RNase A (Shapiro et al., 1989; Shapiro and Vallee, 1989; Harper and
Vallee, 1989; Shapiro and Vallee, 1992, Curran et al., 1993). Moreover, they have
demonstrated that the angiogenic activity of Ang requires not only an intact catalytic site
(Shapiro and Vallee, 1989, 1992; Shapiro et aI., 1989; Curran et al., 1993), but also another
region, thought to constitute a cell or receptor binding site (Hallahan et al., 1991, 1992). In
addition, these and other studies (Lee and Vallee, 1989) have implicated various Ang
residues as important for the tight binding of Ang to human placental ribonuclease inhibitor
(RI), an effective inhibitor of enzymatic and angiogenic activities of Ang (Shapiro and
Vallee, 1987). RI binds to Ang with a K j of 7 x 10- 16 M and the t\12 for dissociation of the
complex is 60 days, making this one of the tightest protein-protein interactions known
(Lee et al., 1989). RI has been shown to suppress the growth of syngeneic mammary
tumours in mice (polakowski et al., 1993) and may be involved in the regulation of the
normal physiological function of Ang in vivo.

2. STRUCTURAL STUDIES ON ANGIOGENIN

2.1 Crystal structures of human angiogenin

We have reported a crystal structure of Ang [Met-(-I)] form at 2.4 A resolution (Acharya
et aI., 1992, 1994, Figure 1) and more recently we have determined the crystal structure of
Ang in its natural form «Glu-l) (Leonidas et a/., unpublished results). [The Met-(-I) and
<Glu-l forms are functionally indistinguishable (Shapiro et aI., 1988).] The refinement of
the high resolution structures to 2.0 A is nearly complete. From the 3D structure of Ang it
is clear that two important regions of Ang are significantly different from the corresponding
parts of RNase A (Figures 2 and 3): the ribonucleolytic active site and the putative
receptor binding site. The pyrimidine-binding subsite (BI) required for RNase activity
(which is completely accessible in RNase A) is 'blocked' in Ang by a glutamine residue,
Gln-ll7 (Figures 4-6). This change in the BI-site architecture is a consequence of the
markedly different orientations and secondary structures of the C-terminal segments of the
two proteins. The blockage of the BI site may underlie much of the decrease in enzymatic
activity for Ang compared to RNase A. It also suggests a valid explanation for the
unexpected activity increases previously seen to accompany mutation of Asp-116 (Harper
and Vallee, 1988): Gln-117 may move more easily out of the active site cleft once the
interactions of Asp-1l6 with Ser-1l8 have been disrupted. This finding provides a
potential mechanism (movement of Gln-l 17) by which Ang might be activated at the
appropriate time and location in vivo to become a more potent RNase. Hence, Gln-117
may act as a 'conformational switch' to trigger movement of the secondary structure
elements and provoke a rearrangement of this site upon ligand/ receptor binding (Figures 4-
6). The blockage of the BI site of Ang by Gln-117 observed in the crystal structure has
been confirmed by mutagenesis (Russo et at., 1994): mutations of Gln-lI7 to Ala and to
Gly were found to increase activity 11- to 18- fold and 21- to 30- fold, respectively,
toward dinucleotide, polynucleotide and cyclic nucleotide substrates. This is a key finding

167
 Hl H2
<---------------------> <-
I 10 20
+ * * + * *
Human ANG q d n s R Y T H F L T Q H Y D A - k P q - G r d
-
Bovine RNase A k e - - - t A A A K F E R Q H M D S s t s a a S s -
1 10 20

H2 Bl H3
--------------------> <-----------> <------------
30 40 50
* * * * * * * * * * * * * *
d R Y C E S I M R R R G L t s - p C K D I N T F I H 9 N K R S I K A
s N Y C N Q M M K S R N L t k d r C K P V N T F V H e S L A D V Q A
30 40 50

H3 B2 B3 B4
----> <-----> <-------> <--------------->
60 70 80
* * * + + * * * *
I C e n k n 9 n P h r - - e n - 1 r i s k s s F Q V T T C K L H
V C s - q k n v a c k n 9 q t n c y q s y s t M S I T D C R E T
60 70 80

B5 B6 B7
<---------------> <---------> <--------
90 100 110
+ + + + + + * * * * + *
9 9 s P P c
P W - q Y R A T A G F R N V V V A C E n - - 9 1 p V H L
9 s s k Y P n - c a Y K T T Q AN K H I I V A C E 9 n P y v P V H F
90 100 110 120

310
-:><------->
120 123
+ +
D q s i f r r p
D a - s v
124

Figure 2: Structure based amino acid sequence alignment for Ang and bovine pancreatic
RNase A. Structurally equivalent residues, represented by capital letters are those whose
Ca positions superimpose within 1.2 A in the two structures. Residues represented in
small letters deviate more significantly. Hyphens indicate that there is no corresponding
equivalent residue in the 3D structure. Positions of the secondary structure elements in
Ang are indicated and solvent accessible residues (less than 20 A2 of exposed surface) are
indicated by the symbol X (Acharya et aI., 1994). Residues involved in crystal packing
contacts for Ang are represented by +.

which has emerged from our crystallographic study and was not predicted from simple
molecular modelling based on the RNase A structure (as discussed by Allen et aI., 1994).
The putative receptor binding site in Ang is thought to include part of the segment 58-70 as
well as residues around Asn-109 (Figure 7), which are nearby in the three-dimensional
structure. This region is significantly different in sequence from the corresponding region
of RNase A (Figure 2) , which in fact constitutes the purine-binding B2 sub site and the

168
Figure 3: Stereo diagram of Ca. superposition for RNase A (thin line) and Ang (thicker
line) based on 116 equivalenced Ca. atoms fitted with 2.26 A r.m.s. deviation. The regions
of Ang that deviate significantly from their RNase counterparts are highlighted.
Particularly noteworthy are the C-terminal segment 117-123 (catalytic site) and 58-70
(receptor site) of Ang. Residues marked correspond to Ang (Acharya et aI., 1994).

peripheral phosphate-binding site Po: there is, at most, one residue conserved and the Ang
segment lacks a disulphide bond present in all known mammalian pancreatic RNases. It is
also shorter by two residues. Our X-ray structure reveals, surprisingly, that there is in fact
an insertion in Ang as well as extensive deletions (Figures 2 and 7). The insertion in Ang
contains the angiogenically critical residue Asn-61 and the deletion includes 3 residues
(Lys-66, Asn-67 and Gln-69) occupying the B2 or Po site (Figure 2). These differences
probably account for the non-angiogenicity of RNase, on the one hand, and the less
stringent B2 site specificity of Ang, on the other.

2.2 Crystal structure of bovine angiogenin

In order to understand further the structure-function relationships of Ang, we have also

determined the 3D structure of bovine Ang (bAng) at 1.5 A (Acharya et aI., 1995). This
protein has 64 % sequence identity to human Ang, virtually the same angiogenic potency,
and similar enzymatic properties (Bond and Vallee, 1988). However, its primary structure
differs in two potentially significant respects: Glu-118 replaces Gln-II7, and it contains a
putative Arg-Gly-Asp (RGD) recognition element that is not present in the human protein.
From the crystal structure of bAng (Acharya et aI., 1995) it is clear that Glu-118 blocks the
pyrimidine binding site in precisely that same manner as does Gln-117 in human Ang and
is, in fact, involved in an even more extensive set of stabilising interactions. Furthermore,
the RGD sequence is not structurally similar to known integrin-dependent recognition
sites. The crystallographic structure is nearly identical to the recently reported NMR
structure of bAng (Lequin et aI., 1996).

169
Figure 4: a) Portion of the 2Fo-Fc electron density map showing the catalytic triad (His-
13, Lys-40, His-II4) in Ang structure. The two water molecules which have been
identified are represented by the symbol * and are denoted by the letter W. A prominent
feature of this structure is the location of the side chain Gln-II7 which occupies the
position corresponding to the B J substrate binding pocket of RNase A . Contours are
drawn at the level of 0.9 (J using the refined structure at 2.4 A resolution. b) A
corresponding view of RNase A (Wlodawer et a!., 1982) showing the catalytic residues
His-I2, Lys-4I, and His-II9. Ala-I22, the residue analogous to Gln-II7 of Ang, is located
far away from the BJ pocket.

170
 \ I~!I
n~ ..•. ".
\ ...•
....
• •
Figure 5: Stereo view of the active site of Ang. All major interactions are indicated by
dashed lines [figure drawn with MOL SCRIPT Program (Kraulis, 1991)].

2.3 Crystal structures of human angiogenin mutants

Recently we have detennined the 3D structure at 2.0 A resolution for a catalytically and
angiogenically inactive mutant in which Lys-40 has been replaced by Gin (K40Q)
(Leonidas et al., unpublished results). A preliminary comparison of the native and mutant
structures indicates that the two structures are virtually indistinguishable and the
orientations of all the active site components are identical. The putative receptor binding
site in the mutant protein is unchanged except for a small shift of Arg-66. These results
support the previous proposal (Shapiro et al., 1989) that K40Q lacks angiogenic activity
solely as a consequence of its decreased enzymatic activity.

We have also detennined the structure of the Q117G Ang mutant at 2.0 A resolution
(Leonidas et aI., unpublished results). The mutant structure differs from that of native Ang
only in that the 117 side chain has been removed, indicating that the single amino acid
replacement is not sufficient to induce a confonnational change in Ang. Modelling with
uridine vanadate (UV), a transition state analog of RNase A, shows that binding of the
uracil ring is still sterically impossible. Thus Q117G adopts an inactive confonnation and
must still undergo a rearrangement to open the blocked pyrimidine site and act on RNA.

3. STRUCTURE-BASED INHmITOR DESIGN

Anti-angiogenic compounds will be of therapeutic value in counteracting pathological

conditions, including cancer, that require blood vessel proliferation. Numerous molecules
have been shown to inhibit angiogenesis in vivo and several of these agents are reported to

171
 T+I
\"

I~~ \.~'"
"

Figure 6: a) Stereo view of the modelled complex of Ang with uridine vanadate (UV) based
on superposition of the Ang structure and the complex of RNase A with UV (Borah et ai.,
1985). b) Corresponding stereo view of the complex of RNase A with UV.

retard tumour growth in animal models (Shapiro and Vallee, 1987; Polakowski et ai., 1993;
O'Reilly et aI., 1994, 1997). Although the pharmacological properties of some of these
molecules preclude their use as drugs in humans, others are currently under investigation
for this purpose.

In the case of Ang, recent studies indicate that Ang antagonists are extremely effective in
preventing growth of solid tumours (Olson et ai., 1994, 1995; Olson and Fett, 1996). Thus,
treatment with non-cytotoxic, neutralising anti-Ang monoclonal antibodies or with the
Ang-binding protein actin at the time of and following the implantation of HT -29 human

172
Figure 7: Stereo view of putative receptor binding regions of Ang (thick lines) and
corresponding residues of RNase A (thin lines). a) Ang residues 60-68 and RNase residues
59-7l. b) Ang residues 108-111 and RNase residues 110-115 (Ang residues are labelled with
one letter amino acid code).

173
colon carcinoma cells in nude mice markedly inhibits their ability to become established. In
a typical experiment, 65% of antibody-treated mice were tumour free after 35 days
compared to only 3% of untreated mice. Histological data strongly suggest that this
inhibition is due to interference with Ang-induced blood vessel formation. Monoclonal
antibodies are also effective against several other human tumor types tested, including lung,
prostate, and fibrosarcoma. These results indicate that Ang is critical for the growth of at
least some human tumours. Since numerous additional human tumour cell lines examined
thus far also produce Ang, these antibodies may be useful as broad-spectrum reagents
against various solid human tumours and especially against their potential metastases. At
the same time, the application of monoclonal antibodies in human therapy is problematic
(Waldmann, 1991) and it is therefore critical to identify or design other types of Ang
inhibitors.

We are following multiple avenues toward this long-term goal, all of them based on the high
resolution 3D structures of Ang described above. Initial nucleotide inhibitors of Ang have
been identified by kinetic mapping of the Ang active site (Russo et aI., 1996) and by
testing and adapting inhibitors of RNase A (Leonidas et at., 1997). The ideal inhibitor
would be a small molecule; this could have significant advantages over protein-based
antagonists in terms of better delivery and lower immunogenicity and cost of production.
The best small inhibitor of Ang identified in our kinetic studies is 5'-diphosphoadenosine
2'-phosphate (ppA-2'-p), with a Ki value of 110 11M. Although ppA-2'-p does not bind
sufficiently tightly to be useful as a drug, its interaction with Ang is strong enough to allow
detailed crystallographic studies. This Ang complex should now provide a basis for
pursuing a "rational design" approach: i.e., using X-ray crystallography and computer
modelling techniques to reveal the actual interactions of the inhibitor with Ang and identify
others that can potentially be formed with additional substituents. Thus ppA-2'-p may
serve as a suitable starting point for the design of tight binding inhibitors of Ang as anti-
angiogenic agents for human therapy.

A second avenue we are pursuing is to explore the structural basis for the remarkably tight
binding of human RI to Ang (see above) and use this information to design modified forms
of RI. RI is a member of a family of 50-kDa, leucine-rich cytoplasmic proteins present in
essentially all mammalian tissues. Although originally isolated and characterised as an
inhibitor of RNase A, it also acts on virtually all other members of the RNase superfamily
examined thus far, and Ang is its most avid known ligand. Several features, however, limit
the potential therapeutic utility of RI- i) its relatively large size, ii) its susceptibility to
oxidation and inactivation by metals, and iii) its lack of specificity, which may require the
use of high doses and result in undesirable side effects. Detailed knowledge of the 3D
structure of the Ang-RI comprex obtained by X-ray crystallography would help in the
identification of specific RI residues that make functionally important contacts with Ang,
and thereby make it possible to design single or multi domain RI derivatives that bind
tightly and specifically to Ang and that are more stable. Toward this goal we have recently
determined the structure of the complex at 2.0 A resolution (papageorgiou et aI.,
unpublished results). The structure reveals significant differences in the modes of
interaction of RI with Ang vs. RNase A and suggests ways in which attempts to
"minimize" RI might proceed.

174
3. ACKNOWLEDGEMENTS

We thank Dr. B.L.Vallee, Dr. IF.Riordan and Dr. J.W.Fett for advice and support.The
research work reported in this article was funded by grants from the Cancer Research
Campaign (UK), the Medical Research Council (UK), the Wellcome Trust (UK), the
National Institutes of Health (USA), and the Endowment for Research in Human Biology,
Inc. (Boston).

4. REFERENCES

Acharya, K. R., Shapiro, R., Allen, S. c., Riordan, J. F. and Vallee, B. L.1994, Crystal
structure of human angiogenin reveals the structural basis for its functional divergence from
ribonuclease. Proc. Nat!. Acad Sci. USA, 91: 2915-2919.

Acharya, K. R., Shapiro, R., Riordan, l F. and Vallee, B. L. 1995, Crystal structure of
bovine angiogenin at 1.5 A resolution. Proc. Natl. Acad Sci. USA, 92: 2949-2953.

Allen, S C., Acharya, K. R., Palmer, K. A., Shaprio, R., Vallee, B. L. and Scheraga, H A.
1994, A comparison of the predicted and X-ray structures of angiogenin. Implications for
further studies of model building of homologous proteins. J. Prot. Chem., 13: 649-658.

Borah, B., Chen, C. W., Egan, W., Miller, M., Wlodawer, A. and Cohen, J. S. 1985, Nuclear
magnetic resonance and neutron diffraction studies of the complex of RNase A with uridine
vanadate, a transition state analog. Biochemistry 24: 2058-2067.

Curran, T. P., Shapiro, R. and Riordan, 1. F.1993, Alteration of the enzymatic specificity of
human angiogenin by site-directed mutagenesis. Biochemistry, 32: 2307-2313 .

Fett, J. W., Strydom, D. J., Lobb, R. R., Alderman, E. M., Bethune, J. L., Riordan, J. F. and
Vallee, B. L. 1985, Isolation and characterization of angiogenin, an angiogenic protein from
human carcinoma cells. Biochemistry, 24: 5480-5486.

Folkman, J. and Cotran, R. S. 1976, Relation of vascular proliferation to tumor growth . Int.
Rev. Exp. Path., 16: 207-248.

Folkman, J. 1971, Tumor angiogenesis: Therapeutic implications. N. Engl. J. Med., 285:

1182-1186.

Folkman, land Shing, Y. 1992, Angiogenesis. J. Bio!. Chem. 256: 10931-10934.

Hallahan, T. W, Shapiro, R. and Vallee, B. L.1991, Dual site model for the organogenic
activity of angiogenin. Proc. Natl. Acad Sci. USA, 88: 2222-2226.

Hallahan, T. W., Shapiro, R. and Vallee, B. L. 1992, Importance of asparagine-61 and

asparagine -109 to the angiogenic activity of human angiogenin. Biochemistry, 31: 8002-
8029.

175
Harper, J. W. and Vallee, B. L. 1988, Mutagenesis of aspartic acid-1l6 enhances the
ribonucleolytic activity and angiogenic potency of angiogenin. Proc. Nat!. Acad Sci. USA,
85: 7139-7143.

Harper, J. W. and Vallee, B. L. 1989, A covalent angiogeninlribonuclease hybrid with a

fourth disulfide bond generated by regional mutagenesis. Biochemistry, 28: 1875-1884.

Hu, G.-f., Riordan. J. F. and Vallee, B. L. 1994, Angiogenin promotes invasiveness of

cultured endothelial cells by stimulation of cell-associated proteolytic activities. Proc. Natl.
Acad. Sci. USA, 81: 12096-12100.

Hu, G.-f., Riordan. J. F. and Vallee, B. L. 1997, A putative angiogenin receptor in

angiogenin-responsive human endothelial cells. Froc. Natl. Acad Sci. USA, 94: 2204-2209.

King, T.V. and Vallee, B.L. 1991, Neovascularization of the meniscus with angiogenin. 1.
Bone Joint Surg., 73-B: 587-590.

Kraulis, P. J. 1991. MOLSCRIPT - a program to produce both detailed and schematic plots
of protein structures. 1. Appl. Crystallogr., 24: 946-950.

Leonidas, D. D., Shapiro, R., Irons, L. L., Russo, N. and Acharya, K. R. 1997, Crystal
structures of Ribonuclease A complexes with 5'-Diphosphoadenosine 3'-phosphate and 5'-
Diphosphoadenosine 2'-phosphate at 1.7 A resolution. Biochemistry, 36: 5578-5588.

Lequin, 0., Albaret, c., Bontems, F., Spik, G. and Lallemand, J. Y. 1996, Solution structure
of bovine angiogenin by 1H nuclear magentic resonance spectroscopy. Biochemistry, 35,
8870-8880.

Lee, F.S., Shapiro,R. and Vallee,B. L. 1989, Tight-binding inhibition of angiogenin and
ribonuclease A by placental ribonuclease inhibitor. Biochemistry, 28: 225-230.

Lee, F. S. and Vallee, B. L. 1989, Binding of placental ribonuclease inhibitor to the active
site ofangiogenin. Biochemistry, 28: 3556-3561.

Moroianu, J. and Riordan, J.F. 1994, Nuclear translocation of angiogenin in proliferating

endothelial cells is essential to its angiogenic activity. Proc. Natl. Acad Sci. USA., 91: 1677-
1681.

Olson, K. A., French, T. C., Vallee, B . L. and Fett, J. W. 1994, A monoclonal antibody to
human angiogenin suppresses tumor growth in athymic mice. Cancer Res., 54: 4576-4579.

Olson, K. A, Fett, J. W., French, T. C., Key, M. E. and Vallee, B. L. 1995, Angiogenin
antagonists prevent tumor growth in vivo. Proc. Natl. Acad Sci. USA, 92: 442-446.

Olson, K. A and Fett, J. W. 1996, Prostatic carcinoma therapy with angiogenin antagonists.
Proc. Amer. Assoc. Canc. Res., 37: 57.

O'Reilly, M. S., Holmgren, L., Shing, Y, Chen, C., Rosenthal, R. A, Moses, M., Lane, W.
S., Cao, Y., Sage, E. H. and Folkman, J. 1994, Angiostatin: A novel angiogenesis inhibitor
that mediates the suppression of metastases by a Lewis lung carcinoma. Cell, 79: 315-328.

176
O'Reilly, M. S., Boehm, T., Shing, Y., Fukai, N., Vasios, G., Lane, W. S., Flynn, E.,
Birkhead, 1. R., Olsen, B. R. and Folkman, 1. 1997, Endostatin: An endogenous inhibitor of
angiogenesis and tumor growth. Cell, 88: 277-285.

Polakowski, I. 1., Lewis, M. K., Muthukkaruppan, V., Erdman, B., Kubai, L. and
Auerbach, R. 1993, A ribonuclease inhibitor expresses anti-angiogenic properties and leads
to reduced tumor growth in mice. Am. J. Pathol., 143: 507-517.

Russo, N ., Shapiro, R., Acharya,K. R., Riordan, 1. F. and Vallee, B. L. 1994, The role of
glutamine-117 in the ribonucleolytic activity of human angiogenin. Proc. Natl. Acad Sci.
USA, 91: 2920-2924.

Russo, N., Acharya, K. R., Vallee,B. L. and Shapiro, R. 1996, A combined kinetic and
modeling study of the catalytic center subsites of human angiogenin. Froc. Natl. Acad Sci.
USA, 93: 804-808.

Shapiro, R., Fox, E. A. and Riordan, 1. F. 1989, Role of Iysines and human angiogenin -
chemical modification and site-directed mutagenesis. Biochemistry, 28: 1726-1732.

Shapiro, R., Harper, J.W., Fox, E. A., Jensen, H-W., Hein, F. and Uhlmann, E. 1988,
Expression ofMet-(-I) angiogenin in Escherichia coli. Anal. Biochem., 175: 450-461.

Shapiro, R., Strydom, D. J., Olson, K.A. and Vallee, B.L. 1987, Isolation of angiogenin from
natural human plasma. Biochemistry, 26: 5141-5146.

Shapiro, R. and Vallee, B. L. 1987, Human placental ribonuclease inhibitor abolishes both
angiogenic and ribonucleolytic activities ofangiogenin. Proc. Natl. Acad. Sci. USA, 84: 2238-
2241.

Shapiro, R. and Vallee,B. L. 1989, Site-directed mutagenesis of histidine-13 and histidine-

114 of human angiogenin. Alanine derivatives inhibit angiogenin-induced angiogenesis.
Biochemistry, 28: 7401-7408.

Shapiro, R. and Vallee, 8. L. 1992, Identification of functional arginines in human angiogenin

by site-directed mutagenesis. Biochemistry, 31: 12477-12485.

Soncin, F. 1992, Angiogenin supports endothelial and fibroblast cell adhesion. Proc. Natl.
Acad Sci. USA, 89: 2232-2236.

Strydom, D. 1., Fett, J. W., Lobb, R. R., Alderman, E. M., Bethune, J. L., Riordan, J. F. and
Vallee, B. L., 1985, Amino-acid sequence of human-tumor derived angiogenin. Biochemistry,
24: 5486-5494.

Vallee, B. L., Riordan, 1. F. , Lobb, R. R., Higachi, N ., Fett, 1. W., Crossley, G., Buhler, R.,
Budzik, G ., Breddam, K., Bethune, 1. L. and Alderman, E. M . 1985, Tumor-derived
angiogenesis factors from rat Walker 256 carcinoma: an experimantal investigation and
review. Experientia, 41: 1-15.

Waldmann, T.A. 1991 , Monoclonal antibodies in diagnosis and therapy. Science, 252: 1657-
1661.

177
 Wlodawer, A., Bott, R. and Sjolin, L. 1982, The refined crystal structure of Ribonuclease A
at 2.0 A resolution. J. Bioi. Chem., 257: 1325-1332.

178
LESIONAL LEVELS OF ENDOTHELIAL CELL STIMULATING ANGIOGENESIS
FACTOR (ESAF) AND VASCULAR ENDOTHELIAL CELL STIMULATING
ANGIOGENESIS FACTOR (VEGF) ARE ELEVATED IN PSORIASIS

B McLaughlin,* M Bushan,** ill Weiss,* C Griffiths**

*Wolfson Angiogenesis Unit, Rheumatic Diseases Centre, University of

Manchester, Hope Hospital, CSB, Manchester M6 SHD. UK
**Department of Dermatology, Hope Hospital, Manchester M6 SHD,
UK.

1. INTRODUCTION

Psoriasis is a common, disabling, chronic dermatosis that affects approximately 2% of the

UK population. Although genetically determined, the usual age of onset of the first lesions
is between the ages of 15 and 30. The redness of the psoriatic plaque is dermal hyperaemia
which results from dilation and increased tortuosity of capillaries in elongated dermal
papillae. Early microvascular changes are amongst the earliest signs of the development of
new psoriatic plaques[l]. Thus, psoriasis may be considered as an example of an
angiogenic disease.
Scaling of the plaques results from enhanced epidermal keratinocyte proliferation[2].
Psoriatic plaques are characterised by epidermal keratinocyte proliferation, intraepidermal
lymphocyte trafficking and dermal vascular proliferation[3]. Over the past several years
there has been considerable interest in the molecular mechanisms responsible for control of
dermal micro vessel formation in psoriatic lesions. The increased expression of pro-
angiogenesis cytokines such as vascular endothelial cell growth factor(VEGF) in active
plaques of psoriasis[ 4] underscores the importance of neovascularisation in the
pathogenesis of this disease.
The angiogenic process is initiated by controlled dissolution of the capillary basement
membrane by neutral matrix metalloproteinases (MMPs)[5]. Endothelial cell Stimulating
Angiogenesis Factor (ESAF), a low molecular weight, non protein angiogenic factor
specific for microvascular endothelial cells activates these neutral MMPs, including the
basement membrane degrading enzyme (Gelatinase-A)[6].
Early, unpublished work with Dr C.R.Payne ( Westminster Hospital, London, UK)
showed cultured fibroblasts from involved psoriatic sites secreted significant amounts of

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis. Plenum Press, New York, 1998 179
ESAF. This early study is now described as Study 1. More recently we have
demonstrated that levels ofESAF and VEGF are significantly elevated in lesional psoriatic
skin (Study 2).

2. MEmODS(Study 1.)

2.1 Clinical History (Studyl.)

Patients with typical chronic plaque psoriasis aged 20-45, male were included in the early
study. Patients with other clinical forms of psoriasis were excluded. Biopsies (2mm) were
taken from uninvolved skin (PN) and from involved skin (PP) from the elbow of patients.
Normal fibroblasts (NN) were obtained from the skin of normal volunteers, sex and age
matched with the psoriatic patients. Lignocaine was not used as an anaesthetic as this had
been previously shown to have a detrimental effect on subsequent fibroblast growth[7].
The technique of a disposable punch kit for the extraction of the biopsy is rapid and no
more uncomfortable than the injection of the local anaesthetic itself. All subjects gave
informed consent and the study had ethical committee approval. PP biopsies were taken
from the edge of lesions that had been untreated for at least seven days and PN biopsies
were taken at least 2cm away from existing lesions.

2.2 Fibroblast Cultures.

Dermal fibroblast lines, PP, PN, NN used in this initial study were established from
samples taken from punch biopsies[8]. The explants were cultured in standard 25cm3
culture flasks in Minimum Essential Medium (MEM) containing 20mM Hepes, 100U/ml
penicillin and 100 ug/ml streptomycin plus 10% heat inactivated foetal calf serum, at 37 0 C
in a 5% CO 2 humidified atmosphere. These primary fibroblasts cultures approached
confluence after four weeks of growth and were subsequently trypsinised with 0.05%
trypsin and 0.02% EDTA with sub passage at weekly intervals until passage number 12.
Fibroblasts in pre- confluence from each passage were incubated in serum free media for 72
hours. The supernatant was collected and stored at _20 0 C and the ESAF content
determined .

2.3. ESAF Assay

ESAF was extracted from the culture supernatant by extraction with w/v 2M MgCI2 in
order to dissociate ESAF from any carrier protein. This method of extraction followed the
method of Cooper et al 1991 [9] Quantification of levels of ESAF was performed using
the widely described activation of procollagenase assay[lO].

3. RESULTS I(Study 1)

Levels ofESAF in all fibroblast cultures, PP, PN, NN were the highest at passage 2 ( Fig
1.) At Passage 2lesional skin (PP) ESAF levels were significantly greater than non-Iesional
skin (PN) or normal (NN) and remained significantly elevated until passage 8.
This difference in ESAF levels between normal and uninvolved skin was highly significant.
There was no significant difference between uninvolved and normal fibroblasts at the early

180
 40 I PP
PP - In"ol ~
. - onnal
ESAF

2° 1~_
3 6 9 12
PA AGE MBER

Figure 1. Elevated ESAF levels(Expressed as % activation of procollagenase) in Involved

Psoriatic skin fibroblasts compared to normal fibroblasts over 12 passages

stage. At later passages the uninvolved skin fibroblasts also secreted a significantly higher
level ofESAF than normal. ( Fig2)

4. METHODS (Study 2)

4.1. Clinical History

For this study into ESAF levels in psoriasis, 15 patients, ( nine males and six females) in
the age range 18-72 with a mean age of 43 years with untreated chronic plaque psoriasis
were enrolled. The severity of psoriasis was assessed using the Psoriasis Area and Severity
Index (PASI)[11]. 6mm punch biopsies were taken from lesion and non lesion skin in a
similar manner to that described in the previous study, snap frozen and stored at _70 0 C
until required.

4.2 Assay of Angiogenic Factors

Direct tissue levels of ESAF were assayed using published methods of extraction and
quantified using the widely published ESAF assay using the activation of
procollagenase. [12, I 0]
VEGF levels were measured using a standard quantikine ELISA kit[13]. (R&D Systems,
Abingdon, UK)

P
40
F

20
'c::::::::::~ _ _ _ _ PN
L-________~==~NN
3 9 12
p 1BER

Figure 2. - Elevated ESAF levels(Expressed as % activation of procollagenase) m

Uninvolved Psoriatic skin fibroblasts compared to normal fibroblasts over 12 passages

181
 5. RESULTS (Study 2)

Levels of ESAF and VEGF were significantly elevated in psoriatic lesional skin: 13.72 ±
8.7 units/mg and 28.12 ± 17.6 respectively as compared to uninvolved, non-lesional skin,
5.27 ±2.9 units/mg; p=O.OOl and 8.8 ± 5.3 p< 0.001 respectively. Both ESAF and VEGF
tissue levels in involved skin were significantly correlated to psoriasis clinical severity
(PASI): r = 0.60 and 0.89 respectively . (Fig3)

6. SUMMARY

It appears from both these studies that ESAF levels are significantly elevated in plaques of
patients with psoriasis.
The elevated ESAF levels in involved psoriatic skin in early passage suggest either the
abnormality in the involved psoriatic skin fibroblast is due to a modification of the cell
phenotype in culture or is a reflection of an induced enzyme responsible for the de novo
synthesis ofESAF.
Study 2, also indicates an involvement of another angiogenic factor, VEGF.
ESAF and VEGF have both been shown to be elevated in the retina or serum of patients or
animal models with proliferative retinopathy[14]. Similarly, both have been shown to be
significantly up regulated in electrically stimulated skeletal muscle,[15] and recently have

Figure 3. Tissue levels ofESAF and VEGF in Psoriasis Patients

DOB SEX ESAF ESAF VEGF VEGF PASJ

UNITS/mg UNITS/mg ng'mg ng/mg SCORES
involved uninvolved involved uninvolved
16-10-24 M 20.80 5.00 46.80 7.38 25 .8
12-4-34 M 15.50 7.40 37.10 9.49 20.0
16-1-42 M 6.45 5.62 11.40 1.07 16.2
2-5-51 M 3.76 0.41 39.20 14.70 17.2
10-7-50 F 4.81 3.57 19.60 7.29 15.6
6-2-54 M 5.91 2.16 14.90 2.45 13.6
26-1-61 F 9.35 1.68 46.70 15.30 19.5
12-11-23 F 7.23 1.07 9.30 7.17 14.1
11-9-77 F 24.60 6.15 35.80 3.07 21.6
17-3-60 M 20.80 9.22 36.70 16.45 19.8
27-9-62 F 2l.l0 5.58 17.50 18.49 16.5
15-9-62 M 27.80 7.10 67.40 9.92 34.8
26-1-52 M 4.03 7.84 7.90 6.78 11.0
19-3-77 M 24.30 7.39 11.90 4.88 11.3
30-11-57 F 9.33 8.83 19.60 7.56 10.0
13.72 5.27 28.12 8.80
+8.7 ±2.9 ±17.6 ±5.3

LEGEND - Patient data for ESAF and VEGF extractions from involved and uninvolved
biopsies from psoriatic skin, correlated with the Psoriasis Area and Severity Index(PASI).

182
been shown to be significantly increased in the serum of patients with peripheral vascular
disease and ischaemic heart disease[16].
It is probable that enhanced ESAF and VEGF production is related to the dermal
hyperaemia and microvascular changes that characterise psoriasis.
A plethora of future studies are indiacted by these observations. It is not known if cell
types, other than fibroblasts are also responsible for the production of ESAF in the skin
and the involvement of the keratinocytes in the epidermis needs to be taken into
consideration.
From the present studies, psoriasis is an example of an angiogenic disease where
microvascular changes are thought to play a crucial role in pathogenesis. Modulation of
these angiogenic factors may be a future therapeutic strategy in the treatment of this
disease.

7. REFERENCES

1. Marks R, 1978, Epiderma activity in the involved and uninvolved skin of patients
with psoriasis. Br.J.Dermatol. 98:399
2. Van de Kerkhof p .c.,van Rennes H.,de Grood RM.,de Jongh GJ.,Bauer F.W.,Mier
P.D., 1983, Response of the clinically uninvolved skin of psoriatic patients to
standardised injury. Br.J.Dermatol. 109:287
3. Bata-Csorgo Z.,Hammerberg C.,Voorhees J.J.,Cooper K.D., 1995, Kinetics and
regulation of human keratinocyte stem cell growth in short term primary ex vivo
culture. Cooperative growth factors from psoriatic lesional T lymphocytes stimulate
proliferation among psoriatic uninvolved, but not normal stem cell keratinocytes.
J.Clin.lnvest. 95:317
4. Dvorak H.F.,Brown L.F.,Detmar M.,Dvorak A.M., 1995, Vascular permeability
factor/vascular endothelial growth factor, microvascular hyperpermeability and
angiogenesis. Am.J.Pathol. 146: 1029
5. AusprunkD.H.,Folkman 1.,1977, Migration and proliferation of endothelial cells in
newly formed blood vessels during angiogenesis. Microvasc.Res 14:53
6. McLaughlin B.,Weiss 1.B., 1996, Endothelial cell stimulating angiogenesis
factor(ESAF) activates progelatinase-A,prostromelysin-l and procollagenase and
reactivates their complexes with tissue inhibitors of metalloproteinases:a role for
ESAF in non-inflammatory angiogenesis. Biochem.J. 317:739
7. Rowland-Payne C.M., 1987, Psoriatic Science. Br.MedJ. 295: 1158
8. Kruse P.F.(Jr), 1974, Production of multilayered tissue cultures. Methods Enzymol.
32:568
9. Cooper RG.,McLaughlin B.,Choo J.J.,Taylor C.M.,Weiss J.B.,1991, Elevated
endothelial cell stimulating angiogenesis factor levels in rat skeletal muscle with
potential for capillary growth. J.Physiol 435: 14
10. Weiss 1.B.,Hill C.R.,Davis RJ.,McLaughlin B.,Sedowofia K.A,Brown RA., 1983,
Activation of procolJagenase by a low molecular weight angiogenic factor.
Biosci.Reps. 3:171
11. Fredriksson T.,Pettersson 0., 1978, Severe psoriasis - oral therapy with a new
retinoid. Dermatologica 157:238
12. Cooper RG.,Taylor C.M.,Choo 1.1.,Weiss 1.B.,I991, Elevated endothelial cell
stimulating angiogenesis factor activity in rodent glycolytic skeletal muscles. Clin.Sci.
81 :267

183
13. Pierce E.A.,Avery R.L.,Foley E.D.,Aiello L.P.,Smith L.E.H.,1995, VEGF expression
in a mouse model ofneovascularisation. PNAS(USA) 92:905
14. Taylor C.M.,Weiss J.B.,1989, Raised endothelial cell stimulating angiogenesis factor
in diabetic retinopathy. Lancet ii: 1329
15. Brown M.D.,Hudlicka O.,Makki R.F.,Weiss J.B., 1996, Low molecular mass
endothelial cell stimulating angiogenesis factor in relation to capillary growth induced
in rat skeletal muscle by low frequency electrical stimulation. Int.J.Microcirc.Clin.
Exp.15:111
16. Blann A.D.,Li lL.,McCollum C.N.,Bate A. ,Weiss lB., 1997, Angiogenesis in
atherosclerosis:possible roles for vascular endothelial cell growth factor, endothelial
cell stimulating angiogenesis factor and soluble E-selectin. Athersclerosis (Submitted
24-2-97)

184
 Role of Cell Adhesion
Molecules and Extracellular
Matrix in Angiogenesis
STRUCTURE AND FUNCTIONAL ROLE OF ENDOTHELIAL CELL-TO-CELL
JUNCTIONS

Pilar Navarro, Maria Grazia Lampugnani and Elisabetta Dejana

Vascular Biology Laboratory, Mario Negri Institute for Pharmacological

Research - Milano, Italy

ABSTRACT

Endothelial cell junctions are complex structures formed by transmembrane adhesive

molecules linked to a network of cytoplasmic/cytoskeletal proteins. These structures have
some features and components in common with epithelial cells but also some which are
specific for the endothelium. During angiogenesis, endothelial cells need first to dissociate
from neighbouring cells and invade the underlying tissues. Indirect evidence suggests that
vascular growth factor(s), besides inducing endothelial cell proliferation, could also change
endothelial junction organization and strength. After the first sprouting the new vascular
structures get organized in a more complex network. At this stage, molecules at junctions
are required for endothelial cell-to-cell anchorage and for vascular remodelling. These
structures are important not only for maintaining adhesion between endothelial cells and,
as a consequence, for the control of vascular permeability, but also for intracellular signal
transduction. The exact pathways of signalling through cell-to-cell contacts are still
obscure but seem to require the release of intracellular molecules from the junctional
complex and their translocation to the cytoplasm and/or the nucleus.

INTRODUCTION

Endothelial cells form the main barrier to the passage of plasma components and
circulating cells between blood and tissues. The integrity of the endothelial layer is
required for this property. Endothelial permeability is mostly regulated by the passage of
solutes through paracellular junctions (paracellular route) (Dejana, Corada, and
Lampugnani, 1995) and by a vesicle mediated uptake of components and their active
transport from the apical to the basal side of the cell membrane (transcellular route)
(Schnitzer, 1993).

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 187
 N-cad

Figure 1 - Schematic representation of endothelial cell-to-cell junctions.

See the text for details on specific molecular components. N-cadherin was found
associated to each of the known cadherins but localized mostly non-junctional. Some
molecular relationships among cytoplasmic components are inferred from epithelial cells.
a-, a-catenin; b-, b-catenin; p-, plakoglobinlg-catenin; VE-cad, VE-cadherin; N-cad, N-
cadherin.

188
In contrast to the vesicular system it is likely that the passage of plasma components
through endothelial junctions does not require specific receptors but is regulated by
diffusion and by the dynamic opening and closure of interendothelial gaps.
The molecular organization of cell to cell junctions in the endothelium has been partly
elucidated during the last few years (Dejana et al., 1995; Staddon and Rubin, 1996).
It is possible to distiguish at least two types of intercellular junctions in endothelial cells
(tight junctions, TJ and adherens junctions, AJ). These structures share the common
characteristic of been formed by specific transmembrane proteins which, through their
extracellular domain, promote cellular homotypic adhesion and through their cytoplasmic
region are anchored to a complex network of cytoplasmic and cytoskeletal proteins
(Aberle, Schwartz, and Kemler, 1996; Anderson and Van Itallie, 1995; Gumbiner, 1996;
Takeichi, 1993).
In addition to these junctional structures the endothelium presents other transmembrane
proteins, such as PECAM (platelet endothelial adhesive molecule), which can concentrate
at intercellular contacts and display homotypic adhesive properties (Newman, 1997).
The function and structure of endothelium can be different along the vascular tree and this
diversity is also reflected by the organization of cell to cell junctions. For instance, in the
brain where permeability needs to be under a strict control the intercellular junctions are
particularly well organized and form a fine complex and intersected seal. In the
postcapillary veins where the interchanges between blood and tissues need to be highly
dynamic intercellular junctions are very simple, frequently TJ are absent or present only a
primitive organization (Simionescu and Simionescu, 1991).
The functional state of interendothelial junctions can also be changed by the situation of
growth and activation of the cells. For instance, growth factors or inflammatory cytokines
can change the expression or phosphorylation state of junctional proteins (Dejana, 1996).
Besides their role in promoting homotypic cell adhesion, intercellular junctions can playa
role in cell-to-cell signal transduction. This likely occurs after the engagement and
clustering of the transmembrane adhesive proteins at junctions and the association with
cytoplasmic signalling molecules. The intracellular molecules which associate to AJ are
different from those of TJ and from those which link other junctional adhesion molecules,
suggesting that should exist a certain specificty of the signalling pathways.

ENDOTHELIAL INTERCELLULAR JUNCTIONS

Adherens junctions (AJ)

AJ are formed by transmembrane glycoproteins belonging to the cadherin family (Aberle

et al., 1996; Gumbiner, 1996; Takeichi, 1993). Cadherins are single chain transmembrane
polypeptides that promote calcium dependent homophilic cell-cell adhesion (figure 1).
The major and specific transmembrane component of endothelial AJ is VE-cadherin
(vascular endothelial-cadherin) (Dejana, 1996; Dejana et al., 1995). This protein belongs to
the type II cadherin subgroup (Aberle et al., 1996) and is encoded by a gene which resides
in a cluster with E-, P- and M-cadherin on the mouse chromosome 8 (Huber, Dalmon,
Engiles, Breviario, Gory, Buchberg, and Dejana, 1996). The cytoplasmic domain of VE-
cadherin is linked to cytoplasmic proteins called catenins, in particular b-catenin,
plakoglobin and p120, which belong to the so called "armadillo" family (Hulsken,
Behrens, and Birchmeier, 1994; Miller and Moon, 1996; Peifer, 1995). Binding of b-
catenin and plakoglobin is localized in the last 80 amino acids of the cytoplasmic tail of

189
 VE-cadherin while p120 associate to a different domain closer to the transmembrane
region (Lampugnani, Corada, Andriopoulou, Esser, Risau, and Dejana, in press).
Only b-catenin and plakoglobin, but not p120, can associate to a-catenin (Daniel, and
Reynolds, 1995). This last protein, which is homologous to vinculin, can in tum bind to the
actin cytoskeleton and promote junction stabilization. Since p 120 does not bind to a-
catenin and actin, cadherin association to it would lead to a more mobile and dynamic
junctional complex. In ras trasformed cells (Kinch, Clark, Der, and Burridge, 1995),
lacking stable intercellular junctions, most ofE-cadherin was bound to p120. Consistently
with this observation, we found that in recently confluent endothelial cells, where
intercellular junctions are still weak and immature, VE-cadherin is bound to p120 while
when junctions stabilize after longer times of confluency, p 120 dissociates from VE-
cadherin (Lampugnani et aI., 1995).
AJ are considered communication centers necessary for transducing signals between
neighbouring cells. Different receptor like phosphatases, including LAR (Kypta, Su, and
Reichardt, 1996), PTPlB (Balsamo, Leung, Ernst, Zanin, Hoffman, and Lilien, 1996) and
hPTPk (Fuchs, Muller, Lerch and Ullrich, 1996), have been found at AJ. Kinases such as
src, lyn and yes and receptor kinases such as EGF receptor and c-erb-B2 are concentrated
at intercellular contacts and in the case ofEGF receptor and c-erb-B2 directly associated to
the cadherin-catenin complex (Kanai, Ochiai, Shibata, Oyama, Ushijima, Akimoto, and
Hiroashi, 1995; Hoschuetzky, Aberle, and Kemler, 1994).
Beta catenin and plakoglobin, as other members of the armadillo family, can act as
signalling molecules. Beta catenin directly participates in the Wnt growth factor signalling
cascade. In particular the binding of Wnt to its receptor leads to inactivation of GSK-3
(glycogen synthase kinase-3) which is responsible for b-catenin phosphorylation and its
consequent rapid inactivation via an ubiquitin/proteasome pathway. This process is
regulated by APC (adenomatous polyposis coli) which can bind to GSK-3 and facilitate b-
catenin phosphorylation and further degradation (Nusse, 1997).
Wnt signalling therefore imply b-catenin stabilization in the cytoplasm. Cytoplasmic b-
catenin or its Drosophila homologue can then bind to at least three HMG (high mobility
group) transcription factors (LEF-l, Xtcf-3 and pangolin) and translocate to the nucleus
(Behrens, von Kries, Kuhl, Bruhn, Wedlich, Grosschedl, and Birchmaier, 1996; Brunner,
Peter, Schweizer, and Basler, 1997; Molenaar, van de Wetering, Oosterwegel, Peterson-
Maduro, Goodsave, Korinek, Roose, Destree, and Clevers, 1996). This process may
regulate expression of a series of homeobox genes (Nusse, 1997) which are likely
responsible for cell growth and differentiation responses.
Cytoplasmic free b-catenin can affect cell behaviour also through other pathways. For
instance, it can affect cellular cytoskeleton in a cadherin independent way. Cytoplasmic b-
catenin directly bind to the actin fasciculation protein, fascin (Tao, Edwards, Tubb, Wang,
Bryan, and McCrea, 1996) and overexpression of a stabilized form of b-catenin leads to its
codistribution with APC in tubulin containing cellular puncta (Barth, Pollack, Altschuler,
Mostov, and Nelson, 1997).
The role of AJ in the control of free cytoplasmic catenin signalling activities is still
obscure. AJ can indirectly modify cellular responses by linking b-catenin and decreasing
its free cytoplasmic pool. However in some conditions the overall levels of b~catenin are
high and therefore the association to cadherins does not significantly affect the amount of
cytoplasmic b-catenin (Orsulic and Peifer, 1996).
VE-cadherin is responsible of specific intracellular signalling in the endothelium. VE-
cadherin limits cell migration and participates in contact inhibition of cell growth in the
endothelium. This last effect requires VE-cadherin binding to catenins (Caveda, Martin-

190
Padura, Navarro, Breviario, Corada, Gulino, Lampugnani, and Dejana, 1996). In addition,
VE-cadherin gene null mutation leads to altered vascular morphogenesis in vitro (Vittet,
Buchou, Schweitzer, Dejana, and Huber, in press). In this system the lack of VE-cadherin
did not alter endothelial cell differentiation but abolished their capacity to organize
vascular like structures.
The mechanism through which VE-cadherin transfers intracellular signal is still unknown.
However the VE-cadherin/catenin complex is very dynamic and its composition rapidly
changes in relation to the functional state of the cells. At early stages of junction assembly
or when the cells are migrating, VE-cadherin is mostly linked to p120 and b-catenin and
heavily phosphorylated in tyrosine. Only a small amount of the complex is bound to the
actin cytoskeleton. When the junctions stabilize, VE-cadherin loses phosphorylation in
tyrosine, p120 and b-catenin tend to detach from the complex and are substituted by
plakoglobin (Lampugnani et al., in press). These changes are accompanied by a stronger
association of the VE-cadherin complex to actin.
The relationship of these changes with intracellular signalling is still a matter of
speculation. It might be that, when detached from AJ, p 120 and b-catenin become
available for signalling. Indeed, p120 is a src substratum and might participate in src
depedent signalling cascade (Daniel and Reynolds, 1995). It is also possible that actin
binding to AJ is responsible for intracellular signalling. Indeed, a-catenin is required for E-
cadherin mediated retardation of growth in a tumor cell line (Watabe, Nagafuchi, Tsukita,
and Takeichi, 1994).
AJ can also participate in signalling by clustering signalling molecules and growth factor
receptors and facilitating their reciprocal interaction through increased proximity effect.
Besides kinases and phosphatases also components of the ras signalling pathway and small
GTP binding proteins can concentrate at cell-to cell junctions. Recent data indicate that the
small GTP binding proteins, Rho and Rac participate in AJ organization and their
inhibition leads to AJ disassembly (Braga, Machaesky, Hall, and Hotchin, in press).

Tight junctions (TJ)

At difference with AJ that represent an ubiquitous type of organized structures at

interendothelial contacts, TJ or occludens junctions are only well developed in those
endothelia that exert a strict control of the exchanges between blood and tissues (typically
at the blood-brain barrier and in the large arteries, Anderson and van Itallie, 1995). At
electron microscopy, TJ present an apparent fusion of the outer leflets of the plasma
membrane of the two contiguous cells, suggesting that they litterally seal the intercellular
space giving rise to a true adhesive belt located towards the apical surface of the cell layer.
Morphologically, TJ in the endothelium present an organization very similar to that of
epithelial TJ. At variance with epithelia, and possibly due to the far more flat aspect of the
endothelium, endothelial TJ can present a less strictly apical localization and can be found
spatially intermingled with AJ (Anderson and van Itallie, 1995; Balda and Anderson,
1993). More recently, several molecular components of TJ as well as their reciprocal
relationships have been defined. Some of these constituents are common to endothelium
and epithelium. However subtle, but significant differences may exist between the two cell
types which still wait experimental analysis.
Occludin is the only transmembrane constituent of both epithelial and endothelial TJ
described up to now (Furuse, Hirase, Itoh, Nagafuchi, Yonemura, Tsukita, and Tsukita,
1993). This is a 65kD protein with four putative membrane spanning domains that locate
both the NH2 and COOH regions of the protein into the cytoplasm. Occludin might bind

191
homophilically an identical molecule present on an neighbouring cell, possibly through the
second extracellular domain (Wong and Gumbiner, 1997). Occludin staining was found
particularly intense in brain capillaries and it was weakly present in heart, muscle and
intestinal endothelium (Furuse et al., 1993).
It has been suggested that occludin needs association to cytoskeletal protein for the
localization at TJ (Furuse, Itoh, Hirase, Nagafuchi, Yonemura, Tsukita, and Tsukita, 1994).
One of the mediators of such an interaction could be the cytoplasmic protein, ZO-1
(zonula occludens-l, 220-195 kD, Anderson, Stevenson, Jesaitis, Goodenough, and
Mooseker, 1988; Stevenson, Siciliano, Mooseker, and Goodenough, 1986), which binds to
the 250 amino acid carboxyterminal region of occludin (Furuse et aI., 1994). Binding of
occludin to ZO-I is necessary, but apparently not sufficient for targetting occludin to TJ.
Some deletion mutants of occludin bind ZO-I, but are not concentrated at TJ (Furuse et al.,
1994). Therefore other factors may be required for the correct junctional distribution of
the molecule. ZO-I through its association to spectrin (Itoh, Yonemura, Nagafuchi,
Tsukita, and Tsukita, 1991) might connect occludin to the actin cytoskeleton and may also
bind actin directly using its COOH terminal domain (Fanning, Jameson, and Anderson,
1996).
ZO-1 binds occludin at its N-terminal region (Fanning et al., 1996). ZO-I in tum interacts
at one (PDZ2) of its three PDZ domains (see below) with ZO-2 (zonula occludens-2, 160
kD, Gumbiner, Lowenkopf, and Apatira, 1991) and both proteins form part of the
cytoplasmic undercoat of TJ. Both ZO-I and ZO-2 are members of the MAGUK family
(membrane-associated guanylate kinase), as they bear a domain homologous to guanylate
kinase, although devoided of enzymatic activity (Kim, 1995). The cytoplasmic proteins
belonging to the MAGUK family in general express organizing and targeting activity
towards cell membrane proteins (Kim, 1995). Beyond the guanylate kinase homologous
region they contain multiple copies of a 80 amino acid repeat, the PDZ (PSD-95,discs
large, ZO-I)/DHR (discs-large homology region) domain, which are involved in protein-
protein interactions.
While ZO-2 is exclusively found at TJ both in endothelia and epithelia, ZO-1 shows a far
less specific distribution. Indeed it is located at cell-cell contacts independently of the
presence of TJ and can be also found in cell types that never develop TJ such as fibroblasts
or cardiac muscle cells (Howarth, Hughes, and Stevenson, 1992; Itoh et aI., 1991).
Targeting of ZO-1 to the plasma membrane can be regulated by its binding to cytoplasmic
components of AJ, tipically catenins (a-, b- and plakoglobin, Rajasekaran, Hojo, Huima,
and Rodriguez-Boulan, 1996). In cells which do not develop TJ, ZO-1 is found associated
to AJ (ltoh et aI., 1991). However, when TJ form, ZO-1 is segregated with them. Therefore
ZO-1 may represent a cross talking element between AJ and TJ.
Interestingly, endothelial cells express a specific ZO-1 isotype, called ZO-1 a- (Balda and
Anderson, 1993). It derives from alternative RNA splicing, and it lacks a 80 amino acid
region (in the proline rich COOH half of the molecule). No alteration of the molecular
relationships of this isoform with other TJ components (for example ZO-2) has been
reported although it was suggested that ZO-la- characterizes more dynamic junctions
(Balda and Anderson, 1993).
Other two cytoplasmic components of TJ common to epithelia and endothelia are cingulin
(140-108 kD, Citi, Sabanay, Jakes, Geiger, and Kendrick-Jones, 1988) and 7H6 antigen
(175-155 kD, Zhong, Saitoh, Minase, Sawada, Enomoto, and Mori, 1993), the expression
of this last molecule being restricted to brain capillaries in vivo. Cingulin is located more
periferally to the plasma membrane than ZO-1 and ZO-2, and its molecular interaction
with the other constituents of TJ remains to be defined.

192
Epithelial cells express molecules at TJ which are absent in endothelial cells such as
symplekin (Keon, Schafer, Kuhn, Grund, and Franke, 1996), this suggests that besides the
strong similarities a certain degree of cell specificity exists in the TJ organization.
TJ have been classically indicated in the control of paracellular permeability and cell
polarity, functions that most of epithelia and many endothelia have to accomplish
(Anderson and Van Itallie, 1995).
Increasing evidence indicates that occludin directly contributes to paracellular barrier
function (McCarthy, Skare, Stankewich, Furuse, Tsukita, rogers, Lynch, and Schneeberger,
1996). Interestingly, when occludin is displaced from its junctional localization, the
cytoplasmic components of TJ (ZO-I, ZO-2 and cingulin) remain in place at the plasma
membrane even if both permeability and electrical resistance are severely affected (Wong
and Gumbiner, 1997).
As expected, cultured endothelial cells from the brain microvasculature express higher
levels of occludin in comparison to aortic endothelium and this parallels their higher
capacity to maintain effective barrier functions (Staddon, saitou, Furuse, Tsukita, and
Rubin, 1996).
As far as cytoplasmic components, the far more studied in terms of functional effects is
ZO-1. In cultured endothelial cells the amount of ZO-1 is upregulated by cells confluency
(Li and Poznan sky, 1990) and is down regulated in response to a vasoactive agent such as
histamine (Gardner, Lesher, Khin, Vu, Barber, and Brennan, 1996). Tyrosine
phosphorylation of ZO-1 correlates with increased paracellular permeability both in
epithelial and endothelial cells (Staddon, herrenknecht, Smales, and Rubin, 1995).
Induction of another cytoplasmic component of TJ, 7H6, with dbcAMP or retinoic acid,
enhances the barrier function of endothelial cells in culture (Satoh, Zhoug, Isomura,
Saitoh, Enomoto, Sawada, and Mori, 1996).
Similarly to AJ, there is indirect evidence that TJ can act as cell-to-cell signalling
structures. Non-classical signalling pathways are delineating for components of TJ. In
epithelial cells ZO-I shows a nuclear localization which is inversely related to the maturity
of cell-cell contacts (Gottardi, Arpin, Fanning, and Louvard, 1996). Symplekin was also
found to be concentrated in the nucleus (Keon et al., 1996).
Many of these possible novel signalling routes make use of the multi domain nature of the
MAGUK proteins (Kim, 1995).
ZO-I, as well ZO-2, contain the same PDZIDHR sequence present in hdgl, the human
homologue of Drosophila discs large (dig) protein, another MAGUK family member,
which are recognized by the tumor suppressor APC (Matsumine, Ogai, Senda, Okumura,
Satoh, Baeg, Kawahara, Kobayashi, Okada, Tohyoshima, and Akiyama, 1996). hdlg, using
the PDZIDHR sequences, binds protein 4.1 (at the 30 kD NH2 terminal domain) and ezrin
(at a similar conserved region) (Lue, Brandin, Chan, and Branton, 1996). At its SH3
domain ZO-I binds a serine protein kinase that can phosphorylate ZO-1 itself (Balda,
Anderson, and Matter, 1996). While the functional consequences of these molecular
association are at the present largely unknown, Drosophila dig, which is a component of
septate junctions, is involved in multiple functions from the control of epithelial
proliferation to maintenance of apicobasal polarity (Woods, Hough, Peel, Callain, and
Bryant, 1996). Interestingly another MAGUK member in C. elegans, Iyn 2, acts in the
pathway homologous to the EGFIEGFR signalling system (Kim, 1995). On the other hand
EGF can induce tyrosine phosphorylation of ZO-1 and ZO-2 in human epidermal cells
(Dehouck, Dehouck, Fruchart, and Cecchelli, 1994). If similar interactions can be
expressed by ZO-I and ZO-2 (and possibly other members of the MAGUK family) at
endothelial TJ is at the present totally ignored. The exploration of this aspect could be of

193
interest to understand the molecular basis of endothelial morphogenesis and its alteration
in pathological conditions.
In endothelial cells, the complexity of TJ organization may be influenced by the
microenvironment. The best example of this is the brain microvasculature where TJ
present an unique complexity. This characteristic is not predetermined but is induced by
the neural microenvironment and in particular by still undefined factors released by
astroglial cells (Dehouck et al., 1994). In brain tumors the neovasculature formed by
angiogenesis does not have the same barrier properties of the rest of brain vessels even if it
originates from them. This is likely due to the interaction with tumor derived factors
which strongly modify the original properties of the vessels.

COMPLEXUS ADHAERENS

In contrast to epithelial cells endothelial cells do not have classical desmosomes. However
they might express similar but possibly simplified desmosomal structures. Endothelial cells
synthesize desmoplakin which is a specific component of desmosomes. Desmoplakin
codistributes with VE-cadherin, plakoglobin and vimentin at confocal microscopy
(Valiron, Chevrier, Usson, Breviario, Job and Dejana, 1996). It is possible that the
association of these molecules (see figure 1) constitute a desmosomal-like structure (also
called complexus adhaerens, Schmelz and Franke, 1993).

OTHER CELL ADHESION MOLECULES

Endothelial cells express high amounts of N-cadherin but this molecule is preferentially
found diffuse on the cell membrane (figure 1) (Salomon, Ayalon, Patel-King, Hynes, and
Geiger, 1992). This suggests that it can playa role in the anchorage of endothelial cells to
other N-cadherin expressing cells such as smooth muscle cells, astrocytes or pericytes.
Adhesive molecules other tahn well-structured junctions have been found to be
concentrated at endothelial intercellular contacts. These include PECAM, S-Endo IlMuc
18, CD34, endoglin (for review see Dejana, 1996). The meaning of this redundacy of
adhesive structures at interendothelial contacts is not clear. It is difficult to believe that it is
simply due to the need of maintaining intercellular adhesion. It is more attractive to
consider that these structures have specific biological activities and possibly intracellular
signalling pathways.
PECAM, platelet-endothelial cell adhesion molecule, is certainly the best studied among
proteins at interendothelial contacts (for review see Newman, 1997). It can transfer and
receive signals possibly through the binding to a tyrosine phosphatase, SHP-2, which may
participate in the signalling cascade of different growth factors.

ROLE OF ENDOTHELIAL INTERCELLULAR JUNCTIONS IN ANGIOGENESIS

Adhesive molecules are required for the achorage of one endothelial cell to another during
vascular cell sprouting. VE-cadherin is expressed in the vessels of tumors with active
angiogenesis. In addition, in the mouse embryo, VE-cadherin (similarly to PECAM) is
found very early in mesodermal cells during their organization in blood islands (Breier,
Breviario, Caveda, Berthier, Schnurch, Gotsch, Vestweber, Risau, and Dejana, 1996). This

194
suggests that this molecule constitutes an early requirement in the acquisition of
endothelial characteristics and in blood vessel architecture. Consistent with this, is the
observation that the block of VE-cadherin expression in vitro, by homologous
recombination experiments, inhibits vascular tube formation.
In addition, the fact that VE-cadherin is ubiquitous and expressed essentially in all type of
vascular endothelium strongly suggests that this molecule is required for a normal
development and maintenance of the vessel structure and morphology.
On the other side, when junctions are fully organized they can limit endothelial cell
proliferation. DNA synthesis is strongly inhibited in confluent endothelium in respect to a
sparse culture. Endothelial cell junction components are therefore good candidates for
transferring migration and growth inhibitory signals. Previous work showed that protein
membrane extracts from confluent endothelial cells were able to inhibit the growth of
sparse endothelium but not of other cell types suggesting the existence of membrane
associated endothelial growth inhibitory protein(s) (Heimark and Schwartz, 1985).
Considering the specificity of expression of VE-cadherin in the endothelium and the fact
that other cadherins have been considered as tumor suppressors (Takeichi, 1993) we
investigated whether VE-cadherin could participate in contact inhibition of cell growth in
the endothelium.
We found that transfection of VE-cadherin in tumor cells could induce contact inhibition
of cell growth. In addition, recombinant fragments of VE-cadherin could inhibit the
growth of sparse endothelial cells (Caveda et al., 1996). The growth negative signal
induced by VE-cadherin required catenin association since a truncated mutant lacking the
intracellular domain responsible for catenin binding was without effect (Caveda et aI.,
1996). All these data are consistent with the possibility that cadherins and catenins could
directly contribute to inhibition of cell growth induced by cell density. This effect might be
of different intensity as a function of cell and cadherin type.As discussed above endothelial
cell junctions are very complex organelles formed by different transmembrane and
cytoplasmic components. Besides VE-cadherin it is likely that other molecules could
contribute to inhibition of cell proliferation induced by density.
Another aspect to consider in angiogenesis is that during the formation of vascular
sprouts, cells need to detach from the monolayer and invade the tissues. For this process to
occur, junctions have to be disrupted. It would be conceivable that proliferating agents
could act on cell-to-cell junctions to make them weaker. Interestingly, vascular endothelial
cell growth factor (VEGF), besides being a growth factor, increases vascular permeability,
suggesting that it can decrease junction strength (Connolly, Heuvelman, Nelson, Olander,
Eppley, Delfino, Siegel, Leimgruber, and Feder, 1989).
In other cell types, growth factors (EGF and scatter factor) or src overexpression (Hinck,
Nathke, Papkoff, and Nelson, 1994; Hoschuetzky et aI., 1994) increase tyrosine
phosphorylation of cadherins and, even more intensely, catenins. This is accompanied by a
decrease in cell-to-cell adhesion.
In endothelial cells, fibroblast growth factor decreases Ca++ dependent aggregation
possibly linked to modification of cadherin activity (Bavisotto, Schwartz, and Heinmark,
1990).
In general the observations discussed might open a new direction in the development of
agents able to modulate endothelial cell growth. For instance, biologically active VE-
cadherin fragments or VE-cadherin clustering agents might be useful tools in inhibiting the
formation of new blood vessels in pathological conditions.

195
CONCLUSIONS

Endothelial cell junctions are important organelles which are essential for the regulation of
vascular permeability, leukocyte extravasation and vascular proliferation. We are now
beginning to understand the molecular architecture of these structures. We still know little
about the pathologic consequences of alterations in the functional behaviour or synthesis
of endothelial cell junction proteins. It is possible that pathologies linked to increased
vascular proliferation like angiogenesis or hemangiomas, are associated with structural
alteration in endothelial junction organization. With the understanding of how these
structures are formed at a molecular level and how they function in organizing new
vascular structures, it will be possible to define new agents able to modulate angiogenesis
and/or vascular permeability to cells and plasma components.

ACKNOWLEDGEMENTS

This work was supported by grants from Associazione Italiana per la Ricerca sui Cancro,
European Community (project CT 960036, BMH4 CT960669, BMH4 CT 950875) and by
Human Frontier Science Program.

REFERENCES

Aberle, H., Schwartz, H., and Kemler, R., 1996, Cadherin-catenin complex: protein
interactions and their implications for cadherin functions, J. Cell. Biochem. 61: 514-523.

Anderson, J. M., Stevenson, B. R., Jesaitis, L.A., Goodenough, D.A., and Mooseker, M.S.,
1988, Characterization of ZO-I, a protein component of the tight junction from mouse
liver and Madin-darby canine kidney cells, J. Cell BioI. 106: 1141-1149.

Anderson, lM. and Van Itallie C.M., 1995, Tight junctions and the molecular basis for
regulation of paracellular permeability, Am. J. Physiol. 269: G467-G475.

Balda, M.S., Anderson, J.M., and Matter, K., 1996, The SH3 domain of the tight junction
protein ZO-1 binds to a srine protein kinase that phosphorylates a region C-terminal to this
domain, FEBS Lett. 399: 326-332.

Balda, M.S. and Anderson, lM., 1993, Two classes of tight junctions are revealed by ZO-1
isoforms, Am. J. Phys. 264: C918-C924.

Balsamo, J., Leung, T.e., Ernst, H., Zanin, M. K. B., Hoffman, S., Lilien, J., 1996,
Regulated binding of a PTP lB-like phosphatase to N-cadherin: control of cadherin-
mediated adhesion by dephosphorylation of b-catenin, J. Cell Bioi. 134: 801-813.

Barth, A. I. M, Pollack, A. L., Altschuler, Y., Mostov, K. E., and Nelson, W. J., 1997,
NH2-terminal deletion ofb-catenin results in stable colocalization of mutant b-catenin with
adenomatous polyposis coli protein and altered MDCK cell adhesion, J. Cell Bioi. 136:
693-706.

196
Bavisotto, L. M., Schwartz, S. M. , and Heinmark, R. L., 1990, Modulation of Ca2+-
dependent intercellular adhesion in bovine aortic and human umbilical vein endothelial
cells by heparin-binding growth factors, J Cell. Physiol. 143: 39-51.

Beherens, J., von Kries, 1. P., Kuhl, M., Bruhn, L., Wedlich, D., Grosschedl, R., and
Birchmeier, W., 1996, Functional interaction of b-catenin with the transcription factor
LEF-l, Nature 382: 638-642.

Braga, V. M. M., Machesky, L., Hall, A. and Hotchin, N. A., The small GTPases Rho and
Rac are required for the establishment of cadherin-dependent cell-cell contacts. JCell Bioi.
in press.

Breier, G., Breviario, F., Caveda, L., Berthier, R., Schnurch, H., Gotsch, u., Vestweber,
D., Risau, W., and Dejana, E., 1996, Molecular cloning and expression of murine vascular
encothelial cadherin in early stage development of cardiovascular system, Blood 87: 630-
641.

Brunner, E., Peter, 0., Schweizer, L., and Basler, K., 1997, Pangolin encodes a Lef-l
homologue that acts downstream of Armadillo to transduce the Wingless signal in
Drosophila, Nature 385: 829-833.

Caveda, L., Martin-Padura, I., Navarro, P., Breviario, F., Corada, M, Gulino, D., Lampugnani,
M. G., and Dejana, E., 1996, Inhibition of cultured cell growth by vascular endothelial cadherin
(cadherin-51VE-cadherin), J Clin. Invest. 98: 886-893.

Citi, S., Sabanay, H., Jakes, R., Geiger, B., and Kendrick-Jones, J., 1988, Cingulin, a new
peripheral component of tight junctions, Nature 333: 272-276.

Connolly, D. T., Heuvelman, D. M., Nelson, R., Olander, 1. V., Eppley, B. L., Delfino, 1. J.,
Siegel, N. R., Leimgruber, R. M., and Feder, J., 1989, Tumor vascular permeability factor
stimulates endothelial cell growth and angiogenesis, J Clin. Invest. 84: 1470-1478.

Daniel, 1. M., and Reynolds, A. B.,1995, The tyrosine kinase substrate p] 20 cas binds
directly to E-cadherin but not to the adenomatous polyposis coli protein or a-catenin, Mol.
Cell. Bioi. 15: 4819-4824.

Dehouck, B., Dehouck, M. P., Fruchart, 1. c., Cecchelli, R., 1994, Upregulation of the low
density lipoprotein receptor at the blood-brain barrier: intercommunictaions between brain
capillary endothelial cells and astrocytes, J Cell Bioi. 126: 465-473.

Dejana, E., Corada, M., Lampugnani, M. G., 1995, Endothelial cell-to-cell junctions,
FASEB J. 9: 910-918.

Dejana, E., 1996, Endothelial cell-cell adherensjunctions: implications in the control of vascular
permeability and angiogenesis, Clin. Invest. 98: 1949-1953

Fanning, A., Jameson, B. T., Anderson, 1. M., 1996, Molecular interactions among the
tight junction proteins ZO-I, ZO-2 and occludin, Mol. Bioi. Cell. 7: 607a

197
Fuchs, M., Muller, T., Lerch, M. M., Ullrich, A, 1996, Association of human protein-
tyrosine phosphatase k with members of the Armadillo family, J. BioI. Chem. 271: 16712-
16719.

Furuse, M., Hirase, T., Itoh, M., Nagafuchi, A, Yonemura, S., Tsukita, S., Tsukita, Sh.,
1993, Occludin: a novel integral membrane protein localizing at tight junctions, J. Cell
BioI. 123: 1777-1788.

Furuse, M., Itoh, M., Hirase, T., Nagafuchi, A., Yonemura, S., Tsukita, S., Tsukita Sh.,
1994, Direct association of occludin with ZO-1 and its possible involvement in the
localization of occludin at tight junctions, J. Cell BioI. 127: 1617-1626.

Gardner, T. W., Lesher, T., Khin, S., Vu, C., Barber, A 1., and Brennan, W. A, 1996,
Histamine reduces ZO-1 tight-junction protein expresssion in cultured retinal microvascular
endothelial cells, Biochem. J. 320: 717-721.

Gottardi, C. 1., Arpin, M., Fanning, AS., Louvard, D., 1996, The junction-associated
protein, zonula occludens-l, localizes to the nucleus before the maturation and during he
remodeling of cell-cell contacts, Proc. Natl. Acad Sci. USA, 93: 10779-10784.

Gumbiner, B., Lowep.kopf, T., and Apatira, D ., 1991, Identification of a 160-kDa

polypeptide that binds to the tight junction protein ZO-I, Proc. Natl. Acad. Sci. USA, 88:
3460-3464.

Gumbiner, B. M., 1996, Cell adhesion: the molecular basis of tissue architecture and
morphogenesis, Cell 84: 345-357.

Heimark, R. L. and Schwartz, S. M., 1985, The role of the membrane-membrane

interactions in the regulation of endothelial cell growth, J. Cell BioI. 100: 1934-1940.

Hinck, L., Nathke, I.S., Papkoff, 1. and Nelson, 1., 1994, Beta-catenin: a common target for
the regulation of cell adhesion by Wnt-l and Src signaling pathways, TIBS 19: 538-542.

Hoschuetzky, H., Aberle, H., and Kemler, R., 1994, Beta-catenin mediates the interaction
of the cadherin-catenin complex with epidermal growth factor receptor, J. Cell Bioi. 127:
1375-1380.

Howarth, AG., Hughes, M. R., and Stevenson, B. R., 1992, Detection of the tight junction-
associated protein ZO-1 in astrocytes and other nonepithelial cell types, Am. J. Physiol.
262: C461-C469.

Huber, P., Dalmon, 1., Engiles, 1., Breviario, F., Gory, S., Buchberg, A M., and Dejana,
E., 1996, Genomic structure and chromosomal mapping of the mouse VE-cadherin gene
(Cdh5), Genomics 32:21-28.

Hulsken, 1., Behrens, J., and Birchmeier, W., 1994, Tumor-suppressor gene products in
cell contacts: the cadherin-APC-armadillo connection, Curro Opin. Cell. Bioi. 6: 711-716.

198
Itoh, M., Yonemura, S., Nagafuchi, A., Tsukita, Sa., Tsukita, Sh., 1991, A 220-KD
undercoat-constitutive protein: its specific localization at cadherin-based cell-cell adhesin
sites. J Cell Bioi. 115: 1449-1462.

Kanai, H., Ochiai, A., Shibata, T., Oyama, S., Ushijima, S., Akimoto, S., Hiroashi, S.,
1995, c-erb-B2 gene product directly associates with b-catenin and plakoglobin, Biochem.
Biophys. Res. Commun. 208: 1067-1072.

Keon, B. A., Schafer, S., Kuhn, c., Grund, c., Franke, W. W., 1996, Symplekin , a novel
type of tight junction plaque protein, J Cell Bioi. 134: 1003-1018.

Kim, S. K., 1995, Tight junctions, membrane-associated guanylate kinases and cell
signalling. Curro Opin. Cell. Bioi. 7: 641-649.

Kinch, M. S., Clark, G. J., Der, C. J., Burridge, K., 1995, Tyrosine phosphorylation
regulates the adhesion of ras-transformed breast epithelia, J Cell BioI. 130: 461-471.

Kypta, R. M., Su, H., and Reichardt, L. F., 1996, Association between a transmembrane
protein tyrosine phosphatase and the cadherin-catenin complex, J Cell BioI. 134: 1519-
1529.

Lampugnani, M. G.,Corada, M., Andriopoulou, P., Esser, S., Risau, W., and Dejana, E.,
Cell confluence regulates tyrosine phosphorylation of adherens junction components in
endothelial cells, J Cell Sci., in press

Li, c., and Poznansky, M. J., 1990, Characterization of the ZO-l protein in endothelial and
other cell lines, J Cell Sci. 97: 231-237.

Lue, R. A., Brandin, E., Chan, E. P. , and Branton, D., 1996, Two independent domains of
hDlg are sufficient for subcellular targeting: the PDZ 1-2 conformational unit and an
alternatively spliced domain, J Cell BioI. 135: 1125-1137.

Matsumine, A., Ogai, A., Senda, T., Okumura, N., Satoh, K., Baeg, G. H., Kawahara, T.,
Kobayashi, S., Okada, M., Tohyoshima, K., and Akiyama, T., 1996, Binding of APC to the
human homolog of the drosophila discs large tumor suppressor protein, Science 272:
1020-1023.

McCarthy, K. M., Skare, I. B., Stankewich, M. c., Furuse, M., Tsukita, S., Rogers, R. A.,
Lynch, R. D., and Schneeberger, E. E., 1996, Occludin is a functional component of the
tight junction, J Cell Sci. 109: 2287-2298.

Miller, J. R., and Moon, R. T., 1996, Signal transduction through b-catenin and
specification of cell fate during embryogenesis, Gen. Dev. 10: 2527-2539.

Molenaar, M., van de Wetering, M., Oosterwegel, M., Peterson-Maduro, J., Godsave, S.,
Korinek, V., Roose, J., Destree, 0., and Clevers, H., 1996, XTcf-3 transcription factor
mediates b-catenin-induced axis formation in Xenopus embryos, Cell 86: 391-399.

Newman, P. J., 1997, The biology ofpECAM-l, J Clin. Invest. 99: 3-8.

199
Nusse, R., 1997, A versatile transcriptional effector of wingless signalling, Cell 89: 321-
323.

Orsulic, S., and Peifer, M., 1996, An in vivo structure-function study of Armadillo, the b-
catenin homologue, reveals both separate and overlapping regions of the protein required
for cell adhesion and for Wingless signaling, J. Cell Bioi. 134: 1283-1300.

Peifer, M., 1995, Cell adhesion and signal transduction: the Armadillo connection, Trends
Cell. BioI. 5: 224-229.

Rajasekaran, A.K ., Hojo, M., Huima, T., and Rodriguez-Boulan, E., 1996, Catenins and
zonula occludens-l form a complex during early stages in the assembly of tight junctions,
J. Cell Bioi. 132: 451-463.

Salomon, D., Ayalon, 0., Patel-King, L., Hynes, R. 0., and Geiger, B., 1992,
Extrajunctional distribution ofN-cadherin in cultured human endothelial cells, J. Cell Sci.
102:1-11.

Satoh, H., Zhoug, Y., Isomura, H., Saitoh, M., Enomoto, K., Sawada, M ., and Mori, M.,
1996, Localization of 7H6 tight junction-associated antigen along the cell border of
vascular endothelial cells correlates with paracellular barrier function against ions, large
molecules and cancer cells, Exp. Cell Res. 222: 269-274.

Schmelz, M., and Franke, W. W., 1993, Complexus adhaerens, a new group of
desmoplakin-containing junctions in endothelial cells: the syndesmos connecting
retothelial cells in lymph nodes, Eur. J. Cell Bioi. 61: 274-289.

Schnitzer, J. E., 1993, Update on the cellular and molecular basis of capillary permeability,
Cardiovasc. Med 3: 124-130.

Simionescu, N., and Simionescu, M., 1991, Endothelial transport macromolecules:

transcytosis and endocytosis, Cell Bioi. Rev. 25: 5-80.

Staddon, J. M., Herrenknecht, K., Smales, C., and Rubin, L. L., 1995, Evidence that
tyrosine phosphorylation may increase tight junction permeability, J. Cell Sci. 108: 609-
619.

Staddon, J. M., and Rubin L. L., 1996, Cell adhesion, cell junctions and the blood brain
barrier, Curro Opin. Neurobiol. 6: 622-627.

Staddon, J. M., Saitou, M., Furuse, M., Tsukita, Sh., and Rubin, L. L., 1996, Occludin in
endothelial cells, Mol. Bioi. Cell. 7: 605a

Stevenson, B. R ., Siciliano, J. D., Mooseker, M. S., and Goodenough, D. A., 1986,

Identification of ZO-I: a high molecular weight polypeptide associated with the tight
junction (zonula occludens) in a variety of epithelia, J. Cell Bioi. 103: 755-766.

200
Takeichi, M ., 1993, Cadherins in cancer: implications for invasion and metastasis, Curro
Opin. Cell BioI. 5: 806-811.

Tao, Y. S., Edwards, R. A., Tubb, B., Wang, S., Bryan, 1., Mc Crea, P. D.,1996, b-catenin
associates with the actin-bundling protein fascin in a noncadherin complex, J Cell BioI.
134: 1271-1281.

Valiron, 0. , Chevrier, V., Usson, Y., Breviario, F., Job, D., and Dejana, E., 1996,
Desmoplakin expression and organization at human umbilical vein endothelial cell-to-cell
junctions. J Cell Sci. 109: 2141-2149.

Van !tallie, C. M., Balda, M. S., Anderson, J. M ., 1995, Epidermal growth factor induces
tyrosine phosphorylation and reorganization of the tight junction protein ZO-I in A431
cells,J Cell Sci. 108: 1735-1742.

Vittet, D., Buchou, T., Schweitzer, A., Dejana, E., and Huber, P., Targeted null-mutation in the
VE-cadherin gene impairs the organization of vascular-like structures in embryoid bodies, Proc.
Natl. Acad. Sci. in press.

Watabe, M., Nagafuchi, A., Tsukita, S., and Takeichi, M., 1994, Induction of polarized
cell-cell association and retardation of growth by activation of the E-cadherin-catenin
adhesion system in a dispersed carcinoma line, J Cell Bioi. 127: 247-256.

Wong V, Gumbiner BM: A synthetic peptide corresponding to the extracellular domain of

occludin perturbs the tight junction permeability barrier. J Cell Bioi 1997, 136 399-409.

Woods DF, Hough C, Peel D, Callain G, Bryant PJ: Dig protein is required for junction
structure, cell polarity and proliferation control in Drosophila epithelia. J Cell Bioi 1996,
134: 1469-1482.

Zhong Y, Saitoh T, Minase T, Sawada N, Enomoto K, Mori M: Monoclonal antibody 7H6

reacts with a novel tight junction-associated protein distinct from ZO-I, cingulin, and ZO-
2. J Cell BioI 1993. 120: 477-483.

201
THE INTERACTION OF HUMAN NEUTROPHlLS WITII TYPE IV COLLAGEN
INVOLVES AN INHIBITORY SIGNAL TRANSDUCTION PATHW A Y

Jean Claude Monboisse+, Georges Bellon+, Roselyne Gamotel+,

Abdelilah Fawzi+, Nobuko Ohno *,
Nicholas A. Kefalides* and Jacques P. Borel+

+ Laboratory of Biochemistry, CNRS UPRESA 6021, University of

Reims, F-51095 Reims, France and
* Connective Tissue Research Institute, and Department of Medicine,
University of Pennsylvania and University City Science Center,
Philadelphia, PA 19104, USA.

1- INTRODUCTION

Type IV collagen is a heterotrimer composed of 3 a chains ; these molecules form

tetramers by overlap at the amino terminus (7S domain) which further aggregate by end-to-
end interaction involving the non-collagenous carboxyl termini (NCI domains). Collagen is
a major component of basement membranes. There are six a chains of type IV collagen
whose genes have been cloned. The most predominent heterotrimer is composed of two a-I
and one a-2 chains. It can be prepared by extraction of a mouse sarcoma (EHS tumor)
(Wisdom, Gunwar, Hudson, Noelken and Hudson, 1992). Molecular forms of the other
type IV collagen chains, a3(IV), a4(IV), a5(IV), a6(1V), have not been isolated (Hudson,
Reeders and Tryggwason, 1993 ; Zhou, Ding, Zhao and Reeders, 1994) and the clones for
the a3(IV), a4(IV), a5(IV) and a6(IV) have been reported (Zhou et al., 1994). There is
evidence that the latter are also distributed in several types of basement membranes
including the kidney glomerulus, the alveolus, the choroid plexus and lens capsule
(Ninomiya, Kagawa, Iyama, Naito, Kishiro, Seyer, Sugimoto, Oohashi and Sado, 1995).
Lens capsule collagen may be obtained in purified forms. It contains all the types of chains
except a6(IV).
Acute inflammation is a complex response to local injury or infection. In response to an
inflammatory stimulus originating within the interstitium, circulating polymorphonuclear
leukocytes (PMN) traverse the wall of the post-capillary venule and invade the
surrounding tissue. In view of the potential for damage to surrounding tissues by the
products of PMN activation and the close contact of PMN with the basement membrane,
as these cells extravasate through the capillary wall, it becomes important to understand the
mechanisms which regulate PMN activation during passage across the basement membrane.
In previous studies, we reported that type I collagen activates PMN (Monboisse, Bellon,
Dufer, Randoux and Borel, 1987; Gamotel, Monboisse, Randoux, Haye and Borel, 1995),

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 203
but, that bovine lens capsule type IV collagen prevents such activation (Monboisse,
Bellon, Perreau, Gamotel and Borel, 1991 ; Monboisse, Gamotel, Bellon, Ohno, Perreau,
Borel and Kefalides, 1994). It was shown that, after a preliminary incubation period of 30
min with type IV collagen, PMNs became insensitive to some of their usual stimuli such as
the peptide f-Met-Leu-Phe, the phorbol ester, phorbol myristate acetate (PMA) or type I
collagen. To determine in which domain of basement membrane collagen this inhibitory
activity resides, we tested bovine lens capsule (ALC) type IV collagen, its NC 1 domain, as
well as synthetic peptides arising from published sequences of the NCI domains of al(IV),
a2(IV), a3(IV), a4(IV) and a5(IV) chains (Zhou et aI, 1994). The initial selection of regions
for which synthetic peptides were synthesized was based on the secondary structure
characteristics of the NCI domains of the aI, a2, a3(IV) chains using the predictive
program of KefaIides et aI. (1993). In this paper, we demonstrate that a 19 amino acid
sequence contained in the NCI domain of the a3(IV) chain is responsible for the inhibitory
effect, a property not shared with the NCI domains of the other chains. EHS type IV
collagen (isolated from the mouse EHS tumor) failed to inhibit PMN activation. In addition,
we show that the whole type IV collagen molecule as well as the peptide sequence of the
a3(IV) chain increase the cytoplasmic levels of cAMP, a second messenger well known for
its ability to inhibit superoxide production and granule secretion by PMN. On the other
hand, we found that adenosine deaminase, an enzyme known to destroy adenosine, when
added to the incubation medium in the presence of the peptide 185-203, suppresses the
inhibiting effect. Accordingly, the inhibitory pathway comprises steps of adenosine
secretion by PMN followed by binding of adenosine on its specific A2 receptors an!1
activation of the adenylate cyclase pathway. The structure of the receptor for peptide 185-
203 and the mechanisms of transduction of the message are presently under investigation.

2 - MATERIALS AND METHODS

2.1. Preparation of PMN

Human blood was obtained by consent from healthy subjects. PMN were isolated
according to a previously published method (Monboisse, Bellon, Randoux, Dufer and
Borel, 1990).

2.2. Isolation of type IV collagen and its fractions

Type IV collagen was isolated and purified in its native state from bovine ALC by
extraction with 0.5 M acetic acid, followed by salt precipitation and ion exchange
chromatography as described previously (Brinker, Pegg, Howard and Kefalides, 1985 ;
Monboisse et ai., 1991). Limited treatment of type IV collagen with pepsin to remove the
NCI domain was performed as described previously (Monboisse et aI., 1991). EHS type
IV collagen was purchased from Sigma, St. Louis, Mo. Type I collagen was prepared by
acetic acid extraction from rat tail and 0.7M NaCl precipitation (Piez, Eigner and Lewis,
1963).

2.3. Preparation of synthetic peptides

Synthetic peptides corresponding to several specific sequences of the bovine and human
NCI domains of the a 1(IV), a2(IV), a3(IV), a4(1V) and a5(IV) chains were prepared
according to the method of Barany and Merrifield (1980). The selection of the specific
regions was based on the secondary structure of each chain and the hydropathy pattern.
The criteria for selection included beta-tum, hydrophilicity and aromatic amino acid
content and are based on the assumption that they represent exposed areas of the molecule
(KefaIides et aI., 1993).

204
 1.1
o PREINCUBATION

type IV collagen or synthetic peptldes

(INHIBITOR)
+ 0.1 ml of PMN suspension
30 min

~I of Oulbecco's solution containing ~ I Centrifugation

Washing of the cells
Suspension In 1.2 ml
of fresh Dulbecco's
(6 x 10' cells)

0.1 ml allquots
(5 x 10' cells)
+ 0.85 ml Dulbecco's solution
+ 0.10 mil mM cytochrome c
+ 0.10 ml stimulating agent
( 0.5 I'M f·Met·Leu·Phe )

INCUBATION
30 I .. tS min

Absorbance evaluation A = 550 nm

Exocytosls evaluation

Figure 1: Principle of preincubation experiments.

2.4. Measurement of superoxide anion (0 2-)

02- release was measured from the SOD-inhibitable reduction of ferricytochrome c
according to the method of English, Roloff and Luckens (1981) as described by Monboisse
et al. (1987). The increase in absorbance at 550 nm was taken as the index of 02-
concentration. The fact that the variation in absorbance depends on 02- was demonstrated
by its suppression elicited by added superoxide dismutase.

2.5. Measurement of the granule secretion

The measurement of the activities of elastase and 92 kDa collagenase (MMP-8) secreted
from the primary granules of PMN were performed using N-methoxysuccinyl-Ala-Ala-
Pro-Val-p-nitroanilide and biotinylated-type IV collagen as substrates during an incubation
period of 1 and 3 h respectively, according to the methods of Nakajima, Powers, Asthre
and Zimmerman (1979) and Wilkinson, Cohen and Schuman (1990). The enzymatic
activities released in the medium were expressed as percent of the corresponding total
enzymatic activity of the preparation. Secretions of elastase and MMP-8 were also
evaluated by ELISA.
Measurements of lactoferrin, secreted in the medium from the secondary granules, were
done by an ELISA technique (Bioxytech, France).

2.6. Incubation of PMN with the Nel domain or the peptide a3(JV) 185-203
The NCl domain or the peptide a3(IV) 185-203 were incubated for 30 min with PMNs. At
the end of this incubation period, the cells were washed and the 2nd period of PMN
incubation with PMA, f-Met-Leu-Phe or type I collagen was conducted as shown on figure
1.

2.7. Separation of the receptor for peptide 185-203

The antigens of the cell membranes from PMN and from HL60 cells were separated and
submitted to an affinity chromatography on a column made of the peptide 185-203 bound
to Sepharose. Elution was performed by a solution of 10 mM EDTA containing 0.6 M
NaCl. The protein fractions eluted by this method were submitted to techniques of SDS-
PAGE and Western blot with mono or polycIonal antibodies directed at the already known
membrane proteins. More details will be given in a forthcoming paper.

205
 '·Me'l.ev·Phe Induced 0 , production
nmolesf1lJ' COllis

23456

1 - Dulbecco's solution 4 - ALC type IV collagen (pepslnlzed)

2 - EHS type IV collagen 5 - NC1 domain from 0 1 and a2(IV) chains
3 - ALC type IV collagen 6 - NC1 domain from a3(1V) chain

Figure 2: Inhibition of 02- production

2.S. cAMP measurements

To 0.25 ml ofDulbecco's solution were added 0.1 ml of a PMN suspension (10 7 cells per
ml), and 0.05 ml of activating ligand solution containing either 20 Ilg of ALC type IV
collagen, 20 Ilg ofEHS type IV collagen or 20 Ilg of the bovine peptide a3(IV) 185-203 and
the mixture incubated for 0.5, 1, 2, 5, 15, 30, 45 and 60 min at 37°e. At the end of the
incubation period, 0.04 ml of a 10 M HCI04 solution was added, the reaction mixtures
were cooled at 4°C and centrifuged for 5 min at 4°C. Aliquots of 0.15 ml of the
supernatants were neutralized with 9 M NaOH and acetylated before analysis.
Cytoplasmic cAMP was then measured according to the method of Cailla, Racine-
Weisbuck and Delaage (1973) slightly modified which consists in a precipitation of cAMP
with a specific antibody.

2.9. Role of a Gi protein on the transduction of the a3(IV) message

PMN were incubated for 15 min with a 100 ng/ml solution of pertussis toxin in Dulbecco's
solution and then the preparation was incubated again with ALC type IV collagen. Then,
the concentration of cAMP in the cytoplasm and the effects of PMA, f-Met-Leu-Phe and
°
collagen I on 2 - release or on granule secretion were measured.

2.10. Intracellular calcium measurements

Briefly, suspensions (10 7 cells per ml) were added to Dulbecco's solution containing 1.3
mM calcium and 0.5 mM magnesium and incubated in the presence of 0.1 11M Fura 2-AM
at 37°C. The PMN were then rinsed in Dulbecco's solution and 2 ml aliquots of this
suspension (2 x 107 cells) were transferred in a spectrofluorometer cuvette and equilibrated
for 2 min at 37°C with gentle stirring. Volumes of 0.05 ml of activating agent solution
(either 0.5 !lM f-Met-Leu-Phe or 20 Ilg of type I collagen or 20 !lg of type IV collagen)

a 3 chain

7S domain helical domain NCI domain

N_

. . ." ' ' ' ' ' " . 203 (01 NCI:m:r C
185 200
CNYYSNSYSFWLASLNPER
(1622 ·1640 01 complete chain)

Figure 3: The sequence of a3(IV) chain that inhibits the PMN oxidative burst.

206
 Table I: Peptidic sequences analogous to peptide 185-203 .
per cent
inhibition

185 190 200

u3(IV) 185·203 : CNYYSNSYSFWLASLNPER 63%
human

185 190 200

u3(IV) 185·203 : CNYYSNSYSFWALSNDPKR 60%
bovine

u1(IV) 185·203 : CNYYANAYSFWLATIERSE 3%

human

u5(IV) 182·200 : CNYYA~SYSFWLATVDVSD 15%

human

u3(IV) 186·203 : NYYSNSYSFWLASLNPER 13%

u3(IV) 194·203 FWLASLNPER <0

u3(IV) 185·191 CNYYSNS

u3(IV) 185·203 : CNYYANSYSFWLASLNPER < 0

(189 S .. > A)

u3(IV) 185·203 CNYYSNA YSFWLASLNPER < 0

(191 S .. > A)

were added. Fluorescence was continuously monitored (I excitation 340 nm ; I emission 510
nm) (Grynkiewicz, Poenie and Tsien, 1985).

2.11. Use of adenosine deaminase

The enzyme adenosine deaminase was used as a tool to demonstrate the role of extracellular
adenosine in the process of inhibition of PMN by peptide 185-203. Adenosine deaminase
(Sigma, 1 U per ml) was added to the medium 10 min prior to the first incubation period,
and the effect of this addition on the formation of OT was studied as described in
paragraph 2.4.

3 - RESULTS AND DISCUSSION

3.1. Inhibition of PMN stimulation by collagens

When PMN were first preincubated with EHS tumor type IV collagen for 30 min and then
incubated with f-Met-Leu-Phe for 15 min, there was no significant change in the activating
effect of this ligand on 02- production. When PMN were first incubated with bovine ALC
type IV collagen, and then exposed to f-Met-Leu-Phe, there was a decrease of about 50 %
in 02- production (fig. 2). When bovine ALC type IV collagen had been previously treated
with pepsin, to remove the NC 1 domain, the inhibitory effect was abolished.

3.2. Effect of synthetic peptides on PMN activation

Since the inhibitory activity appeared to be residing in the NCI domain of the ALC type
IV collagen, we decided to test a series of synthetic peptides for their ability to inhibit the
activation of PMN by the three ligands, PMA, f-Met-Leu-Phe and type I collagen. Only
one of the synthetic peptides, the one comprising residues 185-203 of the NC 1 domain of
the a3(IV) chain, had the ability to decrease significantly the production of superoxide by
PMN when exposed to either PMA, f-Met-Leu-Phe or type I collagen. The degree of
inhibition observed with this peptide was comparable to that seen with the complete NCI
domain, composed of all the chains, as well as the NCI domain of the a3(lV) chain (fig. 3).
As shown on Table I, the peptides a3(1V) 185-203 from human or bovine origin exhibited

207
 Cytoplasmic cAMP
(pmoles 11O"cells)
2.0

ALC type IV collagen

1.0

o.3( IV) 185-203

~==:::==:::$..-.:::::o::!::=~ Dulbecco's
EHS type IV collagen
o L -_ _ _ _ _ _ _ _ _ _~_ _ _ _~.

o 20 40 60
Incubation time (min)

Figure 4 : Induction of cytoplasmic cAMP elevation by type IV collagen and the peptide
185-203

more than 60 % inhibition on the production of 02- by PMN. The peptides reproducing
the homologous regions of aI, a2, a5 chain did not exhibit any significant activity. We
synthesized several peptides partially reproducing the sequence of the active peptides and
found that several residues were necessary to the activity, the N-terminal cystein residues
and the triplet SNS-(residues 189-191) (Table I).

3.3. Search for a membrane receptor to peptide 185-203.

By the use of monoclonal or polyclonal antibodies added to PMN prior to the incubation
with the peptide 185-203, we found that the integrin aVb3 and the protein CD47 were
involved in the process of inhibition. More detail will be given in a forthcoming paper.
On the other hand, the use of classical methods of affinity chromatography of membrane
proteins followed by SDS-PAGE of the separated proteins and Western blot confirmed the
involvement of integrin aVb3 and CD47. Lactoferrin and carcinoembryonic antigens were
found to unspecifically bound onto the peptide.

3.4. Effect of type IV collagen on intracellular cAMP and calcium

To test the nature of the transduction system triggered by the peptide 185-203, we

cAMP '-Met-I.eu-Phe Induced

pmolesl10 6cells
02 production
nmoles/l0 S cells

o. 3(IV)185-203

Dulbecco's

° O~~2~5----~15~------~3~O~
Dul becco's n3(IV)18S-203
IncubatIon tIme (min)
oD D Preincubation without adenosine deamlnase
• • fill!! Preincubation with adenosine deamlnase (1 U I mil

Figure 5: Effect of adenosine deaminase

208
 adenylate cy=cl~ase=-_~
A2 recepto ___ cAMP

~Gipr~tem
adenosine
+
PKA
(Inhiblbng effects
on protein phosphorylabon)

1
Rat p67

integrin aV~3

~47
pepbde a3(IV) 185 - 203
(stmulus)

Figure 6: Tentative picture of the pathways involved in the inhibiting effect of peptide
185-203 on PMN.

measured cAMP after exposure of PMN to these proteins. An increase of cAMP was
elicited in PMN after exposure to intact bovine ALC type IV collagen as well as exposure
to the peptide 185-203 (fig. 4). On the other hand, EHS type IV collagen did not cause any
change in intracellular cAMP levels.
When PMN were preincubated with ALC type IV collagen in the presence of pertussis
toxin, the concentration of cytoplasmic cAMP did not increase under the effect of peptide
185-203. In addition, the preincubation of PMN with pertussis toxin suppressed the
inhibitory activity of ALC collagen IV on 02- production triggered either by PMA or by f-
Met-Leu-Phe or by collagen I.
When PMN were exposed to normal ALC type IV collagen or EHS type IV collagen the
intracellular calcium concentration remained unchanged throughout the experiment.

3.5. Effect of adenosine deaminase on the transduction system.

Based on reports from other laboratories that extracellular adenosine deaminase was
involved in situations where cAMP was elevated (Cronstein, Kramer, Weissmann and
Hirschhorn, 1983; Iannone, Wolberg and Zimmerman, 1989; Cronstein, 1994), we decided
to test the effect of the addition of adenosine-deaminase during the first incubation period
of PMN with the peptide 185-203. Adenosine-deaminase destroys adenosine. We found
that the liberation of 02- was no longer inhibited by the peptide 185-203 when this
enzyme was added to the medium (fig. 5).
Among the three types of membrane receptors for adenosine, only the type A2 triggers the
increase in cyclic AMP in the cytoplasm, suggesting that our peptide elicits the exocytosis
of ATP, ADP or adenosine outside the cells, followed by their conversion into adenosine
through the action of pericellular phosphatases and triggering of the A2 receptors ofPMN.
The activation of A2 receptors is followed by the formation of cAMP and the activation of
protein-kinase A. This enzyme, by phosphorylating some target intracellular proteins, yet
to be characterized, induces the inhibition of the complex system of membrane NADPH-
oxidase, responsible for the formation of 02-, according one of the pathways suggested on

209
figure 6. The same protein-kinase A, by phosphorylating proteins responsible of the
granule exocytosis, inhibits the secretion of lytic enzymes.

ACKNOWLEDGMENTS

This work was supported by grants from the CNRS UPRESA 6021 the University of
Reims (contract DRED) by grants AR-20553, HL-29492 and AR 07490 from the National
Institutes of Health and by the NATO collaborative Research Grants Program We thank Pr
B. Haye (Univ. Reims) for his helpfull assistance in measuring cAMP, and Elisabeth
Deschamps and Sandrine Etienne for typing the manuscript.

REFERENCES

Barany, G. and Merrifield, R.B., 1980, in The peptides, volume 2, (E. Gross and J.
Meinhofer, eds) part A, Academic Press, New York.
Brinker, lM., Pegg, M.T., Howard, P.S. and Kefalides, N.A., 1985, Immunochemical
characterization of type IV procollagen from anterior lens capsule, Collagen Rei. Res.,
5, 133-244.
Cailla, H.L., Racine-Weisbuck, M.S. and Delaage, M.A., 1973, Adenosine 3',5' cyclic
monophosphate assay at 10- 15 mole level, Anal. Biochem., 56, 394-407.
Cronstein, B.N., 1994 Adenosine, an endogenous anti-inflammatory agent. J. Appl.
Physiol., 76, 5-13 .
English, D., Roloff, lS. and Luckens, IN., 1981, Regulation of human
polymorphonuclear leukocyte superoxide release by cellular responses to
chemotactic peptides, J. Immunol. , 126, 165-171.
Garnotel, R., Monboisse, le., Randoux, A., Haye, B. and Borel, lP., 1995, The
binding of type I collagen to lymphocyte function-associated antigen (LF A) 1
integrin triggers the respiratory burst of human polymorphonuclear neutrophils, J.
Bio!. Chem ., 270, 27495-27503.
Grynkiewicz, G., Poenie, M. and Tsien, K.Y., 1985, A new generation of calcium
indicators with greatly improved fluorescence properties, J. BioI, Chem., 260, 3440-
3450.
Hudson, B.G., Reeders, S.T., and Tryggvason, K., 1993, Type IV collagen :
structure, gene organization and role in human disease, J. Bioi. Chem ., 268, 26033-
26036.
Iannone, M.A., Wolberg, G. and Zimmerman, T.A., 1989, Chemotactic peptide
induces cAMP elevation in human neutrophils by amplification of the adenylate
cyclase response to endogenously produced adenosine, J. Bioi. Chem., 264, 20177-
20180.
Kefalides, N.A., Ohno, N ., Wilson, C.B., Fillit, M ., Zabriski, l and Rosenbloom, J.,
1993, Identification of antigenic epitopes in type IV collagen by use of synthetic
peptides. Kidney Int., 43, 94-100.
Monboisse, lC., Bellon, G., Dufer, l, Randoux, A. and Borel, lP., 1987, Collagen
activates superoxide anion production by human polymorphonuclear neutrophils,
Biochem. 1., 246, 599-603.
Monboisse, lC., Bellon, G., Randoux, A., Dufer, l and Borel, lP., 1990, Activation
of human neutrophils by type I collagen. Requirement of two different sequences,
Biochem. 1., 270, 459-462.

210
Monboisse, J.e., Bellon, G., Perreau, c., GamoteI, R and Borel, J.P., 1991, Bovine
lens capsule basement membrane collagen exerts a negative priming on
polymorphonuclear neutrophils, FEBS Lett., 294, 129-132.
Monboisse, J.e., Gamotel, R, Bellon, G., Ohno, N., Perreau, e., Borel, J.P. and
Kefalides, N.A., 1994, The a 3 chain of type IV collagen prevents activation of
human polymorphonuclear leukocytes, J Bioi. Chem., 269, 25475-25482.
Nakajima, K., Powers, J.e., Ashe, B.M. and Zimmerman, M., 1979, Mapping the
extended substrate binding site of cathepsin G and human leukocyte elastase, 1. Bioi.
Chem., 254,4027-4031.
Ninomiya, Y., Kagawa, M., Iyama, K., Naito, I., Kishiro, Y, Seyer, lM., Sugimoto,
M., Oshashi, T. and Sado, Y, 1995, Differential expression of two basement
membrane collagen genes, COL4A6 and COL4A5, demonstrated by
immunofluorescence staining using peptide-specific monoclonal antibodies, J Cell.
BioI., 130,1219-1229.
Piez, K.A., Eigner, E.A. and Lewis, M.S., 1963, The chromatographic separation and
amino acid composition of the subunits of several collagens, Biochemistry, 2,58-66.
Wilkinson, M.I., Cohen, RL. and Shuman, M.A.S., 1990, A non radioactive assay
for type IV collagen degradation, Anal. Biochem., 185, 294-296.
Wisdom, B.I. Jr., Gunwar, S., Hudson, M.D., Noel ken, M.E. and Hudson, B.G.,
1992, Type N collagen of Engelbreth-Holm-Swarm tumor matrix : identification of
constituent chains, Connect. Tissue Res., 27, 225-234.
Zhou, J., Ding, M., Zhao, Z. and Reeders, S.T., 1994, Complete primary structure of
the sixth chain of human basement membrane collagen, a 6(N), J Bioi. Chem., 269,
13193-13199.

211
ANGIOGENIC POLYPEPTIDES IN BREAST CANCER: EXPRESSION OF
MRNA'S IN PRIMARY HUMAN TUMOURS, MCF-7 CELL TRANSFECTION AND
XENOGRAFT MODELS

Hua-Tang Zhang l , Rangana Choudhuri I, Prudence AE Scotti, Lyna

Zhang l ,4, MarinaZiche5, Lucia Morbidelli 5, Sandra Donnini l ,5 Rhys T.
1 I . 2 3
Jagger, Hock-Ye Chan, Kenneth SmIth , Sandra Peak, Margaret C. P.
Rees 4 , Adrian L. Harris2 and Roy Bicknell I *

IMolecular Angiogenesis Group

2Growth Factor Group, Imperial Cancer Research Fund, Institute of
Molecular Medicine, University of Oxford, John Radcliffe Hospital,
Headington, Oxford OX3 9DU
3Imperial Cancer Research Fund, Clare Hall Laboratories, Blanche Lane,
South Mimms, Herts
4Nuffield Department of Obstetrics and Gynaecology, University of
Oxford, John Radcliffe Hospital, Oxford OX3 9DU
5Department of Pharmacology, University of Florence, 50134 Florence,
Italy

I. SCREENING TO IDENTIFY KEY ANGIOGENIC POLYPEPTIDES IN

PRIMARY HUMAN BREAST CANCER.

Screening of 84 primary human breast carcinomas for the mRNA expression of seven
angiogenic polypeptides showed that the most commonly expressed and/or the most
highly expressed when compared to normal breast tissue were vascular endothelial growth
factor (VEGF) and platelet-derived endothelial cell growth factor/thymidine phosphorylase
(PDECGF/TPt The neurokines midkine (MK) and pleiotrophin were also fairly
commonly expressed in tumour but not normal tissue (unpublished data from this group
and ref. I),
Expression of the mRNA for midkine has been shown to correlate with prognosis in
invasive bladder cancers2 Expression of the other factors examined (acidic and basic FGF,

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 213
TGF -b 1 and placental growth factor) was rare. Only VEGF expression was found to have
prognostic significance - high VEGF mRNA correlating with poor patient survival l .

II. THE MCF-7 BREAST CARCINOMA MODEL.

Having identified potential key players in human breast tumour angiogenesis it was
imperative to demonstrate that these factors do indeed play a role in breast carcinoma
tumorigenesis. The initial approach taken to examine this was to prepare stable
transfectants of the MCF-7 human breast carcinoma cell line and then to characterise the
biological activity of these transfectants when xenografted into athymic mice.
The MCF-7 cells form tumours in athymic mice that are (i) slow growing, (ii) poorly
vascularised, (iii) oestrogen dependant, (iv) tamoxifen sensitive and (v) do not metastasise.
This permits the effect of gene transfection and expression on each of these properties to
be assessed. Also important is the fact that early passage (p - 54) MCF-7 cells express
few angiogenic factors. (In the absence of hypoxic or other stress only mouse midkine was
detected, unpublished data from this group). Over the past few years, transfectants of the
following angiogenic factors have been prepared and characterised in detail (i) VEGFI21,
(ii) VEGFI65, (iii) VEGFI89, (iv) PDECGFtrP, (v) PDECGF/TP + IgG secretion signal
(vi) MK, and (vii) PTN. There is little point in repeating the detailed in vitro
characterisation of these transfectants here, it is all published3,4. Instead we will
concentrate on the in vivo behaviour of the transfectants when compared to wild type
cells. In each case the cDNA of interest was cloned into the plasmid pCDNAlneo with
expression under control of the cytomegalovirus promoter. All transfectants have been
isolated as stable cell lines by clonal selection. In ~ did expression of the transfected
cDNA have an effect on in vitro growth.

(A) Vascular Endothelial Growth Factor

The first factor examined was VEGF121 5. Screening by RNase protection analysis had
shown VEGF 121 to be the predominant spliced transcript of VEGF in the primary human
breast tumours 1. Although the mRNA for VEGF121 was the predominant spliced
transcript, western analysis showed VEGF165 to be the predominant protein isoform6.
This is thought to be due to the 'trapping' of VEGF on the surface of the cells that secrete
it due to the heparin binding activity not shown by VEGF121 (see below). High level
expression of VEGF121 conferred a growth advantage on the MCF-7 cells in vivo. The
VEGF expressing tumours contained vascular hot spots as well as larger vessels, neither of
which were ever seen in tumours formed from wild type cells. The vascular hot spots
were very similar to those that have in primary human breast tumours been shown to
correlate with lymph node metastasis (reviewed in 7). However, no evidence of metastasis
was found in the transfectant bearing mice. There may be several reasons for this but the
hypothesis that we favour is based on the lack of lymphatics in the xenografted tumours
and mouse skin compared to a primary human tumour growing in the autologous site. We
propose that the high capillary vascular density leads to high hydrostatic pressure within
the tumour. This high pressure forces tumour clumps to break off into the remaining
functional lymphatics at the edge of the growing tumour. It is the lack oflymphatics in the

214
 a
25

Q,I

""
¢
(J
20

r./)
15
.~
en
=
Q,I

Q,I 10
OJ)
0
. Oil 5
C
~
0
0 5 10 15 20 25 30

Day
b looo
-0- COI'rffiOL . .--
._ VEO 121

....• ... ---...-

. .. .
VEOFI6S VEOF189

II) T/~ 1
00

.;:
...
0
)(

.....
'"E 600

.E
~
T L.· .: /'
:J
~/ Ii

. •••••••
"0 T-
T
•
T T/-
> 400
...
:J
0
~/ •
E
:J ,
f-<
lOO

O+------r-----.----~r-----._----._----_;
10 lO ]0 40 so 60 70

Days after injeclion

Figure 1. (a) Induction of angiogenesis in the rabbit cornea by mock transfected (empty
vector) human MCF -7 breast carcinoma cells and by MCF -7 transfectants
expressing VEGFI21, VEGF165 and VEGFI89. (b) Tumour growth by the
same cells when implanted subcutaneously into the flank of an athymic mouse.

skin, compared to the mammary fat pad that leads to the lack of metastasis in the animal
models.
Preparation of VEGF 165 and VEGF 189 transfectants such that expression of the growth
factor was comparable to that of the VEGF 121 transfectants permitted a direct comparison
of the angiogenic and tumorigenic activity of the isoforms 6 The results were somewhat
unexpected (Figure 1). VEGF 121 was found to be far more angiogenic in the rabbit corneal
assay and to be a much stronger promoter of tumorigenesis than were the other isoforms.
Indeed promotion of tumorigenesis following expression of VEGF 165 was very weak and

215
for VEGF189 there was no statistically significant effect. It is proposed that the reason
for this differential activity is one of 'bioavailability'. Thus, if we accept that each of the
three VEGF isoforms has equal endothelial cell chemotactic and migratory activity (a
strongly disputed tenet by some groups at the present time), then we postulate that
VEGF121 is secreted, diffuses, ligates VEGF receptors on existing endothelial cells and
induces neovascularisation. In contrast the heparin binding VEGF 165 and VEGF 189 are
secreted by the MCF -7 ceU and then bind to the MCF -7 cell surface - diffusing at a much
lower concentration and with a longer time course to the existing endothelium. These
observations constitute some of the first to report in vivo differences between the VEGF
isoforms and have important implications for the role of VEGF in tumorigenesis.

(B) Platelet-derived Endothelial Cell Growth Factorffhymidine Phosphorylase

The angiogenic enzyme isolated from platelets caUed called platelet derived endothelial cell
growth factor is now known to be the intracellular enzyme thymidine phosphorylase
(TP)8,9,1O. TP has long been known (it was identified many years before its isolation as
PDECGF) and has been thought to play a role in thymidine salvage. Thus the
deoxynucleotide synthetic pathway utilises ribose that subsequently has an hydroxyl
group removed, it does not use 2-deoxyribose. 2-deoxyribose enters a unique catabolic path
after phosphorylation, being cleaved by 2-deoxyribose-5-P-aldolase to gylceraldehyde-3-P
and acetaldehyde, that enter glycolysis and Krebs cycle respectively (as acetyICoA).
Death of tumour cells by ischaemic necrosis or apoptosis leads to release of substantial
amounts of DNA into the extracellular mileau. The DNA is cleaved to thymidine giving
rise to the well characterised thymidine fluxes that occur within tumours. Thymidine is
taken up by live ceUs and TP then phosphorolyses this to thymine and 2-deoxyribose-l-P.
The thymine enters nucleotide/deoxynucleotide synthesis, while 2-deoxyribose-1-P may be
dephosphorylated followed by loss of 2-deoxyribose from the cell and subsequent
stimulation of neovascularisation. Like VEGF, TP expression is induced by hypoxia II. The
foUowing are some observations that support an angiogenic role for TP in tumourigenesis:
1) TP is strongly angiogenic in several assays including the rat sponge model 12, 13
wound healing assays12, the CAM13 and in transfectant I xenograft models 12 ,13.
2) TP is strongly expressed in many solid tumour types (reviewed in 14).
3) Expression ofTP correlates with malignancy and blood flow in ovarian cancer 15 .
4) Expression ofTP correlates with patient prognosis in primary human breast cancer
and colorectal cancer 16,17,18,19.
The angiogenic activity ofTP is now known to be mediated by the small sugar molecule 2-
deoxyribose. 2-deoxyribose is strongly angiogenic on the CAM (Figure 2) and stimulates
the migration of capillary endothelial cells on a fibronectin matrix20; and unpublished work
by this group.
A detailed examination of the activity of several small sugar molecules for endothelial cell
migratory activity has revealed, amongst others, the following two significant observations
(R. Choudhuri, unpublished observations):
1) 2-Deoxy-D-ribose (the product of phosphorolysis of thymidine by TP) stimulates
endothelial cell migration but the enantiomer 2-deoxy-L-ribose is inactive (Figure 2).
2) Other sugars such as myo-inositol do not stimulate endothelial cell migration.
The mechanism by which 2-deoxy-D-ribose stimulates endothelial cell migration remains
an observation of considerable interest.

216
 <:
m
G>
::E 200

a;
0

"~
G>

Cl
::E

100
bFGF control D-isomer L-isomer

Figure 2. Endothelial cell migratory and angiogenic activity of 2-deoxyribose. (a)

Capillary endothelial (HMEC-l) cell migration in a Boyden chamber in
response to 2-deoxy-D-ribose, 2-deoxy-L-ribose and a bFGF control. (b)
Chorioallantoic membrane of the chick after a control (left) and 2-deoxy-D-
ribose (right) implant.

(C) Midkine and Pleiotrophin

Midkine (a 13-kDa polypeptide) and pleiotrophin (a 15-kDa polypeptide) are secreted

heparin-binding neurokines that share 50% sequence homology. Several studies have
suggested that MK and PTN may play a role in tumourigenesis, however, MK has not
previously shown to have endothelial cell growth-stimulating or angiogenic activity and the
assertion that PTN is angiogenic has previously rested on the extrapolation of its ability to
induce endothelial cell tube formation in vitro.
In our models4 , mouse MK- and human PTN-transfectatnts of MCF-7 breast carcinoma
cells gave rise to tumours that grew significantly faster than those from the mock-

217
 3000
.!i2 2500
a;
u
c
~ 2000
E
.c 1500
~
C\J

LL 1000
z
I--
0)
0.
500

0
e I--
::;;: I--
::;;:
I--
::;;:
eOJ I--
::;;: I--
::;;:
I--
::;;:
:::J
W ~ (J) 0. W ~ (J)
0. (J) (J)
(J) (J) 0 0
C'l
I-- C'l
I--
C'l
C'l
I-- Z
W
0
Z
0
Z
zW
C'l I-- C'l
W
C'l C'l <J)
<J)
W <J)
<J)

Figure 3. Endothelial specific expression of TNF-a after infection of 3T3 fibroblasts or

mouse skin endothelial cells with virus containing a fragment of the E-selectin
or KDR promoter coupled to the TNF-a gene.

transfected implants. The tumours formed by MK- and PTN-transfectants had a greater
vascular density than those formed by the control cell lines. The vasculature in the MK
and PTN tumours was largely uniform, similar to that in PDECGF/TP-overexpressing
tumours but unlike VEGF I2l overexpressing tumours which display vascular hot spots in a
uniform background. An unsual feature of the vessels in the MK- and PTN-overexpressing
tumours is that they appaered very elongated throughout the tumour, strikingly different
from the vessels in many other tumurs, which display a tortuous corkscrew morphology.
Vasculature in the MK and PTN tumours had a greater endothelial cell proliferative index
than that in the control tumours. We conclude that overexpression of either MK or PTN
leads to release of an endothelial growth-stimulating and angiogenic activity that enhances
the tumour growth and vascular density . The implicated angiogenic activity of MK and
PTN was confirmed in the rabbit corneal micropocket assay4 It is worthy of note that, in
vitro, the effect of MK and PTN on endothelial cell (HUVEC) proliferation was quite
small compared with that of aFGF or bFGF and that, in vivo, while VEGF-expressing cells
elicited a strong angiogenic response within 48 hours those expressing MK and PTN took
two weeks4.

Ill. VASCULAR TARGETING OF ANTI-CANCER GENE THERAPY

The other interest of our group is in vascular targeting. This has recently been extensively
reviewed 21 and so it will be discussed here only in outline. The basic idea is to target a
toxic gene or pro-drug activating gene to the existing tumour vasculature. The aim being to
bring about maximal endothelial cell destruction. It is now known that destruction of the
endothelial cell exposes the extracellular matrix surrounding the capillary that is highly
thrombogenic leading to clotting in all vessels feeding the tumour and ultimately ischaemic
necrosis of the tumour.
We have concentrated on the delivery of toxic genes using retroviral vectors with the gene
under control of endothelial specific promoters. Promoters of genes were chosen for which

218
the mRNA was known to be differentially upregulated in tumour endothelium, specifically,
E-selectin 22 and the VEGF receptor KDR23.
Using self-inactivating vectors we have shown highly specific (~100%) endothelial cell
expression of TNFa after infection of mouse microvascular endothelial (sEND) cells in
vitro (Figure 3).
Our current work is concentrating on attempts to retrovirally transduce microvascular
endothelium in vivo. While there have been several studies that have shown retroviral
transduction of endothelial cells in vivo after isolation of sections of large vessels and
administration of high titre virus within the isolated section of vessel, there have been no
reports to date of transduction of microvascular endothelium in vivo. The rat sponge
angiogenesis model appears eminently suitable for this purpose. bFGF treated sponge will
contain many proliferating endothelial cells and the virus should remain localised in the
sponge longer than it would if administrated into the artery of an isolated perfused tumour.
These experiments are in progress in collaboration with Dr Tai-Ping Fan, University of
Cambridge, UK.

ACKNOWLEDGEMENT

The authors thank Ms Jennifer Mackenzie for her excellent typing of the manuscript.

REFERENCES

1. RelfM, Lejeune S, Scott PAE, Fox S, Smith K, Leek R, Moghaddam A, Whitehouse

R, Bicknell R, Harris AL. Expression of the angiogenic factors vascular endothelial
cell growth factor, acidic and basic fibroblast growth factor, transforming growth
factor b-l, platelet-derived endothelial cell growth factor, placental growth factor, and
pleiotrophin in human primary breast cancer and its relation to angiogenesis. Cancer
Res 57: 963-969, 1997

2. O'Brien T, Cranston D, Fuggle S, Bicknell R, Harris AL. The angiogenic factor

midkine is expressed in bladder cancer and overexpression correlates with a poor
outcome in patients with invasive cancers. Cancer Res 56: 2515-2518, 1996

3. Bicknell R. Mechanistic insights into tumour angiogenesis. in "Tumour Angiogenesis"

Ed. Bicknell R, Lewis CE, Ferrara N, pp 19-28, Oxford University Press, 1997

4. Choudhuri R, Zhang RT., Donnini S, Ziche M, Bicknell R. An angiogenic role for

the neurokines midkine and pleiotrophin in tumourigenesis. Cancer Res 57: 1814-
1819,1997

5. Zhang HT, Craft P, Scott PAE, Ziche M, Weich HA, Harris AL, Bicknell R.
Enhancement of tumour growth and vascular density by transfection of vascular
endothelial cell growth factor into MCF-7 human breast carcinoma cells. J Natl
Cancer Inst 87: 213-219, 1995

6. Scott PAE. Vascular endothelial growth factor in breast carcinoma. D. Phil thesis,
University of Oxford, 1996

219
 7. Gasparini G. Prognostic and predictive value of intra-tumoral microvessel density in
human solid tumours. In "Tumour Angiogenesis" Ed. Bicknell R, Lewis CE Ferrara
N, pp 29-44. Oxford University Press, 1997

8. Moghaddam A and Bicknell R. Expression of platelet-derived endothelial cell growth

factor in E coli and confirmation of its thymidine phosphorylase activity.
Biochemistry, 31: 12141-12146, 1992

9. Usuki K, Saras J, Waltenberger J, Miyazono K, Pierce G, Thomason A, Heldin CH.

Platelet-derived endothelial cell growth factor has thymidine phosphorylase activity.
Biochem Biophys Res Commun 184: l311-l316, 1992

10. Finnis C, Dodsworth N, Pollitt CE, Carr G and Sleep D. Thymidine phosphorylase
activity of platelet-derived endothelial cell growth factor is responsible for
endothelial cell mitogenicity. EurJ Biochem 212: 201-210,1993

11. Griffiths L, Dachs GU, Bicknell R, Harris AL and Stratford II. The influence of
oxygen tension and pH on the expression of platelet-derived endothelial cell growth
factor / thymidine phosphorylase in human breast tumour cells grown in vitro and in
vivo. Cancer Res 57: 570-572, 1997

12. Moghaddam A, Zhang HT, Fan TPD, Hu DE, Lees VC, Turley H, Fox SB, Gatter
KC, Harris AL, Bicknell R. Thymidine phosphorylase is angiogenic and promotes
tumour growth. Proc Natl Acad Sci USA 92, 998-1002, 1995

l3. Ishikawa F, Miyazono K, Hellman U, Drexler H, Wernstedt C, Hagiwara K et al.

Identification of angiogenic activity and the cloning and expression of platelet-derived
endothelial cell growth factors. Nature 338: 557-562, 1989

14. Moghaddam A, Choudhuri R, Bicknell R. Thymidine phosphorylase/platelet-derived

endothelial cell growth factor: an angiogenic factor. in "Tumour Angiogenesis" Ed.
Bicknell R, Lewis CE, Ferrara N, pp 251-260, Oxford University Press, 1997

15. Reynolds K, Farzaneh F, Collins WP, Campbell S, Bourne TH, Lawton F,

Moghaddam A, Harris AL, Bicknell R. Association of ovarian malignancy with
expression of platelet-derived endothelial cell growth factor. J Natl Cancer Inst 86:
1234-1238, 1994

16. Fox SB, Westwood M, Moghaddam A, Comley M, Turley H, Whitehouse RM,

Bicknell R, Gatter KC, Harris AL' The angiogenic factor platelet-derived endothelial
cell growth factor / thymidine phosphorylase is up-regulated in breast cancer
epithelium. Brit J Cancer 73: 275-280, 1996

17. Takahashi Y, Bucana CD, Liu W, Yoneda J, Kitadai Y, Cleary KR et al. Platelet-
derived endothelial cell growth factor in human colon cancer angiogenesis: role of
infiltrating cells. J Natl Cancer Inst 88:1146-1151, 1996

220
18. Takebayashi Y, Akiyama S, Akiba S, Yamada K, Miyadera K, Sumizawa T et af.
Clinicopathologic and prognostic significance of an angiogenic factor, thymidine
phosphorylase, in human colorectal cancer. J Natl Cancer Inst 88: 1110-1117, 1996

19. Folkman J. What is the role of thymidine phosphorylase in tumour angiogenesis? J

Nat! Cancer Inst 88: 1091-1092, 1996

20. Haraguchi M, Kazutaka M, Uemura K, Sumizawa T, Furukawa T, Yamada K,

Akiyama SI. Angiogenic activity of enzymes. Nature 368: 198, 1994

21. Jagger RT, Bicknell R. Vascular targeting of anti-cancer gene therapy. in "Tumour
Angiogenesis" Ed: Bicknell R, Lewis CE, Ferrara N, pp 357-376, Oxford University
Press, 1997

22. Fox SB, Turner GDH, Gatter KC, Harris AL. The increased expression of adhesion
molecules ICAM-3 and E- and P-selectins on breast cancer endothelium. J Pathol
177: 369-376, 1995

23. Brown LF, Berse B, Jackman RW, Tognazzi K, Manseau EJ, Dvorak HF, Senger
DR. Increased expression of vascular permeability factor (vascular endothelial growth
factor) and its receptors in kidney and bladder carcinomas. Am J Pathol143: 1255-
1262, 1993

221
 Role of Thrombosis and
Fibrinolysis in Angiogenesis
THE ROLE OF mROMBIN AND ITS RECEPTORS IN ANGIOGENESIS.
PHYSIOLOGICAL AND PATHOLOGICAL APPLICATIONS.

Michael E. Maragoudakis, Nikos E. Tsopanoglou and Eva Pipili-Synetos

Department of Pharmacology, Medical School

University of Patras
265 00 Rio, Patras, Greece

INTRODUCTION

The original observation by Trousseau (1872), that there is associatIon between

thrombosis and cancer growth and metastasis, has been verified subsequently by many
investigators. Measuring circulating fibrinopeptides Rickles and Edwards (1983) have
shown inappropriately high intravascular coagulation and fibrinolysis in patients with
cancer. Others have shown that rumor cells interact with platelets, leukocytes and
endothelial cells as well as with thrombin - and plasmin - generating systems, all of which
influence clot formation (Sloan et al. 1986). In addition some rumors express the
transmembrane protein tissue factor, which when exposed to factor VII activates factor X
leading to generation of thrombin and fibrin formation (Tapparelli et aI., 1993).
More recently Zacharsky and co-workers (1995) have shown staining of malignant
tissue for thrombin in a variety of tumor types, by immunohistochemical techniques
using hirudin binding by antihirudin antibodies. Staining for thrombin was observed in
tumor cells of small cell lung carcinoma, renal cell carcinoma and malignant melanoma.
In addition tumor associated macrophages, but not tumor cells, were stained for
thrombin in adenosarcoma and squamous cell carcinoma of the lung.
In murine melanoma it was shown that B 16 melanoma cells treated with thrombin
caused 156-fold increase in the number of lung colonies in mice (Nierodzik et aI., 1992).
These animal experiments prompted clinical trials using anti-coagulants in human
cancers (Tapparelli et aI., 1993).
The aforementioned findings provide considerable insight for the cellular and
molecular mechanism of blood coagulation induced by cancer. They do not explain,
however, the mechanism by which thrombin promotes tumor growth and metastasis.
In addition to cancer in many other siruations such as wound healing (Carney et al,
1992), inflammation (Tapparelli et aI., 1993), placenta (Zacharsky et aI., 1995), thrombin
is present. Zacharsky et al. (1995) have shown by immunostaining the presence of
thrombin in rheumatoid synovial fluid, placenta, macrophages and the capillaries of
freshly incised skin, but not in either unperturbed skin or aged incisions.
The presence of thrombin in all the above clinical siruations, where angiogenesis IS

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 225
involved, raises the question whether thrombin is a contributing factor for activation of
angiogenesis.
Bleeding and blood coagulation is a primary event not only in solid tumors, but also
in diabetic retinopathy, within an atherosclerotic plaque, wounds etc. A very common
clinical observation is that when a blood clot forms in a large vessel this is often
recanalized by capillaries which grow into the thrombous. While plasma thrombin is
rapidly inactivated, the thrombin trapped within thrombous persists and is slowly
released. This phenomenon is often seen in portal vein thrombosis and also in tumors
of the kidney. Thrombi can be formed into the renal vein and up to the vena cava. The
tumor cells can be observed to induce long capillaries along the entire length of vena cava
almost to the entrance of the heart. These capillaries can be seen with angiography.
A plausible explanation for these phenomena is our findings that thrombin is a
potent promoter of angiogenesis (Tsopanoglou, Pipili-Synetos & Maragoudakis, 1993;
Haralabopoulos & Maragoudakis, 1996; Tsopanoglou & Maragoudakis, 1997).
In this report we summarise our results and discuss the mechanism by which
thrombin activates the angiogenic cascade. In addition we speculate on the possible
therapeutic implications of these findings.

MATERIALS AND METHODS

The angiogenic action of thrombin was studied using three experimental systems:
a) The chick chorioallantoic membrane (CAM) assay which is used by many
investigators as a convenient model for studying angiogenesis. In our laboratory we are
using collagenous protein biosynthesis as a biochemical index of angiogenesis
(Maragoudakis et aI., 1988). This method has been validated using a variety of promoters
and inhibitors of angiogenesis and comparing the results in parallel experiments with
direct counting of blood vessels (Harris-Hooker et a1., 1983) or by computed assisted
image analysis (Maragoudakis et a1., 1995). This method is sensitive, reproducible,
convenient and can accommodate the quantitation of angiogenesis in a large number of
samples.
b) The Matrigel plug assay was performed as described by Passaniti et aI (1992).
500 f.tl Matrigel at 4°C containing the test material or vehicle was injected subcutaneously
into C57 black female mice. After injection the Matrigel rapidly forms a plug, which is
left in the animal for 7 days. For each treatment 500 [!I of Matrigel was injected into 4
mice and the experiments repeated twice. At the end of the experiment the skin of the
mouse was easily pulled back to expose the Matrigel plug, which remained intact. After
photographing and noting quantitative differences, the plug was dissected out of the mouse
and fixed with 3.7% formaldehyde/phosphate buffered saline and embedded in paraffin.
It was then sectioned and stained with Masson Trichrome stain. Histological slices were
evaluated for the relative area of the cells that had infiltrated into Matrigel plug using
image analysis in a similar manner as the tube assay. This in vivo method of
angiogenesis has the advantage that no surgery is required for implantation and involves a
simple injection of Matrigel which is liquid at 4°C and solidifies rapidly at the body
temperature forming a stable Matrigel plug. In addition the Matrigel does not elicit any
immunological response because it is produced by the same strain of mice.
c) Cell Adhesion Assay. 48-Well plastic plates were coated with collagen type N
(lO[!glwell) oriaminin and then were incubated at 37°C for 60 minutes with Dulbecco's
PBS containing 2 mglrnI bovine serum albumin (BSA) fraction V to block nonspecific
binding. Before each experiment, HUVECs were incubated for three hours with serum-free
M199 medium containing 2mglml BSA. Subsequently cells were washed twice with
Dulbecco's PBS, detached from the culture flasks with EDTA in PBS, pelleted and
resuspended in serum-free M199. Various test substances, including thrombin and

226
PPACK-thrombin, TRAP, hirudin, were added directly to cells suspension and incubated
at 37°C, for time intervals ranging from 5 to 60 minutes. After preincubation the cells were
collected by centrifugation, washed twice and resuspended in serum-free M199. 40000
cells were applied on each well. After 60 minutes of incubation at 37°C the nonadherent
cells were aspirated. The plates with the adhered cells were rinsed twice with Dulbecco's
PBS before being fixed and stained with Diff-Quick. Adhered endothelial cells were
detached by the addition of 0.05% trypsin, 0.75mM EDT A solution and subsequently
were counted under microscope using glass haemacytometer.
The average area covered by adhered HUVECs was measured in triplicate wells with
a computerised digital image analyser (4.12 version of the MCID software from Brock
University, St. Catherine's, Ontario). Each experiment is repeated at least three times.
Results are given as mean % changes of control ±SE. Statistical analyses were performed
with Student's t-test for all experiments.
Endothelial Cell Culture. Human umbilical vein endothelial cells (HUVECs) were
obtained by established methods from freshly delivered umbilical cords from caesarean
births. Balanced salt solution without calcium and magnesium, but with 0.1% collagenase
was used to release endothelial cells from umbilical vein. Cells were cultured in a standard
endothelial cell growth medium consisting of Medium 199 supplemented with 20% foetal
bovine serum, 100 IU/ml penicillin/streptomycin, 50 !-tg/ml gentamycin, 2.5 !-tg/ml
Amphotericin b, 2 mmollL glutamine, 5 IU/ml heparin, and 100 !-tg/ml endothelial cells
growth supplement (ECGS). The latter is an extract of bovine brain, rich in acidic
fibroplast growth factor, that is essential for endothelial cells growth in vitro. Cells used for
experiments were from passages 3 through 6.

RESULTS

Effect of thrombin in the chick chorioallantoic membrane system.

It was shown in this system that both a- and y-thrombin causes a dose-dependent
promotion of angiogenesis. As with other actions of y-thrombin a 10-15 fold-higher
concentrations of y-thrombin than a-thrombin is required to obtain the maximum
angiogenic response (about 80% over that in controls) (Tsopanoglou et aI., 1993). A
maximum of angiogenic response was obtained with 1 IU (8.4 pmoles) of a-thrombin or
130 pmoles ofy-thrombin. As with many actions of thrombin, higher doses of 0.- and y-
thrombin cause a drop in the angiogenic action probably as a result of desensitisation of
the receptor.
Alpha-thrombin is a glycosylated trypsin-like serine protease generated form the
circulating plasma prothrombin. Upon autocatalytic and factor Xa cleavage the functional
two chain molecule of thrombin is generated consisting of 36 and 259 amino acid residues
for A and B chain respectively for human thrombin. The two chains are covalently
interconnected by disulphide bridge, the B chain having the catalytic site. In addition to
the catalytic domain a-thrombin possesses an exosite which is important for binding and
cleavage of fibrinogen to fibrin. This "fibrin-(ogen)-recognising exosite" or "anion binding
exosite" is essential for clotting activity.
Gamma-thrombin is the product of proteolytic or autolytic cleavage of a-thrombin.
In y-thrombin the "anion binding exosite" is missing, therefore, it can not form fibrin or
participate in the blood clotting cascade. However, y-thrombin, like a-thrombin is also a
potent cell activator. It can activate platelets, promote shape and permeability changes
of endothelial cells, cause Ca2+ release, prostocyclin biosynthesis, release of plasminogen
activator etc. The only important difference between a- and y-thrombin is that y-
thrombin has no blood clotting activity.
The fact that both 0.- and y-thrombin promote angiogenesis to the same extent lead

227
us to the conclusion that the action of thrombin on angiogenesis is independent of fibrin
formation (Tsopanoglou et aI., 1993).
The specificity of the effect of thrombin on angiogenesis is shown by the fact that
hirudin, which by itself has no effect, completely abolishes the angiogenic action of
thrombin, when hirudin is used in combination with thrombin. Hirudin binds both the
catalytic and the anion binding exosites, thus inactivating thrombin. Similarly heparin,
which binds to thrombin and accelerates its destruction by antithrombin III also cancels out
the angiogenic action of thrombin, when it was used in combination with thrombin
(Tsopanoglou et aI., 1993). P-PACK thrombin, is chemically modified thrombin in which
the catalytic site is inactivated but retains the anion binding exosite. This analog of
thrombin has no effect on angiogenesis in the CAM indicating that the anion binding
exosite alone is not sufficient to promote angiogenesis. When P-PACK thrombin,
however, was combined with either a- and y-thrombin it prevented the stimulatory
effect of thrombin presumably by interfering with the binding of thrombin via its catalytic
site. Apparently P-PACK thrombin binds to the receptor and prevents its activation by
thrombin.
These results suggest that the angiogenic action of thrombin is specific and that the
catalytic site is essential for this effect.
Recent studies have shown that the thrombin receptor, which is a member of the
seven transmembrane receptor family, has an extracellular amino terminal extension. This
extracellular extension possesses the cleavage site by thrombin. Thrombin proteolytically
cleaves a fragment of the receptor thus generating a new NH2-terminus of the receptor.
This new receptor peptide acts as a tethered receptor agonist by binding to an as yet
unidentified site of the receptor to affect cell activation. This site specific proteolysis of
the thrombin receptor provides all the necessary information for receptor activation. A
synthetic peptide TRAP, consisting of 14-amino acids representing the NH2-terminus of
the cloned receptor elicits most but not all the effects of thrombin (Garcia, 1992 & Garcia
et al., 1993). TRAP can mimic the effects of thrombin in cellular events such as secretion,
mitogenesis, second messenger generation etc. (Tapparelli et aI., 1993). However, not all
actions of thrombin can be seen by substituting TRAP for thrombin e.g. tyrosine kinase
activation, endothelial cell activation etc. It is of interest that TRAP also promotes
angiogenesis in the CAM system in a dose-dependent fashion . However, the maximum
obtained with 7 nmoles is about 50% above control and at higher concentrations there is a
decline in the angiogenesis promoting effect of TRAP.
These results with TRAP suggests that promotion of angiogenesis by thrombin is a
receptor mediated event, which is distinct form the clotting mechanism.

Effect of thrombin on endothelial cell infiltration into Matrigel plug in vivo.

In the Matrigel system in vivo in the absence of thrombin or other angiogenic
substances the Matrigel plugs were completely clear with no visible blood vessels going
into the Matrigel. Plugs containing thrombin (0.3, 1.0 and 3.0 IU/ml Matrigel) appeared
pink and a large number vessels could be seen leading into the Matrigel plug. Histological
evaluation of the sections by image analysis showed a 15-20 fold increase in the area of
cells that have infiltrated into the Matrigel plug containing thrombin (Haralabopoulos et
aI., 1997).

Endothelial cell attachment on collagen type IV and laminin. Inhibition by thrombin.

HUVECs suspended in serum-free M199 were applied on plastic wells coated with either
collagen IV or laminin. Cells adhered to these basement membrane proteins in a specific
manner within 60 minutes to an extent of 32% and 39% in collagen type IV or laminin
respectively. Only 3% of cells were attached on wells coated with BSA. Exposure of
endothelial cells for 15 minutes to thrombin (l IU/ml) or to endothelial cell growth medium

228
modulated the ability of HUVECs to adhere to collagen type IV and to laminin. Thrombin
decreased endothelial cell attachment to 14% and 17%, whereas endothelial cell growth
medium increased the percentage of adhered cells to 74% and 82% in collagen IV and
laminin respectively. The viability of applied endothelial cells is similar in all experimental
groups, as estimated by trypan blue staining. This effect of thrombin on endothelial cell
attachment to extracellular matrix proteins was dose-dependent (IC so at lill/ml) and was
evident after short exposure to thrombin. Even at the earliest time studied of 5 minutes, the
inhibition was about 80% of the maximum and did not increase further after 15 minutes
exposure to thrombin. This change of the phenotype of endothelial cells, manifested in
their limited ability for adhesion, was fully reversible. Cells which have been exposed to
thrombin for 15 min, then washed free of thrombin and subsequently incubated for further
15 min in fresh endothelial cell growth medium, had the same adhesion characteristic as the
non-thrombin treated cells.

Specificity of the effect of thrombin on cell adhesion. Hirudin which had no effect by
itself, completely abolished the action of thrombin on endothelial cell adhesion. The
catalytic site of thrombin was essential for this phenomenon, since PPACK-thrombin,
which has the catalytic site chemically inactivated, lacked this effect. Furthermore
PPACK-thrombin, when combined with thrombin, prevented the inhibitory effect of
thrombin in a dose-dependent manner.

The effect of TRAP on endothelial cell adhesion. The synthetic peptide TRAP
representing the 14 amino residues N-terminus of the activated receptor possesses agonist
activity similar to thrombin in many cell types (Grand, Tumell & Grabham, 1996). TRAP
also mimicked the effects of thrombin on endothelial cell adhesion either on collagen IV or
laminin. This further supports the notion that activation of thrombin receptor was
involved.

DISCUSSION

The studies summarised above suggest a novel action of thrombin on angiogenesis.

The molecular and cellular mechanisms involved are under investigation.
Thrombin in addition to its role in blood coagulation has many other actions on
various cell types, which may be related to its angiogenesis-promoting effect.
Particularly on endothelial cells, thrombin elicits a multitude of effects. For example,
thrombin affects endothelial cell barrier function (Garcia et aI., 1991) and increases
permeability, which is associated with changes in cell to cell junction organisation
(Rubiet et al., 1996). We found that thrombin reduces the ability of endothelial cells to
adhere to basement membrane components (Tsopanoglou & Maragoudakis, 1997). Both
these effects of thrombin are probably related to early events of the angiogenic
process, by allowing the activated endothelial cells to migrate through the locally lysed
basement membrane.
In addition, thrombin has been reported to have effects on endothelial cell
proliferation, migration, the expression of adhesion molecules, thrombomodulin, tissue
factor PDGF, bFGF, plasminogen activator/inhibitor, PGI 2 and EDRF (for review see
Kanthou & Benzakour, 1995). Many of these actions of thrombin are related to the
activation of angiogenesis.
Thrombin also activates gelatinase (Zucker et al., 1996) and secretion of
extracellular matrix proteins (papadimitriou et al., 1997) in cultured endothelial cells.
These events are relevant to the initial and final steps of the angiogenic cascade
respectively.
Thrombin has many cellular actions on smooth muscle cells, platelets, macrophages

229
and fibroblasts (Kanthou & Benzakour, 1995), which may play roles in activating
angiogenesis. On smooth muscle cells thrombin elicits growth and proliferation events
that are mediated, at least in part via, the induction of synthesis of autocrine growth
factors such as b-FGF. Similarly, fibroblasts express thrombin receptor and its
activation leads to mitogenic effects. Thrombin is a potent activator of platelets and
induces shape change, granule release and secretion, phospholipase C activation and Ca2 +
mobilisation. On monocytes and macrophages thrombin synergises and potentiates the
mitogenic effects of colony stimulating factor-I and induces recruitment and proliferation
of these cells. This action of thrombin on inflammatory cells may be important in
relation to angiogenesis in tumors, wound healing, atherosclerosis and inflammation.
It is of interest that the aforementioned effects of thrombin on the various cell
types, depend on a functional thrombin receptor and an intact catalytic site of
thrombin. Furthermore, many of these effects of thrombin can be mimicked by N-
terminal peptide agonists, such as TRAP, which corresponds to the activated thrombin
receptor "tethered" ligand (Grand et ai., 1996).
The question then arises, which one of this multitude of effects of thrombin on
endothelial cell and other cell types is the most important in the activation of
angiogenesis? It is likely that many of the above actions of thrombin play a role
depending on the particular tissue and the pathology involved. Temporal and spatial
factors may play a determining role. Thrombin may be a key factor in orchestrating
the events described in the microenviroment of the site of angiogenesis, where many
factors and cell types participate.
We conclude from our studies and those of others, that promotion of angiogenesis
by thrombin may explain not only the metastatic spread of cancer, but also the role of
thrombin in other conditions such as inflammation, diabetic retinopathy, wound healing,
atherosclerosis, etc. where angiogenesis is thought to playa role (Folkman, 1995).
This new role of thrombin as angiogenic factor opens many possibilities that need
to be worked out experimentally. The fact that y-thrombin, which can not form fibrin
and therefore can not cause blood clotting, promotes angiogenesis implies that there is a
possibility of developing specific inhibitors of the angiogenesis promoting effect of
thrombin, without effects on blood coagulation. Such inhibitors have potential
therapeutic applications in "angiogenic diseases" states (Folkman, 1995).
On the other hand angiogenesis is essential in wound healing and thrombin has been
shown to enhance incisional wound healing in rats (Carney et ai., 1990). The potential,
therefore, exists for therapeutic applications of non-thrombogenic analogs of thrombin
(e.g. y-thrombin) orthrombin-mimetic molecules that promote angiogenesis in situations
where stimulation of angiogenesis is desirable. This may be a novel therapeutic approach
for non-healing wounds and ulcers in diabetic or older patients and also myocardial
infarction.
The localisation of thrombin on tumor cells and tumor macrophages (Zacharski et
ai., 1995) may provide a tool for early detection of cancer by targeting radiolabelled
hirudin to malignant tumors, which are known to generate thrombin within the tumor
mass (e.g. melanomas). It is also conceivable that using hirudin as a homing vehicle for
drug targeting may be possible.

ACKNOWLEDGEMENT

This work was supported in part by grants from the Greek Ministry, the European
Communities "Mechanisms for the regulation of angiogenesis" and Biomed 2 BMH4-
CT96-0669 entitled "Molecular mechanisms of angiogenesis associated to cancer progress"
and a collaborative Research Grant from NATO #940677.
We thank Anna Marmara for typing the manuscript.

230
REFERENCES

1. Carney, D.H., Mann, R, Redin, S.D., Pemia, S.D., Berry l, Heggers lP., Hayward,
P.G., Robson, M.C., Christie l, Annable C., Fenton lW., Genn, K.C. Enhancement
of incisional wound healing and neovascularization in normal rats by thrombin
and synthetic thrombin receptor-activating peptides. l Clinical Investig. 1990; 89,
1599-1606.
2. Folkman, l Angiogenesis in cancer, vascular, rheumatoid and other disease. Nature
Med. 1995; 1,27-31.
3. Garcia, lG.N. Molecular mechanisms of thrombin-induced human and bovine
endothelial cell activation. l Lab. Clin. Med. 1992; 120, 513-519.
4. Garcia, J.G.N., Aschner, J., and Malik, A.B. Regulation of thrombin-induced
endothelial cell prostaglandin synthesis and barrier function. In: Berliner, LI., ed.
Thrombin: Structure and Function. New York: Plenum Publishing, 1991: 397-430.
5. Grand, RI.A., Tumell, A.S., and Grabham, P.W. Cellular consequences of thrombin-
receptor activation. Biochem. J., 1996; 313, 353-368.
6. Grant, D.S., Tashiro, K.I., Segui-Real, B., Yamada, Y., Martin, G.R., and Kleinman,
H.K. Two different laminin domains mediate the differentiation of human
endothelial cells into capillary-like structures in vivo. Cell 1989; 58, 933-943.
7. Haralabopoulos, G.c., Grant, D.S., Kleinman, H.K., Lelkes, P.I., Papaioannou, S.P.,
and Maragoudakis, M.E. Inhibitors of basement membrane collagen biosynthesis
prevent endothelial cell alignment in Matrigel in vitro and angiogenesis in vivo. Lab.
Invest. 1994; 71(4), 575-581.
8. Haralabopoulos, G., Grant, D.S., Kleinman, H.K., and Maragoudakis, ME
Thrombin promotes endothelial cell alignment in Matrigel in vitro and angiogenesis in
vivo. AmJ.Physiol., 1997; 273 (Cell Physiol. 42), C239-C245.
9. Harris-Hooker, S.A., Gajdusek, C.M., Wight, T.N., and Schwartz, S.M. Neovascular
response induced by cultured aortic endothelial cells. J. Cell Physiol. 1983; 114,
302-310.
10. Kanthou, c., and Benzakour, O. Cellular effects of thrombin and their signalling
pathways. Cell. Pharmacol., 1995; 2,293-302.
11. Maragoudakis, M.E., Tsopanoglou, N.E., Sakkoula, E., and Pipili-Synetos, E. On the
mechanism of promotion of angiogenesis by thrombin. FASEB l, 1995; 9, A587.
12. Nierodzik, M.L., Kajumo, F., and Karpatikiu, S. Effect of thrombin treatment of
tumor cells on adhesion of tumor cells to platelets in vitro and tumor metastasis in
vivo. Cancer Res., 1992; 52,3267-3272.
13. Papadimitriou, E., Manolopoulos, v., Hayman, G.T., Maragoudakis, M.E.,
Unsworth, B., Fenton, J.W. II, and Lelkes, P.I. Thrombin modulates vectorial
secretion of extracellular matrix proteins in cultured endothelial cells. Am. Physiol.
Soc. 1997; 272 (Cell Physiol. 41) CII12-C1122.
14. Passaniti, A., Taylor, R.M., Pili, R, Guo, Y., Lowrey, P.v., Haney, lA., Daniel,
RR., Grawd, D.S., and Martin G.R A simple quantitative method for assessing
angiogenesis and antiangiogenic agents using reconstituted basement membrane
heparin, fibroblast growth factor. Lab Invest. 1992; 67, 519-528.
15. Rabiet, M.l, Plantier, J.L., Rival, Y., Genoux, Y., Lampugnani, M.G., and Dejana, E.
Thrombin-induced increase in endothelial permeability is associated with changes in
cell-to-cell junction organization. Arterioscler. Thromb. Vasco BioL, 1996; 16, 488-
496.
16. Rickles, F.R., and Edwards, RL. Activation of blood coagulation In carcer:
Trousseau's syndrome revisited. Blood, 1983: 64,14-31.

231
17. Sloan, BJ., Rozhin, J., Johnson K., Taylor, H., Crissman, J.D., and Honn, KY
Cathepsin B: Association with plasma membrane in metastatic tumors . Proc. Natl.
Acad. Sci. (USA) 1986,83, 2483-2487.
18. Tapparelli, C., Mettemich, R., Ehrhardt, c., and Cook, N.S. Synthetic low
molecular weight thrombin inhibitors: molecular design and pharmacological
profile. TIPS, 1993, 14,366-376.
19. Tapparelli, C., Mettemich, R., and Cook, N.S. Structure and function of thrombin
receptors. TIPS, 1993, 14,426-428.
20. Trousseau, A. Phlegmasia alba delens in Trouseau A. Lectures in clinical medicine,
delivered in Hotel-Dieu, 1872; Paris, London, New Sudenham Society 281-295 .
21. Tsopanoglou, N .E., and Maragoudakis, M.E. On the mechanism of thrombin-induced
angiogenesis: Inhibition of attachment of endothelial cells on basement membrane
components. Angiogenesis J., 1997 (in press)
22. Tsopanoglou, N .E., Pipili-Synetos, E., and Maragoudakis, M.E. Thrombin promotes
angiogenesis by a mechanism independent of fibrin formatiom. Am. J. Physiol., 1993;
264, C1302-C1307.
23. Zacharsky, L-R., Memoli, V.A., Morain, W.D., Schlaeppi, J.M. & Rousseau, S.M.
Cellular localization of enzymatically active thrombin in intact human tissues by
hirudin binding. Thrombosis & Haemostasis, 1995; 73, 793-797.
24. Zucker, S., Conner, C., DiMassmo, B.I., et al. Thrombin induces the activation of
progelatinase A in vascular endothelial cells. Physiological regulation of angiogenesis.
J. BioI. Chern. 1996; 270, 23730-23738.

232
FIBRIN DEGRADATIVE PATHWAYS AND ANGIOGENESIS IN HEALING,
ATHEROSCLEROSIS AND TUMOUR INVASION

W Douglas Thompson, Chris M Stirk *, Andrew J Keating *, Allyson

Reid, Elspeth B Smith, and W T Melvin *

Departments of Pathology, and Molecular and Cell Biology *,

University of Aberdeen Medical School, Aberdeen Royal Infirmary,
Aberdeen AB25 2ZD, UK

INTRODUCTION

Fibrin fragment E has been shown by us to be the component of fibrin degradation

products (FDP) active in stimulating angiogenesis (1, 2). It is abundant in the healing
wound (3), proliferative atherosclerotic plaques (4) and breast cancer (5). It is both a
thrombin and plasmin dependant product, and stimulates cell proliferation in culture in
serum rich medium in contrast with thrombin itself which requires serum free medium (6).
We have attempted to study and locate the active site on this 35-40 kD molecule by
various blocking antibody approaches. Fibrinogen is a highly conserved molecule, and this
results in species similarity problems that confound conventional approaches . The mouse
will not respond with antibodies to human fragment E, although it will to fragment D, the
other major constituent of FDP. The rabbit and the rat will produce polyclonal antibodies
against E that block the angiogenic effect ofFDP.

We have produced a range of rabbit and rat antisera against several synthetic peptides
analogous to sites on human fragment E that we guessed might be related to the active site.
This approach has not worked out, but has yielded some interesting antibodies. Two that
immunostain fibrin only within macrophages in tissue sections have given novel insights
into human atherogenesis (7).

An entirely different approach has been to employ our two assets, the polyclonal rabbit
and rat anti E blocking antisera, with a Fd phage combinatorial library of random epitope
display to select clones binding and common only to both antisera. We describe the
rationale of the methodology, and the current stage of this approach to locate the active site
on fibrin E.

Angiogenesis: Models. Modulators. and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 233
METHODS

The Phage Library

Rat and rabbit polyclonal antibodies were raised against purified human fibrin fragment E
as previously (2). All secondary antibodies were purchased from commercial sources
(DAKO, Glostrup, Denmark, and Diagnostica Stago, Asnieres, France). A sample of the 2
x 108 clone hexapeptide gene VITI library as described by Scott & Smith (1990) (8), was
obtained as a phage suspension from George Smith (University of Missouri Columbia,
USA). 20 ml of this phage suspension was infected into mid-log K91 cells, which were
IPTG induced when the OD reached approximately 0.9 resulting in the formation of 6.6 x
109 tetracycline-resistant transductants. These were amplified by growth in one litre of 2 x
TY medium containing 20 mg tetracycline/ ml for 24 hours at 370 C with shaking.
Before precipitation of the phage, the cultures were first spun for 10 minutes at 8,000
RPM at 4 0 C to remove bacteria. The supernatant was kept, and PEG 600020% (w/v), 0.5
molar NaCI, added resulting in a final PEG concentration of ~4% (w/v). The cultures were
then cooled on ice for 1 hour and centrifuged for 20 minutes at 10,000 RPM at 4 0 C. The
phage were re -suspended in lOml ofTris buffered saline.

Selection of Phage

Selection of phage from the library was carried out using a procedure based on the
biopanning method of Parmley & Smith (1988) (9), as summarised in Figure 1. 1 ml of each
poly clonal antibody in PBS at dilutions from 1: 200 to 1: 1,000 was added to a 60 mm
petri dish (Nunc) and incubated in an orbital incubator at room temperature overnight to
allow antibody adhesion. The next day 5mls of blocking solution (PBS, 0.5% Tween, 5 %
Skimmed Milk), was added to each dish, in order to block any sites not adhered to by the
polyclonals. The dishes were then left to incubate in an orbital incubator for 1 hour at
room temperature, and then washed 5 times in PBST (PBS, 0.5% Tween).

100 ml of the gene VITI library was added to each dish, and the dishes incubated at room
temperature for 1 hour, in an orbital incubator. The unattached phage were then poured
away and the dishes washed 5 times in PBS-Tween. Phage which remained bound to the
immobilised polyclona1s were recovered from the antibodies by the addition of 800 ml 0.1
M HCL (pH. 2.2 with glycine), and allowed to incubate at room temperature for 15
minutes. 48 ml of unbuffered 2 M Tris base was then added to each dish to neutralise the
acid.

For round 1 biopanning, the contents of each dish were added to 3m1 cultures of mid-log E
coli K91, and allowed to incubate at room temperature for 10 minutes. 12 ml of LB
medium (lOg Tryptone, 5g Yeast extract, 109 NaCI in 1 litre H20, pH 7.5) was then added
to each culture. Tetracycline was also added to a concentration of 0.2 mg/ml, and the
cultures incubated at 37 0 C (225 RPM) for 40 minutes. Additional tetracycline was then
added, to a final concentration of 20 mg/ml and the cultures allowed to grow for 16-20
hours at 37 0 C (225 RPM). Phage input and output titres were determined by plating and
incubation on LB agar dishes.

234
A second round of amplification was carned out as for the first, except that 100 ml of
phage recovered from the first rounds of biopanning was used against the blocked
polyclonals. A third round ofbiopanning was also carried out except that 100 ml of phage
recovered from the second round ofbiopanning was used against the blocked polyclonals.

During rounds 2 & 3 negative controls, consisting of K91 cells, with no added phage, were
used to demonstrate that the K91 stocks had not acquired tetracycline resistance by
another means i.e. plasmid transfer. A fourth round of amplification was canied out as for
the third except that 100 ml of phage recovered from the second round of biopanning was
used against polyclonals from a different species, i.e. phage recovered after three rounds
of biopanning against the rabbit polyclonals, was used against the rat polyclonals, and
phage recovered after three rounds of biopanning against the rat polyclonals was used
against the rabbit polyclonals. This was to demonstrate the homogeneity of different
polyclonals raised in two different species against the same antigen and to reject non shared
target epitopes.

Testing of Reactive Clones

Reactive clones were identified from the output of round 4 with an ELISA using a 1: 200
dilution of rat anti fibrin E polyclonal for coating one set of wells, and the second coated
with a 1: 200 dilution of rabbit anti fibrin E polyclonaI. Samples of the selected phage
clones were each run on PAGE and immunoblotted with the existing anti E antisera that

ELISA choice
Run phage on gel
Anti E immunoblot
FINAL
OUTPUl & I I.
.",
Selected
phllge are
Implified
Recycl e all the
way round once
more

Immunise rabbits
~
1
::'etyCle r o u r l
procedure .
Testing by
ELISA
ImmunoblOl Fd
Blocking IblUty ph ge

PHAGE
EPITOPE
LIBRARY Rat polyclonal anti Frag E
Rabbit polyclonal Ab

First stage of affinity selection

"Biopanning""
I Second stage I

~
selected
phage are
~ ~ ,mplified
J
Fig 1 The sequence of steps in the selection process, summarised diagramatically,
illustrating the use of a phage epitope display library to determine the active
site for angiogenic stimulation on the fibrin fragment E molecule employing two
blocking antibodies raised against fibrin E in the rabbit and the rat.

235
 1 2 3 4 S 6 7 8 9 10

Fig 2 An immunoblot of SDS PAGE of phage numbers 1 to 10 detected with rat anti
phage 1-10 combined. At least 8 out of the 10 phage clones contain coat
proteins reactive with the rat anti 1-10.

kD
208
144
87

44
32.7
17
7.7

2 3 4 5 6 7

Fig3 An immunoblot of SDS PAGE of human fibrin degradation products. Each

strip has been exposed to a different antibody.

1 rat anti E 5 rabbit anti E

2 rat anti 1-20 6 rabbit anti 1-20
3 rat anti 21-40 7 rabbit anti 21-40
4 rat anti 41-60

There is a difference in pattern of human FDP band staining between the two
species, reflected by both anti fragment E antibodies and the antibodies against
phage clones selected by both these initial antibodies. The rabbit antisera
immunostain more strongly bands between 45-50 kD which represent plasmin
derived fragments E" ~ and E3 . Rat antibodies immunolabel more strongly
higher molecular weight combinations of DDE, DED, and DE. (both E, and D
and E combinations are angiogenic).

236
were the basis of the selection process. The 60 selected reactive clones were combined into
3 groups, 1-20,21-40, and 41-60, and samples mixed with Freund's complete adjuvant and
used to immunise rats and rabbits. The resultant new antisera were tested initially by
immunoblotting of PAGE of selected phage clone proteins and also human FDP. Then the
rabbit antisera were tested for ability to block the stimulatory activity of human FDP on
the chick CAM model. The assay used was based on quantitative measurement of DNA
synthesis in the CAM after exposure to control and test substances 18 h after application
in liquid form to the whole "dropped" area of each CAM (1, 2). This assay is a parameter
of quantitation of changes in CAM vascularity (10). Anti phage antisera, raised by
relatively short term immunisation of rabbits, were used at a 112 dilution and admixed with
FDP used at a concentration of approximately 1.95 mglml diluted before use to 1110.
Controls included buffer only and antiserum only groups of CAMs. After filter
sterilisation, 0.3 ml was added onto each CAM dropped surface.

RESULTS

Use of the Phage Library

The gene VITI phage display peptide library was relatively simple to use. Extra care needs
to be taken at the stage ofbiopanning, when the phage are been removed from the plate to
infect the log K91 cells. The original protocol allowed five minutes for this infection
procedure, but ten should be allowed to ensure successful transformation of the bacteria. If
this stage is unsuccessful then the phage will not infect the bacteria, and the bacteria will
not have acquired resistance to the antibiotic, and there will be no phage output. It is vitally
important to ensure that the stock of bacteria have not acquired resistance to the antibiotic
from a source other than the phage, especially when in an environment where other
workers are using antibiotics. A simple control experiment at each stage (i.e. K91s with no
phage in media with tetracycline), will determine if the stock of bacteria is still suitable to
use. Another stage which requires care is the phage precipitation step. At this stage it is
important to cool the supernatant before centrifugation. This simply ensures a higher yield
of phage as increasing numbers are precipitated at lower temperatures.

Testing of Reactive Clones

The sequence of checks to ensure that the selected clones are genuinely reactive with the 2
antisera that formed the basis of that selection is summarised in Figure 1. Two examples of
these stages are shown. First, the demonstration of binding of rat anti phage 1-10 combined
to individual phage clone proteins is shown in Figure 2. Second, the binding of to human
fragment E sized bands by several of the new antisera raised is seen in another immunoblot
(Fig 3).

Demonstration of Blocking of FDP Angiogenic Activity

Figure 4 illustrates one experiment where rabbit anti phage 1-20 does not appear to block
the stimulatory effect of FDP after admixture, but rabbit anti phage 21-40 and 41-60 do
inhibit the simulation. Repeat experiments have now been completed a further 2 times for
the latter two antibodies with similar inhibition. A further experiment with anti phage 1-20
has shown some inhibition and this antibody is still under investigation. Antibody alone

237
and buffer alone controls have no effect. Rat antibodies have not yet been tested. The first
3 antibodies raised against individual clones, clones 42, 43, and 44, have been found to be
inhibitory (Fig 5), but the phage inserts have not yet been sequenced. Many earlier
experiments performed in the same way with such polyclonal rabbit antibodies against
synthetic fibrin E epitopes have not shown such inhibition.

DISCUSSION

The diversity of possible sequences which are present in a random peptide library,
represent a potentially wide range of affinities. Selection procedures isolate a number of
clones from the many millions in the library and some of these isolated sequences should
share sequence elements.

The epitopes on the human fragment E molecule that are detectable via the two existing rat
and rabbit polyclonal blocking antibodies must include epitopes within or adjacent to the
active site for stimulating angiogenesis. These antibodies have been used to select phage
clones that display epitopes common to this site on the molecule. Sequential selection and
amplification has yielded 60 clones of phage. At the current stage, rabbit antiserum against
clones 1 to 20 requires further investigation of what may be a minor degree of blocking
activity against the cell stimulatory activity of fragment E. Polyclonal antisera raised
against 21 to 40, and 41 to 60 detect fragment E bands on immunoblots, and more crucially,
substantially block the angiogenic activity of fragment E (Fig 4). These antisera have been
made as further selection tool. Further antisera against each clone are now being made, and
the first three have turned out recently to have blocking activity (Fig 5).

16000

14000

~
W 12000
I/)

i: 10000
~
«
...GI
()
8000
Q.
~
Q. 6000
0
c
ca 4000
GI
~

2000

0
control FDP Anti phage Anti phage Anti phage
1-20 21-40 41-60

Fig4 The effect of admixture of 3 different rabbit anti phage antibodies with
stimulatory FDP. Anti phage 1-20 does not have an effect in this experiment,
whilst anti phage 21-40 and 41-60 have abolished the stimulatory effect of
FDP.

238
 20000
~
w
VI
-;. 15000
~
<II(
(J

~
a.
~
Q.

.
C
5000
"OJ
~

Fig5 Admixture of anti phage antibodies against selected phage clones 42, 43, and 44
abolishes the stimulatory effect of FDP. Student's t test on log transformed
data shows a significant (P<O.05) increase in DNA synthesis compared with
the buffer only control group by FDP. The 3 anti phage groups are
significantly different from the FDP group, but not from the control group.

The sequence information for the finally selected epitopes, derived from analysis of each
phage clone DNA insert, can then be used not only to locate the active site on the molecule,
and to perpetuate blocking antibodies, but also to synthesise large quantities of short
peptides and short peptide analogues. These can be tested for competitive blocking activity
for human fragment E. Such peptides and analogues are potential therapeutic agents in the
longer term for blocking the cell stimulatory effects of fibrin fragment E in vivo in a
potentially wide variety of pathologies (11) without the attendant risks of interfering in
clotting or fibrinolysis. Our previous work has shown that admixture of blocking antisera
to fibrin E will inhibit the angiogenic effect of experimental mouse wound extracts (12) and
extracts of proliferative types of human atherosclerotic plaques (4). We have not yet
attempted inhibition in vivo. Sustained delivery in vivo to inhibit, for example, wound
healing should be more readily achieved by administration of small peptides than
poly clonal antisera from another species.

ACKNOWLEDGEMENTS

We wish to acknowledge the grant assistance of the Wellcome Trust and Tenovus,
Scotland.

REFERENCES

Thompson WD, Campbell R, Evans AT. Fibrin degradation and angiogenesis:

quantitative analysis of the angiogenic response in the chick chorioallantoic
membrane. J Pathol 1985; 145: 27-37.

2 Thompson WD, Smith EB, Stirk CM, Marshall FI, Stout AJ, Kocchar A
Angiogenic activity of fibrin degradation products is located in fibrin fragment E.
J Pathol 1992; 168: 47-53.

239
3 Thompson WD, Harvey JA, Kazmi MA, Stout AJ. Fibrinolysis and angiogenesis
in wound healing. J Pathol 1991; 165: 311-318.

4 C M Stirk, A Kochhar, E B Smith, W D Thompson. Presence of growth-

stimulating fibrin degradation products containing fragment E in human
atherosclerotic plaques. Atherosclerosis 1993; 103: 159-169.

5 Thompson WD, Wang JEH, Wilson SJ, Ganesalingam N. Angiogenesis and

fibrin degradation in human breast cancer. In: Angiogenesis: Molecular Biology,
Clinical Aspects. ME Maragoudakis et ai, (eds), pp 245-252. Plenum Press, New
York 1994.

6 Naito M, Sabally K, Thompson WD, Stirk CM, Smith EB, Benjamin N.

Stimulation of proliferation of smooth muscle cells in culture by fibrin
degradation products. Fibrinolysis 1996; 10 supp 4: 12 (Abs).

7 Thompson WD, Smith EB, Stirk CM, Seath J, McNally SJ, McCallion DSE,
Melvin WT, Gaffney PI. Fibrin in macrophages in inflammation and
atherosclerotic plaques. Fibrinolysis 1996; 10 supp 4: 121.

8 Scott JK, and Smith GP. Searching for peptide ligands with an epitope library.
Science 1990; 249: 186-390.

9 Parmley SF, and Smith GP. Antibody-selectable filamentous Fd phage vectors-

affinity purification of target genes. Gene 1988; 73: 305-318.

10 Thompson WD, Brown FI. Measurement of angiogenesis: mode of action of

histamine in the chick chorioallantoic membrane is indirect. Int J Microcirc
1987; 6: 343-357.

11 Thompson WD, Smith EB, Stirk CM, Wang J. Fibrin degradation products in
growth stimulatory extracts of pathological lesions. Blood Coagulation and
Fibrinolysis 1993; 4: 113-116.

12 Thompson WD, McNally SJ, Ganesalingam N, McCallion DSE, Stirk CM,

Melvin WT. Wound healing, fibrin and angiogenesis. In: Molecular, Cellular and
Clinical Aspects ofAngiogenesis. Ed M Maragoudakis, Plenum Press, New York,
1996. pp 161-172.

240
PROTEASES AND ANGIOGENESIS. REGULATION OF PLASMINOGEN
ACTIVATORS AND MATRIX METALLOPROTEASES BY ENDOTHELIAL
CELLS.

Pieter Koolwijk, Roeland Hanemaaijer, and Victor W.M. van Hinsbergh

Gaubius Laboratory TNO-PG, Zernikedreef 9, 2333 CK Leiden, The

Netherlands.

1. INTRODUCTION

Angiogenesis, the outgrowth of new blood vessels from existing ones, is an essential
process during development, but this normally stops when the body becomes adult. In the
absence of injury, overt angiogenesis in adults is limited to the reproductive system of
females (formation of corpus luteum and placenta). However, the formation of new blood
vessels is an essential factor in tissue repair (formation and regression of granulation
tissue), which is necessary to restore healthy tissue after wounding and/or inflammation,
and is associated with many pathological conditions, such as tumor development and
rheumatoid arthritis (Folkman and Klagsbrun, 1987; Liotta et aI., 1991; Folkman and Shing,
1992; Montesano, 1992; Colvil1e-Nash and Scott 1992). Fibrin (Dvorak et aI., 1992),
inflammatory cells (Polverini, 1989) and angiogenic factors (Broadley et al 1989; Klagsburn
andD'Amore, 1991; Shweiki eta!., 1992; Plate et a!., 1992; Sengeret aI., 1993; Koch et aI.,
1994) are commonly observed in angiogenesis associated with disease in man. A series of
sequential events can be distinguished during the formation of new microvesse1s: (i)
degradation of the vascular basement membrane and the fibrin or interstitial matrix by
endothelial cells; (ii) endothelial cel1 migration; (iii) endothelial proliferation; and (iv) the
formation of new capi\1ary tubes and a new basement membrane (Folkman, 1986).
The plasmin-plasminogen activator system and the matrix metalloproteinases cooperate
in the degradation of extracellular matrix proteins (Liotta et aI., 1991). Figure 1 depicts the
interaction between the two systems. It should be noted, however, that this schematic
picture is based on in vitro data, and that it sti\1 uncertain whether all the depicted steps
also act in vivo. It is general1y assumed that urokinase-type plasminogen activator (u-PA)
and its inhibitor, the plasminogen activator inhibitor I (P AI-I), are involved in the
regulation of the first steps of angiogenesis, i.e. local proteolytic remodeling of matrix
proteins and migration of endothelial ce\1s (Pepper et al., 1987; Bacharach et aI., 1992;

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 241
 ENDOTHELIAL CELL
t
Pig
procoll.g.n . . . · '
proatromolyaln. '
prog.l.lln ... ·B

pl . .m l n -

a2 •• ntlpl . . mln _ _
collagon ••• -1
atrom.lyaln . '
g.I.lIn . . . ·B
+-- liMP

liMP - -

matrix degradation

production convor. ion .ctlvatlon/l nhlbltlon proteolyala

Figure 1. Schematic representation of the interaction between the u-PA-plasmin system

and the presumed activation of matrix degrading metalloproteinases (MMPs).
Abbreviations: PA: plasminogen activator; u-PA urokinase-type PA; sc-u-PA: single-chain
u-PA; tc-u-PA: two-chain u-PA; u-PAR: u-PA receptor; PIg: plasminogen; Plg-R: PIg
receptor; PAl-I: PA inhibitor-I; MT-MMP: membrane-bound MMP; TIMP: tissue
inhibitor ofMMP. +: stimulation; -: inhibition.

Niedbala et al., 1992; van Hinsbergh, 1992; Vassalli, 1994). u-PA converts plasminogen
into the broadly-acting serine protease plasmin (see Fig. 1), which in turn is able to both
degrade fibrin and other matrix proteins, and to activate several matrix metalloproteinases
(MMPs), in particular stromelysin-l (MMP-3), collagenase-l (MMP-I) and gelatinase-B
(MMP-9) (Woessner, 1991; Docherty et aI., 1992; Matrisian, 1992; Nagase, 1994).
We will focus here on the effect of angiogenic growth factors, such as aFGF, bFGF and
VEGF, and the inflammatory mediator TNF on the regulation of these proteases by
endothelial cells and on the expression of their receptors on endothelial cells. The role of
the expression of these proteases and their receptors in the fonnation of capillary-like
structures of human microvascular endothelial cells in three-dimensional fibrin matrices as
parts of the angiogenesis process will also be discussed.

2. COMPONENTS OF THE MMP SYSTEM

The involvement of the MMP system in the process of angiogenesis was demonstrated
both in vitro as well in vivo using MMP-specific inhibitors (Passanti et al.,1992; Fisher et
aI.,1994; Benelli et al.,1994; Braunhut et al.,1994;Galardy et al., 1994; Volpert et al.,1996;

242
Moses et ai., 1992). MMP-activity is needed during migration and invasion, when proteins
from the basal membrane or the underlying matrix, of which the collagens type I, III and IV
and laminin are the most prominent, are degraded and resynthesised. Native triple-helical
folded collagens are extremely stable proteins which are only degraded by MMPs.
The family of matrix metalloproteinases (MMPs) consists of at least 14 members, a
number that is still increasing. The various MMPs are expressed as latent enzymes, and
have similar domain building structures. They all contain a Zn 2 +-ion at the active site of the
catalytic domain, and besides matrilysin (MMP-7), a hemopexin-like domain, which is
thought to play a role in interaction with matrix components, interaction with cells or
inhibitors (Fig. 2). MMPs are secreted, soluble enzymes with the exception of the
membrane-type MMPs (MT-MMPs) (Sato et al., 1994;Takino et ai., 1995; Will et
ai.,1996; Puente et ai., 1996) which contain a transmembrane domain and are located at the
cell surface. Based on their substrate specificity and domain building the various MMPs
can be divided into five groups:

Collagenases (collagenase-1,2,3 or MMP-l, MMP-8, MMP-13, respectively) which can

degrade intact collagen. Neutrophil collagenase (MMP-8) is highly glycosylated in
neutrophils. This enzyme was also detected in non-PMN cells like endothelial cells,
chondrocytes and synovial fibroblasts (Hanemaaijer et ai., 1997;Cole et ai., 1996).

Gelatinases (gelatinase-A,B or MMP-2, MMP-9) can degrade collagen type IV and

unfolded collagens.

Stromelysins (stromelysin-1,2 or MMP-3, MMP-I0) have a broad substrate specificity

and can activate other MMPs like MMp-1 and MMP-9 (Nagase, 1994).

MT-MMPs (MT-1,2,3,4-MMP) playa role in the activation of gelatinase A (MMP-

2)(Sato et ai., 1994).

Other MMPs (matrilysin, stromelysin-3, metalloelastase or MMP-7,MMP-l1,MMP-12,

respectively) which differ in substrate specificity or domain building from the MMPs III
other groups.

MMP-7
'"CD
MMP-1
'"'"coCD MMP-8
CJ)
pre

c::::J pro
~
c:0
._ u MMP-13
~
0; MMP-3
E furin
o
~ CD
MMP-10
'" '" MMP-2 I • __ catalytic
.S
m
'" MMP-9
FN-like

!~~Iii~il~~~~
a;
Cl

MMP-11 c=J collV-like

c=J hemopexin
CD
% MT-MMP1
ci>
c:
MMP-12
MT-MMP2
'"
};
E
CD
MT-MMP3 E:J TM
E

Figure 2. Domain structure of various members of the protease family of Matrix

Metalloproteinases.

243
 Recently some new MMPs (MMP-18, MMP-19) have been identified by sequence
similarity searching of databases containing expressed sequence tags or cDNA libraries
(Cossins et aI., 1996, Pendas et aI., 1997). Their substrate specificity and functional role is
not yet known.

3. REGULATION OF MMP ACTIVITY

The activities of the several MMPs are regulated not only by the protein expression
level and the way of activation, but also by their localization on the cell surface and their
interaction with inhibitors.

3.1. MMP expression

During early steps of angiogenesis degradation and remodeling of basal membrane and
matrix takes place. MMPs have been detected in vivo in proliferating endothelial cells
during development (Karelina et aI., 1995) and in endothelial cells present in atherosclerotic
plaques and growing tumors (Galis et aI., 1994; Nikkari et aI., 1995; Rao et aI., 1996). It
was shown that growth factors like bFGF and VEGF which act as angiogenic stimuli, also
induce MMP expression (MMP-l and MMP-2) in endothelial cells (Tsuboi et aI., 1990;
Moscatelli et aI., 1985; Unemori et aI., 1992). In human microvascular endothelial cells, the
inflammatory mediator TNF increases the mRNA levels and the synthesis of interstitial
collagenase (MMP-l), stromelysin-l (MMP-3), collagenase-2 (MMP-8), and - if protein
kinase C is also activated - gelatinase-B (MMP-9), whereas the mRNAs levels and
synthesis of their physiological inhibitors TIMP-l and TIMP-2 are not changed
(Hanemaaijer et ai, 1993; Case et aI., 1989; Cornelius et ai, 1995; Karelina et aI., 1995;
Hanemaaijer et aI., 1997).

3.2. Localization
Endothelial cell migration and invasion in the matrix is a localized process: only a few
cells really enter the matrix to form new sprouts. Unemori et aI. (1990) showed that the
secretion of MMPs by bovine endothelial cells occurs predominantly towards the
basolateral side of the cells - the site of the basement membrane - (Unemori et ai, 1990)
similarly to the TNF -induced production ofuPA (see below). Recently, it has been shown
that activation of gelatinase A depends on the binding to and activation by the membrane-
type MMPs (MT-MMP) (Sato et ai, 1994; Foda et aI., 1996), and that MMP-2 interacts
with the v 3 integrin (Brooks et aI., 1996) This v 3-integrin, interacts with many matrix
proteins including the A -chain of fibrin and vitronectin, and by this mechanism, cell-bound
gelatinase-A activity is encountered at the cellular focal attachment sites, where cell-matrix
interactions are concentrated. In that way it is comparable to urokinase/urokinase-receptor
(see below, Van Hinsbergh, 1992) by which activated u-PA is accumulated and localized at
the focal adhesion sites of the cell, resulting in an invasive character (Sato and Seiki, 1996).

3.3. Activation of MMPs

It is not exactly known how MMPs are activated in vivo. In vitro studies have shown
that the proteases u-PNplasmin, carthepsins and stromelysin (MMP-3) can activate
various MMPs (except MMP-2) (for a review see Murphy and Docherty (1992». MMP-
2 can be activated by MT -MMPs both in vitro and in vivo (Sato and Seiki, 1996). Also

244
 C P+B P C P P+I

- MMP-9
- MMP-2

- MMP-1I3

A B
Figure 3. Activation of gelatinase A (MMP-2) by human endothelial cells.
A. HMEC-l cells were grown under serum-free conditions (control, lane 1) and stimulated
for 24 h with 10 nM PMA (P) in the absence (lane 2) or presence of the MMP-inhibitor
BB-94 (B). Samples of 10 III of the conditoned media were assayed by gelatin
zymography. Following PMA-treatment of the cells MMP-2 is activated, a process that is
inhibited by BB-94, an inhibitor ofMMP activity.
B. Human umbilical vein endothelial cells (HUVEC) were grown under serum-free
conditions (control, lane 1) and stimulated for 24 h with 10 nM PMA (P) with (lane 3) or
without (lane 2) preincubation for 16 h with 100 Vlml IFN . Following PMA-treatment of
the cells MMP-2 is activated, a process that is inhibited by pretreatment of the cells with
100 Vlml IFN .

lectins like the ulex europeus agglutinin I can mediate MMP-2 activation by endothelial
cells (Gomez et aI., 1995). Human endothelial cells only produce latent MMPs. However,
after in vitro stimulation with TNF or the PKC activator PMA, MMP-2 is activated
(Hanemaaijer et aI., 1993; Zucker et al., 1995). This cell-mediated activation is inhibited by
a synthetic MMP-inhibitor (Fig. 3A), indicating that MMP-activity is needed. In addition,
the activation of MMP-2 by endothelial cells is PKC-dependent (Foda et al., 1996).
Incubation of endothelial cells with the cytokine IFN prevented activation of MMP-2,
probably via down-regulation of PKC (Fig. 3B). In periodontal diseases oxygen radicals
have been shown to be responsible for activation of collagenase-2 (MMP-8) (Desrochers et
aI., 1992).

3.4. MMP inhibitors

MMPs are specifically inhibited by Tissue Inhibitors of MetalloProteinases (TIMPs).
Four different TIMPs (TIMP-l,2,3,4) have been identified. TIMP-2 is constitutionally
expressed in various tissues, whereas TIMP-l is often only induced at sites of
inflammation, co-regulatory with MMPs (Denhardt et aI., 1993). TIMP-3 is an inhibitor
which is matrix-related. In general all TIMPs can inhibit all MMPs, albeit with different
efficiency. As an example MTl-MMP activity is hardly inhibited by TIMP-l, but very
efficiently by TIMP-2 and TIMP-3 (Will et aI., 1996). Very recently a fourth TIMP was
identified, whose expression is tissue-specific. All TIMPs inhibit MMPs in a non-covalent
1:1 complex. In serum MMP-activity is mainly inhibited by 2MacrogIobulin, which is
present in high amounts.

245
4; COMPONENTS OF THE PLASMINIPLASMINOGEN ACTIVATOR SYSTEM.

Fibrin degradation and probably also activation of several MMPs is accomplished by

the serine protease plasmin, which is formed from its zymogen plasminogen by
plasminogen activators (PAs). The actual activities of plasmin and the PAs are regulated
not only by their concentration and activation, but also by their interaction with inhibitors,
cellular receptors, and matrix proteins.

4.1. Proteases
Plasmin is formed from its zymogen plasminogen by proteolytic activation by PAs.
Two types of mammalian PAs are presently known: tissue-type plasminogen activator (t-
PA) and urokinase-type plasminogen activator (u-PA) (Bachmann, 1987; Wallen, 1987).
The fibrinolytic activity in blood is largely determined by t-P A, whereas the activation of
plasminogen in the tissues is mainly mediated by U-P A. The three serine proteases,
plasminogen, t-PA and u-PA, are synthesized as single polypeptide chains, and each of
them is converted by specific proteolytic cleavage to a molecule with two polypeptide
chains connected by a disulphide bond. The carboxy-terminal part of the molecule (the so-
called B-chain) contains the proteolytically active site, whereas the amino-terminal part of
the molecule (the A-chain) is built up of domains that determine the interaction of the
proteases with matrix proteins and cellular receptors. The proteolytic cleavage of
plasminogen and single-chain U-PA to their respective two-chain forms is necessary to
disclose the proteolytically active site and to activate the molecule. The interaction of
plasminogen with fibrin or the cell surface occurs predominantly via binding sites in the
kringle structures, which recognize lysine residues of proteins, in particular carboxy-
terminal lysines. Because B-type carboxypeptidases remove carboxy-terminal lysine
residues from potential binding sites for plasminogen in fibrin or on the cell surface, they
can act as negative regulators of the fibrinolytic system (Redlitz et aI., 1995).

4.2. Inhibitors
The activities of the proteases of the fibrinolytic system are controlled by potent
inhibitors, which are members of the serine protease inhibitor (serpin) superfamily.
Plasmin, if not bound to fibrin, is instantaneously inhibited by 2-antiplasmin (Holmes et
al., 1987). Because this interaction is facilitated by the lysine binding domain of plasmin, it
is attenuated when plasmin is bound to fibrin. The predominant regulators of t-P A and u-
PA activities are PAI-l and PAI-2 (Feams et al, 1996; Bachman, 1995). PAl activity in
human plasma is normally exclusively PAl-I. PAI-l binds to vitronectin, which stabilizes
its inhibitory activity. PAI-l is the main if not the sole inhibitor of PAs synthesized by
endothelial cells, vascular smooth muscle cells and hepatocytes (Sprengers and Kluft, 1987;
Loskutoff,1991).

4.3. Receptors
Regulation of fibrinolytic activity also occurs by cellular receptors. These receptors
direct the action of PAs and plasmin to focal areas on the cell surface, or are involved in the
clearance of the PAs. High affinity binding sites for plasminogen (Miles et al., 1988; Plow
et aI., 1991; Nachman, 1992, Hajjar, 1995), t-PA (Hajjar, 1991, 1995) and u-PA (Vasalli,
1994; Blasi et aI., 1994; Dan0 et al., 1994) are found on various types of cells including
endothelial cells.
Endothelial cells in vitro bind plasminogen with a moderate affinity (120 to 340 nM
depending on whether the Lys- or Glu-form of plasminogen is used) but with a high

246
capacity (3.9 to 14 x 105 molecules per cell) (Hajjar and Nachman, 1988; Plow et aI., 1991).
This binding, which is also observed with many other cell types, is mediated by the lysine
binding sites of kringles 1-3 of the plasmin(ogen) molecule. Because these lysine binding
sites are also involved in the interaction of plasmin with 2-antiplasmin, occupation of
lysine binding sites protects plasmin from instantaneous inhibition by 2-antiplasmin not
only when plasmin is bound to fibrin (see above) but also when it is bound to the cellular
receptors. The nature of the plasminogen receptors is not fully resolved. In addition to
gangliosides, which directly or indirectly contribute to the plasminogen binding (Miles et
aI., 1989), at least eight proteins have been reported to be involved in plasminogen binding.
Among them are members of the low density lipoprotein receptor family, such as gp330
and LRP; annexin II; a not yet identified 45 kD protein; GbIIb/IIIa and -enolase (see
Hajjar, 1995 for review) ..
Binding ofu-PA with the cell surface limits plasminogen activation to focal areas such as
the focal attachment sites and cellular protrusions involved in cell migration and invasion.
Furthermore, u-PA interaction with the cell evokes signal transduction and
phosphorylation of several proteins (Dumler et al., 1993; Rao et al ., 1995). A specific u-
PA receptor has been identified and cloned. It is present on many cell types, including
endothelial cells (Barnathan, 1992). It is a glycosyl phosphatidyl inositol- (GPI-) anchored
glycoprotein (Dan0 et aI., 1994), which binds both single-chain u-PA and two-chain U-P A
via their growth factor domain (Appell a et al ., 1987). The u-PA receptor is heavily
glycosylated. It belongs to the cysteine-rich cell surface proteins. After synthesis it is
proteolytically processed at its carboxyl-terminus and subsequently anchored in the
plasma membrane by a GPI-group (Ploug et aI., 1991). It comprises three domains, which
are structurally homologous to snake venom -toxins (Dan0 et al., 1994). The u-PA
receptor has been found in focal attachment sites, where integrin-matrix interactions occur,
and in cell-cell contact areas (p6llanen et aI., 1988; Conforti et aI., 1994). Human
endothelial cells in vitro contain about 140,000 u-PA receptors per cell (Haddock et aI.,
1991 ).
The u-PA receptor both acts as a site for focal pericellular proteolysis by u-PA and is
involved in the clearance of the u-PA:PAI-l complex. Upon secretion, single-chain u-PA
binds to the u-PA receptor and is subsequently converted to the proteolytically active
two-chain u-PA. Since the endothelial cell also contains plasmin(ogen) receptors, an
interplay between receptor-bound u-PA and receptor-bound plasmin(ogen), and plasmin
formation is likely to happen. The plasmin generated can degrade a number of matrix
proteins. In addition, a direct plasmin-independent proteolytic action of u-PA on matrix
proteins may also occur (Quigley et al, 1987). Like free u-PA activity, receptor-bound
two-chain u-PA is subject to inhibition by PAl-I. As a consequence u-PA is only active
over a short period of time. In contrast to receptor-bound single-chain or non-inhibited
two-chain u-PA, the u-PA:PAI-1 complex is rapidly internalized together with the u-PA
receptor (Olson et aI., 1992), followed by the degradation of the u-PA:PAI-l complex and
the return of the empty u-PA receptor to the plasma membrane. Internalization of the GPI-
linked u-PA receptor occurs probably after interaction with another receptor, such as the
2-macroglobulinllow density lipoprotein receptor-related protein (LRP) (Nykjrer et al.,
1992) or the VLDL-receptor (Heegaard et al., 1995) present on capillary and arteriolar
endothelial cells in vivo (Wyne et al, 1996). Recently it has been found that the U-P A
receptor may have an additional role. The u-PA receptor occupied by u-PA interacts avidly
with vitronectin (Wei et aI., 1994). Hence, cell adhesion may represent an additional
function of the U-P A receptor.

247
5. REGULATION OF u-PA ACTMTY BYION ENDOTHELIAL CELLS

5.1. Regulation of proteases and inhibitors

The cytokines interleukin-l (IL-l) and tumor necrosis factor- (TNF), involved in
angiogenesis, exert many effects on the vascular endothelium. Their most prominent feature
is the induction or increase of the transcription of many genes. It was early recognized that
TNF and IL-l, as well as bacterial lipopolysaccharide (LPS), markedly increase the
production of PAI-l in endothelial cells in vitro (Colucci et al, 1985; Emeis and Kooistra,
1986; Schleef et ai, 1988; van Hinsbergh et ai, 1988). This induction was also demonstrated
at the transcriptional level, and was largely inhibited by the isoflavone compound.genistein
(van Hinsbergh et ai, 1994).
TNF , IL-l and LPS also elicit another effect on the regulation of plasminogen activator
production in endothelial cells. Simultaneously with the increase in PAI-I, these
inflammatory mediators induce the synthesis of u-PA in human endothelial cells in vitro
(van Hinsbergh et al., 1990). Whereas u-PA is normally not found in endothelial cells in
vivo, association of u-PA with the endothelium was observed in acute appendicitis
(Gmndahl-Hansen et ai, 1989) and in rheumatoid arthritis (Weinberg et ai, 1991). Induction
of endothelial U-P A by TNF in vitro is associated with an increased degradation of matrix
proteins (Niedbala and Stein Pi carella, 1992). The enhanced secretion of u-PA occurs
entirely towards the basolateral side of the cell, whereas the secretion of t-PA and PAI-l
proceeds equally to the luminal and basolateral sides of the cell (van Hinsbergh et al.,
1990). The polar secretion ofu-PA suggests that u-PA may be involved in local remodeling
of the basal membrane of the cell. u-PA activity is controlled in space by interaction of u-
PA with its cellular receptor and by the inhibitor PAl-I. The increase of PAI-I induced by
inflammatory mediators may represent, in addition to a role in the modulation of
fibrinolysis, a protective mechanism of the cell against uncontrolled u-PA activity.
The regulation of u-PA by endothelial cells in vitro induced by the angiogenic growth
factors aFGF, bFGF and VEGF is species dependent. It is of interest that, in contrast with
data found with animal endothelial cells, where the FGF's and VEGF induce the expression
of u-PA (Saksela et al., 1987; Pepper et al., 1991, 1993), aFGF (Koolwijk et al,
unpublished data) and bFGF do not have any effect on the u-P A expression (Koolwijk et
aI., 1996) in human endothelial cells. VEGF hardly influences the u-PA production, but
stimulates the expression oft-PA (Bikfalvi et al., 1991; Koolwijk et aI., 1996; Fig. 4).

5.2. Regulation of the u-PA receptor.

The expression of the u-PA receptor is enhanced by angiogenic growth factors including
basic and acidic fibroblast growth factor (bFGF, aFGF) and vascular endothelial growth
factor (VEGF) (Mignatti et al, 1991; Mandriota et al., 1995; Koolwijk et al., 1996). The
number of u-PA receptors on human and bovine endothelial cells is also enhanced by the
activation of protein kinase C and by the elevation of the cellular cAMP concentration
(Langer et aI., 1993; van Hinsbergh, 1992). Preliminary experiments in our laboratory have
demonstrated that the induction of U-P A receptor by VEGF in human endothelial cells is
inhibited by protein kinase C inhibition. The effects of bFGF and VEGF on the induction
of u-PA receptor in human endothelial cells are regulated independently of their effects on
cell proliferation. Similarly, Presta et ai. (1989) have shown that the induction of U-P A by
bFGF in bovine endothelial cells proceeds independently from the stimulation of
mitogenesis by this growth factor.
In addition to FGFs and VEGF, which induce mitogenesis, TNF can also induce
angiogenesis, but this occurs without stimulation of cell proliferation (Leibovitch et al.,

248
 control

bFGF

VEGF

TNF

bFGF·TNF

VEGF.TNF

0 .0 2.0 4 .0 6 .0
u-PA antigen (ng/ml)

control

bFGF

VEGF

TNF

bFGF.TNF

VEGF·TNF

0 500 1000 1500 2000

PAI-1 antigen (ng/ml)

oontrol

bFGF

VEGF

TNF

bFGF·TNF

VEGF·TNF

0 5 10 1S 20
t-PA antigen (ng / ml)

Figure 4. Regulation of u-PA, PAl-I and t-PA expression by human microvascular

endothelial cells. Human microvascular endothelial cells (hMVEC) cultured in M 199
medium supplemented with 10% human serum were stimulated for 24 hours with 20 nglml
bFGF, 25 ng/ml VEGF I65 , or 4 ng/ml TNF . The amount of u-PA, PAI-l and t-PA antigen
in the supernatants was determined by ELISA. The data is expressed as nglml± SD of one
representative experiment out of three performed in triplicate wells.

249
 culture ccnd/tlcn
control ~

bFGF (5 nQIml) ~

VEGF (SO nglml) ~

TNF ~(20 nQIml)

bFGF'TNFE~
TNF
VEGF • OOQ<;;

bFGF.VEGF.TNF ~

o 100 200 300 400 500 600

Total tube lenght (mm/cm,)

Figure 5. Formation of capillary-like tubular structures in a three dimensional fibrin matrix

by human microvascular endothelial cells. HMVEC were cultured on the surface of a three-
dimensional fibrin matrix in M199 medium supplemented with 10% human serum and 10%
NBCS and stimulated with bFGF (5 ng/ml), VEGF 165 (50 ng/ml), and TNF (20 ng/ml)
alone or with the combination of these mediators. Mter 8 days of culture, the total length
of tube-like structures was measured using a microscope equiped with a monochrome CCD
camera (MX5) connected to a computer with image analysis software. The data represent
mean length/cm2 ± SEM of one representative experiments out of three performed in
duplicate wells.

1987; Frater-SchrOder et aI., 1987). TNF increases u-PA receptor in human microvascular
endothelial cells (Koolwijk et aI., 1996) and in monocytes (Kirchheimer et aI., 1988), but
not in endothelial cells from human umbilical vein or aorta (Koolwijk et aI., 1996).
However, simultaneous exposure of the latter cells to TNF (which induces u-PA
synthesis) to bFGF and to VEGF (which enhance the expression of u-PA receptors)
potently increases cell-bound u-PA activity.

6. IN VITRO CAPILLARY-LIKE TUBE FORMATION IN FIBRIN GELS

The formation of capillary structures in three-dimensional matrices of fibrin and collagen

has been studied in vitro with bovine endothelial cells (Pepper et aI., 1990; Madri et aI.,
1991; Montesano, 1992; Goto et aI., 1993) and rat aorta explants (Nicosia and Ottinetti,
1990). Pepper et aI (1990) demonstrated that the formation of capillary-like tubular
structures in a three-dimensional fibrin or collagen matrix is induced by the addition of
bFGF and counteracted by TGF . In these bovine endothelial cells bFGF induces both
u-PA activity and u-PAR expression, whereas TGF predominantly enhances PAl-\,
Recently, we developed an in vitro angiogenesis model, in which the ingrowth of new
capillary-like structures from human microvascular endothelial monolayers into three-
dimensional fibrin matrices can be followed and influenced. In contrast with that was found
using bovine endothelial cells (pepper et aI 1990, 1992), addition of bFGF or VEGF alone
did not induce the formation of tubular structures of human endothelial cells. Only the
combination of the growth factors bFGF and VEGF with the inflammatory mediator TNF
induced capillary-like tube formation in the fibrin matrix. The mediators themselves did not

250
induce the fonnation of these capillary-like structures (Koolwijk et al, 1996, Fig. 5). Cross
sections and electron microscopy have shown that the tube-like structures have a lumen
and are very akin to capillary structures in vivo. The induction of the capillary-like
structures required cell-bound u-PA activity and plasmin activity, while inhibition of
mitogenesis only marginally influenced the fonnation of capillary-like structures (Koolwijk
et ai, 1996). The outgrowth of tubular structures requires u-PA activity and is completely
reduced by anti-u-PA immunoglobulins but not by anti-t-PA antibodies. It is also reduced
by inhibiting the interaction of u-PA with its receptor. Furthermore, proteolytic activation
of plasminogen appears to be involved, because the plasmin inhibitor aprotinin largely
inhibits the fonnation of tubular structures (Koolwijk et ai, 1996). Recent
immunohistochemical studies perfonned on cross-sections of these invading capillary-like
tubular structures showed an enhanced expression of both u-PA and the u-PAR antigen by
invading endothelial cells compared with non-invading endothelial cells (Kroon et al;
manuscript in preparation). These data agree with the data on bovine endothelial cells,
except that in human endothelial cells a second mediator is required to induce U-P A
synthesis.

7. SUMMARY

The endothelial cell uses u-PA for proteolytically changing its interaction with its
underlying matrix and for remodelling of its basement membrane, processes which are
necessary for cell migration and angiogenesis. In vivo evidence of the involvement of the u-
PAlu-PA receptor in the process of angiogenesis was recently shown by Min et al. (1996).
They were able to reduce bFGF-induced in vivo angiogenesis, and the outgrowth of a B 16
melanoma in syngeneic mice, by prevention of the binding of u-PA to the u-PAR by a
fusion product of the epidennal growth factor-like domain of murine u-PA and Fc portion
of human IgG.
The remodeling of the basement membrane is limited in space and time by interaction of
u-PA with its specific receptor and by the presence of PAl-I. As the expression of u-PA
and u-PA receptor are under the control of the monocyte-derived cytokines TNF and IL-I
and by angiogenic growth factors, respectively, monocytes may play an important role in
the control of angiogenesis. In addition the induction and activation of stromelysin and
other matrix metalloproteinases by the monokine TNF and the u-PA plasmin system
further contribute to the complex regulation of the local proteolytic events associated with
endothelial matrix remodelling.

8. ACKNOWLEDGEMENT

The authors would like to thank the technical assistence of Marielle Kroon, Erna Peters,
Bea van der Vecht, and Hetty Visser. The financial support of the Netherlands Heart
Foundation (grant 95.193), Alternatives to Animal Experiments Platfonn ( 94-16), the
Praeventiefonds (grants 28-2621 and 28-2622) and the Netherlands Cancer Society (TNOP
97-1511) is also gratefully acknowledged.

251
9. REFERENCES

Appella, E., Robinson, E. A, Ullrich, S. J., Stoppelli, M. P., Corti, A., Cassani, G., and
Blasi, F., 1987, The receptor-binding sequence of urokinase A biological function for the
growth-factor module of proteases, 1. Bioi. Chem. 262:4437-4440.

Bacharach, E., A Itin, and Keshet, E., 1992, In vivo patterns of expression of urokinase
and its inhibitor PAI-l suggest a concerted role in regulating physiological angiogenesis,
Proc. Natl. Aead. Sci. USA. 89:10686-10690.

Bachmann, F., 1987, Fibrinolysis, in: Thrombosis and Haemostasis 1987, (M. Verstraete,
J. Vermylen, R. Lijnen, and J. Arnout, eds.), pp. 227-265, Leuven University Press,
Leuven.

Bachmann, F., 1995, The enigma PAI-2. Gene expression, evolutionary and functional
aspects, Thromb. Haemostas. 74:172-179.

Barnathan, E. S., 1992, Characterization and regulation of the urokinase receptor of human
endothelial cells, Fibrinolysis, 6: 1-9.

Benelli, R., Adatia, R., Ensoli, B., Stetler-Stevenson, w., Santi, L., and Albini, A., 1994,
Inhibition of AID S-kaposi's sarcoma cell induced endothelial cell invasion by TIMP-2 and
a synthetic peptide from the metalloproteinase propeptide: implications for an anti-
angiogenic therapy, Oneol. Res, 6: 251-257.

Bikfalvi, A, Sauzeau, c., Moukadiri, H., Maclouf, J., Busso, N., Bryckaert, N., Plouet, J.,
and Tobelem, G., 1991, Interaction ofvasculotropinlvascular endothelial cell growth factor
with human umbilical vein endothelial cells - Binding, internalization, degradation, and
biological effects, 1. Cell. Physiol. 149:50-59.

Blasi, F., Conese, M., M011er, L. B., Pedersen, N., Cavallaro, u., Cubellis, M., Fazioli, F.,
Hernandez-Marrero, L., Limongi, P., Munoz-Canoves, P., Resnati, M., Riittinen, L.,
Sidenius, N., Soravia, E., Soria, M., Stoppelli, M., Talarico, D., Teesalu, T., and
Valcamonica, S., 1994, The urokinase receptor: Structure, regulation and inhibitor- mediated
internalization, Fibrinolysis, 8: 182-188.

Braunhut, S., and Moses, M., 1994, Retinoids modulate endothelial cell production of
matrix- degrading proteases and tissue inhibitors of metalloproteinases (timp), 1. Bio/.
Chem.269:13472-13479.

Broadley, K.N., Aquino, A.M., Woodward, S.C., Buckley-Sturrock, A., Sato, Y, Rifkin,
D. and Davidson, J.M., 1989, Monospecific antibodies implicate basic fibroblast growth
factor in normal wound repair, Lab. Invest. 61:571-575.

Brooks, P., Stromblad, S., Sanders, L., Von Schalscha, T., Aimes, R., Stetler-Stevenson, W.,
Quigley, J., and Cheresh, D., 1996, Localization of matrix metalloproteinase MMP-2 to the
surface of invasive cells by interaction with integrin alpha v beta 3, Cell, 85:683-693.

Case, J. P., Lafyatis, R., Remmers, E. F., Kumkumian, G. K., and Wilder, R. L., 1989,

252
Transin Stromelysin Expression in Rheumatoid Synovium - A Transformation-Associated
Metalloproteinase Secreted by Phenotypically Invasive Synoviocytes, Am. 1. Pathol.
135:1055-1064.

Cole, A., Chubinskaya, S., Schumacher, B., Huch, K., Csszabo, G., Yao, J, Mikecz, K.,
Hasty, K., and Kuettner, K., 1996, Chondrocyte matrix metalloproteinase-8 - human
articular chondrocytes express neutrophil collagenase, 1. Bioi. Chem. 271: 11023-11026.

Colucci, M., Paramo, 1.A., and Collen, D., 1985, Generation in plasma of a fast-acting
inhibitor of plasminogen activator in response to endotoxin stimulation, 1. Clin. Invest.
75:818-824.

Colville-Nash, P. R., and Scott, D . L., 1992, Angiogenesis and rheumatoid arthritis -
Pathogenic and therapeutic implications, Ann. Rheumatic Diseases 51:919-925.

Conforti, G., Dominguez-Jimenez, c., Ronne, E., H0yer-Hansen, G., and Dejana, E., 1994,
Cell-surface plasminogen activation causes a retraction of in vitro cultured human umbilical
vein endothelial cell monolayer, Blood, 83:994-1005.

Cornelius, L., Nehring, L., Roby, J, Parks, W., and Welgus, H., 1995, Human dermal
microvascular endothelial cells produce matrix metalloproteinases in response to angiogenic
factors and migration, 1. Invest. Dermatol. 105: 170-176.

Cossins, J, Dudgeon, T., Catlin, G. , Gearing, A., and Clements, J, 1996, Identification of
MMP-18, a putative novel human matrix metalloproteinase, Biochem. Biophys. Res.
Comm.228:494-498.

Dan0, K., Behrendt, N., Brunner, N., Ellis, v., Ploug, M., and Pyke, c., 1994, The
urokinase receptor Protein structure and role in plasminogen activation and cancer invasion,
Fibrinolysis, 8: 189-203.

Denhardt, D.T., Feng, B., Edwards, D.R., Cocuzzi, E.T., Malyankar, U.M., 1993, Tissue
Inhibitor of Metalloproteinases (TIMP, aka EPA) - Structure, Control of Expression and
Biological Functions, Pharmacology & Therapeutics. 59:329-341

Desrochers, P. E., Mookhtiar, K., Van Wart, H. E., Hasty, K. A., and Weiss, S J, 1992,
Proteolytic Inactivation of alphal-Proteinase Inhibitor and alphal-Antichymotrypsin by
Oxidatively Activated Human Neutrophil Metalloproteinases, 1. BioI. Chem.
267:5005-5012.

Docherty, A. J. P., J O'Connell, T. Crabbe, S. Angal, and Murphy, G., 1992, The matrix
metalloproteinases and their natural inhibitors: Prospects for the treating degenerative
tissue diseases, Trends in Biotechnol 10:200-207.

Dumler, I., Petri, T., and Schleuning, W-D., 1993, Interaction of urokinase-type
plasminogen activator (u-PA) with its cellular receptor (u-PAR) induces phosphorylation
on tyrosine of a 38 kDa protein, FEBS Letters, 322:37-40.

Dvorak, H. F., Nagy, J A., Berse, B., Brown, L. F., Yeo, K-T., Yeo, T-K., Dvorak, A. M,

253
Van De Water, L., Sioussat, T. M, and Senger, D. R., 1992, Vascular permeability factor,
fibrin, and the pathogenesis of tumor stroma formation, Ann. N. Y. Acad. Sci., 667:101-111.

Emeis, J.J., and Kooistra, T., 1986, Interleukin-l and lipopolysaccharide induce a fast-
acting inhibitor od tissue-type plasminogen activator in vivo and in cultured endothelial
cells, J. Exp. Med. 163: 1260-1266.

Fearns, C., Samad, F., and Loskutoff, DJ., 1996, Synthesis and localization ofPAI-l in the
vessel wall, in: Vascular Control of Hemostasis, (VW.M. van Hinsbergh ed.), pp. 207-226,
Harwood Acad. Publ., Amsterdam.

Fisher, C., Gilbertson-Beadling, S., Powers, E., Petzold, G., Poorman, R., and Mitchell,
M., 1994, Interstitial collagenase is required for angiogenesis invitro, Develop. BioI.
162:499-510.

Foda, H., George, S., Conner, C., Drews, M., Tompkins, D., and Zucker, S., 1996,
Activation of human umbilical vein endothelial cell progelatinase a by phorbol myristate
acetate: a protein kinase c-dependent mechanism involving a membrane-type matrix
metalloproteinase, Lab. Invest. 74:538-545.

Folkman, J., 1986, How is blood vessel growth regulated in normal and neoplastic tissue?
G.H.A. Clowes memorial Award lecture, Cancer Res. 46:467-473.

Folkman, J., and Klagsburn, M., 1987, Angiogenic factors, Science 235:442-447.

Folkman, J., and Shing, Y, 1992, Angiogenesis,J. Bioi. Chem. 267:10931-10934.

Frater-Schroder, M., Risau, W., Hallman, R., Gautschi, P., and Bohlen, P., 1987, Tumor
necrosis factor type , a potent inhibitor of endothelial cell growth in vitro, is angiogenic in
vivo, Proc. Natl. Acad. Sci. USA, 84:5277-5281.

Galardy, R., Grobelny, D., Foellmer, H., and Fernandez, L., 1994, Inhibition of
angiogenesis by the matrix metalloprotease inhibitor n-2r-2-
(hydroxamidocarbonymethyl)-4- methyl pentanoy I)-I-tryptophan methyl amide, Cancer
Res. 54:4715-4718.

Galis Z.S., Sukhova, G.K, Lark, M.W., and Libby, P., 1994, Increased expression of
matrix metalloproteinases and matrix degrading activity in vulnerable regions of human
atherosclerotic plaques, J. Clin. Invest. 94:2493-2503.

Gomez, D., Yoshiji, H., Kim, J., and Thorgeirsson, u., 1995, U1ex europaeus i lectin
induces activation of matrix- metalloproteinase-2 in endothelial cells, Biochem. Biophys.
Res. Comm. 216:177-182.

Goto, F., K Goto, K Weindel, Folkman, J., 1993, Synergistic effect of vascular endothelial
growth factor and basic fibroblast growth factor on the proliferation and cord formation of
bovine capillary endothelial cells within collagen gels, Lab. Invest. 69:508-517.

Gnmdahl-Hansen, J., Kirkeby, L., Ralflcirer, Kristensen, P., Lund, L.R., Dan0, K, 1989,

254
Urokinase-type plasminogen activator in endothelial cells during acute inflammtion of the
appendix, Am. J. Pathol. 135: 631-636.

Haddock, R c., Spell, M. L., Baker III, C. D., Grammer, J. R, Parks, J. M., Speidel, M.,
and Booyse, F. M., 1991, Urokinase binding and receptor identification in cultured
endothelial cells, J. BioI. Chem., 266:21466-21473.

Hajjar, KA., and Nachman, RL., 1988, Endothelial cell-mediated conversion of Glu-
plasminogen to Lys-plasminogen. Further evidence for assembly of the fibrinolytic system
on the endothelial cell surface, J. Clin. Invest. 82: 1769-1778.

Hajjar, KA, 1991, The endothelial cell tissue plasminogen activator receptor. Specific
interaction with plasminogen, J. BioI. Chem. 266:21962-21970.

Hajjar, K A, 1995, Cellular receptors in the regulation of plasmin generation, Thromb.

Haemostas.,74:294-301.

Hanemaaijer, R, Koolwijk, P., Leclercq, L., De Vree, W. 1. A., and Van Hinsbergh, V. W.
M., 1993, Regulation of matrix metalloproteinase expression in human vein and
microvascular endothelial cells - Effects of tumour necrosis factor- , interleukin-l and
phorbol ester, Biochem. J., 296:803-809.

Hanemaaijer, R, Sorsa, T., Konttinen, Y.T., Ding, Y., Kylmaniemi, M., Visser, H., van
Hinsbergh, V.W.M., Helaakoski, T., Kainulainen, T., Virkkunen, 1., R6nta, H., Tschesche,
H., and Sal 0, T., 1997, Matrix Metalloproteinase-8 is Expressed in Rheumatoid Synovial
Fibroblasts and Endothelial Cells. Regulation by TNF and Doxycycline, J. Bio!. Chem., in
press.

Heegaard, C.W., Wiborg Simonsen, AC., Oka, K., Kj0l1er, L., Christensen, A., Madsen, B.,
Ellgaard, L., Chan, L., and Andreasen, PA, 1995, Very low density lipoprotein receptor
binds and mediates endocytosis of urokinase-type plasminogen activator-type-l
plasminogen activator inhibitor complex, J. Bioi. Chem. 270:20855-20861.

Holmes, W.E., Nelles, L, Lijnen, H.R and Collen, D., 1987, Primary structure of human 2-
antiplasmin, a serine protease inhibitor (Serpin),J. BioI. Chem. 262:1659-1664.

Karelina, T.v., Goldberg, G.I., and Eisen, AZ., 1995, Matrix metalloproteinases in blood
vessel development in human fetal skin and in cutaneous tumors, J. Invest. Dermato!.
105:411-417.

Kirchheimer, J. c., Nong, Y., and Remold, H. G., 1988, IFN- , tumor necrosis factor- , and
urokinase regulate the expression of urokinase receptors on human monocytes, J.
Immunol.,141:4229-4234.

Klagsbrun, M ., and D'Amore, P. A., 1991, Regulators of angiogenesis, Ann. Rev. Physiol.
53:217-239.

Koch, A E., L. A Harlow, G. K. Haines, E. P Amento, E. M. Unemori, W. L. Wong, R.

M. Pope, and Ferrara, N. E., 1994, Vascular endothelial growth factor - A cytokine

255
modulating endothelial function in rheumatoid arthritis, J. Immunol. 152:4149-4156.

Koolwijk, P., van Erck, M.G.M., de Vree, WJ.A., Vermeer, M.A., Weich, H.A.,
Hanemaaijer, R. and van Hinsbergh, V.W.M., 1996, Cooperative effect ofTNF , bFGF and
VEGF on the formation of tubular structures of human microvascular endothelial cells in a
fibrin matrix. Role of urokinase activity,J. Cell Bioi. 132: 1177-1188

Langer, D. 1., Kuo, A., Kariko, K., Ahuja, M., Klugherz, B. D., Ivanics, K. M, Hoxie, I.
A., Williams, W. V., Liang, B. T., Cines, D. B., and Barnathan, E. S., 1993, Regulation of
the endothelial cell urokinase-type plasminogen activator receptor - Evidence for cyclic
AMP-dependent and protein kinase-C dependent pathways, Cire. Res., 72:330-340.

Leibovich SJ., Polverini, PJ., Shepard, H.M., Wiseman, D.M., Shively, V. and Nuseir, N.,
1987, Macrophage-induced angiogenesis is mediated by tumour necrosis factor- , Nature
329:630-632.

Liotta, L.A., Steeg, P.S., and Stetler-Stevenson, W.G., 1991, Cancer metastasis and
angiogenesis - An imbalance of positive and negative regulation, Cell, 64:327-336.

Loskutoff, DJ., 1991, Regulation of PAI-l gene expression, Fibrinolysis 5: 197-206.

Madri,1. A., L. Bell, M. Marx, 1. R. Merwin, C. Basson, and Prinz, c., 1991, Effects of
soluble factors and extracellular matrix components on vascular cell behaviour invitro and
invivo - Models of de-endothelialization and repair, J. Cell. Bioehem. 45: 123-130.

Mandriota, S., Seghezzi, G., Vassalli, I-D., Ferrara, N., Wasi, S., Mazzieri, R., Mignatti, P.
and Pepper, M., 1995, Vascular endothelial growth factor increases urokinase receptor
expression in vascular endothelial cells, J. Bioi. Chem. 270:9709-9716.

Matrisian, L. M., 1992, The Matrix-degrading metalloproteinases, BioEssays 14:455-463.

Mignatti, P., Mazzieri, R., and Rifkin, D. B., 1991, Expression of the urokinase receptor in
vascular endothelial cells is stimulated by basic fibroblast growth factor, J. Cell BioI.
113:1193-1201.

Miles, L.A., Levin, E.G., Plescia, 1., Collen, D. and Plow, E.F., 1988, Plasminogen
receptors, urokinase receptors, and their modulation on human endothelial cells, Blood
72:628-635.

Miles, L. A., Dahlberg, C. M., Levin, E. G., and Plow, E. F., 1989, Gangiiosides interact
directly with plasminogen and urokinase and may mediate binding of these fibrinolytic
components to celis, Bioehem. 28:9337-9343.

Min, H.Y., Doyle, LV, Vitt, C.R., Zandonella, c.L., Stratton-Thomas, 1.R., Shuman,
M.A., Rosenberg, S., 1996, Urokinase receptor antagonists inhibit angiogenesis and
primary tumor growth in syngeneic mice, Cancer Res. 56:2428-2433.

Montesano, R., 1992, Regulation of angiogenesis in vitro, Eur. J. Clin. Invest. 22:504-515.

256
Moscatelli, D., 1985, Collagenase and plasmin activator production by blood
vessel-associated cells in response to angiogenic preparations, in: Intracellular Protein
Catabolism (E A Khairallah, J S Bond and J W C Bird Eds.), pp. 669-671, A. R Liss New
York.

Moses, M. A., Sudhalter, 1., and Langer, R., 1992, Isolation and characterization of an
inhibitor of neovascularization from scapular chondrocytes, J. Cell BioI. 119: 475-482.

Murphy, G., and Docherty, A. 1. P., 1992, The matrix metalloproteinases and their
inhibitors, Am. J. Resp. Cell Mol. Bioi. 7: 120-125.

Nachman, RL., 1992, Thrombosis and atherogenesis: molecular connections, Blood

79: 1897-1906.

Nagase, H. 1994. Matrix metalloproteases- a mini-review, in: Extracellular Matrix in the

Kidney (H. Kiode, and T. Hayashi, eds.), vol. 107, pp. 85-93, Contrib Nephrol. Karger,
Basel.

Nicosia, RF., and Ottinetti, A., 1990, Modulation of microvascular growth and
morphogenesis by reconstituted basement membrane gel in 3-dimensional cultures of rat
aorta - A comparative study of angiogenesis in matrigei, collagen, fibrin, and plasma clot, In
Vitro Cellular & Developmental BioI. 26: 119-128.

Niedbala, M. 1., and Stein-Picareila, M., 1992, Tumor necrosis factor induction of
endothelial cell urokinase-type plasminogen activator mediated proteolysis of extracellular
matrix and its antagonism by -interferon, Blood 79: 678-687.

Nikkari S.T., O'Brien, KD., Ferguson, M., Hatsukami, T., Weigus, H.G., Alpers, C.E., and
Clowes, A.W., 1995, Interstitial collagenase (MMP-l) expression in human carotid
atherosclerosis, Circulation 92: 1393-1398.

Nykjrer, A., Petersen, C. M., M0ller, B., Jensen, P. H., Moestrup, S. K, Hoitet, T. L.,
Etzerodt, M., Thogersen, H. c., Munch, M., Andreasen, P. A., and Gliemann, J., 1992,
Purified 2-macroglobulin receptorlLDL receptor-related protein binds urokinase activator
inhibitor type-l complex - Evidence that the 2-macrogiobulin receptor mediates cellular
degradati on of urokinase receptor-bound compl exes, J. BioI. Chern. 267: 14543 -14546.

Olson, D., P6l1anen, 1., H0yer-Hansen, G., R0nne, E., Sakaguchi, K, Wun, T-C., Appella,
E., Dan0, K and Blasi, F., 1992, Internalization of the urokinase-plasminogen activator
inhibitor type-l complex is mediated by the urokinase receptor, J. Bioi. Chern. 267:9129-
9133.

Passaniti, A., Taylor, R. M., Pili, R., Guo, Y, Long, P. V., Haney, 1. A., Pauly, R. R,
Grant, D. S., and Martin, G. R, 1992, A simple, quantitative method for assessing
angiogenesis and anti angiogenic agents using reconstituted basement membrane, heparin,
and fibroblast growth factor, Lab. Invest. 67:519-528.

Pendas, A., Knauper, v., Puente, X, Llano, E., Mattei, M., Apte, S., Murphy, G., and
Lopezotin, C., 1997, Identification and characterization of a novel human matrix

257
metalloproteinase with unique structural characteristics, chromosomal location, and tissue
distribution, J. BioI. Chem. 272: 4281-4286.

Pepper, M. S., Vassalli, J-D., Montesano, R., and Orci, L., 1987, Urokinase-type
plasminogen activator is induced in migrating capillary endothelial cells, J. Cell Bioi.
105:2535-2541.

Pepper, M. S., Belin, D., Montesano, R., Orci, L., and Vassalli, J., 1990, Transforming
growth factor-B-1 modulates basic fibroblast growth factor induced proteolytic and
angiogenic properties of endothelial cells in vitro, J. Cell BioI. 111:743-755.

Pepper, M. S., Ferrara, N., Orci, L., and Montesano, R., 1991, Vascular endothelial growth
factor (VEGF) induces plasminogen activators and plasminogen activator inhibitor-1 in
microvascular endothelial cells, Biochem. Biophys. Res. Commun. 181:902-906.

Pepper, M. S., Ferrara, N., Orci, L., and Montesano, R., 1992, Potent synergism between
vascular endothelial growth factor and basic fibroblast growth factor in the induction of
angiogenesis in vitro, Biochem. Biophys. Res. Commun.189:824-831.

Pepper, M. S, Sappino, A-P., Stocklin, R., Montesano, R., Orci, L., and Vassalli, J-D.,
1993, Upregulation of urokinase receptor expression on migrating endothelial cells, J. Cell
Bioi. 122:673-684.

Plate, K. H., G. Breier, H. A Weich, and Risau, W., 1992, Vascular endothelial growth
factor is a potential tumour angiogenesis factor in human gliomas in vivo, Nature
359:845-848.

Ploug, M., Behrendt, N., Lober, D., and Dan0, K., 1991, Protein structure and membrane
anchorage of the cellular receptor for urokinase-type plasminogen activator, Seminars in
Thromb. Haemostas. 17:183-193.

Plow, E.F., Felez, J. and Miles, L.A, 1991, Cellular regulation of fibrinolysis, Thromb.
Haemostas. 66:32-36.

P6llanen, J., Hedman, K., Nielsen, L. S., Dan0, K., and Vaheri, A, 1988, Ultrastructural
localization of plasma membrane-associated urokinase-type plasminogen activator at focal
contacts, J. Cell Bioi. 106: 87-95.

Polverini, P., 1989, Macrophage-induced angiogenesis - A review, Macrophage-Derived

Cell Regulatory Factors, 1:54-73.

Presta, M., Maier, J. A M., and Ragnotti, G., 1989, The mitogenic signalling pathway but
not the plasminogen activator-inducing pathway of basic fibroblast growth factor is
mediated through protein kinase C in fetal bovine aortic endothelial cells, J. Cell BioI.
109: 1877-1884.

Puente, X.S., Pendas, A.M., Llano, E., Velasco, G., Lopezotin, c., 1996, Molecular
cloning of a novel membrane-type matrix metalloproteinase from a human breast
carcinoma, Cancer Res. 56:44-949

258
Quigley, 1. P., Gold, L. I., Schwimmer, R., and Sullivan, L. M ., 1987, Limited cleavage of
cellular fibronectin by plasmin activator purified from transformed cells, Proc. Natl. Acad.
Sci. USA 84: 2776-2780.

Rao, N. K., Shi, G-P., and Chapman, H. A., 1995, Urokinase receptor is a multifunctional
protein: Influence of receptor occupancy on macrophage gene expression, J. Clin. Invest.
96:465-474.

Rao, 1.S., Yamamoto, M., Mohaman, S., Gokaslan, Z.L., Stetler-Stevenson, W.G., Roa,
VH., Liotta, L.A., Nicolson, G.I.,and Sawaya, R.E., 1996, Expression and localization of
92 kD type IV collagenase genatinase B (MMP-9) in human gliomas, Clin. Exp. Metastasis
14:12-18.

Redlitz, A., Tan, A. K., Eaton, D. L., and Plow, E. F., 1995, Plasma carboxypeptidases as
regulators of plasminogen system, J. Clin. Invest. 96: 2534-2538.

Saksela, O.D., Moscatelli, D. and Rifkin, D., 1987, The opposing effects of basic fibroblast
growth factor and transforming growth factor beta on the regulation of plasminogen
activator activity in capillary endothelial cells, J.Cell. Bioi. 105:957-963.

Sato, H., Takin, T., Okada, Y., Cao, 1., Shinagawa,A., Yamamoto, E., and Seiki, M., 1994,
A matrix metalloproteinase expressed in the surface of invasive tumour cells, Nature
370:61-65.

Sato, H., and Seiki, M., 1996, Membrane-type matrix metalloproteinases (MT -MMPs) III
tumor metastasis, J. Biochern. 119:209-215.

Schleef, R. R., Bevilacqua, M. J., Sawdey, M., Gimbrone, M. A., and Loskutoff, D. 1.,
1988, Cytokine activation of vascular endothelium. Effect on tissue-type plasminogen
activator and type 1 plasminogen activator inhibitor, J. BioI. Chern. 263:5797-5803.

Senger, D. R., L. Vandewater, L. F. Brown, J. A. Nagy, K. T. Yeo, T. K. Yeo, B. Berse, R.

W. Jackman, A. M. Dvorak, and Dvorak, H. F., 1993, Vascular permeability factor (VPF,
VEGF) in tumor biology, Cancer and Metastasis Rev. 12:303-324.

Shweiki, D., A. Hin, D. Soffer, and. Keshet, E., 1992, Vascular endothelial growth factor
induced by hypoxia may mediate hypoxia-initiated angiogenesis, Nature 359:843-845.

Sprengers, E.D. and Kluft, c., 1987, Plasminogen activator inhibitors, Blood 69:381-387.

Takino, T., Sato, H., Shinagawa, A., and Seiki, M., 1995, Identification of the second
membrane-type matrix metalloproteinase (MT -MMP- 2) gene from a human placenta
cDNA library - MT-MMPs form a unique membrane- type subclass in the MMP family,
J. Bioi. Chern. 270:23013-23020.

Tsuboi, R., Sato, Y., and Rifkin, D. B., 1990, Correlation of cell migration, cell invasion,
receptor number, proteinase production, and basic fibroblast growth factor levels in
endothelial cells, J. Cell Bioi. no: 511-517.

259
Unemori, E. N., Bouhana, K. S., and Werb, Z., 1990, Vectorial Secretion of Extracellular
Matrix Proteins, Matrix- Degrading Proteinases, and Tissue Inhibitor of
Metalloproteinases by Endothelial Cells, J. Bioi. Chem . 265:445-451.

Unemori, E . N ., Ferrara, N ., Bauer, E. A., and Amento, E. P., 1992, Vascular endothelial
growth factor induces interstitial collagenase expression in human endothelial cells, J. Cell.
Physiol. 153:557-562.

Van Hinsbergh, V.W.M., Kooistra, T., Van den Berg, E.A., Princen, H.M.G., Fiers, W . and
Emeis, J.1., 1988, Tumor necrosis factor increases the production of plasminogen activator
inhibitor in human endothelial cells in vitro and in rats in vivo, Blood 72: 1467-1473.

Van Hinsbergh, V. W. M., van den Berg, E. A., Fiers, W., and Dooijewaard, G., 1990,
Tumor necrosis factor induces the production of urokinase-type plasminogen activator by
human endothelial cells, Blood 75: 1991-1998.

Van Hinsbergh, V. W. M., 1992, Impact of endothelial activation on fibrinolysis and local
proteolysis in tissue repair, Ann. N. Y. Acad. Sci. 667:151-162.

Van Hinsbergh, V.W.M., Vermeer, M ., Koolwijk, P., Grimbergen, l, and Kooistra, T.,
1994, Genistein reduces tumor necrosis factor -induced plasminogen activator inhibitor-l
transcription but not urokinase expression in human endothelial cells, Blood 84:2984-2991.

Vassalli, J-D., 1994, The urokinase receptor, Fibrinolysis 8:172-181.

Volpert, 0., Ward, W ., Lingen, M., Chesler, L., Solt, D., Johnson, M., Moiteni, A.,
Polverini, P., and Bouck, N., 1996,Captopril inhibits angiogenesis and slows the growth of
experimental tumors in rats, J. Clin. Invest. 98:671-679.

Wallen, P., 1987, Structure and Function of Tissue Plasminogen Activator and Urokinase,
in: Fundamental and Clinical Fibrinolysis (P.1. Castellino, P.1. Gaffney, M.M. Samama,
and A. Takada, eds.), pp. 1-18, Elsevier, Amsterdam.

Wei, Y, Waltz, D., Rao, N., Drummond, R., Rosenberg, S., and Chapman, H., 1994,
Identification of the urokinase receptor as cell adhesion receptor for vitronectin, J. Bioi.
Chem. 269:32380-32388.

Weinberg, lB., Pippen, A.M.M and Greenberg, C.S., 1991, Extravascular fibrin formation
and dissolution in synovial tissue of patients with osteoarthitis and rheumatoid arthritis,
Arthitis Rheum. 34:996-1005 .

Will, H., and Hinzmann, B., 1995 eDNA sequence and mRNA tissue distribution of a
novel human matrix metalloproteinase with a potential transmembrane segment, Eur. J.
Biochem.231:602-608.

Will, H ., Atkinson, S., Butler, G., Smith, B., and Murphy, G., 1996, The soluble catalytic
domain of membrane type 1 matrix metalloproteinase cleaves the propeptide of
progeJatinase a and initiates autoproteolytic activation - regulation by TIMP- 2 and
TIMP-3, J. Bioi. Chem. 271:17119-17123.

260
Woessner, J. F., 1991, Matrix metalloproteinases and their inhibitors in connective tissue
remodelling, FASEB J. 5:2145-2154.

Wyne, K.L., Pathak, RK., Seabra, M.C., and Hobbs, H.H., 1996, Expression of the VLDL
receptor in endothelial cells, Arterioscl. Thromb. Vase. BioI. 16:407-415.

Zucker, S., Conner, c., Dimassmo, B., Ende, H., Drews, M., Seiki, M., and Bahou, W.,
1995, Thrombin induces the activation of progelatinase a in vascular endothelial cells -
physiologic regulation of angiogenesis, J. Bio!. Chem. 270:23730-23738.

261
CEILULAR AND MOLECUlAR EFFECTS OF THROMBIN IN THE VASCULAR
SYSTEM.

Chryso Kanthou, Vijay Vir Kakkar and Omar Benzakour.

Thrombosis Research Institute, Emmanuel Kaye Building, Manresa Road,

London SW3 6LR, u.K.

1. INTRODUCTION

Two distinct physiological processes lead to the development of new blood vessels:
vasculogenesis and angiogenesis [reviewed by Risau, 1997]. Whereas vasculogenesis is
restricted to embryonic development and consists in the differentiation of mesodermal
precursor cells into endothelial cells followed by their organisation into the capillary plexus,
angiogenesis is the formation of new capillaries from pre-existing blood vessels and takes
place both during prenatal and adult life. Angiogenesis can occur under both physiological
and pathological conditions such as during wound healing and solid tumour development.
Moreover, intra-arterial angiogenesis is evident in atherosclerotic plaques and in recanalised
thrombi [reviewed by Eisenstein, 1991]. The cellular and molecular events that lead to
angiogenesis are not as yet fully elucidated but are known to include (i) breakdown of the
extracellular matrix and of basement membrane of pre-existing blood vessels; (ii) the
migration and proliferation of endothelial cells; (iii) production of extracellular matrix
allowing the reconstitution of the basement membrane; (iv) recruitment of pericytes and
vascular smooth muscle cells (VSMC) to the newly formed blood vessel.
Thrombin, which is generated during the activation of the coagulation cascade is believed to
playa key role in angiogenesis, possibly through its ability to stimulate the migration and
proliferation of endothelial cells [Carney, 1992; Tsopanoglou et al, 1993]. Thrombin also
activates and stimulates the migration and proliferation of VSMC, pericytes and
inflammatory cells [reviewed by Kanthou and Benzakour, 1995], all of which support the
angiogenic process. Moreover, the generation of both promoters and inhibitors of
angiogenesis as well as the stimulation of extracellular matrix production and degradation by
the cell types described above are events known to be modulated by thrombin.
We will first summarise some general notions concerning the generation and structure of
thrombin, the physiopathology of the arterial wall, the nature of thrombin receptors and
activated signalling pathways. The effects of thrombin on some of the major processes

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 263
relating to angiogenesis, including cell proliferation and migration, modulation of
growth factor expression and activity will be discussed. Based on the above we will
speculate on the possible targeting of thrombin as an anti angiogenic approach.

2. ALPHA-THROMBIN GENERATION, STRUCTURE, SUBSTRATES AND

INHIBITORS.

Prothrombin, an inactive proenzyme, is synthesised in the liver and is secreted into the
circulation as a 579 amino acid protein (72 kDa) [reviewed by Mann, 1994]. In the
penultimate step of blood coagulation, enzymatically active a-thrombin is generated by
factor Xa which cleaves two successive bonds of the prothrombin molecule. a-Thrombin
(36 kDa) consists of two chains linked by a disulphide bond: the A chain (36 amino acid
residues) and the B chain (259 amino acid residues) where the serine proteinase domain
resides. Efficient conversion of prothrombin to a-thrombin requires the assembly of a
"prothrombinase complex" which consists of factor Xa, cofactor Va, calcium ions and a
phospholipid surface.
We have recently shown that at least partial prothrombin activation can also be achieved
by a serine protease secreted by cultured human VSMC which may constitute an
alternative pathway to the coagulation cascade [Benzakour et ai, 1995]. Indeed, incubation
of prothrombin with cultured human VSMC or serum-free medium conditioned by these
cells, results in limited proteolysis of this zymogen. A minor but significant enzymatic and
mitogenic activity similar to active thrombin is generated by this cleavage. Cleavage of
prothrombin was analysed by immunoblotting and N-terminal sequencing and was found to
occur at some sites in the prothrombin molecule normally cleaved during activation by
factor Xa. However, cleavage at R320_I321 which, during prothrombin activation produces
two-chain a-thrombin, was not detected by the methods used above although it is likely
that it occurred at lower levels. The use of specific inhibitors for the various classes of
proteases as well as neutralising antibodies to candidate serine proteases suggest that an as
yet unidentified VSMC-secreted serine protease, is responsible for prothrombin cleavage.
The hypothesis that some of the coagulation and fibrinolytic factors may be converted at
the cell surface by proteolytic cleavage to mitogenically active molecules is attractive.
Indeed, several laboratories have described a cell derived prothrombin cleaving activity
similar to the one we reported [Benezra et ai, 1993; Sekiya et ai, 1994]. Whether the
observed prothrombin cleavage by human VSMC implies the generation of an activity
which stimulates mitogenically dormant VSMC and leads to fibrin generation or, in
contrast, represents a mechanism for removing the activation peptide of prothrombin, and
therefore represents a clearance mechanism, remain to be investigated.
Tryptic or autocatalytic cleavage of a-thrombin which disrupts the anion binding exosite
results in the formation of g-thrombin. This exosite disruption abolishes a-thrombin's
affinity for fibrin and hence leads to complete loss of clotting activity. The major function
of thrombin in haemostasis is the conversion of fibrinogen to fibrin monomers that
spontaneously polymerise to form fibrin clots [Mann, 1994]. Thrombin also activates
factors V, Vill, XIII and thus accelerates its own formation. In a negative feedback
mechanism, thrombin through its interactions with membrane associated thrombomodulin
activates protein C which is a potent inhibitor of blood coagulation [Esmon, 1993].
Thrombin's enzymatic activity is controlled in vivo mainly by the serpins (serine
proteinase inhibitors) antithrombin-ill (AT-ill), heparin cofactor II, a2-macroglobulin and
protease nexins [Mann, 1994]. Hirudin which is derived from leeches is a potent and

264
specific natural thrombin inhibitor and interacts with both the active site and anion
binding exosite of thrombin. Several synthetic thrombin inhibitors which interact and
inactivate thrombin's catalytic site such as the tripeptide D-Phe-Pro-Arg
chloromethylketone (PP ACK) and diisopropyl fluorophosphate (DIP-F) have been
developed [Taparelli et al, 1993]. Natural and synthetic inhibitors directed against specific
domains of thrombin have been powerful tools for elucidating the contribution of regions of
the thrombin to its function in blood coagulation and haemostasis.

3. PHYSIOPATHOLOGY OF THE ARTERIES.

The normal adult artery consists of three morphologically distinct layers. The intima is
formed by a single continuous layer of endothelial cells lining the luminal side and a
peripheral internal elastic lamina which is a fenestrated sheet of elastic fibres. Extracellular
connective tissue, occasional VSMC and matrix are found between these two boundaries.
The media or middle layer consists entirely of layers of VSMC surrounded by collagen,
elastic fibres and proteoglycans. An external elastic lamina, separates the media from the
adventitia which is the outermost layer of the vessel and consists primarily of fibroblasts
with some VSMC and connective tissue. During embryonic development blood vessels
arise from the mesenchyme as a budding network of endothelial-lined channels [Sethi and
Brookes, 1971). These endothelial channels become surrounded by mesenchymal cells
which at first lack VSMC specific features. The genes and the signalling pathways which
control the differentiation of VSMC during embryonic development are not yet fully
understood. At the time of birth, most arteries contain their adult number of layers of
differentiated VSMC and any further intimal thickening is due to further proliferation and
production of connective tissue [Gerrity and Cliff, 1975]. The uninjured adult blood vessel
wall is a quiescent tissue with a very low VSMC replication index [Lombardi et al, 1991).
Normal adult vessels have a microvasculature that is confined to the adventitia and the
outer media, which extends to the intima when associated with atherosclerosis [Eisenstein,
1991).
Atherogenesis is a multifactorial and progressive process leading to the formation of
stenotic and obstructive lesions of the arteries [reviewed by Benzakour and Kanthou,
1995). In the normal intact vessel, the endothelium forms a transport barrier and provides a
non-thrombogenic and non-adhesive surface [Pearson, 1994]. Changes in endothelial
permeability following injury result in the appearance of cell surface adhesive glycoproteins
such as selectins, intracellular adhesion molecules and vascular adhesion molecules
[reviewed by Ross, 1993]. One of the earliest events in the development of atherosclerotic
lesions is the adherence of monocytes and T lymphocytes to the endothelium and their
migration beneath the arterial surface, in response to chemoattractants. Monocytes can
then differentiate into macrophages which accumulate lipids and thus become foam cells
leading to the development of the earliest recognised form of lesions known as "fatty
streaks". Mitogenic and chemotactic molecules are released by the injured endothelium, by
platelets, by the adherent leukocytes and the VSMC resulting in the migration and
proliferation of medial VSMC into the intima and the formation of a neointima which is a
characteristic of intermediate lesions. More advanced lesions ultimately form the "fibrous
plaques", which project into the arterial lumen and encroach the blood flow. These lesions
contain VSMC embedded in dense connective tissue surrounding a core of lipid, necrotic
debris and calcified material. Macrophages and other inflammatory cells are found at the
site of the necrotic core. A much accelerated version of this proliferative process appears

265
in restenosis which frequently occurs in patients who have undergone heart
transplantation, coronary graft bypass and percutaneous transluminal coronary angioplasly
(PTCA).
A significant feature of atherosclerotic plaques or restenotic lesions is the presence of
intraplaque microvessels which also penetrate and recanalise intra-arterial thrombi [Barger
et al, 1984]. Little is known about the origins and mechanisms responsible for the
formation of these intra-arterial microvessels and how these microvessels contribute to
lesion formation and to the pathology of the arteries. However, intra arterial microvessels
are likely to have a dual role in plaque pathology. First, the plaque microvasculature may
indirectly promote lesion progression by allowing the entry of inflammatory cells and
maintaining VSMC of the plaque in a biosynthetically active state via the provision of
nutrients and oxygen. Intra-arterial microvessels supplying the atheroma may haemorrhage
and lead to the formation of further thrombi which can become occlusive. Secondly, intra-
arterial microvessels could represent a mechanism of re-establishment of blood flow in
occluded vessels, a process which may be beneficial.

4. EFFECTS OF THROMBIN ON THE VESSEL WALL FUNCTION.

Under physiological conditions, the endothelium constitutes a thromboresistant cell layer

with anticoagulant and profibrinolytic properties [Pearson, 1994]. Thrombin activity is
downregulated by the intact endothelium through several mechanisms. Thrombomodulin, a
membrane glycoprotein expressed by endothelial cells, binds thrombin and neutralises its
proteolytic activity [Esmon, 1993]. Sulphated proteoglycans synthesised by endothelial
cells are incorporated into their extracellular matrix and are postulated to accelerate
thrombin inactivation by AT-III and heparin cofactor-II [Pearson, 1994]. Perturbation of
the endothelium results in a shift in the balance between its anticoagulant and
profibrinolytic properties towards procoagulant activities resulting in the generation of
thrombin [pearson, 1994]. Thrombin upregulates the expression of procoagulant molecules
such as tissue factor and plasminogen activator inhibitor-l (PAl-I) in endothelial cells.
Tissue factor, a transmembrane glycoprotein, binds and activates factor VII which catalyses
the activation of factor X and leads to thrombin generation. PAI-l inhibits tissue type
plasminogen activator (t-PA) and urokinase type plasminogen activator (u-PA) and thus
attenuates fibrinolysis. Thrombin also increases the permeability of endothelial cells and
mediates the expression of adhesion proteins which attract inflammatory cells [Garcia et al,
1992]. Several lines of evidence provide a link between thrombin and the inflammatory and
proliferative processes that occur during atheroma development. Davis et al, 1989, have
demonstrated that thrombus material is often found overlying plaque fissures or entrapped
in plaques which have healed and resealed thus implying that active thrombin is being
generated. Thrombin activity is thought to persist for considerable periods of time in the
vessel as clot entrapped thrombin is protected from inactivation by circulating inhibitors
such as AT-III and may be released slowly following clot lysis [Weitz et al, 1990].
Accordingly, tissue factor procoagulant activity and thrombin generation are detected
during experimental vascular injury [Hatton et al, 1989]. In such experimental models,
treatment of animals with thrombin inhibitors [Walters et al, 1994] or transfection of
injured blood vessels with adenovirus expressing the thrombin inhibitor hirudin [Rade et al,
1996] significantly reduced restenosis. In normal uninjured vessels the expression thrombin
receptor, PAR-I, [Vu et al, 1991] is exclusively localised in the endothelium. After

266
experimental injury or in atherosclerotic tissues the expression of this
receptor increases significantly both in the medial and intimal layers of the vessel [Wilcox et
ai, 1994].

5. THROMBIN RECEPTORS.

The capacity of thrombin to stimulate cells and activate various signalling pathways led to
the search for a thrombin receptor. A cDNA encoding for a functional thrombin receptor
(protease activated receptor 1, PAR-I) was originally isolated from a magakaryocytic cell
line and endothelial cells [Vu et ai, 1991; Rasmussen et al, 1991]. This 3.5 kb long cDNA
comprises a 224bp 5' non-coding region, 1275 bp encoding a 425 amino acid protein and
1980 bp 3' non coding region [Vu et al, 1991]. The receptor is a member of the seven
transmembrane domain receptor family coupled to heterotrimeric guanine nucleotide-
binding proteins ( G proteins) [Vu et al, 1991; Rasmussen et al, 1991]. A novel receptor
activation mechanism was described which involves the binding of thrombin to the
receptor's extracellular N-terminal domain which is then cleaved by thrombin's proteolytic
activity at the specific sequence LDPR-SFLL. This process unmasks a new amino
terminus which activates the receptor by functioning as a "tethered" peptide ligand. A
sequence in the extracellular domain of the receptor, carboxyl to the thrombin cleavage site,
resembles the C-terminal tail of hirudin that interacts with the anion binding exosite of
thrombin, and is thought to act as a binding site for thrombin. The above mechanism of
thrombin receptor activation is supported by evidence obtained from mutant receptors,
synthetic peptides mimicking the newly unmasked amino terminus or the hirudin-like
domain of the receptor and studies using monoclonal antibodies directed against specific
epitopes of the receptor's N-terminus [Vu et ai, 1991; Brass et al, 1992; Bahu et al, 1993].
Synthetic SFLLRN peptides of the new amino terminus termed "thrombin receptor
activating peptides" (TRAP) possess agonist activity similar to thrombin on many cell
types [Vu et ai, 1991; Ngaiza et al, 1991; Vouret-Craviari et ai, 1992; Hung et al, 1992].
These peptides bypass the requirement for receptor proteolysis prior to receptor
activation. The use of Xenopus/human thrombin receptor chimeras expressed in Xenopus
oocytes has allowed the identification of regions in the receptor's second extracellular loop
and N-terminal extension that are important for its activation [Gerszten et ai, 1994].
It is thought that because thrombin acts as an enzyme rather than a classical ligand,
eventually all cell surface receptors would be cleaved and therefore activated. Using
antibodies that distinguish between the native PAR-l and the cleaved form in rat 1
fibroblasts transfected with PAR-I, the rate of cleavage of the receptor was shown to be
proportional to the concentration of thrombin, even though low concentrations of thrombin
could eventually cleave all available receptors [Ishii et al, 1993]. It is proposed that target
cells detect different thrombin concentrations as different rates of receptor cleavage and
hence the yield of receptor activation depends primarily on the rate of receptor cleavage. In
a megakaryoblastic cell line CHRF-288 the ability of individual thrombin molecules to
activate multiple numbers of receptors appears to be limited thus indicating that different
mechanisms may apply to specific cell systems [Brass et ai, 1994].
Like other G-protein-coupled receptors, PAR-l are quickly desensitized after activation so
that a subsequent presentation of thrombin evokes little or no response. Present evidence
suggests that this loss of function may be due to a combination of events including receptor
phosphorylation, dephosphorylation, intemalisation, degradation and recycling. Indeed,
phosphatase inhibitors drastically reduce the recovery of thrombin responsiveness after

267
desensitisation in megakaryoblastic human erythroleukemia (HEL) cells [Brass,
1992]. Moreover, in rat 1 fibroblasts transfected with human PAR-I, thrombin stimulation
leads to a rapid receptor phosphorylation and this process is correlated with receptor
desensitisation [Ishii et ai, 1994]. In Xenopus oocytes, co-expression of the G protein
receptor-coupled kinase BARK2 and PAR-l results in blocked PAR-l signalling thus
suggesting a role for BARK2 or a related kinase in this receptor's desensitisation [Ishii et aI,
1994]. Receptor internalisation and degradation have also been proposed as means of
terminating PAR-l signalling [Brass et aI, 1994; Hoxie et ai, 1993]. However, the handling
and replacement of activated PAR-l seems to be distinct in different cell types. In
megakaryoblastic HEL and CHRF-288 cells, subsequent to thrombin treatment, most of the
thrombin receptors are internalised and degraded [Brass et ai, 1994; Hoxie et ai, 1993]. A
small proportion of internalised receptors return to the cell surface but are no longer
responsive to thrombin. In these cells, replacement of the thrombin receptors seems to be
mainly due to neosynthesis. In human umbilical endothelial cells thrombin cleaves all
available receptors which are then internalised. Recovery of receptors to the cell surface is
fast and is paralleled by the regaining of responsiveness to thrombin [Woolkalis and Brass,
1994]. In these cells, re-establishment of responsiveness to thrombin does not require
protein synthesis and involves rather the mobilisation to the cell surface of pre-existing
cellular pools of non-activated thrombin receptors.
Knockout of the gene encoding PAR-l provided evidence for the existence of other
thrombin receptors [Connolly et aI, 1996]. Recently, a second thrombin receptor belonging
to the same family of receptors as PAR-I, termed PAR-3 has been cloned [Ishihara et ai,
1997]. This receptor shares a 27% amino acid sequence similarity to PAR-l and contains
an N-terminal thrombin cleavage site, LPIK-TFRG. PAR-3 mediates thrombin induced
phosphoinositide hydrolysis and 45Ca release in P AR-3 transfected Xenopus oocytes.
P AR-3 is expressed in a variety of tissues including human bone marrow and mouse
megakaryocytes. Its involvement in thrombin mediated effects in various cell types
remains to be elucidated.

6. SIGNALLING PATHWAYS ACTIVATED BY THROMBIN AND TRAP.

Thrombin receptors (PAR-I) are expressed by a variety of cell types including endothelial
cells, VSMC, fibroblasts, platelets and macrophages [Vu et aI, 1991; McNamara et ai, 1993;
Kanthou et ai, 1995a]. Thrombin acts as a cellular activator, mitogen and/or
chemoattractant for these cells and the signalling pathways involved in these processes
have been the main focus of numerous studies. The repertoire of G proteins coupled to the
thrombin receptor(s) and the cellular effectors activated, determine the nature of the cellular
responses generated. In most cell types thrombin stimulates phospholipases C (PLC) and
A2 (PLA2) [Brass et aI, 1986; Garcia et ai, 1992], protein kinase C (PKC) [Huang and Ives,
1987], mitogen activated protein (MAP) kinases [Kahan et ai, 1992], tyrosine kinases
[Molloy et ai, 1996] and modulates the activity of adenylate cyclases (AC) [Brass et ai,
1986; Magnaldo et ai, 1988]. Although it was originally thought that only the a subunits of
G proteins were responsible for effector activation, recent data shows that G protein bg
dimers can also mediate the activation of certain PLC and AC isoenzymes and MAP
kinases [Clapham and Neer, 1993].
Pertussis toxin (PTX), a bacterial toxin isolated from Bordetella pertussis that specifically
inactivates a subclass of G proteins (Gi) by covalent modification of their a subunit, has
been of a great use for evaluating the repertoire and the role of G proteins in thrombin

268
receptor signalling in various cell systems [reviewed by Obberghen-Schilling and
Pouyssegur, 1993]. Such studies originally established that the thrombin receptor(s)
interact with at least two subtypes of G proteins: a PTX-sensitive Gi subtype coupled to
AC and a PTX-insensitive Gq subtype coupled to PLC. It is now apparent that both these
families of G proteins can interact with both AC and PLC isoenzymes via their a and bg
subunits [Clapham and Neer 1993]. In platelets and VSMC, thrombin activates the PTX-
sensitive Gi proteins; activated Gia subunits inhibit the activity of AC, thus lowering
intracellular cAMP levels [Hung et al, 1992, Kanthou et al, 1992]. In endothelial cells
thrombin activates PLA2 resulting in the synthesis of prostacyclin (PGI2) which then
activates AC via G s protein [Garcia et al, 1992]. In fibroblasts thrombin exerts a
concentration dependent dual effect on AC activity: an inhibition via Gi protein at low
concentrations and a potentiation mediated by PKC at higher concentrations [Magnaldo et
ai, 1988]. In the various cell systems studied, synthetic TRAP mimic, to some extent, the
multiple effects of thrombin on AC activity and their sensitivity to PTX, hence implying
the involvement of P AR-l in these effects [Ngaiza and Jaffe, 1991; Vouret-Craviari et al,
1992; Hung et aI, 1992; Obberghen-Schilling and Pouyssegur, 1993; Kanthou et aI, 1995a].
In most responsive cells, thrombin activates PLC leading to the formation of inositol 1,4,5
triphosphate (IP3) and diacyl glycerol (DAG) [Brass et aI, 1986; Garcia et al, 1992; Huang
and Ives, 1987; Obberghen-Schilling and Pouyssegur, 1993]. IP3 elevates cyctosolic Ca2+
whereas DAG activates PKC. The mechanism(s) involved in PLC activation by thrombin
as deduced from the sensitivity to PTX differ according to the cell type and presumably
through the repertoire of G protein subtypes and PLC isoenzymes. In endothelial cells,
thrombin-induced activation ofPLC occurs in a PTX-insensitive manner and is thought to
be mediated via the Gq family of G proteins [Garcia et al, 1992]. In VSMC the activation
ofPLC is partially sensitive to PTX, whereas in platelets the PLC inhibition is complete in
the presence ofPTX [Brass et aI, 1986]. The sensitivity to PTX of thrombin-induced PLC
activation in platelets is thought to occur via a mechanism involving free bg dimers of PTX-
sensitive Gi. As for AC, thrombin-induced PLC activation is reproduced with TRAP in
various cell systems examined, thus implying the involvement of PAR-l [Vu et al, 1991;
Hung et aI, 1992; Obberghen-Schilling and Pouyssegur, 1993].
Protein phosphorylation and dephosphorylation reactions occur at serine threonine or
tyrosine residues. These reactions constitute major pathways for signal transduction.
Indeed, the stimulation of serine/threonine and tyrosine phosphorylations by thrombin
have been observed in several different cell systems [Kahan et al, 1992; Molloy et al,
1996]. Thrombin stimulation results in the modulation of second messengers such as
cAMP and Ca2+ which in turn affect the activity of protein kinases. MAP kinase and S6
ribosomal kinase are key enzymes involved in receptor mediated initiation of cellular
growth [Kahan et al, 1992]. Although the signalling events or effector systems that lie
between G protein coupled receptors and MAP kinase activation are not well understood it
is thought that this process involves activation of Ras proteins. The MAP kinase isoforms
p42 mapk and p44 mapk are phosphorylated on threonine and tyrosine residues in response
to various mitogens including thrombin [Kahan et aI, 1992]. In quiescent CCL39
fibroblasts activation of MAP kinase by thrombin occurs in a biphasic mode. An initial
rapid peak is followed by a persistent phase of stimulation which correlates with cell cycle
re-entry [Kahan et aI, 1992]. Synthetic SFLLRN peptides induce only the first initial
phase of MAP kinase activation [Vouret-Craviari et al, 1993]. In rat aortic VSMC
thrombin and TRAP, activated MAP kinase with similar kinetics to other mitogens such as
platelet-derived growth factor (PDGF). An initial peak of MAP kinase activity occurred at

269
5 min and was sustained above basal levels for up to 6 h [Molloy et al, 1996].
Thrombin also induces a rapid translocation of MAP kinases to the nucleus and this event
is thought to be a prerequisite for transcription factor activation and cell cycle re-entry
[Vouret-Craviari et al, 1993]. In response to thrombin, MAP kinase phosphorylates and
activates cytosolic PLA2 which is responsible for arachidonic acid release and synthesis of
prostaglandins and leukotrienes which in tum control vascular permeability and inhibit
smooth muscle cell proliferation [Lin et al, 1993]. In quiescent CCL39 cells, thrombin also
activates the S6 ribosomal kinase leading to the stimulation of protein synthesis [Kahan et
aI, 1992].

7. EFFECTS OF THROMBIN ON CELL PROLIFERATION AND MIGRATION.

Although permanently exposed to a whole range of growth factors, endothelial cells remain
out of the cell cycle in a quiescent state which implies the existence of an inhibitory
system(s) that prevents them from responding to mitogens and chemoattractants and from
proliferating. In vitro, endothelial cells are stimulated mitogenically by thrombin [Carney,
1992]. In particular microvascular endothelial cells respond mitogenically to thrombin and
a correlation was described between mitogenic responsiveness to thrombin and presence of
high affinity thrombin receptors [Belloni et al, 1992]. Thrombin was shown to induce the
proliferation of human umbilical vein endothelial cells (HUVEC) via both proteolytic and
non-proteolytic mechanisms [Herbert et ai, 1994]. The mitogenic effect of thrombin was
mimicked by TRAP and was inhibited by inhibitors of thrombin's enzymatic activity
indicating the involvement of a proteolytic mechanism. In addition a 14 amino acid peptide
derived from the sequence of the B-chain of thrombin was shown to be mitogenic for
HUVEC indicating the involvement of non-enzymatic pathways in mitogenesis [Herbert et
ai, 1994]. Thrombin, TRAP and the 14 amino acid peptide-induced growth were inhibited
by a P AR-l antisense oligonucleotide showing that the growth inducing effects of all these
compounds were mediated through the same thrombin receptor, PAR-l [Schaeffer et al,
1997]. The proliferation ofHUVEC induced by thrombin or TRAP was strongly inhibited
by antibodies against bFGF suggesting that thrombin-induced mitogenesis involves
autocrine pathways [Herbert et al, 1994]. Thrombin has also been shown to promote
endothelial cell migration in vitro [Wang et al, 1997].
By virtue of its ability to stimulate microvessel endothelial cell proliferation, thrombin is
now considered as an angiogenic factor [Maragoudakis, 1996; Carney, 1992]. Indeed
Tsopanoglou et ai, 1993, have demonstrated that thrombin is a potent angiogenic agent in
the chick chorioallantoic membrane (CAM) assay and does so in a manner which is
independent of fibrin generation. The angiogenic activity of thrombin was abolished by
hirudin which interacts with both thrombin's catalytic site and the fibrinogen recognition
site and by PPACK which inhibits thrombin's catalytic activity. TRAP also promoted
angiogenesis in the CAM system suggesting the involvement of P AR-1 in this effect. The
angiogenic properties of thrombin were reproduced in two other models of angiogenesis.
First, thrombin promoted the formation of capillary like tubes by HUVEC in matrigel in
vitro and secondly thrombin promoted the formation of blood vessels in a matrigel plug
injected subcutaneously into mice [Haralambopoulos et ai, 1997].
The studies of the mitogenic effects of thrombin on VSMC have produced somewhat
conflicting data. Berk et al, 1991, demonstrated that thrombin stimulates growth related
events but not DNA synthesis or proliferation whereas other investigators demonstrated
that thrombin elicits both growth related events and proliferation in VSMC [McNamara et

270
al, 1993; Huang and Ives, 1987; Kanthou et al, 1992; Bar-Shavit et al, 1990].
These contradictions may be accounted for by differences in experimental conditions,
species differences and the vascular sources from which VSMC were derived. Indeed, the
studies of Bar-Shavit et al, 1990 were done on bovine aortic VSMC, whereas Huang and
Ives, 1987, used neonatal rat VSMC and Berk et al, 1991 and McNamara et al 1993, used
adult rat aortic VSMC. We have performed all our studies on human VSMC [Kanthou et
al, 1992]. It is established that the peak of DNA synthesis in VSMC stimulated with
thrombin is delayed by approximately 8 h when compared to that induced by serum or
other direct acting mitogens [Kanthou et al, 1992; McNamara et al, 1993; Molloy et al,
1996]. It is likely, therefore, that Berk et al, 1991, failed to detect any mitogenic effect of
thrombin, because cells were labelled with 3H-thymidine, 20-24 h post stimulation, during
which time the induction of DNA synthesis by thrombin may not have yet began. It is
important to note that in some of these studies, defined serum-free media were used which
contained factors such as transferrin and insulin that might synergise with thrombin or act
as progression factors [Weiss and Maduri, 1993; McNamara et al, 1993]. In addition,
certain culture conditions may favour the constitutive expression of some growth factors
that may synergise with thrombin in inducing VSMC division. Mitogenesis in both human
and rat aortic VSMC is dependent on thrombin's catalytic site as inhibition with PP ACK,
hirudin or AT-III abolishes these effects [Kanthou et al, 1992; McNamara et al, 1993;
Kanthou et al, 1995b]. Gamma thrombin which retains enzymatic activity also induces
mitogenesis but does so with a reduced potency [McNamara et aI, 1993; Kanthou et al,
1992]. All VSMC systems examined have been shown to express PAR-l and TRAP
peptides mimic, at least in part, the mitogenic effects of thrombin implying the involvement
of this receptor in the transmission of mitogenic signals [McNamara et aI, 1993; Kanthou et
aI, 1995a].
Several early studies have shown thrombin to be a potent mitogen for cultured fibroblasts
from many species [Chen and Buchanan, 1975]. Fibroblasts express P AR-l and synthetic
TRAP have been shown to induce mitogenesis in these cells [Vouret-Craviari et al, 1992].
P AR-l appears to be necessary and sufficient for activation of mitogenesis in mouse
fibroblasts as the response to thrombin was selectively lost in fibroblasts derived from
P AR-l knockout mice [Trejo et aI, 1996]. Mitogenesis in fibroblasts was also reported to
be activated by a peptide corresponding to residues 178-200 of thrombin which was also
shown to accelerate wound healing in vivo [Stiernbert et aI, 1993].
Thrombin binds to monocytes and macrophages and elicits a wide range of responses
including cell migration and proliferation [reviewed by Bar-Shavit et aI, 1992]. Some
studies have shown that esterolytically inactivated thrombin is as effective as a-thrombin in
inducing macrophage migration which suggest that the proteolytic activity is not a
requirement for the chemotactic effect [Bar-Shavit et al, 1992]. A peptide, CB67-129,
which represents residues 33-84 of the human B-chain of thrombin, promotes significant
chemotaxis for both human peripheral blood monocytes and murine macrophages and
competes for the same binding site on the monocytes membrane as a-thrombin [Bar-Shavit
et al, 1992]. Several murine macrophage-like cell lines such as 1775 and P388Dl respond
mitogenically to a-thrombin, proteolytically inactive thrombin and the chemotactic peptide
CB67-129 described above. It is important to note here that the thrombin and peptide
mitogenic effect is observed only in transformed macrophages which unlike non-
transformed macrophages are colony stimulating factor-I-independent. Thrombin
synergises and potentiates the mitogenic effects of colony stimulating factor-l in non-
transformed macrophages. Monocytic cell lines such as U937 express PAR-I and respond
to synthetic SFLLRN peptides. Thrombin activation of these monocytic cells can be

271
entirely accounted for by a proteolytic mechanism of thrombin receptor
activation. The ability of thrombin to recruit and induce the proliferation of monocytes and
macrophages is of particular importance for several processes such as wound healing,
atherosclerosis and in angiogenesis.

8. EFFECTS OF THROMBIN ON GENE EXPRESSION AND GROWm

FACTOR ACTIVITIES.

The synthesis and secretion of growth factors, and chemoattractants are thought to be
important mechanisms for the regulation of the angiogenic process. Concording
experimental evidence supports the concept that growth factors stored in the extracellular
matrix and released upon its degradation are mediators of angiogenesis. Their association
with the matrix may represent a mechanism of sequestration of these factors at the cell
surface offering protection from degradation by circulating proteinases. Some groups of
growth factors including basic fibroblast growth factor (bFGF), transforming growth factor-
b (TGF -b), vascular endothelial growth factor (VEGF), and PDGF can be found in close
association with the extracellular matrix [Vlodavski et ai, 1987; Taipale et al, 1992]. bFGF
is found in the subendothelial extracellular matrix, bound to heparan sulphate proteoglycans
and protected from inactivation [Vlodavski et aI, 1987]. Exposure of subendothelial matrix
to thrombin results in the release of heparan sulphate proteoglycan-bound bFGF which is
biologically active and resistant to degradation. TGF-b, in its latent form, is also associated
with the extracellular matrix, is released by thrombin and activated by plasmin [Taipale et
al, 1992]. Some alternatively spliced forms of VEGF bind to extracellular matrix and
heparan sulphate proteoglycans and can be released by heparanases and plasmin [Houck et
ai, 1992]. Whether thrombin can release VEGF bound to extracellular matrix remains to be
investigated. PDGF, a potent mitogen and chemoattractant for mesenchymal cells can also
be found associated with extracellular matrix proteins. An alternatively spliced form of
PDGF which contains some basic C-terminal amino acids forming a 'retention motif' which
interacts with extracellular matrix heparan sulphate proteoglycans [Raines and Ross, 1992;
Andersson et ai, 1994]. The cell-associated long PDGF-A variant can be released from the
cells by synthetic peptides corresponding to the sequence encoded by exon 6 [Raines and
Ross, 1992].
Thrombin also mediates the expression and or secretion of growth factors by different cell
types. bFGF is induced by thrombin in VSMC [Weiss and Maduri, 1993]. and endothelial
cells [Ku and D'Amore, 1995] and is thought to be a key endogenous mediator of
thrombin's mitogenic activity in VSMC. TGF-b is also regulated by thrombin at the gene
and protein level in various cell types including VSMC [Bachhuber et al, 1997] and may
participate in autocrine activation of cell proliferation and migration. In both VSMC and
endothelial cells PDGF gene and protein expression are upregulated following thrombin
treatment [Kanthou et ai, 1995; Shankar et ai, 1994]. Although PDGF may not have direct
effects on endothelial cell migration and proliferation, it is thought to play a direct role in
angiogenesis by promoting the recruitment of mural cells, pericytes in small vessels and
VSMC in larger vessels in a paracrine fashion. PDGF is a dimeric molecule consisting of
homo- and hetero-dimers of two chains, PDGF-A and PDGF-B [reviewed by Benzakour
and Kanthou, 1996]. The receptor for PDGF consists of two subunits, a and b. PDGF is a
potent mitogen and chemoattractant for a variety of mesenchymal cells. The response of
any particular cell type is dependent on its receptor complement. The a subunit binds both
A and B chains whereas the b subunit binds only the B chain. Endothelial cells express

272
PDGF-B but have no detectable PDGF-b receptors whereas VSMC express
PDGF-b receptors at high levels and respond strongly to PDGF-BB.
The mitogenic effects of thrombin on VSMC are shown to be mediated at least in part via
the induction of synthesis of autocrine growth factors. Indeed, antibodies to bFGF abolish
the mitogenic effects of thrombin but not those of other growth factors such as PDGF-BB
in neonatal VSMC [Weiss and Maduri, 1993]. We have shown that antibodies to PDGF-
AA also result in a partial inhibition in thrombin induced mitogenesis in human VSMC
[Kanthou et ai, 1995]. In various VSMC species, thrombin induces a rapid increase in the
level of expression of many "immediate-early" genes such as c-fos and c-myc thus
suggesting that thrombin acts also as a direct mitogen [Berk et ai, 1991; Kanthou et al,
1992; Bar-Shavit et ai, 1992; Kanthou et al, 1995a]. During experimental arterial injury the
pattern of PDGF-A and PDGF receptor gene expression undergoes a rapid change
characterised by an increase in the level of PDGF-A mRNA and a rapid loss of both
PDGF-a and PDGF-b receptor transcripts. [Majesky et al, 1990]. Out of more than 12
agonists tested, thrombin and serotonin were the only agonists which reproduced, in
cultured VSMC, the injury pattern of PDGF ligand and receptor transcript expression
[Okazaki et ai, 1992]. The induction of c-fos and PDGF-A gene expression in human and
rat aortic VSMC is dependent on thrombin's catalytic site as inhibition with PP ACK or
hirudin abolishes these effects [Okazaki et ai, 1992; McNamara et al, 1993; Kanthou et al,
1995b]. The enzymatic activity of thrombin does not seem to be essential for the induction
of mitogenesis or c-fos expression in bovine VSMC as catalytically inactivated thrombin
forms induce both these effects [Bar-Shavit et al, 1990]. Whether these contradictory
findings represent species differences or are due to variability in experimental conditions
remains to be determined. Synthetic TRAP mimic thrombin-induced events such as c-fos
and PDGF expression and mitogenicity in several VSMC systems, hence implying the
involvement of PAR-1 in these processes [McNamara et ai, 1993; Kanthou et ai, 1995a].
The plasminogen activator/plasmin enzyme system has been implicated in extracellular
proteolysis events associated with cell migration and proliferation in relation to
angiogenesis [Pepper et al, 1996]. Both t-PA and u-PA have been shown to be potent
mitogens and chemoattractants for VSMC [Herbert et al, 1994; Kanse et ai, 1997; Okada et
al, 1996]. u-PA and its receptor u-PAR as well as t-PA and PAI-I are expressed by
endothelial cells, VSMC as well as some inflammatory cells. Thrombin regulates the
production and release of these components of the plasminogen activator system and may
thus modulate the angiogenic process in this respect. Both t-PA and u-PA as well as PAI-l
are upregulated in VSMC and endothelial cells in response to thrombin [Reuning et al,
1994; Wojta et ai, 1993; Noda-Heiny et al, 1992]. The u-PAR is also upregulated by
thrombin in VSMC but downregulated in endothelial cells [Reuning et ai, 1994; Li et al,
1995]. Urokinase bound to its receptor facilitates localised plasminogen activation and cell
migration and hence the regulation of this receptor by thrombin may modulate the
angiogenic process.

9. CONCLUDING REMARKS.

Some of the effects of thrombin on vessel wall function, cell proliferation and migration,
growth factor gene expression and synthesis, release and/or secretion have been discussed
above. Therefore in addition to its direct effects on endothelial cell migration and
proliferation, the capacity of thrombin to modulate gene expression may constitute a key
step in the regulation of vascular remodelling or angiogenesis in vivo. As thrombin activity

273
is essential for the maintenance of haemostasis, we raise the question whether it
would be possible to uncouple thrombin clotting activity from its proliferative/cellular
effects with an aim to specifically target the latter.
Numerous potent and well characterised thrombin inhibitors exist [reviewed by Taparelli et
ai, 1993]. However, such inhibitors suppress both thrombin's clotting and proliferative
activities, as thrombin-induced proliferation is mainly dependent on the presence of an
active catalytic site [Kanthou et ai, 1995b; McNamara et ai, 1993; Vu et ai, 1991]. The link
between thrombosis and vessel wall lesion development is well established as was
discussed above. It is, therefore, not possible using classical thrombin inhibitors to
determine whether the reduction in vessel wall lesion formation is due to the reduction in
the generation of thrombotic material and associated fibrin deposition or alternatively due
to inhibition of direct thrombin proliferative effects or a combination of both. The work of
Zoldhelyi et ai, 1994, demonstrating persistent thrombin generation in animal models
treated with hirudin, together with our in vitro data showing that the full thrombin
proliferative signal is transmitted within minutes after the addition of thrombin to human
VSMC (the time necessary for thrombin to cleave and activate its cell surface receptor(s),
highlight the limits of classical anti thrombotic strategies [Kanthou et ai, 1995b]. Moreover,
various laboratories including ours, have recently demonstrated that cultured cells secrete a
serine proteinase that can proteolytically cleave prothrombin with the generation of
thrombin activity independently of the coagulation cascade [Benzakour et ai, 1995; Sekiya
et ai, 1994]. Together these various experimental findings emphasise the need to elaborate
alternative pathways for the specific inhibition of thrombin proliferative effects. Such
alternative strategies could include the development of antagonists for the thrombin
receptor, PAR-I. However, the identification of a recently described second thrombin
receptor (PAR-3) [Ishihara et ai, 1997] that transmits at least some of the signals of
thrombin in different cell systems together with reports showing that inhibition of the
expression of PAR-l does not completely abolish thrombin cellular effects [Connolly et ai,
1996], demonstrate that PAR-l may not be the most appropriate component to target in
the chain of cellular events that are activated by thrombin.
In an earlier report [Kanthou et ai, 1992] we provided the first evidence for the link
between thrombin-induced mitogenicity in human VSMC and PDGF-A gene expression. In
three other recent reports, we have compared the signalling pathways involved in thrombin-
and TRAP-mediated c-fos, c-myc and PDGF-A gene induction in human VSMC and
investigated the structural domains of thrombin involved in these activities [Kanthou et ai,
1995a; Kanthou et ai, 1995b; Kanthou et ai, 1996]. Taken together, our data point out
clearly a primordial role for PDGF-A gene expression in thrombin-induced human VSMC
proliferation. Taken into account that PDGF being a potent chemotactic and mitogenic
factor may be considered as a proangiogenic factor, it will be extremely interesting to
investigate whether thrombin's angiogenic properties are mediated at least in part by
PDGF.
The molecular mechanisms underlying thrombin stimulation of transcription of the PDGF-
A chain gene or, in fact, thrombin induction of any gene in VSMC as well as in any other
cell types, have remained largely uninvestigated. Recently, Scarpati and DiCorieto, 1996,
reported the presence of thrombin responsive elements within PDGF-B promoter elements
and of a thrombin-inducible nuclear factor (TINF) in extracts from thrombin-treated, but
not control endothelial cells. This study was performed on endothelial cells, since under
physiological conditions, human VSMC transcribe the PDGF-B gene at only very low
levels. The thrombin responsive element consists of a repeat of CCACCC in an ABBA
configuration. Localisation of thrombin responsive consensus sequences within the PDGF-

274
A promoter region should lead to the identification of the transcnptIOn factor(s)
specifically activated by thrombin in human VSMC and which mediate the observed
increase in PDGF-A gene level. In the last decade various responsive elements were
identified within the promoter region of several genes and such consensus sequences appear
to confer specificity for the responsiveness to a given stimulus. Our computer searches
have identified the presence of a thrombin responsive sequence CCACCC in the promoter
elements of several genes which are known to be activated by thrombin, such as PDGF -A,
bFGF, u-PA, t-PA, PAI-1 etc and its absence in various other genes that are not known to
be modulated by thrombin. Antisense oligonucleotides directed to the thrombin responsive
consensus sequence would prevent the binding of TINF and hence the transcriptional
activation of thrombin responsive genes. Such a strategy would inhibit gene induction by
thrombin and hence suppress or at least reduce its involvement in various pathological
situations such as angiogenesis.

ACKNOWLEDGEMENTS.

This work was supported by the British Heart Foundation (PG951138 and PG961130), and
The Thrombosis Research Trust.

REFERENCES.

Andersson, M., Ostman, A., Westermark, B., Heldin, C-H., 1994. Characterisation of the
retention motif in the C-terminal part of the long splice form of platelet-derived growth
factor A-chain. 1. Bioi. Chem. 269: 926-930.

Bachhuber, B.G., Sarembock, U., Gimble LW., and Owens, G.K, 1997. Alpha thrombin
induces transforming growth factor beta-1 mRNA and protein in cultred vascular smooth
muscle cels via a proteolytically activated recetpro. 1. Vasco Res. 34:41-48.

Bahou, W.F., Coller, B.S., Potter, c.L., Norton KJ., Kutok J.L., and Goligorsky, M.S.,
1993. The thrombin receptor extracellular domain contains sites crucial for peptide ligand-
induced activation. 1. Clin. Invest. 91: 1405-1413.

Bar-Shavit, R., Benezra, D., Eldor, A., Hy-Am, E., Fenton, J.W., II, and Vlodavsky, I.,
1990. Thrombin immobilized to extracellular matrix is a potent mitogen for vascular
smooth muscle cells: Nonenzymatic mode of action. Cell. Regul. 1: 453-463.

Bar-Shavit, R., Benezra, M., Sabbah, V. , Dejana, E., Vlodavsky, I., and Wilner, GD., 1992.
Functional domains in thrombin outside the catalytic site. In: Berliner, LJ., (ed)
"Thrombin: structure andfunction." Plenum Press: NY and London, pp. 315-350

Barger, A., Beeuwkes, R., Lainey, L., Silverman, K, 1984. Hypothesis: vasa vasorum and
neovascularization of human coronary arteries. N. Eng. 1. Med. 310: 175-177.

Belloni, P.N., Carney, D.H., and Nicolson, G.L., 1992. Organ-derived endothelial cells
express differential responsiveness to thrombin and other growth factors. Microvasc. Res.
43: 20-45.

275
Benezra, M, Vlodavsky, I., Bar- Shavit, R., 1993. Prothrombin conversion to
thrombin by plasminogen activator residing in the subendothelial extracellular matrix.
Semin. Thromb. Haemost. 19:405-411.

Benzakour, 0 ., and Kanthou, C., 1996. Cellular and molecular events in atherogenesis:
basis for pharmacological and gene therapy approaches to restenosis. Cell. Pharmacol. 3:
7-22.

Benzakour, 0., Kanthou, c., Lupu, F., Dennehy, u., Goodwin, c., Scully, M.F., Kakkar,
V.V., and Cooper, D.N., 1955. Prothrombin cleavage by human vascular smooth muscle
cells: a potential alternative pathqway to the coagulation cascade. J. Cell. Biochem. 59:
514-528.

Berk, B.C., Taubman, M.B., Cragoe, EJ., Fenton, lW., II, and Greindling K.K., 1991.
Thrombin-stimulated events in cultured vascular smooth-muscle cells. Biochem. J. 274 :
799-805

Brass, L.F., 1992. Homologous desensitisation of HEL cell thrombin receptors:

distinguishable roles for proteolysis and phosphorylation. J. BioI. Chem. 267: 6044-6050.

Brass, L.F., Laposata, F., Banga, HS., and Rittenhouse, S.E., 1986. Regulation of the
phosphoinositide hydrolysis pathway in thrombin-stimulated platelets by a pertussis
toxin-sensitive guanine nucleotide-binding protein. J. BioI. Chem. 261: 16833-16847.

Brass, L.F., Pizarro, S., Ahuja, M., Belmonte, E., Stadel, l, and Hoxie, J.A. , 1994.
Changes in the structure and function of the human thrombin receptor during receptor
activation, internalisation, and recycling. J. BioI. Chem. 269: 2943-2952.

Brass, L.F., Vassallo R.R., Belmonte, E., Ahuja, M., Cichowski, K., and Hoxie, lA., 1992.
Structure and function of the human platelet thrombin receptor. Studies using monoclonal
antibodies directed against adefined domain within the receptor N terminus. J. Bioi. Chem.
267: 13795-l3798.

Carney, D.H 1992. Postclotting cellular effects of thrombin mediated by interactions with
high affinity thrombin receptors. In: Berliner, LJ. (ed) "Thrombin: structure and
function". Plenum Press: NY and London, pp. 351 -396.

Chen, L.B., and Buchanan, lM., 1975. Mitogenic activity of blood components.
Thrombin and Prothrombin. Proc. Natl. Acad. Sci. USA 72: 131-135.

Clapham, D.E., and Neer, EJ., 1993. New roles for G-protein bg-dimers in transmembrane
signalling. Nature 365 : 403-406.

Connolly, A., Ishihara, H, Kahn, M., Farese, R.V. Jr., and Coughlin S.R., 1996. Role of
thrombin receptor in development and evidence for a second receptor. Nature 381:516-
519.

Davies, MJ., Bland, M.J., Hangartner, W.R., Angelini, A., and Thomas, A.c., 1989.
Factors influencing the presence of acute coronary thrombi in sudden ischaemic death . Eur.
Heart. J. 10: 203-208.

276
Eisenstein, R., 1991. Angiogenesis in arteries. Pharmacol. Ther. 49: 1-19.

Esmon, C. T., 1993. Cell-mediated events that control blood coagulation and vascular
injury. Annu. Rev. Cell Bioi. 9: 1-26.

Garcia, lG.N., Aschner, lL., and Malik, A.B., 1992. Regulation of thrombin-induced
endothelial barrier dysfunction and prostaglandin synthesis. In: Berliner, L.J, (ed)
"Thrombin: structure andfunction" Plenum Press: NY and London, pp. 397-430.

Gerrity, RG., and Cliff, W.l, 1975 . The aortic tunica media of the develooping rat. I.
Quantitative stereologic and biochemical analysis. Laborat. Invest. 32: 585-600.

Gerszten, R.E., Chen, l, Ishii, M., Ishii, K, Wang, L, Nanevicz, T., Turck, C.W., Yu, T-
KH., and Coughin S.R., 1994. Specificity of the thrombin receptor for agonist peptide is
defined by its extracellular surface. Nature 368 : 648-651.

Haralabopoulos, G.c., Grant, D.S, Kleinman, H.K, Maragoudakis, ME , 1997.

Thrombin promotes endothelial cell alignment in matrigel in vitro and angiogenesis in vivo.
Am. J Physiol. (in press)

Hatton, M.W., Moar, SL, and Richardson, M., 1989. Deendothelialization in vivo
initiates a thrombogenic reaction at the rabbit aorta surface. Correlation of uptake of
fibrinogen and antithrombin III with thrombin generation by the exposed subendothelium.
Am. 1. Pathol. 135: 499-508.

Herbert, J-M., Dupuy, E., Laplace, M-C., Zini, J-M., Bar-Shavt, R, and Tobelem G.,
1994. Thrombin induces endothelial cell growth via both a proteolytic and a non-
proteolytic pathway. Biochem.1. 303: 227-231.

Herbert, J-M., Lamarche, I., Prabonnaud, Y. , Dol,F ., and Gauthier, T., 1994. Tissue type
plasminogen activator is a potent mitogen for human aortic smooth muscle cells. 1. Bioi.
Chem. 269: 3076-3080.

Houck, KA., Ferrara, N., Winer, l , Cachianes, G. , Li, B., and Leung, D.W., 1992. The
vascular endothelial growth factor family : identification of a fourth molecular species and
characterization of alternative splicing of RNA. Mol. Endocrinol. 5: 1806-1814.

Hoxie, J.A., Ahuja, M., Belmonte, E., Pizarro, S., Parton, R, and Brass L.F., 1993.
Internalisation and recycling of activated thrombin receptors. 1. BioI. Chem. 268 : 13756-
13763 .

Huang, c.L., and Ives, H.E., 1987. Growth inhibition by protein kinase C late In
mitogenesis. Nature 329: 849-850.

Hung, D.T., Wong, YH., Yu, T-KH., and Coughlin, S.R, 1992. The cloned platelet
thrombin receptor couples to at least two distinct effectors to stimulate phosphoinositide
hydrolysis and inhibit adenylyl cyclase. 1. BioI. Chem. 267: 20831-20834.

Ishihara, H., Connolly, A.l, Zeng, D., Kahn, M.L. , Zheng yw., Timmons, c., Tram, T.,
and Coughlin S.R, 1997. Protease-activated receptor 3 is a second thrombin receptor in
humans. Nature 386: 502-506.

277
Ishii, K., Chen J., Ishii, M., Koch, W.J., Freedman, N.J., Lefkowitz, R.J., and
Coughlin S.R., 1994. Inhibition of thrombin receptor signalling by a G-protein coupled
receptor kinase: functional specificity among G-protein coupled receptor kinases. J. BioI.
Chem. 269: 1125-1130.

Ishii, K., Rein, L., Kobilka, B., and Coughlin, S.R., 1993. Kinetics of thrombin receptor
cleavage on intact cells. Relation to signaling. J. BioI. Chem. 268: 9780-9786.

Kaetzel, D.M., Coyne, D.W., and Fenstermaker, RA., 1993. Transcriptional control of
the platelet derived growth factor subunit genes. Bio/actors 4: 71-81.

Kahan, C., Seuwen, K., Meloche, S., and Pouyssegur, J., 1992. Coordinate, biphasic
activation of p44 mitogen-activated protein kinase and S6 kinase by growth factors in
hamster fibroblasts. J. BioI. Chem. 267: 13369-13375.

Kanse, S.M., Benzakour, 0., Kanthou, C., Kost, c., Lijnen, R., and Preissner, K.T., 1997.
Urokinase-type plasminogen activator stimulates hyman vascular smooth muscle cell
proliferation. Atherosc!. Thromb. Vasco BioI. (in press).

Kanthou, C., and Benzakour, 0., 1995. Cellular effects of thrombin and their signalling
pathways. Cell. Pharmacol. 2: 293-302.

Kanthou, C., Benzakour, 0., Patel, G., Deadman, J., Kakkar, v.v., and Lupu, F., 1995a.
Thrombin receptor activating peptide (TRAP) stimulates mitogenesis, c-fos and PDGF-A
gene expression in human vascular smooth muscle cells. Thromb. Haemost. 74: 1340-1347.

Kanthou, C., Kanse, S.M., Kakkar, V.v., and Benzakour, 0., 1996. Involvement of
pertussis toxin sensitive and insensitive G proteins in thormbin signalling on cultured
human vacular smooth muscle cells. Cell. Signall. 8: 59-66.

Kanthou, C., Kanse, S.M., Newman, P., Kakkar, v.v., and Benzakour, 0., 1995b.
Structural domains ofthormbin involved in the induction of mitogenesis in cultured human
vascular smooth muscle cells. Blood Coag. Fibr. 6: 634-642.

Kanthou, C., Parry, G., Wijelath, E., Kakkar, V.v., and Demoliou-Mason, C., 1992.
Thrombin-induced proliferation and expression of platelet-derived growth factor-A chain
gene in human vascular smooth muscle cells. FEBS Lett. 314: 134-148.

Ku, p.r., and D'Amore, P.A., 1995. Regulation of basic fibroblast growth factor (bFGF)
gene and protein expression following its release from sublethally injured endothelial cells.
J. Cell. Biochem. 58: 328-343.

Li, X.N., Varma, v.K., Parks, J.M., Benza, RL., Koons J.C., Grammer, J.R., Grenett, H.,
Tabengwa, E.M., and Booyse F.M., 1995. Thombin decreases the urokinase receptor and
surface localized fibrinolysis in cultured endothelial cells. Arterosc/er. Thromb. 15: 410-
419.

Lin, L-L., Wartmann, M., Lin, A.Y., Knopf, J.L, Seth, A.A., and Davis RJ., 1993. cPLA2
is phosphorylated and activated by MAP kinase. Cell 72: 269-278.

278
Lombardi, D.M.M., Reidy, M.A., and Schwartz, S.M., 1991. Methodologic
considerations important in the accurate quntitation of aortic smooth muscle cell replication
in the normal rat. Am. 1. Pathol. 138: 441-446.

Magnaldo, I., Pouyssegur, J., and Paris, S., 1988. Thrombin exerts a dual effect on
stimulated adenyl ate cyclase in hamster fibroblasts: an inhibition via a GTP-binding protein
and a potentiation via activation of protein kinase C. Biochem.1. 253: 711-719.

Majesky, M.W., Reidy, M.A., Bowen-Pope, D.F., Hart, C.E., Wilcox, J.N., and Schwartz,
S.M., 1990. PDGF ligand and receptor gene expression during repair or arterial injury. 1.
Cell Bioi. Ill: 2149-2158.

Mann, K.G., 1994. Prothrombin and Thrombin. In Colman, R.W., Hirsh, J., Marder,
V.J., and Salzman, E.W., (eds): "Haemostasis and Thrombosis: Basic PrinCiples and
Clinical Practice". Lippincott Company: Philadelphia, pp. 184-199.

Maragoudakis, M.,E., 1996. The role ofthormbin in angiogenesis. In: "Molecular, Celluar
and Clinical Aspects of Angiogenesis" Ed. Maragoudakis, Plenum Press, ppI15-124.

McNamara, c.A., Sarembock, IJ, Gimple, L.W., Fenton, J.W., II, Coughlin, S.R, and
Owens, G.K., 1993 . Thrombin stimulates proliferation of cultured rat aortic smooth
muscle cells by a proteolytically activated receptor. 1. Clin. Invest. 91: 94-98.

Molloy, c.J., Pawlowski, J.E., Taylor, D.S., Turner, C.E., Weber, H ., Peluso, M., and
Seiler, S.M., 1996. Thrombin receptor activation elicits rapid protein tyrosine
phosphorylatin and stimulation of the raf-1lMAP kinase pathway preceding delayed
mitogenesis in cultured rat aortic smooth muscle cells. 1. Clin. Invest. 97: 1173-1183.

Ngaiza, J.R., and Jaffe, E.A., 1991. A 14 amino acid peptide derived from the amino
terminus of the cleaved thrombin receptor elevates intracellular calcium and stimulates
prostacycline production by human endothelial cells. Biochem. Biophys. Res. Commun.
179: 1656-1661.

Noda-Heiny, H., fujii, S., and Sobel, B.E., 1993. Induction of vascular smooth muscle cell
expression of plasminogen activator inhibitor-l by thrombin. Cire. Res. 72: 36-43.

Obberghen-Schilling, EV, and Pouyssegur, 1., 1993. a-Thrombin receptors and growth
signalling. Semin. Thromb. Haem. 19: 378-385.

Okada, S.S., Grobmyer, S.R., and Bamathan E.S., 1996. Contrasting effects of
plasminogen activators, urokinase receptor and LDL receptor -related protein on smooth
muscle cell migration and invation. Arterioscler. Thromb. Vasco BioI. 16: 1269-1276.

Okazaki, H., Majesky, M.W., Harker, L.A., and Schwartz, S.M., 1992. Regulation of
platelet-derived growth factor ligand and receptor gene expression by a-thrombin III
vascular smooth muscle cells. Circ. Res. 71: 1285-1293.

Pearson, J.D., 1994. Endothelial cell function and thrombosis. Bail/ieres Clinical
Haematology 7: 441-452.

279
Pepper, M.S., Montesano, R., Mandriota, S., Orci, L., and Vassalli, J-D., 1996.
Angiogenesis: a paradigm for balanced extracellular proteolysis during cell migratio and
morphogenesis. Enzyme Protein 49: 138-162.

Rade, 1.1., Schulick, A.H., Virmani, R., and Dichek, D.A., 1996. Local adenoviral-mediated
expression of recombinant hirudin reduces neointima formation after arterial injury. Nature
Med 2: 293-298.

Raines, E.W., and Ross, R., 1992. Compartmentalisation ofPDGF on extracellular binding
sites dependent on exon 6 encoded sequences. J. Cell Bioi. 116: 553-543.

Rasmussen, ll.B., Vouret-Craviari, v., Jallat, S., Schlesinger, Y, Pages, G., Pavirani, A.,
Lecocq, J-P., Pouyssegur, 1., and Van Obberghen-Schilling, E., 1991. cDNA cloning and
expression of a hamster a-thrombin receptor coupled to Ca 2+ mobilization. FEES Lett.
288: 123-128.

Reuning, ll., Little S.P., Dixon, E.P., and Bang, N.ll. 1994. Effect of thormbi, the
thrombin receptor activation peptide, and other mitogens on vascular smooth muscle cell
urokinase receptor mRNA levels. Blood 84: 3700-3708.

Risau, W., 1997. Mechanisms of angiogenesis. Nature 386: 671-674.

Ross, R., 1993. The pathogenesis of atherosclerosis: a perspective for the 1990s. Nature
362: 801-809.

Scarpati, E.M., and DiCorieto, P.E., 1996. Identification of a thrombin response element in
the human platelet-derived growth factor B-chain (c-sis) promoter. J. Bioi. Chem. 271:
3025-3032.

Schaeffer, P., Riera, E., Dupuy, E., and Herbert 1.M., 1997. Non-proteolytic activation of
thro thrombin receptor promotes human umbilical vein endothelial cell growth but not
intracellular Ca2+ prostaclyclin or permeablility. Biochem. Pharmacol. 53: 487-491.

Sekiya, F ., Usui, H., Inoue, K., Fukudome K, Morita, T., 1994. Activation of prothrombin
by a novel membrane-associated protease. J. BioI. Chem. 269: 32441-32445.

Sethi, N., and Brookes, M., 1971. Ultrastructure of the blood vessels in the chick allantois
and chorioallantois. J. Anal. 109: 1-5.

Shankar, R, de la Motte, c.A., Poptic, E.J., and Dicorletto, P.E., 1994. Thrombin
receptor-activating peptides differentially stimulate platelet-derived growth factor
production, monocytic cell adhesion, and E-selectin expression in human umbilical vem
endothelial cells. J. Bioi. Chem. 269: 13936-13941.

Stiemberg, 1., Redin, W.R., Warner, W.S., and Carney, D.H., 1993. The Role of thrombin
and the thrombin receptor activating peptide (TRAP-508) in initiation of tissue repair.
Thromb. Haemost. 70: 158-162.

Taipale, J, Koli, K., and Keski-Oja, 1., 1992. Release of transforming growth factor-bl
from the pericellular matrix of cultured fibroblasts and fibrosarcoma cells by plasmin and
thrombin . J. BioI. Chem. 267: 25378.

280
Taparelli, c., Mettemich, R., Ehrhardt, c., and Cook, N.S., 1993. Synthetic
low-molecular weight thrombin inhibitors: molecular design and pharmacological profile.
TIPS 14: 366-376.

Trejo, lA., Connolly, A.l, and Coughlin, SR, 1996. The cloned thrombin receptor is
necessary and sufficient for activtion of mitogen-activated protein kinase and mitogenesis
in mouse lung fibroblasts. J. Bioi. Chem. 271: 21536-21541.

Tsopanoglou, N.E., Pipili-Synetos, E., and Maragoudakis, ME, 1993. Thrombin

promotes angiogenesis by a mechanism independent of fibrin formation. Am. J. Physiol.
264: C1302-1307.

Vlodavsky, I., Folkman, l, Sullivan, R., Fridman, R., Ishai-Michaeli, R., Sasse, l, and
Klagsbrun M., 1987. Endothelial cell-derived basic fibroblast growth factor: Synthesis and
deposition into subendothelial extracellular matrix. Proc. Nat!. Acad Sci. USA 84: 2292-
2296.

Vouret-Craviari, v., Grall, D., Chambard, J-C., Rasmussen u.B., Pouyssegur, 1., and Van
Obberghen-Schilling, E., 1995. Post translational and activation-dependent modifications
of the G-protein coupled thrombin receptor. J. BioI. Chem. 270: 8367-8372.

Vouret-Craviari, v., Van Obberghen-Schilling, E., Scimeca J-C., Van Obberghen, E., and
Pouyssegur 1., 1993. Differential activation of p44mapk (ERKl) by a-thrombin and
synthetic receptor peptide agonists. Biochem. J. 289: 209-214.

Vouret-Craviari, V., Van Obberghen-Schilling, E., Rasmusse, VB., Paviran, A., Lecocq, J-
P., and Pouyssegur, l, 1992. Synthetic a-thrombin receptor peptides activate G-protein
coupled signalling pathways but are unable to induce mitogenesis. Mol. BioI. Cell 3: 95-
102.

Vu, T-K.H., Hung, D.T., Wheaton, V.I., and Coughlin, S,R., 1991. Molecular cloning of a
functional thrombin receptor reveals a novel proteolytic mechanism of receptor activation.
Cell 64: 1057-1068.

Walters, T.K., Gorog, D.A., and Wood, R.F.M., 1994. Thrombin generation following
arterial injury is a critical initiating event in the pathogenesis of the proliferative stages of
the atherosclerotic process. J. Vasco Res. 31: 173-177.

Wang, H.S., Li, F, Runge, M.S., and Chaikof, E.L., 1997. Endothelial cells exhibit
differential chemokinetic and mitogenic responsiveness to alpha thormbin. J. Surg. Res. 68:
139-144.

Weiss, R.H., and Maduri, M., 1993. The mitogenic effect of thrombin in vascular smooth
muscle cells is largely due to basic fibroblast growth factor. J. BioI. Chem. 268 5724-5727.

Weitz, 1.1., Hudoba, M., Massel, D., Maraganore, lM., and Hirsh, l, 1990. Clotbound
thrombin is protected from inhibition by heparin-antithrombin III but is susceptible to
inactivation by antithrombin III-independent inhibitors. J. C/in. Invest. 86: 385-391.

Wilcox, IN., Rodriguez, l, Subramanian, R., Ollerenshaw, l, Zhong, c., Hayzer, D.J.,
Horaist, c., Hanson, S.R., Lumsden A., Salam, T.A., Kelly, A.B., Harker, L.A., and Runge,

281
M ., 1994. Characterisation of thrombin receptor expression during vascular
lesion fonnation. Circ. Res. 75: 1029-1039.

Wojta, 1., Gallicchio, M., Zoellner, H., Hufnagi, P., Last, K., Filonzi, EL, Binder, B.R.,
Hamilton, 1.A., and McGrath, K., 1993. Thrombin stimulates expression of tissue-type
plasminogen activator and plasminogen activator inhibitor type 1 in cultured human
vascular smooth muscle Cells. Thromb. Haemost. 70: 469-474.

Woolkalis, MJ., and Brass, L.F., 1994. Thrombin receptor cleavage and intemalisation in
human umbilical endothelial cells. FASEB J. 8: A 1385.

Zoldhelyi, P., Bichler, 1., Owen, W.G., Grill, D.E., Fuster, v., Mruk, 1.S., Chesebro, 1.H.,
1994. Persistent thrombin generation in humans during specific thrombin inhibition with
hirudin. Circulation 90: 2671-2678.

282
Regulation of
Angiogenesis
ENDOGENOUS REGULATION OF ANGIOGENESIS IN VITRO

Roberto F. Nicosia

Department of Pathology, Allegheny University of the Health Sciences,

Philadelphia, PA, U.S.A.

1. ABSTRACT

Rings of rat aorta embedded in collagen gel and cultured under serum-free conditions
produce a self-limited angiogenic response in the absence of exogenous growth factors.
Aortic rings respond to the injury of the dissection procedure by generating outgrowths of
branching microvessels and fibroblasts. The microvessels originate primarily from the
aortic intima whereas fibroblasts arise from the adventitia. The endothelium of the rat aorta
switches to a microvascular phenotype and recruits pericytes from a subpopulation of
smooth muscle cells located in the intimal/subintimallayers. Formation of microvessels is
due to the combined effect of injury, exposure of the apical surface of the endothelium to a
hydrated collagen matrix, and paracrine interactions between endothelial and non-
endothelial cells. Endogenous growth factors involved in the regulation 0 angiogenesis
include basic fibroblast growth factor (bFGF), vascular endothelial growth factor (VEGF).
and platelet-derived growth factor (pDGF). bFGF and VEGF stimulate angiogenesis
directly. PDGF stimulates fibroblasts and pericytes, which in turn modulate the growth,
differentiation and survival of microvessels. Fibroblasts, which are the first cells to migrate
out of the explants, promote angiogenesis by secreting growth factors and by stabilizing the
newly formed microvessels. Pericytes, which increase in number during the maturation of
the microvessels, contribute to the differentiation and remodeling of the neovasculature.
These results indicate that the rat aorta model can be used to study the cellular and
molecular mechanisms by which the vessel wall regulates microvessel formation at different
stages of the angiogenic process.

2. INTRODUCTION

Angiogenesis plays an important role in physiologic, reactive, inflammatory, and neoplastic

conditions. During angiogenesis, preexisting blood vessels generate new vessels in response

Angiogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 285
to soluble growth factors which stimulate endothelial cell migration, proliferation,
proteolytic activity and capillary tube formation (Folkman and Shing, 1992). Since they
produce angiogenic factors (Schiffers, Fazzi, van Ingen Schenau, and De Mey, 1994), blood
vessels activated by injury have the capacity to generate an angiogenic response without
the stimulatory intervention of nonvascular cells. Endogenous autoregulation of
angiogenesis can be demonstrated in vitro by culturing explants of rat aorta under serum-
free conditions in the absence of exogenous growth factors (Kawasaki, Mori and Awai,
1989; Nicosia and Ottinetti, 1990). Angiogenesis in this model is a self-limited process
mediated by growth factors secreted by aortic cells in response to the injury of the
dissection procedure. In this chapter we review the phenomenon of rat aortic angiogenesis
and discuss the role of growth factors-and cell-cell interactions in the system.

3. MATERIALS AND METHODS

Aortic rings obtained from Fischer 344 male rats were embedded in collagen gels and
cultured in serum-free MCDB-131 growth medium (Clonetics, San Diego CA) at 35.5 DC
in a 5% CO 2 incubator (Nicosia and Ottinetti, 1990). Antibodies against bFGF and VEGF
for immunochemical and neutralizing experiments were from R&D Systems (Minneapolis,
MN) (Villaschi and Nicosia, 1993; Nicosia, 1997; Nicosia, Lin, Hazelton and Quian, 1996).
Anti-PDGF antibody for immunoblot experiments was from R&D Systems. Neutralizing
anti-rat PDGF B antibody was a kind gift from Dr. Volkhard Lindner (Maine Medical
Center Research Institute, South Portland, ME). VEGF enzyme linked immunosorbent
assay (ELISA) was from Cytimmune Services Inc., College Park, MD (Nicosia, 1997;
Nicosia et aI. 1996). For stimulation experiments, the growth medium was supplemented
with basic fibroblast growth factor (bFGF, R&D Systems), vascular endothelial growth
factor (VEGF, R&D Systems), natural platelet-derived growth factor (PDGF, UBI, Lake
Placid, NY), recombinant PDGF AA (UBI) or recombinant PDGF BB (UBI) (Nicosia,
Nicosia, and Smith, 1994a; Villaschi and Nicosia, 1993). Reverse transcriptase polymerase
chain reaction (RT-PCR) was used to demonstrate the mRNA for VEGF and the VEGF
receptor jlk-l in aortic cultures MD (Nicosia, 1997; Nicosia, Lin, Hazelton and Qui an,
1996). For co-culture experiments, endothelial cells were isolated nonenzymatically from
aortic rings (Nicosia, Villaschi and Smith, 1994b). Fibroblasts were isolated by enzymatic
digestion from rat tail tendons (Villaschi and Nicosia, 1994). Smooth muscle cells were
isolated by nonenzymatic methods from the intimal aspect of the rat aorta or from the
media (Villaschi, Nicosia and Smith, 1994). Microvascular networks were obtained by
overlaying isolated endothelial cells cultured on a collagen gel with a second layer of
collagen (Nicosia et al., 1994b). For co-cultures experiments, fibroblasts or smooth muscle
cells were embedded in the bottom collagen gel before plating endothelial cells and
overlaying them with collagen (Villaschi and Nicosia, 1994; Nicosia and Villaschi, 1995).
Rat aortic cultures and co-cultures with isolated cells were studied by
immunohistochemical and ultrastructural methods.

4. RESULTS

4.1 Angiogenesis in Serum-free Collagen Gel Culture of Rat Aorta

Aortic rings embedded in collagen gels and cultured under serum-free conditions generated
outgrowths composed of microvessels and fibroblasts (Nicosia and Ottinetti, 1990).

286
Fibroblasts, which were the first cells to migrate into the collagen, were observed after 2-3
days of culture. Microvessels started sprouting at day 3-4, primarily from the wounded
edges of the aortic ring and its collateral branches. During the early stages of angiogenesis,
the vascular outgrowth was primarily composed of endothelial cells. As they matured, the
microvessels became surrounded by pericytes which migrated and proliferated at the
abluminal surface of the endothelium (Nicosia and Villaschi, 1995). The angiogenic growth
phase lasted approximately one week and was followed by a phase of neovascular
regression and remodeling. During this phase, the number of microvessels decreased while
pericytes continued to proliferate. The endothelial cells of the tips and branches retracted
into the main stems of the microvessels. As a result of endothelial retraction and pericyte
proliferation, the microvessels became thick and unbranched. The endothelial cells were
positive for factor VIII-related antigen and the Griffonia simplicifolia isolectin B4. The
pericytes were positive for alpha-smooth muscle actin (Nicosia and Ottinetti, 1990;
Nicosia and Villaschi, 1995). Approximately 10-15 % of the fibroblasts expressed alpha-
smooth muscle actin indicating myofibroblastic differentiation. By electron microscopy
the microvessels had a patent lumen and were lined by polarized endothelial cells with
distinct luminal and abluminal surfaces. Pericytes exhibited electron dense cytoplasm and
thin cytoplasmic processes which formed a discontinuous coating around the endothelium
(Nicosia and Villaschi, 1995) (Figure 1).

Figure 1. Electron micrograph of newly formed microvessel in serum-free collagen gel

culture of rat aorta. L, lumen; E, endothelial cells; P, pericytes. Magnification: X 1250.

287
Figure 2. Electron micrograph showing reorganization of intimal endothelial cells (E) into
a microvascular tube with a patent lumen (L). Microvascular morphogenesis occurs above
the internal elastic lamina (IEL), at the opening of a collateral branch of the rat aorta.
Pericytes (P) originate from subendothelial smooth muscle cells (S). Magnification: X 1600.

During the angiogenic response, endothelial cells of the aortic mtIma switched to a
microvascular phenotype in response to the combined effect of injury and exposure of their
apical surface to collagen fibrils. Adventitial vasa vasorum were not required for this
process since angiogenesis was observed also in cultures of everted thoracic aortas whose
vasa vasorum were sequestered inside the aortic tubes and had no access to the collagen gel
(Nicosia, Bonanno and Villaschi, 1992). Ultrastructural studies demonstrated a dramatic
reorganization of the aortic endothelium into microvascular tubes above the internal elastic
lamina. This was particularly evident at the opening of the collateral vessels of the thoracic
aorta. The smooth muscle cells of the intima or the subintimal layers of the aorta migrated
toward the endothelium and became incorporated as pericytes into the subendothelial space
of the newly formed microvessels (Figure 2).

4.1.1. Role of bFGF, VEGF and PDGF in the Angiogenic Response of the Rat Aorta
Since the rat aorta was capable of generating new microvessels in the absence of serum or
exogenous growth factors, we hypothesized that angiogenesis in this model was mediated
by endogenous growth factors produced by the explants. Immunohistochemical staining of
the rat aorta showed bFGF in the cytoplasm of endothelial cells and smooth muscle cells.
Aortic rings mechanically injured by the dissection procedure released bFGF which was
demonstrated in-conditioned medium by both slot and Western blot analysis. bFGF
release was maximal during the first days of culture and decreased over time. Medium
conditioned by freshly cut aortic explants stimulated the angiogenic response of the rat

288
 -
~
w
(/)
z
o
11.
(/)
W
II:
!.!
z
w
"Q
"~
IgG

Figure 3. Effect of neutralizing antibodies against bFGF, VEGF and PDGF B on

angiogenesis in the rat aorta model. Cultures were treated with 50 micrograms/ml anti-
bFGF antibody, 40 micrograms/ml anti-VEGF antibody or 50 micrograms/ml anti-PDGF
B antibody. Angiogenesis was quantitated by counting the number ofmicrovessels at day 7.
Values are expressed as percentage of control treated with 50 micrograms/ml nonimmune
/gG. N = 6 per experimental group. Angiogenesis was inhibited by anti-bFGF and anti-
VEGF antibodies.p < 0.001 (**); P < 0.05 (*). The anti-PDGF B antibody had no anti-
angiogenic effect..

aorta in collagen gel culture by 90% over control values. Neutralizing antibodies against
bFGF caused a 40% reduction in the number of microvessels generated by the aortic
explants but were unable to completely abolish the angiogenic response (Figure 3; Villaschi
and Nicosia, 1993)
The inability of anti-bFGF antibodies to completely suppress angiogenesis suggested that
additional angiogenic factors were present in the system. Immunoblot, ELISA and R T-
peR studies demonstrated VEGF and its receptor flk-l in the aortic cultures. Secretion of
VEGF was maximal during the early stages of angiogenesis and decreased over time as seen
for bFGF (Nicosia, 1997; Nicosia et al., 1996). Inhibition of VEGF with a neutralizing
antibody caused a 70% reduction in the angiogenic response of the rat aorta (Figure 3).
The aorta conditioned medium contained also PDGF, which is known to be overexpressed
by the arterial wall after injury (Majesky, Reidy, Bowen-Pope, Hart, Wilcox and Schwartz,
1990) . PDGF was more abundant during the angiogenic growth phase, as seen for bFGF
and VEGF (Figure 4). An antibody directed against the PDGF B chain, which binds to
both the ex and ~ subunits of the PDGF receptors (Majesky et al., 1990), failed to inhibit
angiogenesis (Figure 3).
To further evaluate the role of growth factors in the system, collagen gel cultures of rat
aorta were treated with purified bFGF, VEGF and PDGF. Exogenous bFGF, VEGF and
PDGF stimulated the proliferation and elongation of microvessels in a dose-dependent
manner (Villaschi and Nicosia, 1993; Nicosia et al. 1994a). Maximal stimulation was
observed in cultures treated with 60 nglml bFGF (140% over control values), 10 nglml
VEGF.(160%), 2 nglml platelet derived-PDGF (93%) and 0.5 nglml PDGF AA (96%) or
PDGF BB (186%) (Figures 5 and 6). PDGF promoted also migration and proliferation of
fibroblasts. The angiogenic effect of PDGF became apparent during the second week of
culture when levels of endogenous PDGF were low and microvessels were no longer

289
 -OOIIYROL
-PDGtr

I
100

I-
~

DAYS

6 10 IS

Figure 4. The effect of exogenous PDGF (0.2 nglml) on the angiogenic response of the rat
aorta over time (N = 3) is compared to the immunoreactivity of aortic ring-conditioned
mediumfor endogenous PDGF. Stimulation of angiogenesis by exogenous PDGF became
significant during the second week ofculture when secretion ofendogenous PDGF by aortic
rings was no longer detectable by slot immunoblotting. Data in graph are expressed as
mean +1- standard error of the mean.

growing in the untreated controls (Figure 4). Aortic explants embedded in collagen gels 10-
14 days after excision from the animal, when endogenous secretion of growth factors was
low, had a markedly reduced angiogenic response (10% of that observed with freshly
wounded explants). Exogenous bFGF, VEGF or PDGF stimulated vascular proliferation
from these quiescent explants restoring the angiogenic response to values comparable to
those obtained with freshly cut explants (Nicosia, 1997; Nicosia et aI., 1996).

4.1.1.1. Interactions between Isolated Endothelial Cells and Fibroblasts or Smooth

Muscle Cells
Since rat aortic endothelial cells did not respond directly to PDGF (data not shown), we
hypothesized that the angiogenic effect of exogenous PDGF was mediated by fibroblasts,
which were greatly stimulated in PDGF-treated aortic cultures. Both isolated fibroblasts
and fibroblast-conditioned medium promoted the angiogenic response of the rat aorta by 90
% as compared to the untreated control cultures (ViIIaschi and Nicosia, 1994). Fibroblasts

290
 )()()
**
~ ** **
W
."
Z
0 200
~
."
W
a:
u
Z
w
8 100
G
z
«

Figure 5. Effect of exogenous bFGF. VEGF, natural PDGF. PDGF AA and PDGF BB on
rat aortic angiogenesis. Angiogenesis was scored by measuring the number of newly
formed microvessels. Stimulation of angiogenesis by growth factors is expressed a s
percentage of untreated contro/.. The histogram shows the maximal stimulatory effects
which were obtained with 60 ng/ml bFGF (N = 9), 10 ng/ml VEGF (N = 6), 2 ng/ml PDGF
(N = 6), 0.5 ng/ml PDGF AA (N = 6), and 0.5 ng/ml PDGF BB (N = 4). P < 0.001 (**); P
< 0.007 (*). Adaptedfrom Villaschiand Nicosia (1993) and Nicosia et al. (1994a).
co-cultured with aortic explants in collagen gels were more etTective than fibroblast-
conditioned medium in prolonging the survival of the neovessels, suggesting that cell-cell
interactions played an important role in the angiotrophic activity ofthese cells . The role of
endothelial-fibroblast interactions in angiogenesis was further evaluated by coculturing
isolated fibroblasts and endothelial cells in collagen gels under serum-free conditions.
Endothelial cells cultured on collagen gel and overlaid with a second layer of collagen
reorganized into microvascular networks. The bottom collagen gel, which served as an
adhesive substrate for the endothelial cells, was prepared with or without fibroblasts .
Microvessels cultured in the absence of fibroblasts degenerated within 3-4 days .
Conversely, microvessels co-cultured with fibroblasts had a markedly prolonged life-span
and survived up to three weeks which was the longest time studied (Figure 7) .
Ultrastructural studies demonstrated increased deposition of extracellular matrix in the
subendothelial space ofmicrovessels stabilized by fibroblasts . Endothelial cells promoted
the transformation of fibroblasts into myofibroblasts which have been implicated in the
contraction of granulation tissue and wound closure (Gabbiani , Ryan and Majno, 1971).
Endothelial cell-conditioned medium and the endothelial cell-derived peptide endothelin-l
promoted the contraction of the collagen gel by fibroblasts . Similarly, fibroblast-
conditioned medium promoted the contraction of the collagen gel by endothelial cells
(Villa schi and Nicosia, 1994).'
Ultrastructural studies of rat aorta cultures suggested that pericytes originated from
subendothelial smooth muscle cells (Figure 2). The capacity of aortic smooth muscle cells
to differentiate into pericytes was studied by coculturing intimal- or medial-derived smooth
muscle cells (Villaschi et a1 ., 1994) with endothelial cells in the collagen gel overlay assay
previously used to study the interactions between endothelial cells and fibroblasts (Nicosia
and Villaschi, 1995). Intimal-derived smooth muscle cells migrated toward the endothelium
and differentiated into pericytes whereas smooth muscle cells isolated from the deep layers
of the media had no significant endothelial tropism . The microvessels formed by
endothelial cells in the presence of intirnal-derived smooth muscle cells were more stable
than control microvessels which rapidly disintegrated as discussed earlier. Thus, the

291
 A

B
Figure 6. Photomicrographs 0/ VEGF-treated (A) and control (B) serum-jree collagen gel
cultures 0/ rat aorta. Exogenous VEGF (10 ng/ml) stimulated the proliferation and
elongation 0/ microvessels. Microvessels are marked by arrows. Magnification: J0 X

smooth muscle cells located in the intima or subintimal layers of the tunica media
responded to paracrine signals generated by the endothelium and became pericytes
contributing to the survival, maturation and differentiation ofthe microvessels .

5. DISCUSSION

The angiogenic response of the rat aorta in serum-free culture is a self-limited process
activated by the injury of the disseetion procedure and mediated by juxtacrinelparacrine
interactions between endothelial cells, fibroblasts and pericytes . We have identified three
growth factors, bFGF, VEGF and PDGF, which contribute to the endogenous regulation of
this system . bFGF and VEGF are direct angiogenic factors capable of stimulating
endothelial cell migration, proliferation and proteolytic activity (Ferrara, Houck, Jakeman,
and Leung, 1992; Folkman and Shing, 1992). PDGF, as proposed by others (Sato, Beitz,
Kato, Yamamoto, Clark, Calabresi, and Fraekelton, 1993), may act indirectly since large
vessel endothelial cells do not respond to PDGF (Fox and Di Corleto, 1991). This is in
contrast with microvascular-derived endothelial cells which have been reported to have
PDGF receptors (Smits, Hermansson, Nister, Kamushina, HeIdin, Westermark, and Funa,
1989). A1though the rat aortic endothelium switches to a microvascular phenotype in
collagen gel culture (Nicosia et al., 1992) and expresses the PDGF alpha receptor in

292
Figure 7. Photomicrograph of Z-day-old endothelial-fibroblast co-culture in collagen gel
overlay assay. Fibroblasts (arrowheads) stabilized the network 0/ microvessels formed by
endothelial cells (arrows). Microvascular networks formed in the absence 0/ fibroblasts
disintegrated within 3-4 days (not shown). The culture was fixed in 3% glutaraldehyde,
postfixed in osmium tetroxide and whole mounted on a glass slide with 20% glycerol.
Magnification: 20 X

response to injury (Lindner and Reidy , 1995), we have been unable to directly stimulate
this cell type with PDGF . The effect of exogenous PDGF may be mediated by fibroblasts
which have angiogenic activity in the rat aorta model (Villaschi and Nicosia , 1994). lt is
also possible that smooth muscle cells respond to PDGF by secreting angiogenic factors
such as bFGF and VEGF (Brogi, Wu, Namiki, and Isner, 1994). The role of endogenous
PDGF in angiogenesis, however, needs further investigation since a neutralizing antibody
against PDGF B failed to inhibit the angiogenic response of the rat aorta . One possible
explanation is that when PDGF BB is neutralized, its function is taken over by PDGF AA ,
which is also upregulated after vascular injury (Lindner and Reidy, 1995). Alternatively,
endogenous PDGF may regulate the migration and proliferation of nonendothelial cells and
may not be required for the angiogenic activity of endothelial cells. In vivo studies have
shown that injured arterial endothelial cells express the PDGF B chain while the underlying
smooth muscle cells upregulate both PDGF Band the PDGF beta receptor (Lindner and
Reidy, 1995). The coordinated expression of the PDGF B ligand and ß receptor in these
two cell types creates a paracrine loop which may play an important role in the
recruitment of pericytes by endothelial cells.
The rat aorta model and the co-culture experiments with isolated cells suggest that both
fibroblasts and pericytes play important roles in angiogenesis. Fibroblasts, which are the
first cells to migrate out of the aortic explants promote angiogenesis by secreting angiogenic
factors and by cooperating with endothelial cells in the deposition of the perivascular
extracellular matrix (Villaschi and Nicosia, 1994). Pericytes, which follow the sprouting
endothelium and increase in number when microvessels stop growing, contribute to the
arrest of the angiogenic response and the remodeling of the neovasculature (Nicosia and
Villaschi, 1995). Both fibroblasts and pericytes have the capacity to stabilize microvessels
formed by isolated endothelial cells.
Previous studies have suggested that pericytes may originate from fibroblasts (Clark and
Clark, 1925; Rhodin and Fujita, 1989). Although our studies indicate that pericytes derive
from subendothelial smooth muscle cells, we cannot rule out the possibility that some

293
fibroblasts became incorporated into the microvessels and differentiated into pericytes
during the late stages of angiogenesis . Isolated pericytes and smooth muscle cells have been
shown to inhibit endothelial cell migration and proliferation (Orlidge and D' Amore, 1987;
Sato and Rifkin, 1989). This inhibitory effect has been attributed to the activation of
transforming growth factor beta-l (TGF beta-I) which occurs when endothelial cells and
pericytes establish physical contact with each other (Orlidge and D ' Amore, 1987; Sato
and Rifkin, 1989). TGF beta-l has an anti-angiogenic effect in the rat aorta mode (Nicosia,
Nicosia and Smith, 1994) and may be one ofthe signals that turn angiogenesis off when the
microvessels are coated by pericytes during regression and remodeling of the
neovasculature. TGF beta-I mayaiso contribute to the differentiation and survival of
microvessels because of its ability to promote extracellular matrix organization and
deposition as weil as junctional complex formation between endothelial cells (Merwin,
Anderson, Kocher, Van Itallie, and Madri, 1990; Nicosia, 1997).

6. CONCLUSIONS AND FUTURE DIRECTIONS

Our studies indicate that serum-free collagen gel culture can be used to study the
endogenous vasoformative properties of the vessel wall and the mechanisms of interaction
between vascular endothelial and nonendothelial cells. The recent observations that
segments of rat renal vein (Nicosia et al., 1997) and human placental vessels (Brown ,
Maynes, Bezos, Maguire, Ford, and Parish, 1996) generate new vessels in culture suggest
that the approach used for studying angiogenesis in the rat aorta model can be applied to
other blood vessels. By studying how vascular explants regulate their vasoproliferative
activity, we may identify critical molecular checkpoints in the angiogenic cascade which can
be targeted pharmacologically for the treatment of angiogenesis-dependent disorders. It is
also possible that, as we learn more about the angiogenic capacity of blood vessels isolated
from different vascular beds, we will find heterogeneity of vasoformative responses which
will warrant an organ-specific approach for both basic and clinical applications of
angiogenesis research.

7. ACKNOWLEDGEMENTS

1 gratefully acknowledge the many colleagues and co-workers who have contributed to the
studies reviewed in this chapter. This work was supported by NIH Grant HL52585.

8. REFERENCES

Brogi, E., Wu, T., Namiki, A., and Isner I.M., 1994, Indirect angiogenic cytokines
upregulate VEGF and bFGF gene expression in vascular smooth muscle cells, whereas
hypoxia upregulates VEGF expression only . Circulation, 90:649-652

Brown, J., Maynes, S.F., Bezos, A., Maguire, DJ., Ford, M .D., and Parish, R., 1996, A
novel in vitro assay ofhuman angiogenesis, Lab. Invest . 75:539-555.

Clark, ER, and Clark, E.L., 1925, The development of adventitial (Rouget) cells on the
blood capillaries of amphibian larvae, Am J. Anal. 35:239-264.

294
Ferrara, N ., Houck, K., Jakeman, L., and Leung D., 1992, Molecular and biological
properties of the vascular endothelial growth factor. Endocr Rev, 13:18-32.

Folkman, 1., Shing, Y., 1992, Angiogenesis,J Biol. Chem. 267 :10931-10934.

Fox, P .L. and Di Corleto, P .E., 1991, Endothelial cell production ofplatelet derived growth
factor, Seminars in Perinatol. 15:34-39.

Gabbiani, G ., Ryan, G.R and Majno, G., 1971, Presence of modified fibroblasts in
granulation tissue and their role in wound contraction, Experientia, 27 :549-550.

Kawasaki , S., Mori M ., Awai M. , 1989, Capillary growth of rat aortic segments cultured
in collagen without serum. Acta Pathol Japonica 39 :712-718 .

Lindner, V. and Reidy, M .A ., 1995, Platelet-derived growth factor ligand and receptor
expression by large vessel endothelium in vivo, Am. J. Pathol. 146 :1488-1497.

Majesky, M .W ., Reidy, M .A ., Bowen-Pope, D .F., Hart, C . E., Wilcox, 1. N ., Schwartz,

S.M., 1990, PDGF ligand and receptor gene expression during repair of arterial injury, 1.
Cell Biol. , 111:2149-2158.

Merwin, 1.R., Anderson, 1.M ., Kocher, 0 ., Van ltallie, C.M ., and Madri , 1.A. , 1990 ,
Transforming growth factor beta-I modulates extracellular matrix organization and cell-cell
junctional complex formation during in vitro angiogenesis, J Ce!l Physiol, 142:117-128.

Nicosia, R .F., 1997, The rat aorta model of angiogenesis and its applications, in Vascular
Morphogenesis in Vivo, in Vitro and in Sapio, (Little, H . Sage and V. Mironov eds .),
Birkhauser, Cambrige, MA (in press)

Nicosia, RF., Lin YJ., Hazelton D ., Quian. X, 1996, Role of vascular endothelial growth
factor in the rat aorta model of angiogenesis, 1. Vasc. Res. 33 (S 1): 73

Nicosia, R . F ., Nicosia S. V. and Smith, M, 1994a, Vascular endothelial growth factor,

piatelet derived growth factor, and insulin-like growth factor-I promote rat aortic
angiogenesis in vitro, Am . 1. Pathol. 145:1023-1029

Nicosia, R . F ., Ottinetti A., 1990, Growth of microvessels in serum -free matri x culture of
rat aorta: a quantitative assay of angiogenesis in vitro . Lab . Invest. 63 :115-122.

Nicosia, RF ., Villaschi , S., 1995, Rat aortic smooth muscle cells become pericytes during
angiogenesis in vitro, Lab. Invest. 73 :658-666

Nicosia, R.F., Bonanno, E ., Villaschi, S., 1992, Large-vessel endothelium sw itches to a

microvascular phenotype during angiogenesis in collagen gel culture of rat aorta.
Atherosclerosis, 95 :191-199.

Nicosia, R F., Villaschi, S., Smith , M . 1994b, Isolation and characterization of

vasoformative endothelial cells from the rat aorta, In Vitro Ce!1. Dev. Biol. 30A:394-399

Orlidge, A. and D'Amore, P ., 1987, Inhibition of capillary endothelial cell growth by

pericytes and smooth muscle cells , J Ce" Biol, 105:1455-62

295
Rhodin, J.A .G. and Fujita, H., 1989, Capillary growth in the mesentery of normal young
rats. Intravital vdeo and electron microscope analysis, J. Submicroscop. Cytol. Pathol. 21: 1-
34.

Sato, N ., Beitz, J.G., Kato, 1., Yamamoto, M., Clark, 1.W., Calabresi., P., and Frackelton,
R. Jr., 1993, Plate1et-derived growth factor indirectly stimulates angiogenesis in vitro. Am.
J. Pathol. 4 :1119-1130.

Sato, Y, Rifkin , D.B ., 1989, Inhibition of endothelial cell movement by pericytes and
smooth musc1e cells : activation oflatent transfonning growth factor beta-l-like molecule by
plasmin during co-culture. J Cell Biol, 109:309-15.

SchitTers, P. M. H., Fazzi, G. E., van Ingen Schenau, D. and De Mey, 1. G. R., 1994,
EtTects of candidate autocrine and paracrine mediators of growth responses in isolated rat
arteries, Arterioscler. Thromb . 14:420-426 .

Villaschi S.and Nicosia R .F., 1993, Angiogenic role of basic fibroblast growth factor
released by rat aorta after injury, Am. J. Pathol. 143:182-190 .

Villaschi, S., Nicosia, R.F., 1994, Paracrine interactions between fibroblasts and
endothelial cells in a serum-free co-culture model: modulation of angiogenesis and collagen
gel contraction. Lab. Invest. 71 :291-299.

Villaschi, S., Nicosia, R .F., Smith, M., 1994, Isolation of a morphologically and
functionally distinct musc1e cell type from the intimal aspect of the normal rat aorta .
Evidence for smooth musc1e cell heterogeneity, In Vitra CelloDev. Biol. 30A : 589-595

296
NITRIC OXIDE AND ANGIOGENESIS

Marina Ziehe

Dept. Phannaeology, University of Florenee, Viale Morgagni 65, 50134

Florenee, Italy

1. BACKGROUND

The steps required for new vessel growth are biologieally eomplex and require
eoordinate regulation of eontributing eomponents, including modifieations of cell-cell
interaetions, proliferation and migration of endothelial eells and matrix degradation involving
urokinase-type plasminogen aetivator (uPA) (Mignatti et al., 1991). The observation that in
vivo angiogenesis is aecompanied by vasodilation and that many angiogenesis effeetors
possess vasodilat ing properties, prompted us to seareh for evidenee of a moleeularl
biochemieal link between vasodilation and angiogenesis. Indeed both events oeeur under a
striet eontrol exerted by the endothelial eells on the surrounding eellular eomponents.
Endothelium-dependent relaxation is known to arise from endothelium derived nitrie
oxide (NO) indueing eyclie GMP (eGMP) in the vaseular smooth muscle eell (Moneada and
Higgs , 1993). The synthesis ofNO by endothelial eells ean be bloeked by L-arginine analog s
sueh as NW-mono-methyl-L-arginine (L-NMMA) and LW-nitro-L-arginine methyl ester (L-
NAME), while D-isomers are ineffeetive.
Three experimental observations attraeted our attention to study the role of NO , the
most potent mediator of endothelium -dependent vasodilation (Moneada et al., 1991), in
angiogenesis. First, in experimental models where angiogenesis ean be direetly monitored,
vasodilation and hyperemia of the pre-existing eapillaries and the persistenee of a dilation
state of the newly fonned vessels are typieal findings (Clark and Clark, 1939). Seeond,
several angiogenesis faetors /mediators promote relaxation in vaseular preparations and
peptides known to indueed endothelium-mediated vasorelaxation are angiogenie (Ziehe et
al., 1990; Hu and Fan, 1993). Third, persistent vasodilation is a speeifie feature in tumor
vaseulature and in the surrounding tissue . Furthermore in studying endothelium dependent
vasodilating peptides we demonstrated a role for NO in angiogenesis.

2. THE NO/cGMP PATHWAY IN ANGIOGENESIS

Several reports have implieated a role for NO synthase in angiogenesis. For example,
it was shown that angiogenie aetivity was only released from baeterial endotoxin treated
human monoeytes in the presenee of L-arginine (Leiboviez et al., 1994). Although the

Angiogenesis: Models. Modulators. and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 297
source of angiogenic activity was not identified in the above study, it was nonetheless
blocked by the NO synthase inhibitors. More recently the role of NO in mediating the
angiogenic activity elicited by different cytokines has been described (Montrucchio et al.,
1997). Similarly L-Arginine was shown to favour healing and angiogenesis in gastric
ulcerations while NO synthase inhibitors delayed it (Brozowski et al., 1995). The in vivo
progression of murine haemangiomas, induced by middle T antigen of polyomavirus, is
reduced by canavanina, an NO synthase inhibitor (Ghigo et al., 1995). Other observations
have indicated a cytotoxic/cytostatic effect of NO by melanoma cell line transfected with
NO synthase gene (Xie et al., 1997) and on the vascular development of the chorioallantoic
membrane (pipili-Synetos et al., 1994) suggesting of a dual role ofNO on cell biology.
To asses the role ofNO in angiogenesis we investigated whether NO could act as an
angiogenesis effector per se and as a modulator of the activity and/or the production of
angiogenesis faetors. The relevance of NO in angiogenesis was assessed in vitro by testing
the activity of NO producing agents on the cellular events of angiogenesis in cultured
capillary endothelium, and in vivo in rabbit corneas receiving angiogenesis effector and
tumor cell implant during systemic treatment with NO-synthase inhibitors .
Following this approach we could demonstrate that chemical mediators which
activate the constitutive NO synthase, like substance P (SP) as weil as NO donors, such as
sodium nitroprusside (NaNP), promoted endothelial cell proliferation and migration in vivo
and in vitro , while inhibitors of NO synthase suppressed these responses (Ziehe et al.,
1993; 1994). The growth promoting effect of NO appeared to be linked to cGMP
generation in cultured endothelium (Ziehe et al., 1993), suggesting the existence of an
autocrine/paracrine loop in NO effect. In vivo NO synthase inhibitors suppress
angiogenesis induced by vasodilating effectors like SP and prostagiandin EI (PGE1) but did
not affect angiogenesis induced by fibroblast growth factor-2 (FGF-2) (Ziehe et al., 1994).

3. THE MECHANISM UNDERLYlNG NO-INDUCED ANGIOGENESIS INVOLVES

ENDOGENOUSFG~2PRODUCTION

From the above experiments we learned that the release ofNO at endothelial cell level
was important in triggering the angiogenic cascade. The mechanism by which NO controls
angiogenesis, however, was not c1arified. A possible explanation was that NO could
modulate the activity and/or the production of angiogenesis factors as indicated for other
peptides (Kourembanas et al., 1993). FGF-2 is a potent angiogenesis factor and in vitro
studies have demonstrated that FGF-2 can initiate the cellular responses associated with
angiogenesis including uPA production and cell division (presta et al., 1986; Mignatti et al.,
1989, 1991). However, this cytokine is not secreted in the c1assical sense and probably does
not diffuse, so its role may be limited to stimulation of endothelial cell outgrowth in
autocrine fashion (Schweigerer et al., 1987; Sato and Rifkin, 1988; Tsuboi et al., 1990) or as
a result of endothelial disruption during wounding (McNeil et al., 1989).
In order to elucidate the mechanism of action of NO in angiogenesis, we investigated
the interaction ofNO and FGF-2 in signalling proliferation and uPA production in cultured
coronary endothelium . The vasodilator NaNP, which provides an exogenous source of NO
(Feelish and Noak, 1987), and the neuropeptide SP, which induces NO production in
endothelial cells (Ziehe et al., 1993), were used to increase intracellular NO. In addition,
insight into the basic question was provided by the use of blockers of NO synthase. The
contribution ofFGF-2 was deterrnined from measurements of the levels of FGF-2 peptide
and mRNA, and the use ofneutralizing antibodies. To investigate the mechanisms by which
NO contributed to the angiogenesis process we then assessed whether NO had the ability to
affect cell proliferation and upregulation ofuPA induced by FGF-2 .

298
 2000 400

1500 300 T
>-~
~(ä
_ <lI
.~ CI)

(,l.Q
<lI
1000 200
<0
~~
:::l~

500 100

O-'--'----,--....L....-'---,.---'-
Basal FGF·2 Basal NaNP

Figure 1
Effect of FGF-2 and NaNP on u-PA activity in postcapillary endothelial cells.
Cell-associated uPA activity was measured in subconflu ent postcapillary endothelial cells
after 24 hexposure to 10 ng/ml FGF-2 and 100 11M NaNP , by a spectrophotometric assay
using a plasmin chromog enic substrate (Ziehe et al., 1997). Data are expressed as % increase
over basal respon se and represent the results of at least 3 experiments run in duplicate .

In postcapillary venular endothelium we found that dose-dependent increase in uP A

activity paralleled DNA synthesis in response to FGF-2 and to the NO-donor drug NaN P
(Fig. 1). Western blot analysis revealed that NaNP treatment induced the endothelial cells to
produce endogenous FGF-2. Neutralizing anti-FGF- 2 antibodies blocked uPA produ ction
and DNA synthesis following nipride treatmen t. Similar results were obtained when
increase in NO levels were produced by the use of SP (Fig. 2). Endothelial cells exhibited
increased expression of mRNA and protein of FGF-2 and the proliferat ion induced by SP
was blocked by anti-FGF-2 antibodies. NO synthase inhibitors reduced the increase in
FGF-2 expression due to SP treatmen t. Furthermore , our results indicated that the etTect of
exogenous FGF- 2 on postcapillary endothelial cell proliferation and uPA activity was
independent of NO synthase activation (Table 1), in keeping with the observation that
exogenous FGF-2 elicited angiogenesis in vivo despite the block of NO production by
capillary endothelium (Ziehe et al., 1994). Thus NO, both exogenous and endogenous,
induced the endothelial cells to produce FGF-2 and to respond to the endogenous growth
factor revealing a feedback loop between vasodilation and angiogenesis controlled by the
endothelium via NO and FGF-2 (Ziehe et al., 1997).

4. VEGF, NO AND TUMOR ANGIOGENESIS

Increased NO levels have been found in human tumors (Thomsen et al., 1994; Cobbs
et al., 1995). Most of the cellular components of the tumor mass (the tumor cell themselves
and the immune cell infiltrate) have been shown to generate NO in vitro. However the role
of NO in tumor biology remains poorly understood . Transfeetion of the inducible NO
synthase into a colon adenocarcinoma line gave a celliine that, despite growing more slow ly
in vitro, promoted tumors which grew more rapidly and were more vascularized than wild
type cells (Jenkins et al., 1995). Other observations in agreement with NO being a specific
signal for tumor vascularization, show that blocking NO synthase activity retards the

299
 TABLEI
Role of nitric oxide on cGMP accumulation, proliferation and migration of
postcapillary endothelial cells induced by VEGF and FGF-2
cGMP accumulation 1 Cell proliferation- Cell migration 3
(fmol/mg of proteins) (Total cell number/well) (Migrating cell/filter)

Control conditions

Basal 36.5±5 1287±82 60±3

FGF-2 10 ng!ml 50.4±5 1917±62 94±7
VEGF 10 ng!ml 96.5±15 ** 2042±96 ** 85±5
NaNP 0.1 mM 68±13 ** 1930±98 ** 90±6

+L-NMMA 200 11M

Basal 42±5 1317±72 63±7

FGF-2 10 ng!ml 48±10 1883±136 85±7
VEGF 10 ng!ml 51.4±8 # 1445±70 # 62±6
NaNP 0.1 mM 63+10 ND ND

+LY 83583 1 11M

Basal 53±8 850±150 60±4
FGF-2 10 ng!ml ND 1662±200 75±6
VEGF 10 ng!ml 52±5 1050±235 42±2
NaNP 0.1 mM 59±8 ND ND

1 Cyclic GMP (cGMP) levels were evaluated by radioimmunoas say in coronary post-
capillary endothelial cell (CVEC) extracts. Stimuli were added in the presence of 60 V/mi
superoxide dismutase (SOD) for 10 min. Data are reported as fmol/mg of proteins . Numbers
represent means±SEM of 6 determinations.
2 CVEC proliferation was evaluated as total cell number counted/well after 48 h incubation
with the stimuli . At the end of incubation cells were fixed with methanol and stained with
Diff-Quik. Cells were counted at magnification of 10X with the aid of an ocular grid. Data
represent the meanS±SEM obtained from 2 experiments run in triplicate .
3 The Neuro Probe 48-well microchemotaxis chamber was used to assess cell migration . The
stimuli, prepared in 1% CS medium, were placed in the lower wells and 1 x 104 cell in the
upper weil . The chamber was incubated at 37°C for 4 h. After fixation non-migrating cells
on the upper surface ofthe filter were removed and migrated cells were stained and counted
(at a magnification of 40 X) in 10 random fields per weil. Migration was expressed as the
number oftotal cells counted/well.
FGF -2= fibroblast growth factor-2, VEGF= vascular endothelial growth factor , NaNP=
sodium nitroprusside, ND=not done. Cells were pretreated with 200 11M of the NO
synthase inhibitor L-NMMA or 1 11M of the guanylate cyclase inhibitor LY 83583 (Mulsh
et al., 1988) 1 h before the addition of stimuli. ** P<O.OI vs basal condition and # vs VEGF
alone, ANOV A.

300
 2750
m +anli-FGF-2 mAb

T T
2500 T
Gi
~
-.:: 2250
c CI>
o.c
:= E
...
III ;,
e 2000
~-
... 0
öGi
T
0.
cu
~
1750
I
.......
T ......
.......
T ........
....
........ "'.
...........
........
...........
"'. T
........
......
".",-"
....'""'" .... "'. .......
",.",.

"'....... ...."'. "-"'.r/'

1500
......
....... ........ ........
..-"'-.-
......... ."'......
......... . ....
..-"'-'"
."'..... "'.......
-",-",.

-"'-'" ....
1250
.......
"-"'-tl' ........ "'.......
Basal FGF-2 NaNP SP

Figure 2
Role of endogenous FGF-2 on the proliferative effect of NaNP and SP.
Proliferation of postcapillary endothelial cells was assessed after 3 days of exposure to the
stimuli . The neutralizing anti-FGF-2 MAb or non immune IgGs were added at the dose of 5
ug/ml IO min before the addition of test substances (10 nglml FGF-2 , 100 11M NaNP and
10 nM SP). Data are expressed as means ±SEM of the total cell number counted/well.
(n=4) .

growth of xenografted tumors (Kennovin et al., 1994; Orucevich and Lala, 1996) and
excessive production of NO sustains tumor growth (Doi et al., 1996). Consistently,
dexamethasone exerts antiangiogenic effect leading to reduced tumor growth (Wolff et al.,
1993). Thus, several data exist in support of NO as a signal transducer in tumor
angiogenesis. Based on these considerations we hypothesize that NO released by capillaries
in proximity of a tumor under the control of a local growth factor may favour tumor
angiogenesis.
To then assess the role ofNO in tumor angiogenesis we investigated whether the L-
argininelNO pathway participated to vascular endothelial growth factor (VEGF) and FGF -2
activity . VEGF is a secreted endothelial-specific growth factor that is strongly angiogenic in
vivo (Ferrara and Henzel, 1989; Keck et al., 1989). The observation that VEGF expression
is induced by hypoxia (Schweiki et al., 1992), together with e1evated expression of VEGF in
solid tumors point to a key role for VEGF in tumor angiogenesis (plate et al., 1992; Brown
et al. 1993; O'Brien et al., 1996; Toi et al., 1996). This is supported by transfection of
human VEGF121 into MCF-7 human breast carcinoma cells or mouse VEGF164 into SK-
MEL-2 human me1anomacells where it was shown that VEGF expression enhanced tumor
growth and vascular density (Zhang et al., 1995; Claffey et al., 1996). Moreover, VEGF
effects on permeab ility and vascular tone are coupled to nitric oxide production (Ku et al.,
1993; Wu et al., 1996; Morbidelli et al., 1996)We started by investigating whether the NO
pathway participated to VEGF activity . We found that NO production significantly
contributed to the growth-promoting effect of VEGF, but not for that of FGF-2 (Table 1).
Postcapillary endothelial cell mobilization, adhesion and growth induced by VEGF were
blocked by the NO synthase inhibitor L-NMMA and by the guanylate cyc1ase inhibitor LY
83583, while these inhibitors did not modify FGF-2 effects (Table 1). We then assessed
whether tumor angiogenesis in vivo induced by VEGF could be affected by NO synthase
inhibitors . Transfeetion of VEGF 121 into human breast carcinoma cell line (V12 cells) has

301
 20 20
[S3 + L-NAME • + D-NAME

T
.
GI
0
15
T
T 15

~
()
Ul
T
C
GI
Cl
10 10
.2
Cl
c
Cl:

5 5

o
4 11 16 4 11 16
Time (days) Time (days)

Figure 3
EfTect NO synthase inhibition on VEGF-induced tumor angiogenesis.
Corneal assays were perfonned in female New Zealand albino rabbits (Charies River, Calco,
Corno, Italy). Systemic administration of L-NAME (0.5 g1 in the drinking water) (left
panel)was compared to that of the inactive enantiomer D-NAME (right panel) . Rabbits
received in the left cornea 2.5 x 105 wild-type MCF-7 cells and in the right cornea an equal
number of V12 cells. Subsequent daily observation of the implants was made with a slit
lamp stereomicroscope without anaesthesia . The potency of angiogenic activity was
evaluated on the basis of the number and growth rate of newly formed capillaries, and an
angiogenic score was calculated [vessel density x distance from limbus] (Ziehe et al., 1994,
1997). A density value of 1 corresponded to 0 to 25 vessels per cornea, 2 from 25 to 50, 3
from 50 to 75, 4 from 75 to 100 and 5 for more than 100 vessels. The distance from the
limbus was graded with the aid of an ocular grid. Only the angiogenic response induced by
V12 cells is reported.

been previously shown to enhance tumor growth and vascular density in vivo and prornotes
a strong angiogenic response (Zhang et al., 1995). V12 cells induced a strong angiogenic
response which was efficiently blocked by NO synthase inhibition (Table 2, Fig. 3). VEGF
efTect appears to be selectively linked to the NO pathway. In fact, L-NAME, but not D-
NAME, completely blocked neovascularization induced by the VEGF transfectants, while
the cells remained donnant in the cornea (Fig. 3). Consistently, VEGF induced angiogenesis
was inhibited by NO synthase inhibitors while the angiogenic activity of FGF-2 was not
affected. Thus , nitric oxide appeared to be a downstream imperative ofVEGF but not FGF-
2 induced angiogenesis (Ziehe et al., 1997).

5. CONCLUSIONS

In summary a number of evidence converge in indicating a role for NO in

angiogenesis. Angiogenesis involves the proliferation of endothelial cells under the control of
local peptide growth factors and physical forces. Vasodilation is a component of
angiogenesis leading to the possibility that elevated shear rate may playa role as a stimulus
for neovascularization (Clark and Clark, 1939). The NO synthase/guanylate cyclase is an

302
 TABLE 2
EfTect of L-NAME treatment on the angiogenic activity elicited by wild type (WT)
and VEGF-transfected (VI2) human breast carcinoma cells

Treatment [Positive/total implants perfonned]

1 week 2 weeks 3 weeks

Control
WT 0/4 0/4 114
V12 4/4 4/4 4/4

L-NAME
WT 0/4 0/4 2/4
V12 0/4 0/4 114

Angiogenesis was evaluated in the rabbit corne assay as described in Fig. 3. Data are
expressed as positive implants exhibiting neovascularization over the total implants
perfonned over time (days). An angiogenic response was scored positive at 3-4 days when
the budding vessels travelled approximately 0.5-0.8 mm from the limbal plexus; at 7-10
days when capillaries progressed in the corneal stroma to 1.0-1.8 mm and at 12-20 days
when the new capillaries reached the implanted cell suspension.

ideal signalling mechanism for integrating both chemical and physical influences and is
involved in angiogenesis. Local chemical mediators, such as vasodilating peptides and
VEGF, activate the signalling pathway via associated surface receptors . By contrast, NO
production can be increased in response to elevated shear rate, thereby linking angiogenesis
to flow rate. Our work document that in the acquisition of angiogenic phenotype by
microvascular endothelium, NO production significantly contributes for the growth-
promoting effect of vasodilating peptides and VEGF, but not for that of FGF-2. The
specificity of the NO synthase inhibitors on VEGF induced angiogenesis and not FGF-2
induced angiogenesis provides new evidence for the existence of the two angiogenic
pathways defined by the aebj and avb5 integrins (Frielander et al., 1995). Thus, although
the NO pathway integrates several chemical and physical modulators of the angiogenic
process, not all angiogenic factors depend on this signalling cascade. This highlights the
problem that arises as a result of tumors secreting multiple angiogenic factors, namely, that
blocking of any single angiogenic factor is unlikely to be an effective antitumor strategy.
More work with specific inhibitors of the NO synthase isofonns is needed to fully
elucidate the role of NO in angiogenesis and tumor growth. Nevertheless the nitric oxide
pathway stays as a promising target for consideration in pro- and anti-angiogenic
therapeutical strategies .

6. ACKNOWLEDGEMENTS

This work was supported : by funds from Italian Ministry for the University and for
Scientific and Technological Research (MURST), National Research Council of Italy (CNR ,
grant n. 95.02983.CTl4), AIRC (Special Project "Angiogenesis") and European
Communities (Biomed 2 Programme "Angiogenesis and cancer" -PL 950669).

303
7. REFERENCES

Brown , L.F., Berse, B., Jackman, RW., Tognazzi, K , Manseau, E.J., Dvorak, H.F., and
Senger, D.R. 1993. Increased expression of vascular permeability factor (vascular
endothelial growth factor) and its receptors in kidney and bladder carcinomas. Am. J.
Pathol. 143: 1255-1262.
Brzozowski, T., Kounturek, S.J., Drozdowicz, D., Dembinski, A , and Stachura, J. 1995.
Healing of chronic gastric ulcerations by L-arginine. Digestion 56: 463-471.
Claffey, KP., Brown, L.F., del Aguila, L.F ., Tognazzi, K , Yeo, K-T., Manseau , E.J., and
Dvorak, H.F. 1996. Expression of vascular permeability factor/vascular endothelial
growth factor by melanoma cells increases tumor growth, angiogenesis, and
experimental metastasis . Cancer Res. 56: 172-181.
Clark, E.R., and Clark, E.L. 1939. Microscopic observations on the growth of blood
capillaries in the living mamma!. Am. J. Anat. 64:251-299
Cobbs, C.S., Brenman, J.E., Alpade, KD., Bredt, D.S., and Israel, M.A. 1995. Expression
ofnitric oxide synthase in human central nervous system tumors. Cancer Res. 55: 727-
730.
Doi, K, Akaike, T., Horie, H., Noguchi, Y., Fuji, S., Beppu, T., Ogawa, M., and Maeda, H.
1996. Excessive production ofNO in rat solid tumor and its implication in rapid tumor
growth . Cancer 77: 1598-1604.
Feelish, M., and Noack, E. 1987. Nitric oxide (NO) formation from nitrovasodilators occurs
independently ofhemoglobin or non-heme iron. Eur. J. Pharmacol. 142:465-469.
Ferrara , N., and Henzel , WJ. 1989. Pituitary follicular cells secrete a novel heparin-binding
growth factor specific for vascular endothelial cells. Biochem. Biophys. Res. Commun.
161: 851-858 .
Friedlander, M., Brooks, P.C., Shaffer, RW., Kincaid, C.M., Varner, J.A ., and Cheresh,
D.A 1995. Definition of two angiogenic pathways by distinct av integrins. Science
270: 1500-1502.
Ghigo, D., Arese, M., Todde, R., Vecchi, A, Silvagno, F., Costamagna, C., et a!. 1995.
Middle T antigen-transformed endothelia! cells exhibit an increased activity of nitric
oxide synthase. J. Exp. Med. 181: 9-19.
Hu, D.E., and Fan, T.-P.D. 1993. [Leu8]des-Arg9-bradykinin inhibits the angiogenic effect
ofbradykinin and interleukin-l in rats. Br. J. Pharmacol. 109: 14-17.
Jenkins, D.C., Charles, I.G., Thomsen, L.L., Moss, D.W., Holmes, L.S., Baylis, SA,
Rhodes, P., Westmore, K, Emson, P.C., and Moncada ., S. 1995. Role of nitric oxide in
tumor angiogenesis. Proc. Natl. Acad. Sei. USA 92: 4392-4396.
Keck, PJ., Hauser, S.D., Krivi, G., Sanzo, K, Warren, T., Feder, J., and Conolly , D.T.
1989. Vascular permeability factor, an endothelial cell mitogen related to PDGF .
Science 246: 1309-1312.
Kennovin, G.D., Hirst, D.G., Stratford, M.R.L., and Flitney, F.W. 1994. Inducible nitric
oxide synthase is expressed in tumor associated vasculature; inhibition retards tumor
growth in vivo. In Biology of Nitric Oxide Vol. 4 Enzymology, Biochemistry and
Immunology. S. Moncada, M. Feelish, R Busse, and EA Riggs, editors . Portland
Press, London . 473-479 .
Kourembanas, S., McQuillan , L. P., Leung, G. K., and Faller, D. V. 1993. Nitric oxide
regulates the expression of vasoconstrictors and growth factors by vascular
endothelium under both normoxia and hypoxia. J. Clin. Invest. 92:99-104.
Ku, D.D., Zaleski, J.K , Liu, S., and Brock, TA 1993. Vascular endothelial growth factor
induces EDRF-dependent relaxation in coronary arteries. Am. J. Physiol. 265: H586-
H592 .

304
Leibovicz, JS, Polverini , PJ., Fong , TW., Harlow, LA, and Kock , A,E. 1994. Production
of angiogenic activity by human monocytes requires an L-arginine/nitric oxide-
synthase-dependent effector mechanism. Proe. Natl. Acad. Sei. USA 91: 4190-4194.
McNeil, P.L., Muthukrishnan, L., Warder, E., and D'Amore, P.A 1989. Growth factors are
released by mechanically wounded cells. J. Cell Bio!. 109: 811-822.
Mignatti, P., Mazzieri, R., and Rifkin, D.B. 1991. Expression of the urokinase receptor in
vascular endothelial cells is stimulated by basic fibroblast growth factor. 1. Cell Biol.
113: 1193-1201.
Mignatti, P., Tsuboi, R., Robbins, E., and Rifkin, D.B . 1989. In vitro angiogenesis on the
human amniotic membrane : requirement for basic fibroblast growth factor-induced
proteinases.1. Cell Biol. 108: 671-682 .
Moncada, S., and Higgs, A. 1993. The L-arginine-nitric oxide pathway . New Eng!. J. Med.
329 : 2002-2012.
Moncada, S., Palmer, R . M. J, and Higgs, A. 1991. Nitric Oxide: physiology,
pathophysiology, and pharmacology. Pharmaeo!. Rev. 43:109-142.
Montrucchio, G., Lupia, E., De Martino, A , Battaglia, E., Arese, M ., Tizzani, A.,
Bussolino, F., and Camussi, G. 1997. Nitric oxide mediates angiogenesis induced in
vivo by platelt-activating factor and tumor necrosis factor. Am. J. Pathol. 151 : 557563.
Morbidelli, L., Chang, c.-H., Douglas, lG., Granger, HJ., Ledda, F., and Ziche, M. 1996.
Nitric oxide mediates the mitogenic effect of VEGF on coronary venular endothelium.
Am J. Physiol. 270(39): H411-H415 .
Mulsh, A , Buse , R., Liebau, S., and Fostermann, U. 1988. LY 83583 interferes with the
release of endothelium-derived relaxing factor and inhibits soluble guanylate cyclase. 1.
Pharmaeo!. Exp. Ther . 247 : 283-288 .
O'Brien, T., Cranston, D., Fuggle, S., Bicknell, R., and Harris, AL. 1996. Different
angiogenic pathways characterize superficial and invasive bladder cancer. Cancer Res.
55 : 510-513 .
Orucevic, A , and Lala., P.K. 1996. NG-nitro-L-arginine methy l ester, an inhibitor of nitric
oxide synthesis, ameliorates interleukin 2-induced capillary leakage and reduces tumour
growth in adenocarcinoma-bearing mice. Br. 1. Cancer 73: 189-196 .
Pipili -Synetos, E., Sakkoula, E., Haralabopoulos, G., Andiopuolou, P., Peristeris, P., and
Maragoudakis, M .E. 1994. Evidence that nitric oxide is an endogenous antiangiogenic
mediator. Br. 1. Pharmaco!. 111 : 894-902 .
Plate, K .H., Breier, G., Weich, H.A ., and Risau, W. 1992. Vascular endothelial growth
factor is a potential tumor angiogenesis factor in human gliomas in vivo . Nature 359 :
845-848.
Presta, M ., Moscatelli, D., Joseph -Silverstein, J, and Rifkin, D.B. 1986. Purification from a
human hepatoma cell line of a basic fibroblast growth factor-like molecule that
stimulates capillary endothelial cell plasminogen activator production, DNA synthesis
and migration . Mo!. Cell. Bio!. 6:4060-4066.
Sato, Y , and Rifkin, D .B. 1988. Autocrine activities of basic fibroblast growth factor:
regulation of endothelial cell movement , plasminogen activator synthesis and DN A
synthesis. J. Cell Biol. 107:1199-1205 .
Schweigerer, L., Neufeld , G., Freidman, J , Abraham , JA ., Fiddes, Je. , and
Gospodarowicz, D. 1987. Capillary endothelial cells express basic fibroblast growth
factor, amitogen that prornotes their own growth . Nature 352 :257-259 .
Shweiki, D ., Hin, A., Soffer, D., and Keshet, E. 1992. Vascular endothelial growth factor
induced by hypoxia may mediate hypoxia-initiated angiogenesis. Nature 359 :843-845 .
Thomsen, L.L., Lawton, F.G., Knowles, R.G., Beesley, JE , Riveros-Moreno, v., and
Moncada, S. 1994. Nitric oxide synthase activity in human gynecological cancer.
Cance r Res . 54: 1352-1354 .

305
Toi, M., Kondo, S., Suzuki, H., Yamamoto, Y., Inada, K., Imazawa, T., Taniguehi, T., and
Tominaga, T. 1996. Quantitative analyses of vaseular endothelial growth faetor in
primary breast cancer . Cancer Res. 77: 1101-1105.
Tsuboi, R., Sato, Y., and Rifkin, D.B. 1990. Correlation of cell migration, cell invasion,
reeeptor number, proteinase produetion , and basic fibroblast growth faetor levels in
endothelial eells. J. Cell Biol. 110:511-517.
Wolff, J.E.A, Guerin, C., Laterra, J., Bressler, J., Indurti, R.R., Brem, H ., and Goldstein,
G.W. 1993. Dexamethasone reduces vaseular density and plasminogen aetivator
aetivity in 9L rat brain tumors. Brain Res . 604: 79-85.
Wu, H.M. , Qiaobing, H., Yuan, Y., and Granger, H.J. 1996. VEGF induees NO-dependent
hyperpermeability in eoronary venules . Am. J. Physiol. 269 (38): C1371-C1378, 1995.
Xie, K" Huang, S., Dong, Z., Juang, S.-H" Wang, Y., and Fidler, I.J. 1997. Destruetion of
bystander eells by tumor cells transfected with indueible nitrie oxide (NO) synthase
gene. J. Natl. Cancer Inst . 89: 421-427.
Zhang, H.-T., Craft, P., Seott, PAE., Ziehe, M., Weich, HA, Harris, AL., and Bicknell,
R. 1995. Enhaneement of tumor growth and vaseular density by transfeetion of
vaseular endothelial eell growth faetor into MCF-7 human breast eareinoma cells. J.
Natl. Cancer Inst . 87: 213-217.
Ziehe, M ., Morbidelli, L., Choudhuri, R., Zhang, H.-T ., Donnini, S., Granger, HJ., and
Bieknell, R. Nitrie oxide-synthase lies downstream of Vaseular Endothelial Growth
Faetor but not basic Fibroblast Growth Faetor indueed angiogenesis. J. Clin. Invest. ,
99: 2625-2634, 1997.
Ziehe, M., Morbidelli, L., Geppetti , P., Maggi, CA, and Dolara, P. 1991. Substanee P
induees migration of eapillary endothelial eells: a novel NK- I seleetive reeeptor
mediated aetivity . Life Sei . 48:PL7-PL11.
Ziehe, M., Morbidelli, L., Masini, E., Amerini, S., Granger, HJ., Maggi , CA, Geppetti, P.,
and Ledda, F. 1994. Nitrie oxide mediates angiogenesis in vivo and endothelial cell
growth and migration in vitro promoted by substanee P. J. Clin. Invest. 94 :2036-2044.
Ziehe, M., Morbidelli, L., Masini, E., Granger, HJ., Geppetti, G., and Ledda, F. 1993.
Nitrie oxide prornotes DNA synthesis and eyclic GMP formation in endothelial eells
from posteapillary venules . Biochem. Biophys. Res. Commun . 192 (3): 1198-1203 .
Ziehe, M., Morbidelli, L., Paeini, M., Geppetti , P., Alessandri, G., and Maggi, CA 1990.
Substanee P stimulates neovaseularization in vivo and proliferation of cultured
endothelial eells. Microvasc. Res . 40: 264-278.
Ziehe, M., Morbidelli, L., Parenti, A., Amerini, S., Granger, HJ., and Maggi, CA 1993.
Substanee P inereases eyclie GMP levels on eoronary posteapillary venular endothelial
eells. Life Sei. 53: 1105-1112.
Ziehe, M., Parenti , A, Ledda, F., Dell'Era, P., Granger, H.J., Maggi, C.A, and Presta, M .
1997. Nitrie oxide prornotes proliferation and plasminogen aetivator produetion by
eoronary venular endothelium through endogenous bFGF. Circ. Res ., 80: 845-852

306
NITRIC OXIDE: A PROMOTER OR INHIBITOR OF ANGIOGENESIS

Eva Pipili-Synetos & Michael E. Maragoudakis

Department ofPharmacolgy, Medical School,

University of Patras, Patras 261 10, GREECE

INTRODUCTION

Nitric oxide (NO) has been shown in the last decade to be an ubiquitous messenger
with a remarkably wide diversity of biological actions. Its role as a major regulator in the
nervous immune and cardiovascular systems as weil as in pathophysiological states (septic
shock, hypertension, stroke and neurodegenerative diseases) is increasingly appreciated
(Gross & Wolin, 1995).
Nitric oxide has been implicated in a number of studies in the progression of tumour
growth and metastasis. The results from these studies however are rather contradictory. A
large amount of evidence suggests that tumour growth and metastasis is followed by
depressed NO synthesis (Umansky et a1., 1996a,b, Hidetsugu et al., 1996, Xie et al.,
1995, Rollo et al., 1996, Robertson et al., 1996, Yim et al., 1993, Lee & Wurster, 1994,
Lejeune et al., 1994, Yu et al., 1996, Sanchez-Bueno et al., 1996, Xie et al., 1995). On the
other hand, Thomsen et al., (1994,1995) have shown that NO synthase is present in fresh
human tumour tissue and its activity correlates positively with tumour grade .
Furthermore, Jenkins et al., (1995) showed that cells engineered to generate NO
continuously, grew more slowly in vitro than the wild type parental cells. In nude mice
however , the tumours from these cells grew faster than those derived from wild type cells
and were more vascularized. Similar data have been obtained by Edwards et al., (1996) .
These observations suggest a dual role for nitric oxide in tumour growth, pro- and anti-
tumour depending on the local concentration of the molecule. Such a dual role of NO on
specific immunity and particularly on Iymphocyte proliferation, has also been proposed to
be due to quantitative differences in NO production, i.e. low concentrations are required for
Iymphocyte replication while excess amounts are cytotoxic (Rodeberg et al., 1995, Efron et
al., 1991, Albina et al., 1991).
Nitric oxide deiived from the inducible NO-synthase can be produced in large and
continuous amounts which mediate several aspects of cytotoxicity against microbes and

Angiogenesis: Models . M odulat ors. and Clinical Applications

Edited by Maragoudak is, Plenum Press, New York, 1998 30-7
tumour cells (Nathan & Hibbs, 1991, Stuher & Griffith, 1992). The mechanism of these
NO effects remains unclear although several pathways have been proposed . Superoxide
anion and NO interact in vitro to form peroxynitrite, aprecursor to two potent oxidants,
hydroxide anion and nitrite (Beckman et al., 1990, Hogg et al., 1992). Although there is no
evidence that this reaction takes piace in vivo, it is possible that NO cytotoxicity relies on
these and other reactive nitrite and oxygen intermediates. NO is known to bind to haeme
groups . It has therefore been proposed that NO binds to the prosthetic groups of
aconitase and complexes I and 11 of the mitochondrial electron transport chain resulting in
cell death (Hibbs et al., 1990, Stuher & Nathan, 1989). In addition, NO causes DNA
fragmentation which contributes to apoptotic cell death (Filep et al., 1996, Nishio et al.,
1996, Sanchez-Bueno et al., 1996).
Nitric oxide causes vasodilatation which is a c-GMP-dependent event (Murad,1986,
Miki et al., 1977, Katsuki et al., 1977) and is followed by increases in vascular
permeability. Increased vascular permeability is considered as a fascilitatory step in the
angiogenic process . New vessel growth is aprerequisite for tumours to grow and
metastasize (Folkman & Singh, 1992). In agreement with the observations which relate NO
to vasodilatation and vascular permeability, NO has recently been shown mediate
angiogenesis induced by substances which increase vascular permeability, such as vascular
endothelial growth factor (VEGF) and substance P (Ziehe et al., 1994, Morbidelli et al.,
1996).
Nitric oxide is also involved in the inhibition of platelet aggregation and inhibition of
platelet and monocyte adhesion to the vessel wall (Moncada et al., 1988, Bath et al., 1991).
Angiogenesis and metastasis are known to share several common features (Liotta et al.,
1991). They both require the adhesion ofhomologous or heterologous cell aggregates on to
the vessel wall with subsequent invasion of the extracellular matrix. Koch et al., (1995)
have shown that soluble E-selectin and VCAM-l are chemotactic to endothelial cells in
vitro and angiogenic in rat cornea in vivo. Cheresh and co-workers (Brooks et al., 1994,
Friedlander et al., 1995) have recently discovered a role for a, integrins in angiogenesis and
were able to inhibit basic fibroblast growth factor (bFGF) - tumour necrosis factor alpha
(TNF-a) induced angiogenesis using antibodies to a;a3 and transforming growth factor alpha
(TGFa)-VEGF-induced angiogenesis using antibodies to a,as integrins respectively .
Invasion following adhesion requires the production of proteolytic enzymes (Liotta et al.,
1991) and relaxation of intercellular junctions which is a result of vasodilatation induced
distention and increase in permeability. Nitric oxide being a vasodilator as weil as an
inhibitor of adhesion, again may have a dual role in angiogenesis and metastasis by
potentially facilitating or inhibiting vasodilatation, permeability and adhesion.
It is weil established that many angiogenesis dependent diseases (tumour growth ,
psoriasis) are accompanied by increased thrombogenesis (Rickles & Edwards ., 1983,
McDonald, 1991). A major component of thrombogenesis is platelet aggregation and
adhesion. Recently it was shown that thrombin, one of the most potent activators of
platelets, stimulates angiogenesis (Tsopanoglou et al., 1993). These result suggest that
platelet activation may be implicated in the angiogenic process. As mentioned above, nitric
oxide inhibits both platelet aggregation and adhesion and in this capacity it is a potential
antiangiogenic mediator. This possibility along with the above mentioned reports where
NO correlates inversely with tumour growth and metastasis , prompted our investigation of
the role ofNO on angiogenesis and tumour progression. This was done by testing he effect
ofincreasing or decreasing NO availability in in vitro and in vivo systems of angiogenesis
and tumour growth (Pipili-Synetos, 1994,1995).

308
METHODS

The in vivo chick chorioallantoic membrane (CAM) angiogenesis model, initially

described by Folkman (1985) and modified as previously reported (Maragoudakis et aI.,
1988) was used . Briefly, fresh fertilized eggs were incubated for 4 days at 370C when a
window was opened on the egg shell exposing the CAM . The window was covered with
sterile cellophane tape and the eggs were retumed to the incubator until day 9 when the test
materials were applied . The test materials or vehicle and 0.5 j.tCi [U)4C] labeled proline,
were placed on sterile plastic discs and were allowed to dry under sterile conditions.
The control discs (containing vehicle and radiolabelIed proline) were placed on the CAM
one cm away from the disc containing the test material. A sterile solution of cortisone
acetate (249 nmoles/disc) was routinely incorporated in all discs in order to prevent an
inflammatory response . The loaded and dried discs were inverted and placed on the
CAM, the windows were covered and the eggs incubated until day 11 when assessment
of angiogenesis took pIace.

Biochemical evaluation 0/ angiogenesis

Collagenous proteins represent 80% of the total basement membrane proteins
formed by the CAM during the chick embryo development (Maragoudakis et al., 1988).
The extent of their biosynthesis has been shown to correlate weil with new vessel
formation . This biosynthesis reaches a maximum between days 8 and 11 and coincides
with maximal angiogenesis in the CAM as shown by morphological evaluation of
vascular density . Furthermore, at day 10, collagenous protein biosynthesis is l l-fold
higher than that of day 15 when angiogenesis has reached a plateau . Biochemical
evaluation of newly formed vessels was performed by determining the extent of
collagenous protein biosynthesis in the CAM Iying directly under the discs . Briefly , the
area under the disc was cut off, placed in an appropriate buffer and protein biosynthesis
was stopped. Non protein bound radioactivity was removed by washing with
trichloroacetic acid. Discs containing radioactivity were resuspended and subjected to
collagenase digestion . The resulting radiolabeled tripeptides corresponding to basement
membrane collagen and other collagenous material synthesized by the CAM from
[U- 14C]-proline, were counted and expressed as cpm/mg protein.

Morphological evaluation 0/ angiogenesis

For morphological evaluation, eggs were treated as above in the absence of
radiolabeled proline . At day 11 the eggs were flooded with 10% buffered formalin, the
plastic discs were removed and the eggs were kept at 370C until dissection. A large area
around the discs was cut off and placed on a gIass slide and the vascular density index
(expressed as number ofblood vessels) was measured by the method of Harris-Hooker et
al. (1983) . Harris-Hooker evaluation underestimates by approximatel y 10% (compared to
the biochemical evaluation of angiogenesis) the changes in the vascular network. This is an
expected limitation ofthis method as some vessels are probably collapsed and do not show
up under the stereoscope.

Determination 0/ NO releasefrom the CAM in vitro

The CAM from day 9 embryos was used for the determination of NO release in
vitro . CAM from 20 eggs was dissected into 4 pieces each into a Petri dish containing

309
Hanks balanced salt solution (HBSS) pH 7.4. Thirty six (36) of these pieces were then
divided between three beakers (control and two treatment groups) containing 10ml of
HBSS alone (wet weight of tissue 0.83g), or the appropriate amounts of isosorbide
mononitrate (ISMN) (wet weight of tissue 0.83g) or isosorbide dinitrate (ISDN) (wet
weight of tissue 0.94g) dissolved in HBSS and were maintained at room temperature.
Sampies were taken 5 min after the introduction of the tissue into the HBSS, according to
the following protocol. 100~1 from each sampie were added to polypropylene vials
containing the reaction mixture which consisted ofH202 (500~M), luminol (30~M) and an
appropriate amount of HBSS to make up a total volume of 500~1. The vial was then
stirred vigorously and the emission was recorded in a Berthold Autolumat LB953
luminometer. Chemiluminescence peaks were converted to nmoles of NO by fitting them
on to a standard curve constructed with increasing concentrations of pure NO as
previously described (Delikonstantinos et al., 1995). Results were expressed as nmoles.g-l
of wet weight of tissue .

Matrigel tube formation assay and assessment ofthe area.

Human umbilical vein endothelial cells were isolated from fresh umbilical cords
according to the method previously described by laffe et. al. (1987). Cells were maintained
in medium (supplemented with ECGS, FBS, and antibiotics) at 370C, 95% humidity, and
5% CO 2. Matrigel consists of reconstituted basement membrane enriched with laminin
extracted from the Engelbredt-Holm Swarm (EHS) tumour (Timpl et al., 1979). When
endothelial cells are cultured on this material, they stop proliferating, they align and within
12-18 hours they form an anastomosing network of tube-like structures resembling a
capillary plexus (Grant at al., 1989). The Matrigel tube formation assay was performed as
previously described (Kubota et al., 1988). Briefly, 300~I/well Matrigel was used to coat
24-well Limbro cluster plates at 40C. After the Matrigel had formed a gel, approximately
40,000 cells were seeded in each weil in medium containing the test substances and 10%
FBS. After 18 hours of incubation the wells were fixed, stained and the tube area was
quantified using the MCID image analysis system from Brock University, Ontario,
Canada, according to Grant et. al. (1989).

Determination ofcGMP release from the CAM in vitro

At day 4 the windows were opened as described above and the eggs were retumed for
a further 7 days to the incubator without placing any discs on the exposed surface. At day
11, the exposed CAMs from 15 eggs were dissected into 4 pieces each and divided
between two Petri dishes (control and treatment group) containing ice-cold Krebs
physiological salt solution (PSS) and 0.5mM isobutyl-methyl-xanthine (ffiMX) to prevent
the degradation of cGMP . Subsequently, they were placed in round bottomed
polypropylene centrifuge tubes standing on ice and containing 5ml of Krebs PSS and
ffiMX with or without 2.8 x 10-5 M SNP. The zero time sampie was transferred
immediately in a deep freeze (-400C) until further processing. The remaining sampies
(with or without SNP) were incubated for 2,5 and 10 min at 370C in a shaking water bath .
At the end of the incubation period, all sampies were homogenized in a Polytron
Homogenizer at 40C in the presence of 6% trichloroacetic acid (TCA) and centrifuged at
2,00Oxg for 15min at 40C. The supematant was recovered, and washed 4 times with 5
volumes ofwater saturated diethyl ether. The aqueous extract remaining was Iyophillized
and reconstituted in a suitable volume of assay buffer immediately prior to analysis .

310
 a
40 b
. ISMN 0 ISDN

:J
0 50
a:
I-

5o 30 e- 4O
u. c:
0 8
~ 20 ~ 3O
al 1Il
a, 0-
U o
13 ö 20
g 10 s
g
.s J!.
z ~ 10
E

8.050.1 3 03 10 300
log NMOLES/DISC nmok../d,SC

F ig. Ia.b, (a) Effect of sodium nitroprusside (SNP) on angiogenesis in the chick
chorioallantoic membrane (CAM) in vivo, expressed as collagenous protein
biosynthesis (CPB). (b)Effect of ISMN (closed circles) and ISDN (open
circles) on angiogenesis in the chick chorioallantoic membrane in vivo ,
expressed as collagenous protein biosynthesis (CPB) . Results are expressed as
mean±s .e ofthe mean % of control.

cGMP determ ination was performed b y radioimmunoassay (RIA) using a

commercially available kit (cGMP[l25I] assay system, Amersham UK) . Results are
expressed as pmoles /mg ofprotein.

RESULTS AND DISCUSSION

The NO donors SNP, ISMN and ISDN caused a dose-dependent inhibition of

angiogenesis in the CAM in vivo (Fig. la , 1b). In contrast the inhibitors of NO synthase
NG-monomethyl -L-arginine (L-NMMA) and NG-nitro-L-arginine meth ylester (L-
NAME), increased angiogenesis in the CAM dose-dependently an effect which was
completely reversedin both cases by the natural substrate L-arginine (Fig. 2). Sodium
nitroprusside (SNP) stimulated cGMP formation in the Cam in vitro. In addition in the
matrigel tube formation assay both SNP and the permeable cGMP analog Br-cGMP
inhibited tube formation dose-dependently fl'tptli -Syne toa These results suggest that the
observed effects of NO donors on angiogenesis may be at least partly mediated via an
increase in c-GMP . Nitric oxide-induced cGMP increases are associated with
antiproliferative phenomena in various cell types including vascular endothelial cells (Yang
et al., 1992), smooth muscle cells (Kariya et al., 1989, Garg & Hassid , 1989, Assender et
al., 1992, Cornwell et al., 1994) and with inhibition of Ca++ mobilization , a process
necessary in a variety of cell signalling pathwa ys. There is, however , some evidence
pointing to the opposite effect since Ziehe and co-workers (Ziehe et al., 1993, 1994,
Morbidelli et al., 1996) have shown that in cultured microvascular endothelium isolated
from coronary postcapillary venules, NO-donors stimulated mitogenesis via a cGMP-
dependent mechanism . These conflicting effects of cGMP could be explained by the fact
that this nucleotide has more than one protein targets (McDonald and Murad , 1996)

3 11
 • L.NAME 0 L·NMMA

70

'e
c:
56

8
'ö
!. 42
CD
c,
U
'ö 28

i'tl"
E 14

10 40
nmolesrdisc

Fig.2. Effect of increasing amounts of (a) NG-monomethyl-L-arginine (L-NMMA,

open circles) and of (b) NG-nitro-L-arginine methyl ester (L-NAME, closed
circles) on angiogenesis in the chick chorioallantoic membrane in vivo,
expressed as collagenous protein biosynthesis (CPB) . Results are expressed as
meanrs.e of the mean % of contro\.

including the G-kinases, the c-GMP gated Ca++ channels and the cyclic nucleotide
phosphodiesterases. Whethe r cGMP will stimulate or inhibit a process depends on
complex interactions and equilibria in a given biological system and the specific
microenvironment under study although more often than not the balance appears to tip in
favour of cGMP causing inhibition rather than activation of cellular functions (Schmidt et
a\., 1993).
SNP, ISMN and ISDN not only inhibited angiogenesis in the CAM but were capable
of reversing completely the angiogenic effects of thrombin . This effeet was mimicked by
superoxide dismutase (SOD) which increases the half life ofNO, suggesting that an increase
in the availability ofNO is a negative regulator signal for both basal and thrombin-induced
angiogenesis in the CAM .
Angiogenesis is a potential target for controlling tumour growth and metastasis. It
was examined whether NO which inhibits angiogenesis in the CAM, might also inhibit
tumour progression. For this purpose the long acting vasodilators ISMN and ISDN which
act through NO release were tested in mice implanted with Lewis Lung carcinoma (LLC).
For this purpose the long aeting vasodilators ISMN and ISDN , which act through NO
release were used and their antitumour and metastatic effects were evaluated . Both
vasodilators in addition to their inhibitory effect in angiogenesis in the CAM , reduced
tumour growth and metastatic foci in the lungs of mice implanted with LLC (Fig. 3a,b).
Neither agent affected the proliferation of LLC in culture indicating that the antitumour
effects of ISMNIISDN were not a consequence of a direct cytotoxic action on tumour cells.
Furthermore, neither compound had any effect on the growth of endothelial cells from
different origins in culture, suggesting that their antiangiogenic effect may not be due
to inhibition of endothelial cell proliferation.

312
 ISDN

a b
4
3

3
M
.Q.
'"
E
.
E
E
2

">
ö 2 '"
'0
.2
:; Ö
0
E 0
c:
2

o
2OOug/day 600ug/day centrot ISMN ISDN

Fig.3a,b. (a) Effect of 200/tg and 600/tg of ISMN or ISDN injected daily (for 14days)
into mice inoculated with 2xl06 cells of LLC, on tumour size(measured with a
microvemier using the formula (axbxc)/2 crrr', where a=length, b=width and
c=depth at the site of inoculation). Results are expressed as meanrs.e crrr' and
are compared to the controls by unpaired t-test, Effect of 200/tg of ISMN or
ISDN injected daily (for 14 days) into mice inoculated with 2xlO6 cells ofLLC,
on the development of pulmonary metastatic foci. Mean±s .e number of foci
was calculated by dividing the sum ofthe number ofmetastases for each animal
within a group, by the number of animals of the group . Asterisks denote
statistical significance between control and test. * p<O.05, *** p<O.OI . (b)

Since NO does not affect endothelial cell proliferation, which vascular does it interact
with? A possible target for NO might be the blood platelet since both platelet aggregation
and adhesion are inhibited by NO (Radomski & Moncada, 1991). Angiogenesis and
tumour growth and metastasis are processes which are believed to involve platelet
activation (Tsopanoglou et al., 1993, Rickles & Edwards, 1983, Karpatkin et al., 1988,
Gasic et al., 1968, 1973, Pearlstein et al., 1984, Nierodzik et al., 1991, 1992, Honn et al.,
1992). Activated platelets may adhere to the vascular endothelium, increase cell
permeability, and initiate proliferative phenomena through the release of growth factors
(page, 1988) which are weil characterized angiogenic molecules (Folkman & Shing, 1992).
Furthermore, antibody- induced thrombocytopenia has been shown to markedly reduce
metastases in experimental tumour models including LLC (Gasic et al.,1968, Pearlstein et
al., 1984). The effect however of antiplatelet agents such as aspirin and prostacyclin, on
metastasis has been controversial and non-reproducible (Gasic et al., 1973, Honn et al.,
1981, Karpatkin et al., 1988). Interestingly, in all these studies, the antiplatelet agents
used had little or no effect on platelet adhesion in contrast to NO which inhibits this
process. Karpatkin et al. (1988) were able to show that agents inhibiting the interaction
between platelets and adhesive proteins such as factor VIII and fibronectin, were able to
inhibit pulmonary metastases in mice induced by three different tumour cell lines.

313
Furthermore, Brooks et al. (1994) have been able to establish a strong connection between
a3
the expression ofthe vascular integrin av (the endothelial cell receptor for factor VIII and
fibronectin) and angiogenesis in the CAM. It is therefore likely that the effects of NO in
angiogenesis and tumour growth and metastasis are at least partly mediated via inhibition of
platelet adhesion (a) to the vascular wall and/or (b) to tumour cells creating aggregates
which then interact with the vascular wall. The fact that NO-releasing compounds
completely reverse the angiogenic effect of thrombin further supports the above
hypothesis since thrombin, a major platelet activator, has been shown to enhance platelet-
tumour adhesion (Nierodzik et al., 1991, 1992) and metastasis in tumour animal models ..
To test whether platelets are indeed involved in the angiogenic process, it was
examined wheher they affect tube formation on matrigel . In this system unstimulated
platelets increase tube formation by about 80% over the control (no platelets), whereas
thrombin-stimulated platelets increase tube formation by about 130% (unpublished
observations). Under these conditions SNP and br-c-GMP reduce the effect of thrombin .
The involvement of adhesion molecules in these process is also investigated
The above results, are in agreement with reports by others (Umansky et al., 1996a,b,
Hidetsugu et al., 1996, Xie et al., 1995, Rollo et al., 1996, Robertson et al., 1996, Yim et al.,
1993, Lee & Wurster, 1994, Lejeune et al., 1994, Yu et al., 1996, Sanchez-Bueno et al.,
1996, Xie et al., 1995a) which suggest that NO restricts tumour growth and metastasis. In
contrast, Jenkins et al., 1995, Thomsen eta1.1994,1995 and Edwards et al., 1996, report
data suggesting that NO may favour tumour growth and vascularization. These authors,
however, state that the NO effects depend on the local concentration ofthe molecule in the
microevironment of the system under study and suggest a dual role for NO, both pro- and
anti-tumour.
Can therefore NO be considered as a promoter or an inhibitor of angiogenesis? It is
likely that it may have a dual role in this process since, in contrast to our observations,
there are reports by Ziehe and coworkers (Ziehe et al., 1994, Morbidelli et al., 1996) and
Leibovich et al., (1994) who suggest that in the cornea assay of angiogenesis a NO-
synthase effector mechanism is necessary for the expression of the angiogenic activity
of substance P, PGE1, VEGF and LPS-stimulated macrophages . These authors do not
show a direct effect of NO-donors in this system, which suggests that the mechanisms
involved are complex and the interpretation may not be straightforward.
A parameter which should be considered in all cases is the local concentration of NO
in different biological settings. A difference of two orders of magnitude might produce
opposing results, stimulating and inhibitory respectively as suggested by Jenkins et al.,
(1995) and Rodeberg et al., (1995) . In a biological setting where only eNOS is activated,
giving rise to small amounts of NO angioproliferative effects may depend on the NOS
pathway. In a setting where the iNOS continuously produces high amounts ofNO, as in
the case oftumorigenesis, the balance might be tipped in favour of the inhibitory effects of
NO .
In conclusion, the data from our laboratory suggest that NO is an inhibitor of
angiogenesis, tumour growth and metastasis. However, the possibility that NO may
promote a process which fascilitates angiogenesis in different biological settings and
conditions to the ones used in our study, cannot be excluded. When NO functions in the
capacity of an inhibitor of new vessel growth, NO-donors may be useful in controlling
angiogenesis-dependent disease. In addition the resulting vasodilatation may facilitate the
availability of drugs within the ischemic areas of a tumou r. On the other hand, NOS
inhibitors, by decreasing vasodilatation, reduce the ability oftumours to meet its metabolic
needs . Nitric oxide therefore appears to have a dual role in angiogenesis, tumour growth

314
and metastasis. Whether NO will promote or inhibit a particular process, depends on local
NO generation, temporally and spatially and interaction with target molecules.

REFERENCES

I. Albina, J.E ., Abate , JA , & Henry, w.L. (1991) . Nitric oxide production is required
for murine resident peritonel macrophages to suppress mitogen-stimulated T cell
proliferation. 1. Immunol. 147, 144-148.
2. Assender, 1. W., Southgate, K.M., Hallett, M.B. & Newby , AC. (1992) . Inhibition
of proliferation, but not of Ca 2+ mobilization by cyclic AMP and GMP in rabbit
aortic smooth muscle cells. Biochem . 1., 288, 527-532.
3. Bath, P.M.W., Hassei , D.G., & Gladwin, AM . (1991) . Nitric oxide and
prostacyclin: Divergence ofinhibitory effects on monocyte chemotaxis and adhesion
to endothelium in vitro. Arterioscler. Thromb ., 11,254-260.
4. Beckman, 1.S., Beckman, I.W., Chen. J., MarshalI, PA & Freeman, RA & (1990) .
Apparent hydroxyl radical production by peroxynitrite: implications for endothelial
injury from nitric oxide and superoxide . Proc. Natl. Acad. Sci. U.S.A., 87, 1620-
1624.
5. Brooks, P.C., Montgomery, AM ., Rosenfeld, M., Reisfeld, R.A., Hu, T., Klier, G .
& Cheresh, D.A (1994) . Integrin alpha v beta 3 antagonists, promote tumour
regression by inducing apoptosis of angiogenic blood vessels . Cell, 79, 1157-1164 .
6. Cornwell , I.L., Arnold, E., Boerth, N.L. & Lincoln, I.M., (1994) . Inhibition of
smooth muscle cell growth by nitric oxide and activation of cAMP-dependent protein
kinase by cGMP . Am. J. Physio\., 267, CI405-1413 .
7. Delikonstantinos, G., Villiotou, V. & Stavrides, 1. C. (1995) . Release by
ultraviolet B (u.v.B) radiation of nitric oxide (NO) from human keratinocytes: a
potential role for nitric oxide in erythema production . Br. J. Pharmaco\. , 114, 1257-
1265.
8. Edwards, P., Cendan , J.c., Topping, D.B., Moldawer, L.L., Mackay, S., Copeland,
E.M . & Lind, D.S. (1996). Tumour cell nitric oxide inhibits cell growth in virtro but
stimulates tumourigenesis and experimental lung metastasis in vivo . 1. Surg. Res ., 63
(1),49-52.
9. Efron, D.T., Kirk., SJ. Regan M.C., Wasserkrug, H.L. & Barbul, A (1991). Nitric
oxide generation from L-arginine is required for optimal human peripheral blood
Iymphocyte DNA synthesis. Surgery, 110, 327-334.
10. Filep, J.G ., Baron, C; Lachance, S., Perreault, C. & Chan, J.S.D.(1996). Involvement
ofnitric oxide in target-cell lysis and DNA fragmentation induced by murine natural
killer cells . Blood, 87, 5136-5143.
11. Folkrnan,1. & Shing, Y. (1992) . Angiogenesis. 1. Bio\. Chem . 267,10931-10934.
12. Folkman, 1. (1985) . Tumorangiogenesis. Adv. Cancer. Res., 43,172-203 .
13. Friedlander, M., Brooks, P.c., ShatTer, R.W., Kincaid, M., Vamer, J. A. & Cheresh ,
D.A (1995) . Definition of two angiogenic pathways by distinct a, integrins.
Science 270, 1500-1502.
14. Garg, u.c. & Hassid , A. (1989) . Nitric oxide-generating vasodilators and 8-bromo-
cyclic GMP inhibit mitogenesis and proliferation of cultured rat vascular smooth
muscle cells . 1. Clin Invest., 83,1774-1777.
15. Gasic, EJ., Gasic, I.B ., Galanti, N., Johnson, I. & Murphy, S. (1973). Platelet
tumour interaction in mice. The role of platelet in the spread of malignant disease.
Int. 1. Cancer, 11, 704-708 .

315
16. Gasic, GJ., Gasic, TB. & Stewart, C.C. (1968). Antimetastatic effects associated
with platelet reduction . Proc. Natl. Acad. Sei. USA 61,45-62.
17. Grant, D.S., Tashiro, K.I., Segui-Real, B., Yamada, Y., Martin, G.R & Kleinman,
HK (1989). Two different laminin domains mediate the differentiation of human
endothelial cells into capillary-like structures in vitro. Cell, 58, 933-943.
18. Gross, S.S. & Wolin, M.S. (1995). Nitric oxide: Pathophysiological mechanism.
Ann. Rev. Physiol 57, 737-769.
19. Harris-Hooker, S.A., Gajdusec, C.M., Wight, TN. & Schwartz, S.M. (1983).
Neovascular response induced by cultured aortic endothelial cells. 1. Cell Physiol.,
114,302-310.
20. Hibbs, 1. B., Taintor, RR. ,Vavrin, Z., Granger, D.L.,Drapier, J-C, Aber, 1.1. &
Lancaster, 1.R, jr. (1990). Synthesis of nitric oxide from a terminal guanidino
nitrogen atom of L-arginine: a molecular mechanism regulating cellular proliferation
that targets intracellular iron. In: Nitric oxide from L-arginine: A cellular Bioregulatory
system, ed. S. Moncada , 1.B. Hibbs, p.p. 189-223, Amsterdam : Elsevier Sei,
21. Hidetsugu, S., Kurose, I. Ebinuma, H., Fukumura, D. ,Higuchi, H., Atsukawa , K.,
Tada, S., Kimura, H., Yonei, Y., Masuda, T,Mura, S. & Ishii, H. (1996). Kupffer
cell-mediated cytotoxicity against hepatoma cells occur through production of nitric
oxide and adhesion via ICAM-l/CDI8. Intern. Immunol. 8 (7), 1165-1172.
22. Hirata, M & Murad, F. (1994). Interrelationships of cyclic GMP , inositol
phosphates and calcium. Adv. Pharmacol. 26, 195-216.
23. Hogg, N., Darley-Usmar, v., Wilson, M.T.& Moncada, S. (1992). Production
of'hydroxyl-radicals from the simultaneous generation of superoxide and nitric oxide.
Biochem . 1. 281,419-424.
24. Honn, K. V., Tang, D. G. & Chen, Y.Q. (1992). Platelets and cancer metastasis:
more than an epiphenomenon. Semin. Thromb. Hemostas., 18,392-415 .
25. Jaffe, E.A., Grulich, 1., Welksler, B.B., Hampel, G. & Watanabe, K., (1987) .
Correlation between thrombin-induced prostacyclin production and inositol
triphosphate and cytosolic free calcium levels in cultured human endothelial cells. J.
Biol. Chem., 262, 8557-8565.
26. Jenkins, D.C., Charles, I.G., Thompson, L.L., Moss, DW., Holnes, L.S. , Bayliss,
S.A, Rhodes, P., Westmore, K., Emson, P.C. & Moncada, S. (1995). Roles of nitric
oxide in tumour growth. Proc. Natl. Acad. Sei. U.S.A. 92, 4392-4396 .
27. Kariya, K., Kawahara, Y., Araki, S., Fukuzaki, H. & Takai, Y., (1989).
Antiprolifarative action of cGMP-elevating vasodilators in cultured rabbit aortic
smooth muscle cells. Atherosklerosis, 80, 143-147.
28. Karpatkin, S. ,Pearlstein, E., Arnbrogio, C. & Coller, B.S. (1988). Role of adhesive
proteins on platelet tumour interaction in vitro and metastasis formation in vitro . J.
CIin Invest., 81,1012-1019.
29. Katsuki, S., Amold , W., Mittal, C. & Mural, F. (1977). Stimulation of guanylate
cyclase by sodium nitroprusside, nitroglycerin and nitric oxide in various tissue
preparation and comparison to the effects of sodium azide and hydroxylamine . J.
Cyclic Nucleotide Res. 3, 23-35.
30. Koch, AE., Halloram, M.M., Haskell, CJ., Shah, M.R. & Polverini, PJ. (1995) .
Angiogenesis mediated by soluble forms of E-selectin and vascular cell adhesion
molecule-1. Nature, 376, 517-519.
31. Kubota, Y., Kleinman, HK, Martin, G.R. & Lawley, TJ. (1988). The role of
laminin and basement membrane in the morphological differentiation of human
endothelial cells into capillary-Iike structures. 1.Cell Biol., 107, 1589-1598.

316
32. Lee, YS. & Wurster, R.D. (1994). Potentiation on anti-proliferative effect of
nitroprusside by ascorbate in human brain tumour cells. Cancer Letters, 78 (1-3), 19-
23.
33. Leibovich, SJ., Polverini, PJ., Fong, T.W., Hardow, LA & . Koch, A.E., (1994) .
Production of angiogenic activity by human monocytes requires an L-arginine-nitric
oxide synthase dependent effector mechanism. Proc. Natl. Acad. Sei. USA., 91,
4190-4194.
34. Lejeune, P., Lagadec, P. ,Onier, N., Pinad, D., Ohshima, H. & Jeannin, J.F. (1994).
Nitric oxide involvement in tumour-induced immunosuppression. J. Immunol., 152,
5077-5083 .
35. Liotta, L.A., Steeg, P.S. & Stetler-Stevenson, W.G. (1991). Cancer metastasis and
angiogenesis: An imbalance ofpositive and megative regulation. Cell, 64, 327-336.
36. Maragoudakis, M.E., Sarmonika, M., & Panoutsakopoulou, M., (1988). Rate of
basement membrane biosynthesis as an index to angiogenesis. Tissue & Cell, 20,
531-539.
37. McDonald, CJ. (1991). Cardiovascular complications ofpsoriasis. In: Psoriasis, ed.
Roenigk, H.H. & Mailbach, HJ ,. p.p. 97-11, New York, Basel, Hong-Kong: Marcel
DekkerInc.
38. McDonald, LJ. & Murad, F. (1996). Nitric oxide and cyclic GMP signalling. Proc .
Soc. Exp. Biol. Med. 211 (1), 1-6.
39. Miki, W., Kawabe, Y, & Kurijama, K. (1977). Activation of cerebral guanylate
cyclase by nitric oxide. Biochem. Biophys. Res. Commun., 75, 851-856.
40. Moncada, S., Radomski, MW. & Palmer, R.M. (1988). Endothelium derived relaxing
factor : identification as nitric oxide and role in the control of vascular tone and
platelet function . Biochem. Pharrnacol., 37, 2495-2501.
41. MorbideIli, V., Chang, C-H, Douglas, J.G. Granger, HJ., Ledda, F. & Ziehe, M .
(1996) . Nitric oxide mediates mitogenic effect of VEGF on coronary renular
endothelium . Am. J. Physiol., 270 (Heart Circ. Physiol. 39): H411-H415 .
42. Murad, F. (1986). Cyclic guanosine monophosphate as a mediator of vasodilatation.
J. Clin. Invest. 78, 1-5.
43. Nathan, C. & Hibbs, J.BJr. (1991). Role of nitric oxide synthesis in macrophage
antimicrobial activity. Curr. Opin. Immunol., 3, 65-70.
44. Nierodzik, M.L., Kajumo, F. & . Karpatkin, S. (1992). Effect of thrombin tretament
oftumour cell on adhesion oftumour cells to platelets in vitro and tumour metastasis
in vivo. Cancer Res., 52, 3267-3272.
45. Nierodzik. M.L., Plotkin, A., Kajumo, F. & Karpatkin, S., (1991). Thrombin
stimulates tumour-platelet adhesion in vitro and metastasis in vivo . J. Clin. Invest.,
87, 229-236 .
46. Nishio, E., Fukushima, K. ,Shiozaki, M. & Watanabe, Y (1996). Nitric oxide donor
SNAP induces apoptosis in smooth muscle cells through cGMP-independent
mechanisms . Biochem . Bophys. Res. Commun., 221, 163-168.
47. Page, c.P. (1988). The involvement ofplatelet in non thrombotic processes. Trends
in Pharmacol. Sei., 9, 66-71.
48. Pearlstein, E., Ambrogio, C. & Karpatkin, S. (1984). Effect of antiplatelet antibody
on the development of pulmonary metastasis following injection of CT26 colon
adenosarcinoma, Lewis Lung carcinoma and B16 a melanotic melanoma tumour cells
into mice. Cancer Res. 44, 3384-3387

317
49 . Pipili-Synetos, E. Sakkoula, E., Haralabopoulos, G., Andriopoulou, P., Peristeris, P.,
& Maragoudakis, M.E., (1994). Evidence that nitric oxide is an endogenous
antiangiogenic mediator. Br. J. Pharmacol., 11,894-902.
50. Pipili-Synetos, E., Papageorgiou, A., Sakkoula, E., Sotiropoulou, G., Fotsis, T. ,
Karakioulakis, G., & Maragoudakis, M.E. (1995) . Inhibition ofangiogenesis, tumour
growth and metastasis by the NO-releasing vasodilators, isosorbide mononiitrate and
dinitrate. Br . J. Pharmacol., 116, 1829-1834 .
51. Radomski, MW., & Moncada, S. (1991). Biological role of nitric oxide in platelet
function In : Clinical Relevance of nitric oxide in the cardiovascular system e.d.
Moncada, S., Higgs, E.A. & Berrazueta, lR., p.p.45-56. Madrid: Edicomplet.
52. Rickles, FR & Edwards, R.L. (1983) . Activation of blood coagulation in cancer:
Trousseau's syndrome revised . Blood , 62, 14-31.
53. Robertson, F.M., Long, BW., Tober, K.L., Ross , M.S. & Oberyszyn. (1996). Gene
expression and cellular sources of inducible nitric oxide synthase during tumour
promotion. Carcinogenesis, 17 (9), 2053-2059
54. Rodeberg, A.D., CHaet, M .S., Bass, R.C ., Arkovitz, M.S. & Garcia, V.F. (1995).
Nitric oxide: An overview. Am . J. Surgery, 170,292-303
55. Rollo, E.E., Laskin , D.E., & Denhardt, D.T., (1996). Osteopontin inhibits nitric
oxide production and cytotoxicity by activated RAW 264 .7 macrophages. J.
Leucocyte Biology 60 (3), 397-404 .
56. Sanchez Bueno, A., Verkhusha., V. ,Tanaka, Y., Takikawa, D . & Yoshida, R .
(1996). Interferon-gamma-dependent expression of inducible nitric oxide synthase,
interleukin-12 and interferon-gamma-inducing factor in macrophages elicited by
allografted tumour cells . Biochem . Biophys. Res . Commun., 224 (2),555-563 .
57. Schrnidt, H.H.HW., Lohmann, S. M. & Walter, U. (1993) . The nitric oxide and
cGMP signal transduction system . Regulation and mechanism of action . Biochim .
Biophys. Acta, 1178, 153-175 .
58. Stuher, DJ. & Griffith, D .W. (1992). Mammalian nitric oxide synthases. Adv .
Enzymol, Relat. Areas . Mol. Biol. 65, 287-346
59. Stuher, DJ. & Nathan, C. (1989) . Nitric oxide. A macrophage product responsible
for cytostasis and respiratory inhibition in tumour cells. J. Exp . Med . 169, 1543-
1555.
60. Thomsen, L.L., Lawton, F.G., Knowles, R.G., Beesly, lE., Riveros-Moreno, V. &
Moncada, S. (1994). Nitric oxide synthase activity in human gynaecological cancer.
Cancer Res ., 54, 1352-1354 .
61. Thomsen, L.L., Miles, D.W., Happerfield, L. ,Bobrow, L.G., Knowles, R.G. &
Moncada, S. (1995) . Nitric oxide synthase activity in human breast cancer. Br . J.
Cancer, 72, 41-44 .
62. Tirnpl, R., Rhode, H., Gehron Robey, P., Rennard,S.I., Foidart, lM. & Martin, G .R.
(1979). Laminin a glycoprotein from basement rnembranes. J. Bio!. Chern., 254,
9933-9937.
63. Tsopanoglou, N .E., Pipili-Synetos, Eva & Maragoudakis, M.E. (1993a). Thrombin
prornotes angiogenesis by a mechanism independent of fibrin formation . Amer. J.
Pysiol. 264 (Cell Physio1.33) : C:1302-C 1307.
64 . Urnansky, V., Rocha , M. & Schirrmacher, V. (1996a) . Liver endothelial cells:
Participation in most response to lymphoma metastasis. Cancer & Metastas. Rev.
15, 273-279.
65. Urnansky, V., Schirrmacher, V., & Rocha, M. (1996b) . New insights into turnour-
host interactions in lymphoma rnetastasis . J. Mol. Med . 74, 353-363 .

318
66. Xie, K., Dong, Z . & Filder, I.J. (1996). Aetivation of nitrie oxide synthase gene for
inhibition of cancer metastasis. 1. Leucoc. Biol., 59, 797-803
67. Xie, K. P., Huang, S.Y., Dong, Z.Y., Juang, S.H., Gutman , M., Xie, o.w., Nathan, C.
& Fidler, I.J. (1995). Transfeetion with the indueible nitric-oxide synthase gene
suppresses tumourigenieity and abrogates metastasis by M-1735 murine melanoma
cells. J. Exp . Med. , 181, 1333-1343.
68 . Yang, W., Ando.J. Korenaga, R. Toyo-oka, T., Kamiya. A. (1994). Exogenous nitric
oxide inhibits proliferation of eultured vascular endothelial eells. Biochem. Biophys.
Res. Commun., 203, 1160-1167.
69 . Yim, CY, Bastian, N.R., Smith, J.c., Hibbs, 1.B. & Samlowski (1993) . Maerophage
nitric oxide synthesis delays progression of ultra-violet light-induced murine skin
cancers. Cancer Res ., 53 (22), 5507-5511.
70. Yu, W.G., Yamamoto, N ., Takenaka, H., Mu, 1., Tai, x.o. ,Zou , J.P., Ogawa M ,
Tsutsui , T., Nijesuriya R. ,Yoshida , R., Herrman, S., Fujuvara, H . & Hamaoka, T .
(1996). Moleeular mechanisms underlying tumour immunotherapy with IL-12 .
lntemationallmmunol., 8 (6)855-865 .
71. Yuang, F., Leuing, M. Berk, DA & Jain, R.K. (1993). Microvascular permeability
of albumin, vascular surface area and vaseular volume measured in human
adenocarcinoma, LS 174T using dorsal chamber in SCID mice. Microvasc. Res.
45 ,269-289.
72. Ziehe, M ., Morbidelli, L., Masini, E. Amerini, S., Granger HJ ., Maggi, CA,
Geppetti , P. & Ledola, F. (1994). Nitrie oxide mediates angiogenesis in vivo and
endothelial eell growth and migration in vitro promoted by substance, P. 1. Clin.
Invest, 94, 2036-2044.
73 . Ziehe, M ., Morbidelli, L. , Masini, E., Granger, HJ., Geppetti, P. & Ledda, F. (1993) .
Nitric oxide promotes DNA synthesis and cyclic GMP formation in endothelial eells
from posteapillary venules. Bioehem. Biophys. Res. Commun., 192, 1198-1203.

319
HYPOXIAlREOXYGENATION ENHANCES TUBE FORMATION OF CULTURED
HUMAN MICROVASCULAR ENDOTHELIAL CELLS: THE ROLE OF REACTIVE
OXYGEN SPECIES

Peter 1. LeIkes, Kenneth A. Hahns, Soverin Karmiol", and Donald H.

Schmidr'

Laboratory of Cell Biology, Department of Medicine, University of

Wisconsin Medical School, Milwaukee Clinical Campus .
$: Section of Cardiology, Department of Medicine, University of
Wisconsin Medical School,
Milwaukee Clinical Campus .
*: current address : BioWhittaker, Inc. 8830 Biggs Ford Road,
Walkersville, MD 21793.

INTRODUCTION

Angiogenesis, the generation of new blood vessels, is a ubiquitous process which is tightly
regulated in normal physiological situations. The cellular and molecular mechanisms
controlling the initiation and termination ofthe angiogenic process are only partially known
(Folkman and Klagsbrun, 1987; Folkman and Shing, 1992; Maragoudakis, 1994; Ferrara,
1996; Montesano et al., 1996; Pepper et al., 1996). The pathophysiology of many diseases
involves uncontrolled growth of new blood vessels, prompting the search for
therapeutically effective inhibitors of angiogenesis (Maragoudakis, Sarmonika, and
Panoutsacopoulou, 1988; Folkman and Ingber, 1992; Fotsis et al., 1993; D'Amato et al.,
1994; O'Reilly et al., 1994; Polverini, 1994; Chen et al., 1995; Gradishar, 1997; O'Reilly et
al., 1997). Converse1y, in other clinical settings, promotion of neovascularization is
desirable, e.g, after myocardial infarction and/or in peripheral blood vessel occlusion, thus
calling for appropriate stimulators of' rtherapeutic" angiogenesis (Röcke I et al., 1993; Isner
et al., 1995).

Numerous prior studies have firmly established that hypoxia enhances angiogenesis in vivo,
presumably by upregulating the expression/production of angiogenic growth factors , such
as vascular endothelial cell growth factor (VEGF) in non-endothe1ial cells, e.g. in smooth
muscle cells or fibroblasts, concomitant with the upregulation of cognate VEGF receptors,
such as KDR/flk, and flt on endothelial cells (Shweiki et al., 1992; Millauer et al., 1993;
Seetharam et al., 1995; Brogi et al., 1996; Takagi et al., 1996; Tian, McKnight, and RusselI,

Angiogen esis: Model s, Modulators, and Clinical Applic ations

Edited by Maragoudakis, Plenum Press, New York, 1998 321
1997). Recently, Shono and coworkers have shown that chronic exposure to hypoxia ( at
least 3 days) enhances the formatiori of capillaries in an in vitro model of angiogenesis
(collagen gel), presumably by up-regulating VEGF, and/or its cognate receptor(s),
activating PKC-dependent signaling, or inducing the secretion of angiogenic cytokines
(Shono et al., 1996).

In contrast to these lengthy studies in the collagen gel system, culturing endothelial cells on
a complex, laminin-rich reconstituted basement membrane (Matrigel) extracted from the
Englebreth-Holm Swarm (EHS) tumor, will rapidly yield a capillary-like network oftubular
structures, termed herein tube-formation or tubular morphogenesis (Kubota et al., 1988;
Grant et al., 1991). The Matrigel system represents a particular in vitro angiogenesis
system, which adequately models certain stages of the angiogenic cascade, such as EC
migration, elongation and assembly of tubes (Baatout, 1997). In spite of these limitations,
the Matrigel system has been widely used, to considerable success, to elucidate some ofthe
basic mechanisms of in vitro angiogenesis. Recently cuItured endothelial cells on Matrigel
under experimental conditions of hypoxia/reoxygenation similar to those encountered in
clinical situations of transient ischemia and observed, enhanced tubular morphogenesis
(Hahn et al., 1995a). We hypothesized that exposure ofHMVEC to hypoxia/reoxygenation
might lead to the intracellular generation of reactive oxygen species (ROS) and that
endogenous ROS might directly and rapidly affect the angiogenic properties of the cells
cultured on a permissive 3-D extracellular matrix, such as Matrigel (Hahn et al., 1995b). In
this communication we will detail some of the salient experimental findings and provide
evidence for the involvement ofROS in the initiation ofthe angiogenic cascade.

MATERIALS AND METHODS

1. Cell Culture: Human dermal microvascular endothelial cells (from Clonetics

Corporation, San Diego, CA) were grown in an optimized cell cuIture medium (MCDB
131 supplemented with 10 ng/ml EGF, 1 Ilglml hydrocortisone, 30 ug/ml bovine brain
extract, 10 ug/ml heparin, 50 ug/ml gentamicin, 50 ng/ml amphotericin B and 5 % fetal calf
serum, as previously described (Manolopoulos, Samet, and Leikes, 1995; Silverman et al.,
1996; Kanda et al., 1998).The endothelial nature of these cells has been previously
established by a number of criteria (immunostaining for vWF, PECAM-l and uptake of
acetylated LDL). Based on previous experience, most of the phenotypic traits of these
cells, including their capability to rapidly form tubes on Matrigel (in vitro angiogenesis),
were maintained through the 9th passage, corresponding to 25-30 population doublings
(Leikes et al., 1994; Leikes et al., 1996). Accordingly, only cells from the 4th through the 9th
passage were used for these studies.

2. In Vitro Assay for Tube Formation on Matrigel: Matrigel was prepared in house
from EHS tumor (generously provided by Dr. Hynda K. Kleinman, NIDR, NIH, Bethesda
MD) according to published procedures (Grant et al., 1991; Haralabopoulos et al., 1994).
To generate asolid, three-dimensional Matrigel base, 12 well cell cuIture plates (surface
area = 3.5 cm2/well, from Costar) were coated with 100 ul/cm? ice cold Matrigel and
incubated overnight at 37° C.

322
3. Generation of Hypoxia/Reoxygenation: Hypoxia was generated by flooding a
conventional CO 2 incubator with a mixture of 95%N2/5%C02 . Dissolved oxygen in the
media was measured on line with a calibrated oxygen electrode (from ORlON). HMVEC
were seeded at the densities indicated below into 12 weil plates coated with Matrigel. The
cells were exposed for the indicated time periods to hypoxia, and then transferred into
normoxia . The length ofthe ensuing tubes was determined 8-24 h after the initiation of the
experiments. In some of the biochemical experiments detailed below, confluent HMVEC
monolayers were exposed to the same regimen ofhypoxia/reoxygenation.

4. Quantitation of tube length: The total length of the tubes formed by HMVEC on
Matrigel was quantitated by computer aided image analysis using established
morphometric methods (Grant et al., 1991; Haralabopoulos et al., 1994).Unless indicated
otherwise, all experimental conditions (hypoxia and normoxia) were tested in triplicate in
12 weil tissue culture plates. At the end of a particular experiment, the cultures were fixed
briefly (I min) with 3.7% paraformaldehyde and examined using phase contrast
microscopy . Images of at least three randomly selected fields from each weil were captured
onto video tape via a contrast enhancing camera (DAGE MTI 66). The video images were
digitized and processed in a PC-based image analysis system (ImagePro Plus, Media
Cybernetics, Silver Spring, MD). The total length of the ensuing tubes per unit area was
quantitated and analyzed statistically using a dedicated software program (Grant et al.,
1991; Samet and Leikes, 1993).

5. PKC Activity was assayed (in the absence or presence of the ubiquitous PKC inhibitor
1-(5-isoquinolinylsulfonyl)-2-methyl-piperazine (H7, 10 11M), or of the PKC activator
phorbol myristoyl acetate (PMA, 100 ng/ml) using a non-radioactive fluorescent kit (from
Panvera, Madison, WI), as previously described (Papadimitriou et al., 1997). Due to the
limited sensitivity ofthis assay, 1 x 106 HMVEC were plated in 100 mm diameter culture
dishes . Upon confluence, the cultures were washed with phenol red-free Medium M 199
and either maintained for 120 min at normoxia (controls) or exposed for 60 min to hypoxia
and then placed into the normoxic incubator for another 60 min. After termination of the
experiments, the cells were hypotonically Iysed on ice. PKC activity was determined in the
celllysates, according to the manufacturer's instructions.

6. Assessment of ROS generation: The formation of reactive oxygen species (ROS) was
measured fluorimetrically using 2,7-dichlorodihydrofluorescein diacetate (DCF, 20 11M,
from Molecular Probes) as an indicator (Royall and Ischiropoulos , 1993; Zulueta et al.,
1997). ROS inhibitors/scavengers (all from Sigma) used were pyrollidine-dithiocarbamate
(PDTC, 0.01 - 100 11M), superoxide dismutase ( SOD, 1 - 500 U/ml) and catalase (1 -
1200 U/ml) . The cells, grown in 96 or 24 weil plates, were preloaded for 30 min with the
membrane-permeant DCF, washed to remove excess dye and then incubated in
hypoxia/normoxia for the times indicated with or without the various ROS scavengers.
After termination of the experiments, DCF fluorescence was measured in a fluorescence
micropiate reader (Cytofluor from Millipore) using the appropriate fluorescein
excitation/emission filter combinations.
7. Statistical analysis: All experiments were carried out in triplicate and repeated at least
three times, unless indicated otherwise . The results, presented are means ± standard error

323
 'i'
D.
D.

::a
;:I
Ö
10

6
0\ "OXYGENATION

I 000°
~~
.000
!;j
15 tSl OOOOd

\00000~0
rn 4
f-o
Z
~ 2

I
Z
0
U
Z 0
til HYPOXIA
t.'
~ -2
0
0 20 60 100 140 160 220
TIME [ Yl N]

Figure 1: Time Course for the Generation of Hypoxia and Re-oxygenation in a 12

Weil Plate.
The amount of dissolved oxygen in the medium was measured on-line with an oxygen
sensitive electrode immersed in one ofthe wells ofthe culture plate. This is a characteristic
curve of a single experiment, which was repeated three more times, yielding similar results.

of the means (SEM), were evaluated by ANOVA and Bonferroni's post-test analysis
using SigmaStat (Jandel), with p < 0.05 being considered as statistically significant.

RESULTS AND DISCUSSION

Using an oxygen electrode immersed into the culture medium in a weil of a 12 weil culture
plate, we first ascertained that the experimental conditions described in Materials and
Methods will indeed generate hypoxia. As seen in Figure 1, exposure to 95 %N 2 / 5% CO2
rapidly reduces the partial oxygen pressure in the medium (normoxia : = 8.5 ppm dissolved
oxygen; p02 = 150 mm Hg) reaching an equilibrium at approx. 0.5 ppm (p02 =7.5 mm Hg)
after 30-40 min. Subsequently, placing the plate into a conventional air/C0 2 incubator,
rapidly increased the level of p02 in the medium, followed by a slower phase of
equilibration, restoring "normal" p02 within approx 1 h.

All models of angiogenesis have their inherent shortcomings (Auerbach, Auerbach, and
Polakowski, 1991; Zimrin, Villeponteau, and Maciag, 1995). For example, the formation of
"capillary-like tubes" on of endothelial cells plated onto Matrigel, models only certain
steps in the angiogenic cascade, such as endothelial cell activation and migration, but not
proliferation (Baatout, 1997). Manipulations of Matrigel- induced tubular morphogenesis
have to account for the fact that this is a "spontaneous" event, which is presumably
initiated by signals derived from the complex mixture of extracellular matrix proteins and
growth factors contained in Matrigel. In the past, the efficiency of anti-angiogenic
compounds to inhibit tube formation on Matrigel has been easily assessed working under
optimal, standardized experimental conditions (Grant et al., 1991; Haralabopoulos et al.,
1994). By contrast, detecting the effects of pro-angiogenic compounds/conditions requires
more sophisticated approaches, for example using growth factor-depleted Matrigel or
working under "sub-optimal" conditions, in which tube formation still can occur, albeit to a
reduced extent and/or at a slower pace.

324
 ErrECT or CEU SEECING CENSIlY ON HIIVEC TUBE rORIlAllON ON IIATRIGEl
200 -r-- - - - - - - - - - - - - - -..

X
...
lI'I 150

......
'"
~

o
;;: 100 0 - 0 8 HOURS
...
Cl
z
. - . 2. HOURS
-'

50 -1---__+_-~-_+_--__+_---_+_--l.
o 50 100 150 200
NUMBER or CEUS SEEDEO. [ • 10- J ]

Figure 2: Cell Density-Dependence of Tubular Morphogenesis on Matrigel. Human

microvascular endothelial cells (HMVEC) were plated at various densities in 12 weil plates
coated with Matrigel. The progress of tubular network formation under conventional,
normoxic conditions was quantitated after 8 and 24 h, as described in Materials and
Methods. The data represent mean ± SEM from three independent experiments, each
performed in triplicate.

As shown in Figure 2, the progression of tubular morphogenesis, as assessed by the

overalliength oftubes per viewing field, is a function of the density, at which the cells are
plated . To evaluate this progression, we plated the cells at densities between 25,000 and
200,000 cells/well and evaluated tube formation after 8 and 24 h. When plated at 25,000
cells/well (:=7,500 cells cm"), tube formation was minimal in 8 hand did not progress
significantly through 24 h. Increasing the cell density lead to a significant initial
acceleration of tube formation. Thus, at plating density of 200,000 cells/well, the
formation of tubes was essentially complete in 8 h, rather than after 24 h. The total length
of the capillary like network, however, was virtually independent of the cell density , as
long as the cells are plated at densities are ~ 50,000 cells/well ("" 15,000 cells cm").

To assess the effects of the duration of exposure to hypoxia on the subsequent tube
formation, 50,000 cells/well were plated onto Matrigel-coated 12 weil plates, incubated for
various amounts oftime (20 -120 min) to hypoxia, and then transferred to normoxia. In a11
instances, the ensuing tube length was quantitated after a total experimental duration (Ttotal
= thypoxia + tnonnoxia) of 8 h. Exposure of the cells to hypoxia yielded a bell-shaped
dose/time -response curve: Enhancement of tube formation, already discernable after only
20 min of exposure to hypoxia, was maximal at 40 -60 min, while longer times of exposure
decreased the extent of this effect (Figure 3).

Based on these results, we chose to further study the effects of hypoxialreoxygenation on

this in vitro model of angiogenesis by plating 50,000 cells/well, exposing the cultures for
40-60 min to hypoxia, and assessing tube formation after 8 h. When working under these
standardized, "sub-optimal" conditions, it becomes evident that abrief exposure to
hypoxia significantly augments endothelial cell tube formation on Matrigel. The micrograph
(Figure 4) shows the capillary like tubular network formed by HMVEC seeded at equal
densities and exposed for 8 h to normoxia (panel A) or to 1 h hypoxia followed by 7 h
normoxia (panel B).

325
 Durallan af Hypoxla Affecls Tube Forrnctlon

1 = B hrs
~ 100
.....,
s:
a,
c 75
11
-'
CD
.0
....
::J
50
11
~
"0
"i 25
""
0
o 20 40 60 90 120
Time of Exposure 10 Hypoxla (min)

Figure 3: Length of Exposure to Hypoxia Determines the Enhancement of Tube

Formation by HypoxiaIReoxygenation. HMVEC were seeded at 50,000 cells/well in 12
weil plates, exposed to hypoxia for the times indicated and then transferred into normoxia.
Tubular morphogenesis (overall length of tubes) was evaluated 8 h after the beginning of
the experiment by computer-aided image analysis, as described in Materials and Methods .
The data (mean ± SEM from 3 independent experiments, each in duplicate) are normalized
to the maximal enhancement oftube formation at 40 min hypoxia.

Computer-aided quantitation of micrographs, such as shown in Figure 4, further sheds light

on the mechanisms by which hypoxia/reoxygenation might enhance angiogenic processes.
By using the " sub-optimal" conditions, detailed above, we've sufficiently slowed down the
workings of this process to clearly discem the hypoxia/reoxygenation-fostered initial
acceleration of capillary tube formation. Evaluation of the time course indicates that at the
"suboptimal " seeding density of 50,000 cells/well, it takes approximately 24 h for
establishing a "complete" capillary network under conventional, normoxic conditions
(Figure 5).

A similar time course of tube formation was found, when the cells were maintained in
hypoxia for the entire duration ofthe experiments, viz. for up to 24 h (data not shown) . By
contrast, exposure for 40 - 60 min to hypoxia, with subsequent transfer to normoxia was
sufficient to greatly accelerate this process, so that the capillary network was completed
within 8 h. Significantly, in terms of the overall tube lengths, the two networks, attained
under normoxia and hypoxia/normoxia , respectively, were essentially indistinguishable.

Our results suggest that the effects ofhypoxia/reoxygenation are rapid and, most probably,
affect some ofthe initial processes ofthe angiogenic cascade. In particular we note that our
findings in part support, but also contrast some of the data recently published by Shono et
al. (1996) . Using the collagen gel angiogenesis model for capillary invasion, these
investigators showed enhancement of capillary invasion in response to 3 days of continued
exposure to hypoxia. In our hands and in our system, hypoxia alone was insufficient to
enhance tubular morphogenes is on Matrigel.

Mechanistically, the rapid enhancement of capillary tube formation can be caused by

autocrine/ paracrine effects, e.g. through the induction of angiogenic growth factors, such as

326
 o

Figure 4: Hypoxia/Reoxygenation Enhances Tube Formation on MatrigeI. Phase

Contrast Micrographs of HMVEC , plated at 50,000 cells/well into Matrigel- coated 12
Weil Culture Plates taken after 8 h of culture. Panel A: nonnoxia; Panel B: Hypoxia;
Original magnification: 125 x

bFGF or VEGF, or augmentation of some of the known (or as yet unknown) angiogenic
signaling pathways, e.g. PKC. We tested for the first possibility by pre-incubating the
cultures with neutralizing polyclonal antibodies to VEGF (5 ug/rnl, generousl y provided b y
Dr. Napoleone Ferrara , Genentech , South San Francisco) , bFGF (10 ug/rnl, from Upstate
Biotechnology, Inc. Lake Placid, NY), and, as a control, an unrelated polyclonal antibody
against human IgG (20 ug/ml from Sigma). In preliminary experiments we established the
"neutralizing" capacity ofthe antibodie s to VEGF and bFGF, as inferred from their abilit y
(at the concentrations listed above) to abrogate the respective growth factor-indu ced
proliferation of serum-starved HMVEC. With these antibod ies present during the
subsequent treatments, capillary tubular morphogenesis was measured, as detailed above.
None ofthese antibodies atTected"basal" tube formation either under normoxia (not show n)
or its enhancement by hypoxia/reoxygenation (Figure 6). In preliminary experim ents we
established that 1) these antibod ies, even at 10 fold higher concentrations, did not abrogate

327
 EITECT OF HYPOXIA ON HIIVEC TUBE FORIIATION ON IIATRIGEL

--
!
200 l- - - -l- - __ l
In n.s ,

~ 150
0------'
~
v·
10
10 0 0 / _ 0 normoo lo

~
. - . hypoolo (~O mln)

5 10 4 ce lla /12 .ell plote

-+
0
50~--------<1-------+---
o 10 20 30
TlIIE AFTER SEEOINC. [hrs]

Figure 5: Time Course of Tube Formation on MatrigeI. Human microvascular

endothelial cells (HMVEC, 50,000/well) were plated onto Matrigel-coated wells in a 12
weil plate. After exposure to hypoxia for 40 min, the plates were transferred into normoxia.
Control plates were not exposed to hypoxia. Tube length was monitored 8, 16 and 24 h
after the initiation ofthe experiments. In preliminary studies we did not find any difference
in the rate of tube formation between cells exposed to hypoxia (without re-oxygenation)
for up to 24 hand control , normoxic cells . Data represent mean ±SEM from 3 experiments,
each performed in triplicate; ** p < 0.01; *: p < 0.05; n.s.: statistically not significant ( p >
0.05)

-- 100
~
.s::.
Co
c
80
•
....J

•
.tI
60
::I
I-
•
..
~
40
;;
IX 20

0
ctn Ig G bFCF VEcF

Figure 6: Tubular Morphogenesis on Matrigel is not Inhibited by Neutralizing

Antibodies to VEGF or bFGF. HMVEC were plated in Matrigel-coated tissue culture
plates in the presence of neutralizing polyclonal antibodies to VEGF (5 ug/ml) and
bFGF(lO ug/ml) or an irrelevant anti human IgG (20 ug/ml), The cells were either exposed
to hypoxia for 1 h with subsequent 7 h reoxygenation or left for 8 h under normoxic
conditions. Tube length was evaluated as indicated in Materials and Methods . The data ,
shown for hypoxia/reoxygenation, are normalized to the control sampies without
antibodies and are expressed as mean ± SEM (n=3)

328
 _ ORIIOXIC

!
150

~
0 100
z
~

~ 50
~
0
clrl H7 PWA

Figure 7: HypoxialReoxygenation-Induced Enhancement of Tube Formation is

Independent of PKC Activity. Tubular morphogenesis of HMVEC on Matrigel under
nonnoxic (8 h) and hypoxia (1 h) / reoxygenation (7 h) conditions was determined as
described in Materials and Methods . In some ofthe cultures, the activity of prote in kinase
C (PKC) was modulated by incubating the cells at the time of plating with, respectively,
the PKC inhibitor H7 (10 IlM) and with the PKC activator PMA (100 ng/ml ). As
expected, tube fonnation was inhibited by H7 and accelerated by PMA . However, none of
these treatments affected the additional acceleration of tubular morphogenesis by
hypoxia/reoxygenation, suggesting that this effect is independent ofPKC.

tubular morphogenesis and 2) HMVEC do not express VEGF mRNA (Hayman and Leikes ,
unpublished observations). Parenthetically, the lack of the inhibition of tube fonnation
under nonnoxic conditions, might suggest a limited role for exogenous, Matrigel-contained
bFGF or VEGF in the fundamental mechanisms, by which HMVEC form a capillary like
network on Matrigel.

The angiogenic process, both in vitro and in vivo, involves activation of PKC (Montesano
and Orci, 1985; Kinsella et al., 1992; Davis et al., 1993; Maragoudakis et al., 1993;
Tsopanoglou, Pipili-Synetos, and Maragoudakis , 1993; Tsopanoglou, Haralabopoulos, and
Maragoudakis, 1995). Previous studies using the Matrigel model had confinned that
activation ofPKC by phorbol myristoyl acetate (PMA) leads to enhanced tube fonnation
whereas PKC inhibition at the time of plating abrogates it. Furthermore, previous studies
with fibroblast suggested, that chronic exposure to hypoxia augments the activit y of PKC
by as yet unidentified mechanisms (Sahai et al., 1994). To evaluate the possible
involvement ofPKC in the our model ofhypoxia/reoxygenation induced tube formation, we
plated the cells in the absence and presence ofthe PKC inhibitor H7 (10 IlM) . In line with
previous studies , inhibition of PKC by H7 significantly impaired tubular morphogenesis
(Figure 7). However, this effect was partially reversed by exposing the culture s to our
standard regimen of hypoxia/reoxygenation . Conversely, PMA (100 ng/mlj-mediated
augmentation of tube fonnation (in nonnoxia) was further enhanced by hypoxia/
reoxygenation . Additional studies confinned that while H7 and PMA at the doses used in
this work, indeed, inhibit or enhance PKC activity, respectively, and that our regimen of
hypoxialreoxygenation did not affect PKC activity.

Taken together these results indicate that brief exposure of HMVEC on Matrigel to
hypoxia with subsequent re-oxygenation enhances tube formation by a mechanism which
apparently does not involve some of the "classical" agonists /promoters of angiogenesis,

329
 C 100 11M POTC

E200 20 11M ocr

•uc
u
~ 150 A- B o I:
0
::J
L;:

.
b 100
0

~
;; 50
"'i
Ir

0
clT1 .c ' hypo. .c ' hypo. .c ' hypo.
+ + +
20 ' Mrrnox 20' nonnox PDTe
(+ pOTe) .. 20· normo.

Figure 8: HypoxialReoxygenation Induces Formation of Reactive Oxygen Species

(ROS) . HMVEC were plated in 12 well plates and grown to confluence. The cells were
loaded with the ubiquitous fluorogenic ROS indicator dye, 2,7-dichlorodihydrofluorescein
diacetate and the generation of ROS was assessed in a fluorescence micropiate reader, as
detailed in Materials and Methods. A: "Basal" level of DCF fluorescence in control cells
maintained for 60 rnin in normoxia. B: DCF fluorescence in normoxia (60 min) in the
continued presence of 100 11M PDTC . C: DCF fluorescence in cells incubated for 40 min in
hypoxia and then for 20 min in normoxia. D : DCF fluorescence in cells treated as in c, but
with PDTC present during the entire experiment. E: DCF fluorescence in cells treated as is
c, but with DCF added 1 min before transferring the plates into normoxia. The results are
normalized to basal DCF fluorescence in control/normoxic cells in the absence of PDTC .
Data are given as mean ± SEM for four experiments carried out in quadruplicate.

such as angiogenic growth factors and activation of PKC . We also noted, that it was not
hypoxia, per se, but rather, brief exposure to hypoxia followed by reoxygenation, which
caused the observed effects . We therefore hypothesized, that the this regimen of
hypoxia/reoxygenation might induce the intracellular generation of ROS, and that ROS
might serve as nove1, angiogenic signaling mechanism.

To test the first part of our hypothesis, we used a recently described , fluorescent techn ique
to quantitate the generation of ROS in confluent monolayers of HMVEC grown in tissue
culture-treated 12 well plates (Royall and Ischiropoulos, 1993; Zulueta et al., 1997). For
these experiments, the cells were loaded with the ubiquitous ROS indicator 2,7-dichloro-
dihydrofluorescein diacetate (DCF, 20 J.1M). After extensive washing to remove excess dye,
the sampies, either untreated or in the presence of the ubiquitous ROS scavenger
pyrollidine-dithiocarbamate (PDTC, 100 J.1M), were either left for 60 min under control,
normoxic conditions, or after exposure to hypoxia for 40 min reoxygenated for 20 min.
DCF fluorescence was measured immediate1y after termination of the experiments in a
fluorescence micropiate reader. As seen in Figure 8, PDTC alone significantly decreased
basal ROS production in normoxic cells (bar B). In the absence of PDTC,
hypoxia/reoxygenation induced an approx . 2-fold increase in DCF fluorescence, suggesting
a significant increase in the generation of ROS (bar C). This increase was abrogated, when
the cells were subjected to the same regimen in the continued presence ofPDTC (bar D), or
when the cells were exposed to hypoxia in the absence of PDTC, which was then added
just prior to reoxygenation (bar E). As positive controls, we verified the dose dependent

330
 c::J .Iri
CXJ + 50 1'10I POTC
g 100
_ + 100 U/ml SOO
..c
0. KleI + JOO U/ ml Calala..
e
•
...J
75
•
.tl
.....::J 50
•
..
~
ö
C<
25

Figure 9: ROS Inhibitors/Scavengers Abrogate Tube Formation on Matrigel .

HMVEC were plated into Matrigel-coated wells of a 12 weil culture plate, as described in
Materials and Methods . The various ROS scavengers/inhibitors were added at the time of
plating. The cells were exposed for 60 min to hypoxia followed by 7 h normoxia The
results are normalized to the total tube length in control cells exposed to
hypoxia/reoxygenation in the absence of the ROS scavengers. The data are expressed as
mean ± SEM form duplicate experiments performed in quadruplicate.

increase in DCF fluorescence upon exposure of the cells (for 15 min) to 1 -100 mM H 2 0 2
(data not shown). These data c1early confirm that our regimen of hypoxia/reoxygenation
generates intracellular ROS.

To assess, whether ROS do indeed playa role in hypoxia/reoxygenation- enhanced tube

formation, we plated HMVEC onto Matrigel in the presence of 50 11M PDTC . As
predicted by our hypothesis, PDTC did indeed inhibit hypoxia/reoxygenation induced tube
formation (Figure 9). Surprisingly, however, PDTC similarly inhibited the "basic"
formation of capillary like tubes also under normoxic conditions . This finding strongly
implies ROS-dependent signaling as a novel angiogenic second messenger pathway.

Inhibition of "normoxic" in vitro angiogenesis as weil as of hypoxia/reoxygenation-

enhanced tube formation was not restricted to PDTC, but could also be obtained with other
ROS scavengers. While tube formation was nearly completely abrogated by PDTC in a
dose-dependent fashion with an IC50 of 0.3 11M (data not shown), two other, weil
characterized ROS scavengers, SOD and catalase, were less efTective (Figure 9). Since
PDTC is cell permeant, while SOD and catalase are not, our data might suggest that part of
the ROS generated intracellularly, might freely cross the plasma membrane, where they are
inactivated by the membrane-impermeant ROS scavengers.

CONCLUSIONS AND FUTURE DIRECTIONS

Our studies provide evidence that brief exposure of HMVEC on Matrigel to hypoxia
followed by reoxygenation, enhances/ accelerates the formation of a network of capillary-
like tubes. Importantly, this efTect does not involve enhanced activation of PKC or the
secretion/release of angiogenic growth factors. Thus, molecular/cellular basis for the
phenomena described in this communication are quite distinct from those triggered by
(prolonged) exposure to hypoxia alone, which calls for activation of hypoxia-sensitive cis-

331
acting elements, e.g. in the promoter of the VEGF gene . Preliminary studies, using RT -PCR
indicate that HMVEC do not express VEGF (LeIkes, et al., manuscript in preparation).
Recent studies, however, indicate an increasing number of hypoxia-inducible genes, which
in this context may be relevant for inducing the angiogenic process, such as growth factor
receptors, enzymes involved in regulating the intracellular redox potential, cytokines, etc.
In view of the rapid, transient nature of the enhancement of tube formation by
hypoxia./reoxygenation, our results strongly suggest a direct activation of ROS-
transcription factors, which, in turn, might lead to the enhanced production of known (or as
yet unknown) promoters of angiogenesis. Clinically, the observed enhancement of
angiogenesis by hypoxiaJreoxygenation may be relevant for recent attempts at "ischemic
preconditioning"(Maulik et al., 1996; Przyklenk and Kloner, 1996; Strasser, Htun, and
Schaper, 1996). In these studies, the severity of the ischemic injury to the myocardium is
mitigated by brief, repetitive episodes of occlusion(hypoxialischemia) followed by
reperfusion/reoxygenation. Our studies suggest that ROS generation during preconditioning
might facilitate therapeutic angiogenesis

Mechanistically, our studies c1early implicate ROS as a noveI, angiogenic signaling

mechanisms. In this context, initiation of angiogenesis may be taken as a sequence of
injury-like activation signals, causing the EC to abandon the quiescent phenotype in favor
of an activated one. In this sense, generation of non-cytotoxic levels of ROS might serve
as a central switchboard, which is turned on by numerous compounds, known to contribute
to endothelial cell activation . Thus, angiogenesis seems to share some of the common
pathways with other EC "injuries" that activate endothelial cells, notably the activation of
ROS-dependent transcription factors, such as NF6B . Our regimen of
hypoxiaJreoxygenation might also trigger a plethora of other mechanisms , such as
production of cytokines or fibrinolytic/matrix-degrading enzymes, which indirectly could
enhance tube formation . Currently, we are testing these possibilities.

ACKNOWLEDGMENT

This study was not supported by the American Heart Association, Wisconsin Affiliate,
but, in part, by agrant from the Milwaukee Heart Research Foundation. We thank Dr.
Hynda Kleinman, NIDR, NIH, for generously providing the EHS tumor and Dr.
Napoleone Ferrara, Genentech, for the anti VEGF antibody .

REFERENCES

Auerbach, R ., Auerbach, W., and Polakowski , I. (1991) Assays for angiogenesis: a review .
Pharmac. Ther. 51: 1-11.

Baatout, S. (1997) Endothelial differentiation using Matrigel. Anticancer Res. 17: 451-455.

Brogi, E., Schatteman, G., Wu, T., Kim, E.A., Varticovski, L., Keyt, B., and Isner, J .M .
(1996) Hypoxia-induced paracrine regulation of vascular endothelial growth factor receptor
expression. J. Clin. Invest. 97: 469-476 .

Chen, C., Parangi, S., Tolentino, M.J., and Folkman, J. (1995) A strategy to discover
circulating angiogenesis inhibitors generated by human tumors. Cancer Res. 55: 4230-4233.

332
D'Amato, RJ ., Loughnan , M .S., Flynn, E., and Folkman, J. (1994) Thalidomide is an
inhibitorofangiogenesis. Proe. Nat/. Aead Sei. USA 91: 4082-4085 .

Davis, C.M ., Danehower, S.C., Laurenza, A., and Molony, 1.L. (1993) Identification of a
role of the vitronectin receptor and pratein kinase C in the induction of endothelial cell
vascular formation . J. Cello Bioehem. 51: 206-218.

Ferrara, N. (1996) The biology of vascular endothelial growth factor. In: Moleeular,
eellular, and clinieal aspeets 01angiogenesis, pp. 73-84 . Edited by Maragoudakis, M . New
York and London, Plenum Press .

Folkman, 1. and Ingber, D. (1992) Inhibition of angiogenesis. Semin. Cancer Bio/. 3: 89-96.

Folkman, 1. and Klagsbrun, M. (1987) Angiogenic factors . Seience 235: 442-447.

Folkman,1. and Shing, Y. (1992) Angiogenesis . J. Biol. Chem. 267, No. 16: 10931-10934.

Fotsis, T., Pepper, M ., Adlercreutz, H., Fleischmann, G., Hase, T., Montesano, R ., and
Schweigerer, L. (1993) Genistein, a dietary-derived inhibitor of in vitro angiogenesis . Proe.
Natl. Acad Sei. USA 90: 2690-2694 .

Gradishar, WJ. (1997) An overview of c1inical trials involving inhibitors of angiogenesis

and their mechanism of action . Invest. New Drugs 15: 49-59 .

Grant, D.S ., LeIkes, P.I ., Fukuda, K., and Kleinman, H.K. (1991) Intracellular mechanisms
involved in basement membrane induced blood vessel differentiation in vitro, In Vitro Cello
Dev. Biol. 27A : 327-336.

Hahn , K.A., Schmidt, D .H., and Leikes, P.l. (1995a) Hypoxia induces in vitro angiogenesis
in cultured human microvascular endothelial cells. FASEB J. 9: A587

Hahn, K.A. , Schmidt, D .H., and LeIkes, P.l. (1995b) Hypoxia enhances in vitro
angiogenesis of human microvascular endothelial cells cultured on Matrigel. In: Moleeular,
Cellular and Clinieal Aspects 01Angiogenesis, pp. 260-261 . Edited by Maragoudakis, M .E.
New York and London, Plenum Press .

Haralabopoulos, G.c., Grant , D.S., Kleinman , H.K., Leikes, P.l. , Papaioannou, S.P., and
Maragoudakis, M .E. (1994) Inhibitors of basement membrane collagen synthesis prevent
endothelial cell alignment in Matrigel in vitro and angiogenesis in vivo . Lab. Invest. 71:
575-582.

Höcke!, M., SchIenger, K., Doctraw, S., Kissel, T., and Vaupel, P. (1993) Therapeutic
angiogenesis . Areh. Surg. 128: 423-429 .

Isner, J.M., Walsh, K., Symes, 1.F., Pieczek, A., Takeshita, S., Lowry, 1., Rossow, S.,
Rosenfield, K., Weir, L., Brogi, E., and Schainfeld, R. (1995) Arterial gene therapy for
therapeutic angiogenesis in patients with peripheral artery disease . Cireulation 91:
2687-2692.

333
Kanda, K., Hayman, G.T., Silvennan, M.D., and Leikes, P.l. (1998) Comparison of
ICAM-I and VCAM-I in various human endothelial cell types and smooth muscle cells.
Endothelium (in press)

Kinsella, lL., Grant, D.S., Weeks, B.S., and Kleinman, HK (1992) Protein kinase C
regulates endothelial cell tube formation on basement membrane matrix, Matrigel. Exp. Cell
Res. 199: 56-62.

Kubota, Y, Kleinman, HK, Martin, G.R., and Lawley, TJ. (1988) Role of laminin and
basement membrane in the morphological differentiation of human endothelial cells into
capillary-like structures . 1. Cell Biol. 107: 1589-1598.

Leikes, P.l., Manolopoulos, V.G., Silvennan, M., Zhang, S., Karmiol, S., and Unsworth,
B.R. (1996) On the possible role of endothelial cell heterogeneity in angiogenesis. In:
Molecular, cellular, and clinical aspects of angiogenesis, pp. 1-18. Edited by
Maragoudakis, M.E. New York, Plenum Press.

Leikes, P.I., Silvennan, M., Wankowski, D.M., Zhang, S., Varani, J., Dame, M., Shen, J.,
and Karmiol, S. (1994) A comparison of umbilical vein endothelial cells (HUVEC) and
dermal microvascular endothelial cells (HMVEC) with respect to adhesion molecule
expression, tissue factor activity, and neutrophil killing. Mol. Biol. Ce1l5: 372a

Manolopoulos, V.G., Samet, M.M., and Leikes, P.I. (1995) Regulation of the adenylyl
cyclase signalling system in various types of cultured endothelial cells. 1. Cello Biochem.
57: 590-598.

Maragoudakis, M.E. (1994) Angiogenesis. Ann. Card. Surg. 8: 13-19.

Maragoudakis, M.E., Sarmonika, M., and Panoutsacopoulou, M. (1988) Inhibition of

basement membrane biosynthesis prevents angiogenesis. 1. Pharmacol. Exp. Ther. 244:
729-733.

Maragoudakis, M.E., Tsopanoglou, N.E., and Haralabopoulos, G. (1993) Regulation of

angiogenesis via protein kinase C. In: Vascular Enthothelium, pp. 81-85. Edited by
Catravas, lD. New York, Plenum Press.

Maulik, N., Watanabe, M., Zu, YL., Huang, C.K., Cordis, GA, Schley, J.A., and Das,
D.K. (1996) Ischemic preconditioning triggers the activation of MAP kinases and
MAPKAP kinase 2 in rat hearts. FEBS u« 396: 233-237.

Millauer, B., Wizigmann-Voos, S., Schnüreh, H., Martinez, R., Moller, N.P.H., Risau, W.,
and Ullrich, A. (1993) High affinity VEGF binding and developmental expression suggest
Flk-l as a major regulator ofvasculogenesis and angiogenesis. Cell72: 835-846.

Montesano, R., Kumar, S., Orci, L., and Pepper, M.S. (1996) Synergistic effect of
hyaluronan oligosaccharides and vascular endothelial growth factor on angiogenesis in vitro .
Lab. Invest. 75: 249-262 .

334
Montesano, R. and Orei, L. (1985) Tumor-promoting phorbol esters induee angiogenesis in
vitro . Ce1l42: 469-477.

O'Reilly , M.S., Boehm, 1., Shing, Y, Fukai, N., Vasios, G., Lane, W.S., Flynn , E.,
Birkhead, J.R., Olsen, B.R., and Folkman, 1. (1997) Endostatin : An endogenous inhibitor of
angiogenesis and tumor growth . Ce1l88: 277-285 .

O'Reilly , M.S., Holmgren, L., Shing, Y , Chen, C; RosenthaI, RA , Moses , M., Lane,
W.S., Cao, Y, Sage, E.H., and Folkman, 1. (1994) Angiostatin : a novel angiogenesis
inhibitor that mediates the suppression ofmetastases by a Lewis lung eareinoma. Cell 79:
345-328 .

Papadimitriou, E., Manolopoulos, V.G., Maragoudakis, M.E., Unsworth , B.R., and LeIkes,
P.1. (1997) Thrombin modulates veetorial seeretion of extraeellular matrix proteins in
cultured endothelial eells. Am. 1. Physiol. 272: CII12-CI122.

Pepper, M.S., Montesano, R., Mandriota, SJ., Drei, L., and Vassalli, 1.D. (1996)
Angiogenesis: A paradigm for balanced extracellular proteolysis during cell migration and
morphogenesis. Enzy me and Protein 49: 138-162.

Polverini, P. (1994) Inhibitors of neovascularization : critical mediators in the coordinate

regulation of angiogenesis. In: Angiogenesis: molecular biology, clinical aspe cts, pp. 29-38 .
Edited by Maragoudakis, M.E., Gullino, P.M., and LeIkes, P.1. New York, Plenum Press .

Przyklenk, K. and Kloner, RA (1996) Role of protein kinase C in ischemic

preeonditioning: in search ofthe "pure and simple truth". Basic Res. Cardiol. 91: 41-43 .

RoyalI, 1.A. and Ischiropoulos, H. (1993) Evaluation of 2',7'-diehlorofluorescin and

dihydrorhodamine 123 as fluorescent probes for intracellular H202 in euItured endothelial
eells . Arch. Biochem. Biophys. 302: 348-355.

Sahai, A , Patel, M.S., Zavosh, AS., and Tannen, RL. (1994) Chronic hypoxia impairs the
differentiation of 3T3-L 1 fibroblast in culture : role of sustained protein kinase C activation .
1. CelloPhysiol. 160: 107-112.

Samet, M.M. and LeIkes, P.1. (1993) Flow patterns and endothelial cell morphology in a
simplified model of an artificial ventricle. Cell Biophys. 23: 139-163.

Seetharam, L., Gotoh, N., Maru, Y, Neufeld, G., Yamaguehi, S., and Shibuya, M. (1995) A
unique signal transduction from FLT tyrosine kinase, a receptor for vascular endothelial
growth factor VEGF . Oncogene 10: 135-147.

Shono, 1., Ono, M., Izumi, H., Jimi, S.I., Matsushima, K., Okamoto, 1., Kohno , K., and
Kuwano, M. (1996) Involvement of the transcription factor NF-kappaB in tubular
morphogenesis ofhuman microvascular endothelial cells by oxidative stress. Mol. Cell Biol.
16: 4231-4239.

335
Shweiki, D., Hin, A., Soffer, D., and Keshet, E. (1992) Vascular endothelial growth factor
induced by hypoxia may mediate hypoxia-initiated angiogenesis . Nature 359: 843-845 .

Silverman, M .D., Manolopoulos, V.G., Unsworth, B.R., and Lelkes, P.1. (1996) Tissue
factor expression is differentially modulated by cyclic mechanical strain in various human
endothelial cells . Blood Coagul. Fibrinolysis 7: 281-288.

Strasser, R., Htun, P., and Schaper, W. (1996) Salvage of jeopardized myocardium by
ischemic preconditioning: Is the quest over? Mol. Cell. Bioehem. 161: 209-215 .

Takagi, H., King, G.L., Ferrara, N., and Aiello, L.P . (1996) Hypoxia regulates vascular
endothelial growth factor receptor KDRlFlk gene expression through adenosine A 2
receptors in retinal capillary endothelial cells . Invest. Ophthalmol. Vis. Sei. 37: 1311-1321.

Tian, H., McKnight, S.L., and RusselI, D.W. (1997) Endothelial PAS domain protein 1
(EPAS1), a transcription factor selectively expressed in endothelial cells. Genes &
Development 11: 72-82 .

Tsopanoglou, N.E., Haralabopoulos, G.c., and Maragoudakis, M.B. (1995) Opposing

effects on modulation of angiogenesis by protein kinase e and cAMP-mediated pathways.
J. Vase. Res. 270: 8367-8372.

Tsopanoglou, N.E., Pipili-Synetos, E., and Maragoudakis, M.E. (1993) Protein kinase e
involvement in the regulation ofangiogenesis. J. Vase. Res. 30: 202-208.

Zimrin, A.B ., Villeponteau, B., and Maciag, T. (1995) Models of in vitro angiogenesis:
Endothelial cell differentiation on fibrin but not Matrigel is transcriptionally dependent.
Bioehem. Biophys. Res. Commun. 213: 630-638.

Zulueta, J.J., Sawhney, R., Yu, F.S., Cote, c.e., and Hassoun , P.M. (1997) Intracellular
generation of reactive oxygen species in endothelial cells exposed to anoxia-reoxygenation .
Am. J. Physiol. 272: L897-L902.

336
TUMOUR ANGIOGENESIS AND METASTASIS: THE REGULATORY ROLE OF
HYALURONAN AND ITS DEGRADAnON PRODUCTS.

D.C. West, D.M. Shaw and M. Joyce .

Department of Immunology, Faculty of Medicine, University of

Liverpool, Liverpool L69 3BX.

Hyaluronan (HA, previously termed hyaluronic acid) is a high molecular weight

glycosaminoglycan, a linear polysaccharide made up of a repeating disaccharide unit [D-
glucuronic acid (I-ß-3) N-acetyl-D-glucosamine] linked by l-ß-4 glycosidic bonds . It is
present in the extracellular matrix of most animal tissues and this ubiquitous distribution,
its capacity to bind large amounts of water, and its simple structure led to the general belief
that it was essentially a structural, or space filling, molecule. However, in the last ten yea rs
HA has been shown to profoundly influence cell behaviour. Transient increases in tissue
HA levels coincide with rapid cell proliferation and migration, during embryonie
development and the regeneration and remodelling of adult tissues . Localized accumulation
has been reported in association with tissue damage, organ rejection and many
inflammatory diseases, notably psoriasis and scleroderma, and it is a major component of
many tumours( Laurent and Fraser, 1992; Knudson et aI., 1989). The temporal and spatial
distribution ofHA during embryogenesis, and tissue remodelling, suggests that the sythesis
and degradation of HA plays an important regulatory role in these processes, and certain
pathological conditions such as tumour growth .
It is now widely accepted that the growth of most solid tumours is dependent on
their ability to induce and maintain an adequate blood supply, through the stimulation of
new capillary vessel growth (angiogenesis), (Folkman, 1990; Folkman, 1995). In addition ,
the level of tumour vascularization has been shown to correlate with the numbers of
circulating tumour cells and metastasis, in a transplantable mouse fibrosarcoma model, and
tumour metastasis and poor prognosis in many human tumours (Weidner et al., 1991; Hart
and Saini, 1992). Both in vivo and in vitro studies suggest that tissue hyaluronan
metabolism may be an important regulator of angiogenesis (West and Kumar, 1989).

HYALURONAN AND ANGIOGENESIS IN VIVO

The high concentration of HA found in avascular tissues, such as cartilage and

vitreous humour, and at relatively avascular sites, such as the desmoplastic region of

Angiogenesis: Mod els, Modulators, and Clinical Applications

Edited by Marag oudaki s, Plenum Pres s, New York, 1998 337
 Table 1. Anti-angiogenic properties oe high- molecular weight hyaluronan
In vivo:

• inhibits granulation tissue formation and vascularization [Balazs & Darzynskiewicz ' 73]
• inhibits vascularization ofimplanted fibrin gels [Dvorak et al '87]
• causes regression ofthe capillary plexus in chick limb buds [Feinberg & Beebe]
• Endogenous HA inhibits angiogenic activity ofhuman wound fluids in CAM assay .

In vitro:

• Inhibits endothelial cell proliferation [West et aI ' 89] / migration [Sattar et aI '92;
Foumier & Doillon ' 92; Watanabe et aI '93] / cell-cell adhesion [West et aI '89]- at
physiological concentrations i.e. lOO-200mg/ml.

Table 2.Angiogenic nature oe OHA (4-20 disaccharides)

In vitro

stimulates endothelial proliferation [West '89 ; Arnold '95] , migration [Sattar et al '94] ,
migration into collagen! fibrin gels [Trouchon et aI '96 ; Montesano et aI '96] , and tube
formation [Hirata et aI '93]
Induces synthesis of proliferation- associated proteins, type VIII collagen, uPA and PAI-I
[West '89; Rooney '93 and '95; Montesano et al ' 96]

In vivo models

• CAM [West et al '85], rabbit corneal [Hirata et al '93]

• rat skin : topical application [Sattar et aI '94] ; graft [ Lees et aI '95], and polyvinyl
sponge implant [West et al, submitted; Smither et a1, submitted], rabbit: subcutaneous
[West & Kumar, unpublished]
• pig: full thickness wound [Amold et al '95]

338
invasive tumours, suggests that extracellular matrix HA can inhibit angiogenesis (West and
Kumar, 1989; Barsky et al., 1987). In vivo studies indicate that implantation of high
concentration of macromolecular hyaluronan inhibits blood vessel formation in granulation
tissue (Balazs and Darzynkiewcz, 1973; Dvorak et al., 1987), in a concentration and size-
dependent manner, and induces regression of the capillary plexus in the developing chick
limb bud (Feinberg and Beebe, 1983).Table 1. In contrast, we have found that low-
molecular-weight HA-oligosaccharides (OHA; 4-20 disaccharides in length) stimulate
angiogenesis on the chick chorioallantoic membrane, in rat band pig wound and graft models
and after subcutaneous implantation or topical application. Table 2. Hirata et al (1993)
have independently confirmed the angiogenic activity of this range of OHA, using the
rabbit corneal assay .
Recent studies on the relationship between angiogenesis and tissue HA metabolism
in a rat sponge-implant wound healing model (West et al., 1997), and a freeze-injured rat
skin-graft model (Lees et al., 1995), have shown that tissue angiogenesis coincided with the
degradation of matrix hyaluronan, evidenced by a rapid fall in tissue HA size and content,
and increasing hyaluronidase levels. The addition of exogenous OHA accelerated both the
onset and rate ofHA degradation, in parallel with angiogenesis, supporting the hypothesis
that macromolecular HA is an important inhibitory regulator of angiogenesis. In foetal
wounds, the addition of exogenous streptomyces hyaluronidase decreases wound HA
content and increased both fibroplasia and capillary formation (Mast et al., 1992). A
similar change, from foetal, regenerative healing to an adult fibrotic type ofhealing,occurs in
late gestation and we have recently shown that this is associated with an increase in
hyaluronidase activity and a decrease in both hyaluronan concentration and size (West et
al., 1997). Interestingly, HA has been reported to inhibit transforming growth factor b,
(TGFb 1; Locci et al., 1995), a promoter of tissue angiogenesis, and to stimulate the
secretion of tissue inhibitors of metalloproteinases (TIMP; Yasui et al., 1992), inhibitors of
angiogenesis, at concentrations similar to those in the early wound . Both studies found
these effects to be size dependent, decreasing significantly with the size of the HA . Such
reports imply that HA degradation must precede the onset ofvascularization.

IN VITRO ANGIOGENESIS AND HYALURONAN

In vitro studies on cultured endothelial cells have shown that macomolecular HA

inhibits endothelial proliferation and migration (West and Kumar, 1988, 1989 a, b, 1991)
and disrupts newly formed monolayers, at physiologically relevant concentrations
(~100 mg/mi) i.e. concentrations present in avascular tissues and during tissue remodelling.
The inhibitory effect ofHA decreases with reducing size, and the HA- mediated inhibition
can not be reversed by addition of exogenous growth factors , such as basic Fibroblast
Growth Factor (bFGF; West and Kumar, 1989a, 1991). Watanabe et al (1993 ) have
recently confirmed these findings in a 3-dimensional collagen gel endothelial cell culture
system . Furthermore, high concentrations (lmg/ml) ofHA have been reported to inhibit the
migration Of human adipose capillary endothelial cells into fibrin gels, in vitro , and blood
vessel formation in fibrin gels implanted subcutaneously in guinea pigs (Dvorak et al.,
1987; Fournier and Doillon, 1992).
In our studies, OHA 2-8 kDa (4-20 disaccharides) specificall y stimulate both
proliferation, and to a lesser extent migration, of both bovine and human endothelial cells
(West and Kumar, 1989a, b; Sattar et al., 1992). Hirata et al (1993) have reported that
aniogenic OHA markedl y stimulated endothelial tube formation , on "rnatrigel".

339
Furthermore, two recent studies, and our own preliminary studies, have shown that OHA
(4-20 disaccharides in length) stimulate angiogenesis in 3-dimensional collagen and fibrin
matrices (Montesano et al., 1996; Trouchon et al., 1996). Trouchon et al (1996) also
confirmed that the 3-disaccharide oligosaccharide is not angiogenic.

HYALURONANMETABOLISM AND TUMOUR METASTASIS

Gven the antiangiogenic nature of macromolecular HA, the increased levels of

hyaluronan often found in human and animal tumou rs (Knudson et al., 1989), are difficult
to equate with the increased angiogenesis reported for many tumours, especially metastatic
tumours. Our earlier studies on a bone metastasising form of Wilm's tumour, Bone
Metastasising Renal Tumour of Childhood (BMRTC), showed that it secretes high levels
of low molecular weight HA into the patients circulation (Kumar et al., 1989) . Whilst
Wilm 's tumours are rarely metastatic and produce high levels ofhigh molecular weight HA .
These data suggested that analysis of tumour HA, especially its degradation, may shed
some light on this conundrum. To this end, we have analysed the HA contentJsize and
hyaluronidase activity of syngeneic murine and rat tumours, human xenograft tumours. HA
levels were increased in most of the 22 different tumours studied (38-4938 ~g tissue),
compared with normal adult tissues and, although their HA showed a wide variation in size
(ranging from 200-2,000 kDa) , there was a significant shift to a lower molecular mass,
compared with normal tissues. The size of HA in both the syngeneic and xenografts
showed a loose relationship to the relative levels of tumour HA and hyaluronidase activity
(Shaw et al., 1994; Shaw, 1996) i.e. those tumours with high hyaluronidase activity
generally had smaller HA. This suggests that increased tumour hyaluronidase levels may
reduce tumour HA size, and potentially content, and should be indicative of increased
angiogenesis and a poor tumour prognosis. Some support is given to this hypothesis by a
recent report by Lokeshwar et al (1996) showing an association between elevated
hyaluronidase levels and prostate cancer progression. Closer examination ofthe data for the
rat and murine tumours highlighted a group of three tumours with low levels «100Jlg/g
tissue) ofhigh molecular weight HA (2-3 x 106 Daltons). These tumours, the B 16-FI0 and
B 16-BL6 murine melanomas and a transplantable spontaneous rat uterine tumour, were all
highly metastatic in vivo, suggesting that both HA degradation and low tumour HA content
are indicative oftumour metastasis and probably increased tumour angiogenesis.
Tumours are complex mixtures of both tumour and strom al cells, and tumour-
stromal cell interactions are reported to increase tumour HA, especially in invasive tumours
(Knudson et al., 1989) . Thus we were interested to compare HA synthesis and
hyaluronidase activity of cultured tumour cell-lines and have analysed over 30 murine and
human cell-lines. Most cell-lines produced medium to high levels of HA (1-20 Jlg/106cells),
compared with normal fibroblast cultures, of mainly high molecular weight HA (> 800
kDa) . Only 5-10% ofthe HA produced remained cel1-associated and this fraction tended to
be slightly smaller than that in the medium. Hyaluronidase activity was difficult to detect,
or absent, in most cultures and was generally only present in the culture medium (Shaw,
1996) . However, there were several notable exceptions.
Metastatic mouse melanoma BI6-FI0 and BI6-BL6 cells produced very little HA
(58-87 ng/Iü'cells). The BI6-FlO cells produced mainly large HA (900 kDa) and low levels
ofhyaluronidase in the medium . In contrast, the more metastatic BI6-BL6 cells produced
HA of 100 kDa, or less, and had ten-fold higher levels of hyaluronidase, some of which was
cell-associated. Table 3.

340
Table 3. Hyaluronan and hyaluronidase measurements in extracts of murine melanoma cell
lines, B 16 Fl , FlO, BL6, ofincreasing metastatic potential.

Tumour Total HA HASize HA Size Total HAse

Cell Line ng/10 6 Medium Cells NFU/l0 6
Cells (kDa) (kDa) Cells
B16 F1 49 .8 2000 2000 1.10 x10-4
B16FlO 58 900 30 8.22 x10-4
B16BL6 87 100 100-30 6.58 x10 -3

Table 4. HA, hyaluronidase and surface receptor expression by colorectal

carcinoma celliines

Tumour Total HA HA Size HA Size Total HAse CAM

Cell Line ng/106 Medium Cells NFU/l0 6 assay
Cells (kDa) (kDa) Cells
HCTl16 588.6 300 100 4.44 +++
A(+) xor 2
HCTl16 1986.5 1800 900 - +/-
B(+)
2010 3675.3 2500 1000 - +/-

Six of the human tumour cell-lines were found to produce low amounts of HA
«500 ng/ 106cells) . Three were aggressive cervical carcinoma cell-lines and three (A2058
melanoma, Du145 prostatic carcinoma and HCTlI6(A+) colon carcinoma cells) are highly
metastatic in animal models. Similarly, Timar et al (1987) have reported that highly
metastatic variants ofLewis lung tumour cell-lines produce significantly lower amounts of
HA. Recently, we have carried out a more detailed study of the relationsh ip between HA
synthesis and angiogenic activity, using aseries of related human colon carcinoma cell-lines.
Table 4. Three colon carcinoma cell-lines, HCTl16 A(+) and HCTl16 B(+), originally
isolated from the same colorectal carcinoma by Dr Brattain (Houston) , and 2010-1, a
subclone of the more invasive HCTl16 B(+) line, selected by Dr Kinsella (Liverpool) for
its increased invasiveness in vitro (Kinsella et al., 1994). However, resuIts in an in vivo
metastasis model contradicted those observed in vitro [Dr A Kinsella, personal
communication]. Cells injected into the ceacum wall of nude mice estab1ished a primary
tumour and eventually secondary liver metastases. The greatest number of liver metastases
were found in animals bearing HCT116 A(+) tumours . Analysis of these cell-lines for HA
production and hyaluronidase activity gave some interesting resuIts. HCT116 A(+) cells
produced considerably less HA than either of the other two cell-lines, 0.5 Ilg/ 106cells vs .
2-41lg/ 106cells . Furthermore, the HA in HCT116 A(+) cuItures was significantly smaller
than that synthesised by the other cell-lines, 100 kDa and < 300 kDa, cell-surface and
medium respectively, compared with 1,800-3,000 kDa and 900- 1500 kDa, respectively .
Interestingly , HCT116 A(+) cells also had high levels ofhyaluronidase activity, which was
mainly cell-associated, whereas none could be detected in the other cell-lines, suggesting a
causative relationship between high enzyme levels and the reduced HA size (Shaw, 1996).

341
Figure 5. A schematic representation of the hypothesized role of HA and hyaluron idase
in regulation of angiogenesis.

We have hypothesised that areduction in HA size and concentration would allow

endogenous angiogenic cytokines to stimulate angiogenesis. Furthermore, the range of HA
secreted by these cells, < 30- 300 kDa suggests that angiogenic OHA are probably present.
Assay ofthe angiogenic activity of conditioned serum-free media from the three cell-lines in
the Chick Chorioallantoic Membrane (CAM) assay showed that only HCTl16 A(+)-
conditioned medium was angiogenic. When this was mixed with HCTl16 B(+)- conditioned
medium the angiogenic activity was marginal, suggesting that HCTl16 B(+)- conditioned
medium contained an anti-angiogenic activity. Centrifugation throgh a 100 kDa cut-off filter
removed both the macromolecular HA and the inhibitory activity from the HCT 116 B(+)-
conditioned medium. Although our data supports the hypothesis that hyaluronidase
mediated breakdown ofHA is necessary for increased tumour angiogenesis and metastasis
(Fig.l), hyaluronidase is though to act only in the Iysosomal environment. Liu et al (1996)
have found a potential "neutral " surface-associated hyaluronidase , similar to the glycosyl-
phosphatidylinositol (GPI)-anchored sperm PH-20 hyaluronidase, expressed by angiogenic
tumour cell-lines. Recent examination has shown that HCTl16 A(+)-cells also express a
GPI- anchored " neutral" surface-associated hyaluronidase which probably mediates their
synthesis of low molecular weight HA which may itself stimulate tumour vascularization
in vivo . Montesano et al (1996) have recently reported that angiogenic OHA acts
synergistically with Vascular endothelial growth factor (VEGF), arguably the main
angiogenic cytokine in tumours, but not basic Fibroblast growth factor (bFGF) . However,
using rat aorta cultured in a serum-free collagen gel matrix (Nicosia and Ottinetti, 1990),
preliminary data suggests that whilst both VEGF and OHA stimulate microvessel growth
in this model, their combined effect is at best additive and not synergistic .(Burbridge and
West, unpublished results) .

342
MECHANISM OF ACTION OF ANGIOGENIC HA-OLIGOSACCHARIDES

Preliminary studies have shown that OHA also induce proliferation related
proteins, stimulate the phosphorylation of a 32 kDa protein and increase Type I and VIII
collagen synthesis by 4-6 fold (West and Kumar, 1989a; Rooney et aI, 1993). These data
indicate that both OHA and macromolecular HA interact directly with endothelial cells,
inducing or inhibiting angiogenesis respectively.
Several workers, inc1uding ourselves, have reported that endothelial cells possess
high affinity HA-binding proteins, or receptors, on their cell surface, that both bind and
internalise HA (West and Kumar, 1989b; West, 1993; Smedrod et al., 1984; Madsen et al.,
1989). Initially, using macromolecular HA (106 Da), we detected 2,000 receptors/cell, with
a Kd of 10-12 M. More recently we have repeated this study, using a defined 42 kDa HA
fraction, and have found approximately 105 receptors/ cell and a Kd of 10-12 M, the
disparity being due to steric exc1usion of large HA molecules from the cell surface (West,
1993; Smedrod et al., 1984). Using HA-affinity chromatography, 1125-Iabeling of surface
proteins and SDS-PAGE analysis ofhuman and bovine endothelial cells, we have identified
five major cell-surface HA-binding proteins , between 90 and 125 kDa, with two minor
bands at 78 and 46 kDa (West and Kumar, 1991 ; West, 1993). Western blotting with anti-
CD44 and anti-ICAM-l antibodies, together with PCR determination of CD44 splice-
variants, indicates that the 78-125 kDa proteins are forms of non-variant CD44, differing in
their glycosylation, and ICAM-l isoforms [West, unpublished results]. Other cell-surface
HA-receptors, such as RHAMM and hyaluronectin, have been characterised, but
antibodies to these do not interact with human or bovine endothelial cells.
Although anti-CD44 antibodies have been reported to inhibit the migration and tube
formation of bovine and porcine endothelial cells in vitro (Banerjee and Toole, 1992;
Trouchon et al., 1996), suggesting that CD44 plays an important role in angiogenesis, it is
unlikely to be the primary "angiogenic" receptor for the active oligosaccharides. Whilst
will bind HA oligosaccharides of this size range, it also binds the 3-disaccharide HA-
oligosaccharide which is not angiogenic. However, Noble et a1 (1996) have reported that
HA-oligosaccharides, between the hexasaccharide and 440 kDa, rapidly induce and activate
NFKB and reduce IKBa expression in cultured macrophages. In an earlier study , they
found that the binding of HA (80 kDa in size) to CD44 induced TNFa and IL-Iß
expression within Ihr, and IGF-l after 12hrs (Noble et al., 1993).
In arecent study we compared the effect of angiogenic cytokines (VEGF, bFGF),
OHA, TNF-a and phorbolmyristyl acetate (PMA) on endothelial cell expression of CD44,
ICAM-l and E-Se1ectin, by flow cytometry . As expected, TNF-a and PMA greatly
upregulated ICAM-l and E-selectin expression (400%;taken as maximal expression) and to
a lesser extent CD44. (90%). VEGF increased both ICAM-l and E-Selectin expression by
150%, whilst bFGF upregulated ICAM- I expression 200% but had only a marginal effect
on E-selectin. Similar findings were recently reported by Melder et al (1996). Both VEGF
and bFGF increased CD44 by 70-75%. OHA had a similar effect on CD44 expression, but
greatly increased E-Selectin (350%) and down- regulated ICAM-l . The latter is significant
in that it is probably due to phagocytosis of the ICAM-l, possibly with bound HA-
oligosaccharide, and suggests that ICAM-l may mediate the angiogenic activity of these
oligosaccharides . However, inhibition of endothelial cell ICAM-l expression using the ISIS
1939 antisense oligonuc1eotide (Bennett et al., 1994) failed to inhibit the upregulation of E-
selectin by OHA, indicating that CD44 mayaiso play a role (West, prelimin ary data).
Preliminary results also indicate that OHA also upregulate Flk-I and Fit VEGF- receptor

343
expression and rapidly increase IL-8 mRNA levels (West and Noble, prelim inary data).
The latter result is similar to that previously reported (Noble et al, 1996) and appears to
indicate a rapid CD44 activation of NFKB transcription factor. Thus it is likely that both
CD44 and ICAM-l are involved in the activation of endothelial cells by OHA.

CONCLUSIONS

Studies on transplantable tumours and cultured tumour cell-lines suggests that tumour
metastasisl angiogenesis is associated with increased HA degradation . This appears to be
mediated by the increased expression of a cell-surface "neutral" hyaluronidase, possibly
re1ated to the sperm PH-20 enzyme . Present evidence suggests that angiogenic HA-
oligosaccharides bind to both CD44 and ICAM-l , but that ICAM-l-binding is probably
the most important for the stimulation of angiogenesis.

ACKNOWLEDGEMENTS

The authors thank the North West Cancer Research Fund (DCW, MJ) and the
MRC (DCW, DMS) for their support .

REFERENCES

Arnold , F., Jia, C.Y., He, C.F., Cherry , GW., Carbow, B., Meyer-Ingold, w.,
Bader, D .
and West, D.C., 1995, Hyaluronan, heterogeneity and healing: The effects of ultra-pure
hyaluronan of defined molecular size on the repair of full thickness pig skin wounds.
Wound Rep Reg. 3: \0-2 1.

Balazs, E.A., and Darzynkiewcz, Z., 1973, The effect of hyaluronic acid on fibroblasts,
mononuclear phagocytes and lymphocytes, in: Biology 0/ the fibroblast (E. Kulonen, and J .
Pikkarainen, eds), pp. 237-52, Acad Press, New York

Banerjee, S.D. and Toole , B.P., 1992, Hyaluronan-binding protein In endothe1ial

morphogenesis. J Cell Bio!' 119: 643-52 .

Bennett, C.F., Condon, T.P., Grimm, S., Chan, H., and Chiang M-Y ., 1994, Inhibition of
endothelial cell adhesion moleeule expression with antisense oligonucleotides. 1. Immunol.
152: 3530-40.

Dvorak, H.F., Harvey , S., Estralla, P., Brown, L.F., McDonagh, 1., and Dvorak, A.M.,
1987, Fibrin containing gels induce angiogenesis, Lab Invest. 57: 673-86.

Feinberg, R .N. and Beebe, D.C., 1983, Hyaluronate in vasculogenesis, Science. 220:
1177-1179.

Folkman,1., 1990, What is the evidence that tumors are angiogenesis dependent?, 1. Nat/.
Cancer. Inst. 82: 4-6 .

344
Folkman, J., 1995, Angiogenesis in cancer, vascular, rheumatoid and other disease , Nature
Medicine . 1: 27-31.

Foumier, N. and Doillon, C.J., 1992, In vitro angiogenesis in fibrin matrices containing
fibronectin or hyaluronic acid. Cel/ Biol Int Reports. 16: 1251-63.

Hart, I.R . and Saini, A., 1992, Biology oftumour metastasis , Lancet. 339: 1453-7.

Hirata, S., Akamarsu, T., Matsubara, T., Mizuno, K. and Ishikawa , H., 1993, Arthritis &
Rheum .36:S247.

Kinsella, A.R ., Lepts, G.c., Hili, CL. and Jones, M., 1994, Reduced E-cadherin expression
correlates with increased invasiveness in colorectal carcinoma cell lines. Clin Exp
Metastasis. 12: 335-342

Knudson, W., Biswas, c., Li, X-Q., Nemec, R.E. and Toole, B.P ., 1989, The role and
regulation of tumour-associated hyaluronan . in: The Biology of Hyaluronan, Ciba
Foundation Symposium 143 (0. Evered, and J. Whelan, eds), pp. 150-169, John Wiley and
Sons, Chiehester

Kumar, S., West, D.C., Ponting, J. and Gattamaneni, H.R., 1989, Sera of children with renal
tumour contain low molecular mass hyaluronic acid. Int. 1. Cancer. 44: 445-8.

Lees, V.c., Fan , T-P .D. and West , D.C., 1995, Angiogenesis in a delayed revascularization
model is accelerated by angiogenic oligosaccharides of hyaluronan, Lab Invest. 73:259-266 .

Liu, D., Pearlman, E., Diaconu, E., Guo, K., Mori, H., Haqqi, T., Markowitz, S., Willson ,
J., and Sy, M-S., 1996, Expression of hyaluronidase by tumor cells induces angiogenesis in
vivo . Proc. Natl. Acad Sei. U S.A. 93: 7832-7837 .

Locci, P., Marinucci, L., Lilli, c., Martinese, D. and Becchetti, E., 1995, Transforming
growth factor beta 1- hyaluronic acid interaction . Cell Tissue Res . 281: 317-24 .

Lokeshwar, V.B., Lokeshwar, B.L., Pham, H.T. and Block, N.L., 1996, Association of
elevated levels of hyaluronidase, a matrix- degrading enzyme , with prostate cancer
progression. Cancer Res . 56: 651-7 .

Madsen, K., Schenholm, M., Jahnke, G. and Tengblad, A., 1989, Hyaluronate binding to
intact comeas and cultured endothelial cells. Invest Opth Vis Sei . 30: 2132-7.

Mast , B.A., Haynes, J.H ., KrummeI, T.M., Diegelman, R.F. and Cohen , I.K ., 1992, In vivo
degradation of fetal wound hyaluronic acid results in increased fibroplasia, collagen
deposition, and neovascularization. Plastic. Reconstructive Surgery. 89: 503-9 .

Melder, RJ., Koenig, G.c., Witwer, B.P., Safabakhsh, N., Munn , L.L. and Jain, R.K.,
1996, During angiogenesis, vascular endothelial growth factor and basic fibroblast growth
factor regulate natural killer cell adhesion to tumour endothelium . Nature Medicine. 2: 992-
997 .

345
Montesano, R., Kumar, S., Orci, L. and Pepper, M.S ., 1996, Synergistic effect of
hyaluronan oligosaccharides and vascular endothelial growth factor on angiogenesis in vitro.
Lab Invest. 75: 249-62

Nicosia, R.F ., and Ottinetti, A., 1990, Growth ofmicrovessels in serum-free matrix culture
ofrat aorta. Lab Invest. 63 : 115-22 .

Noble, P .W., McKee, C.M ., Cowman, M . and Shin, H.S., 1996, Hyaluronan fragments
activate an NFkB/IkBa autoregulatory loop in murine macrophages. J Exp Med. 186: 2373-
2378 .

Noble, P .W ., Lake, F .R., Henson, P .M. and Riches, D .W.H., 1993, Hyaluronate activation
of CD44 induces insulin-like growth factor-l expression by tumor necrosis factor-o-
dependent mechanism in murine macrophages. J Clin Invest. 91: 2368-2377.

Rooney, P ., Wang, M ., Kumar, P . and Kumar, S., 1993, Angiogenic oligosaccharides of

hyaluronan enhance the production of collagen by endothelial cells . J. Cell Sei. 105: 213-8

Sattar, A., Kumar, S. and West D .C., 1992, Does hyaluronan have a role in endothelial cell
proliferation ofthe synovium. Semin. Arth. Rheum . 21: 43-49.

Sattar, A., Rooney , P ., Kumar, S., Pye, D., West, D.C., Scott, I. and Ledger, P ., 1994,
Application of angiogenic oligosaccharides of hyaluronan increase blood vessel numbers in
rat skin. J. Invest Dermatol. 103: 576-579

Shaw, D .M . (1996) Ph .D. Thesis, University ofLiverpool

Shaw, D .M ., West, D .C. and Hamilton, E., 1994,Hyaluronidase and hyaluronan in tumours
and normal tissues, quantity and size distribution. Int J Exp Pathol. 75: A67

Smedrod, B ., Pertoft, H., Eriksson, S., Fraser, J.R.E. and Laurent, T., 1984, Studies in vitro
on the uptake and degradation of sodium hyaluronate by rat liver endothelial cells. Biochem
J. 223: 617-26.

Timar, J., Moczar, E ., Timar, F ., Pal, K., Kopper, L., Lapis, K. and Jeney, A., 1987,
Comparative study on Lewis lung tumor lines with "low" and "high" metastatic capacity .
II. Cytochemical and biochemical evidene for differences in g1ycosaminoglycans. Clin. Expl.
Metastasis . 5: 79-87.

Trouchon, v., Mabilat, c., Bertrand, P., Legrand, Y, Smadia-Joffe, F ., Soria, C., Delpeche,
B. and Lu, H., 1996, Evidence of involvement of CD44 in endothelial cell proliferation,
migration and angiogenesis in vitro . Int. J. Cancer. 66: 664-668

Watanabe, M ., Nakayasu, K. and Okisaka, S., 1993, The effect of hyaluronic acid on
proliferation and differentiation of capillary endothelial cells. Nippon Ganka Gakkai
Zasshi . 97: 1034-9.

346
Weidner, N ., Sempie, 1.P., Welch, W.R., and Folkman, 1., 1991, Tumour angiogenesis
correlates with metastasis in invasive breast carcinoma, N Engl. J. Med. 324 : 1-8.
West, D .C., 1993, Hyaluronan receptors on human endothelial celIs: the effect of
cytokines. in: Vascular Endothelium: Physiological Basis 01 Clinical Problems II (J .D.
Catravas, ed), pp .209-210, Plenumm Press, New York

West, D .C., Hampson, LN., Arnold, F. and Kumar, S., 1985, Angiogenesis induced by
degradation products ofhyaluronic acid, Science. 228: 1324-8.

West, D .C. and Kumar, S., 1988, Endothelial proliferation and diabetic retinopathy . Lancet.
1: 715-6.

West, D .C., and Kumar, S., 1989a, Hyaluronan and angiogenesis, in: The Biology 01
Hyaluronan, Ciba Foundation Symposium 143 (D. Evered, . and 1. Whelan, eds), pp . 187-
207 , John Wiley and Sons, Chichester

West, D .C. and Kumar, S., 1989b, The effect of hyaluronate and its oligosaccharides on
endothelial proliferation and monolayer integrity . Exp Cell Res . 183: 179-96 .

West, D .C. and Kumar, S., 1991, Tumour-associated hyaluronan: a potential regulator of
tumour angiogenesis. Int J Radial. 61/62 : 55-60.

West, D .C, Shaw, D.M., Lorenz, P., Adzick, N.S/ and Longaker M ., 1996, Fibrotic healing
of adult and late gestation fetal wounds correlates with increased hyaluronidase activity and
removal ofhyaluronan. Int. J. Biochem Cell Biol. 29: 201-10 .

Yasui, T., Akatsuka, M., Tobetto., K, Umemoto, J., Ando, T., Yamashita, K., and
Hayakawa, T., 1992, Effects of hyaluronan on the production of stromelysin and tissue
inhibitor of metalloproteinase-l (TIMP-I) in bovine articular chondrocytes. Biomed Res.
13: 343-8.

347
CYTOKINES AND GROWTH FACTORS. EARLY GENE EXPRESSION
DURING ANGIOGENIC STIMULI

David BenEzra t, Shai Yarkoni", Genia Maftzir', Haya Loberboum-

Galski?

Department of Ophthalmology! and Department ofMolecular Biology

Hadassah Hebrew University Hospital and Medical School ,
Jerusalem, Israel

From accumulating experimental observations, it has become evident that angiogenesis - the
sprouting of new blood vessels from other mature vessels - can be induced by a myriad of
angiogenic factors' :" Despite their potential roles in initiating the process, these factors,
most probably, influence a single (or a few) events within a very complex cascade of
pathways which finally lead to angiogenesis as observed in vivo' .
This study was undertaken in order to investigate early genes exp ression of potential
angiogenic factors . Stimulation of angiogenesis was induced by 500 ng of
lipopolysaccharide (LPS) or basic fibroblast growth factor (bFGF) sequestered within
Elvax 40 and implanted in the rabbit cornea':' .

MATERIALS AND METHODS

Corneas ofmongrel albino rabbits weigh ing 2.5 to 3.0kg were used throughout.
Preparation of stimulants, implantation and monitoring of the angiogenic reaction were
carried out as described earlier': 4-6
At intervals of _' 1, 3, 6 and 24 hours after initiation of the angiogenic stimulus, rabbits
were sacrificed and their corneas subjected to fluorescent in situ hybridization technique
(FISH) . For each time interval, six corneas were studied simultaneously. Two corneas
stimulated by LPS, two corneas stimulated by bFGF and two corneas sham operated
receiving either an empty implant of Elvax 40 or an implant sequestering 500 ng of human
albumin were used as controls.
cDNA probes for the detection of various cytokines and growth factors expression within
the corneal limbal vessels endothel ial cells were carried out in a double masked manner.
The results obtained with cDNA probes for interleukin-l beta (IL-1 ß) and vascular
endothelial growth factor (VEGF) are i1Iustrated.

Angiogenesis: Model s, Modulators, and C/inical App/ications

Edited by Maragoudakis, Plenum Press, New York, 1998 349
 IL-l ß Probe
2.0
(LPS Implant)
IL-lb Probe
Control

1.5

1.0

0.5

0.0 -I-LLL.."-'
0.5 3 6
Time
Figure 1. Intensity ofmRNA signals for IL-Iß in limbal vessels endothelial cells as
observed after stimulation of angiogenesis by LPS implants . These signal
intensities for IL-l ß are compared to those detected in control comeas (sham
operated with empty implants ofElvax-40) The signals detected early after
implantation in the controls iIIustrate the production ofIL-lß following the
surgical trauma.

VEGF Probe
2.0 (LPS Implant>

III
";;j
:: 1.5 VEGF Probe
Oll
'in • Control
<:
Z
c:::
E 1.0
.....o
)(
QJ

-
"'0
::
0.5

0.0

Time
Figure 2. Intensity of mRNA signals for VEGF in limbal vessels and endothelial cells is
observed after stimulation of angiogenesis by LPS implants. These signal
intensities for VEGF are compared to those detected in control comeas. Note
that the signal intensity for VEGF in control comeas is similar to that observed
for IL-l ß probe in figure 1 and observed only during the first hour are
stimulation.

350
 IL-l ß Probe
1.0
(bFGF Implant)

0.8
s:"=
C
CO • IL-b Probe
"üi
-< 0.6 • Control
Z
c::
...E
0 0.4
><
CU
-0

'-C

0.2

0.0
0.5 3 6 24
Time
Figure 3. Intensity of mRNA signals for IL-1ß in limbal vessels endothe1ial celIs as
observed after stimulation of angiogenesis by bFGF implants. These signal
intensities for IL-1ß are compared to those observed in contro l corneas
(implants sequestering 500 ng of human albumin). In these control corneas,
mRNA for IL-1ß positive celIs were detected even three hours after
implantat ion ofthe experiments. The signal intensities however are comparable
to those observed for the sham operated comeas with empt y implants of
Elvax-40.

VEGF Probe
1.0 (bFGF Implant>

0.8
III
r: VEGFProbe
=
eo • Control
'üi 0.6
<:
zc::
c
004
'""'0
><
cu
-0
.5 0.2

0.0
0.5 3 6 24
Time
Figure 4. mRNA signals for VEGF in limbal vessels endothelial celIs as observed after
stimulat ion of angiogenesis by bFGF implants. The signals for VEGF in
mRN A in these comeas are only slightly higher than those observed in control
corneas with implants sequestering human albumin.

351
RESULTS

Figure 1 illustrates the intensity of expression of mRNA to IL-Iß in corneas stimulated

with LPS and those sham operated. The control Iimbal vessels endothelial cells did not
demonstrate any significant fluorescence over background . In corneas stimulated with
LPS , an early and strong signal is detected as early as _ hour after initiation of the stimulus.
The intensity of the signal increases at one hour and remains present during a11 tested
intervals. The signals for VEGF mRNA expression on the other hand was undetectable _
hour after LPS stimulation, similar to control corneas . Positive signals for VEGF mRNA
were detected one and three hours after LPS stimulation with an expression slightly over
background controls at six and 24 hours (figure 2).
When stimulation was carried out with bFGF , strongest signals for IL-1ß mRNA were
detected _ hour after initiation of the experiments. The signal for IL-1ß mRNA remained
strong during the first three hours decreasing thereafter (Figure 3). A weak expression of
VEGF mRNA was detected during the first _ hour after stimulation of angiogenesis bFGF
(Figure 4). These signals increased slightly at one hour decreasing at 3 hours and
disappearing thereafter (Figure 4).

DISCUSSION

These observations and additional unpublished results demonstrate unequivocally that

stimulation of angiogenesis in vivo is set in motion by the expression of multiple factors . In
the system used for this study , different stimulants of angiogenesis (LPS and bFGF)
induced the early gene expression of similar cytokines and growth factors . Based on the
timing and strength of expression ofthe various cytokines and growth factors , it is difficult
at the present stage to speculate on the exact roles of these factors in the cascade of events
leading to the sprouting of new blood vessels. However, it can easily be deducted from the
results reported herein that VEGF mRNA expression is only one event among many
others. Possibly, a factor's intensity of mRNA expression and its timing of appearance
can represent the possible importance of the role(s) fulfilled by this factor in angiogenesis.
Ifyes, of all the factors studied , IL-Iß can be singled out as one of those playing a pivotal
role in the cascade of events leading to the sprouting of new blood vessels - angiogenesis.
Additional experiments and studies are needed in order to ascertain this postulation or
refute it.

REFERENCES

BenEzra D. Neovasculogenic ability of prostaglandins, growth factors and synthetic

chemoattractants. Am . 1. Ophthalmol. 86:455-461 , 1978.
Folkman J, Klagsbrun M. Angiogenic factors . Science . 235: 442-447, 1987.

BenEzra, D. Neovasculogenesis. Triggering factors and possible mechanisms.Survey

Ophthalmol. 24 :167-176,1979.

BenEzra, D., Hemo, 1., and Maftzir, G. In vivo angiogenic activity ofinterleukins. Arch .
Ophthalmol. 108:573-576, 1990.

352
BenEzra, D., Griffin, B.W ., Maftzir, G. and Aharonov, O. Thrombospondin cnhanee s in
vivo angiogenesis indueed by bFGF or lipopolysaeeharide. Invest. Ophthalmol. Vis. Sei.
34:3601-3608, 1993.

BenEzra , D. The eorneal model for the study of angiogenesis. Angiogenesis : Moleeular
Biology, Clinieal Aspeets Maragoudakis ME (Ed.) Plenum Press, New York, 1994 pp.
315-320 .

353
THE ROLE OF THE LAMININ PEPTIDE SIKVAV IN THE

REVASCULARIZATION OF ISCHEMIC TISSUE

R.Wesley Rose', Richard C. Morrisorr', Michael G. Magno/, John Manniorr',

Hynda K. Kleinmarr', and Derrick S. Grant!"

1 The Cardeza Foundation for Hematologic Research, Thomas Jefferson

University, Philadelphia, PA
2 Department of Surgery, Thomas Jefferson University, Philadelphia, PA
3 National Institute for Dental Research, National Institutes of Health, Bethesda,
MD . .

INTRODUCTION

Angiogenesis is a processes of new capillary formation from preexistrng vessels in

response to cytokine stimuli. It involves the breakdown of the surrounding basement
membrane, proliferation and migration of the endothlial cells which comprise the vessel wall
toward the angiogenic stimulus, and subsequent secretion of basement membrane, leading to the
formation of a new capillary branch . This process is dynamic, rather than occurring in discrete
steps. Angiogenesis is necessary for many physiological (development, reproductive cycle) and
pathological (tumor growth and metastasis, wound healing) processes. Previously, several
investigators have explored the use of angiogenic agonists in the revascularization of ischemic
tissue. We have recently exarnined in vivo the ability of SIKVA V (a peptide derived from the
alpha chain of larninin-l) to revascularize ischemic tissue. In this manuscript, we review the
work done with SIKV AV and its role in angiogenesis as it pertains to the revascularization of
ischemic tissue .

The role oflaminin in angiogenesis

The basement membrane is a biologically active structure in direct contact with the
endothelium of the vessels, and has a strong influence on the maintenance and stability of the
endothelium (lngber et al., 1989), as well as on the regulation of angiogenesis (Grant et al.,
1990). The basement membrane consists of numerous glycoproteins and proteoglycans, namely
larninin, collagen IV, entactin, and perlecan (Kleinman et al., 1987) . The laminins are a family
of large trimeric glycoproteins, several members of which have been shown to promote cell
adhesion, migration, proliferation, and differentiation (Kleinman et al., 1985). Currently, the

Angiogenesis: Mod els, Modulators, and Clinical Applications

Edited by Maragoudakis, Plenum Press, New York, 1998 355
laminin family consists of ten isoforms (Malinda et al., 1996). One predominant isoform,
laminin-I (EHS-laminin; 800,000- 1,000,000 Da) consists of three disulfide-linked chains ,
designated alpha 1 (Mr=400,000), beta 1 (Mr=21O,000) and gamma 1 (Mr=200,000 ), all of
which have been cloned and sequenced (Yarnada et al., 1987). Specific biologically active
regions of laminin-I have been identified based on the activity of synthetic peptides
corresponding to sequences in the protein produced by elastase digestion (Engel, 1991).
Previously, our studies showed that endothelial cells utilize the laminin-I RGD sequence for
attachrnent to laminin, whi1e the YIGSR site is involved in both attachrnent and cell-cell
interactions during morphogenesis . A third active sequence in the a1 chain domain, designated
PA22-2 (CSRARKQAASIKVAVSADR), promotes neurite outgrowth of PC12 cells and
increases the number of metastatic lesions to the lung by melanoma cells (Sephel et al., 1989;
Kanemoto et al., 1990). The peptide also induces an increase in melanocyte motility,
invasiveness into Matrigel, and most importantly, an increase in collagenase IV (gelatinase)
activity, a key enzyme in basement membrane degradation (Goldberg et al., 1990, Kanemoto et
al., 1991). The peptide also increases plasminogen activation (reviewed in Grant et al., 1992)
(Table 1). These observations led to investigation of whether SIKVAV played a role in the
stimulation of angiogene sis, since cell motility, invasiveness, and collagenase activity are all
important mechanisms of angiogenesis.

Angiogenic activity of SIKVAV

In order to Table 1. Activities of SIKVAV
deterrnine whether
I n vitro In vivo
SIKVAV was an angio-
genie agonist , the syn-
• sprouting from tubes • angiogenic

thetic SIKVAV peptide • cell adhesion • subcutaneous tumor growth

was examined for its • cell migration • experimental tumor metastasis

effects on endothelial • cell proliferation • netrophil chemoattractant
cell activities related to • neurite outgrowth • increase cytokine release from
neutrophils
steps in the angiogenic • collagenase IV activity
process inc1uding ar- • p1asminogen activator activity • mast cell adhesion
tachment, migration,
collagenase production, and tube formation on Matrigel (Grant et al., 1992). We also observed
that different synthetic peptides containing SIKVAV could elicit a different degree of HUVEC
attachrnent. It should be noted that if the isoleueine in the peptide is changed to glycine there is a
complete loss of activity, demonstrating the specificity of this peptide for endothelial cell
binding . Cells plated on SIKVAV-coated plastic appear less spread, exhibiting elongated cell
extensions, and are more migratory. Generally in vivo, endothe1ia1 cells change their appearance
and become elongated prior to mobilizing and becoming angiogenic; this effect is also seen when
cells are treated with TNF or scatter factor (Rosen et al., 1993), or after physical denuding of a
confluent layer. When endothelial cells were seeded on tissue culture plastic coated with either
laminin or Matrigel in the presence of the peptide, a 25% increase in cell binding was observed

356
within a 30 min period. Preliminary studies indicate that several integrins (including a2ß I, a4ß6
and roßt ) and the laminin bindin g protein , LBP 32/67 increase on the cell surface when cells are
incubated on the SIKV AV peptide. It was therefore concluded that SIKVAV play s a major role
as an angiogenic stimulator.
SIKV AV has also been shown to be angiogenic in vivo (Grant et al., 1992; Kibbe y et
al., 1992). The Matrigel mouse assay (Passaniti et al., 1992) is a useful model for the
assess ment of angiogenesis in vivo. Brietly, a bolu s of Matrigel (4°C) containing the test factor
is injected subcutaneously into the groin region of a C571B16 mouse, where it gels. Seven to ten
days post-injection, the plug is exised and sectioned, and the number of vessels can be
quantitated. SIKV AV increased mean vessel density in the Matrigel plug s in a dose -dependent
mann er as compared to contr ol, at doses as low as 10 ug (Kibbey et al., 1992). It was therefore
clear that SIKVAV played a role in the stimulation of angiogenesis in vivo. The next step was to
determine the role of SIKV AV in the revascularization of ischemic tissue.

Tissue ischemia
Tissue ischemia due to blood vessel occlusion is a major factor in many pathological
conditions (such as myocardial infarction and stroke ), and is extremel y problematic for diabetic s
(rev iewed in Stonebridge et al. , 1993). The distribution of macrovascular occlusive disease in
diabetic patient s seems to differ from that of non-di abet ics. The problems assoc iated with
occlusive arteries in diabetics is further exacerbated by the fairly low number of native and
functional collateral vess els found below the knee. In addition, existing collateral s often have
dysfunctional shunting mechan isms due to in part to diabetic peripheral neuropathy. In many
patients infrainguinal bypass has been very successful and the literature cont ains significant data
that sugges t that an early aggressive surgical treatment of lower-limb occlusive arterial disease
may reduce the frequenc y of ischemia or amput ation in diabetic patients. It should be pointed out
that althou gh 80% of infrainguinal bypass surgical treatment are success ful, anoth er 10-15% of
case s fail and lead to arnputation (Kwolek et al. , 1992; Reichle et al., 1974 , Oliveira et al.,
1993 ). Therefore, an alternative treatment is needed for patients who may be at risk of losing
patency in the bypasss vessel and subsequently undergoing amputation.
Diabetic patients also have a higher incidence of cardiovascular occlu sion and myocardial
infarction. The most common form of heart disea se (80 %) that results in heart failure is due to
vascular insufficiency leading to ischemia and subsequent infarct. Most hearts have some
intercoronary anastomoses which may be as large as 100 11m in diameter. In most cases, chronic
ischemia leads to extensive collateralization; however blood tlow is insufficient to supply the
cardiac tissue . Ischemia usually affects the inner 1/3 to 2/3 of the wall of the heart (the
subendocardial region ). Localized branching of coronary arterioie s and capillaries usually helps
to decrease the degeneration of the local muscle mass but may only result in sustaining 10-20 %
of the ischemic area.

357
Vascular growth factors and SIKVA V in the revascularization of ischemic tissue
A possible treatment for the prevention of tissue infarction might involve the local
administration of an angiogenic compound in the area of ischemia. This local administration
would stimulate revascularization of the ischemic tissue and could rescue it from the irreversible
damage caused by infarction. Several studies have investigated the efficacy of enhanced
revascularization of ischemic tissue stimulated by local release of angiogenic factors. Two
angiogenic growth factors , basic fibroblast growth factor (bFGF) and vascular endothelial cell
growth factor (VEGF), have been successfully used to revascularize ischemic tissue (Harada et
al., 1994; Banai et al., 1994). The mechanism of tissue ischemia in the heart, the angiogenic
responses and the resulting expression of growth factors is very similar to the ischemic limb. It
is unclear, however, which cells are producing the angiogenic factors or if this local increase in
angiogensis can stimulate sufficient angiogenesis and blood flow to prevent long-term ischemia
or infarcts leading to death of the animal. Therefore, exogenous addition of growth factors has
been investigated as a possible treatment for tissue ischemia. For example, bFGF has been
shown to be a potent angiogenic factor. Others (Yanagisawa-Miwa et al., 1992) have exarnined
the effect of localized administration of bFGF to ischemic canine myocardium , and found an
increase in coronary capillaries in ischemic hearts. However, bFGF has a short half life (in the
order of seconds to minutes) in vivo and under surgical procedures requires either an extremely
large initial dose or repeated or continous administration (Hickey et al., 1994). For these
reasons, its use in a clinical setting might be cost-prohibitive . Finally, bFGF has been shown to
have an effect on cell types other than vascular cells (Banai et al., 1994). VEGF has shown
some promise as to its specificity, and has been shown to induce revascularization in the
ischemic hindlimb of rabbits (ref). These growth factors appear to require heparin, which plays
an important role in the interaction with their receptor; this includes (although to a lesser extent)
VEGF . The larninin derived peptide SIKVAV seems not to require heparin and has a direct
effect on endothelial ceHbehavior. The SIKVAV peptide is smaller than the other growth factors
and can be synthesized; therefore it has potential clinical applications and may be practical to use
alone or in combination with VEGF in myocardial revascularization. Therefore, it was our intent
to deterrnine both in vitro and in vivo whether SIKVAV could be used in the revascularization of
ischemic tissue.

Activity of SIKVAV in vitro on human coronary artery endothelial

cells
Recently we exarnined the effect of the SIKVAV peptide on coronary artery endothelial
cells, since we previously examined its effect on venous cells (HUVEC) . In order to assay the
effect of the peptide on angiogenesis in vitro , we exarnined cell attachment and migration, two
integral mechanisms of angiogene sis. We also exarnined the effect of SIKVAV on human left
internal mamary artery (LIMA) sprouting.

358
Attachment
We examined the effect of SIKVAV on the attachrnent of human coranary artery
endothelial cells (HCAEC, passage 200 • C I '
"C ontro peptlde • p<0.05
12-14; a gift fram Dr. Jeffrey Stadel, Qj
;:
180 0 SIKVAV
SmithKline Beecham) to laminin. Ci 160
c.
HCAEC were incubated with either "C
.c
GI 140
SIKVAV or control peptide at o
III 120
concentrations of 0, 50 or 100 ug/ml =:
III
100
at 3TC for 20 minutes. Twenty-four .!!! Qj
o 80
weil plates (Nunclon , Denmark) were ö
60
incubated for 1 hour with laminin
~ 40
(20~g/ml) in serum free medium at §
3rc. 40,000 cells per weIl were
z 20

added and incubated at 37°C for 30

o
o 50 100
minutes. All wells were then washed, Peptide (1l 9)
and the attached cells were fixed and
Figure 1. Attachrnent of HCAEC (preincubated in
stained for quantitation. The number
SIKVAV or Control peptide) to laminin
of cells attached per square milimeter
were quantitated micrascopically.
SIKVAV demonstrated a dose-dependent increase in HCAEC attachrnent over the control
peptide (Figure I). No signifcant difference in cell attachment at any dose of control peptide was
observed . SIKVAV at both 50 and 100 ug doses resulted in a significant increase in attachrnent
over no peptide, as weIl as contral 200 • Control peptide
'0 • p<0.05
peptide, at the same dose. Qj
180 0 SIKVAV
;:
Therefore, preincubation of Ci 160
HCAEC with SIKVAV resulted in ~
GI 140
a dose-dependent increase in cell ~
attachrnentto laminin .~ 120
E 100
.!!!
Migration 1j 80
The effect of SIKVAV on ö... 60
cell migration was also examined .8 40
using a modified Boyden chamber § 20
Z
assay. Falcon tissue culture
inserts (8 um pores; Becton-
o
o 250 500
Dickinson, Lincoln Park, NJ) Peptide (119)
were placed in twenty-four weIl
culture plates. Serum free medium Figure 2. Chemotactic migration of HCAEC
was placed in the lower chamber toward a SIKVAV or contral peptide gradient.
with SIKVAV or control peptide at
0, 250 or 500 ug/ml . Aliquots of serum-free medium with suspended HCAEC were added to

359
the upper chamber. Inserts were fixed and stained after 3 hours. All cells which had not
migrated through the pores were removed by wiping the upper surface of the membrane .
A dose-dependent increase in migration for both 250 and 500 ug doses of SIKVAV was
observed for HCAEC (Figure 2). AdditionaIly, the increase of SIKVAV dose from 250 to 500
ug resulted in a significant increase in cell migration. Therefore, it was determined that SIKVAV
was a potent chemoattractant for HCAEC.

FibrinlLejt Internal Mammary Artery (UMA) Ring Assay

The human LIMA assay was also used to demonstrate the effect of SIKVAV on
angiogenesis, since this in vitro model more accurately approximates the in vivo environment.
Human internal mammary
artery, obtained at time of Table 2. Total number of sprout s from the human LIMA rings
coronary artery bypass in the presence of SIKVAV, control peptide, or no peptide.
grafting , was stripped of Treatment Day 4 Day 5 Day 7
adventitia, cut into lmm No peptide 26 75 218
lengths and washed SIKVAV 46 125 312
extensively. Purified Control peptide 25 78 214
fibrinogen Cl mg/rnl,
9OOuVweIl) was aliquotted into twenty-four weIl plates, and thrombin was then added (final
concentration of 1 V/mi), and the arterial ring placed in the weIl so that the lumen was parallel to
the base of the weIl. After formation of the fibrin clot, low-serum medium containing SIKVAV
or control peptide (300ug/weIl) was added to the wells . The number of vascular sprouts from
the cut ends of the rings was quantitated using phase contrast microscopy.
Rings incubated in SIKVAV exhibited a greater number of sprouts than rings incubated
in either control peptide or no peptide, as early as day 4 (Table 2). The SIKVAV-induced
sprouts were longer and demonstrated a greater degree of branching as weIl. Therefore,
SIKVAV enhanced human LIMA sprouting as compared to no treatment or treatment with a
control peptide .

Activity of SIKV AV in vivo

Since the in vitro data showed that SIKVAV was a potent inducer of angiogenic
behavior in HCAEC, we tested the ability of the peptide using an in vivo rat model to enhance
the formation of collateral circulation in the ischemic hindlimb, similar to the one previously
described by Chleboun et al. Chleboun et al., 1992).
Briefly, rats were anesthetized, and a longitudinal incision was made, extending
inferiorly from the inguinal ligament to a point 7 mm proximal to the patella. Though this
incision, the femoral artery of one leg was ligated proximal to the adductor hiatus (Figure 3) ,
while the contralateralleg remained unligated. A slow-release pellet consisting of Elvax only ,
control peptide (500 ug), bFGF (20 ug), or SIKVAV (500 Ilg) was placed subcutanously in the

360
region of the popliteal fossa. The incision
was then sutured, and after two weeks,
the thoracic aorta was cannulated and
perfused first with heparinized saline and
then with a radiopaque silicon latex
compound. Radiographs were taken of
the lower limbs, and sections of several
tissues (skin, foot pad) and muscles
(gastrocnemius, quadriceps) were excised
and fixed in 4% paraformaldehyde for
histological processing. Twenty rats per
experiment were used.
Medial
Analysis of the radiographs
showed an increased degree of Lig lure Plaeed Aroun
the Femoral Ar1ery
revascularization of ischemic limb by
SIKVAV as compared to no peptide or
control peptide (Figure 4A-C). Increased
tortuosity of vessels supplying the
Figure 3. Placement of the silk ligature in the in
ischemic tissue was also observed,
vivo model of ischemia
indicating neovascularization (Figure 4C).
An increased supply of blood (as compared to control peptide) to the lower limb and foot pad

A B

ischemla

/'
Sham operated Contral peptide
C

SIKVAV 2 weeks Soleus , Control Soleus, SIKVAV

Figure 4. Radiograph of radiopaque latex-perfused rats showing revascularization of ischemic

hindlimb, Shown are rats treated with a) nothing, b) control peptide, and c) SIKVAV. Panel
D shows histology of ischemic soleus treated with nothing (left) and SIKVAV (right)

361
was observed (Figure 4C) . Control peptide, however, showed little revascularization of the
hindlimb, as indicated by the non-perfused areas in the lower limb and foot pad (Figure 4B ,
arrows ). The degree of revascularization in response to bFGF treatment was similar to that
observed for control peptide (data not shown). Histologically, there were a greater number of
perfused vessels in the musculature of the ischemic leg treated with SIKVAV than the
contralateralleg treated with nothing (Figure 4d). When the calf regions of the radiographs were
compared using computer-enhanced morphometric analysis, a greater increase in vessel density
normalized to the contralateral limb was observed for ischemic tissue treated with SIKVAV as
compared to ischemic tissue treated with control peptide and bFGF (Figure 5). Therefore , this
indicated that SIKVAV was able to induce greater revascularization of ischemic hindlimb tissue
in this in vivo model than bFGF or control peptide.

SUMMARY
SIKVAVis a potent angiogenic peptide
100
derived from laminin-l. It stimulates ceIl
invasion, attachment, migration and
~
tlI
80
proliferation not only in human venous cells 30 _tl1
but arterial endothelial ceIl as weIl. These IIlE
tlI(I) 60
>~
findings indicate that SIKVAV can stimulate Co
- 1Il
(1)-
arterial vessel growth in vivo , similar to 1Il"
tlI(I) 40
(1):::
~tlI
growth factors such as VEGF , to become
.s
angiogenic and form coIlateral branche s. This :.e
0
20

is an important finding , since these processes

are integral to angiogenesis. In the human
0
U.
Ö
u.
..
N

>
...
>
co
:2
-c -c Ci.
LIMA capillary sprouting assay, SIKVAV
.c
> > a
increased not only the number of sprouts, but
Treatment '"Vi '"Vi ec
8
the length and branching of those sprouts as
Figure S. Effectiveness of the various
compared to control , which agrees with the agents in revascularizing ischemic tissue .
attachment and migration findings . In vivo, Shown is the percent decrease in vessel
density of the ischemic limb compared to
SIKVAV was shown to be a potent angiogenic the normaliimb. The rats receiving elvax
agent in the revascularization of the ischemic alone (vehicle) showed no appreciable
increase in vascularity after ischemia.
hindlimb of the rat, as verified by radiography
and computer assisted image analysis.

CONCLUSI ON

In this study, we investigated in vivo the ability of SIKVAV to revascularize ischemic

tissue in the rat model. Currently, we are evaluating the use of SIKVAV in conjunction with
bFGF and VEGF. Preliminary data show an upregulation of the mRNA for the KDR receptor
for VEGF in HUVEC incubated with SIKVAV. If the vasculature of an ischemic region could
be made more sensitive to VEGF by SIKVAV treatment, then a lower dose of VEGF could be

362
used to attain revascularization, thus alleviating some of the problems associated with the use of
VEGF clinically. Therefore, SIKVAV rnight prove to be a valuble clinical treatment for tissue
ischemia due to its relative ease of synthesi s and cost-effectiveness,
Therefore, we conclude that the laminin-I derived peptide SIKVAVis a potent
angiogenic factor that shows prornising potential clinical applications in the revascularization of
ischemic tissue, either alone or in conjunction with one or more growth factors.

REFERENCES

Banai, S., Jaklitsch, M.T., Shou, M., Lazarous, D.F., Scheinowitz, M ., Biro, S .,
Epstein, S.E., & Unger, E.F. (1994) . Angiogenic-induced enhancement of collateral blood flow
to ischernic myocardium by vascular endothelial growth factor in dogs . Cireulation , 89(5) ,
2183-2189.
Chleboun, J.O ., Martin s, R.N., MitchelI, c.A., & Chirila, T.V. (1992) . bFGF
enhances the development of the collateral circulation after acute arterial occlusion. Bioehern
Biophys Res Cornrn , 182(2) , 510-516.
Engel, 1. (1991) . Domains in protein s and proteoglycans of the extracellular matrix with
functions in assembly and cellular activities. Int J Biol Macrornol, 13(3), 147-151.
Goldberg, G.I., Frisch, S.M ., He, C., Wilhelm, S.M., Reich , R., & Collier, I.E.
(1990) . Secreted proteases. Regul ation of their activity and their possible role in metastasis. Ann
N. Y. Aead. Sei ., 580, 375-384.
Grant, D.S., Kinsella, J.L. , Fridman, R. , Auerbach, R., Piasecki, RA ., Yamada, Y .,
Zain, M ., & Kleinman, H.K. (1992). Interaction of endothelial cells with a larninin Achain
peptide (SIKV AV) in vitro and induction of angiogenic behavior in vivo. J Cell Physiol, 153(3),
614-25.
Grant, D.S., Kleinman, H.K., & Martin, G.R. (1990). The role of basement
membranes in vascular development. Ann N Y Aead Sei , 588(61), 61-72 .
Harada, K., Grossman, W., Friedman, M ., Edelman, E.R., Prasad, P.V., Keighley,
C.S ., Manning, WJ ., Sellke, F.W., & Simons, M. (1994). Basic fibroblast growth factor
improves myocardial function in cronically ischernic porcine hearts . J. Clin. Inv est., 94, 623-
630.
Hickey, M .J., & Morrison, W.A. (1994). An improved matrix-type controlled release
system for basic fibroblast growth factor. Bioehern. Biophys. Res. Corn., 201(3), 1066-1071.
Ingber, D.E ., & Folkman, J. (1989) . Mechanochernical Switching between Growth and
Differentiation during Fibroblast Growth Factor-stimulated Angiogenesis In Vitro: Role of
Extracellualr Matrix . J. Cell Biol., 109,317-330.

363
 Kanemoto , T. , Martin, G.R. , Hamilton, T.C., & Fridman, R. (199 1). Effects of
synthetic peptides and protease inhibitors on the interaction of a human ovarian cancer cell line
(NIH:OVCAR-3) with a reconstituted basement membrane (Matrigel). Invasion Metastasis,
11(2),84-92.
Kanemoto, T., Reich, R., Royce, L., Greatorex, D., Adler, S.H ., Shiraishi, N . ,
Martin, G.R., Yamada, Y., & Kleinman, H.K (1990). Identification of an amino acid sequence
from the larninin Achain that stimulates metastasis and collagenase IV production. Proc. Natl.
Acad. Sei. USA, 87, 2279-2283.
Kibbey, M.C. , Grant, D.S., & Kleinman, H.K. (1992). Role of the SIKVAV site of
larninin in promotion of angiogenesis and tumor growth: An in vivo Matrigel model. J. Nalt.
Cancer Inst., 84, 1633-1637.
Kleinman, H.K , Cannon , F.B., Laurie, G.W., HasselI, J.R., Aumailley, M.,
Terranova, V.P., Martin, G.R., & Dalcq, M.D.B. (1985). Biological activities of laminin. J.
Cell Bio/., 27, 317-325.
Kleinman, H.K ., Graf, J., Iwamoto, Y., Kitten, G.T., Ogle, R.C., Sasaki, M.,
Yamada, Y., Martin , G.R., & Luckenbill-Edds, L. (1987). Role of basement membranes in cell
differenti ation . Ann . N. Y. Acad. of Sci., 513, 134-145.
Kwolek , c.J., Pomposelli, F.B., Tannenbaum , G.A., Broph y, C.M., Gibbons, G.W .,
Campbell, D.R., & LoGerfo , F.W. (1992). Peripheral vascular bypass in juvenile-onset
diabetes mellitus: are aggrassive revascularization attempts justified? J. Vase. Surg. , 15, 394-
401.
Malinda, KM. , & Kleinman, H.K (1996). The Laminins. Int. J. Biochem. Cell Bio!. ,
28(9), 957-959.
Oliveira, M., Wilson, S.E., Williams, R., & Freischlag, J.A. (1993). lliofemoral
bypass: a l O-year review. Cardiovascular Surgery, 1(2), 103-106.
Passaniti , A., Taylor, R., Pili, R., Guo, Y., Long, P., Haney, J., Pauly, R., Grant , D. ,
& Martin, G. (1992). A simple, quantitative ' method for assessing angiogenesis and
antiangiogenic agents using reconstituted basement membrane, heparin, and fibroblast growth
factor. Lab Invest, 67(4), 519-28.
Reichle, F.A., & Tyson, R.R . (1974) . Femoroperoneal bypass: evaluation of potential
for revascularization of the severely ischemic lower extremity. Ann Surg, 181(2), 182-185.
Rosen, E.M ., Zitnik, RJ., Elias, J.A ., Bhargava, M.M., Wines, J., & Goldbe rg, LD .
(1993). The interaction of HGF-SF with other cytokines in tumor invasion and angiogenesis.
Exs, 65(301 ), 301-10 .
Sephel , G.c., Tashiro, K-I. , Sasaki, M., Greatorex, D., Martin, G.R., Yamada, Y., &
Kleinman, H.K. (1989). Larninin Achain synthetic peptide which supports neurite outgrowth.
Bioch em. Biophys. Res. Com. , 162(2), 821-829 .

364
 Stonebridge , P.A., & Murie, J.A. (1993). Infralingual revascularization in the diabetic
patient. Br. J. Surg., 80, 1237-1241.
Yamada, Y., Albini, A., 1. Ebihara, Graf, J., Kato, S., Killen, P., Kleinman, H.K .,
Kohno, K., Martin, G.R ., Rhodes, C, Robey, F.A., & Sasaki, M. (1987). Structure,
expression, and function of mouse laminin. (eds. J .R. Wolff et al. Springer-Verlag, Berlin,
Heidelberg).
Yanagisawa-Miwa, A., Uchida, Y., Nakamura, F., Tomaru, T., & Kido, H. (1992) .
Salvage of infarcted myocardium by angiogenic action of basic fibroblast growth factor. Science ,
257(5075), 1401-3.

365
OXIDIZED LIPOPROTEINS ENHANCE THE IN VITRO TUBE FORMAnON BY
ENDOTHELIAL CELLS CUL TURED ON MATRIGEL.

Bruce Sundstrom, Kala Venkitesnaran, Periasamy Selvaraj and Demetrios

Sgoutas.

Department of Pathology and Laboratory Medicine , School of Medicine,

Emory University Atlanta, GA. 30322, USA .

When U937 monocyte-like cells were pretreated with oxidized low density lipoproteins
(LDL) or oxidized lipoprotein (a) [Lp(a») and subsequently the U937 cells were co-
cultured with human microvascular endothelial cells (HMVECs), or human umbilical
vascular endothelial cells (HUVECs) seeded on Matrigel, angiogenic differentiation and
formation of microtubules were enhanced. The effect was depended on the concentration of
the oxidized lipoproteins and it was negligible when U937 was treated with native
lipoproteins. Under optimum conditions, 250 ug LDL or 200 ug oxidized Iipoprotein(a)
[Lptaj] , respectively, preincubated with 106 U937 cells per ml of media, for 24 hours ,
followed by 24 hours of co-culture with endothelial celIs (EC) seeded on Matrigel (50,000
cells per 300 uL Matrigel), gave similar results with the average capilIary length of the
microtubules increasing from 160±25 to 400±40um . When potentially active components
of oxidized LDL, like malondialdehyde, hexanal, 4-hydroxynonenal and Iysophospholipids
were pre-incubated with U937 cells prior to co-culturing them with HMVECs and
HUVECs on Matrigel, only malondialdehyde and Iysophospholipids, gave an effect similar
in size to the oxidized lipoproteins. Antibodies (moAb Mac-I) directed against the
CDllb/CDI8 epitopes inhibited adhesion ofU937 celIs to EC and formation ofmicrotubes
when used in conjunction with moAbs towards ICAM-I(CD54). Blockage of tube
formation by antibodies to Mac-I and ICAM-I seems to suggest that there is a link
between the formation of a complex between adhesion molecule CD 11b/CD 18 and
adhesion moleeules CD54 from ICAM-I and tube formation as observed in co-cultures of
activated U937 and HMVECs or HUVECs grown on Matrigel.

1. INTRODUCTION

We have shown recently (Ragab, Selvaraj, and Sgoutas, 1996) that equimolar concentrations
of oxidized low-density lipoproteins (LDL) and oxidized lipoprotein (a) [Lptaj], inhibited
the growth of U937 monocyte-like cells by 64±7.1% and 84.3±8.2%, respectively, when

Angiogenesis: Models, M odulators, and Clinical Applications

Edited by Marago udakis, Plenum Press, New York, 1998 367
compared to native LDL and native Lp(a). In addition, we showed that equimolar
concentrations of oxidized Lp(a) and oxidized LDL induced adhesion molecule, Mac -1
(CD-11b) expression in U937 cells by 64.7±15 and 58±6.1% (p>0.05), respectively, of the
effect produced by the phosphokinase C activator, phorbol 12-palmitate-13-acetate ester
(PMA) (p<O.Ol). U937 cells incubated with oxidized LDL and oxidized Lp(a), showed an
adherence to cultured endothelial cells at 42±5.2% and 34±4.8%, respectively (p<0.05), of
the adherence shown by the same cells activated by PMA (p<O.Ol).
The aim ofthis study was to investigate whether oxidized lipoproteins by activating U937
cells can induce endothelial cells to produce capillary tubes, when cultured on a complex
artificial basement membrane, Matrigel. Our hypothesis is that Mac-1 expressed in
monocytes is a potent inducer of new blood vessel growth .

2. MATERIALS AN» METHODS

2.1. Lipoprotein Preparation.

Whole blood was collected in vacutainers containing 1.5 mg of dipotassium EDT Alm I of
blood, from ovemight fasting, healthy individuals with high LDL and/or high Lp(a) plasma
concentrations. Plasma was obtained by centrifugation and the solvent density of the
plasma was adjusted with solid KBr to 1.050 glml and centrifuged at 100000 x g at 10°C
for 24 hours in a Ti70 fixed-angle rotor on a Beckman L5-75B ultracentrifuge (Havel, Eder,
and Bragdon, 1955). The floating LDL particles were removed and stock LDL suspensions
containing EDT A were stored at 4°C under nitrogen in the dark for a maximum of 5 days .
In order to isolate Lp(a), after removal of LDL, the infranatant of those individuals with
high Lp(a) was adjusted with KBr to 1.100 glrnl and recentrifuged under the same
conditions for 48 hours. The 1.050-1.100 glml, Lp(a) enriched fraction , was further purified
by column chromatography over Biogel A-15m (Bio-Rad Laboratories) in order to exclude
any LDL, as previously described (Ragab, et al., 1996). The purified LDL and Lp(a) gave
single bands on agarose electrophoresis . Before the experiments, EDT A was removed from
the LDL and Lp(a) sampies by chromatography on sephadex columns (PD-lO, Sephadex
G25M ; Pharmacia). Oxidative modification was achieved by incubation in 7.5 uM copper
ions for 6 hours at 37°C. It has been repeatedly shown that this method gives similar
results with those obtained during incubation of lipoproteins with cultured cells
(Henriksen, Mahoney, and Steinberg, 1981). The copper ions were removed by EDTA (1.5
mg of dipotassium EDT Alm I of lipoprotein suspension) and column chromatography (PD-
10, Sephadex). Lipoprotein preparations were sterilized by passage through a membrane
filter (Millex-GV, pore size 450nm; Millipore) and stored under nitrogen at 4°C.

2.2. U937 Cell Culture.

The monocytic U937 cell-line was grown in RPMI 1640 medium supplemented with 10%
v/v heat inactivated fetal calf serum (FCS) and gentamycin sulfate (50 uglml) and kept in an
atmosphere of5% carbon dioxide and 95% air, as previously described (Fostegard , Nilsson ,
Haegerstrand, Hamsten, Wigzell, and Gidlund, 1990).

2.3. Expression of Surface Antigens.

U937 cells were counted and 1 rnl of cell suspension (500000 cells/ml) was added to each
weil ofa multiweIl plate containing native, oxidized LDL, native and oxidized Lp(a) at the
indicated concentrations each and PMA at 1.6 x 10-7 M. The incubation at 37°C lasted for
24 hours . The U937 cells were washed three times in PBS containing 0.1% bovine albumin
and 0.02% sodium azide. The pellet was then gently suspended, the monoclonal fluorescein
isothiocyanate-conjugated antibodies were added to the cell suspension, and the suspension

368
Figure 1. (a) Control; (b) and (c), effect of oxidized LDL (10 and 20 nM, respectively);
(d) oxidized Lp(a) (10 nM) on tube formation. 50,000 HMVE cells per weil were seeded
onto 300u1 ofMatrigel and incubated for 24 hours with the test substance. The cells were
co-cultured with V937 cells as described in the text.

was covered and placed on ice for 30 min. Thereafter, the cells were washed twice with
PBS and fixed with 1% formalin. The expression of surface antigens was determined in a
FACS Scan from Becton Dickinson.

2.4. Endothelial Cells.

HDMVE (foreskin) cells and HUVEC's were purchased from the Emory Skin Disease
Research Center, Emory Vniversity. They were provided frozen or as contluent cultures .
Cells as contluent cultures were harvested, pooled, centrifuged at 800 x g, and suspended in
medium 199 with 20% FCS, penicillin (50 ug/ml) and streptomycin (50 units/ml). Primary
cultures, were grown in gelatin coated (0.2% gelatin in PBS) for 30 min, tissue culture
flasks until contluence was reached (2-3 days). Cells were then detached with a 0.1%
trypsin in 0.02% EDTA solution and employed for further experimentation accordingly.
Cells were assessed in culture for morphologic, metabolic, and immunologic characteristics
considered in concert to indicate microvascular endothelial origin. Phase contrast
microscopy was used to determine "cobblestone" morphology on plastic and tube
formation when grown on Matrigel. Flow cytometry was used to examine the expression of
endothelial cell (EC) surface epitopes by binding monoclonal antibodies (MoAbs)
purchased from Harlan, Bioproducts for Science, Inc.. These were MoAb to EC marker
CD31, and MoAbs to adhesion molecules CD36, CD54 for ICAM-l. Surface antigen
expression on HMVEC and HVVEC was quantified by indirect immunotluorescence and
flow microcytotluorometry using a FACScan. To obtain EC single-cell suspensions,
contluent monolayers of HUVEC or HMVEC were treated with 50 V/mI dispace

369
Table 1. Comparison of tube formation on Matrigel using mock or U937 co-cultured
cells.

U937 Cell Treatment with Different Substrates

Cell PMA LDL ox-LDL Lp(a) ox-Lp(a)
Treatment Plastic 16 nM 10 nM 10 nM 10 nM 10 nM

HUVECs 285 * 76 95 87 110

HMVECs 266 110 88 95 98

HUVECs Co-incubated 510 100 287 108 320

with U9 37

HMVECs Co-incubated 525 86 275 112 260

with U937

* CapiUary tube length in 11m.

(Collaborative Biomedical Products), which was previously shown not to affect expression
of the antigens to be examined. Cells were incubated with primary mAbs, rinsed, then
incubated with an FITC-labeled antimouse IgG secondary Ab. Isotype controls were run in
parallel.

Ta ble 1 compares tube formation ofHUVECs on Matrigel in the presence and absence of
preactivated by oxidized lipoproteins U937 cells. HUVECs added to wells without
Matrigel (plastic) did not produce any tubes. When HUVECs were added to wells
containing Matrigel , they produced capillary tubes which were significantly increased when
the U937 cells pretreated with PMA were co-cultured with the ECs. The effect was the
same for both HMVECs and HUVECs. Pretreatment of U937 with oxidized LDL and
oxidized Lp(a) for 24 hours enhanced tube formation in both HMVECs and HUVECs .
Native, unmodified lipoproteins, when preincubated with U937 did not show an effect at
least for the time period tested (24 hours incubation). Oxidized LDL and oxidized Lp(a) as
activators ofU937 in equimolar concentrations, gave similar results (Table 1).

2.5. Ad hesi on to En dothelial Cells.

In order to study adhesion to endothelial cells, U937 cells were incubated for 24 hours with
native and oxidized LDL, native and oxidized Lp(a) as previously described (Frostegard et
al, 1990; Ragab et al, 1996) (Frostegard, Nilsson, Haegerstrand, Hamsten, Wigzell and
Gidlung, 1990; Ragab, et al., 1996).

2.6. Tu be Formation Assay

Matrigel, the complex basement membrane isolated from tumors (Collaborative Biomedical
Products), is an established system for the study of the cellular and molecular mechanisms
of in vitro angiogenesis. Matrigel, is suitable for a wide range of celI cultures and
biochemical applications. EC grown Matrigel (50 ul/crrr' ) migrate, align in parallel bundles
by making cytoplasmic extensions and finally form tubes after 24 hours. Time course
experiments demonstrated that tube formation was completed after 24 hours, and this time
point was therefore used for read out.
Tube formation was analyzed by phase contrast microscop y or the resulting tubes were
fixed and stained (Leucostain Kit, Fischer), and quantitation was done by measuring the
total tube area in 5 random areas of each weil. In addition, five random images from each
slide flask were scanned with a confocal laser scan microscope at the same setting for all

370
Table 2. Comparison of tube formation on Matrigel using lipid peroxidation products as
activators to U937 cells .

Cell Treatment with Different Substrates

4-Hydroxy ox- ox- Lysophospho-
Cell PMA MOA Hexanal nonenal LDL Lp(a) lipids
Treatment Plastic 16 nM 10 nM 10 nM 10 nM 10 nM 10 nM 10 nM

HUVECs 150 125 102 96 140 128 135

*
HUVECs Co- 440 230 136 112 255 224 242
incubated with
U937

* Capil lary tube length in um ,

'"C t4
~
<.: 12
d 10
1:1
l.o
~ 8
1:1
.0 6
....~ 4
::
~ 2
1:1
~ 0
0 5 10 15 20 30 40 50

Figure 2. Dose effect of oxidized LDL and oxidized Lp(a) on tube formation of
HMVECs.

images in each treatment group and transferred to a Maclntosh computer. Images were then
analyzed by counting fragments oftubes using the public domain , Nll-l Image program . The
program can be found at the Internet as http://rsb.info.nih.gov/nih-image. Data were
statistically analyzed by the Student's t test. For control purposes, EC cobblestone
mono layers grown on gelatin-coated flasks were cultured in the presence and absence of
blocking antibodies and observed by a phase contrast-phase microseope.

2.7. Monoclona l Ant ibody Anti-Mac-l.

We used the commercia lly available monoclonal antibody anti-Mac-l (Collaborative
Biomedical Products) originally inlended for flow cytometry and immunohistochemistry .
Endotoxin contamination of the antibod y preparat ion was below 0.01 lU/IOD ug (limulus
assay). In order to remove the azide the antibody was dialyzed overnight against a large
volume ofphosphate buffered saline. We also used CD54 (an ICAM -l blocking antibody)
available from Boehringer Ingelheim Pharmaceuticals . Antibody blocking was perforrned as
folIows; thirty minutes before the addition of stimulated U937 cells to the ECs, 50 ug/ml
anti-Mac-l were added to control and oxidized LDL treated U937 cells . Unbound ant ibody
was washed away before the stimulated U937 cells were added. The CD54 blocking
antibody was added to the EC culture media thirty minutes before the addition of the
stimulated U937 cells .

371
3. RESULTS

Culturing HMVECs and HUVEC on Matrigel resulted in the elongation and reorganization
of the cells into tube-like structures . In 24 hours, wells containing HMVECs seeded on
Matrigel showed a network of capillary like tubes (Figure 1), panels b, c and d; panel a
shows HMVECs plated on plain plastic. HUVEC's cultures gave similar results (not
shown). In pre1iminary experiments (not shown) it was determined that 50,000 cells seeded
on 300 uL ofMatrigel gave optimum results.
The effect was dose dependent. Figure 2 indicates that at 20 nM of oxidized LDL and
oxidized Lp(a) the effect was 35% higher than at 10 nM. At higher concentrations the
obtained results as measured by the tube area per field showed that oxidized lipoproteins
were less tubogenic and that the results were not as reproducible as results obtained at
lower concentrations.
The effect was also time dependent (results not shown) with maximum effects achieved at
24 hours preincubation. The effects of seleeted protein-free end products from the
modified lipids ofLDL, were tested as U937 activators along with PMA and oxidized LDL
and oxidized Lp(a) (Table 2).
Malondialdehyde (MDA), an arachidonate derived lipid oxidation product, an hexanal
product and 4-hydroxynonenal, oxidation products of linoleate, the major polyunsaturated
fatty acid (pUF A) in LDL were tested . All these oxidized substances were employed to
activate U937 cells for 24 hours prior to their co-culture with HUVECs . Table 2 shows
that in the absence ofU937 cells, all ofthe oxidized substances induced tube formation but
not as much as PMA, at least for the duration of the incubation in our experiments (24
hours). In the presence of activated U937, production of tubes by the HUVECs was
remarkably higher, in the order of PMA > oxidized LDL and Lp(a), MDA ,
lysophospholipids > hexanal and 4-hydroxy nonenal, in that order .
To investigate whether CDIIb/CDI8 was involved in ox-LDL-induced U937 adhesion, we
pretreated the culture media (Matrigel) with antibodies against epitopes of CDlI/CDI8 by
using moAbs toward Mac-I . These antibodies have been used extensive1y within the last
years in various pathophysiologic conditions (DiCorleto and de la Motte, 1985; Lehr,
Krober, Hubner, Vajkczy, and Menger, 1993). Phase-contrast micrographs of ECs grown
Matrigel in the presence of isotype control antibodies, mAb towards Mac-I (CD-lI b),
moAb towards IMAC-I and in the presence of both antibodies were tested . Figure 3
shows that whereas each individual antibody alone did not alter tube formation as
compared with control, the combination ofmAb towards CD-Ilb plus antibodies towards
ICAM-I (CD54) produced a fragmented network of cells and significantly inhibited the
tube formation ofboth HMVECs and HUVECs.

20
I:::
.~ 15
lU
"E
'"eeo 10

~
..
lU
5

0
Control Ab MAC-I Ab Combined Abs,
ICAM-I MAC.I, & ICAM-I

Figure 3. Inhibition of in vitra tube formation by endothelial cells by blocking antibodies

towards Mac-I and ICAM-l.

372
In contrast, cobblestone growth patterns on gelatin-coated dishes were not affected by the
presence ofthe antibodies. Using the image analysis system (see Materials and Methods),
numbers of tube fragments were counted in five randomly selected fields, and the mean
SEM was calculated and shown in Figure 3.

4. DISCUSSION

The ability of EC to form capillary tubes is a specialized function of this cell type and a
prerequisite for the formation of a continuous vessel lumen. Tube formation is thought to
depend on cell-cell adhesion moleeules but the specific proteins subserving this function
remain unknown . (Folkman and Shing, 1992; Polverini, 1996) have demonstrated the
involvement of monocytes and their products as mediators in the production of
angiogenesis. In the past several years, there has been dramatic progress in the elucidation
ofthe molecular basis of leucocyte adherence to endothelium, more than a dozen adhesion
molecules involved in leucocyte endothelial interaction have been immunologically
characterized and molecularly cloned. These investigations have increased our
understanding of the biology and c1inical relevance of leukocyte adherence to endothelium
in pathological conditions as weil as normal physiology . Most importantly, a new
approach to the treatment of a wide spectrum of clinical disorders may emerge from these
studies.
In the present study , by employing HMVECs and HUVECs for an in vitra capillary tube
formation in a process that we were able under in vitra conditions to show that activated
monocytic cells of the U937 line induce a rapid leukocyte-endothelium interaction.
Activation of the monocytic U937 line was accomplished only when oxidized LDL and
Lp(a) were used and not the unmodified lipoproteins. The pre-incubation period lasted
only 24 hours . It is of interest to note, that recently it was reported (Smaleley, Lin, Curtis,
Kobari, Stemerman, and Pritchard , 1996), that chronic exposure of EC to native LDL for at
least 4 days increased the adherence among the ECs by an ICAM-I-dependent mechanism .
However, and a1though unmodified LDL also increased E-selectin and VCAM-l
transcripts, such increases did not lead to statistically significant differences in
immunogenic-detectable proteins.
In the oxidized lipoproteins that we experimented with, the LDL constituents most
susceptible to oxidation are the PUFA of the LDL lipids (Steinberg, Parthasarathy, and
Carew, 1989). PUF A oxidation is complex and can produce a variety of substances,
including epoxides, alcohols, and fragmentation products such as aldehydes, alkanes , and
cyclic substances. The arachidonate oxidation end product MOA and other end products of
lipid oxidation generated within oxidized LDL are capable of diffusion from the LDL
particle. For this reason we studied the effects of these soluble substances known to be
formed within the oxidized LDL.
The experiments with the blocking antibodies clearly suggested that the activated
monocytic cells produce the leukocyte-endothelial cells interaction through a mechanism
involving the CDllb/CD18 receptor complex (DiCorleto and de la Motte, 1985; van der
wal, Das, Tigers , and Becker, 1992). Tube formation in vitra was only partially prevented
by adhesion blocking mAbs either to CD-Ilb or to ICAM-l. However, it was completely
inhibited, when mAbs towards Mac-I and ICAM-I, were used in conjunction . Since one of
the identified ligands for CDllb/CDI8, the intracellular adhesion molecule (ICAM), is
constitutively expressed on the endothelial surface (Lehr, et al., 1993), it is conceivable that
oxidized LDL may upregulate or activate CDllb/CD18 on the surface of circulating
leukocytes within a definite time course .
Very recently, (Matsumura, Wolff, and Petzelbauer, 1997) identified cadherin-5 and CD-31
as two molecules critical for angiogenesis. In addition , and using immuno-recipitation

373
methods, they showed that cadherin-5, CD-31, beta-catenin, and F actin shape a functional
complex that controls endothelial cell tube formation and coneluded that EC tube
formation depends on cadherin 5 and CD 31 interactions with filamentous actin .
The implication of these studies is that severe atheroselerotic changes are often
accompanied by a dense plexus of microvessels (neovascularization) extending from the
adventitia through the media and into the base of the plaque, where they may give rise to
small hemorrhages, regardless of the plaque surface integrity . Oxidized lipoproteins are
detected within atheromata and they exert several proatherogenic effects. Oxidized
lipoprotein-derived lipids and their degradative end products, like those we tested in the
present study can also be released from oxidized lipoprotein partieIes and generate
biological effects that have the potential to mediate cell recruitment. Theoretical concern is
that adhesion molecules , growth factors Iike vascular endothelial growth factor (VEGF)
(polverini, 1994) and cytokines may exacerbate plaque angiogenesis and thus may
adversely affect progression of coronary disease or plaque stability. In particular, VEGF' s
promotion of permeability could theoretically lead to plaque expansion or rupture, since
interplaque vessels tend to be quite permeable even without exogenous VEGF . In addition,
these growth factors might stimulate progression of coronary stenosis by stimulating
growth of ftbroblasts and medial smooth museIe cells in vitro. Since smooth museIe cells
produce both an inhibitor and a stimulator of endothelial cell growth, there is a change in
the balance of synthesis of these two molecules, with predominance of the stimulator in
advanced atherosc1erosis . It is not known to what extent angiogenesis plays a role at the
site of atherogenesis, but there is evidence that these mechanisms are relevant to the
development and the pathophysiology of atherosc1erotic plaque.

5. REFERENCES

DiCorieto, P.E., and de la Motte, C.A., 1985, Characterization of the adhesion of the
human monocytic cellline U937 to cultured endothelial cells, J Clin Invest 75 : 1153.

Folkman, 1., and Shing, Y., 1992, Angiogenesis, Mini review, J Biol Chem 267 : 10931.

Frostegard, 1., Nilsson, 1., Haegerstrand, A., Hamsten, A., Wigzell, R, and Gidlund, M .,
1990, Oxidized low-density lipoprotein induces differentiation and adhesion of human
monocytes and the monocytic celliine U937, Proc Natl Acad Sei USA 87: 904 .

Havel , R.J ., Eder, HA, and Bragdon, 1.R, 1955, The distribution and chemica1
composition ofultracentrifugally separated lipoproteins in human serum , J Clin Invest 34:
1345.

Henriksen, T., Mahoney, E., and Steinberg, D., 1981, Enhanced macrophage degradation of
low density lipoprotein previously incubated with cultured endothelial cells; recognition by
receptors for acetylated low density lipoproteins, Proc Natl Acad Sei USA 78: 6499.

Lehr, RA. , Kröber, M ., Hübner, C., Vajkoczy, P., and Menger, M.D ., et al., 1993,
Stimulation of leucocyte/endothelium interactions induced by oxidized low-density
lipoprotein in hairless mice : involvement of CD11b/CDI8 adhesion receptor complex , Lab
Invest 68 : 388 .

Matsumura, T., WolfT, K., and Petzelbauer, P., 1997, Endothelial cell tube formation
depends on cadherin 5 and CD31 interactions with filamentous actin, J Immunoll58 : 340 .

374
Polverini , P.l, 1996, Cellular adhesion molecules: newly identified mediators of
angiogenesis , Amer J Pathol 148: 1023.

Ragab, M.S., Selvaraj, P., and Sgoutas, D.S., 1996, Oxidized lipoprotein(a) induces cell
adhesion moleeule Mac-I (CD l 1b) and enhances adhesion of the monocytic celliine U937
to cultured endothelial cells, Atherosclerosis 123: 103.

Smalley, D.M., Lin, J. u-c., Curtis, M.L., Kobari, Y., Stemerman, M.B ., and Pritchard,
K.A Jr., 1996, Native LDL increases endothelial cell adhesiveness by inducing intercellular
adhesion molecule-1 , Arteriosclerosis, Thrombosis, and Yascular Biology, 16: 585.

Steinberg, D., Parthasarathy, S, and Carew, T.E., et al., 1989, Beyond cholesterol ,
modification of low-density lipoprotein that increases its atherogenecity, N Engl J Med
320: 915.

van der Wal, AC., Das, P.K., Tigers, Al, and Becker, A.E., 1992, Adhesion molecules in
the endothelium mononuclear cells in human atherosclerotic lesions, Amer J Pathol 141 :
1427.

375
A PEPTIDE FROM mE NCl DOMAIN OF TUE a3 CHAIN OF TYPE IV
COLLAGENPREVENTSDAMAGETOBASEMENTMEMBRANESBYPMN.

Zahra Ziaie". Jean-Claude Monboisses. Abdelilah Fawzi §, Georges Bellon§ ,

Jacques P. Borel § and Nicholas A. Kefalides *

*Connective Tissue Research Institute and Department ofMedicine, University

ofPennsylvania and University City Science Center, Philadelphia, Pennsylvani a
19104 USA and
§Laboratory ofB iochem istry, CNRS UPRESA, 6021, Universit y ofReims, F-
51095 , Reims , France

INTRODUCTION

During the process of acute inflammation, circulating polymorp honuclear leukocytes

(PMN) transmigrate from the vascular lumen to the site of infection or inju ry. Thi s
process involves the interaction of PMN with endothelial cells (EC) and subsequent
diapedesis ofPMN through the subendothelial basement membrane (BM) . A10ng the path
oftransmigration, PMN come across and interact with numerous macromolecules of BM .
Extensive research has been carried out to understand how various components of the BM
e.g. type IV Collagen (COL (IV», laminin, entactin, and proteoglycans mediate the
physiological function of PMN (Matzner, Bar-Ner, Yahalom, Ishai-Michaeli, Fuk s, and
Vlodavsky, 1985; Matzer, Vlodavsky, Michaeli , and Eldor, 1990; Pike, Wicha, Yoon,
Mayo, and Boxer, 1989; Senior, Hinek, Griffin , Pipoly, Crouch , and Mecham, 1989;
Senior, Gresham, Griffin , Brown , and Chung, 1992). Studies by Huber and Weiss (1989)
suggest that the transmigrating PMN cause a transient focal disruption of the BM which
allows PMN to traverse it. These disruptions are rapidl y repaired by the overlying EC . In
our laboratory, the focus has been on the role of COL (IV), a major component of the
BMs, on pMN function .

COL (IV) is a heterotrimer moleeule composed of three a chains . The predominant

molecular species is [al(IV)h a2(IV). The presence of additional chains a3(IV), a4(1V),
a5(IV) and a6(IV) has been demonstrated (Hudson , Reeders, and Trygg vason , 1993;
Ninomiya, Kagawa, Iyama, Naito, Kishiro, Seyer, Sugimoto, Ooahashi , and Sado, 1995).
Although the genes for the variou s COL (IV) chains have been cloned, the exact molecular
structure of the COL (IV) involving the new o-chains is not yet clear (Kamagata , Mattei ,
and Ninomi ya, 1992; Mari yama, Leinonen, Mochizuki , Tryggvason, and Reeders, 1994;
Zhou , Hertz, Leinonen, and Tryggvason , 1992; Zhou , Ding, and Reeders, 1994).

Angiogenesis: Models, Mod ulators . and Clinical Applications

Edited by Maragoudakis, Plenum Press . New York, 1998 377
PMN produce superoxide 02- and release proteolytic enzymes upon activation by various
agents such as PMA, tMLP and even collagen type I (Monboisse, Bellon, Randoux, Dufer,
and Borei, 1990). Previous In vitro studies from our laboratory have established that in
contrast to type I collagen, COL (IV) does not activate PMN but instead inhibits their
activation by PMA, fMLp and collagen type I (Monboisse, Bellon, Perreau, Gamotel, and
Borei, 1991). The inhibitory activity was localized in the non-collagenous domain (NCl
domain) of the a3 chain. Additional studies using synthetic peptides showed that the
inhibitory activity resides within a sequence comprising residues a3(IV) 185-203
(Monboisse, Garnotel, Bellon, Ohno, Perreau, Borei, and Kefalides, 1994.) The N-terminal
cycteine and the triplet -SNS- (residues 189-191) were absolute requirements for such
activity.

In this study, we used an in vitro model of avessei wall construct and examined the
integrity of subendothelial BM after PMN, treated with various peptides of COL (IV),
came in contact with it. The extent of the integrity of BM deposited over the collagen gel.
by EC was assessed by its resistance to the penetration of a colloidal pigment into the
collagen gel. The results suggest that PMN treated with the a3(IV) 185-203 peptide from
the NCI domain of COL (IV) have an impaired ability to damage the subendothelial BM .
The results also show that the triplet -SNS- (residues 198-191) of the a3(IV) peptide is
essential for the observed effect. However, this ability to prevent damage to BM by
nonactivated PMN was abolished when the experiment was carried out in the presence of
the activator tMLP . We have noted that the PMN ability to damage was similarly reduced
by an antibody to CD47 antigen which is present on both cell types and is involved in EC-
PMN interaction . On the other hand, sequential treatment of PMN with anti-CD47
antibody followed by the a3(IV) 185-203 peptide abolished the effect of either ligand.
This result is corroborated by unpublished observations from our laboratory indicating that
the CD47 antigen in association with the avß3 integrln serves as part of the receptor
complex for a3(IV) peptide.

2. MATERIALS AND METHODS

2.1. Cell Culture: Endothelial cells were isolated from human umbilical cord according to
Gimbrone, Cotran and Folkman (1974) with some modifications. The cells were grown in
tissue culture flasks coated with 1% gelatin and fed with modified M199 medium
containing fetal bovine serum (20%), L-g1utamine (2 mM), gentamiein (10 ug/ml),
amphotericin B (2.5 ug/ml) and supplemented with heparin (90 ug/ml) and endothelial cell
growth factor (30 Ilgl ml) (Ziaie, Friedman, and Kefalides, 1986).

2.2. PMN Preparation : Blood was obtained from healthy individuals and PMN were
isolated through a Ficoll-Hypaque gradient. The PMN layer was collected, diluted with
PBS and centrifuged two cycles to remove the gradient materials. The contaminating
erythrocytes were Iysed with a hypotonic solution (O.lx PBS). The cells were then washed
three more times and finally resuspended in the appropriate medium.

2.3. Cell Culture in the Vessel Wall Model: Double chamber tissue culture dishes
(Becton Dickenson, FrankIin Lakes, NJ.) were used to grow EC on the collagen gel.
Collagen gels were essentially prepared according to and Huber and Weiss(1989) . Collagen
type I stock solution was mixed with reconstitution buffer (0.05 M NaOH, 0.2 M Hepes,
and 0.26 M NaHC03), M199 medium and water to obtain a 2% collagen solution. NaOH
(15-30 111 of 1 N) was added to adjust the pH. All solutions were kept at 4° C until the
collagen solution was dispensed onto the inserts. The inserts (2.3 cm in diameter) with a
polyester membrane (3 11 pore) were overlaid with 0.8 ml of ice-cold collagen solution and

378
 Table 1. Amino acid sequence of synthetic peptides of the NCl domain of the a
chains of human type IV collagen.

Peptides 185 203

al(IV) 185-203 C N Y Y A N A y S F W L A T I E R S E
a2(IV) 185-203 C H Y Y A N K Y S F WL T T I P E Q S
a3(IV) 185-203 C N y y S N S y S F W L A S L N P E R
a3(IV) 185-191 C N Y Y S N S - - - - - - - - - - - -
a3(IV) 194-203 - - - - - - - - - F WL A S L N P E R

___ ~. ::~liW Cells

~. Upper Weil

............................................... . Collagen Gel

Lover Weil

Figure 1. Schematic diagram of vessel wall construct (for details see Materials and
Methods).

incubated at 3T C for 2 hours or overnight. Endothelial cells (5x104) in 2 ml growth

medium were then seeded on the collagen gel with 3 ml modified M 199 medium in the
lower chamber. This three-dirnensional construct (Figure 1) was incubated at 37 °C for 18-
21 days . The cultures were fed with growth medium until a confluent EC monolayer was
formed over the gel. Subsequently, the amount of growth factors(heparin and ECGF) was
lowered to 50% or 25% ofthe original growth medium .

2.4. Preparation of Synthetic Peptides: Peptides corresponding to specific sequences of

the NC 1 domain of a chains of COL (IV) (Table 1) were synthesized according to the
method of Barany and Merrifield (1980) by the Protein Chemistry Laboratory of the
University of Pennsylvania, Philadelphia, PA (Kefalides, Ohno, Wilson, Fillet, Zabriski ,
and Rosenbloom, 1993).

2.5. PMN Preincubation with Peptides and/or Antibody: PMN (5xI06/ml) were
incubated with peptides (50 ug/ml M199 medium unless otherwise specified) for 30 min at
3T C. For antibody treatment, PMN were either incubated with the antibody alone for 1
hr at 3TC or incubated with the antibody for 30 min and then peptide was added and the
incubation was continued for an additional 30 min. After removal of the peptide and/or
antibody solution by centrifugation, PMN were resuspended in EC growth medium at
5x105/ml.

2.6. PMN Incubation in the Double Chamber Tissue Culture: 106 PMN,
preincubated as described in section 2.5., in 2 ml EC medium were added to the upper
chamber with or without EC monolayer. The lower chamber contained 3 ml modified
M199 medium with or without chemoattractant tMLP(10-7M) . The culture was incubated
overnight in a humidified C02-incubator at 31" C.

379
2.7. Preparation of Denuded Basement Membranes: Following incubation of PMN
with the EC monoalyer, the subendothelium BMs, deposited by EC over a three week
period on the collagen gel, were denuded according to Huber and Weiss (1989). After
removal ofPMN, the monolayer was washed with PBS and then incubated at 37°C for 10
min in hypotonic buffer (10 mM Tris-HCI, 0.1% BSA, 0.1 mM CaCIz, 1 mM NEM, ImM
PMSF , and 5 mM EDTA, pH 7.5). EC were then Iysed with I ml of 0.5% NP-40 in
hypotonic buffer for 2 min followed by one PBS wash. The residual debris were removed
by treatment with 2 ml of 0.2% deoxycholate in hypotonic buffer for 2-5 min and a final-
wash with PBS .

2.8. Colloidal Permeability of the DM: The integrity of the BM was assessed by its
ability to retain a colloidal pigment, i.e., monastral blue B, and prevents its penetration into
the underlying collagen gel. 3 ml of a suspension of monastral blue B, (0.03% in HBSS
containing 1 mM NEM , 1 mM PMSF and 1 mM EDTA) was added to the denuded
membrane in the upper chamber. The medium in the lower chamber was replaced with 2 ml
of an identical buffer without the pigment. Following an ovemight incubation at 31' C, the
culture was washed with PBS several times to remove the pigment particles which did not
penetrate the collagen gel. The gel was then digested in bacterial collagenase (Sigma
Chemical Company, Saint Louis, MO) (100 Ilg/ml in 50 mM Tris-HCI, 4 mM CaCIz, pH
8.0) and absorbenc y was measured at 605 nm. Absorbency values correspond to the
pigment concentration in the gel. Therefore, a higher level of pigment accumulation in the
gel indicates that the integrity of the basement membrane is damaged by transmigrating
PMN .

RESULTS AND DISCUSSION

3.1. Effect of a.3(IV) 185-203 Peptide on PMN-Generated Damage to DM.

The a3(1V) 185-203 peptide inhibits PMN activation as measured by superoxide

production. We tested the hypothesis whether the same peptide affects PMN ability to
damage BMs as they traverse it. PMN (5X106/ml) were incubated for 1 hr at 31'C in the
presence or absence ofthe a3(IV) 185-203 peptide (25 ug/ml). The peptide solution was
removed and PMN were resuspended in EC growth medium and added (106
PMN/chamber) to the EC monolayer grown in a double chamber system . Following an
ovemight incubation at 31'C, the PMN were washed out, the BM was denuded and
incubated with colloidal pigment and finally after digestion of the gel with collagenase,
absorbency was measured . The results are shown in Figure 2. As is evident, the pigment
penetrates the gel in the absence of the EC monolayer (Figure 2, bar 1). However, when
EC are allowed to grow on the gel, they deposit a continuous BM which resists the passage
of pigment into the gel (Figure 2, bar 2). Incubation of PMN with the EC monolayer
results in the interaction between the two cells and the subsequent contact of PMN with
BM allowing the colloidal pigment to cross through the BM barrier and enter the gel
(Figure 2, bar 3). The presence of the chemoattractant agent fM.LP in the lower chamber
generates a gradient and causes more PMN to move across the membrane and results in
greater damage (Figure 2, bar 4). When PMN are treated with a3(IV) 185-203 peptide, a
significant drop in the pigment accumulation in the gel is observed (Figure 2, bar 5). This
suggests that the peptide has impaired the capacity of PMN to damage the BM barrier.
However, when fM.LP is present, PMN overcome the inhibitory activity of a3(IV) 185-
203 peptide (Figure 2, bar 6). Although the mechanism and the pathways leading to this
observation are not clear at present, we speculate that the inhibitory activity of the peptide

380
 1.00 -,--- - - - - - - - - - - - ----,;----,

0.80
E
=
In
e
0.60
\C

OAO -

I
0.20

i I
~
5 I>
Bar Number + +
GE L +
EC - + + + + +
I'MN - - + + + +
r;\ILI' - - - + - +
a J (l V) 185 · 20J - + +

Figure 2. Effect of a,3(IV) peptide-treated PMN on the integrity of subendothelial BM .

EC monolayers were grown in the double chamber system. PMN incubated in the
presence or absence of a.3(IV) 185-203 peptide (25 ug/rnl) fOT 1 hr at 37°C were added to
the upper chamber. The experiment was carried out as described under Materials and
Methods. Absorbency values correlate with the amount of pigment passed through the
BM and entered the gel. The values represent mean ± SD of duplicate sampies . Plus (+)
or minus (-) signs represent the presence or absence of the features listed on the left,
respectively.

and the stimulatory activity of tMLP might depend on separate and independent
pathways.

3.2. Specificity of Peptide Inhibitory Effect.

We had established that the inhibitory activity of NCI domain of COL (IV) on PMN
activation, as measured by superoxide production, was due to a,3(IV) 185-203 peptide
(Monboisse et al., 1994). A comparable region of a,1(IV) and a.2(IV) chains did not inhibit
PMN activation . We, therefore, compared the effect of a,1(IV), a,2(IV) peptides with that
of a,3(IV) peptide on PMN in the vessel wall construct and assessed the degree of damage
to the BM. The results are shown in Figure 3. The presence of EC monolayer poses a
barrier to the pigment penetration into the gel (compare bars 1 and 2 in Figure 3).
Treatment ofPMN with a,1(IV) and a.2(IV) peptides (Figure 3, bars 5 and 6 and bars 7 and
8 respectively) regardless ofthe presence or absence offMLP in the lower chamber results
in a comparable degree of damage to the BM as that generated by non-treated PMN (Figure
5, bars 3 and 4). In contrast, a,3(IV)-treated, non-tMLP treated PMN cause the least
damage (Figure 3, bar 9) to the membrane integrity.

3.3. Sequence Specificity of a.3(IV) 185-203 Peptide.

Previous observat ions from our laboratories (Monboisse et al., 1994) have shown that the
N-terminal cysteine and the triplet -SNS- are essential for the inhibitory action of a,3(IV)

381
 1.0

0.8
E

e
=
!Tl 0.6
~

@J
tIl
0.4
.:::;;
~
0.2

0.0
Bar Nu rnbe r 6 7 II 9
GE L + + + + + + + +
EC + + + + + + + +
PMI' + + + + + + +
r;\ I L P + + +
u 1( I V) 185 · 203 + +
a 2(I V) 185 ·203 + +
u 3( I V) 185 · 203 + +

Figure 3. Comparison ofthe effect ofvarious peptides ofthe NCI domain of a chains of
type IV collagen on PMN-generated damage to the subendothelial BM. Conditions are as
described in Figure I except that PMN were incubated with peptides at 50 ug/ml for 30
min at 3 TC . The values represents mean ± SD of duplicate sampies. Plus (+) or minus (-)
signs represent the presence or absence ofthe features on the left, respectively .

185-203 peptide on PMN activation. Therefore, the integrity of BM was tested in the
double chamber system using two synthetic peptides of a3(IV); one comprising residues
185- I91 and contains the cysteine residue and the triplet -SNS-, and another comprising
residues 194-203 which lacks both ofthese features. The results are summarized in Figure
4. The least damage to the membrane was observed when PMN were treated with a3 (IV)
185-203 (Figure 4, bar 4). Treatment with the shorter peptide, a3(IV) 185-19 I (Figure 4,
bars 6 and 7), was also as effective. Treatment with the C-terminal portion of the peptide,
a3(IV) 194-203, not only was it not effective in inhibiting PMN damage to the BM barrier,
it resulted in more damage than non-treated PMN (Figure 4, compare bars 2 and 3 with
bars 8 and 9).

3.4. EfTect of Anti-CD47 Antibody on the Inhibitory Function of the a3(IV) Peptide
on PMN.

We speculated that the a3(IV) peptide inhibitory activity on PMN is mediated through a
receptor. Unpublished observations from OUf laboratory suggest that the CD47 antigen in
association with the a Vß3 integrin forms part of a receptor complex for the a3(IV) 185-203
peptide . Cooper, Lindberg, Gamble, Brown and Vadas (1995) showed that in addition to
other molecules, integrin-associated protein (CD47 antigen), which is present on both EC
and PMN , is essential for PMN migration through endothelium and into the site of
inflammation. Since the a3(IV) 185-203 peptide inhibits migration of PMN through the
BM barrier, we hypothesized that the peptide binds to its putative receptor complex
(CD47 antigen and aVß3 integrin) and, therefore, inhibits interaction between the two cell
types . 1fthis is the case, then an anti-CD47 antibody might function in a similar manner as
the a3(IV) peptide in the double chamber system . Figure 5 shows the results of such an

382
 1.5 11

---
11")
1.2 0

U.9 0
e
"=
I::
0 .611
~
-e O.JO

0.00

lI:lr :-.'umber 6 9
<a:t + + + + + + + + +
E' + + + + + + + +
" 1\1:" + + + + + + + +
f:\II .1' + + + +
0 3( I V) 11I5- 203 + +
0 3( I V) 18 5· 19 1 + +
0 3 (1 V ) 19 01 . 203 + +

Figure 4. Sequence requirement for inhibitory activit y of u3(IV) peptide on

transmigration of PMN . PMN preincubated with u3(IV) 185-203 peptide or its partial
sequences, a3(IV) 185-191 and u3(IV) 194-203 at 50 ug/ml for 30 min at 3TC. Peptides
were removed and PMN were then added to the EC monolayer in the double chamber
culture. The experiment was carried out as described under Materials and Methods. The
values are mean ± SD of duplicate sampies . Plus (+) or minus (-) signs represent the
presence er absence of the features on the left, respectively .

experiment, where PMN were incubated with anti CD47 monoclonal antibody (1 :500
dilution; MAB 1796, Chemicon International Inc., Temecula, CA) for 1 hr or alternati vely
incubated with antibody for 30 min followed by another 30 min incubation in the presence
of the a3(IV) 185-203 peptide (50 ug/ml) or with peptide alone for 30 min. PMN were
then incubated with the EC monolayer in the double chamber system and the extent of
damage to the BM was assessed by pigment accumulation into the collagen gel. The results
are shown in Figure 5. Treatment of PMN with the a3(IV) peptide or anti-CD47 antibody
reduces the damage to the BM (Figure 5, bars 5 and 7). Howeve r, when PMN are first
incubated with anti CD47 antigen followed by incubation with the pepti de, the inhibitory
activity of the peptide is abolished and a higher degree of damage to the membrane is
observed (Figure 5, bar 9). In contrast , incubation of PMN with an irrelevant antibody
(med 13D 12H), does not affect the inhibitory activity of the u3(IV) peptide on PMN
induced damage to BM (Figure 5, bar 11). When tMLP is present in the lower chamber,
PMN induced damage and presumably the transmigration of PMN are unaffected by the
presence of u3(IV) peptide or anti-CD47 antibody (Figure 5, bars 6, 8 and 10).

SUMMARY

Previous studies from our laboratories have demonstrated that a specific amino acid
sequence(185-203) of the NC 1 domain of the u3 chain of COL (IV) inhibits activation of
PMN by ligands such as fMLP , PMA, or collagen type I. In the present study we
examined the rote of this peptide on the ability of PMN to damage BM in a vessel wall

383
 1.20

0.90
T
T
IT
E
-=
lO
\C 0.60

:IJ
.:::.
~ 0.30

0.00
Ilar ;';umbcr .\ ~ 5 6 7 8 9 10 11
GEL + + + + + + + + + + +
EC + + + + + + + + + +
l':\ tN + + + + + + + + +
rxu.r: + + + +
a 3( I V) 185·203 + + + +
Anli·CD47MAII + + + +
J\tcd 130 1211 +

Figure 5. Effect of a3(IV) peptide on PMN-generated damage of BM in the presence or

absence ofanti-CD47 antibody. PMN were incubated in the presence or absence of anti-
CD47 antibody for 1 hr or aiternative1y incubated with the antibody for 30 min followed
by the addition of a3(IV) 185-203 peptide (50 ug/ml) and a further 30 min incubation at
3TC. The antibody and peptide were removed by centrifugation and PMN were added to
the double chamber culture system . The experiment was carried out as described under
Materials and Methods . The values represent mean ± SD of duplicate sampies . Plus (+)
or minus (-) signs represent the presence or absence of the features on the left,
respectively.

model using a three dimensional collagen gel overiaid with endothelial cells. Our results
show that the presence of the EC monolayer decreases damage to the BM by unstimulated
PMN . The damage to BM is increased if PMN are activated by tMLP . Treatment of
PMN with the a3(IV) 185-203 peptide reduces the capacity of unstimulated PMN to
damage the BM . However, when tMLP is present, PMN become activated and overcome
the inhibitory effect of the a 3(IV) 185-203 peptide. That this inhibitory activity is unique
to the a3 chain was confirmed by using peptides from the same region of the NC 1 domain
of the al(IV) and a2(IV) chains. The al(IV) and a2(IV) peptides did not prevent the
damage to the BM by either unactivated- or tMLP activated-PMN. As with our previous
studies showing that the triplet -SNS- (residues 189-191) is necessary for the inhibitory
activity, shorter peptides , one containing the -SNS- triplet [a3(IV) 185-19 I] and another
lacking it [a3(IV)194-203], were tested. Treatment ofPMN with the peptide a3(IV) 185-
191 was effective in reducing the damage to the BM whereas treatment of PMN with the
peptide a3(IV) 194-203 was not effective. Unpublished observation from our laboratories
suggest that the a3(IV) 185-203 peptide bonds to the CD47-avß3 complex on PMN .
Reports from other laboratories show that CD47 antigen is present in both EC and PMN
and is essential for PMN migration across the EC into the site of inflammation. A
monoclonal antibody against CD47 antigen was capable of reducing the damaging effect of
PMN on BM via a manner similar to that observed with the a3(IV) 185-203 peptide.
However, incubation of PMN with the anti-CD47 antibody followed by incubation with
the peptide abolished the protective effect ofthe peptide on BM damage.

384
We conc1ude, therefore, that in vivo BMs and the EC monolayer provide a barrier to
circulating PMN and that any such cells which cross the vascular EC monolayer provide
minimal or no damage to the underlying BM. However, during inflammation, when a
variety of cytokines and bacterial products are liberated , PMN become activated, releasing
proteolytic enzymes which damage the BM, a1lowing passage of PMN into the
interstitium .

ACKNOWLEDGMENTS

The authors would like to thank Jeffrey Edwards for excellent technical assistance in
preparation of PMN and endothelial cells and the double-chamber cultures and Marilyn
Videtto for typing the manuscript. This work was supported by the grants from the
National Institutes of Health, AR-20553, HL-29492 and AR-07490, by a NATO
Collaborative Research Grant and by grants from the CNRS UPRESA 6021 to the
University ofReims (contract DRED).

REFERENCES

Barany, G., Merrifield, R. B. , 1980, Solid-phase peptide synthesis, in: The peptides, (E.
Gross, and 1. Meinhofer, eds.) p. I, Academic Press, Inc ., New York.

Cooper, D ., Lindberg, F. P., Gamble, 1. R ., Brown, E. 1., and Vadas, M . A ., 1995 ,

Transendothelial migration of neutrophils involves integrin-associated protein (CD47),
Proc . Nat!. Aead. Sei. U.S.A. 92:3978-3982 .

Gimbrone, M . A., Jr., Cotran, R. S., and Folkman, 1., 1974, Human vascular endothelial
cells: Growth and DNA synthesis, 1. Cell Bio!. 60: 673-684

Huber, R ., and Weiss S. J., 1989, Disruption of the sunendothelial basement membrane
during neutrophil diapedesis in an in vitro construct of a blood vessel wall, 1. Clin. Invest.,
83 :1122-1136.

Hudson, B. G., Reeders, S. T., and Tryggvason, K., 1993, Type IV collagen : Structure,
gene organization, and role in human diseases, 1. Bio!. Chem. 268:26033-26036.

Kamagata, Y., Mattei, M .~G., and Ninomiya, Y., 1992, Isolation and sequencing of cDNAs
and genomic DNAs encoding the a4 chain of basement membrane collagen type IV and
assignment of the gene to the distal long arm of human chromosome 2, 1. Bio!. Chem.
267:23753-23758.

Kefalides, NA, Ohno, N., Wilson, C. B., Fillet, M. , Zabriski, J . and Rosenbloom, J., 1993 .
Identification of antigenic epitopes in type IV collagen by use of synthetic peptides,
Kidney Int., 43 :94-100.

Mariyama, M ., Leinonen, A., Mochizuki, T., Tryggvason, K., and Reeders, S. T., 1994,
Complete primary structure of the human a3(IV) collagen chain . Co-expression of the
a3(IV) and a4(IV) collagen chains in human tissues, J. Bio!. Chem. 269:23013-23017.

Matzner, Y., Bar-Ner, M ., Yahalom, 1., Ishai-Michaeli, R ., Fuks, Z ., and Vlodavsky, 1.,
1985, Degradation of haparan sulfate in the subendothelial extracellular matrix by a readily
released heparanase from human neutrophils, 1. Clin. Invest, 76: 1306-1313 .

385
Matzner, Y., Vlodavsky, I, Michaeli, R, and Eldor, A., 1990, Selective inhibition of
neutrophil activation by the subendothelial extracellular matrix : Possible role in protection
ofthe vessel wall during diapedesis, Exp. Cell Res. 189:233-240.

Monboisse, 1. C., Bellon, G., Randoux, A., Dufer, 1., and Borei, J. P ., 1990, Activation of
human neutrophils by type I collagen, Biochem. 1. 270:459-462.

Monboisse, 1. C ., Bellon, G., Perreau, C., Garnotel, R ., and Borei , 1. P ., 1991, Bovine lens
capsule basement membrane collagen exerts a negative priming on polymorphonuclear
neutrophils, FEBS u« 294: 129-132.

Monboisse, 1. c., Garnotel, R., Bellon, G., Ohno, N., Perreau, C ., Borei, 1. P ., and
Kefalides, N . A., 1994, The a3 chain of type IV collagen prevents activation of human
polymorphonuclear leukocytes, 1. Biol. Chem. 269:25475-25482.

Ninomiya, Y., Kagawa, M ., Iyama, K., Naito, 1., Kishiro, Y., Seyer, 1. M ., Sugimoto, M .,
Ooahashi, T ., and Sado, Y., 1995, Differential expression of two basement membrane
collagen genes, COL4A6 and COL4A5 , demonstrated by immunofluorescence staining
using peptide-specific monoclonal antibodies, 1. Cell Bio!' 130: 1219-1229.

Pike, M. C ., Wicha, M. S., Yoon, P., Mayo, L., and Boxer, L. A., 1989, Laminin promotes
the oxidative burst in human neutrophils via increased chemoattractant receptor expression,
1. Immunol. 142:2004-2011.

Senior, R. M ., Hinek, A., Griffin, G. L., Pipoly, D. 1., Crouch, E . C., and Mecham, R. P.,
1989 , Neutrophils show chemotaxis to type IV collagen and its 7S domain and contain a 67
kD type IV collagen binding protein with lectin properties, Am. 1. Respir. Cell Mol . Biol.
1:479-487.

Senior, R . M ., Gresham, H. D., Griffin, G. L., Brown, E. J., and Chung, A. E., 1992,
Entactin stimulates neutrophil adhesion and chemotaxis through interactions between its
Arg-Gly-Asp (RGD) domain and the leukocyte response integrin, 1. Clin. Invest. 90:2251-
2257 .

Zhou, 1., Hertz, 1. M ., Leinonen, A., and Tryggvason, K., 1992, Complete amino acid
sequence ofthe human a5(IV) collagen chain and identification of a single-base mutation in
exon 23 converting glycine521 in the collagenous domain to cysteine in an Alport
syndrome patient, 1. Biol. Chem. 267: 12475-12481.

Zhou, 1., Ding, M. , and Reeders, S., 1994, Complete primary structure of the sixth chain of
human basement membrane collagen, a6(IV),1. Bio!. Chem. 269:13193-13199.

Ziaie, Z , Friedman, H. M ., Kefalides, N . A., 1986, Suppression ofmatrix protein synthesis

by herpes simplex virus type 1 in human endothelial cells, Collagen Rel. Res . 6:333-350 .

386
Human Pathology and
Clinical Developments
TUMORAL VASCULARTIY: WHAT DOES IT TELL US ABOUT THE GROWTH
AND SPREAD OF CANCER?

Noel Weidner, MD

Department of Pathology;
University of Califomia, San Francisco ;
San Francisco , CA 94143-0102

A. TUMOR GROWTH IS ANGIOGENESIS DEPENDENT

Without blood vessels, tumor cells continue to grow until passive diffusion no
longer allows enough nutrients to enter or metabolic waste products to exit (1-17) .
Furthermore, intratumoral endothelial cells prolife rate faster than those in the adjacent
benign stroma (45-fold faster in breast carcinoma and 30-fold in prostate carcinoma)
(18, 19), and the rate of tumor progression increases with increased intratumoral vascularity
(19-92). Also , different techniques to specifically inhibit angiogenesis (i.e., not cytostatic
to tumor cells in vitro) c1early inhibit tumor growth in vivo (93-104). For example, an
analog of fumagillin (a.k.a. AGM-1470 or TNP-470) inhibits endothelial proliferation in
vitro and tumor-induced angiogenesis in vivo (95) TNP-470 and other angiogenesis
inhibitors are now in various phases of clinical trials as therapeutic agents for a variety of
malignant solid tumors, leukemias , and infantile hemangiomas (1,93,94) . Moreover, Kim et
al. (97) found that inhibition of vascular endothelial growth factor (VEGF)-induced
angiogenesis suppressed tumor growth in vivo . This group injected human malignant cell
lines into nude mice followed by treatment with a monoclonal antibody specific for VEGF .
The antibody inhibited the tumor growth and reduced tumor vessel density , but had no
effect on the growth rate of the tumor cells in vitro. Millauer et al. (99) noted marked
suppression of tumor growth with the introduction of defective VEGF receptors into
tumor endothelial cells . He also noted that a single intravascular injection of antagonists of
the v 3 integrin disrupted ongoing angiogenesis in the chick chorioallantoic membrane .
This resulted in the rapid regression of human tumors transplanted into the chorioallantoic
membrane (100) . Recent and quite compelling evidence that tumor growth and metastases
are angiogenesis-dependent is that the recently discovered substance angiostatin, which
specifically inhibits proliferation of vascular endothelial cells, causes metastases to remain
small and dormant « 0.2 mm) (101). Also, angiostatin significantly inhibits growth of
animal and human tumors (102).

Angiogenesis: M odels, Mod ulators, and Clinical App lications

Edited by Maragoudakis, Plenum Press, Ncw York, 1998 389
B. HOW TUMOR GROWTH IS STIMULATED BY ANGIOGENESIS

New tumor vessels allow exchange of nutrients, oxygen, and waste products by a
crowded cell population for which simple diffusion is inadequate. Also, activated
intratumoral endothelial cells release important paracrine growth factors for tumor cells
(e.g., basic fibroblast growth factor [bFGF], insulin growth factors , platelet derived growth
factor, and colony stimulating factors)(103 ,105-107) . Furthermore, activated endothelial
cells located at the tips of growing capillaries secrete collagenases, urokinases, and
plasminogen activator (108-111), which allow capillary ingrowth and the spread of tumor
cells into and through the adjacent fibrin-gel matrix, connective tissue stroma, and into the
lymphatic and vascular spaces . In breast carcinoma patients, increased intratumoral levels
ofurokinase-type plasminogen activator (uPA) and plasminogen activator inhibitor-l (PA-
1) are reported to be independent predictors of poor prognosis and Fox et al. and
Heldenbrand et al. (109, 110) have shown a significant association of uPA and PA-l with
intratumoral microvessel density. Thus, the additive impact of the perfusion and paracrine
tumor effects, plus the endothelial-cell derived invasion-associated enzymes, all likely
contribute to a phase of rapid tumor growth and signal a switch to a potentially lethal
angiogenic phenotype.

C. MEDIATORS OF TUMOR ANGIOGENESIS

Tumor neovascularization is similar to neovascularization in wound healing (112)

and it is likely mediated by similar and specific angiogenic molecules. These mediators are
released by the tumor cells andlor host immune cells (i.e., macrophages) or are possibly
mobilized from substances within the tumor stroma (1,6,113,114) . We reported that some
breast ducts containing duct carcinoma in situ (DCIS) have a cutT or ring of microvessels
around the duct. The close proximity of this cutT of neovascularization to the DCIS cells
suggested that it formed in response to angiogenic factor(s) released by the DCIS cells. The
cuff was limited to ducts containing DCIS and did not correlate with periductal
inflammation, suggesting that the angiogenic stimulus was a ditTusible factor coming from
the DCIS cells and not from inflammatory cells. Guidi et al. (115) subsequently reported
similar observations in 38% of DCIS cases . Likewise, Smith-McCune et al. (116) found
that the number of microvessels was greater immediately beneath foci of cervical
intraepithelial neoplasia (CIN) when compared to adjacent normal epithelium. Yet, the
neovascularization was not related to the number of macrophages within the CIN lesions,
indicating that the production of the angiogenic factor(s) is likely a property of the
dysplastic epithelial cells themselves (116).
The factor(s) andlor cell(s) causing tumor angiogenesis have not been determined,
but, current leading candidates include bFGF(l17, 118) and vascular permeability factor
Ivascular endothelial growth factor (VPFNEGF)(119). VPF/VEGF is a potent vascular
permeability factor , a selective endothelial-cell growth factor, and likely an important
tumor angiogenic factor (120,121) . Moreover, VPFNEGF has been shown to cause
endothelial cells to express plasminogen activator, plasminogen activator inhibitor,
interstiti al collagenase, and procoagulant activity (119,122) . By increasing vascular
permeability, VPFNEGF prornotes extravasation of plasma fibrinogen, leading to fibrin
deposition within the tumor matrix. This matrix or scatTold prornotes the ingrowth of
macrophages, fibroblasts, and endothelial cells (119,122) . It is possible that molecules
derived from the fibrin gel matrix have important angiogenic properties. Thompson et al.

390
(123) found that fibrin degradation products have angiogenic potential. Furthermore, the
work of Kim et al. (97), Millauer et al. (99), and Warren et al. (104) strongly support
VPFNEGF as an important tumor angiogenic factor. Toi et al. (124) found that VPF/VEGF
expression was higher in breast tumors with higher intratumoral microvessel densities and
that relapse-free survival was shorter in patients with tumors showing relatively high
VPFNEGF expression. But, VPFNEGF may not act alone, and Goto et al. (125) have
shown that VPFNEGF and bFGF can act synergistically to cause angiogenesis. Lately,
platelet-derived endothelial growth factor (a.k.a. thymidine phosphorylase) has gained
more recognition as a potentially important angiogenic factor and its expression has been
associated with an adverse outcome (126). Finally, various low molecular weight, non-
peptide angiogenic factors are known (e.g., nitric oxide and 12(R)-hydroxy-eicosatrienoic
acid) (6,113 ,127,128). Their significance is unclear at present, yet they might be
important.
Although probably necessary, angiogenic stimulators may not be sufficient for
tumor angiogenesis. Bouck et al. (129,130) reported that tumor angiogenesis results from a
net balance between positive and negative regulators ofblood vessel growth . Zajchowski et
al. (131) reported that somatic hybrid cells (i.e., MCF -7 human breast-carcinoma cells
fused with normal human mammary epithelial cells) did not form tumors in nude mice. The
hybrids showed increased expression of thrombospondin, an angiogenesis inhibitor,
suggesting that angiogenic capability contributes to tumorigenicity in human breast
carcinoma. Also, Dameron et al. (137) showed that the switch to the angiogenic phenotype
by fibroblasts cultured from Li-Fraumeni patients coincided with loss of the wild-type
allele of the p53 tumor suppressor gene and was the result of reduced expression of
thrombospondin-l. Finally, O'Reilly et al. (101,102) reported that a novel angiogenesis
inhibitor, angiostatin, is produced by the primary tumor mass of a Lewis lung carcinoma. In
this model, when the primary tumor is present, metastatic tumor growth is suppressed by
angiostatin. After primary tumor removal, the metastases neovascularize and grow . This
mechanism may explain one form of dormancy, but some metastatic deposits appear to
remain dormant in spite of the fact that the primary tumor had been previously removed
(133) . It is probable that the "angiogenic/anti-angiogenic cocktail" produced by various
tumors will vary between tumor types and, temporally, within the same tumor, becoming
more complex as tumors become more genetically unstable and progress to higher grade,
more angiogenic lesions.

D. ANGIOGENESIS IS NECESSARY FOR METASTASES TO DEVELOP

To spread to distant body sites, a tumor cell must overcome aseries of physical
barriers and biochemical deficiencies. Likely, less than one cell in -100,000 or 1,000,000
has a1l of the properties necessary to develop a viable metastasis (134-136). Tumor cells
must induce angiogenesis to gain access to the vasculature from the primary tumor, survive
the circulation , escape immune surveillance, localize in the microvasculature of the target
organ, escape from or grow from within the vasculature into the target organ, and induce
tumor angiogenesis (134,135) . If the primary tumor is highly angiogenic, then its daughter
metastases (i.e., clones) are very likely to be highly angiogenic. Growth and spread are
amplified when the newly established angiogenic metastases shed additional angiogenic
tumor cells to form even more metastases by following the same cascade of events (135).
Angiogenesis is necessary early in malignant transformation, because without it,
tumor cells are only rarely shed into the circulation (137-139). Liotta et al. (I 38, 139), using

391
 a transplantable mouse fibrosarcoma model, showed that the number of tumor cells shed
into the bloodstream correlated with intratumoral microvessel density, increased tumor
volume, and the number of established lung metastases (6,138,139). McCulloch et al. (140)
studied vascular density and cell shedding in women having surgery for breast carcinoma
and found that the extent of tumor cell shedding correlated with intratumoral microvessel
density . Obviously , these observations suggest that intratumoral microvessel density might
correlate with metastases and other measures of aggressive tumor behavior, a relationship
subsequently shown to be true (19-92). New, proliferating capillaries have fragmented
basement membranes and are leaky, making them more accessible to tumor cells than
mature vessels (141) . Furthermore, the invasive chemotactic behavior of endothelial cells at
the tips of growing capillaries is facilitated by the secretion of collagenases, urokinases , and
plasminogen activator (111). These degradative enzymes may facilitate the escape of tumor
cells into the tumor neovasculature. Indeed, the "invading" capillaries may actively
partieipate in the metastatic process by engulfing, and thus, facilitating the entry of tumor
cells into vascular spaces, allowing systemic spread. Indeed, Sugino et al. (142) observed in
a naturally occurring mouse mammary carcinoma model (C3H/He mice) that intravasating
tumor cells and tumor emboli within blood vessel lumina retained their nested architecture
within a continuous basement membrane and were also invested by an endothelial-cell
layer. These investigators believed the findings indicated a "passive" mechanism of tumor
cell intravasation distinct from the invasive properties of tumor cells wherein endothelial
cells in sinusoidal vessels can envelop tumor cell nests, which then become detached into
the blood . Arrest of such encapsulated emboli in pulmonary arterioies downstream could
form new metastatic tumor foei.
Also, supporting the role of angiogenesis in the metastatic process is the
observation that India ink injected into the rabbit cornea will stay at the injection site
indefinitely as a tattoo, unless neovascularization is induced in the cornea (143,144) . As
new capillaries approach the ink spot, the ink breaks up and reappears in the ipsilateral
Iymph nodes. Tumors cells can invade adjacent Iymphatics that form concomitantly with
blood capillaries or, hypothetically, they can pass from the blood stream into the
lymphatics via lymphaticovenous junctions (6, 145). Also, tumor angiogenesis may
facilitate this process by increasing tumor volume, thus enhancing tumor cell-lymphatic
contact at the growing edge of the tumor.

E. MICROVESSEL DENSITY AND TUMOR AGGRESSIVENESS

Although Brem et al. (146) were among the first to suggest that the intensity of
intratumoral angiogenesis may correlate with tumor grade and aggressiveness, the first c1ear-
cut evidence that tumor angiogenesis in a human solid tumor could predict the probability
of metastasis was reported for cutaneous melanorna. Srivastava et aI. (63) studied the
vascularity of20 intermediate thickness skin melanomas. Vessels were highlighted and the
stained histologic sections analyzed with an image analysis system. The 10 cases that
developed metastases had a vascular area at the tumor base that was more than twice that
seen in the 10 cases without metastases (p=0.025).
In 1990, my colleagues and I asked if the extent of angiogenesis (as measured by a
spectrum ofintratumoral microvessel densities) in human breast carcinoma correlated with
metastasis (20). Such information might prove valuable in selecting subsets of breast
carcinoma patients for aggressive adjuvant therapies . When the microvessel counts in a
number of invasive breast carcinomas are sorted in ascending order on a log scale, the

392
spectrum of low to high microvessel densities becomes apparent. The densities are an
evenly distributed continuum, extending from about 10 to 200 microvessels per 0.74 mm2
(200x) field.
Invasive breast carcinomas from patients with metastases (either Iymph nodal or
distant site) had a mean microvessel count of 101 per 200x field . For those carcinomas from
patients without metastases the corresponding value was 45 per 200x field (p=0 .003) . We
plotted the percent of patients with metastatic disease in whom avessei count was carried
out within progressive 33 vessel increments. The plot showed that the incidence of
metastatic disease increased with the number of microvessels, reaching 100% for patients
having invasive carcinomas with > 100 microvessels per 200x field.
To further define the relationship of intratumoral microvessel density to overall and
relapse-free survivaI and to other reported prognostic indicators in breast carcinoma, a
blinded study of 165 consecutive carcinoma patients was performed using identical
techniques to measure intratumoral microvessel density (23). The other prognostic
indicators evaluated were metastasis to axillary Iymph nodes, patient age, menopausal
status, tumor size, histologic grade (i.e., Scarff-Bloom-Richardson criteria), peritumoral
Iymphatic-vascular invasion (pLVI), flow DNA ploidy analysis, flow S-phase fraction,
growth fraction by Ki-67 binding, c-erbB2 oncoprotein expression, pro-cathepsin-D
content, estrogen-receptor content, progesterone-receptor content, and EGFR expression.
We found a highly significant association of intratumoral microvessel density with overall
survival and relapse-free survival in aIl patients, incIuding node-negative andnode-positive
sub sets. All patients with breast carcinomas having > 100 microvessels per 200x field
experienced tumor recurrence within 33 months of diagnosis , compared to <5% of patients
who had <33 microvessels per 200x field. Moreover, intratumoral microvessel density was
the only significant predictor of overall and relapse-free survivaI among node-negative
women .
Other studies performed on different patient databases by different investigators at
different medical centers have observed this same association of increasing intratumoral
vascularity with various measures of tumor aggressiveness, such as higher stage at
presentation, greater incidence of metastases , and/or decreased patient survival. This has
been shown in studies ofpatients not only with carcinoma ofthe breast (20-44), but also in
those with prostate carcinoma (45-50), head-and-neck squamous carcinoma (51-55), non-
small-cell lung carcinoma (56-61) , maIignant melanoma (62-68) , gastrointestinal carcinoma
(69-72), testicular germ-cell maIignancies (73), multiple myeloma (74), central nervous
system tumors (75,76), ovarian carcinoma (77,78), cervicaI squamous carcinoma (79-81),
endometrial carcinoma (82), and transitionaI-cell carcinoma ofthe bladder (88-91) . In many
of these studies intratumoral vascularity was found to have independent prognostic
significance when compared to traditional prognostic markers by multivariate analysis.
Thus far, tumors originating in the liver, biliary system, or pancreas have not been
extensively studied by these techniques . Nonetheless, Shimoyama et al. (83) found that the
expression of angiogenin in pancreatic carcinoma is related to cancer aggressiveness, Egawa
et al. (84) discovered that tumor angiogenesis plays an important role in growth of a
hamster pancreatic celliine, and several groups have reported an association between tumor
angiogenesis and a propensity of gastrointestinal carcinomas to metastasize to the liver (85-
87). Additional work needs to be done.
Although assaying intratumoral microvessel density has been the most common
means of determining tumor aggressiveness, some investigators have found similar
associations with tumor aggressiveness using image analysis to measure vascular area or
doppler ultrasound to measure blood flow . The optimum technique for assaying

393
intratumoral vascularity has not been completely defined. Finally, it is important that some
studies have found that high intratumoral microvessel density can be significantly
associated with a favorable outcome, apparently when specific forms of therapy are given
wherein therapeutic effectiveness depends directly on the extent of blood flow (81,92) .
Kohno et al. (91) found that cervical carcinomas treated with hypertensive intra-arterial
chemotherapy are more responsive when highly vascular, and Zatterstrom et al. (92) found
that highly vascular squamous carcinomas of the head-and-neck are more responsive to
radiation therapy when highly vascular.F. WHY IS MICROVESSEL DENSITY
RELATED TO TUMOR AGGRESSIVENESS?
The association between intratumoral microvessel density and tumor aggressiveness
could be explained in the following ways . First, a highly angiogenic primary tumor with a
high intratumoral microvessel density is more likely to seed distant sites with highly
angiogenic clones (1,176). Second, solid tumors are composed of two discrete yet
interdependent components (i.e., the malignant cells and the stroma they induce), and
measuring intratumoral microvessel density could be a valid measure of the success that a
particular tumor has in forming this very important stromal component. Also, the
endothelial cells of this stromal component may be stimulating the growth of the tumor
cells in areverse paracrine fashion. 1fthis is true, the more microvessels there are, the more
endothelial cells there will be, thus, the greater this paracrine growth stimulation. Third, the
density ofthe microvessel bed within a tumor is likely a direct measure of the size of the
vascular window through which tumor cells pass to spread to distant body sites. The larger
that window, the greater the number of circulating tumor cells from which a metastasis
could develop . Finally, if it is true that endothelial cells play a very active role in the
metastatic process and that tumor cells are actually more "passive" than previously
thought, then intratumoral microvessel density could be a measure of those endothelial-
derived forces promoting metastases . I believe all of these factors are acting together to
encourage tumor growth and metastasis . Indeed, it is no surprise that intratumoral
microvessel density correlates with various measures oftumor aggressiveness.

G. CONCLUDING COMMENTS

Tumor angiogenesis alone is not sufficient to cause metastases . Tumor cells must
also proliferate, penetrate host tissues and vessels, survive within the vasculature, escape
the host's immune system, and then begin growth at a new body site. The behavior of
typical bronchial carcinoids illustrates this point: they are highly vascular tumors, yet they
rarely metastasize to distant sites. It remains to be seen whether intratumor microvessel
density will be universally reproducible and continue to hold up as a predictor of
metastasis or patient outcome when utilized in different centers. As therapies become
more effective in preventing tumor recurrence, the ability of a prognostic test to stratify
patients into various prognostic categories becomes diminished. With a 100% eure rate (or
death rate for that matter), all prognostic tests for predicting patient survival become
meaningless. Nonetheless, the well-documented association of increasing intratumoral
microvessel density with various measures of tumor aggressiveness have increased our
understanding about the critical role of angiogenesis in human tumor growth and metastasis .

394
REFERENCES
1. Folkman J. Clinical applications of angiogenesis research . N Eng) J Med 1995;
333,1757-1763 .
2. Folkman J. Tumor angiogenesis: therapeutic implications. N Eng! J Med
1971;285 :1182-1186.
3. Folkman J, Hochberg M, Knighton D. Self-regulation of growth in three
dimensions: the role of surface area limitations. In: Clarkson B, Baserga R, eds,
Control of Proliferation in Animal Cells, Cold Spring Harbor: Cold Spring Harbor
Laboratory, 1974;pp 833-842.
4. Sutherland RM . Cell and environment interactions in tumor microregions: the
multicell spheroid model. Science 1988;240:177-184.
5. Sutherland RM , McCredie JA, Inch WR . Growth of multicell spheroids in tissue
culture as a model of nodular carcinomas. J Natl Cancer Inst 1971;46:113-120.
6. Blood CH, Zetter BR . Tumor interactions with the vasculature: angiogenesis and
tumor metastasis. Biochimica et Biophysica Acta 1990;1032 :89-118.
7. Gimbrone MA , Leapman SB, Cotran RS, Folkman J. Tumor dormancy in vivo by
prevention of neovascularization. J Exp Med 1972;136:261-276.
8. Gimbrone MA , Cotran RS, Leapman SB, Folkman J. Tumor growth
neovascularization : An experimental model using rabbit cornea . J Natl Canc Inst
1974;52:413-427.
9. Folkman J. What is the evidence that tumors are angiogenesis-dependent? J Natl
Cancer Instit 1990;82:4-6.
10. AntoneIli-Orlidge A, Saunders KB, Smith SR, D'Amore PA . An activated form of
transforming growth factor-beta is produced by co-cultures of endothelial cells and
pericytes. Proc Natl Acad Sei USA 1989;86:4544-4588.
11. Ausprunk DH, Folkman J. Migration and proliferation of endothelial cells in
preformed and newly formed blood vessels during tumor angiogenesis . Microvasc
Res 1977;14 :53-65.
12. Adam JA, Maggelakis AA. Diffusion of regulated growth characteristics of a
spherical prevascular carcinoma. Bull Math Biol 1990;52 :549-582 .
13. Roberts AB, Sporn MB, Assoian RK, Smith JM, Roche NS, Wakefield LM , Heine
UI, Liotta LA, Falanga V, Kehrl JH, Fauci AS. Transforming growth factor type-
beta : rapid induction offibrosis and angiogenesis in vivo and stimulation of collagen
formation in vitro. Proc Natl Aca Sei USA 1986;83:4167-4171.
14. Knighton D, Ausprunk D, Tapper D, Folkman J. Avascular and vascular phases of
tumor growth in the chick embryo. Brit J Canc 1977;35 :347-356.
15. Lien W, Ackerman N. The blood supply of experimentalliver metastases. H. A
microcirculatory study of normal and tumor vessels of the liver with the use of
perfused silicone rubber. Surgery 1970;68:334-340 .
16. Thompson WD , Shiach KJ, Fraser RA, McIntosh LC, Simpson JG . Tumors
acquire their vasculature by vessel incorporation, not vessel ingrowth. J Pathol
1987; 151:323-332.
17. Skinner SA, Tutton PJM , O'Brien PE. Microvascular architecture of experimental
colon tumors in the rat. Cancer Res 1990;50: 2411-2417.
18. Vartanian R, Weidner N. Correlation of intratumoral endotheIial-ceII proliferation
with microvessel density (tumor angiogenesis) and tumor-cell proliferation in breast
carcinoma. Am J Pathol 1994;144:1188-1194 .

395
 19. Vartanian RK, Weidner N. Endothelial cell proliferation in human prostatic
carcinoma and hyperplasia: correlation with microvessel density, epithelial cell
proliferation, and Gleason's score. Lab luvest 1995;73:844-850 .
20. Weidner N, Sempie JP, Welch WR, Folkman 1. Tumor angiogenesis and metastas is -
correlation in invasive breast carcinoma . N Engl J Med 324:1-8, 1991.
21. Bosari S, Lee AKC, DeLellis RA, et al. Microvessel quantitation and prognosis in
invasive breast carcinoma. Hum PathoI23 :755-761,1992.
22. Horak E, Leek R, Klenk N, LeJeune S, Smith K, Stuart N, Greenall M, Stepniewska
K, Harris AL. Angiogenesis, assessed by plateletlendothelial cell adhesion molecule
antibodies, as indicator of node metastases and survival in breast cancer. Lancet
340 :1120-1124 , 1992.
23. Weidner N, Folkman J, Pozza F, Bevilacqua P, Allred EN, Moore DH, Meli S,
Gasparini G. Tumor angiogenesis: a new significant and independent prognostic
indicator in early-stage breast carcinoma . J Natl Cancer Inst 84:1875-1887, 1992.
24. Visscher DW, Smilanetz S, Drozdowicz S, Wykes SM. Prognostic significance of
image morphometric microvessel enumeration in breast carcinoma. Anal Quant
Cytolol Histol 15:88-92, 1993.
25. Toi M, Kashitani J, Tominaga T. Tumor angiogenesis is an independent prognostic
indicator ofprimary breast carcinoma. Int J Cancer 55:371-374,1993.
26. Gasparini G, Weidner N, Bevilacqua P, Maluta S, Dalla Palma P, Caffo 0,
Barbareschi M, Boracchi P, Marubini E, Pozza F. Tumor microvessel density, p53
expression, tumor size, and peritumoral Iymphatic vessel invasion are relevant
prognostic markers in node-negative breast carcinoma. J Clin Oncol 12:454-
466,1994.
27. Obermair A, Czerwenka K, Kurz C, Buxbaum P, Schemper M, Sevela P.
Influenceof Tumoral Microvessel Density on the Recurrence-Free Survival 10
Human Breast Cancer: Preliminary Results. Onkologie 17:44-49,1994.
28. Gasparini G, Bevilacqua P, Boracchi P ,Maluta S, Pozza F, Barbareschi M, Dalla
Palma P, Mezzetti M, Harris AL. Prognost ic value of p53 expression in early-stage
breast carcinoma compared with tumour angiogenesis, epidermal growth factor
receptor, c-erbB2, cathepsin D, DNA ploidy , parameters of cell kinetics and
conventional features . IntJ OncoI4:155-162, 1994.
29. Fox SB, Leek AK, DeLellis RA, et al. Tumor angiogenesis in node-negative breast
carcinomas : relationship with epidermal growth factor receptor, estrogen receptor,
and survival. Breast Can Res Treat 29: 109-116, 1994.
30. Gasparini G, Barbareschi M, Boracchi P, Bevilacqua P, Verderio P, Dalla Palma P,
Menard S. 67-kDa laminin-receptor expression adds prognostic information to
intra-turnoral microvessel density in node-negative breast cancer. Int J Cancer
60:7604-610,1995 .
31. Inada K, Toi M, Hoshina S, Hayashi K, Tominaga T. Significance of tumor
angiogenesis as an independent prognostic factor in axillary node-negative breast
cancer. Gan To Kagaku Jap J Cancer Chemo 22 (suppl 1):59-65, 1995.
32. Charpin C, Devictor B, Bergeret D, Andrac L, Boulat J, Horschowski N, Lavaut
MN, Piana L. CD 31 quantitative immunocytochemical assays in breast
carcinomas. Correlation with current prognostic factors . Am J Clin Pathol 103:443-
448, 1995.

396
33 . Barbareschi M, Gasparini G, Weidner N, Morelli L, Forti S, Eccher C, Fina P,
Leonardi E, Mauri F, Bevilacqua P, Dalla Palma P. Microvessel density
quantification in breast carcinomas: assessment by manual vs . a computer-assisted
image analysis system . Appl Immunohistochem 3:75-84,1995 .
34 . Obermair A, Czerwenka K, Kurz C, Kaider A, Sevelda P. Tumor vascular density
in breast tumors and their effect on recurrence free survival. Chirurg 65 :6 11-6 15,
1994 .
35 . Toi M, Hoshina S, Yamamoto Y, Ishii t, Hayashi K, Tominaga T. Tumor
angiogenesis in breast carcinoma: significance of vessel density as a prognostic
indicator. Gan To Kagaku Ryoho Japanese J Cancer Chemo 21 (Suppl 2) :178-182,
1994 .
36 . Gasparini G, Barbareschi M, Boracchi P, Verderio P, Caffo 0, Meli S, Palma PD ,
Marubini E, Bevilacqua P. Tumor angiogenesis may predict c1inical outcome of
node-positive breast cancer patients treated either with adjuvant hormone therapy
or chemotherapy . Cancer J Sei Am 1:131-141, 1995.
37 . Ogawa Y, Chung Y-S, Nakata B, Takatsuka S, Maeda K, Sawada T, Kato Y,
Yoshikawa K, Sakurai M , Sowa M. Microvessel quantitation in invasive beast
cancer by staining for factor Vlll-related antigen . Brit J Cancer 1995; 71: 1297-
1301.
38 . Obermair A, Kurz C, Czervenka K, Thoma M, Kaider A, Wagner T, Gitsch G,
Sevelda P . Microvessel density and vessel invasion in Iymph node negative breast
cancer: effect on recurrence-free survival. Int J Cancer 62: 126-130, 1995 .
39 . Bevilacqua P, Barbareschi M, Verderio P, Boracchi P, Caffo 0 , Dalla Palma P, Meli
S, Weidner N, Gasparini G. Prognostic value of intratumoral microvessel density, a
measure of tumor angiogenesis, in node-negative breast carcinoma - results of a
multiparametric study . Breast Cancer Res Treat 1995; 36 :205-217.
40 . Toi M , Inada K, Suzuki H, Tominaga T. Tumor angiogenesis in breast cancer: its
importance as a prognostic indicator and the association with vascular endothelial
growth factor expression. Breast Cancer Res Treat 1995; 36 :193-204.
41. Gasparini G, Barbareschi M, Boracchi P, Bevilacqua P, Verderio P, Calla Palma P,
Menard S. 67-kDa laminin-receptor expression adds prognostic information to
intra-turnoral microvessel density in node-negative breast cancer. In J Cancer
1995;60 :604-610.
42 . Fox SB, Turner GDH, Leek RD, Whitehouse RM, Gatter KC, Harris AL. The
prognostic value of quantitative angiogenesis in breast cancer and role of adhesion
moleeule expression in tumor endothelium . Breast Cancer Res Treat 1995;36 :219 -
226.
43 . Scopinaro F, Schillaci 0 ,Scarpini M, Mingazzini PL, diMacio L, Banci M, Danieli
R, Zerilli M , Limiti MR, Centi Colella A. Technetium-99m sestamibi : an indicator
ofbreast cancer invasiveness. Eur J Nucl Med 1994;21 :984-987.
44 . Lipponen P, Ji H, Aaltomaa S, Syrjanen K. Tumor vascularity and basement
membrane structure in breast cancer as related to tumor histology and prognosis . J
Cancer Res Clin Oncol 1994; 120:645-650.
45 . Wakui S, Furusato M, Itoh T, Sasaki H, Akiyama A, Kinoshita I, Asano K, Tokuda
T, Aizawa S, Ushigome S. Tumor angiogenesis in prostatic carcinoma with and
without bane marrow metastasis: a morphometric study. J Pathol 168:257-
262 ,1992.

397
46. Weidner N, Carroll PR, Flax J, Blumenfeld W, Folkman J. Tumor angiogenesis
correlates with metastasis in invasive prostate carcinoma. Am J Pathol 143:401-
409,1993 .
47. Fregene TA, Khanuja PS, Noto AC, Gehani SK, Van Egmont EM, Luz DA, Pienta
KJ . Tumor- associated angiogenesis in prostate cancer. Anticancer Res 13:2377-
2381,1993 .
48. Brawer MK, Deering RE, Brown M, Preston SD, Bigler SA. Predictors of
pathologic stage in prostate carcinoma. Cancer 73:678-687,1994.
49. Vesalainen S, Lipponen P, Talja M, A1hava E, Syrjanen K. Tumor vascularity and
basement membrane structure as prognostic factors in Tl-2MO prostatic
adenocarcinoma . Anticancer Res 14:709-714,1994.
50. Hall MC, Troncoso P, Pollack A, Zhau HY, Zagars GK, Chung LW, von
Eschenbach AC. Significance of tumor angiogenesis in c1inically localized prostate
carcinoma treated with external beam radiotherapy. Urology 44:869-875, 1994.
51. Mikami Y, Tsukuda M, Mochimatsu I, Kokatsu T, Yago T, Sawaki S. Angiogenesis
in head and neck tumors. Nip Jib Gak Kai 96:645-50, 1991.
52. Gasparini G, Weidner N, Bevilacqua P, Maluta S, Boracchi P, Testolin A, Pozza F,
Folkman J. Intratumoral microvessel density and p53 protein: correlation with
metastasis in head-and-neck squamous-cell carcinoma. Int J Cancer 1993;55:739-
744.
53. Albo D, Granick MS, Jhala N, Atkinson B, Solomon MP. The relationship of
angiogenesis to biological activity in human squamous cell carcinomas of the head
and neck. Ann Plast Surg 32:588-594,1994.
54. Williams JK, Carlson GW, Cohen C, Derose PB, Hunter S, Jurkiewicz MJ . Tumor
angiogenesis as a prognostic factor in oral cavity tumors . Am J Surg 168:373-380,
1994.
55. A1calde RE, Shintani S, Yoshihama Y, Matsumura T. Cell proliferation and tumor
angiogenesis in oral squamous cell carcinomas. Anticancer Res 1995;15:1417-1422.
56. Macchiarini P, Fontanini G, Hardin MJ, Hardin MJ, Squartini F, Angeletti CA.
Relation of neovasculature to metastasis of non-small-cell lung cancer. Lancet
340:145-146,1992
57. Macchiarini P, Fontanini G, Dulmet E, de Montpreville V, Chapelier AR, Cerrin J,
Le Roy Ladurie F, Dartevelle PG. Angiogenesis: an indicator of metastasis in non-
small-cell lung cancer invading the thoracic inlet. Ann Thorac Surg 57:1534-
1539,1994.
58. Yamazaki K, Abe S, Takekawa H, Sukoh N, Watanabe N, Ogura S, Nakajima I,
Isobe H, Inoue K, Kawakami Y. Tumor angiogenesis in human lung
adenocarcinoma. Cancer 1994;74:2245-2250.
59. Yuan A, Yang P-C, Yu C-J, Yao Y-T, Lee Y-C, Kuo S-H, Luh K-T. Tumor
angiogenesis correlates with histologie type and nodal metastasis in non-small cell
lung carcinoma . Amer J Resp Critical Care Med Dec. 1995, 152(6 Pt 1): 2157-62 .
60. Fontanini G, Bigini D, Vignati S, Basolo F, Mussi A, Lucchi M, Chine S, Angeletti
CA, Bevilacqua G. Recurrence and death in non small cell lung carcinomas: a
prognostic model using pathological parameters, microvessel count, and gene
protein products . J Pathol 1995;177:57-63.
61. Angeleui ,CA, Lucchi M, Fontanini G, Mussi A, Chella A, Ribechini A, Vignati S,
Bevilacqua G. Prognostie significance of tumoral angiogenesis in completely
resected late stage lung carcinoma (stage IIIA-N2). Impact of adjuvant therapies in
a subset ofpatients at high risk ofrecurrence. CA, Aug 1, 1996; 78(3):409-15.

398
62. Srivastava A, Laidler P, Hughes LE, Woodcock J, Shedden EI. Neovascularization
in human cutaneous melanorna: a quantitative morphological and Doppler
ultrasound study . Eur J Cancer Clin Oncol 22:1205-1209, 1986.
63 . Srivastava A, Laidler P, Davies R, Horgan K, Hughes LE. The prognostic
significance oftumor vascularity in intermediate-thickness (0.76-4 .0 mm thick) skin
melanoma. Am J Pathol133:419-23, 1988.
64 . SmolIe J, Soyer H-P, Hofmann-Wellenhof R, Smolle-Juettner FM, Kerl H. Vascular
arch itecture ofmelanocytic skin tumors . A quantitative immunohistochemical study
using automated image analysis. Path Res Pract 185:740-745, 1989.
65 . Fallowfield ME, Cook MG . The vascularity of primary cutaneous melanoma . J
Pathol 1991; 164:241- 244.
66. Barnhill RL, Levy MA. Regressing thin cutaneous malignant melanomas «1 .0 mm)
are associated with angiogenesis. Am J Pathol 1993;143:99-104 .
67. Vacca A, Ribatti D, Roncali L, LospalIuti M, Serio G, Carrel S, Dammacco F .
Melanocyte tumor progression is associated with changes in angiogenesis and
expression ofthe 67-kilodalton laminin receptor . Cancer 72:455-461, 1993.
68 . Graham CH, Rivers J, Kerbel RS, Stankiewicz KS, White WL . Extent of
vascularization as a prognostic indicator in thin «0.76 mm) malignant melanomas.
Am J Pathol 145:510-514 , 1994
69. Saclarides Tl, Speziale NJ, Drab E, Szeluga DJ, Rubin DB. Tumor angiogenesis and
rectal carcinoma. Dis Colon Rectum 37:921-926,1994 .
70. Maeda K, Chung Y-S, Takatsuka S, Ogawa Y, Sawada T, Yamashito Y, Onoda N,
Kato Y, Nitta A, Arimoto Y, Kondo Y, Sowa M. Tumor angiogenesis as a predictor
ofrecurrence in gastric carcinoma. J Clin Oncol 13:477-481, 1995.
71. Maeda K, Chung YS, Takatsuka S, Ogawa Y, Onoda N, Sawada T, Kato Y, Nitta
A, Arimoto Y, Kondo Y, Sowa M. Tumor angiogenesis and tumor celI proliferation
as prognostic indicators in gastric carcinoma. Br J Cancer 1995;72 :319-323 .
72. Takebayashi Y, Akiyama S-I, Yamada K, Akiba S, Aikou T. Angiogenesis as an
unfavorable prognostic factor in human colorectal carcinoma. Cancer
1996;78(2) :226-31 .
73. Olivarez D, Ulbright T, DeRiese W, Foster R, Reister T, Einhorn L, Sledge G .
Neovascularization in c1inical stage A testicular germ cell tumor: prediction of
metastatic disease. Cancer Res 54:2800-2802,1994.
74. Vacca A, Ribatti D, Roncali L, Ranieri g, Serio G, Silvestris F, Dammacco F. Bone
marrow angiogenesis and progression in multiple myeloma. Brit J Haematol 87:503-
508, 1994.
75. Li VW, Folkerth RD, Watanabe H, Yu C, Rupnick M, Bames P, Scott RM, Black
PM, Sallan SE, Folkman I. Microvessel count and cerebrospinal fluid basic
fibroblast growth factor in children with brain tumors. Lancet 334:82-86, 1994.
76. Leon SP, Folkerth RD, Black PM. Microvessel density is a prognostic indicator for
patients with astroglial brain tumors. Cancer 1996;77(2) :362-72 .
77. Hollingsworth HC, Kohn EC, Steinberg SM, Rothenberg ML , Merino MI. Tumor
angiogenesis in advanced stage ovarian carcinoma . Am J Pathol 147:33-41, 1995.
78. Volm M, Koomagi R, Kaufmann M, Mattem J, Stammler G. Microvessel density,
expression of proto- oncogenes, and resistance-related proteins and incidence of
metastases in primary ovarian carcinomas. Clin Exp Met 1996;(1n press) .
79. Wiggins DL, Granai CO, Steinhoff MM, Calabresi P. Tumor angiogenesis as a
prognostic factor in cervical carcinoma. Gynecol Oncol 56:353-356, 1995.

399
80. Bremer GL, Tiebosch ATMG, van der Putten HWHM, Schouten HJA, de Haan J,
Arends J-W . Tumor angiogenesis: an independent prognostic parameter in cervical
cancer. Amer J Obstet Gynecol ,Jan 1996; 174(1 Pt 1):126-31.
81. SchlengerK, Hockel M, Mitze M, Schaffer U, Weikel W, Knapstein PG, Lambert
A. Tumor vascularity - A novel prognostic factor in advanced cervical carcinoma .
GynecolOncol 1995;59:57-66.
82. Abulafia 0 , Triest WE, Sherer DM, Hansen CC, Ghezzi F. Angiogenesis in
endometrial hyperplasia and stage I endometrial carcinoma. Obstet Gynecol
1995;86:479-485.
83. Shimoyama S, Gansauge F, Gansauge S, Negri G, Oohara T, Beger HG . Increased
Angiogenin Expression in Pancreatic Cancer Is Related to Cancer Aggressiveness.
CA Res lune 1996; 56: 2703-2706.
84. Egawa S, Tsutsumi M, Konishi Y, Kobari M, Matsuni S, Nagasaki K, Futami H,
Yamaguchi K . The role of Angiogenesis in the Tumor Growth of Syrian Hamster
Pancreatic Cancer Cell Line HPD-NR. Gastroenterology 1995;108:1536-1533 .
85. Konno H, Tanaka T, Matsuda I, Kanai T, Maruo Y, Nishino N, Nakamura S, Baba
S. Comparison ofthe Inhibitory Effect of the Angiogenesis Inhibitor TNP-470 and
Mitomycin C on the Growth and Liver Metastasis of Human Colon Cancer . Int 1
Ca 1995; 61:268-271,
86. Tomisaki S, Ohno S, Ichiyoshi Y, Kuwano H, Maehara Y, Sugimachi K .
Microvessel Quantification and Its Possible Relation with Liver Metastasis is
Colorectal Cancer. CA (Suppl) April 15, 1996;77(8) :1722-1728.
87. Maeda K, Chung Y, Ogawa Y, Takatsuka S, Kang S, Ogawa M, Sawada T, Sowa
M . Prognostic Value of Vascular Endothelial Growth Factor Expression in Gastric
Carcinoma. CA, March 1, 1996;77(5) :858-863 .
88. laeger TM , Weidner N, Chew K, Moore DH, Kerschmann RL, Waldman FM,
Carroll PR . Tumor angiogenesis correlates with lymph node metastases in invasive
bladdercancer.1UroI154:59-71;1995.
89. Bochner BH, Cote RJ, Weidner N, Groshen S, Chen S-C, Skinner DG , Nichols PW.
Angiogenesis in bladder cancer : relationship between microvessel density and tumor
prognosis. INati Cancer Instit 1995;87:1603-1612.
90. Dickinson AJ, Fox SB, Persad RA, Hollyer 1, Sibley GN, Harris AL . Quantification
of angiogenesis as an independent predictor of prognosis in invasive bladder
carcinomas. Br 1 Urol 1994;74 :762-766.
91. Kohno Y, Iwanari 0, Kitao M. Prognostic importance ofhistologic vascular density
in cervical cancer treated with hypertensive intraarterial chemotherapy. Cancer
1993;72:2394-2400.
92. Zatterstrom UK, Brun E, Willen R, Kjellen E, Wennerberg 1. Tumor angiogenesis
and prognosis in squamous cell carcinoma of the head and neck. Head & Neck
1995; 17:312-318 .
93 . Folkrnan 1. Angiogenesis and its inhibitors. In: DeVita VT, Hellman S, Rosenberg
SA, eds, Important Advances in Oncology . Philadelphia: 1.B. Lippincott Co,
1985;pp 42-62 .
94. Harris AL, Fox S, Bicknell R, Leek R, Reif M, LeJeune S, Kaklamanis L. Gene
therapy through signal transduction pathways and angiogenic growth factors as
therapeutic targets in breast cancer. Cancer 1994;74(3 Suppl) :1021-1025 .
95. Ingber D, Fujita T, Kishimoto S, Katsuichi s, Kanamaru T, Brem H, Folkrnan 1.
Synthetic analogues of fumagillin that inhibit angiogenesis and suppress tumor
growth. Nature 1990;348:555-557.

400
96 . Gross JL , Herbl in WF, Dusak BA, Czemiak P, Diamond M, Dexter DL.
Modulation of solid tumor growth in vivo by bFGF . Proc Am Assoc Cancer Res
1990 ;31 :79(Abst).
97 . Kim KJ , Li B, Winer J, Armanini M, Gillett N, Phillips HS, Ferrara N. Inhibition of
vascular endothelial growth factor-induced angiogenesis suppresses tumor growth in
vivo. Nature 1993;362 :841-844.
98 . Hori A, Sasada R, Matsutani E, Naito K, Sakura Y, Fujita T, Kozai Y. Suppression
of solid tumor growth by immunoneutralizing monoclonal antibody against human
basic fibroblast growth factor. Cancer Res 1991;51:6 180-6 184.
99. Millauer B, Shawver LK, Plate KH, Risau W, Ullrich A. Glioblastoma growth
inhibited in vivo by a dominant-negative Flk-l mutant. Nature 1994;367 :576-579.
100. Brooks PC , Montgomery AMP, Rosenfeld M, Reisfeld RA, Hu T, I1ier G, Cheresh
DA . Integrin alpha.beta, antagonists promote tumor regression by inducing
apoptosis of angiogenic blood vessels . Cell 1994;79 :1157-1164.
10 1. O'Reilly MS , Holmgren L, Shing Y, Chen C, RosenthaI RA, Moses M, Lane WS,
Cao Y, Sage EH, Folkman 1. Angiostatin: a novel angiogenesis inhibitor that
mediates the suppression of metastases by a Lewis lung carcinoma. Cell
1994;79:315-328.
102. O'Reilley MS, Holmgren L, Chen K, Folkman 1. Suppression of angiogenesis in
mice by angiostatin inhibits murine and human primary tumor growth . (Submitted).
103. Nicosia RF, Tchao R, Leighton 1. Interactions between newl y formed endothelial
channels and carcinoma cells in plasma elot culture. Clin Exp Metast 1986 ;4:91-
104.
104. Warren RS, Yuan H, Matli MR , Gillett NA , Ferrara N . Regulation by vascular
endothelial growth factor of human colon cancer tumorigenesis in a mouse model of
experimentalliver metastasis. J Clin Invest 1995;95:1789-1797 .
105. Rak JW, Hegmann EJ, Lu C, Kerbel RS. Progressive loss of sensitivity to
endothelium-derived growth inhibitors expressed by human melanoma cells during
disease progression . J Cell Physiol 1994; 159:245-255 .
106. Hamada J, Cavanaugh PG, Lotan O. Separable growth and migration factors for
large-cell lymphoma cells secreted by microvascular endothel ial cells derived from
target organs for metastasis. Br J Cancer 1992;66 :349-354.
107. Folkman J. Angiogenesis and breast cancer . J Clin OncoI1994;12:441-443.
108. Pepper MS , Vassalli JD, Montesano R, Orci L. Urokinase type plasminogen
activator is induced in migrating capillary endothelial cells. J Cell Biol
1987; 105:2535-254 1.
109. Fox SB, StuartN, Smith K, BrunnerN, Harris AL. High levels ofuPA and PA-l are
associated with highly angiogenic breast carcinomas. J Pathol 993;170 (Suppl):388a .
110. Hildenbrand R, Dilger I, Horlin A, Stutte H1. Urokinase and macrophages in tumor
angiogenesis. Br J Cancer 1995;72:818-823 .
111. Moscatelli D, Gross J, Rifkin D. Angiogenic factors stimulate plasminogen
activator and collagenase production by capillary endothelial cells . J Cell Biol
1981;91:20Ia.
112. Dvorak HF . Tumors: wounds that do not heal. Similarities between tumor stroma
generation and wound healing . N Engl J Med 1986;315 :1650-1659.
113. Folkrnan J, Klagsbrun M . Angiogenic factors . Science 1987; 235:442-7.
114. Polverini PJ , Leibovich SJ. Induction of neovascularization in vivo and endothelial
proliferation in vitro by tumor associated macrophages. Lab Invest 1984 ;51:635-42.

401
115. Guidi AJ, Fischer L, Harris JR, Schnitt SJ. Microvessel density and distribution in
ductal carcinoma in situ ofthe breast. J Natl Cancer Inst 1994;86:614-619.
116. Smith-McCune KK, Weidner N . Demonstration and characterization of the
angiogenic properties of cervical dysplasia. Cancer Res 1994;54:800-804.
117. Kandel J, Bossy-Wetzel E, Radvani F, Klagsburn M, Folkman J, Hanahan D .
Neovascularization is associated with a switch to the export of bFGF in the multi-
step development of fibrosarcoma. Cell 1991;66:1095-1104.
118. Nguyen M, Watanabe H, Budson AE, Richie JP, Folkman J. Elevated levels of the
angiogenic peptide basic fibroblast growth factor in urine of bladder cancer patients.
J Natl Cancer Instit 1993;85:241-242.
119. Dvorak HF, Brown LF, Detmar M, Dvorak AM. Vascular permeability
factor/vascular endothelial growth factor, microvascular hyperpermeability, and
angiogenesis . Am JPathoI1995;146:1029-1039.
120. Brown LF, Berse B, Jackman RW, Tognazzi K, Manseau EJ, Dvorak HF, Senger
DR. Increased expression of vascular permeability factor (vascular endothelial
growth factor) and its receptors in kidney and bladder carcinomas. Am J Pathol
1993;143:1255-1262.
121. Brown LF, Berse B, Jackman RW, Tognazzi K, Manseau EJ, Senger DR, Dvorak
HF . Expression ofvascular permeability factor (vascular endothelial growth factor)
and its receptors in adenocarcinomas of the gastrointestinal tract. Cancer Res
1993;53:4727-4735.
122. Senger DR, Van De Water L, Brown LF, Nagy JA, Yeo K-T, Yeo T-K, Berse B,
Jackman RW, Dvorak AM, Dvorak HF. Vascular permeability factor (VPF, VEGF)
in tumor biology . Cancer Met Rev 1993;12:303-324.
123. Thompson WD, Campbell R, Evans T. Fibrin degradation and angiogenesis:
quantitative analysis of the angiogenic response in the chick chorioallantoic
membranes. J Pathol 1985;145:27-37.
124. Toi M, Hoshina S, Takayanagi T, Tominaga T. Association of vascular endothelial
growth factor expression with tumor angiogenesis and with early relapse in primary
breast cancer . Jpn J Cancer Res 1994;85:1045-1049.
125. Goto F, Goto K, Weindel K, Folkman J. Synergistic effects of vascular endothelial
growth factor and basic fibroblast growth factor on the proliferation and cord
formation of bovine capillary endothelial cells within collagen gels. LabInvest
1993;69:508-517.
126. Maeda K, Chung Y-S, Ogawa Y, Takatsuka S, Kang S-M, Ogawa M, Sawada T ,
Onoda N, Kato Y, Sowa M. Thymidine phosphorylase/platelet-derived endothelial
cell growth factor expression associated with hepatic metastasis in gastric
carcinoma. Br J Cancer 1996;73:884-888.
127. Leibovich SJ, Polverini PJ, Fong TW, Harlow LA, Koch AE. Production of
angiogenic acitivity by human monocytes requires an L-arginine/nitric oxide-
synthase-dependent effector mechanism. Proc Natl Acad Sei USA 1994;91:4190-
4194 .
128. Laniado-Schwartzman M, Lavrovsky Y, Stotlz RA, Conners MS, Falck JR,
Chauhan K, Abraham NG. Activation of nuclear factor kappa Band oncogene
expression by 12(R)-hydroxyeicosatrienoic acid, an angiogenic factor in microvessel
endothelial cells. J Biol Chem 1994;269 (39): 2432-2437 .
129. Rastinejad F, Polverini PJ, Bouck NP. Regulation of the activity of a new inhibitor
ofangiogenesis by a cancer suppressor gene. CelI1989;56:345-355.

402
130. Bouck NP. Tumor angiogenesis: the role of oncogenes and tumor suppressor genes .
Cancer Cells 1990;2:179-185.
131. Zajchowski DA, Band V, Trask DK, Kling D, Connoly JL, Sager R. Suppression of
tumor-forming ability and related traits in MCF -7 human breast cancer cells by
fusion with immortal mammary epithelial cells. Proc Natl Acad Sei USA
1990;87 :2314-2318.
132. Dameron KM, Volpert OV, Tainsky MS, Bouck N. Control of angiogenesis in
fibroblasts by p53 regulation of thrombospondin-1. Science 1994;265 :1582-1584.
133. Folkman J. Angiogenesis in cancer, vascular, rheumatoid, and other disease. Nature
Medicine 1995;1:27-3 1.
134. Fidler IJ, Gersten DM, Hart IR. The biology of cancer invasion and metastasis.
Adv Cancer Res 1978;28:149-250.
135. Nicolson G. Cancer metastasis. Sei Am 1979;240 :66-76.
136. Weiss L. "Biophysical aspects of the metastatic cascade ." In: Weiss L, ed,
Fundamental Aspects ofMetastasis, Amsterdam : N. Holland, 1976; pp . 51-70.
137. Bernstein LR, Liotta LA. Molecular mediators of interactions with extracellular
matrix components in metastasis and angiogenesis. Curr Opin Oncol 1994;6: 106-
113.
138. Liotta L, Kleinerman J, Saidel G. Quantitative relationships of intravascular tumor
cells, tumor vessels, and pulmonary metastases following tumor implantation.
Cancer Res 1974;34:997-1004 .
139. Liotta L, Saidel G, Kleinerman J. The significance of hematogenous tumor cell
c1umps in the metastatic process . Cancer Res 1976;36:889-894 .
140. McCulloch P, Choy A, Martin L. Association between tumour angiogenesis and
tumour cell shedding into effluent venous blood during breast cancer surgery . Lancet
1995;346 :1334-1335.
141. Nagy JA, Brown LF, Senger DR, Lanir N, Van de Water L, Dvorak AM, Dvorak
HF . Pathogenesis oftumor stroma generation: a critical role for leaky blood vessels
and fibrin deposition. Biochim Biophys Acta 1989;948 :305-26 .
142. Sugino T, Kawaguchi T, Suzuki 1. Stromal invasion is not essential to blood-borne
metastasis in mouse mammary carcinoma. In: Scientific Program Booklet of the
Pathological Society of Great Britain and Ireland; 170th Meeting, January 1995;
(Abstract #161).
143. Smolin G. Hyundiuk RA . Lymphatic drainage from vascularized rabbit cornea . Am
J Ophthalmol 1971;72:147-151.
144. Folkman J. Angiogenesis . In: Verstraete M, Vermylen J, Lijnan R, Arnout J, eds,
Thrombosis and Haemostasis, Leuven: Leuven University Press . 1987;24 :583-96 .
145. Liotta LA, Stracke ML. "Tumor invasion and metastasis: biochemical mechan isms"
In: Lippman, ME and Dickson RB, eds, Breast Cancer: Cellular and Molecular
Biology, Boston : Kluwer Academic Publishers, 1988, pp.223-238.
146. Brem S, Cotran R, Folkman J. Tumor angiogenesis: a quantitative method for
histologie grad ing. J Natl Cancer Inst 1972; 48:347-356.
147. Folkman J, Watson K, Ingber D, Hanahan D. Induction of angiogenesis during the
transition from hyperplasia to neoplasia. Nature 1989;339 :58-61 .
148. Weidner N . "The relationship of tumor angiogenesis and metastasis with emphasis
on invasive breast carcinoma." In: Advances in Pathology and Laboratory Medicine,
Weinstein RS (ed), Mosby Year Book, Chicago, Vol. 1992;5:101-121.

403
149. Hall NR, Fish DE, Hunt N, Goldin RD, Guillou PJ, Monson JRT. Is the
relationship between angiogenesis and metastasis in breast cancer real? Surg Oncol
1:223-229, 1992.
150. Van Hoef MEHM, Knox WF, Dhesi SS, Howell A, Schor AM. Assessment of
tumor vascularity as a prognostic factor in Iymph node negative invasive breast
cancer. Eur J Cancer 29A :1141-1145, 1993.
151. Miliaras D, Kamas A, Kalekou H. Angiogenesis in invasive breast carcinoma : is it
associated with parameters of prognostic significance? Histopathol 26:165-169,
1995.
152. Axelsson K, Ljung BMB, Moore DH, Thor AD, Chew KL, Edgerton SM, Smith
HS, Mayall BH . Tumor angiogenesis as a prognostic assay for invasive ductal
carcinoma. J NatJ Cancer Instit 87:997-1008, 1995.
153. Camochan P, Briggs JC, Westbury G, Davies AJ. The vascularity of cutaneous
melanoma: a quantitative histologic study of lesions 0.85-1.25 mm in thickness. Br
J Cancer 1991;64:102-107 .
154. Leedy DA , Trune DR, Kronz JD, Weidner N, Cohen JI. Tumor angiogenesis, the
p53 antigen, and cervical metastasis in squamous carcinoma. Otolaryngol Head
Neck Surg 1994; 111:417 -422.
155. Rutger JL, Mattox TF, Vargas MP . Angiogenesis in uterine cervical squamous cell
carcinoma. Int J Gyn Path 1995;14:114-118 .
156. van Diest PJ, Zevering JP, Zevering LC, Baak JPA . Prognostic value of
microvessel quantitation in cisplatin treated Figo 3 and 4 ovarian cancer patients.
Path Res Pract 1995;191:25-30.
157. Siitonen SM, Haapasalo HK, Rantala IS, Helin HJ, Isola JJ . Comparison of
different immunohistochemical methods in the assessment of angiogenesis: lack of
prognostic value in a group of 77 selected node-negative breast carcinomas. Mod
Pathol 1995;8:745-752 .
158. Goulding H, Rashid NFNA, Robertson JF, Bell JA, Elston CW, Blamey RW, Ellis
10. Assessment of angiogenesis in breast carcinoma: An important factor in
prognosis? Hum PathoI1995;26:1196-1200.
159. Costello P, McCann A, Camey DN, Dervan PA. Prognostic significance of
microvessel density in Iymph node negative breast carcinoma . Hum Pathol
1995;26 :1181-1184.
160. MacLennan GT, Bostwick DG. Microvessel density in renal cell carcinoma: lack of
prognostic significance. Urology 1995;46:27-30.
161. Tahan SR, Stein AL. Angiogenesis in invasive squamous cell carcinoma of the lip:
tumor vascularity is not an indicator of metastatic risk. J Cut Pathol 1995;22:236-
240 .
162. Dray TG, Hardin NJ, Sofferman RA . Angiogenesis as a prognostic marker in early
head and neck cancer. Ann Otol Rhinol LaryngoI1995;104 :724-729 .
163. Bamhill RL, Busam KJ, Berwick M, Blessing K , Cochran Aj, Eider DE, Fandrey
K, Daraoli T, White WL. Tumor vascularity is not a prognostic factor for cutaneous
melanoma [letter]. Lancet 1994;344:1237-1238 .
164. Kainz C, Speiser P, Wanner C, Obermair A, Tempfer C, Sliutz G, Reinthaller A,
Breitenecker G. Prognostic value of tumor microvessel density in cancer of the
uterine cervix stage IB to 00. Anticancer Res 1995;15:1549-1551.
165. DeYoung BR, Wick MR, Fitzgibbon JF, Sirgi KE, Swanson PE. CD31: An
immunospecific marker for endothelial differentiation in human neoplasms . Appl
Immunohistochem 1993;1:97-100 .

404
166. Longacre TA, Rouse RV. CD31: A new marker for vascular neoplasia. Advan Anat
Pathol 1994; 1:16-20.
167. van de Rijn M, Rouse RV. CD34 : A review. Appl Immunohistochem 1994;2:71-80 .
168. Traweek ST, Kandalaft PL, Mehta P, Battifora H. The human hematopoietic
progenitor cell antigen (CD34) in vascular neoplasia. Am J Clin Pathol 1991;96:25-
31.
169. Schlingemann RO, Rietveld FJR, Kwaspen F, van de Kerkhof PCM, de Waal
RMW, Ruiter D1. Differential expression ofmarkers for endothelial cells, pericytes ,
and basal lamina in the microvasculature of tumors and granulation tissue. Am J
PathoI1991;138 :1335-1347 .
170. Wang JM, Kumar S, Pye D, Haboubi N, AI-Nakib L. Breast carcinoma:
comparative study of tumor vascu1ature using two endothelial-cell markers . J Natl
Cancer Inst 1994;86:386-388.
171. Wang JM, Kumar S, Pye D, van Agthoven AJ, Krupinski J, Hunter RD . A
monoclonal antibody detects heterogeneity in vascular endothelium of tumors and
normal tissues . IntJ Cancer 1993; 54:363-370 .
172. Watanabe 11, Nguyen M, Schizer M. Basic fibroblast growth factor in human serum
- a prognostic test for breast cancer. Mol Biol Cell 1992;3:324a.
173. Nguyen M, Watanabe 11, Budson AE. Elevated levels of an angiogenic peptide,
basic fibroblast growth factor, in the urine of patients with a wide spectrum of
cancers. J Natl Cancer Inst 1994; 86:356.
174. Li VW, Folkerth RD, Watanabe H, Yu C, Rupnick M, Barnes P, Scott RM, Black
PM, Sallan SE, Folkman 1. Microvessel count and cerebrospinal fluid basic
fibroblast growth factor in children with brain tumors . Lancet 334 :82-86, 1994.
175. Esserman L, Hylton N, George T, Yassa L, Weidner N. Constrast-enhanced
magnetic resonance imaging (cMRI) provides a window to visualize anatomic extent
and tumor angiogenesis in breast carcinoma. Cancer Res (submitted).
176. Herlyn M, Clark WH, Rodeck U, Mancianti ML, Jambrosic J, Koprowski H.
Biology oftumor progression in human melanocytes. Lab Invest 1987;56:461-7.

405
TUMOR VASCULARITY, HYPOXlA, AND MALIGNANT PROGRESSION IN
SOLID NEOPLASMS

Michael Höckel, Karlheinz Schienger, Billur Aral, Uwe Schäffer, and

Wolfgang Weikel

Dept of Ob/Gyn, University of Mainz, Mainz, Gennany

1. INTRODUCTION

Malignant progression designates the biologic process which transforms a phenotypically

normal cell fixed and cooperating within a tissue into a disseminated therapy-resistant
lethai disease . In c1inical terms this process consists of three major steps (Fig. 1):
( ) the transition from regulated to deregulated cell proliferation,
( ) the emerging ability of the neoplastic cell collectives to induce angiogenesis and to
invade other tissues,
() the development ofmetastases and ofresistance towards anti-tumor therapies.
Whereas the -step can be reversible resulting in spontaneous regression of dysplastic
tissues, the -step irreversibly leads to a growing and infiltrating neoplasm , which can
however be completely eliminated by standard treatment in case of early detection . The -
step usually renders the neoplastic disease incurable.
We have devoted our research to the understanding ofbiological mechanisms ofthe -stepof
malignant progression. Although cancer is a genetic disease and malignant progression is
characterized by the accumulation of genomic mutations , the level of our investigations is
not molecular or intracellular. Since malignant progression is an evolutionary biological
phenomenon we expect to find Darwinian principles at supracellular levels. We use the
methods of population ecology and have defined a tumor volume/area element ("tuxel ") as
the smallest unit to study parameters representing tumor cell population dynamics. We
have chosen advanced cancer of the uterine cervix with increasing tumor sizes as a model
for -malignantprogression in human solid neoplasms . So far we could demonstrate that -
malignant progression in cervical cancer is associated with tumor core hypoxia (Höckel,
Schlenger, Knoop , and Vaupel, 1991; Höckel, Knoop, Schienger, Vomdran, Baussmann,
Mitze, Knapstein, and Vaupel, 1993a; H öckel, Vorndran, Schlenger, Baussmann,
Knapstein, and Vaupel, 1993b; Höckel, Schlenger, Aral, Mitze , Schäffer, and Vaupel,
1996). Indeed, the tumor oxygenation profile proved to be the most powerful

Angiogenesis: Models, Modula tors. and Clinica l Applications

Edited by Mara goudaki s, Plenum Press, New York, 1998 407
 Regulated proliferation

@ I Gersi~
Deregulated proliferation

®l~
Angiogenes is
Invasion

0l~
Metastasis
Chemoreslstance
Radioresistance
Fig.l
The three c1inical steps of malignant progression in solid neoplasms.

pretherapeutical predictor of survival and recurrence-free survival in a cohort of more than

100 cervical cancer patients studied prospectively. Moreover, the prognostic re!evance of
tumor oxygenation was independent of the mode of primary treatment, radical surgery or
standard radiotherapy .
With this paper we introduce recent findings on the emergence of tumor hypoxia during -
malignant progression in human cervical cancer.

2. MATERIALS AND METHODS

2.1 Patients
All patients with cervical cancers of at least 2 cm diameter as determined by c1inical and CT
or MRI investigation treated with surgery at the Department of Ob/Gyn, University of
Mainz Medical Center were eligible for the prospective study. The study design was
approved by the ethics committee. Patients were accrued once informed consent had been
obtained. Fourteen patients without cervical pathology volunteered to participate in the
study as control subjects representing normal cervices .

2.2 Intratumoral p02 histography

Tumor oxygenation was measured pretherapeutically with the Eppendorf histograph
system adhering strictly to the standard procedure as developed and validated earlier
(Röcke! et al.,1993a; Höckel et al.,1993b). Immediately after calibration of the ethylene
oxide sterilized needle e!ectrode, the conscious patient was placed into a defined lithotomy
position . p02 readings were performed along linear tracks, first in the normal fatty tissue of
the mons pubis followed by measurements at the 12 o'clock and 6 o'c1ock sites of the
central cervical tumor or normal cervices. Twenty-five to 35 p02 measurements were taken
on each tumor track, (50-70 readings in total) starting at a tissue depth of 5 mm. The
measuring points were placed 0.7 mm apart from each other resulting in an overall
measuring track length of approximately 2 to 2.5 cm. By an on-line computing system the
p02 data of each track were displayed (i) as absolute values of oxygen partial pressure
related to the location ofthe measuring point along the track and (ii) as relative frequencies
within a p02 histogram ranging from 0 to 100 mmHg with a dass width of2.5 mmHg .
After the p02 measurements, cylindrical punch biopsies of 2 mm diameter and 2 cm length

408
were taken from those tumor areas where the p02 determinations had been made using a
BioptyTM device (Radiplast AB, Uppsala, Schweden). The biopsies were fixed in buffered
formalin and processed for histology and immunohistochemistry as specified below .
Concomitantly with the p02 determinations intravaginal temperature, heart rate, arterial
blood pressure, hemoglobin concentration and hematocrit were monitored. The p02
measurements were usually performed 1-5 days prior to the oncologic treatment. The
oxygenation status of each tumor was represented by the median p02 and the low p02
fraction (corresponding to the relative frequency Ofp02 readings from 0 - 2.5 mmHg or 0 -
5 mmHg) derived from the pooled histograms.

2.3 Tumor biopsies

The biopsies were immediately fixed in a buffered formalin solution . Standard paraffin
blocks were prepared and 5- m sections were stained with hematoxylin and eosin (H&E)
for conventional histologic evaluation .
To highlight proliferating cells immunohistochemical Iocalization of the Ki-67 antigen was
perfromed in 5- m sections offormalin-fixed and paraffin-embedded tissue after dewaxing
and microwave oven treatment using the MIB I monoclonal antibody (Dioanova GmbH,
Hamburg, Germany) according to Cattoretti, Becker, Key, Duchrow, Schlüter, Galle, and
Gerdes (1992)
For the specific staining of blood vessels five-micrometer-thick paraffin sections of the
biopsy cylinders were deparaffinized. After treatment with trypsin, nonspecific binding
sites were blocked with goat serum . The sections were then treated with rabbit anti-human
von Willebrand factor (Dako Diagnostica, Hamburg, Germany) diluted to 1:300 and
incubated at 37% C for 30 min. For the detection of the primary antibody, the avidin-
biotin-peroxidase-diaminobenzidine system was applied (Elite Peroxidase Kit , Vector
Laboratories, Burtinggame, CA). Positive staining served to identify blood vessel
endothel ium. Tumor cells were highlighted by counterstaining the sections with
hematoxylin.

2.4 Morphometry
The quantitative determination of (i) the density of tumor ceIIs, (ii) the density of
proliferating tumor ceIIs, (iii) the median distances from tumor cells to the closest
microvessel within the tumor core was performed in immunohistochemically stained
sections ofthe 12 o'clock and 6 o'clock core biopsies of each tumor. The three parameters
were determined in each ofthe tuxels in linear array outlined by overlaying a grid with Imm
by Imm calibration .
Tumor cells and proliferating tumor cells were identified by their nuclear dimensions
without and with peroxidase stain in MIB-I stained sections counterstained with
hemato xylin. Within a tuxel all tumor cells and a11 MIB-I positive tumor cells in two
random subfields of 0.005 mm? displayed at IOOx magnification were counted to estimate
the mean cell number per tuxel.
Distances from tumor cells to the closest microvessel were measured according to the
method described by Kayar, Archer, Andrew, Lechner, and Banchero (1982) at 50 random
sites per tuxel in core biopsy sections stained for factor VIII-reiated antigen and
counterstained with hematoxylin. Differences were measured by a computerized image
analysis system . Image acquisition was performed using a frame grabber board (ITI-MFG-
3M-V, Imaging Technology Inc., Bedford, MA) and a RGB camera (DXC-15IP, Sony
Corp. , resolution 756X581 picture elements), which was attached to a laboratory
microscope (BH-2 , Olympus Corp .). The microscopic picture was displayed (at IOOx

409
 100
P < 0.05
Ci 90 n =14
I
E 80
oS 70
....
Q)
:J
l/l 60 r - P < 0.05 - - ,
l/l
....
Q)
50
n =26 n =23
0-
CU
:eCU 40 •
• •
••
.
0- 30
c:
Q) • ..
••
• -••-
Ol 20
>.
x •
• • • • . ..
• ..
0 10

0
··
0 10
•
• •
20
......
....
30 40
.. ...... - - 1 ..
.. , t .... ..
50 60 70 80
Normal cervix Maximum tumor diameter [mm]

Fig .2
Median p02 measured in normal cervica1 tissue and in the core regions of cervical cancers
with increasing sizes representing y-malignant progression.

magnification) on a video monitor (FA3435KL Mitsubishi Corp .). The actual magnification
was determined by calibrating the system using a stage micrometer. Image analysis was
performed with an interactive software (Optimas Bio Scan Inc., Edmonds, WA) . To
exclude incorrect detenninations for random points located at the field margins, the
boundaries for the random tumor points were set at 100 m smaller than the vessel field
perimeters . Tumor microvessels were identified by their positive staining for Factor VIII-
related antigen without the necessity of either lumen or erythrocyte filling . At each random
point a neoplastic cell and its closest blood vessel were marked with the cursor. The
distances were measured and further processed by the computer. The mean distance from
tumor cells to the closest microvessel was calculated for each tuxel.

2.5 Histopathological tumor size deter mination

The surgical specimens obtained by radical hysterectomy were dissected by serial slicing of
the amputated cervix and parametria according to Schmidt-Mathiesen (1988) and then fixed
in fonnalin and processed for routine histological examination. Tumor diameters were
detennined in craniocaudal, anterior-posterior and lateral directions verified
microscopically. Tumor maximum diameter was used for size classification.

2.6 Data analysis

For all four parameters determined in the tuxel histograms were calculated representing the
individual tumor. The medians of the tumor histograms were plotted against the maximum
tumor size measured histopathologically in the surgical specimen. The Mann -Whitney-
Wilcoxon test (U test) was applied for the comparison of the medians. A p value ce 0.05
was considered to indicate statistical significance.

4 10
 100

E 90
2: • ..
Q) 80 ..
l/l
l/l • • ..
70
..
Q)
>
....
0 •• .... ..
o 60 • .. .. ,
E
iii 50 • .- _. .. .. t
....
..
Q)
l/l
• • • ....
0 40
• • •
U
30
•
.8
Q)
n =17 n =20
o
c
ro
20
n =11 l.---
n.s. ----.J

iii 10 n.s.
0
0
0 10 20 30 40 50 60 70 80
Normal cervix Maximum tumor diameter [mm]
Fig.3
Median distances to the closest microvessel from cells of normal cervical tissue and from
neoplastic cells of the core regions of cervical cancers with increasing sizes representing -
malignant progression .

~ All tumor cells

.. Proliferating tumor cells

....
o
E
....
::J

--....
o
Q)
'J

..
.0
E
::J
C
..
C
ro
Q) . ..
~

I I I I
o 10 .w m ~ ~ ~ 70 80
Maximum tumor diameter [mm]
Fig.4
Median densities of neoplastic cells (open squares) and of proliferating neoplastic cells
(triangles) in the core regions of cervical cancers with increasing sizes representing -
malignant progression.

411
3.RESULTS

From 1989 until 1997 49 patients with cervical cancers of > 2 cm dinical diameter treated
by radical surgery entered the prospective study . At this writing oxygenation
measurements have been carried out for all 49 patients, whereas vascularity, tumor cell
density and proliferating tumor cell densities have been determined for a subset of 37 and
26 patients only .
To analyze the dynam ics of these tumor core parameters during -malignant progression,
their integral values characterizing each tumor have been plotted against the maximum
tumor size determined histopathologically. Despite pronounced intertum oral variations
within the same size groups any dependence on -malignant progression should become
evident by this way of evaluation . As shown in Fig. 2 median tumor core p02 decreases
with increasing tumor size/progres sion. Generally, median tumor core p02 values are
significantly lower than the median p02 values measured in normal cervices. However,
median tumor cell - microvessel distances within the tumor center do not change during -
malignant progression and vary within the range of median cell - microvessel distances in
normal cervices (Fig. 3). Likewise, median densities of all neoplastic cells as weil as median
densities ofproliferating tumor cells within the tumor core did not show a dependence on
tumor size/progression (Fig. 4).

4. DISCUSSION

According to the present understanding cancer is adonai genetic disease characterized by

the accumulation of mutations in the genome of neoplastic cells during the disease course .
Whereas the molecular pathology at the beginning (- step) of the malignant progression,
i.e. the emergence of deregulated proliferation, is becoming to be partl y understood as a
result of the intense reduction istic biological research during the last decade, the molecular
basis for the - and -steps is largely unknown .
Because of the complexity of a malignant tumor as a biological system with mostly non-
linear behavior it is questionable if a reductionistic ("top-down") approach will elucidate
the biology of advanced malignant progression at all. From a dinical standpoint however,
the understanding of -malignant progression is ofhigh priority, since the -step usuallyleads
to the incurability ofthe malignant disease .
We try to study some aspects of -malignant progression with the "bottom-up" approach of
tumor population ecology in cancer ofthe uterine cervix as a model for human solid cancers.
It has been demonstrated earlier that the tumor size determined histopathologically strongly
correlates with malignant progression in cervical cancer (Burghardt, Pickel, Haas, and
Lahousen, 1987).
The probability of having achieved the -step of malignant progression (i.e. the ability to
metastatize and to become radioresistant) significantly increases with tumor sizes greater
than 2 cm. Whereas about 85-90% ofthe patients with histopathological tumor sizes of up
to 2 cm can be cured with standard surgery or radiation, patients with tumors larger than 5
cm as determined histopathologically have only a small chance to survive 5 years and
longer. Therefore we have focussed our investigation on cervical cancers of 2 cm and larger
sizes which represent increasing -malignant progression. To study biological mechanismsof
-malignant progression we have defined a tumor volume/area element ("tuxel ") as the
smallest unit of investigation. The tuxel is supracellular in dimension composed of viable
tumor cells, stroma and necrotic tissue . Its actual size is determined by technical features of

412
the method of investigation, e.g. the field size of microscopic observation. Within a tuxel
various parameters influencing population ecology are determined. Tuxels are sampled in
systematic random fashion to be representative for the whole tumor.
We have shown that tumor core oxygenation in advanced (>2 cm) cervical cancer is
significantly lower than normal cervix oxygenation and decreases with increasing tumor
size. -malignant progression is clearly associated with tumor core hypoxia.
What is the reason for that phenomenon? Since the median distances from tumor cells to
the closest microvessels do not decrease with increasing tumor size/ progression and overall
tumor cell densities as weil as proliferating cell densities do not increase , the emerging
tumor hypoxia is most probably the consequence of the more and more insufficient
function ofthe tumor vascularity with waxing lesion sizes .
In experimental tumors a variety of mechanisms impairing tumor core microcirculation have
been demonstrated: structural abnormal ities ofthe vessel wall and the vascular architecture
as weil as the lack of a functioning Iymphatic drainage system lead to increased vascular
permeability, high inte rstitial pressure, altered rheology, arteriovenous shunt perfusion and
various other phenomena which all result in diminished nutritive blood flow (Vaupel ,
Kallinowski, and Okunieff, 1989).
Our observation of the association between tumor hypoxia and -malignant progression in
cervical cancer does not prove a causal interrelationship. Nevertheless, recent findings in
vitro and in animal tumor models support the view that tumor hypoxia per se may have an
impact on the emerging clinical aggressiveness in solid neoplasms: Tumor cells, like normal
cells, when starved in a hypoxic microenvironment res pond with the expression of a variety
of genes coding for oxygen regulated proteins (Heacock, and Sutherland, 1986; Sutherland,
Ausserer, Murphy, and Laderoute, 1996). Several potential mechanisms have been
suggested for how some of these oxygen-regulated proteins, including p53 and VEGF,
influence tumor aggressiveness through increased tumor dissemination and decreased
responsiveness to therapy (Takahashi, Kitadai, Bucana, Cleary, and Ellis, 1995; Giaccia,
1996) . Moreover, there is strong evidence that hypoxia enhances genetic instability and the
selective pressure leading to the evolution of metastatic and intrinsically resistant tumor cell
variants (Rice, Hoy , and Schimke, 1986; Young, MarshalI , and Hili, 1988; Russo, Weber,
Volpe, Stoler, Petrelli , Rodriguez-Bigas, Burhans, and Anderson, 1995).
The growth advantage of tumor cell populations lacking the ability for apoptosis under
hypoxic conditions may represent an important biological mechanism of - malignant
progression (Graeber, Osmaninan, Jacks , Housman, Koch, Lowe, and Giaccia, 1996 ).

REFERENCES

1. Burghardt, E., Pickel , H., Haas, 1., and Lahousen, M ., 1987, Prognostic factors and
operative treatment of stages IB to IIB cervical cancer, Am. 1. Obstet. Gynecol.
156:988-996.
2. Cattoretti, G., Becker, M .H.G., Key, G., Duchrow, M ., Schlüter, c., Galle, 1., and
Gerdes, J., 1992, Monoclonal antibodies against recombinant parts of the Ki-67
antigen (MIB 1 and MIB 3) detect proliferating cells in microwave-processed
formalin-fixed paraffin sections,1. Pathol. 168:357-363.
3. Giaccia, AJ., 1996, Hypoxie stress proteins: survival ofthe fittest, Semin. Radiat.
Oncol. 6:46-58.

413
4. Graeber, T.G ., Osmaninan, C., Jacks, T., Housman, D.E., Koch, C.K Lowe, S.W.,
and Giaccia, AJ., 1996, Hypoxia-mediated selection of cells with diminished
apoptotic potential in solid tumors, Nature 379:88-91 .
5. Heacock, C.S., and Sutherland, RM., 1986, Induction characteristics of oxygen
regulated proteins. Int. J. Radiat. Onco/. Biol. Phys. 12: 1287-1290.
6. Höcke1, M ., Schlenger, K , Knoop , C., and Vaupel, P., 1991, Oxygenation of
carcinomas of the uterine cervix: evaluation by computerized O2 tension
measurements, Cancer Res. 51:6098-6102.
7. Höckel, M., Knoop, C., Schlenger, K, Vomdran B., Baussmann, E., Mitze, M .,
Knapstein, P.G., and Vaupel, P, 1993a, Intratumoral p02 predicts survival in
advanced cancer ofthe uterine cervix, Radiother. Oncol. 26:45-50 .
8. Höcke1, M., Vomdran , B., Schlenger, K, Baussmann, E., Knapstein, P.G., and
Vaupe1, P., 1993b, Tumor oxygenation: a new predictive parameter in locally
advanced cancer of the uterine cervix, Gynecol. Oncol. 51: 141-149.
9. Höckel, M., Schlenger, K, Aral, B., Mitze, M., Schäffer, U., and Vaupe1, P., 1996,
Association between tumor hypoxia and malignant progression in advanced cancer
of the uterine cervix, Cancer Res. 56:4509-4515 .
10. Kayar, S.R, Archer, P.G., Andrew, A., Lechner, 1., and Banchero, N ., 1982, The
c1osest-individual method in the analysis of the distribution of capillaries,
Microvasc. Res. 24:326-341 .
11. Rice, G.c. , Hoy , C., and Schirnke, RT., 1986, Transient hypoxia enhances the
frequency of dihydrofolate reductase gene amplification in chinese hamster ovary
cells, Proc. Natl. Acad Sei. USA 83:5978-5982.
12. Russo, C.A ., Weber, T.K, Volpe, C.M., Stoler, D.L., Petrelli N.J., Rodriguez-
Bigas, M ., Burhans, W.c., and Anderson G.R., 1995, An anoxia inducible
endonuc1ease and enhanced DNA breakage as contributors to genomic instability in
cancer, Cancer Res. 55: 1122-1128.
13. Schmidt-Mathiesen, H., 1988, Histopathologische Basisinformationen als
Voraussetzung für individuelle gynäkologische Therapie, Pathologe 9:251- 257.
14. Sutheriand , R.M., Ausserer W., Murphy, B., and Laderoute K., 1996, Tumor
hypoxia and heterogeneity : Challenges and opportunities for the future, Semin.
Radiat. Oncol. 6:59-70 .
15. Takahashi, Y , Kitadai , Y, Bucana C.D., Cleary, KR., and Ellis L.M ., 1995,
Expression of vascular endothelial growth factor and its receptor, KDR, correlates
with vascularity, metastasis, and proliferation of human colon cancer, Cancer Res.
55:3964-3968.
16. Vaupel, P., Kallinowski , F., and Okunieff, P., 1989, BloOO flow, oxygen and
nutrient supply, and metabolic microenvironment of human tumors: A review,
Cancer Res. 49:6449-6465.
17. Young, S.D., MarshalI, R.S., and Hili R.P. , 1988, Hypoxia induces DNA
overreplication and enhances metastatic potential of murine tumor cells, Proc. Natl.
Acad Sei. USA 85:9533-9537.

414
SCATTER FACTOR AS A MEDIATOR OF TUMOR ANGIOGENESIS

Recent C1inical and Experimental Studies

Eliot M. Rosen, " Katr in Lamszus.!" John Laterra." Peter J. Polverin i;'
Jeffrey S. Rubin ," and Itzhak D. Goldbe rg '

IDep artment of Radiat ion Onc ology, Long Island Jewi sh Medical Ce nter
270-0 5 76th Avenue New Hyd e Park , New York 11040
2 Departments of Neurology, Oncol ogy, and Neurosc ience, Th e Joh ns
Hopkins Sc hoo l of Med icin e and Kenn edy-Krieger Research Instit ute
Baltimore, Mary land 2 1205
3 Unive rsity of Mic higa n Schoo l of Dent istry, Departm ent of Ora l Pathology,
Room 52 17
101 1 N. Univers ity Ann Arbor, Mic higan 48 109-1078
4 Laboratory of Ce llu lar and Mo lecular Biology, National Ca ncer Institute
Build ing 37, Room I E24 Bethesda , Maryland 20892

INTRODUCTION

Sca tter factor (SF) (a lso known as hepatocyte grow th factor (HGF» is an invasogen ic
and ang ioge nic cyto kine that has been implicat ed in diverse biologie proc esses, inc ludi ng or-
gan regeneration , embryogenesis, wound healing, and carcin og ene sis (review ed in Rosen et
al., I 994c ). SF-induced biologie responses are medi ated throu gh the canonical SF receptor, a
tran smembrane prot ein tyro sine kin ase encoded by a proto-oncogene (e-Met) (Bottaro et al.,
1991 ; Go nzatti-Haces et al., 1988). Two years ago, at the NATO Advanced Study Institut e
confe rence on " Molecular, Ce llular, and Clinical Aspe cts of Angioge nes is", we present ed
our studies on the ability ofSF to induce an angiog enic phenotyp e in cultured vasc ular endo -
thel ial ce lls and to induce new blood vessel form at ion in seve ral in vi vo ass ays of angiogene-
sis, incl uding the mou se Matr igel assay and the rat corneal neovascular izat ion assa y. We a lso
presented data suggesting a prominent role for SF and the c-Me t recept or in the pat hogenes is
of AIDS-related Kaposi 's sa rco ma, a tum or associated wi th prom inen t co mpone nts of endo -
theli al ce ll proliferation and angiogenesis . Th ese studies are reviewed in two chap te rs in the
Na to Se ries A : Life Seiences Vol. 285 (see Rosen and Go ldberg, 1996; Rosen et al., 1996b ).

Addre ss all correspo ndence to Dr. Rosen .

t Dr. Lamszus ' curre nt add ress is: Department 01' Neuropatho logy, University Hospital Eppendor f, Mart ini
Strasse, #52 20246 Harnburg. Germa ny.

Angiogenes is: Mod els. M odulators. and Clinical App lications

Edited by Maragoudakis, Plenum Press, New York, 1998 4 15
Additional information can be found in the original references (see Rosen et al., 1991a and b;
Grant et al. , 1993; Naidu et al., 1994; Rosen and Goldberg, 1995) .
In this chapter, we will discuss some ofour more recent studies relating to the potential
role ofSF as a tumor angiogenesis factor. We will describe studies on the following topics: I) in-
hibition ofSF-mediated angiogenesis using highly specific recombinant reagents; II) expres-
sion ofSF and its receptor, c-Met, during human cancer progression; III) clinical studies on the
role ofSF in human tumor angiogenesis; and IV) experimental animal studies on the role ofSF
in tumor angiogenesis. The data presented suggest that SF may function to promote the malig-
nant progression ofhuman tumors not only by enhancing the motility and invasiveness ofthe
tumor cells but also by stimulating an angiogenic response within the tumor stroma.

I) INHIBITION OF SCATTER FACTORMEDIATED ANGIOGENESIS

Several peptide fragments ofthe SF a-chain, designated NKI and NK2 , are generated
by alternative mRNA processing. These peptides consist of the N-terminal pre-kringle re-
gion plus the first kringle domain (NK1) or the first two kringle domains (NK2) ofthe SF
a-chain (Chan etal., 1991; Lokkeretal., 1992; Cioce etal., 1996). Recombinant human NKI
and NK2 peptides were found to bind with high affinity to the c-Met receptor and to exert par-
tial agonistic and antagonistic activity for full-length heterodimeric SF (Cioce et al. , 1996;
Lokker et al., 1992). Utilizing the rat corneal neovascularization assay, we found that neither
NKI nor NK2 were capable of inducing angiogenesis, but each of these peptides inhibited

Table I. Inhibition of scatter factor-mediated angiogenesis in

the rat cornea by scatter factor o-chain peptides NKI and NK2
and by chimeric c-Met IgG '
Agent(s) added Positive assays (%)
A. Inhib ition ofSF-Mediated Angiogenesis by NKI and NK2 Peptides
Control assays
Hydron + PBS 0/2 (0)
bFGF (50 ng) 2/2 (100)
rhSF (50 ng) 3/3 (100)
NKI (25 ng) 0/2 (0)
NK2 (25 ng) 0/2 (0)
rhSF ± Peptide
bFGF (50 ng) + NKI (25 ng) 2/2 (100)
bFGF (50 ng) + NK2 (25 ng) 2/2 (100)
rhSF (50 ng) + NKI (25 ng) 0/2 (0)
rhSF (50 ng) + NK2 (25 ng) 0/2 (0)

B. Inh ibition ofSF-Mediated Angiogenesis by c-Met IgG

Control assays
Hydron + PBS 0/2 (0)
bFGF (100 ng) 2/2 (100)
rhSF (100 ng) 3/3 (100)
c-Met IgG (I ug) 0/4 (0)
rhSF ± c-Met IgG
rhSF (100 ng) 3/3 (100)
rhSF (100 ng) + c-Met IgG (250 ng) 3/6 (50)
rhSF (100 ng) + c-Met IgG (500 ng) 1/5 (20)
rhSF (100 ng) + c-Met IgG (I ug) 0/5 (0)
1Abbreviations : bFGF, basic fibroblast growth factor ; PBS, phosphate-
buffered saline; rhSF, recombinant human scatter factor. Assays were per-
formed and responses scored as described before (Toisma et al., 1993).

416
 Table 2. Inhibition of chemotaxis of human dermal
microvascular endothelial cells toward scatter factor (SF)
by recombinant human SF a-ch ain peptides NKI and NK21
Agent(s) add ed (ng /ml) Cell s/ten 40x grids (N) % Inhibition
None 18±3 (4)
rhSF (20) 136± 12 (4) o
NK I (20) 21 ± 4 (3)
NKI (50) 26 ± I I (3)
NK I (200) 12±4 (3)
rhSF (20) + NKI (20) 130 ± 10 (3) 8
rhSF (20) + NK I (50) 101 ±4 (3) 36
rhSF (20) + NK I (200) 56 ± 6 (3) 63
NK2 (20) 15±3 (3)
NK2 (50) 9±3 (3)
N K2 (200) 17±10 (3)
rhSF (20) + NK2 (20) 94 ± I (3) 33
rhSF (20) + NK2 (50) 100± 10 (3) 23
rhSF (20) + NK2 (200) 54 ± 5 (3) 68
'Chemotactic mig ration of human derm al microvascul ar endothe lial
cells toward different combinations ofrhSF and NKI or NK2 was as-
sayed using microwell-modified Boyden chambers, as de scribed be fore
(see Lams zus et al., 1997). Migration values (me ans ± standard errors)
repre sent the number of cells that had migrated acro ss porous coll agen-
coated filter s over a four hour assay incubation period . Values in paren-
theses represent the number of repl icate wells tested .

angiogenesis induced by full-length SF (Table JA) . In contrast, neither ofthese peptides in-
hibited angiogenesis induced by basic fibroblast growth factor .
A chimeric antibody consisting of a human IgG heavy chain fused to bivalent c-Met
receptor extracellular binding domains [designated "c-Met IgG" (Mark et al., 1992)] also
blocked SF-mediated angiogenesis in the rat cornea (Table 1B). One ug of c-Met IgG ablated
the angiogenic action of 100 ng ofrecombinant human SF, a dose sufficient to induc e a maxi-
mal angiogenic response . Consistent with these findings, NK land NK2 inhibited chemotac-
tic migration of human dermal microvascular endothelial cells toward SF (Table 2). In
another assay ofcell motility, NK land NK2 inhibited the migration ofendothel ial cells from
microcarrier beads to tlat plastic surfaces (Table 3). In similar assays, c-Me t IgG (5 ug/rnl) in-
hib ited 87% of the migration induced by 20 ng/ml of SE These find ings sugg est that NK 1,
NK2, and c-Met IgG may be useful reag ents for specifically blocking SF-induced endothelial
cell act ivation and angiogenesis.

11) EXPRESSION OF SF AND C-MET DURING HUMAN CANCER

PROGRESSION

Breast Cancer

Non-invasive human breast cancers [ductal carcinoma-in-situ (DCIS)] rarely spread

beyond the breast and are highly curable by local therapies, including mastectomy or radia -
tion . On the other hand , invasive breast cancers frequently spread to regionallymph nodes or
distant sites (eg ., lung , bone , liver) . Consistent with theidea that SF may act as a tumor pro-
gression factor, we have found that SF and c-Met expression are each up-regulated during the
transition from non-invasive tumors to invasive cancers. By semi-quantitative immuno-
chemical staining of different types of human breast lesions, we found that the levels of SF

417
 Table 3. Inhibition of SF-induced m ig rati on of vasc ular
e ndo thelia l ce lls from carri er beads to flat surfaces
by recombinant human SF a -chain peptides NK l and NK2 1
Agent(s) added (ng/ml) CelIs/5 IOx fields % Inhib ition
Contro l (0) 20 I ± 5
NK I (200) 186 ± 7
NK I (500) 218 ± 12
NK2 (200) 208 ± 18
NK2 (500) 189 ± 1l
rhSF (20) 376 ± 8 0
rhSF (20) + NK I (200) 220 ± 10 81
rhSF (20) + NKI (500) 190 ± 15 116
rhSF (20) + NK2 (200) 250 ± 10 76
rhSF (20) + NK2 (500) 219 ± 9 85
'Calf pulmonary artery endothelial cell s cultured on Cytodex 2 m i-
croc arrier beads (Pharmacia) were seeded into 24-well plastic
dishes (100 ,000 cells per weil) in 0.5 ml of DMEM containing 5%
fetal calf serum. Dishes were incubated @ 37"C for one day, after
which the beads were removed by gentle rinsing with phosphate-
buffered saline. Cells that had migrated offthe bead s and onto the
flat bottom surfaces ofthe wells were stained with crystal violet and
counted. Values are means " standard errors of N=3-6 replicate de-
terminat ions. In all case s, migration values for (rhSF + pept ide)
were significantly less than those for rhSF alone (P < 0.00 I ) (tw o-
tailed Student's t-test s).

and c-Met proteins expressed b y benign or malignant mammary epithelial cells in creased
during progression from normal/benign tissue 6 non-invas ive cancer ( D CI S ) 6 infiltratin g
(invasive) ductal carcinoma (Jin et al., 1997) (see Table 4). Since staining intensity is n ot nec-
essarily linearl y related to antigen content, these studies a r e onl y useful for rank order corn-
parisons of antigen levels rather than for comparing relative s toic h io m e tr ic quantities o f the
antigens in the d ifferent lesion types . Within the DCIS and in vasive cancer groups , hi gh SF
staining intensity was s ig n ifi c a nt ly correlated with high c-Met sta ining in tensity.
A more quantitative method for d e te rmin ing the SF content of breast tumors is to
measure immunoreactive SF titers (relative to total p rotein content) by E LISA ofprotein ex-
tracts ofthe tumors . In a study of258 invasive b r ea s t cancers, high SF content in tumor tiss ue
extracts was fo und to be a powerfu l independent predictor of relapse and d e a th (Yamashita et

Table 4. Expression of SF an d c-Met prot e ins in h uman breast lesion s

det erm ined by semi-quantitative immunochemical sta ining analysis'

Type of lesion SF staining score c-M et staining score

Infiltrating Ducta l Carcinoma (IDC) 3.3 ± 0.2 (43) 2.8 ± 0.2 (43)
Ducta l Carcin oma- In-Situ (DCIS ) 2.7 ± 0.1 (40) 2.2 ± 0.2 (39)
Atypic al Hyperpla sia (AH) 1.5 ± 0.3 (6) l.l ± 0.3 (5)
Hyperpl asia (H) 1.7 ± 0.2 (32) 1.0 ± 0.1 (32)
Normal (NL) 1.2 ± 0.2 (9) 1.2 ± 0.3 ( 10)

' Imrnunoperoxidase staining intensities ofmalignant or benign mammary epithelial

celIs were scored semi-quantitatively, on ascale of O(no staining) to 4.0 (maximum
staining ) (see Jin et al. , 1997 for details ). Values are means ± SEMs; numbers in pa-
renthe ses represe nt number of patients . The following compari sons were statisti-
calIy significant for SF and for c-Met: (AH, H, or NL) vs DCIS (P < 0.0 1); (AH, H,
or NL) vs IDC (P < 0.01); and DCIS vs IDC (P=0.03). Note that for (H or NL) vs.
(DCIS or IDe), alI comparisons were significant at P < 0.00 I.

4 18
 Table 5. Content of scatter factor (SF), interleukin-1 ß (I L-I ß),
and von Willebrand's factor (vWF) in tissue extracts of
different types of human breast lesions I
Lesion type SF (ng/mg) IL-I ß (pg/mg) vWF (ng/mg)
Benign 0.74 ± 0.15 (9) 2.4 ± 1.1 (7) 45 ± 22 (9 )
DCIS 0.63 ± 0.14 (24) 13.6 ± 3.4 (18) 45 ± 15 (24)
Benign+DCIS 0.66 ± 0.11 (33) 10.4 ± 2.7 (25) 45 ± 12 (33)
Invasive 1.56 ± 0.11 (230) 21.1 ± 4.1 (204) 95 ± 15 (22 7)

'DClS = ductal carcinoma-in-situ, Tissue extracts wer e prepared ; and the SF,
IL-I ß, and vWF contents were measured using specific ELi SAs, as des crib ed
before (Jin et al., 1997; Jin et al., in press). Values were normali zed to total
protein cont ent of the extract and tabulated as means ± standard erro rs
(number of tumors). Benign tissu e sampi es wer e obtained from lesion s biop-
sied to rule out canccr, including proliferative fibroc ystic disease, usual and
aytpical hyperplasia, papillomatosis, fibroadenoma, scle rosing adeno sis, or
combinations of these . Comp arisons of invasive canc er vs (beni gn+DCIS)
were as follows (two- tailed t-test) : 11-1 ß. P = 0.029; SF. P < 0.0001 ; vWF. P =
0.01.

al., 1994) . We measured the SF content of over 250 breast cancers from the Long lsland Jew-
ish Medical Center's Frozen Tumor Bank. In these studies, we found that SF content was sig-
nificantly greater in invasive cancers, as comp ared with a group of non-invasive cancers
(DCIS) and benign breast lesions (Yao et al., 1996; Jin et al., in press) (see Table 5) . In these
studies, there were too few benign lesions to determine ifthe SF content ofbenign pro li fera-
tive lesions differed from that of DCIS.
It was initially though that tumor-derived SF was produced by tumor-associated stro-
mal cells (eg ., fibroblasts, smooth muscle cells , macrophages), since the vast majority ofcar-
cinoma cell line s, including a number ofhuman breast cancer cell lines , do not produce any
detectable SF in vitro (Stoker et al ., 1987; Rosen et al., 1994a). However, more recent stud ies
indicate that brea st carcinoma cells express both SF protein and SI" mRNA in vivo, suggest-
ing that loss ofthe ability to produce SF may be an art ifact ofcell culture (Jin et al., 1997;
Tuck et al., 1996; Wang et al ., 1994) . Tumor cells also express c-Me t receptor in vivo , sug-
gesting the existence ofboth autocrine and paracrine loops involving the SF:c-Met ligand .re-
ceptor signalling pathway (Jin et al., 1997).

Transitional Cell Bladder Cancer

Tran sitional cell carcinomas ofthe urinary bladder exhibit a spectrum ofbiologic ag-
gressiveness, but tend to occur in two dist inct forms : superficial tumors (which are usually
low grade) vs . muscl e-invasive carcinomas (wh ich are usually high grade). The former cate-
gory of'turnors, which do not invade the bladder wall musculature, tend to follow an indolent
clinical course; whereas the latter category of tumors that penetrate into the bladder wall
muscularis often exhibit rapid local tumor growth, dissemination to regional Iymph nodes,
and distan t metastasis. Immunoreactive SF and c-Met proteins were also detected in hum an
bladder carcinoma tissue sections (Joseph et al. , 1995) . We have analyzed the SF content of
protein extr acts of aseries ofbenign and malignant human bladder lesions. We found signifi-
cantly high er titers of SF in tissue extracts of musc1e-invasive bladder cancers than in non-
musc1e invasive tumors or non-tumor tissue (Rosen et al., 1997) (see Table 6). Invasion was
more closely related to tissue SF content than tumor grade, since high grade non-invasive
cancers had similar SF content to low grade non-invasive cancers. SF was also detetecd in the
urine ofpatients with bladder cancer; with the highest levels observed in patients with clini-
cally active high grade or invasive cancers (Rosen et al., 1997) .

419
 Table 6. Tissue scatter factor (SF) contcnt ofhuman bladder cancersI
Disease category SF content (ng/mg) Comparisons
I. Non-muscle invasive cancers 0.52 ± 0.12 (16) vs 2. P < 0.001
la. Low grade (I-li) 0.58 ± 0.16 (10) vs l b, P = NS
Ib. High grade (IlI-IV) 0.42 ± 0.20 (6) vs 2. P < 0.02
2. Muscle-invasive cancers (all grade IlI-IV) 2.65 ± 0.56 (15)
3. Non-cancer bladder tissue 0.82 ± 0.18 (11) vs 2, P < 0.02
3a. NI bladder adjacent to TCC 0.81 ± 0.20 (5)
3b. Cuff ofbladder neck from radical prostatectomy 0.83 ± 0.31 (6)
I All bladder cancers were transitional cell carcinomas (TCes), with the exception ofthree tumors (two
adenocarcinomas and one squamous cell carcinoma) . One patient had a superficial recurrencc of an
invasive TCC and was excluded from the analysis. Values listed are means ± standard errors ; and num-
bers in parentheses represent the number of different tissue sampies studied . The overall Kruskal-
Wallis H-statistic was 9.52, corresponding to P < 0.01. Pairwise comparisons in the Table were made
using the Mann-Whitney U-test. NS means P > 0.05.

Glial Tumors

Low grade (I-II) glial tumors (eg., astrocytomas, gliomas) tend to have pushing bor-
ders and grow relative1y slowly; while high grade glial tumors (eg., anaplastic astrocytoma,
gliobastoma multiforme) tend to invade the surrounding tissues, grow more rapidly, and ex-
hibit both tumor necrosis and angiogenesis. Glioma tumor cells expressed immunoreactive
SF as weil as c-Met protein, with stronger immunostaining observed in high grade as com -
pared with low grade lesions (Rosen et al., 1996a). In a study of77 human brain tissue ex-
tracts, high grade tumors had significantly greater SF content than did low grade tumors or
non-neoplastic tissue (Lamszus et al., submitted) (see Table 7). We also found that SF stimu-
lated DNA synthesis, proliferation, and chemotactic migration ofthree different lines of hu-
man brain-derived microvascular endothelial cells (Rosen et al., 1996a; Lamszus et al.,
submitted), suggesting that SF might contribute to the high levels of tumor angiogenesis
found in glioblastomas and other high grade gliomas.

Table 7. Tissue scatter factor (SF) content of human gliaJ

tumor extracts as a function of tumor gradeI
Tumor type SF content (ng/mg)
High grade (llI-IV) tumors
Glioblastoma multiforme 1.83 ± 0.67 (27)
Anaplastic astrocytoma /malignant glioma 1.63 ± 0.89 (10)
Malignant oligodendroglioma 1.54 ± 0.56 (10)
All high grade tumors: 1.73 ± 0.43 (47)

Low grade (1-11) tumors

Low grade astrocytoma/glioma 0.81 ± 0.29 (8)
Benign oligodendroglioma 0.49 ± 0.16 (16)
Ependymoma 0.32 ± 0.32 (2)
Alliow grade tumors: 0.58 ± 0.15 (26)
Non-tumor tissue
Temporal lobe 0.49 ± 0.12 (4)
'Tabulated values represent means ± standard errors (number of sam-
pies). Statistical comparison ofthe SF content of rall high grade" vs
"all low grade" tumors gave P < 0.03 (Mann Whitney U-test) .

420
III) ROLE OF SF IN TUMOR ANGIOGENESIS: CLINICAL STUDIES

Recent studies suggest that tumor angiogenesis, indicated by large numbers of

microvessels in tumor stroma, is an independent indicator ofpoor prognosis in patients with
invasive breast cancer and other tumor types (Weidner et al., 1991, 1992; Bosari et al., 1992).
In the studies ofSF content ofhuman breast tumors described above, we also measured the
content ofvon Willebrand's factor (vWF), an endothelial-specific marker protein, and corre-
lated the values ofvWF with those ofSF (Yao et al., 1996; Jin et al., in press). We found that :
1) invasive cancers had significantly higher vWF content than did non-invas ive cancers (see
Table 5); and 2) the vWF content ofinvasive cancers was strongly correlated with SF content
(see Table 8A). Thus, invasive tumors in the highest SF content range had about 8 times as
much vWF as did invasive tumors in the lowest SF range . Admittedly, this assay does not
distinuish between pre-existing host vasculature and newly formed tumor microvessels.
Nevertheless, our findings do suggest that the overall level oftumor vascularity, as indicated
by the vWF content, is related to the SF content in breast cancer tissue .
Interleukin-I (IL-l) is a pleiotropic med iator of biologie responses related to infec-
tion, immunity, and intlammation (reviewed in Dinarello, 1996). IL-l can induce target cells
to produce various cytokines, including SF (Tamura et al., 1993) and several other angio-
genie mediators [eg., vascular endothelial growth factor (VEGF) and interleukin-8 (lL-8)]
(Dinarello. 1996; Ben Av et al., 1995). IL-I itself can induce angiogenesis in several model

Table 8. Relationship ofvon Willbebrand factor content (vWF), scatter factor

content (SF), and interleukin-I ß Content (IL-I ß) in tissue extracts of invasive human
breast cancers
1
A . vWF vs SF alone
SF content range (ng/mg) vWF content (ng/mg) Comparisons
I. 0.00-0.50 26 ± 7 (50) (1+11) vs (II1+IV), P < 0.001
11. 0.51-1.00 49 ± 8 (62) 11 vs I, P = 0.040
III. 1.01-2.00 93 ± 27 (57) III vs IV, P = 0.061
IV. 2.01-up 203 ± 51 (55) IV vs I, P < 0.001

B. vWF vs (SF and IL-I ß)2

SF Range (ng/mg) IL-I ß Range (pg/mg) VWF Content (ng/mg) Comparison
0.00---1.00 All 38±6(103) P < 0.0001
1.01-up All 139±22(101)
All 0---10.0 54 ± 11 (107) P = 0.002
All 10.I -up 125 ± 20 (97)
0.00---1.00 0---10.0 23 ± 6 (60) P = 0.0008
0.00---1.00 10.I-up 60 ± 10 (43)
1.01-up 0---10.0 95 ± 24 (47) P = 0.05
l.OI-up IO.I -up 177 ± 34 (54)

C. SF and vWF vs IL-I ß3

Parameter Low IL-Iß (0---10 pg/mg) High IL-I ß (~I 0 pg/mg) Comparison
SF Content (ng/mg) 1.25 ± 0.12 (108) 1.85 ± 0.22 (96) P = 0.015
vWF Content (ng/mg) 54 ± 11 (108) 126 ± 20 (96) P = 0.0022
IAll values tabulated are means ± standard errors, with numbers oftumor specimens included in
parcntheses. Statistical comparisons were made using the two-tailed Student's t-test .
2This analysis was restricted to tumors for which SF, IL-I ß, and vWF content values were each
available . One-way ANOVA for vWF revealed significant differences (P < 0.01).
3Statistical comparisons of parameters for low vs high IL-I ß range were made using the two-
tailed t-test.

421
systems (Ben Av et al., 1995; Fan et al., 1993). As part ofthe above studies, we exam ined our
breast tumor extracts to determine if immunoreactive IL-1 ß was present, and if the levels
were correlated with those ofSF or vWF (Jin et al., in press).
We found that IL-1 ßcould be detected in about 90% of breast cancers and that IL-1 ß
levels were significantiy higher in invasive cancers than in a group of earlier 1esions (DCIS +
benign) (see Table 5). To determine ifvWF conte nt was related to that of IL-I ß in invasive
cancers, IL-1 ßcontent was divided into two ranges (0-10 pg/mg and > 10 pg/mg), such that
each range contained close to half of the values. vWF titers were found to be significantiy
higher in tumors with high IL-I ß as compared to those with low IL-I ß (Table 8B and C). The
vWF content appeared to be independently related to the SF and IL-1 ß contents, so that the
tumors with highest levels ofboth SF and IL-Iß also had the highest levels ofvWF (Table
8B). Finally, tumors in the high IL-1ß range had about l.5-fold higher SF content than did
those in the low IL-1ß range, and this difference was statistically significant (P < 0.02) (see
Table 8C). Thus it is possible that IL-1 ß may contribute both to the intratumoral accumula-
tion of SF and to the angiogenic phenotype of the tumor.

IV) ROLE OF SF IN TUMOR ANGIOGENESIS: EXPERIMENTAL ANIMAL

MODELS

Human Breast Carcinoma

We have examined the ability of SF to stimulate angiogenesis in an animal model of

human breast cancer. We transfected human SF cDNA into MDAMB231 human breast can-
cer cells so that they overexpressed SF. SF-transfected (SF+) cell clones secreted large quan-
tities of SF into the culture medium (ca. 50 ng/ml in 24 hr). Tumors generated from SF+
clones grew much more rapid1y in the mammary fat pads ofnude mice than did tumors gener-
ated control-transfected (SF-) clones (Lamszus et al., 1997) (see Table 9). Extracts from the
SF+ tumors contained 50-fold higher SF content than did extracts prepared from SF- control
tumors (Table 9). In contrast to the in vivo results, in vitro proliferation rates and cloning effi-
ciencies were no greater in SF+ vs SF- clones . We performed further studies to investigate the
discrepancy between the in vivo and in vitro growth results and found that: a) microvessel
densities were significantly higher in SF+ vs SF- tumors (see Table 10); b) chemotactic activ-
ity for capillary endothelial cells was greater in conditioned media and primary tumor ex-
tracts from SF+ vs SF- clones ; c) angiogenic activity in the rat cornea assay was much higher

Table 9. Final size and scatter factor (SF) content ofprimary tumors
generated from SF-transfected vs control-transfected MDAMB231
human breast cancer cells 1

Primary tumor Primary tumor Tumor SF content (ng

Tumor cell clone weight (g) volume (mnr') SF/mg protein)
21 (SF-transfected) 0.76 ± 0.11 (25) 723 ± 93 (25) 39.3 ± 3.2 (25)
29 (SF-transfected) 1.01 ± 0.24 (17) 866 ± 193 (17) 47.4 ± 3.8 (17)
32 (neo-transfected) 0.19 ± 0.05 (16) 121 ± 22 (16) 1.1 ± 0.3 (16)
34 (neo-transfected) 0.36 ± 0.08 (16) 321 ± 67 (16) 0.8 ± 0.3 (16)
ISF+ and SF- clones ofMDAMB23I human breast cancer cells were grown as tumors in
thc mammary fat pads of athymic nude mice. At the time of sacrifice, primary tumors
were excised for determinations ofwet weight and oftumor tisse SF content . For details,
see Lamszus et al. (1997). All tabulated values are expressed as means ± standard errors
(number of tumors assayed) . Statistical comparisons of pooled data for SF+ tumors
(clone 21+29) vs SF- tumors (clone 32+34) were as folIows: tumorweight, P < 0.001; tu-
mor volume, P < 0.001; SF content, P < 0.001 (two-tailed Student's t-tests).

422
 Table 10. Microvessel counts (MYCs) ofpri mary tumors
generated from SF-transfected vs co ntro l-transfected
MDAMB23 1 human brea st ca ncer ce lls '
Clone Peak MYC Average peak MYC
21 (SF-transfected) 6.4 ± 0.4 (23) 4.7 ± 0.3 (23)
29 (SF-transfected) 6.4 ± 0.8 ( 17) 4.6 ± 0.5 ( 17)
32 (neo-transfected) 4.1 ± 0.5 ( 15) 3.2 ± 0.4 ( 15)
34 (neo-transfected) 3.5 ± 0.5 (15) 2.5 ± 0.3 ( 15)
At the time of sacrifice, tumors were excised. fixed in formalin.
1

embedded in paraffin, and sectioned. Sections corresponding to

the maximum dimension of the tumor were stained with anti-
laminin antibody and examined for microvessels (see Lamszus er
al., 1997). Peak MYC represents the number of microvessels per
field (40X objectivc, 10X ocular) in the region of most active tu-
mor angiogenesis. Average peak MYC is the mean MYC of the
3-4 fields with the highest microvessel counts. Yalues listed are
means ± standard errors (number of tumors assayed). Compari-
sons of pooled clones (21+29) vs (32+34) gave P < 0.00 I (peak
MYC) and P < 0.00 1 (average peak MYC).

in SF+ vs SF- tumorextract s (see Tabl e 11); d ) chemot ac tic a nd a ng ioge nic a c tivi ty in SF+ tu -
m or ex tra c ts was not ex p la ined b y seco ndary c hang e s in le vel s of ot her re gulat o rs (vasc u lar
e ndo the lial g rowth factor a nd th rombo spondin-l ); a nd e) th o se ac tivi ties we re neutra lized by
a n a n ti-SF mono cl onal ( La mszus e t a l., 19 97). Thu s, the increased g ro w th ra tes of SF+
M DAM B23 1 tumors a p pea rs to be due, in part, to a n a ngiogenic re sp onse ind uced b y SF.
In th e a bove d e scrib ed st udy, we a lso exami ne d th e a bi lity ofSF to sti m u la te reg io na l
d is semin ation of b rea st cancer cells to ax illary Iy m ph nodes and distant m et a st a si s to th e
lung . A nima ls with SF+ tu m o rs had signi ficant ly h ighe r ra te s ofaxillary no da l meta st a si s
th an did a nima ls wi th SF- tumo rs ( La mszus e t al. , 19 9 7 ). However, it is no t c1ea r if th is was

Table 11. Angiogenic responses indu ced by scatter

factor-tra nsfe cted (SF +) and co ntro l-transfected (SF -)
MDA MB23 \ hum an breast ca ncer tumor extract s '
Corneal neovascularization,
proportion of positive
Test sampie responses (%)
Controls
Hydron + PBS 0/4 (0)
SF (50 ng) 4/4 ( 100)
M0 294 (500 ng) 0/3 (0)
SF (50 ng) + M0294 (2.5 ug) 0/3 (0)
Tumor extracts ± M0294
SF+ Clone 2 1 (5 u g) 4/4 ( 100)
SF+ Clone 2 1 + M0294 1/4 (25)
SF- Clone 32 (5 ug) 1/3' (33)
SF- Clone 32 + M0294 1/4' (25)
'Abbre viations: SF = recombinant human scatter factor; M0294 =
anti-SF neutralizing monoclonal; PBS = phosp hate-b uffe rcd
saline. Assay methods and response criteria are described in
Lamszus et al. ( 1997). Neovascular responses induced by SF-
tumor extracts without or with M0294 were substantially weaker
in intensity than responses induced by SF+ extracts.

423
related to the presence ofSF or simply to the larger size ofthe SF+ tumors . There were no de-
tectable pulmonary metastases in animals with SF+ or SF- tumors .
In a prior study, we reported that EMT6 mouse mammary tumor cells treated in vitro
with SF gave rise to significantly more pulmonary nodules following intravenous injection
into isogeneic Balb/c-Rw mice than did control cells (Rosen et al., 1994b). The SF-treated
EMT6 cells did not exhibit any increases in cell size or in the in vitro rates ofDNA synthesis,
cell proliferation, and clonogenic efficiency as compared with control cells . However, these
cells did exhibit significantly increased cell motility, expression ofthe matrix-degrading en-
zyme urokinase on the cell surface, invasion through a reconstituted matrix of basement
membrane (Matrigel), and adhesiveness to certain subastrates (collagen land laminin) (Ro-
sen et al., 1994b). Thus SF may enhance a number oftumor characteristics associated with a
more invasive and metastatic phenotype.

Rat and Human Gliomas

In addition to the breast cancer model, we have developed several rodent models for
the orthotopic growth of glial tumors . We found that 9L rat glioma cells transfected with hu-
man SF cDNA grew much more rapidly than control-transfected cells both as brain tumors
and as subcutaneous tumors in syngeneic host rats (Laterra et aJ., 1997). Analysis ofbrain tu-
mor sections using a digital image analysis system revealed significantly higher microvessel
densities in SF+ tumors than in SF- control tumors . SF+ tumor s also showed significantly
higher expression ofurokinase, an enzyme required for proteolytic degradation of extracel-
lular matrix and invasion ofmigrating endothelial cells during the angiogenic process. In par-
allel studies, SF also induced increased growth and microvessel formation in brain tumors
from SF-transfected human glioblastoma cells (U373MG) grown in athymic nude mice , as
compared with control U373MG cells (Laterra et aJ., in press) . These findings suggest that
overexpression ofSF stimulates in vivo tumor growth, in part by conferring an angiogenic tu-
mor phenotype. In other studies, we found that SF stimulates chemotactic migrations, prolif-
eration, and DNA synthesis in various cultured cell strains of human neuromicrovascular
endothelial cells derived from temporal lobe biopsies, consistent with the idea that SF may be
an authentic angiogenesis factor in the central nervous system (Rosen et al., 1996a; Lamszus
et aJ., submitted).

SUMMARY

Scatter factor (SF) [also known as hepatocyte growth factor (HGF)] is a multifunc-
tional cytokine that stimulates the motility and invasiveness ofvarious types ofepithelial and
carcinoma cells . These responses are mediated by the c-Met proto-oncogene product, a trans-
membrane tyrosine kinase receptor. Several years ago, we showed that SF is a potent angio-
genic molecule, and that its angiogenic activity is mediated primarily through direct actions
on vascular endothelial cells. These include stimulation of cell migration, proliferation,
urokinase production, invasion through basement membrane, and organization into
capillary-like tubes. More recently, we have found that SF is overexpressed in high grade and
invasive human tumors relative to low grade and non-invas ive tumors or to benign lesions.
These cancers include breast carcinoma, transitional cell carcinoma ofbladder, and primary
brain tumors (gliomas). For invasive breast cancers, the content ofSF was strongly correlated
with the content ofvon Willebrand 's factor (an endothelial cell marker), suggesting a rela-
tionship between the SF content and the overall level of tumor vascularity. Furthermore,
transfection ofhuman breast cancer or glioma cells with SF cDNA greatly enhanced their or-
thotopic growth as tumors in syngeneic or immunocompromized animals. The increased
growth rate of the SF-transfected tumor cells was attributable, in part, to increased tumor

424
angiogenesis, as demonstrated by significantly higher tumor microvessel counts in these tU-
mors as compared with vector-transfected control tumors. These findings suggest that SF
may function as a tumor progression factor, in part, by stimulating tumor cell invasiveness
and, in part, by stimulating an ang iogenic response in the host stroma.

ACKNOWLEDGMENT

This research was supported in part by grants from the United State s Public Health
Service (CA-64869 and CA-644l6), the American Cancer Society (EDT-l 02), and the Ch il-
dren 's Brain Tumor Foundation ofNew York. We are indebted to Dr. Ralph Schwall, Genen-
tech, Inc. (South San Francisco, CA) for providing sheep antiserum against SF as weIl as
recombinant human two-chain human SF. We are grateful to Drs. Donald Bottaro, Stephen
Stahl, and Paul Wingfield, Laboratory ofCeIlular and Molecular Biology, National Cancer
Institute (Bethesda, MD) for provid ing recombinant human truncated SF peptides (NKI and
NK2) .

REFERENCES

Ben Av P,Crofford LJ, Wilder RL, Hla T. Induction ofvascular endothelial growth factor exp ression in sy novial
fibroblasts by pro stagiandin E and interl eukin-I : a potential mechanism for inflammator y ang iog ene-
sis . FEBS Let1372: 83-87, 1995 .
Bosari S, Lee AK , DeLellis RA, Wiley BO, Heatley GJ, Silverman ML. Microvessel quantitation and prognosis
in invasive breast carcinoma. Hum Patho!23 : 755- 76 1, 1992.
Bottaro OP, Rubin JS , Faletto DL, Chan AM-L , Kmiecik T E, Vande Woud e GF, Aar on son SA . Identificati on of
the hep atocyte growth factor receptor as the c-met proto-oncogene product. Science 251 : ~W2-804 ,
1991 .
Chan AM-L Rubin JS , Bott aro OP, Hirschfield DW, Chedid M, Aaronson SA : Identification of a competitive an-
tagoni st encoded by an alternative transcript. Science 254 : 1382-1385, 1991 .
C ioc e V, Csaky KG , Chan AM-L , Bottaro OP, Taylor WG, Jemsen R, Aaronson SA, Rubin JS . HG F/NK I is a
naturally occurring HGF /SF variant with partial agon ist/ant agoni st activity. J Bio! Chem 271 :
13110-13115,1 996 .
Dinarello. Biologie basis for interleukin-I in disease. B!ood 87 : 2095-2147, 1996.
Fan TP, Hu DE, Guard S, Gresham GA, Watling KJ. Stimulation of angiogenesis by substance P and
interleukin-I in the rat and its inhibition by NK 1 or inter leukin- I receptor antagonists. Brit J Pharma-
co! 110: 43--49, 1993 .
Gon zatti-Haces M, Seth A, Park M, Copeland T, Oroszlan S, Vande Woude GF. Characterization ofthe TPR-
METoncogene p65 and the MET protooncogene p 140 protein tyrosine kinase s. PNAS USA 85 : 21-25 ,
1988 .
Grant DS, Kleinman HK, Goldberg 10, Bhargava M, NickoloffBJ , Polverini P, Ros en EM. Scatter facto r in-
duces blood vessel formation in vivo . PNAS USA 90 : 1937- 194 1, 1993.
Jin L, Fuchs A, Schnitt SS , Yao Y,Jos eph A, Lamszus K, Park M, Goldberg 10, Rosen E M. Express ion ofSF and
c-m et receptor in benign and malignant breast tissue. Cancer 79: 749-760, 1997 .
Jin L, Yuan R-Q , Fuchs A, Yao Y, Jo seph A, Schwall R, Guida A, Schnitt SJ, Hastings HM, Andres J, Turkel G,
Polverini PJ, Goldberg 10, Rosen EM. Expre ssion oflL-l ß in human breast cancer. Cancer. in press .
Joseph A, Weiss GH, Jin L, Fuch s A, Chowdhury S, O'Shaughnessy P, Goldberg 10, Ro sen EM. Ex pression of
scatter factor in hum an bladder carcinoma. J Nat! Cancer !nst 87 : 372-377 , 1995 .
Lamszus K, Jin L, Fuch s A, Shi YE , Chowdhury S, Yao Y, Polverini PJ, Goldberg 10, Rosen EM. Scatter factor
stimulates tumor growth and tumor angiogenesis in human breast cancers in the mammary fat pads of
nude mice . Lab lnvest 76: 339-353 , 1997 .
Lamszus K, Schmidt NO, Jin L, Laterra J, Zag zag 0, Way 0, Witte M, Weinand M, Goldberg 10, Westfal M, Ro-
sen EM . Scatter factor prornotes mot ility of human glioma and neuromicrovascular endothel ial cells.
Submitted for publication.
Laterra J, Rosen E, Nam M, Rangathan S, Fielding K, Johnston P. Scatter factor/hepatocyte growth factor ex-
pression enhances human glioblastoma tumorigenicity and growth. Bioch em Biophys Res Commun, in
press.

425
Laterra J, Nam M, Rosen EM, Rao JS, Lamszus K, Johnston P. Scatter factor/hepatocyte growth factor gene
transfer enhances glioma growth and angiogenesis in vivo. Lab lnvest 76: 565-577,1997.
Lokker NA, Mark MR, Luis EA, Bennett GL, Robbins KA, Baker 18, Godowski PJ. Structure- function analysis
ofhepatocyte growth factor : Identification ofvariants that lack mitogenic activity yet retain high affin-
ity receptor binding. EMBO J 11: 2403-2410, 1992.
Mark MR, Lokker NA, Zioncheck TF, Luis EA, Godowski PJ. Expression and characterization ofhepatocyte
growth factor receptor-lgG fusion proteins . Effects ofmutations in the potential proteolytic c1eavage
sites on processing and ligand binding . J Biol Chem 267 : 26166-26171, 1992.
Naidu YM, Rosen EM, Zitnik R, Goldberg I, Park M, Naujokas M, Polverini PJ, NickoloffBJ. Role ofscatter
factor in the pathogenesis of AIDS-related Kaposi's sarcoma. PNAS USA 91: 5281-5285,1994.
Rosen EM, Goldberg ID: Scatter factor and angiogenesis. Adv Cancer Res 67: 257-279. 1995.
Rosen EM, Goldberg ID. Scatter factor as a potential tumor angiogenesis factor. In: "Molecular, Cellular, and
Clinical Aspects of Angiogenesis", Maragoudaikis ME, editor, NATO ASI Series , Series A: Life Sci-
ences, Vol. 285, Plenum Press, New York, 1996, pp. 85-94.
Rosen EM, Grant D, Kleinman H, Jaken S, Donovan MA, Setter E, Luckett PM, Carley W. Scatter factor stimu-
lates migration ofvascular endothelium and capillary-like tube formation. In: "Ce 11 Motility Factors",
Goldberg ID, Rosen EM, eds ., Birkhauser-Verlag, Basel. 1991a, pp 76-88.
Rosen EM, Jaken S, Carley W, Setter E, Bhargava M, Goldberg ID. Regulation ofmotility in bovine brain endo-
thelial cells . J Cell Physio/146 : 325-335 , 1991b.
Rosen EM, Joseph A, Jin L, RockweIl S, Elias JA, Knesel J, Wines J, McClellan J, Kluger MJ, Goldberg ID, Zit-
nik R. Regulation of scatter factor production via a soluble inducing factor. J Cell Biol 127: 225-234,
I994a .
Rosen EM, Knesel J, Goldberg ID, Bhargava M, Joseph A, Zitnik R, Wines J, Kelley M, RockweIl S. Scatter fac-
tor modulates thc metastatie phenotype of the EMT6 mouse mammary tumor. Int J Cancer 57:
706-714, 1994b.
Rosen EM, Nigam SK, Goldberg ID. Mini-Review: Scatter factor and the c-met receptor: A paradigm for mes-
enchymal :epithelial interaction. J Cell Bio1127 : 1783-1787, 1994c.
Rosen EM, Joseph A, Jin L, Yao Y, Chau M-H, Fuchs A, Gomella L, Hasings H, Goldberg ID, Weiss GH. Uri-
nary and tissue levels ofscatter factor in transitional cell carcinoma ofbladder.J UroI157 :72-78, 1997.
Rosen EM, Laterra J, Joseph A, Jin L, Fuchs A, Way D, Wilte M, Weinand M, Goldberg ID. Scatter factor ex-
pression and regulation in human glial tumors . Int J Cancer 67: 248-255, 1996a.
Rosen EM, Polverini PJ, NickoloffBJ, Goldberg ID. Role ofscatter factor in pathogenesis of AIDS-related Ka-
posi sarcoma. In: "Molecular, Cellular, and Clinical Aspects of Angiogenesis", Maragoudaikis ME,
editor, NATO ASI Series , Series A: Life Sciences, Vol. 285, Plenum Press. New York, 1996b , pp.
181-190.
Stoker M, Gherardi E, Perryman M, Gray 1. Scatter factor is a fibroblast-derived modulator ofepithelial cell mo-
bility. Nature 327 : 238-242, 1987.
Tamura M, Arakaki N, Tsoubouchi H, Takada H, Daikuhara Y. Enhancement ofhuman hepatocyte growth factor
production by interleukin-I alpha and -I beta and tumor necrosis factor-alpha by fibroblasts in culture.
J Biol Chem 268: 8140-8145,1993 .
Toisma SS, Volpert OV, Good DJ, Frazier WA, Polverini PJ, Bouck N. Peptides derived from two separate do-
mains ofthe matrix protein thrombospondin-I have antiangiogenic activity. J Cell Bioll22: 497-511,
1993.
Tuck AB, Park M, Sterns EE, Boag A, Elliott BE. Coexpression ofhepatocyte growth factor and receptor (Met)
in human breast carcinoma. Am J Pathol 148: 225-232, 1996.
Wang Y, Seiden AC, Morgan N, Stamp GW, Hodgson H1. Hepatocyte growth factor/scatter factor expression in
human mammary epithelium. Am J Pathol144: 67~82, 1994.
Weidner N, Sempie JP, Welch WR, Folkman J. Tumor angiogenesis and metastasis - correlation in invasive
breast carcinoma. New Engl J Med 324: 1-8, 1991.
Weidner N, Folkman J, Pozza F, Bevilaqua P, Allred EN, Moore DH, Meli S, Gasparini G. Tumor angiogenesis:
a new significant and independent prognostic indicator in early stage breast carcinoma. JNCI 84:
1875-1887,1992.
Yamashita J, Ogawa M, Yamashita S, Nomura K, Kuramoto M, Saishoji T, Sadahito S. Immunoreactive hepato-
cyte growth factoris a strong and independent predictor ofrecurrence and survival in human breast can-
cer. Cancer Res 54: 1630-1633,1994.
Yao Y, Jin L, Fuchs A, Joseph A, Hastings H, Goldberg ID, Rosen EM. Scatter factor protein levels in human
breast cancers : C1inieopathologic and biologie correlations. Am J Pathol 149: 1707-1717, 1996.

426
Clinical Applications
THE VASCULARIZATION OF EXPERIMENTAL AND HUMAN PRIMARY
TUMORS: COMPARATIVE MORPHOMETRIC AND MORPHOLOGIC
STUDIES

M.A. Konerding ', E. Fait l , A. Gaumann-, Ch. Dirnitropoulou' , and W.

Malkusch'

Institute of Anatomy, Johannes Gutenberg-University Mainz, D-55099

I

Mainz,
2Institute ofPathology, University Clinics , Mainz
3Kontron Elektronik GmbH , Oskar-van-Miller Str. 1, D-85385 Eching

Significance of the tumor vascular system. The importance of the blood vessel
system in solid tumors has given rise to an increasing interest in this system as a direct
target for tumor therapy, i.e. vascular targeting (Denekamp, 1984) . Furthermore, its
importance as a route for delivery of anticancer drugs (chemo- and immunotherapies) or
photosensitizers (photodynamic laser therapy), as well as its modulatory influences on
radiotherapy and hyperthermia - the former greatly depending on the amount of oxygen
available, the latter on heat transfer - are evident. Numerous studies on the energy
metabolism of solid tumors (Vaupel et al., 1987, 1989) have pointed out the functional
importance of the blood vessel system and stress the need for further thorough investi -
gations of its functions in terms of transport capacities for nutrients, oxygen, catabolite
removal, delivery oftherapeutic substances, and heat transfer.

Tumor blood vessels derive from two sourees, namely from pre-existing host vessels,
and newly formed ones from tumor angiogenesis (neoangiogenesis). Tumor angiogenesis
is defined as the growth ofnew vessels from pre-existing (mature) ones - exclusively from
capillaries and/or venules - by means of sprouting as well as dilatation and elongation of
established mature and newly evolved immature ones . For selected reviews on the role of
endothelial cells, extracellular matrix (ECM) components and tumor cells in tumor
angiogenesis, see Klagsbrun and D'Amore (1991), Carey (1991), Fajardo (1989), Folkman
(1985), and Folkman and Shing (1992) .

Scanning electron microscopy of corrosion casts in the study of tumor

vascularity. The vasculature in solid tumors is most frequently studied using the

Angi ogenesis: Models, Modulators, and Clinical Applications

Edited by Maragoudaki s, Plenum Press, New York, 1998 429
Fig. 1: Schematic drawing of the architectonic characteristics of the tumour vascular
system after removal of parts of the vascular capsule of a human xenotransplant. Note the
dilated major vessels (m) with circumferential and centripetal courses, constrictions
(arrows) and tortuosities (arrowheads). These vessels undergo transition to peripheral and
central vessels without any hierarchy, whereby avascular regions (H) are seen next to weil
vascularised ones. Evasates (e) are also seen in part.

transparent rabbit ear chamber, the hamster cheek-pouch , the rabbit cornea, and the chick
chorioallantoic membrane (CAM) assays (for reviews see Auerbach et al., 1991; Folkman ,
1985) . These allow for intravital light microscopy and enable detailed analyses of
sprouting events to be made like in immature tissue (Rhodin and Fujita, 1989). Despite the
tumor biological and therapeutic implications, studies on primary tumors and comparisons
of different tumor models, which confirrn their reliability and validity, are still scarce.

At least partly due to the limitations of the methods used, our knowledge of the
vascular architecture is still limited. Intravital microscopy is effective in the observation of
two-dirnensional vascular networks as given in most angiogenesis assay systems.
However, it cannot visualize the tremendous heterogeneity of the tumor vascular system

430
found even in different areas of the same tumor. Classical injection methods (Spalteholtz,
1914) with light microscopic evaluation are tirne -consuming, require laborious
reconstruction work and are still of limited value, unless sophisticated computerized
techniques are used.

Scanning electron microsocpy (SEM) ofmicrovascular corrosion casts as described as

early as in 1971 by Murakami can overcome this problem . SEM of corrosion casts is a
powerful method providing information on the geometry of the tumor vascular system in
terms of a network defined primarily by the number of vessels, their branching behaviour
(modes, angles , frequencies, and interbranching distances), and their characteristic course
as a whole (Fig . 1-3). The microvascular corrosion cast/SEM method has proven to be an
extensive, powerful method to convincingly document the 3-D arrangement ofthe vascular
system with a) an high depth of focus , b) a reasonably high resolution allowing
visualization of the microvascular bed as weil as the reliable differentiation between
arteries and vcins (Miodonski et al., 1976). and c) comparatively little risks in producing
artifacts (Lametschwandtner et al., 1989; Konerding, 1991).

Several studies using scanning electron microscopy of microvascular corrosion casts

have shown that this method is suitable for studies of blood vessel development both in
physiological vascular systems, e.g., the uterus or the placenta (Kaufmann et al., 1985) and
in tumors (Christoffersson and Nilsson, 1992). However, until now published work on the
tumor vascular system as revealed by the vascular corrosion cast/SEM method is mainly
descriptive. The terms used are subjective, qualitative, and sometimes prone to
misunderstandings and misinterpretations. For reference of all relevant papers on tumor
microvascularisation using corrosion casting, see Konerding et al. (1995) .

Recently we have shown that numerous parameters defining the microvascular unit
such as the diameter, length, surface, volume, branching pattern, course and looping can be
easily quantified in SEM micrographs provided that stereo images with an exactly defin ed
tilting angle and infinite working distances were made. After digitization ofthe images, the
information of depth can be calculated for each interactively defined measuring point using
the common parallax (Malkusch et al., 1995). Reproducibility tests and tests for greatest
error possibilities resulted in a maximum deviation of 2.5% to bc expected (Malkusch et
al., 1995).

This methodology can also be applied in human material using biopsy material ,
surgical specimens or even in cadavers (Konerding, 1991). It should be noted that the
injection ofthe casting medium does not impair subsequent histopathological diagnosis. In
arecent study using human large intestine segments we have seen that histopathology
ineluding light- and transmission electron microscopy as weil as immunostaining is not
affected by the rinsing, fixation and casting ofthe vascularity . However, the preparation of
regionary Iymph nodes is more laborious since the surrounding tissue contains vessels
filled with the casting medium.

Characteristics and remodeling of the tumor vascular system. As pointed out

earlier in detail (Konerding et al., 1992a, 1995), in experimental and primary tumors a
elose side by side of undifferentiated, newly formed and destroyed or compressed vessels
can be seen in tumors at all times. A real differentiation into morphological distinct
arteries, veins and capillaries with typical vesscI wall structure does not take place . The

431
432
Fig. 2a-c: Yascular architecture and remodeling of the capsular tumour vessels. a: The
tumor vascular envelope consists of multilayered , plexus-like vascular system consisting
mainly of sinusoidal vessels devoid of all hierarchy spreads out from several venous majo r
vessels (m), which are in part flattened (H). e=evsates. Note the varying diameters. b:
Instead of sinusoidal vessels, more highly differentiated vessels are found 15 days after
transplantation originating from preformed host vessels . Here, two arteries (a) and a vein
(v) are shown . The newly formed vascular corona appears to have a higher degree of
hierarchy despite the atypical course . c: On day 23, the vascular capsule shows a higher
degree of differentiation. Note the capillary patterns (c) as weil as the filling artifacts
(arrows). Squamous cell carcinoma 4197; a, 4 days p.t., magn .: x 65; b, 15 days p.t.,
magn .: x 55; c, 23 days p.t., magn.: x 45.

most apparent architectonic characteristics are the chaotic lack of hierarchy and
heterogeneous vascular densities. In xenografts and experimental tumors frequently a
tumor vascular envelope can be seen formed by pre-existing host vessels . The majority of
the vessels have a capillary-like sinusoisal vessel wall construction with flattened endo -
thelium and low structural stability . The lack of contractile cell elements explains for the
missing blood flow regulation and numerous discontinuities and new sprouts for frequent
evasates. Sinuoids originating and draining into large caliber veins show that the
intravascular p02 must be far lower than in normal tissues' vascularities. Fig. 1 summarizes
the typical features.

Areal remodeling ofthe tumor vasular system exists only in the periphery . Sequential
studies of xenografted human tumors on nude mice show that practically all vascular
features can be seen to different extent at all times after transplantation. the vessels of the
tumor vascular envelope (Figs. 2a-c) form a multilayered plexus-like system consisting of
sinusoidal vessels on day 3 after transplantation which is remodeled later on showing at
least a certain degree of hierarchy . Cork screw like vessels indicating for new vesel
formation, as frequently seen in the tumor periphery (Figs. 3a-c), are not replaced in later
stages as in all other kinds of secondary angiogenesis . The sinusoidal plexuses seen in the
tumor center (Figs. 4a-c) remain widely unchanged at all times. Comparisons of
morphological features

Tumor-type specific vascular structure. In a previous study (Konerding et al.,

1992b) we examined the question as to whether individual tumor entitites express tumor
type specific or characteristic structural features . Transmission electron microscopy of 8
different human tumors (undifferentiated amelanotic melanoma, 2 leiomyosarcomas, soft
tissue sarcoma, 2 spindie cell sarcomas,neurofibrosarcoma, squamous cell carcinoma)
transplanted onto 80 nude mice showed that the exam ined parameters (continuity of
endothelium, gaps, fenestrations, endothelial cell organelle contents, contact structure
differentiation, basal lamina, pericytes, microthromba etc.) were not significant differently
expressed in the individual entities and cell lines . We concluded that it was thus
permissible to extrapolate the results of therapeutic measures aimed at the vessel system to
other entities (Konerding et al., 1992b) This was, according to our today's knowledge, at
least in part a wrong conclusion , since it did not take into account the tumor type specific
differences in the vascular architecture, which morphologically determines the tumor blood
flow, oxygenation and metabolite removal.

433
434
Fig. 3a-c: Architecture and remodeIing of peripheral tumor vessels. a: Subcapsular
capillaries arising from a large vein (v) with early stages of sprout formations in form of
numerous protrusions (H) and furrows (arrowheads). Note the numerous blind ends with
variable diameters. b: Cork screw-like tortuous subcapsular vesseI with circular furrows
(arrows) and several sprout formations (arrowheads), which are clearly discernible from
evasate (e) . a: artery . c: On day 15, a dense sinusoidal system can be demonstrated
subcapsularly around retained tortuous vesselsüt). No differentiated arteries supplying this
system can be seen. Squamous cell carcinoma 4197, a: 4 days p.t., magn.: x 745; b: 7 days
p.t., magn. : x440; c: 15 days p.t., magn.: x 130.

Evidence for a tumor type specific vascular architecture. Up to now the question
of the specificity of the tumor vascular system already posed in the last century has not
been finally resolved: Lewis (1929) was of the opinion that each tumor formed its own
characteristic vascular system . According to Lindgren's (1945) findings, tumors develop a
specific capillary angioarchitecture to an extent which was not more closely defined . Only
undifferentiated tumors showed such an uncharacteristic angioarchitecture that it was not
possible to determine the origin of the tumor solely from the vascular anatomy. Other
authors, however, contended that the vascular morphology of a certain tumor was
characteristic for, but not necessarily unique for this tumor (Shubik, 1982; Vaupel and
Gabbert, 1986).

Comparisons of the intervascular and interbanehing distances of three different

murine tumors (two carcinomas and a sarcoma) using the above mentioned techniques
prove that the architectural differences are highly significant (Figs 5a, b). No differences
were found in the branching angles (Fig. Sc). From this experiment, we conclude that
tumors of an individual cellline develop a specific vascular architecture .

Individual factors may alter the tumor type specific vascular architecture. A
variety of angiogenic and antiangiogenic factors determine the structure and architecture of
the tumor vascularity. Among those, basic fibroblast growth factor (bFGF) is one of the
most potent and widely investigated angiogenic factors (Folkman, 1995; Dellian et al.,
1996). Coltrini et al. (1995) obtained different bFGF-expressing clones after transfection
with an expression vector harboring a human bFGF cDNA under the control of the human
ß-actin gene promoter. One ofthe clones (bFGF-B9 clone) showed the capacity to secrete
significant amounts of the growth factor . This clone formed highly vascularized tumors
growing faster than the non-bFGF releasing clones and the parental cells when injected s.c.
into nude mice (Coltrini et al., 1995).

In a collaboration we examined tumors derived from this bFGF secreting clone and
compared it to the tumors derived from non BFGF secreting clones (Konerding et al.,
unpublished findings) . Comparisons ofthe intervascular and interbranching distances and
branching angles of the tumors originated from the different bFGF clones did not reveal
significant interindividual differences . However, group comparisons ofthe casted luminal
vessel diameters revealed significant differences between bFGF secreting tumors and the
controls. The most striking differences among the individual groups were oberved when
the variability of the luminal vascular diameters was considered (Fig. 6a, b). Within
individual vessel segments, i. e. in the vessel course between two ramifications, only Iittle
changes of the diameters were seen in the non-bFGF secreting tumors (Fig 6a), whereas

435
436
Fig. 4a-c: Central tumor vessels . a: Irregular, multiform sinusoids originating from large
calibre venous vesse1s (v) extending from the capsule into the centre and which are often
flattene d (H) . b : Central sinusoidal vessels forming a heterogeneous network with
numerous extravasates (e). Note the variability of vesse1 diameters (arrows) . c: During the
course of time, the sinusoids remain unchanged in course and shape at the corresponding
sites, extrava sates however occur more frequently (e). Squamous cell carcinoma 4 197, a: 4
days p.t., magn .: x 160; b: 15 days p.t., magn.: x 160; c: 23 days p.t., magn.: x 210.

bFGF secreting tumors showed a strikingly high variability of the vessel diameters (Fig.
6b). The differences proved to be significant on the p_O.OOOI level. From this, it can be
concluded that bFGF does not only increase the angiogenicity and tumorigenicity, but also
affects the tumor microvascular architecture. The impact of other angiogenic and
antiangiogen ic factors on the vascular architecture was not yet studied up to now.

Tumor type speci fic architect ures exist in primary tumors, too. Adenocarcinomas
grown in the large intestine (colon, sigmoidal colon and rectum) displaya typica l vascu lar
architecture irrespective of the localisation of the tumor within the large intestine (Fig . 7a-
d). The normal honeycomb pattern ofthe mucosal capillary plexus (Fig. 7a) is not retained
in these tumors . Only in the peripheral areas adjacent to the normal, not invaded tissue the
previous hexagonal pattern is retained despite the high angiogenic activity (Fig . 7b). The
tumor center, howe ver, is made up by tumor vessels displaying all the features described
above in experimental or xenografted human tumors (Fig. 7c, d).

14 c CaX
• SAS
12
• Ca T

10

. 6

o 18 36 54 n 90 108 126 144 162 180

branching angle (0)

Fig. 5 a-c: Morphometry of vascular parameters of different tumors . a: logarithmic

distribution of intervascular distances, b: logarithmic distribution of interbranching
distances, c: branching angles . All values are smoothed values. The differences in a and b
are significa nt (p=O .OOO 1).

437
 20
c Cax
18 • ' A'

16

14

12

10
~

·2
10 1110 1000
interv ascular dist an cc (u m)

20
n Cax
18 • S S
16 • C.....U

14

12

10
~

·2
10 100 1000
intcr bra nching distancc (u m)

438
 a 100 -,

G; 80
~ 60
'"
i5
c
40
'"'"
E
20

=:'" o
ö -20
c
Q -40
<ii
.S:

'"
u
-60
~ -80 1
-100 ~

b 100 ~

80
G;
Q;
"1
°0 0 0
•
!.
.. , '.

:~
E
'"
i5
c
'"'E" 2~ ]
=:'"
ö -20
c
0
.~ -40
.S:
u'" -60
~ -80
- 100 ~

F ig. 6a and b : Changes of the vessel diameters within individual vessel segments
expressed as percental deviation from the individual mean vessel segment diameter. The
individual vessel segment was defined as the segment between two con secutive
branchings. Each point visualizes the percental decrease or increase from the mean vessel
segment diameter. The mean vessel diameters were normalized to zero . For each group,
110 vessel segments were measured at five distinct points each, resulting in 550 values per
group. a : bFGF-A8 tumors with no secretion of bFGF show a lower variab ility of the
vascular diameters than bFGF-B9 tumors with high secretion levels (b).

Morphomelric evaluations of the tumor corrosion casts show that the intervascular
distances clearly differ between control and tumor. Further rnore, the vascularisation in the
peripheral tumor surface is far denser with mean intervascular distances of 52.2 um
compared to 105 um in the adjacent control tissue (Table 1, Fig. 8a). The central , luminal
tumor surfaces of the colonic adenocarcinomas and the tumor centers also show highl y
significant differences in intervascular distances (Table 2). These differences within the
tumor areas exist also for the interbran ching distances (Fig. 8b). On contrast, the branching
angles are similar in all parts of the tumor, but different from the control tissue (Fig. 8c).
The vascular diameters are bigger in the tumor center than in the peripheral and central
surface as weil as in the control mucosa (Fig. 8d). For the individual mean values , standard
deviations, minima and maxima see Table 1, for the significances Table 2.

Summing up, this paper shows that studies on the microvascular architecture can be
best accomplished using scanning electron microscop y of corrosion casts, both in primary

439
Table 1: Mean values, standard deviations (SD), minima (min) and maxima (max) as weil
as the number of measurements done in the tumors depicted in Fig. 8a-d.
Intervascular distances (IJm) Interbranching distances (IJm)

contro l PTS CTS TC control PTS CTS TC

mean 105.0 52.2 112.2 180.1 55.0 49.8 74.5 132.0

SO 23.3 34.1 88.4 154.9 29.4 28.6 59.1 100.7

min 8.9 3.9 8.5 20.0 2.2 7.0 6.7 17.3

max 181.2 494.6 551.9 1056.3 238.5 240.5 576.1 733.9

measurements 773 2291 1353 1013 1662 1911 1911 983

Branching angles n Vessel diameter (IJm)

control PTS CTS TC control PTS CTS TC

mean 87.8 73.6 72.1 76.1 12.2 19.6 17.8 30.9

SO 28.6 33.1 32.3 34.8 2.0 6.6 8.2 23.3

min 15.0 1.8 1.5 4.4 6.6 6.5 2.8 1.5

max 161.7 175.6 166.3 165.7 20.9 46.8 84.5 161.6

measurements 803 1194 1194 874 2500 2500 2500 2500

Table 2: Significances of the individual measurements listed in Table 1. IVA =i

ntervascular distance, ffiRA = interbranching distance, BRA = branching angle , DIA =
diameter.

IVA IBRA BRA DIA

. control PTS p<0 .OOO1 n.s, p<O.OOO1 p<O.OOO1

control CTS p<O.OOO1 p<O.OOO1 p<O.OOO1 p<O.OOO1

control TC p<O.OOO1 p<O.OOO1 p<O.OOO1 p<O.OOO1

PTS CTS p<O.OOO1 p<O.OOO1 n.s. p<O.OO5

PTS TC p<O.OOO1 p<O.OOO1 n.s. p<O.OOO1

TC CTS p<O.OOO1 p<O.OOO1 n.s. p<O.OOO1

440
441
Fig. 7 a-d: Microvascular corrosion casts of the human colonic mucosa (a) and of tumors
ofthe large intestine (b-d). a was obtained from the ascending colon of a 75 year old male
10 cm distal of a colonic carcinoma. The mucosal capillary plexus is arranged in a
honeycomb pattern . Arrows mark numerous intercapillary connections . The supplying
arteriole (fat arrow) and draining veins (star) take a straight course to and from the
underlying submucosal vessels (rounded star). b: Peripheral tumor surface of a rectum
carcinoma (pT3 , pNO, pMx , G2). c: Central tumor surface of of a sigmoidal
adenocarcinoma (pT3, pN3, pMl , G2). d: Vessels of the tumor center in a rectum car-
cinoma (pT2, pNO, pMx, G2). Note the absence of the normal vascular pattern as seen in
the control. Bars = 200 um

442
 35 i 25
--0-C
30 -:1 .......&.- PT<:: I -Ä- PTS
- CTS
25 -+- TC -+- TC
~
0 ~
e, 15
;: 20
o
c o>-
c
"1
CI) CI)
;:) ;:)
CT
15
CT 10
~ ~
10
5
5

o , .-- , ~ ==sr • o f ......1'8 0

. I
10 20 40 100 250 500 900 10 20 40 100 250 500 900

Inlervascular dislance (~ m) Interbranching distance (prn)

t
v>
 25 35
-0-
C
--- PTS 30 r~ PTS
___ c
20 -l _ CTS
-CTS
25 -+- TC
;;e
e, ;;e
15 J ~ TC 0
>- ;:: 20
g u
Ql c:
::J Ql
CI' ::J 15
10 CI'
~ ~
10

I I \ .,. 1 0
10
:1-4 20 30 50 80

branch ing ang le (0)

120 180 270 1 2 4 7 12 20 30 50 80 120 180

vesse l diameter (IJm)

Fig. 8 a-d: Common logarithmic distribution of the intervascular (a) and interbr anching distances (b), bran ching angles (e) and vessel diameters
(d) in ten human colonic and rectal carcinomas. Th e data were pooled since no significant interindividual differences were observed. Note the
differences between thc control tissue (C), the per ipheral tum or surface (PTS), centra l tum or surface (CTS) and the tumor center (TC).
tumors as weil as in experimental tumors. Experimental and primary tumors have many
features in common and they induce a tumor type inherent specific vascularity.

ACKNOWLEDGEMENT

This study was supported by the European Community within the framework of the
Human Capital and Mobility Program "Mechanisms for the Regulation of Angiogenesis"
(ERB CHRXCT 940593) and by the Robert- Müller-Stiftung for Cardiovascular Research
Mainz.

The authors thank the Vascular Targeting Group ofthe Gray Laboratory, Northwood,
England, for providing the tumors shown in Fig. 5 and M . Presta, University of Brescia,
Italy, and R . Giavazzi, Ist ituto Mario Negri , Bergamo, Italy, for providing tumors grown
from bFGF transfected cells .

REFERENCES

Auerbach R ., Auerbach W., and Polakowski 1., 1991, Assays for angiogenesis: a review .
Pharmacol. Ther., 51: 1-11.

Carey DJ., 1991, Control ofgrowth and differentiation ofvascular cells by extracellular
matrix proteins. Annu.Rev.Physiol., 53 : 161-177.

Christofferson R. H ., Nilsson B. 0 ., 1992, Microvascular corrosion casting in angiogenesis

research. In : Motta P. M ., Murakami T., Fujita H. (Eds .) Scanning electron Microscopy of
Vascular Casts: methods and Applications. 27-37 . Kluwer Academic Publishers, Boston-
Dordrecht-London.

Coltrini D ., Gualandris A., Nelli E. E., Paroli ni S., Molinari-Tosatti M . P ., Quarto N.,
Ziehe M ., Giavazzi R ., Presta M ., 1995, Growth advantage and vascularization induced by
basic fibroblast growth factor overexpression in endometrial HEC-I-B cells: an export-
dependent mechanism of action . Cancer Res ., 55 : 4729-4738.

Dellian M ., Witwer B . P., Salehi H. A., Yuan F., Jain R. K ., 1996, Quantitation and
physiological characterization of angiogenic vessels in mice . Effect of basic fibroblast
growth factor, vascular endothelial growth factor/vascular permealility factor, and host
microenvironment. Am . 1. Pathol ., 149: 59-71.

Denekamp J ., 1984, Vascular endothelium as the vulnerable element in tumors. Acta

Radio!. Oncol., 23 : 217-225 .

Fajardo L . F., 1989, The complexity of endothelial cells . A review. Am . J. Clin. Pathol. ,
92 : 241-250.

Folkman J ., 1985, Tumor angiogenesis. Adv . Cancer Res ., 43 : 175-203 .

Folkman, J., 1995, Tumor angiogenesis. The molecular basis of cancer. 1. Mendelsohn, P.
H. Howley, M . A. Israel and L. A. Liotta (eds .). Philadelphia, W. B. Saunders, 206-232.

445
Folkman J., Shing Y., 1992, Angiogenesis. J. Biol. Chem ., 267 : 10931-10934.

Kaufmann P., Bruns u., Leiser R., Luckhardt M ., Winterhager E., 1985 , The fetal
vascularisation of term human placental villi . II Intermediate and terminal villi. Anat.
Embryol. Berl ., 173 : 203-214.

Klagsbrun M ., D'Amore P. A, 1991, Regulators ofangiogenesis. Annu . Rev . Physiol. , 53 :

217-239.

Konerding M . A., 1991, Scanning electron microscopy of corrosion casting in medicine.

SanningMicrosc., 5: 851-865 .

Konerding M . A., van Ackern C., Steinberg F., Streffer c., 1992a, The Development ofthe
Tumour Vascular System : 2-D and 3-D Approaches to Network Formation in Human
Xenografted Tumours. In: Maragoudakis, M . (ed) : Angiogenesis in Health and Disease.
Plenum Press, New York , 173-184 .

Konerding M . A., Steinberg F., van Ackern C., Budach V., Streffer C. 1992b,
Comparative ultrastructural studies of the vascularity in different human xenografted
tumours. Vol. 42 . In : Fiebig RH, Berger DP (eds .) Immunodeficient mice in oncology,
169-179. Karger, Basel: Contributions to oncology.

Konerding M . A , Lametschwandtner A., Miodonski A 1., 1995, Microvascular corrosion

casting in the study oftumorvascularity: a review . Scanning Microsc., 9: 1233-1244.

Lametschwandtner A , Weiger T., Bernroider G., 1989, Morphometry of corrosion cast. In :

Motta, PM (ed .), Cells and Tissues : A Three-dimensional Approach by Modern
Techniques in Microscopy. Progress in Clinical and Biological Research., Vol. 295 : 427-
433 .

Lewis W . H. , 1927 , The vascular patterns oftumors. Johns Hopkins Hosp BuH, 41 : 156-
175.

Lindgren A G. H ., 1945, The vascular supply of tumours with special reference to the
capillary angioarchitecture. Acta Pathol Microbiol Scand , 22 : 433-452.

Malkusch W ., Konerding M . A , Klapthor B ., Bruch J., 1995, A simple and accurate

method for 3-D measurements in microcorrosion casts illustrated with tumour
vascularization. An Cell Pathol, 9: 69-81

Miodonski A., Hodde K . C., Bakker C., 1976, Rasterelektronenmikroskopie von Plastik-
Korrosions-Präparaten, morphologische Unterschiede zwischen Arterien und Venen . Beitr.
Elektronenmikroskop. Direktabb. Oberfl., 9: 435-442.

Murakami T., 1971 , Application of the scann ing electron microscope to the study of the
fine distribution ofthe blood vessels. Arch . histol. Jap ., 32: 445-454.

Rhodin J.A .G., Fujita H ., 1989, Capillary growth in the mesentery of normal young rats .
Intavital video and electron microscope analysis. J. Submicros. Cytol. Pathol., 21: 1-34.

Shubik P., 1982, Vascularisation of tumours : A review. J Cancer Res Clin Oncol, 103:
211-226.

446
Spalteholtz W., 1914, Über das Durchsichtigmachen von menschlichen und tierischen
Präparaten nebst Anhang über Knochenfärbung. 2. ergänzte Auflage, Leipzig, Hirzel.

Vaupel P., Gabbert H., 1986, Evidence for and against a tumor type specific vascularity.
Strahlenther. Onkol. , 162: 633-638 .

VaupeI P., Fortmeyer H. P., Runkel 1987, Blood flow, oxygen comsumption, and tissue
oxygenation ofhuman breast cancer xenografts in nude rats., Cancer Res., 47: 3496-3503.

Vaupel P., Kallinowski F., Schlenger K., Fortmeyer H. P., 1989, Blood flow and oxygen
consumption rates of human gynecological tumors xenografted into rnu /rnu-rats .
Strahlenther. Onkol ., 165: 502.

447
"CMIOI - AN ANTI-PATHOANGIOGENIC AGENT: PRE-CLINICAL AND
CLINICAL EXPERIENCES"

Carl G. Hellerqvist, Department of Biochemistry, Vanderbilt University

School ofMedicine, Nashville, Tennessee 37232-0146 , USA

INTRODUCTION

The pathophysiology of Group B Streptococcal (GBS) disease depends on the age of the
neonate infected . Neonates infected at birth by GBS colonizing the birth canal may be
diagnosed with early-onset disease. The symptoms are respiratory distress and
cardiovascular shock (Baker and Edwards, 1995) and the gravity may vary from
asymptomatic to lethaI. In infants older than I week, late-onset disease is often associated
with meningitis. GBS rarely affects normal adults, but may present in old adults with
impaired peripheral circulation (diabetics) or with a compromised immune system due to
cancer or other debilitating disease (Farley, Harvey, StulI, Smith, Schuchat, Wenger, and
Stephens, 1993).

GBS is an ovine pathogen in adult sheep with the lung as the only target organ. We
demonstrated in a sheep model (Hellerqvist, Rojas, Green, Seil, Sundell, and Stahlman,
1981) that a polysaccharide, which we isolated from culture media of GBS isolate s from
infected neonates and which we called "GBS toxin," would induce the pathophysiology
seen in the infected neonate (Rojas, Green, Hellerqvist, Olegard, Brigham, and Stahlman,
1981; Sandberg, Engelhardt, Hellerqvist, and Sundell, 1987; Rojas, Larsson, Ogletree,
Brigham, and Stahlman, 1983; Engelhardt, Sandberg, Bratton , Van den Abbeele, Grogaard,
Hellerqvist, and SundelI, 1987; Rojas, Larsson, Hellerqvist, Brigham, Gray, and Stahlman,
1983).

We have recently completed a large study on human newboms diagnosed with GBS or
other bacterial or viral infections at birth. Using an ELISA assay, we demonstrated that
CMI0l (see below) is exclusively present in the plasma and urine in GBS-infected
newboms in the high ~lg/m l range. As the neonates recover and the respiratory distress
symptoms resolve , the amount of plasma CMI0l may remain elevated. This observation
supports the notion that embryonie receptors are necessary for the inflammatory reaction
in the lung neovasculatu re. This is further supported by the fact that even the non-

Angiogenesis : Models, M odulators, and Clinica l Applicalions

Edited by Marago udaki s, Plenum Press, New York, 1998 449
symptomatic earlier GBS-infected newborns secreted CMIOI in the urine with a half-life
of 2.5 days . CMl 01 isolated from the urine had the same molecular weight and sugar
composition as the clinical CMI0l (see below) and was biologically active in the sheep
model.
Further fractionation and purification of GBS toxin lead to the identification of a more
potent GBS toxin (HeIlerqvist, 1991) with approximately 30 - 60 times the specific
activity as compared to the original GBS toxin (Sandberg, Edberg, Fish, Parker, HeIlerqvist,
and SundelI, 1992. The more potent GBS toxin, now referred to as CMI01 , has a
molecular weight of 300,000 Dalton and a sugar composition similar to the previously
reported high molecular weight GBS toxin (HeIlerqvist et al., 1981) minus the mostly
mannose-containing GBS toxin component (Hellerqvist, SundelI, and Gettins, 1987).

RATIONALE

Based on the observation that the neonatal lung endothelium, the last endothelium to
develop, is the only target for CMI01, we speculated that embryonie receptors were still
expressed for a few days after birth. We then made the assumptions that hypoxia-driven
pathologic and embryonie angiogenesis would express similar receptors which would
appear in the de-differentiated stage during angiogenesis and sprouting of new vessels, but
would be down-regulated following completion ofthe vessel as endothelial cells return to a
fully differentiated stage (Hellerqvist, 199I).

Using mouse anti-CMI01 MoAbs and rabbit anti-CMI01 IgG, we demonstrated with
immunohistochemistry that, in fact, neovasculature from different types of human tumors
binds CMI0l (Hellerqvist, Thurman, Page, Wang, RusselI, Montgomery, and SundelI,
1993).

EFFECT OF CMIOI ON ANGIOGENESIS IN MOUSE MODELS

Physiologie Angiogenesis: Mice, as humans, show the same age-dependent

susceptibility to GBS infection. Neonatal mice are susceptible to GBS within 12 hours
after birth and afterwards become totally unsusceptible to GBS. In our hands, CM 101 can
be infused intravenously in normal adult mice at 40 mg/kg with no apparent toxicity or
discomfort to the animals (HeIlerqvist et al., 1981). Likewise, CMIOI can be infused i.v.
during pregnancy in mice with no effect on the physiologic, presumably hormonally
regulated , neovascularization governing the reproductive process in the mother and result in
normal size live litter (Thurman , Liu, Wamil, Page, and Hellerqvist, 1997 Unpublished).
CMIOI has also been shown not to affect wound healing as measured in a sponge model
(Quinn, Thurman , SundelI, Zhang, and HelIerqvist, 1995).

Pathologie Angiogenesis: We have reported on the inhibitory effect on the growth of

human xenograft in nude mice ofi .v. infusion of CMIOI in picomole amount s (HeIlerqvist
et al., 1993) and on the ability to ablate tumors in immunocompetent mice by repeat
infusions every other day over 11 weeks (Thurman, RusselI, York, Wang, Page, SundelI,
and Hellerqvist, 1994). Tumor endothelium was the apparent target which we established
in aseparate study . CMIOI could induce neovascularitis in developing neovasculature in
tumors and indirectly cause infiltration of inflammatory cells and apoptosis of tumor cells

450
within 60 minutes post-infusion (Thurman, Page, Wamil, Wilkinson, Kasami, and
Hellerqvist, 1996).
Recent studies have also demonstrated that CMIOI can inhibit and reverse pannus
formation and joint degradation, in a mouse model for rheumatoid arthritis, by inhibiting the
hypoxia-induced angiogenesis (Wamil and Hellerqvist, 1997 Unpublished). A similar
inhibition of angiogenesis induced by wounding could be demonstrated by MRI (Neeman,
Abramovitch, Wamil, and HeIIerqvist, 1997 Unpublished). The effect was an improved
wound healing with lack of scarring, a phenomenon also explored in spinal cord inj ury
(Warnil, Wamil, and HeIIerqvist, 1997 Unpublished) .

TOXICOLOGY

Histological examination oflung, liver, kidney, spleen and heart tissues from mice subjected
to CMI0l treatment have revealed no apparent toxicity ofCMI0l at the different dosages
used in the different types of experiments.

MECHANISM OF ACTION

Our working hypothesis is that CMI01 binds and cross links receptors expressed only on
de-differentiated endothelial cells engaged in pathologic or embryogenic angiogenesis and
neovascularization. Complement C3 is then activated by the alternative pathway and C3a
initiates a cytokine-driven inflammatory response with phagocytosis of targeted
endothelial cells . This effect on pathologic angiogenesis is supported by the data presented
below and by arecent study by Vetvicka, Thornton, and Ross (1996) which shows that
the C3b receptors, present on normal leukocytes, will recognize C3b but subsequent
phagocytosis of the target requires an interaction with a lectin component of the C3b
receptor and a polysaccharide. CMIOI is such a polysaccharide allowing the phagocytosis
ofthe targeted endothelial cells (Hellerqvist et al., 1993).

Balb/c - Madison Lung Tumor Model

We have tested the hypothesis in tumor-bearing and normal mice . First, we had previously
established that Balb/c mice respond to CMIOI with a systemic cytokine response only if
tumor- bearing. IL-la, as weil as TNFa levels, are e1evated 1 hour post-CMIOI infusion.
No cytokine response to CMIOI is seen in control Balb/c mice.

To further our understanding ofthe mechanism of action, Balb/c mice were inoculated with
Madison lung tumors propagated to approximately 100 cmm in size. Mice were then
treated with CMI0l and sacrificed 5,15,30 and 60 minutes post a single CMIOI infusion.
Tumors and livers were harvested, split in half, and the halfs were subjected to
immunohistochemical and RT-PCR analysis for CMIOI binding, complement activation
and cytokine mRNA expression, respectivel y.

CMIOI Binding: The results show that CMIOI is detected by a monoclonal antibody to
CMIOI in the tumor neovasculatu re at the 5 minute time point. The intensity decreases

451
progressively over the 60 minutes. No signal is detected in the liver or in the tumor of the
untreated mice.

Complement Activation: Control and treated tumor sections from the same time points
were stained for mouse C3 using a sheep anti-mouse C3 IgG. Data showed that
complement was activated but only in the CMI0l-treated tumors . To further substantiate
that CM101-C3b initiates complement activation, we incubated additional treated and
control tumor sections with human serum. The results showed that anti-human C3 IgG
detected human C3 binding to the tumor endothelium where CM101 was detected but did
not bind to the tumor cells where mouse complement could also be detected.

Leukocyte Infiltration and Cytokine Levels: Tumor and liver sections from treated and
untreated animals were analyzed immunohistochemically for IL-6, IL-6R, TNFa and
TNFaR(I).

Liver Tissue: The liver tissue showed no evidence of infiltrating inflammatory cells nor
signs oftoxicity. IL-6 could be barely detected in Kupffer cells with no increase with time.
IL-6R could be detected in the endothelial lining at 15 minutes and by 30 minutes the
hepatocytes were strongly positive presumably as a consequence of an acute phase
response.

TNFa could be detected at 15 minutes in the large vessels and a diffuse stain was evident in
the hepatocytes which increased by 60 minutes.

The lining of large vessels stained for TNFaR(l) 5 minutes post-infus ion and intensity
increased over the 60 minute time course. Hepatocytes were positive at 15 minutes only
and had faded by 60 minutes.

The cytokine response in the liver to an inflammatory agent targeting a site, in this case a
tumor, reflects the engagement ofthe liver in a managed inflammatory response.

Tumor Tissue: There was evidence of infiltrating inflammatory cells in the tumor
sections already 5 minutes post-CMlOI treatment. Infiltrating macrophages stained
positive for IL-6 at 15 minutes and the stain intensified with time. Small vessel staining for
IL-6 peaked at 30 minutes.

TNFa was detected in infiltrating macrophages and the intensity increased with time. By
15 minutes, a weak cytoplasmic staining was detected in the tumor cells and intensity
increased with time.

TNFaR(I) was detected only in the tumor vasculature, appeared at 15 minutes and was
strongest by 60 minutes .

The observations in the tumor are consistent with an inflammatory response targeting the
tumor vasculature expressing TNFaR(1) interacting with its ligand delivered by the
infiltrating activated macrophages. The appearance of apoptotic tumor cells at the 60
minute time point is consistent with the cytokine response.

452
The RT-PCR analyses gave data consistent with the appearance of the cytokines and their
receptors.

Rat Glioma Model

We postulated that the anti-pathoangiogenic properties of CM101 would be reflected in an

ability to impair angiogenesis as measured by expression of VEGF and Fit-I. For this
purpose, we implanted rat glioma spheroids (Wamil, Abramovitch, Wang, Neeman, and
Hellerqvist, 1997) of approximately 2 mm size in C57 nude mice. Mice were treated with
CMI0l or saline every other day from day 9. Tumors were measured and harvested 1
hour post 4th and 7th infusion and the mRNA levels for Rat VEGF and Mou se Flt-I as
weil as ß-actin were determined quantitatively and qualitativel y.

The results demonstrated a 70-80% reduction in tumor volume already after 4 infusions.
RT-PCR analyses showed that, when normalized to the ß-actin message, the message for
both Rat VEGF and Mouse Flt-I was down-regulated 50-70% . The mechanism by which
an inflammatory cascade targeting the tumor neovasculature can down-regulate VEGF and
Flt-l is unclear, but we speculate that the induced apoptosis of the tumor cells leads to an
early down-regulation of VEGF, whereas basic survival function s reflected in the message
for ß-actin are sustained . Lower levels of Rat VEGF would result in lower levels of
message for mouse Fit-I (Warnil et 01., 1997) .

CLINICAL EXPERIENCES

Clinical Phase I trials have been conducted with CM101 in patients with refractory disease.

Data obtained from cancer patients participating in c1inical trials supports a similar
mechanism of action as discussed above and supports an engagement of tumor
neovasculature in the inflammatory process (DeVore, Hellerqvist, Wakefield, Wamil,
Thurman, Minton, SundelI, Yan, Carter, Wang, York, Zhang, and Johnson, 1997; Wamil,
Thurman, DeVore , Wakefield, Johnson , and Hellerqvist, 1997; DeVore, Johnson,
Hellerqvist, Wakefield, Browning, Page, SundelI, and Johnson, 1996).

Fifteen patients in groups of 3 were treated with i.v. infusions of CM101 starting at
7.5!lglkg = 1 Unit and escalating to 2U, 3.3U and 5U. All patients experienced flu-like
symptoms with fever and chilis. Chilis lasted from 45 minutes to 90 minutes post-
infusion . Analysis of serum sampIes demonstrated a dose- and tirne-dependent
inflammatory cytok ine cascad e. RT-PCR analysis ofthe isolated leukoc ytes allowed us to
map the inflammatory response .

Elevated C3a at 30 minutes post-infusion established a complement activation in response

to CMlO 1. The initiation of an inflammatory cascade was evident in a decrease by 50-60%
at 60 minutes of circulating granulocytes, monocytes by 80-100%, and Iymphocytes b y
60-80%. Macrophage inflammatory protein-I (MIP-I) produced by granulocytes peaked
at 60-90 minutes, whereas IL-8 initially produced by the granulocytes and by 6 hours by
all leukocytes peaked at 120-180 minutes. MIP-l recruits Iymphocytes by producing
TNFa which peaked at 90-120 minutes . IL-6 peaked at 180 minutes and the inflammatory

453
response , which for reasons given below targets the patient's tumors , was gradually turned
off by up-regulation of IL-l 0 which peaked at 240 minutes .

Patients experienced no hepatic or renal toxicity nor was there any evidence of vascular
leakage or long-term toxicities due to elevated cytokines. It is worth mentioning that
newborns who overcome the GBS infection had CMlOl levels 2000 times those given to
the patients. Furthermore, during the infectious state the babies had cytokine levels 10-
fold those measured at the highest dose tested in the patients (Sundell et al., 1996).

Whereas tumor responses in Phase I trials are rare (1-2%), tumor responses in 15 patients
in response to 3 infusions of CMI0l were seen in 5 patients at the 2U and 3.3 U levels.
Three patients had tumor shrinkage and 2 had stable disease. A patient with c1assic
Karposi sarcoma who had 30-50 new tumors per month experienced a dramatic change.
His I cm marker tumor disappeared in 2 weeks with no scar and his other lesions "dried up
and fell off in the shower." The patient participated for 5 months over which time he had
only 1 new lesion. He could wear his shoes and walk again. A patient with hepatocellular
cancer demonstrated the limited effectiveness on bulky disease of an anti-angiogenic agent
alone. The patients 2 cm tumor shrunk by 80-90%, his 6 cm by 50%, and a 12 cm tumor
shrunk only by 25%. This is, however, consistent with the anticipated mechanism of
action where CM 10 1 acts to target only neovasculature via an induced intlammatory
response , which was documented in most patients by intense localized tumor pain.

The third responding patient was diagnosed in 1992 with adenocarcinoma of the duodenum
resistant both to chemo and radiation therapy . After surgery and radiation, she presented
in 1994 with a 2 cm metastasis to the suprac1avicular node. The patient responded to
treatment and after 10 infusions the tumor that had shrunk by 90% was removed. The
biopsy of the responding tumor showed c1ear evidence of infiltration of intlammatory
leukocytes consistent with our proposed mechanism of action . The responding tumor was
apoptotic by tunnel assay, whereas the original tumor was not (Hellerq vist, Wang, Wamil,
Price, Yan, Carter, Wang, DeVore, Johnson , and Lloyd , 1997). Interestingly, the treated
tumor showed no positive stain for FasL which the original tumor did and p53 was readily
detected in the responding apoptotic tumor whereas the original tumor was negative for
p53. The origin of the p53 is unc1ear. It is, however, of interest that the cancer patient's
granulocytes are negative for p53 message but highly positive 30 minutes post-CM 101
infusion . In contrast, normal individuals express p53 message in their leukocytes. The
biopsy suggests that the activated leukocytes, which have up-regulated pS3 in response to
CMIOI migrated to the tumor which became apoptotic and shrunk by 90% . We are
currently investigating this interesting possibility of leukocyte delivery of p53 to the p53
negative tumor cells.

CONCLUSIONS

CMI01 , an apparent anti-pathoangiogenic agent infused i.v., targets and binds within
minutes to de-differentiated endothelial cells responding to hypoxia-induced growth stimuli
but not to physiologically governed angiogenesis. This makes CMlOl unique, although
endostatin may have similar properties (Folkman, Judah, 1997 Personal Communications).
Binding of CMlOI to endothelium results in immediate activation of complement by the
alternative pathway, evidenced by immunohistochemical evaluation of tumor biopsies from

454
mice and from systemic C3a in patients serum. The resulting inflammatory response is
documented in a dose- and time-dependent cytokine cascade with activated leukocytes
migrating in a rapid but orderly fashion to the tumors. The infiltration of the tumors b y
activated leukocytes is evidenced from the mouse models and from patient biopsies and the
inflammatory response in the patients is supported by the severe tumor localized pain
experienced by the patients especially at the 3D level. Thus , the abilit y of GBS
toxinlCMIOI to induce fatal inflammation targeted to the lung in the human neonate is
parallelled in effect ofCM101 on the neovasculature recruited by human tumors.

The additional observations that CMIOI down-regulates mRNA in the tumor cells and
consequently the mRNA for Flt-l in the rat g1ioma model and in an unclear way provides
pS3 to a pS3 negative tumor in a patient rendering the tumor apoptotic and shrinking, only
adds strength to our beliefthat CMIOI has great potential to play an important role in the
fight against cancer and other diseases amplified by pathoangiogenic elements .

455
REFERENCES

Baker, C.J. and Edwards M.S., 1995, Group B streptococcal infections, in: Infectious
Diseases ofthe Fetus and Newborn Infant, 4th Edition (Remington, J.S. and Klein, J.O.,
eds.), pp. 980-1054, W.R Saunders Co., Philadelphia, PA.

DeVore, R., Johnson, D.R., Hellerqvist, C., Wakefield, G., Browning, P., Page, D.,
SundelI, H., and Johnson, D.H., 1996, A phase I study of the anti-neovascularization drug
CM101, Proceedings of ASCO Meeting, 15:1558.

DeVore, R.F., Hellerqvist, c.o., Wakefield, G.R, Wamil, B.D., Thurman, G.B., Minton,
PA, SundelI, H.W., Yan, H.-P., Carter, C.E., Wang, Y-F., York, G.E., Zhang, M.-H., and
Johnson, D.H, 1997, A phase I study ofthe antineovascularization drug CMI0l, J. Clin.
Can. Res. 3:365-372.

Engelhardt, R , Sandberg, K, Bratton, D., Van den Abbeeie, A., Grogaard, J., Hellerqvist,
C., and SundelI, C., 1987, The role of granulocytes in the pulmonary response to group B
streptococcal toxin in young lambs, Pediatr. Res. 21: 159.

Farley, MM., Harvey, RC ., StulI, T., Smith, J.D., Schuchat, A., Wenger, J.D., Stephens,
D.S., 1993, A population-bases assessment of invasive disease due to group B
streptococcus in nonpregnant adults, N Engl. J. Med 328: 1807-11.

Hellerqvist, c.o., Rojas, J., Green, RS., Seil, S., SundelI, H , and Stahlman, MT, 1981,
Studies on group B ß-hemolytic streptococcus . I. Isolation and partial characterization of
an extracellular toxin, Pediatr. Res. 15:892-898.

Hellerqvist, C.G., SundelI, H., and Gettins, P., 1987, Molecular basis for group B 13-
hemolytic streptococcal disease, Proc. Natl. Acad. Sei., U.S.A. 84: :51.

Hellerqvist, C.G., 1991, Therapeutic agent and method of inhibiting vascularization of

tumors, U.S. Patent 5,010,062.

Hellerqvist, c.o., Thurman, G.B., Page, D.L., Wang, Y-F., RusselI, BA, Montgomery,
C.A., and SundelI, H.W., 1993, Anti-tumor effects ofGBS toxin: a polysaccharide exotoxin
from group B ß-hemolytic streptococcus, J. Can. Res. Clin. Oncol. 120:63-70.

Hellerqvist, C.G., Wang, E., Wamil, B.D., Price, J.O., Yan, H-P., Carter, C.E., Wang, Y-
F., DeVore, R.F., Johnson, D.H, and Lloyd, 1997, Evidence ofinduced apoptosis in cancer
patients treated with CMlOl , Proceedings of ASCO Annual Meeting.

Quinn, T.E., Thurman, G.B., Sundell, A.K , Zhang, M., and Hellerqvist, C.G., 1995,
CMlOI , a polysaccharide anti-tumor agent, does not inhibit wound healing in murine
models, J. Can. Res. Clin. Oncol. 121:253-256.

Rojas, J., Green, RS., Hellerqvist, C.G., Olegard, R , Brigham, KL., and Stahlman M.T.,
1981, Studies on group B ß-hemolytic streptococcus . n. Effects on pulmonary hemo-
dynamics and vascular permeability in unanesthetized sheep, Pediatr. Res. 15:899-904.

456
Rojas , 1., Larsson, L.E ., Ogletree, M.L., Brigham, KL., and Stahlman, MT., 1983, Effects
of cyclooxygenase inhibition on the response to group B streptococcal toxin in sheep,
Pediatr. Res. 17: 107-110.

Rojas , J., Larrson, L.E., Hellerqvist, c.G., Brigham , KL., Gray, ME, and Stahlman, MT. ,
1983, Pulmonary hemodynamic and ultrastructural changes associated with group B [3-
hemolytic streptococcal toxemia in adult sheep and newbom lambs, Pediatr. Res. 17: 1002-
1008.

Sandberg, K , Engelhardt, B., Hellerqvist, C.G., and SundeIl H., 1987, Pulmonary response
to group B streptococcal toxin in young lambs, 1. Appl. Physiol. 63:2024-2030 .

Sandberg, K., Edberg, KE. , Fish, w., Parker, RA, Hellerqvist, C.G., and SundelI, H.W.,
1992, Thromboxane receptor blockage (SQ 29548) in group B streptococcal (GBS) toxin
challenge in young lambs, Pediatr. Res. 35:571-579.

SundelI, H.W., Yan, H.-P., Wu, K, Wamil, B.D., Gaddipati , R., Carter, c.E., Stahlman,
MT., and Hellerqvist, c.G., 1996, Isolation and identification of group B [3-hemolytic
streptococcal (GBS) toxin from septic newborn infants, Pediatr. Res. 39:302A.

Thurman, G.B ., RusselI, BA, York, G.E., Wang, Y-F., Page, D.L., SundelI, HW., and
Hell erqvist, C.G., 1994, Effects of GBS toxin on long-term survival of mice bearing
transplanted Madison lung tumors , 1. Can. Res. Clin. Oncol. 120:479-484 .

Thurman, G.B., Page, D.L., Wamil, B.D., Wilkinson, L.E., Kasami, M., and Hellerqvist,
c.G., 1996, Acute inflammatory changes in subcutaneous microtumors in mice ears induced
by intravenous CM101 (GBS toxin), 1. Can. Res. Clin. Oncol. 122:549-553.

Vetvicka, V, Thomton, B.P., and Ross, G.D., 1996, Soluble beta-glucan polysaccharide
binding to the lectin site of neutrophil or natural killer cell complement receptor type 3
(CDllb/CDl8) generates a primed state of the receptor capable of mediating cytotoxicity
of iC3b-opsonized target cells, 1. Clin. Invest. 98(1):50-61 .

Wamil, B .D., Abramovitch, R, Wang, E., Neeman, M ., Hellerqvist, c.o., 1997, CMI0l
Inhibits VEGF Induced Tumor Neovascularization as Determined by MRI and RT -PCR,
Proceedings ofthe 88th Annual American Association for Cancer Research Meeting.

Wamil, B.D ., Thurman , G.B., DeVore, RF., Wakefield, G., Johnson, D.H., and Hellerqvist,
c.G., 1997, Soluble E-Selectin in cancer patients as a marker of the therapeutic efficacy of
CM101, a tumor inhibiting agent, evaluated in phase 1 clinical trial. J. Can. Res. Clin.
Oncol. 123(3): 173-179.

457
MAPPING NEOVASCULARIZA TION AND ANTINEOVASCULARIZATION
THERAPY

Michal Neeman, Gila Meir, Catherine Tempel , Yael Schiffenbauer and

Rinat Abramovitch

Department ofBiological Regulation

The Weizmann Institute of Science
Rehovot, 76100 Israel

1. INTRODUCTION

The switch of tumors from avascular to the vascular phase and the on set of tumor
angiogenesis mark a critical checkpoint in tumor progression. Avascular tumor dormancy
can sometimes extend over many years, while upon vascularization, tumors show
increased invasiveness, elevated metastatic potential and significantly worse prognosis
(Weidner and Folkman, 1996). Regulation ofthe transition to the vascular phase depends
on the balance between the production of promoters and inhibitors of angiogenesis
(Hanahan and Folkman, 1996) . The goal of our work was to define phys iological
seenarios which perturb this balance and can thus drive angiogenesis to pre viousl y
dormant tumors . Such me chanisms that promote angiogenesis ma y explain
epidemiological observations regarding age specific probabilities of certain tumors,
environmental efTects and the effects of trauma on tumor growth. In order to follow
angiogenesis quantitati vely , we developed an experimental system that relies on detection
ofvessel density by magnetic resonance imaging (MRI) .
Most experimental systems for in vivo analysis of angiogenesis, such as the cornea assay
and the window chamber assays, tend to be extremely invasive and are thus prone to false
positive errors. Wh ile these methods are very effective in identification of positive or
negative angiogenic activity, they do not yield quantitati ve measure ofthis activity, and do
not provide the sensiti vity necessary for detect ion of subtle changes in the extent or the
kinet ics of the angiogenic respon se. An additional problem in quantificati on of the
angiogenic kinetics is intrinsic to the large biological variance in the response of different
(though geneticall y identical) laboratory animals, to inoculation of the same numb er of
cell s, even when great care is taken to ensure ident ical handling. The sev erity of this
heterogeneity and the experimental approach for overcoming this variab ility will be
described in the following sections .

Angiogenesis: Model s, M odulators, and C/inical App/ications

Edited by Maragoudakis, Plenum Press, New York, 1998 459
 2D Perivascular cuITs 3D Avascular nodules
Figure 1 The avascular solid tumor. Two primary modes of avascular tumor initiation
can be envisioned. The perivascular growth implies that tumor progresses as a two
dimensionallayer covering and advancing along the surface of existing vasculature. Tumor
progression in this case will be relatively independent of angiogenesis . The second mode
of growth will create a three dimensional spherical nodule. Tumor progression in this case
is expected to be absolutely dependent on angiogenesis . The implanted spheroid model
can be regarded as an experimental simulation ofthis 3D tumor growth mode.

1.1 The angiogenic switch in tumor progression

In order to study the onset of angiogenesis, it is necessary to design a reproducible

experimental model of the avascular tumor. A conceptual limitation in designing such a
model lies in the fact that human tumors are diagnosed only after the angiogenic switch
took piace, and therefore the properties ofthe very early stages ofhuman tumor growth are
not known . Assuming that one of the earliest changes in ' phenotype with malignant
transformation is reduced suppression of cell proliferation, we can envision two general
modes of avascular tumor progression (Figure I). One mode would include the growth of
cells along the surface of existing blood vessels, in a small number of cell layers. This 2D
mode of cell proliferation would be relatively independent of angiogenesis, but requires
active invasion ofthe tumor cells into the normal tissue.
The second mode of growth leads to the creation of 3D avascular nodules . Such nodules
would theoretically reach a point in which nutrient and oxygen supply, and removal of
waste products by diffusion are not sufficient to support cell proliferation and viability .
While diffusion is very efficient over small distances , and very effectively provides
accessibility of metabolites within cells, diffusion becomes extremely ineffective over
large distances. The exact rates of diffusion of water were measured by MRI as will be
outlined in the following section, and provide insight to the problem of diffusional nutrient
and oxygen supply. Tumor growth will therefore arrest at a certain size, a necrotic center
will develop, and further tumor progression will absolutely depend on angiogenesis. We
were interested in developing a model to reproduce this 3D mode of tumor progression .

1.2 The implanted multicellular spheroid model

Multicellular tumor spheroids were developed as a system for analysis of the chronic
microenvironmental stress characteristic of solid tumors (Sutherland, 1988). Typically, the
spheroids will contain a viable rim, 250 11m thick , consisting of an outer layer of
proliferating cells, and inner layer of hypoxie, quiescent cells that are capable of resuming
proliferation if rescued from the spheroid. The inner core of spheroids larger than 500 11m
in diameter is usually necrotic. A major advantage ofthe spheroid system is the excellent

460
 10' r - - - . - - - - - , - - - , . - - - , - - - , - - - , - - - r - - . , , . - - - , - - " ' " 7 1
Tumor volume (ul)

5 6 7 8 9 10 11 12 13 14 15

Time after Implantation (days)

Figure 2 Scatter of tumor growth in the implanted spheroid model. Spheroid growth was
assumed to follow a delayed Gompertz kinetics, in which an initial lag period corresponds
to the development of the vascular bed (Abramovitch et al., 1995). Analysis included a
random 10% Gaussian distribution in the initial volume of the spheroid and a 10%
variance in the angiogenic process . The large distribution of the data at the transition from
growth arrest to rapid growth implies that it would be impossible to deduce the connection
between angiogenesis and tumor growth from a population average, while such
information will be very easy to deduce from non invasive studies monitoring both growth
and angiogenesis in each mouse.

reproducibility of preparation and the almost perfect spherical symmetry. These features
provide an experimental advantage in creating extremely weil defined initial conditions for
studying the onset of angiogenesis (Abramovitch et al., 1995; Borgstrom et al., 1996;
Farrell et al., 1987; Torres Filho et al., 1995). Moreover, the expression of angiogenic
factors can be determined pre implantation, in a manner not possible with tumors initiated
by inoculation of suspended cells (Shweiki et al., 1995; Waleh et al ., ] 995). The gradients
in nutrient and oxygen concentration found across spheroids (Acker et al., 1987; Casciari
et al., 1988; Mueller Klieser and Sutherland, 1983) may reproduce weil the expected
microenvironments within 3D avascular tumor nodules .
Despite the greatly improved reproducibility in initiation of tumors by spheroid
implantation relative to inoculation of cells, the biological variability may still be too large
to allow analysis of kinetics of angiogenesis from population average. The problem is
demonstrated in simulations of tumor growth kinetics, assuming a growth delay
corresponding to the duration of vascular recruitment (Figure 2) . This anal ysis was
performed assuming a 10% variance in the initial volume of the spheroid and 10%
variance in the rate of vascular development between different mice. The experimental
accuracy of spheroid selection can be up to 5% in diameter which will imply a 15%
variability in spheroid volume. Thus this simulation provides a lower limit for variabil ity
and the scatter in the actual experiment is expected to be significantly larger.

461
We can see that even iffor each individual tumor the escape from dormancy and initiation
of growth exactly matches the establishment of a vascular bed , the population average,
which is the parameter that can be deduced from post mortum invasive protocols, will not
contain this information (Figure 2). Non invasive mapping of angiogenesis and tumor
volume are absolutely essential for obtaining such kinetic data from each individual tumor.
Both parameters were monitored by MRI primarily because of the three dimensional
capacity of this methodology, and the fact that this type of monitoring does not impose a
physiological significant burden on the mice.

1.3 Following angiogenesis and anti-angiogenesis by MRI

Vascularization of implanted spheroids in nude mice occurs over aperiod of 4-10 days .
Frequent analysis ofthe mice during this period is only possible ifthe measurement itself
perturbs the mice as little as possible. We decided to develop the methodology for
monitoring vascularization using an intrinsic MRIcontrast that requires the shortest scan
time and thus the shortest duration of anesthesia (Abramovitch et al., 1995). The validity
of the measurement was assessed by comparison of the apparent vessel density determined
by MRI with the analysis of the density of vessels determined from the mice post mortum
photograph, using a number of angiogenic stimuli inc1uding implanted spheroids, fuH
thickness dermal incisions, and agarose beads containing a defined angiogenic growth
factor.
The physical basis for this approach of determination of the apparent vessel density by
MRI (AVDMRJ) lies in the paramagnetic properties of deoxyhemoglobin (Boxerman et
al., 1995; Ogawa et al., 1990a; Ogawa et al., 1990b). The bulk magnetic susceptibility of
blood vessels that contain deoxyhemoglobin is very different from that of tissue resulting
in the distortion of the magnetic field in the vicinity of the vessels. Water within this
region will show line broadening and faster T2 * relaxation . The MRI gradient echo
imag ing pulse sequence is extremely sensitive to changes in T2* , and regions with large
magnetic inhomogeneity will appear dark in the images . Induction of angiogenesis leads to
major increases in the local density of blood vessels, and these will be translated to signal
loss in gradient echo images .
Quantitative determination of T2* can be obtained using pixel by pixel fitting of images
obtained with increasing echo time. This type of analysis implies relatively long
measurement time and accordingly, longer duration of anesthesia. Alternatively, the
degree of signal loss can be deduced using the signal intensity of another tissue as a
reference (Abramovitch et al., 1995). In the experiments reported here, signal intensity
was scaled to the signal of the back musc1e of the mice at a distance of about 1 cm from
the angiogenic stimulus (Figure 3).
At the end of MRI measurements the skin surrounding the angiogenic stimulus was
removed and photographed (Figure 3). The apparent vessel density was determined from
the skin photograph (AVDphoto) by analysis of the mean signal inten sity in a scanned
photograph (using NIH Image) . The highest density of vessels was found in the 1 mm
periphery of the angiogenic stimulus. Comparison of the two independent measures of
vessel density (AVDMRI and AVDphoto) gave a highly significant correlation (r=0.9,
p=O.OOOI, n=35).
Functionality of the neovasculature, namely the ability to serve as oxygen carrier, was
determined by allowing mice to breath 95% oxygen 5% C02 (carbogen). The elevation of
blood oxygenation results in enhanced signal intensity in gradient echo MRI (Karczmar et
al., 1994; Robinson et al., 1995). The extent of MRI signal enhancement depends on the
volume fraction of blood and indeed , we found that regions of high vesseldensity also
showed large signal enhancement with carbogen breathing.

462
Figure 3 Determination of the apparent vessel density . A. Gradient echo image of an
implanted MLS ovarian cancer spheroid . S(a) is the mean signal intensity in a region of
high angiogenic activity . S(O) is the mean signal intensity in a control region. AVDMRI =
-In (S(a)/S(O» . B. Photograph of a skin removed at the end of the MRI scan shown in A.
AVDphoto = mean signal intensity in a region of active angiogenesis.

2. ANGIOGENESIS OF IMPLANTED C6 RAT GLIOMA SPHEROIDS

C6 rat glioma cells were cultured in DMEM (Dulbecco's Modified Eagle's Medium)
supplemented with 5% fetal calf serum (FCS, Biological Industries Israel), 50 unit mI-I
penicillin, 50 ug ml- I streptomycin and 125 ug ml- I fungizone (Biolab Ltd .). Aggregation
of cells into small spheroids was initiated in agar coated bacteriological plates . After 4-5
days the spheroid suspension was transferred to aspinner flask (Bell co, USA) and the
medium changed every other day for approx imately 6 weeks as reported previously
(Abramovitch et al., 1995).
Spheroids were implanted in CD1-nude mice (2 month old, male, 30 g body weight) . Mice
were anesthetized with 75 ug s' Ketamine + 3 ug g-I Xylazine (i.p.), and placed in a
sterile laminar flow hood. A single spheroid per mouse, I mm in diameter, was implanted
subcutaneously in the lower back using a Teflon tubing through a 4 mm incision as
reported previously (Abramovitch et al., 1995). The incision was formed by fine surgical
scissors, and closed with cyanoacrylate (SUPER GLUE-3, Loctite, Ireland) or with an
adhesive bandage (Tegaderm'[N, USA).

463
 1000

E •
100
:L
.::;
~

=
~ 10

monclayer vpheroid divsec iated

pheroids

Figure 4 Growth advantage of spheroids in vivo. The capacity to form subcutaneous

tumors in nude mice was evaluated using three different culture variants of C6 rat glioma .
Spheroids of 800 um in diameter (5xl0 4 cells) were implanted as described previously
(Abramovitch et al., 1995) . Cells from log phase monolayer were inoculated
subcutaneously; and cells from dissociated spheroids were cultured in monolayer for 24 h
and then inoculated subcutaneously . Eighteen days after implantation the tumors were
removed and their wet weight was detennined .

2.1 Growth advantage ofC6 glioma spheroids in vivo

Spheroid growth in vitro follows Gompertz rather than logarithmic kinetics. Such growth
kinetics are in accord with the inhibition of cell proliferation as the spheroid grows. While
cells within large spheroids proliferate significantly slower than in monolayer culture in
vitro, an opposite behavior was seen in vivo (Figure 4). In fact, sphero ids show a
significant growth advantage in vivo (Figure 4). Such an advantage of spheroids might be
due to selection of more aggressive subset of cells during the long spheroid culture. We
therefore dissociated spheroids and plated the cell suspension in monolayer culture for 24
h. The cells were then harvested, and a comparable number of cells per mouse as one
spheroid was inoculated (50,000 cells /mouse). The growth ofthese cells in vivo was very
similar to the inoculation of cells from the parent C6 glioma line . Thus the growth
advantage of spheroids in vivo was not due to selection, but rather due to the properties of
the intact spheroid (Figure 4).
Furthennore, the growth advantage observed for spheroids relative to inoculation of a
comparable number of cells collected from logarithmic culture suggests that tumor
progression is not dictated by the rate of cell proliferation . One possible explanation for
this finding could be enhanced angiogenic capacity for spheroids. The
microenvironmental heterogeneity of spheroids can support elevated expression of genes
regulated by hypoxia and glucose deprivation. Following the reports ofthe association of
expression of vascular endothelial growth factor (VEGF) with necrosis (Shweiki et al.,
1992), we tested the patterns of VEGF expression in C6 glioma multicellular spheroids
(Shweiki et al., 1995; Stein et al., 1995).

2.2 Generation of oxygen gradients and their efTect on the angiogenic potential of
spheroids:

The diffusion of water in multicellular spheroids was mapped by diffusion pulsed spin
echo NMR microscopy (Neeman et al., 1991). This type of measurement, detects the
mean displacement of molecules during an experimentally controlled 'diffusion time' .

464
'""'
~ 100 T"""-------------, 1.0
e
-
s 0.8

-
e >
Qj

0.6
10 ~
"'5 (JA 0
>
r...
C 0.2
s:::::l •
0.0
E-
o .. 8 12 16 20 0 .. 8 12 16 20

Time after implantation (Days)

Figure 5 Tumor growth and vascularization. Growth of a C6 glioma spheroid implanted
subcutaneously in nude mice, initiated only 4 days after implantation. This lag in tumor
growth matched the duration of development of a vascular bed around the implanted
spheroid. Tumor growth (Jeft) as weil as vascularization (right) were measured in vivo
using gradient echo MRI. The kinetics of vascular development are consistent with a
negative feedback control , leading to a steady state in vascular .density,

Measurements of water diffus ion report the presence of diffusion barriers that will hinder
translocation of molecules dissolved in water. In the spheroids we found that water
diffusion in the viable rim occurred in two distinct compartments with different diffusion
values but both diffusing significantly slower than free water (2.5x 10- 5 cm 2 s-1) . One
compartment assigned to intracellular water had a diffusion coefficient of 2x 10- 6 cm 2 s-1 ,
10 fold slower than free water. The second compartment assigned to the extracellular
water diffused at 1.7x10- 5 cm 2 s-1 (Neeman et al., 1991). Such slow diffusion coupled
with the consumption of oxygen generates steep grad ients in the concentration of oxygen
across the viable rim . At a certain distance from the spheroid surface nutrient, oxygen
supply and removal ofwaste products are not sufficient to maintain cellular viability, and
necrosis develops.
In situ hybridization maps of spheroids revealed elevated expression of VEGF as weil as
glucose transporter 1 (GLUTI) in the inner layers of the viable rim of the spheroids
(Shweiki et al., 1995 ; Stein et al., 1995) . A similar pattern of VEGF expression was
previously observed in tumor biopsies suggesting that expression of VEGF is upregulated
by metabolic stress. The role of oxygen in this upregulation of VEGF and GLUTl was
consistent with the observation that acute hyperoxygenation of the spheroids led to down
regulation ofboth VEGF and GLUTl.

2.3 Kinetics of tumor vascularization and growth:

As outlined above, one of the primary reasons for employing MRI for following tumor
angiogenesis, is the ability to follow tumor growth and vascularization non invasively, and
determine the dependence of release from growth arrest on the angiogenic switch. Such
kinetic studies were done for C6 glioma spheroids implanted in nude mice (Abramovitch et
al., 1995). An example of one such mouse is shown in Figure 5.
MRI experiments were performed on a horizontal 4.7 T Bruker-Biospec spectrometer
using a 2 cm RF decoupled surface coil and a whole body excitation coil. Mice were
anesthetized with 75 Ilg g-l Ketamine + 3 ug g-l Xylazine (i.p.), and placed supine with

465
the tumor located at the center of the surface coil. Gradient echo images were acquired
with a slice thickness of 0.5-0.6 mm, TE of 20 ms, TR of 100 ms and 256x256 pixels
matrix resulting in resolution of 110 um .
Growth ofthe capillary bed was reflected by reduction ofthe mean intensity at a region of
interest of 1 mm surrounding the spheroid or the incision. Data is reported here as apparent
vessel density (AVD), where AVD = - In (S(a)/S(O)). S(a) is the mean intensity at a region
of interest of 1 mm surrounding the implanted spheroid and S(O) is the mean intensity of a
distant muscle (Figure 3).
Spheroids typically showed a growth delay of 3-8 days followed by Gompertz growth
(Abramovitch et al ., 1995). We found that for each mouse the growth delay matched the
duration in which the apparent vessel density reached a steady state . Approximately 1
week after initiation of growth , the rate of tumor growth reduced significantly (Figure 5).
The kinetic analysis of spheroid growth and vascularization raises the following two
questions: First, why does vascular density reach a steady state 4 days after implantation;
and second, why is the tumor growth rate reduced 10 days after implantation?
Assuming that the spheroid is a source that is continuously producing angiogenic growth
factors , we can predict as a first approximation that the amount of angiogenic promoters
produced, and thus the rate of angiogenesis, will be directly proportional to the number of
viable cells in the tumor. In this case, after vascularization, tumor growth is accelerated,
and as the spheroid grows we see increasing rate of angiogenic activity . If however, the
angiogenic response of the host is limited and reaches a maximal rate with the output of
angiogenic stimuli from the implanted spheroid, we should see a constant rate of
angiogenesis regardless of spheroid growth . Obviously these two options do not agree
with the experimental finding .
The data supports a negative feedback regulation in which establishment ofthe vasculature
around the spheroid leads to down regulation of angiogenesis. Such down regulation can
result either from inhibition in the production of angiogenic stimuli or from the induction
of angiogenic inhibitors. The elevation of expression of VEGF by hypoxia, and down
regulation of VEGF with restoration of normoxia as described earlier are consistent with
this explanation, and indeed we found that VEGF expression in vivo returned to baseline
upon vascularization (Shweiki et al ., 1995).
The second question, namely the reduced tumor growth rate 10 days after implantation,
could perhaps be linked to the first question . The inhibition of angiogenesis 4 days after
implantation creates a vascular bed that can support rapid tumor growth for a while.
Eventually however, the vasculature will not be sufficient and growth will be delayed until
additional vessels form.
In order to respond to these questions, we developed experimental protocols for analyzing
the kinetics of spheroid growth and vascularization in a number of more complex systems
as will be outlined in the following sections.

2.4. Systemic inhibition of angiogenesis by the tumor,

Some tumors secrete systemic inhibitors of angiogenesis that can suppress the growth of
metastases. A number of such inhibitors were identified including thrombospondin
(Bouck, 1996; Dameron et al., 1994), angiostatin (O'Reilly et al., 1994) and endostatin
(O'Reilly et al., 1997). Secretion of such inhibitors by the tumor, explains the induction of
metastatic growth upon removal ofthe primary tumor. Kinetic analysis of spheroid growth
might provide information on the role of this pathway . The two spheroid experiment
follows the following time course : At day 0 the first spheroid is implanted, and its growth
and vascularization are followed. After the vascular bed is established, a second spheroid
is implanted within the MRI detection volume . Both spheroids are now followed to detect

466
 lmplantation oi Removal of
secend spheroid firs t . phcr oid

\ I ~ t- volume
2-volume
.....
<
e
~

E
10

-;
1 ' - ---'-----'---'--'--'---'--'--..................- - '
11 ~ X 12 16 20 2~ 2X 32 36 ~o

Du y s

Figure 6 The two spheroid model of primary-metastatic interaction. This experiment was
designed to determine the influence of a primary tumor on the growth kinetics of a
secondary tumor. A single C6 glioma spheroid was implanted and its growth and
vascularization were followed . Twelve days later a second spheroid was implanted at a
distance of more than 1 cm from the first spheroid . A prolonged growth delay was
observed for this second spheroid. After 10 days, the first spheroid was surgically
removed . Subsequently, the second spheroid showed rapid growth.

any specific inhibition of growth and vascularization of thesecond spheroid. The first
spheroid can now be removed in order to see if this will rescue the second spheroid from
growth arrest. An example of tumor growth in such an experiment is consistent with
systemic inhibition of tumor progression (Figure 6).

2.5. Tissue injury and its effect on tumor angiogenesis and growth:

In addition to metastatic induction upon removal of the primary tumor due to systemic
inhibitors such as angiostatin, surgical removal of tumors can induce tumor progression
through the direct effect of the wound milieu on residual local tumor. The induction of
tumor growth by proximal wounds was studied in a system of a multicellular spheroid
implanted at different distances from an incision (Abramovitch et al., 1996). When
spheroids were implanted at least more than 5 mm from the incision, wound healing
followed the same time course as in the absence of a spheroid, and tumor growth was
relatively slow . On the other hand spheroid growth was accelerated and the initial lag was
shortened when the spheroid was implanted on the incision. In this case, tumor- wound
angiogenesis was significantly higher than for each alone, and the regression of vessels
that occurs in normal wound healing a week after injury, did not occur. 3 weeks after
injury, the wound was still open, and tumor growth was faster than in tumors implanted
further from the wound (Abramovitch et al., 1996).
In addition to the importance of this finding with respect to surgical intervention in cancer
therapy, we see here that the angiogenic capacity ofthe host does not limit the angiogenic
response to an implanted spheroid, and conditions that induce the production of angiogenic
stimuli can have an important impact on tumor progression. Stimulation of tumor
progression in the site of injury is only one side of the coin . The other side includes the
inhibition of wound healing in the presence of a local tumor. The wounds remain in this
case open and reepithelialization, which is usually completed by day 7 after injury, is not
finished for tumor-wounds even 3 weeks after injury. Analysis of vessel density by MRI

467
shows elevated vascular density persisting a long time after injury in these cases. Possibly
the presence of a local tumor inhibits vascular regression that is an essential step in
healing, and maintains the wound in a stage in which it continuously stimulates tumor
progression. The elevated angiogenic activity associated with wound-tumor systems could
potentially provide a c1inical tool for detection ofresidual tumor after surgery.

3. HORMONAL MILIEU AND TUMOR ANGIOGENESIS

Changes in the hormonal milieu affect angiogenesis in a huge number of normal and
pathological examples. Using the implanted spheroid model, we saw that removal of the
ovaries in female CDl nude mice, induces the angiogenic response to human epithelial
ovarian cancer (Schiffenbauer et al., 1996). The effects of ovariectomy can be related to
induction of production of angiogenic growth factor in the ovarian cancer cells in response
to elevated levels of the gonadotropin hormones LH and FSH. This hypothesis is being
studied now for a number of human ovarian cancer cell lines. In contrast with these
findings reduction in the circulating levels of estrogen, achieved by ovariectomy results in
general, in reduced angiogenic response to defined stimuli (Morales et al., 1995).

4. REGULATION OF ANGIOGENESIS IN THE NORMAL OVARY: THE ROLE

OF HYPOXIA, HORMONAL STIMULATION AND WOUND REPAIR

Angiogenesis occurs in the ovary as an integral part of the menstrual cycle, and is thus a
fascinating system for studying the regulation of angiogenesis in normal physiology .
Within the periodic ovarian cycle we can distinguish three major angiogenic events, which
involve perhaps the different angiogenic promoting seenarios outlined earlier for solid
tumors . The ovarian follicle expands significantiy in volume, from the preantral stage to
the mature preovulatory graafian follicle . This process , stimulated by LH leads to
ovulation . Angiogenesis at this stage is restricted to the theca cell layers that surround the
follicle, and inner cell layers rely on diffusion for nutrient and oxygen supply . The
diffusional properties of the follicle show striking similarity to the multicellular spheroid
(Neeman et al., 1997; Tempel et al ., 1995). Moreover, as in the spheroid, VEGF
expression in the follicle is primarily induced in the inner cell layers . This pattern of
VEGF expression in combination with restricted diffusion may suggest a role for hypoxia
as a regulator of angiogenesis at this stage.
Perfusion of the preovulatory cycling ovary has a complex spatial organization . On one
hand the follicles develop large avascular centers which are not perfused and depend on
diffusion . On the other hand, stimulated angiogenesis creates a very dense capillary bed
surrounding the follicles . This spatial heterogeneity in ovarian perfusion was followed by
MRI spin tagging, in which magnetization ofwater in the arterial inflow to the ovaries was
suppressed (Tempel et al., 1997). This study showed that the increased blood flow to the
ovary during follicular development until ovulation exactly matched the increased
follicular volume .
Ovulation marks the initiation of two additional angiogenic processes. The tear formed in
the follicular wall as weil as in the outer epithelium ofthe ovary, during ovulation, invokes
angiogenesis associated with wound repair. It is important to note here that the frequent
and repetitive induction of repair with ovulations may be responsible for the initiation of
malignant transformation of the ovarian epithelium. Following ovulation, development of
the corpus luteum results in a third wave of angiogenesis.

468
Figure 7 Pathways accelerating the angiogenic switch in solid tumors. Transition of
tumors from the avascular to the vascular phase can be induced by a number of pathways
including a genetic change to constitutive over expression of angiogenic growth factors;
induction of angiogenesis by hypoxie stress; interaction with proximal wounds; and
induction of angiogenesis associated with a change in the hormonal milieu .

5. DISCUSSION

We saw here a number of different physiological pathways that can promote the
angiogenic switch in tumors (Figure 7). Hypoxia, developing very early in three
dimensional solid tumor nodules, proximal injuries either accidental or associated with
treatment, as weil as changes in the hormonal milieu, can all initiate the transition and
invoke initiation of growth of previously dormant tumors . The same regulatory
mechanisms function also in normal physiology for matehing nutrient supply to demand
and for vascular remodeling. The ovary provides an excellent system fo r studying the
regulation ofthese angiogenic pathways in normal physiology .
The different pathways promoting the angiogenic switch suggest that tumors may be able
to overcome or become resistant to a specific antiangiogenic therapy . One possible
resistance mechanism may include massive overproduction of angiogenic growth factors .
Another mechanism could include induction of pericytes or smooth muscle cell migration
so as to obtain a more mature, and less sensitive vascular bed. A third mechanism might
include a selection to a more invasive growth pattern, in which tumor cells will grow along
the existing vascular bed replacing normal tissue . The theoretical possibility of these

469
pathways should be taken into consideration in the development of antiangiogenic therap y,
and also calls for the development of efficient methods for patient follow-up .
The wound-tumor system presents a major dilemma with respect to ant i angiogenic
therapy . Angiogenesis is an essential and integral part of the healing process. If
ang iogenesis in wound healing and in tumors share a common molecular pathway,
inhibition of one may affect the other. On the other hand it is possible that the excessive
angiogenic activity associated with tumor-wounds is beyond what is required for healing,
and tissue recovery could take piace even if angiogenesis is significantl y inhibited . Lastly ,
it is conceivable that wound healing as a cardinal life supporting process will include a
number of compensatory alternative pathways, and wound angiogenesis will occur despite
inhibition of the activity of a particular angiogenic factor . Tumors may be limited to a
single angiogenic pathway, and may thus show increased sensitivity to therapy. The
embryonie lethality of heterozygous VEGF knockouts (Carmeliet et al., 1996; Ferrara et
al., 1996) suggests that this may not be the case and VEGF may be a critical common
signal mediating angiogenesis in both systems. On the other hand however, the anti-
neovascular activity of CM101, a toxin produced by group B Streptococcus, was effective
against tumors (Hellerqvist et al., 1993; Thurman et al., 1996; Thurman et al. , 1994;
Warnil et al., 1997) but did not inhibit wound healing (Quinn et al., 1995).
Elucidation of the kinetics of angiogenesis provides a powerful tool for analysis of the
regulation of vascular modeling in vivo . The importance of three dimensional tumor
growth was a major consideration for the development ofthe experimental system reported
here . Tumor microenvironments including hypoxia, hypoglycemia and necrosis can be
markedly affected by tumor geometry , and will thus affect the extent and the kinetics of
angiogenesis. MRI provides an elegant research tool for non invasive monitoring of tumor
growth and angiogenesis. Moreover, MRI can in principle detect a tumor implanted at any
location, so studies are less restricted as with other experimental systems .
Beyond the application of MRI as a research tool for studying ang iogenesis and
antiangiogenic therapy as described here, MRI can provide an important tool for clinical
determination of angiogenesis that can be important for diagnosis, prognosis and follow-up
of antiangiogenic therapy .

6. REFERENCES

Abramovitch, R., Marikovsky, M., Meir, G. and Neeman , M., 1996, Direct and indirect
mechanisms for acceleration of tumor growth by wounds : NMR studies of implanted C6
glioma spheroids, in: ISMRM, Vol. 1. pp. 369: New York .

Abramovitch, R., Meir, G. and Neeman, M., 1995, Neovascularization induced growth of
implanted C6 glioma multicellular spheroids : magnetic resonance microimaging. Cancer
Res. 55: 1956-1962 .

Acker, H., Carlsson, J., Mueller Klieser, W. and Sutherland, R. M., 1987, Comparative
p02 measurements in cell spheroids cultured with different techniques. Br .J Cancer . 56:
325-327 .

Borgstrom, P., Hillan, K. 1., Sriramarao , P. and Ferrara, N ., 1996, Completc inhibition of
angiogenesis and growth of microtumors by anti-vascular endothelial growth factor
neutralizing antibody : novel concepts of angiostatic therapy from intravital
videomicroscopy. Cancer Res. 56: 4032-4039.

Bouck, N ., 1996, PS3 and angiogenesis . Biochim Biophys Acta. 1287: 63-66 .

470
Boxerman, 1. L., Hamberg, L. M ., Rosen, B. R. and Weisskoff, R. M ., 1995 , MR contrast
due to intravascular magnetic susceptibility perturbations. Magn Reson Med. 34: 555-566.

Carmeliet, P ., Ferreira, V., Breier, G., Pollefeyt, S., Kieckens, L. , Gertsenstein, M ., Fahrig,
M ., Vandenhoeck, A., Harpal, K ., Eberhardt, C., Declercq, C., Pawling, J. , Moons, L. ,
CoIIen, D. , Risau, W . and Nagy, A., 1996, Abnormal blood vessel development and
lethality in embryos lacking a single VEGF allele. Nature. 380: 435-439.

Casciari , 1. 1., Sotirchos, S. V . and Sutherland, R . M ., 1988, Glucose diffusi vity in

multicellular tumor spheroids. Cancer Res . 48: 3905-3909.

Dameron, K. M. , Volpert, O . V., Tainsky, M . A . and Bouck, N ., 1994, Contral of

angiogenesis in fibroblasts by p53 regulation of thrombospondin-1. Science . 265: 1582-
1584 .

FarrelI , C. L., Stewart, P. A and DeI Maestro, R. F., 1987, A new glioma model in rat: the
C6 spheroid im plantati on technique permeability and vascular characterization. J
Neurooncol. 4: 403-415 .

Ferrara, N ., Carver Moore, K , Chen, H. , Dowd, M ., Lu, L., O'Shea, K S., Powell Braxton,
L. , Hillan, K J. and Moore, M . W ., 1996, Heterozygous embryonie lethality induced by
targeted inactivation ofthe VEGF gene. Nature. 380: 43 9-442.

Hanahan, D . and Folkman, J ., 1996 , Patterns and emerging mechanisms of the angiogen ic
switch during tumorigenesis. Ce/I. 86: 353-364.

HeIIerqvist, C. G ., Thurman, G . B., Page, D. L. , Wang, Y. F., RusseII , B. A , Montgomery,

C . A . and SundeII , H . W ., 1993 , Antitumor effects of GBS to xin : a polysaccharide
exotoxin from group B beta-hemolytic streptococcus. J Cancer Res Clin Oncol. 120: 63-
70 .

Karczmar, G . S., River, 1. N ., Li, 1., Vijayakumar, S., Goldman, Z . and Lewis, M . Z ., 1994 ,
Effects of hyperoxia on T2* and resonance frequency weighted magnetic resonance
images ofrodent tumours. NMR Biomed. 7: 3-11.

Morales, D . E., McGowan, K . A, Grant, D. S., Maheshwari , S., Bhartiya, D ., Cid, M . c.,
Kleinman, H . K. and Schnaper, H. W ., 1995 , Estragen promotes angiogenic activity in
human umbilical vein endothelial cells in vitro and in a murine model. Circu/ation . 91:
755-763 .

Mueller Klieser, W . and Sutherland, R . M ., 1983, Frequency distribution histograms of

oxygen tensions in multicell spheraids. Adv Exp Med Bio/ . 159: 497-508 .

Neeman, M ., Abramovitch, R ., Schiffenbauer, Y. and Tempel, c., 1997, Regulation of

angiogenesis by hypoxie stress: from solid tumours to the ovarian follicle. Int J Exp
Pathol. 78 In press.

Neeman, M ., Jarrett, K . A., SilIerud, L. O. and Fre yer, J. P., 1991, Self-diffusion ofwater
in multicellular spheraids measured by magnetie resonance microimaging. Cancer Res . 51:
4072-4079.

O'Reilly, M. S., Boehm, T. , Shing, Y., Fukai , N ., Vasios, G., Lane, W. S ., Fl ynn, E.,
Birkhead,1. R ., Olsen, B . R. and Folkman, 1., 1997, Endostatin : an endogenous inhibitor of
angiogenesis and tumor growth . Ce/I. 88: 277-285 .

471
O'Reil1y, M . S., Holmgren, L., Shing, Y., Chen, c., Rosenthai, R. A, Moses, M ., Lane, W .
S., Cao, Y., Sage, E. H. and Folkman, 1., 1994, Angiostatin: a novel angiogenesis inhibitor
that mediates the suppression of metastases by a Lewis lung carcinoma, Cell o79: 315-328.

Ogawa, S., Lee, T. M ., Kay, A. R. and Tank, D. W ., 1990a, Brain magnetic resonance
imaging with contrast dependent on blood oxygenation. Proc Natl Acad Sei USA 87:
9868-9872.

Ogawa, S., Lee, T. M ., Nayak, A S. and Glynn, P., 1990b, Oxygenation-sensitive contrast
in magnetic resonance image of rodent brain at high magnetic fields. Magn Reson Med.
14: 68-78 .

Quinn, T. E ., Thurman, G. B ., Sundell, A. K., Zhang, M . and Hellerqvist, C. G., 1995,

CM101 , a polysaccharide antitumor agent, does not inhibit wound healing in murine
models. J Cancer Res Clin Oncol. 121: 253-256.

Robinson, S. P ., Howe, F. A. and Griffiths, 1. R ., 1995, Noninvasive monitoring of

carbogen-induced changes in tumor blood flow and oxygenation by functional magnetic
resonance imaging. Int J Radiat Oncol Biol Phys. 33: 855-859.

Schiffenbauer, Y . S., Abramovitch, R ., Meir , G., Nevo, N., Holzinger, M ., Itin, A, Keshet,
E. and Neeman, M ., 1996, Gonadotropin contribution to the angiogenic potential of MLS
spheroids. 31P NMR spectroscopy and MR microscopy, in: ISMRM, Vol. 1. pp . 371 : New
York.

Shweiki, D., Itin, A., Soffer, D. and Keshet, E., 1992, Vascular endothelial growth factor
induced by hypoxia may mediate hypoxia-initiated angiogenesis. Nature . 359: 843-845.

Shweiki, D ., Neeman, M ., Itin, A and Keshet, E., 1995, Induction ofvascular endothelial
growth factor expression by hypoxia and by glucose deficiency in muIticell spheroids:
implications for tumor angiogenesis. Proc Natl Acad Sei USA . 92: 768-772.

Stein, 1., Neeman, M., Shweiki, D. , Itin, A and Keshet, E., 1995, Stabilization ofvascular
endothelial growth factor mRNA by hypoxia and hypoglycemia and coregulation with
other ischemia-induced genes. Mol Cell Biol. 15: 5363-5368.

Sutherland, R . M. , 1988, Cell and environment interactions in tumor microregions: the

muIticell spheroid model. Seience . 240: 177-184 .

Tempel, C., Meir, G. and Neeman, M., 1997, Modulation ofblood flow in the preovulatory
rat ovarian follicle, in : ISMRM, Vol. 2. pp. 986 : Vancouver.

Tempel, C., Schiffenbauer, Y. S., Meir, G. and Neeman, M ., 1995, Modulation of water
diffusion during gonadotropin-induced ovulation: NMR microscopy of the ovarian follicle.
MagnResonMed. 34: 213-218. .

Thurman, G. B ., Page, D. L., Wamil, B. D., Wilkinson, L. E., Kasami, M . and Hellerqvist,
C. G., 1996, Acute inflammatory changes in subcutaneous microtumors in the ears ofmice
induced by intravenous CM101 (GBS toxin) . J Cancer Res Clin Oncol. 122: 549-553 .

Thurman, G. B ., Russe!, B. A, York, G. E., Wang, Y. F., Page, D . L., Sundell, H. W . and
Hellerqvist, C. G., 1994, Effeets of group B Streptococcus toxin on long-term survival of
mice bearing transplanted Madison lung tumors. J Cancer Res Clin Oncol. 120: 479-484.

472
Torres Filho, I. P., Hartley Asp, B. and Borgstrom, P., 1995, Quantitative angiogenesis in a
syngeneic tumor spheroid model. Microvasc Res . 49: 212-226 .

Waleh, N. S., Brody, M. D., Knapp, M . A., Mendonca, H. L., Lord, E. M., Koch, C. J .,
Laderoute, K. R. and Sutherland, R. M., 1995, Mapping of the vascular endothelial growth
factor-producing hypoxie cells in multicellular tumor spheroids using a hypoxia-specific
marker. Cancer Res . 55: 6222-6226.

Wamil, B. D., Thurman, G. B., SundelI, H. W., DeVore, R. F., Wakefield, G., Johnson, D.
H., Wang, Y. F. and Hellerqvist, C. G., 1997, Soluble E-selectin in cancer patients as a
marker of the therapeutic efficacy of CMlOl, a tumor-inhibiting anti-neovascularization
agent, evaluated in phase I clinical trial. J Cancer Res Clin Oncol. 123: 173-179 .

Weidner, N. and Folkman, 1.,1996, Tumoral vascularity as aprognostic factor in cancer.

Important Adv Onco1167-190.

473
CLINICAL TRIALS OF ANGIOGENESIS-BASED THERAPIES: OVERVIEW AND
NEW GUIDING PRINOPLES

William W. Li , M.D., Vincent W. Li, M.D., Richard Casey, M.D. ,

Dimitris Tsakayannis", M.D., Erwin A. Kruger, Andrew Lee, Yee-Li
Sun, Christopher A. Bonar, Y.M.D., and Shawna Comelius.

The Angiogenesis Foundation, P.O. Box 383011, Cambridge,

Massachusetts 02238 U.S.A.
*The Angiogenesis Foundation, Europe, 113 Marathonodromon Street
15452 Athens GREECE

1. INTRODUCTION

Angiogenesis, new blood vessel growth, is a "common denominator" pathologie

feature in many serious diseases, including cancer, coronary artery disease , stroke,
blindness, arthritis, and more than one dozen other major conditions. Against the backdrop
of largely unsatisfactory current treatments for these conditions, a new form of medical
therapy aimed at controlling angiogenesis is emerging after three decades of basic scientific
research. In simplest terms, angiogenesis modulating drugs restore growth control over the
vascular system - by either turning "on" or tuming "off' angiogenesis. A number of
angiogenesis-modulating compounds, both stimulators and inhibitors, are currentl y in
c1inical trials in the United States, Canada and Europe. The reality of these trials, and some
exciting early positive results , heraids an exciting era of medical therapy for the new
millennium, and offers renewed hope for patients suffering from currently incurable
diseases such as cancer .
The Angiogenesis Foundation - a private, nonprofit institution created to coordinate
global efforts to develop new angiogenesis-based therapies - has undertaken a systematic
study and "real-time" analysis of the progress , challenges, and hurdles encountered by
c1inicians, patients and pharmaceutical firms during the current, early stage of drug
development. This report , based upon an Angiogenesis Foundation-sponsored study , was
presented at the 1997 NATO Advanced Study Institute symposium on angiogenesis, held
in Rhodes , Greece.
The following discussion is not intended to be a comprehensi ve review of the science
of angiogenesis, but rather a broad overview of the current c1inical trials of angiogenesis-
based treatments and the therapeutic insights g1eaned from them. Based on the lessons
leamed so far, the Angiogenesis Foundation is advancing ten recommendations for

Angiogenesis: Models, Modulators, and Clinical App lications

Editcd by Maragoudak is, Plenum Press, Ncw York, 1998 ~7 5
 Table 1. Known angiogenic factors

Fibroblast growth factors : acidic (aFGF) and basic (bFGF)

Vascular endothelial growth factor (VEGF)/vascular permeability factor (VPF)
Transforming growth factor-alpha (TGF-alpha)
Transforming growth factor-beta (TGF-beta)
Tumor necrosis factor-alpha (TNF-alpha)
Angiogenin
Hepatocyte growth factor (HGF) /scatter factor (SF)
Interleukin-8 (IL-8)
Granulocyte colony-stimulating factor (G-CSF)
Platelet-derived endothelial cell growth factor (PD-ECGF)
Pleiotrophin (PTN)
Proliferin
Placental growth factor
Follistatin
Midkine

developers of angiogenesis-based therapeutics to facilitate drug development progress in

this area.

1.1. A Historical Perspective

Angiogenesis-based therapies regulate the nourishment of tissues by controlling their
blood supply . As long as 400 years ago, physicians recognized the critical dependence of
tissues, both healthy and diseased, on blood vessels. Leonardo da Vinci observed in Anal.
Ms. B. fol.28r that, "the body of anything whatsoever that receives nourishment
continually dies and is continually renewed ... unless therefore you supply nourishment
equivalent to that which has departed, the life fails in its rigors; and ifyou deprive it of this
nourishment, the life is completely destroyed."
The term "angiogenesis" was coined in 1935 by Harvard anatomist Arthur Tremain
Hertig to describe the neovascularization he observed in the placenta and in the developing
embryo of monkeys 19. Thirty-six years later, Judah Folkman, a pediatric surgeon from
the same institution, observed similar growth of new capillary blood vessels associated
with solid tumors and he proposed a then-avant-garde hypothesis that "tumor growth is
angiogenesis-dependent."10. Greeted with initial skepticism, Folkman then launched a
career-spanning research prograrn, assisted by numerous colleagues, post-doctoral fellows
and medical students, that led to: (i) the purification of the first angiogenic cytokine
molecule; (ii) the discovery of numerous inhibitors of angiogenesis; (iii) the cloning of
capillary endothelial cells; (iv) innovative experimental laboratory assays that enable the
study of new blood vessel growth in tissue culture and in anlmal models; and more
recently , (v) the discovery of endogenous antiangiogenic molecules that are sufficiently
potent to regress human tumors to microscopic size and to a dormant state 21,22,35.
These fundamental discoveries have ignited global interest and participation in the study of
angiogenesis.
Currently more than three thousand laboratories and one hundred and twenty
biopharmaceutical companies are conducting various types of angiogenesis research. At
least 15 angiogenic cytokine factors have been sequenced and cloned.
More than three hundred inhibitors of angiogenesis have been reported . In 1989, the
first patient was successfully treated with the antiangiogenic drug interferon alpha-2a 48.
In 1997, the first complete response of a human cancer, metastatic cervical carcinoma, to

476
antiangiogenic therapy was reported 25. As of this writing, more than 140 medical centers
are engaged in the clinical testing of some twenty-five new drug candidates, in diseases
ranging from cancer to diabetic blindness, from rheumatoid arthritis to coronary artery
disease.

1.2. Tbe Contemporary Perspective: The Angiogenesis Model for Homeostasis

It is now recognized that the process of angiogenesis is critical to health and
homeostasis in humans . It is often stated that angiogenesis in adults occurs only in
response to wounding, and in females, during the reproductive cycle. Growing evidence,
however, points to angiogenesis as a more dynamic state. Angiogenesis occurs in skeletal
muscIe in response to exercise and training, and in tissues, such as the myocardium , that are
being actively remodeled under conditions ofhypoxia and stress. Furthermore, there are at
least three distinct mechanisms for blood vessel formation in the adult: sprouting
angiogenesis-, intussusceptive microvascular growth 5, and vascular formation from
circulating endothelial progenitor sterns cells 2,40. Existing dogma states that vascular
endothelial cells are inactive, or quiescent. Yet, in the human body, endothelial cells are
literalIy bathed in a sea of pro-angiogenic molecules (including cytokines , prostaglandins,
eiconsanoids, and fat-derived molecules) and anti-angiogenic modulators (including
thrombospondin-I , angiostatin, and endostatin) 12.
We feel that a holistic and operational model for angiogenesis may serve as a more
useful guide for designing new therapeutic strategies. This new model, which we term the
"Angiogenesis Model for Homeostasis" describes the folIowing schema: "the healthy state
reflects an exquisite physiological balance of endogenous positive and negative regulators of
angiogenesis . Loss ofthis balance - resulting in either excessive or insufficient new blood
vessel growth - leads to disease or its progression" 6. Accordingly, restoration of the
balance leads to restoration to the healthy state.
New angiogenesis modulator drugs may thus be considered from an operational point
of view : they either inhibit excessive angiogenesis (in order to deprive diseased tissues of a
blood supply) , or stimulate angiogenesis (in order to nourish and repair damaged tissues),
with the goal ofrestoring vascular homeostasis and retuming a patient to health. With this
concept in mind, we will now review the existing clinical trials of angiogenesis-based
therapies.

2. OVERVIEW OF ANGIOGENESIS-BASED THERAPIES IN CLINICAL TRIAL

The current clinical applications of angiogenesis research fall into three broad
categories: (I) stimulation of angiogenesis; (2) inhibition of angiogenesis; and (3) the use of
angiogenic markers, such as microvessels and angiogenic cytokines, as tools to aid in the
prognosis of patients .

2.1. Tberapeutic Stimulation of Angiogenesis: Molecular Revascularization

Angiogenic growth factors (cytokines) are being tested in clinical trials to stimulate
therapeutic angiogenesis in patients sufTering from hypoxic tissues . This approach, which
we term "molecular revascularization" aims to achieve the necessary physiologie response
(increased perfusion to tissues) without the risks inherent in a major surgical operative
procedure.
Four ofthe 15 known angiogenic cytokines are available in recombinant form and are
now in cIinical trials: acidic fibroblast growth factor (aFGF); basic fibroblast growth factor
(bFGF) , vascular endothelial growth factor (VEGF), and platelet-derived growth factor
(PDGF).

477
 Four major diseases are now being treated m clinical trials of molecular
revascul arization:

(1) Ischemic heart disease . Cardiologists and cardiothoracic surgeons are now delivering
recombinant acidie FGF , basic FGF and VEGF to the hearts of patients with severe
inoperable coronary artery disease. The clinical goal is stimulate neovascularization and
collateral channels to bypass atherosclerotic vessels and perfuse the areas of heart at risk
for infarcti on 18,28,49.

(2) Severe peripheral vascular disease. Vascular surgeons and cardiologists are delivering
recombinant VEGF and its naked DNA to the ischemic limbs of patients with severe,
inoperable peripheral vascular disease. The clinical goal is to stimulate angiogenesis and
increase blood perfusion in limbs that face amputation due to the threat of gangrene
resulting from atherosclerotic blockages 23,38.

(3) Non-healing diabetic ulcers Surgeons are using topical preparations of recombinant
human PDGF, basic FGF , and sucralfate to treat the non-healing ulcers on the feet of
diabetic patients . The clinical goal is to accelerate wound angiogenesis to speed closure and
healing of the ulcers 4 1,44. Bercaplermin (PDGF) is the first FDA-approved topical
cytokine for the treatment of poorly healing skin ulcers.

(4) Peptic ulcer disease . Gastroenterologists are delivering oral cocktails of acid-stable
bFGF to the stornach and intestinal tracts of patients with refractory gastric ulcers . The
clinical goal is to stimulate angiogenesis in the ulcer bed and promote wound healing 14.

2.2 Therapeutic Inhibition of Angiogenesis: Antiangiogenesis

Antiangiogenic drugs are the focus of most biopharmaceutical companies. The
primary target of antiangiogenic therapy is human cancer. Ophthalmie, rheumatologie,
pediatric and AIDS applications are also under development.

(1) Cancer. Oncologists at over 140 medical centers throughout North America and
Europe are testing 20 antiangiogenic drugs in patients with a wide variety of solid
tumors 36 . Several of the agents are well known to angiogenesis researchers , such as
interferon-alpha, TNP-470 (AGM-1470), thalidomide, CM101, Batimastat, and
Marimastat. Other Iesser-known agents in clinical trial include: razoxane (a topoisomerase
11 inhibitor), flavopiridol (an analog of the natural isoflavanoid genistein), and CAI
(carboxy-amido imidazole, an inhibitor of cellular calcium influx in endothelial cells). Table
2 contains a fulliist of compounds in trial as of this writing .
The clinical goal is tumor regression or stabilization. Currently, these agents are being given
in Phase 1 and 2 studies to patients with metastatic cancer, who have failed conventional
cytotoxic chemotherapy or radiation therapy.

(2) Ocular neovascularization. Ophthalmologists are testing interferon alpha,

thalidomide, and ISV-120 (Batimastat) in patients suffering from age-related macular
degeneration , diabetic retinopathy, and recurrent pterygiumf. The clinical goal is
stabilization ofthe disease, or regression of ocular neovascularization.

(3) Hemangiomas of childhood. Pediatricians, dermatologists and plastic surgeons are

now administering interferon alpha to children suffering from massive, or life- or vision-
threaten ing hemangiomas. This therapy is now widely accepted for lesions that are

478
 Table 2. Antiangiogenic agents that have entered c1inical trial"

TNP -470 (AGM-1470)

Thalidomide
Vitaxin
Interleukin-12
Marimastat
ISV-l20 (Batimastat)
Carboxy-amido imidazole (CA!)
Paclitaxel (Taxol, Docetaxel, Paxene)
Neovastat
NEORETNA
Interferon alfa-2a
SUIOI
SU5416
Replistatin (platelet factor-4)
Pentosan polysulfate
Suramin
Octreotide
Tecogalan sodium
Razoxane
Flavopuridol

('Source : Clinical Trials Registry , The Angiogenesis Foundation)

unresponsive to conventional steroid treatment. The c1inical goal to regress these benign
vascular tumors has been achieved 9,42.

(4) AIDS -Kaposi's sarcoma. AIDS specialists are administering hCG (human
choriogonadotrophin), Captopril (an angiotensin converting enzyme inhibitor) , TNP-470,
thalidomide, and taxol to patients with Kaposi's sarcoma. The c1inical goal is regression of
these vascular lesions I6,37 ,47

(5) Rheumatoid arthritis. Rheumatologists are testing thalidomide and paclitaxel in

patients with rheumatoid arthritis . The c1inical goal is reduction of joint pain and swelling
by inhibiting the angiogenesis that accompanies the invasive rheumato id pannus 11,30,34.

2.3 Clinical Use of Angiogenesis Markers

Two types of markers for angiogenesis are presently being evaluated for potential
c1inical use. The quantitative detection of angiogenic cytokines in human body fluids is
being developed to moniter the progress of antiangiogenic therapy . In addition , specific
histochemical stains are being used to detect and quantify (as microvessel density) the
degree of angiogenesis present in a tumor biopsy specimen. Both applications have been
shown to have c1inical correlates in human disease. For example, in a prospective study,
children with brain tumors were found to have basic FGF circulating in their cerebrospinal
fluid. The presence of bFGF correlated with biologie activity , increased intratumoral
angiogenesis, and poor prognos is-P. Elevated basic FGF was also found in the serum of
patients with a wide variety of cancer, including those ofthe brain, breast, lung, colon, head

479
and neck, and genitourinary tract 13. Cancer patients also excrete abnorrnally elevated
amounts of bFGF in their urine 33.
The serial quantification ofurinary bFGF has also been used assist the monitoring of
therapy in children with massive hemangiomas being treated with interferon alpha-2a.
Response to treatment appears to be correlated with a decline in urinary bFGF levels,
suggesting that bFGF maya useful surrogate marker for this disease 13 Other angiogenic
factors, such as VEGF are also under investigation for their potential use as surrogate
markers, but areas yet ofunproved clinical utility.

3. EMERGING INSIGHTS INTO ANGIOGENESIS-BASED THERAPIES

Some important insights are emerging about the nature of angiogenesis-based

therapies, as clinicians begin to gather data from phase 1/2 clinical trials of drugs that
control blood vessel growth. The Foundation believes it is critical to analyze these insights
and pay heed to the operational conclusions derived from them, so that appropriate
adjustments and crucial directional changes may be made along the lengthy and costly path
of translating basic science into successful clinical treatments . As the biotechnology
industry has demonstrated in the 1980's with the failure of "anti-sepsis" drugs, the
pharrnaceutical development pathway is paved with pitfaJls. Drug candidates that work
weJl in the laboratory often fail to have the desired effect in human patients.
To help accelerate the prospects of developing safe and effective angiogenesis-based
therapies, the Angiogenesis Foundation is developing risk reduction strategies for
scientists, clinicians, corporations, and regulatory bodies that are navigating the obstacles to
drug design and therapeutic development. These strategies are based upon the following
ten criticallessons:

3.1. Lesson No. 1: Disease Selection is Critical

The success of a particular antiangiogenic drug, interferon alpha-2a, has been
criticaJly dependent upon the disease selected for clinical trial. For example, the first
disease selected was pulmonary hemangiomatosis, a condition characterized by vascular
tumors affecting the lung. Interferon alpha-2a was highly effective in regressing the
vascular lesions and reversing this condition48. Subsequently, other hemangiomas of
childhood, including massive life-threatening or vision-threatening lesions affeeting the liver
and orbit, respectively, have responded weil to interferon therapy9,29 .
As a result of the hemangioma studies, ophthalmologists began to administer
interferon therapy to patients suffering vision loss from age-related macular degeneration
(AMD) . Despite initial case reports of success in a few individuals, a large, randomized,
prospective study of interferon for macular degeneration failed to reveal any evidence for
efficacy by either visual testing or fluorescein angiography15,17. The reason for this
failure remains unknown, although it has been speculated that interferon may be too weak
an angiogenesis inhibitor, or that the process of subretinal neovascularization in macular
degeneration has some unique, resistant property to interferon .
However, retinal neovascularization in diabetic patients does appear to respond to
interferon treatment. Skowsky and colleagues reported that their patients with
proliferative retinopathy experienced stabilization of their disease after receiving
interferon 43 . But, the mode of drug administration was different than in ARMD , and this
was reported only in a smaJl number of patients.
Clearly, more research is needed to understand the fundamental mechanisms of the
angiogenic switch in each disease. Furtherrnore, the molecular mechanisms of each drug

480
require elucidation, so that the selection of drug and disease for c1inical trials can be more
precise and systemat ic.

3.2. Lesson No. 2: Multiplicity of Angiogenic Factors

Diseased tissues are known to contain a diverse array of angiogenic cytokines. Yet,
the scientific literature reflects the trend for investigators to focus on a particular angiogenic
factor, such as basic FGF or VEGF or TNF-alpha, in studying pathologic angiogenesis.
The work ofRelf and colleagues suggests that this "tunnel vision" approach may be
misleading, especially as the basis for the rational design of antiangiogenic agents for
cancer 39 In their study, biopsy specimens from 64 patients with primary breast cancer
were examined for seven angiogenic factors: VEGF, acidic and basic FGF, TNF -beta, PD-
ECGF , placenta growth factor, and pleiotrophin . Interestingly, a11 patients expressed at
least six of the seven angiogenic factors they tested, and significantly, the only factor
expression that correlated with microvessel density in the breast cancer biopsy , was PD-
ECGF (thymidine phosphorylase). This type of study teaches us that human cancers
likely use redundant mechanisms to stimulate tumor angiogenesis, and that antagonism of a
single growth factor may not be sufficient for therapy . Rather, a "multiple warhead"
approach might ultimately be required to optimize success. A case study of this potential
problem exists in the field of ophthalmology, where considerable scientific investment has
been placed on labeling VEGF as the "culprit molecule" in ocular neovascularization i-.
VEGF has been described as "necessary and sufficient" to induce angiogenesis in the eye 46
Studies showed that neutralization of VEGF (by a monoclonal antibody ) can inhibit
angiogenesis in experimental models of eye neovascularization1.
Yet, historically it has been demonstrated that ocular tissues contain a diverse array
of angiogenic factors, including: basic FGF (tears), EGF, TGF-alpha, TGF-beta, acidic and
basic FGF, IL-1 (cornea); HGF, KGF, basic FGF, EGF, TGF-beta1 (trabecular
meshwork) ; basic FGF and VEGF (vitreous humor), basic FGF (retinal pigment
epithelium, choroid) . Thus, a drug directed solely against VEGF may be inadequate to
neutralize the angiogenic effects of other cytokines, and therefore generate a c1inical
response . Additionally, a major obstacle to the development of ocular antiangiogenic
therapies has been the lack of reliable and convincingly relevant animal models for human
neovascular eye disease.

3.3. Lesson No. 3: Toxicity of Therapy

There is a common misconception that antiangiogenic therapies are "nontoxic". This
has arisen, in part, from the reports of angiogenesis researchers in basic science laboratories
because of the absence of obvious adverse effects in animals receiving experimental
angiogenesis inhibitors. Nevertheless, in human clinical trials, toxicities have been observed
with virtually a11 angiogenesis inhibitors tested. Such adverse reactions are expected,
because of the dosing strategies being employed to find the so-called Maximal Tolerated
Dose (MTD) , and indeed, traditional Phase 1 c1inical trials are designed specifically to
reveal the type, magnitude and reversibility of toxicities that can be caused by a new drug
compound. Table 3 lists the major side effects that have been observed to date with several
compounds.

3.4. Lesson No. 4: Host Response to Treatment

The response time to antiangiogenic drugs in cancer patients appears to differ
significantly from conventional chemotherapeutic agents. Based on c1inical experience with
interferon a1fa-2a for treating infantile hemangiomas, and with TNP-470 for treating solid
tumors, it appears that after initiation of treatment there is aperiod of continued (albeit

481
 Table 3. Some associated toxicities observed in clinical trials of antiangiogenic
agents"

Flu-l ike symptoms

Depression
Fatigue
Leukopenia
Thrombocytopenia
Cerebellar ataxia
Sensory neuropathies
Sustained paralysis
Spastic diplegia
Nausea
Muscle spasms
Sedation
Anticholinergic etTects
Anorexia
Fever
Hypertensi on
Headache
Liver function abnormalities
Anemia

r*Source: Clinical Trials Registry, The Angiogenesis Foundation)

slowed) growth and progression of the tumor. This period of disease progression may
persist for weeks before tumor growth plateaus and then regresses. For example, a
complete response to TNP-470 was obtained in a patient with metastatic cervical
carcinoma only after 18 weeks of intravenous treatment->. Regression of hemangioma
lesions after initiating interferon alfa-2a also appears to take weeks if not months of
therapy l-'.
Insight into the delayed host response to antiangiogenic therapy is critically
important because ofthe conventional standards for evaluating new cancer chemotherapies.
Oncologists generally gauge success of a drug by the rapidity with which it induces a tumor
response . Iftumor progression in a patient is noted after a new treatment is initiated , the
standard protocol is to discontinue the drug and declare the trial a failure in that patient.
The clinical trial design for testing antiangiogenic therapy may require a different paradigrn:
in order to obtain tumor regression in response to angiogenesis inhibition, it may be
necessary to continue treatment for a prolonged period of time, even in the face of
continued tumor progression. Such a delayed host response to treatment may be due to the
indirect etTectof angiogenesis inhibition on tumor cells. Antiangiogenic therapy may have
to generate a sufficient "critical mass" etTect on depressing the tumor microcirculation
before tumor cells become hypoxie and apoptotic . Therefore , early withdrawal of
antiangiogenic therapy appears to be inadvisable .

3.5. Lesson No. 5: Importance of the Interdisciplinary Approach

Modem medical research is characterized by its narrow focus and its specialization,
traits that may result in missed data and inefficient use of resources . Based on the
Foundation's survey of clinicians testing antiangiogenic therapy in cancer patients, it has
become evident that most clinical trial protocols involve only oncologists . But the
enrollment criteria have allowed many studies to include cancer patients sutTering from

482
other angiogenic conditions, including diabetes, psonasis, rheumatoid arthritis, and
coronary artery disease . Such patients receiving an antiangiogenic agent, like TNP-470 or
thalidomide, offer several platforms from which to evaluate a drug's c1inical effects. For
example, oncologists should be collaborating with ophthalmologists to examine the drug's
effects on pre-existing diabetic retinopathy, or collaborating with dermatologists to examine
psoriasis lesions, or with the rheumatologist to look for possible beneficial effects on
arthri tic joints.
An interdisciplinary approach to testing angiogenesis therapies is important because
one drug might fail to be an etTective anticancer agent, but might demonstrate a potent
etTect on suppressing diabetic retinopathy, or psoriasis, or arthritis . Such serendipitous
therapeutic etTects should be systematically looked for by combining multiple c1inical
disciplines in the c1inical trial design of antiangiogenic agents. Given the considerable
expense of drug development, failure of the drug in one discipline can easily lead to its
abandonment altogether, unless c1inicians take an interdisciplinary approach .

3.6. Lesson No. 6: Unexpected Drug Interactions

It appears that certain common medications may interfere with angiogenesis-based
therapies . Arecent c1inical study reporting failure of VEGF gene therapy to salvage the
ischemic leg of a patient with end-stage peripheral vascular disease 23 led the Angiogenesis
Foundation to analyze the effects of concurrent medications being administered to this
patient27 . While angiographically-proven new vascular channeIs appeared in the patient's
leg after VEGF-treatment, it was noted that these new vessels regressed after 12 weeks.
Indeed, several new vascular lesions described as "spider angiomas" regressed at 8 weeks
after VEGF-treatment.
Careful examination of the patient's hospital course revealed that significant
peripheral edema occurred in the treated leg, an etTect consistent with the known effect of
VEGF to increase vascular permeability. The edema was promptly and aggressively
treated with a standard intravenous diuretic, bumetanide (Bumex). Within the Foundation's
database, it was discovered that bumetanide is an inhibitor of endothelial cell proliferation
in vitro, as are other common diuretic medications , such as furosemide and spironolactone.
Therefore, the administration ofbumetanide to treat leg edema may have led to inadvertent
inhibition of the successful, therapeutically-induced angiogenesis. As a result, the leg
became increasingly ischemic, gangrenous and ultimately required amputation. Clearly,
patients who are receiving angiogenic therapy should not take medications that have
antiangiogenic activity .
Another example of the antiangiogenic effect of common medication was recently
reported for the angiotensin-converting enzyme inhibitor, Captopril, an antihypertensive
agent. Captopril is a known inhibitor of angiogenesis. Bruno Vogt and Felix Frey
deliberately administered Captopril, at an oral dose of 50 mg/day, to an AIDS patient with
Kaposi's sarcoma, and successfully regressed these angiogenic lesions'l". The take-home
lesson is that the desired etTect of angiogenesis therapy may be undone by common
medications that possess an opposite angiogenic effect. The Angiogenesis Foundation now
keeps a running database of a1l common medications that atTect endothelial cell
proliferation.

3.7. Lesson No. 7: Drug Delivery

Each therapeutic target in c1inical trials of angiogenesis therapies, whether a tumor,
the retina , the skin, or joints possess unique tissue barriers that can limit penetration of the
drug to the desired site . Example of such tissues barriers include the blood-brain barrier ,
the blood-retina barrier, and the keratinized epithelium . Because biopharmaceutical
companies often consider the mode of drug delivery only at late stages of drug
development, it is important to emphasize the need for early identification of etTective

483
methods to deliver the drug to its target site24. The drug Batimastat, for example, was
originally developed as an anticancer agent that required intraperitoneal or intrapleural
injection. Aithough the compound proved unsuccessful for treating ovarian and lung
tumors in this mode, it was later reformulated as a topical solution for treating the eye
disease, recurrent pterygium. In topical ophthalmic form, Batimastat (renamed ISV-120)
appears to be effective for treating this condition.
Another issue linked to drug delivery systems is the growing concern with the cost of
conducting clinicai trials. Experimental medications that must be injected intravenously
usually require patients to be admitted to the hospital for an overnight stay. In the United
States, even such abrief stay can incur acharge of more than $1,000 per day per patient.
Alternatively, drugs that can be administered orally, such as thalidomide, CAI, Marimastat,
and Neovastat, can be taken at horne with physician follow-up in the outpatient setting,
saving many thousands of dollars. Such practical aspects of drug delivery impact on the
development costs and ultimate affordability of angiogenesis-based therapies .

3.8. Lesson No. 8: Clinical Trial Design

As previously alluded to, clinicai trial design is a complex issue in the testing of
angiogenesis-based therapies . In addition to concerns regarding drug delivery, drug
interactions, treatment response, toxicity, and disease selection, it is important to
standardize clinical protocols for clinicians that are conducting the trial.
For example, in clinical trials of topical preparation of basic FGF to promote
angiogenic healing of diabetic leg ulcers, both endocrinologists and vascular surgeons have
been recruited to examine and treat patients. When vascular surgeons treat a diabetic ulcer,
they routinely debride the wound, a technique that surgeons have long-believed to "prime"
the wound for healing. Endocrinologists, on the other hand, do not routinely debride
wounds . This disparity in managing bFGF-treated diabetic ulcers has been used to explain
the inconsistent results obtained by different medicai centers conducting bFGF wound-
healing trials.

3.9. Lesson No. 9: Vascular Stabilization

Although stimulation or inhibition of angiogenesis are the current goals of clinicians
and drug developers, vascular stabilization is a new area of angiogenesis research that may
be important in treating conditions charaeterized by pathologic neovascularization . For
example, ophthalmologists who treat diabetic retinopathy and macular degeneration, and
neurosurgeons who treat Moya moya disease, regard the neovascularization occurring in
these conditions as pathologic and undesirable. The angiogenic vessels in such patients
tend to be aneurysmal and fragile, leading to hemorrhage in the eye or in the brain. We
hypothesize that in diabetes, macular degeneration, and in moya moya, the angiogenesis
may actually be an appropriate physiological response to a pathologic stimulus, such as
hypoxia . In such a situation antiangiogenesis may not be the ideal approach . Rather,
vascular stabilization therapy may serve to prevent bleeding complications, while enabling
tissue perfusion.
The work of John Mulliken and George Yancoupoulos and their colleagues is
shedding light on mechanisms that underlie this therapeutic approach. Studies of the
molecular genetic defects in patients with heredity venous vascular malformation have
revealed that a mutation in the Tie-2 gene on chromosome 9, encoding a receptor tyrosine
kinase in endothelial cells, leads to a deficiency in smooth muscle cells that surround and
stabilize new blood vessels. Tie-2 gene deletions result in fragile, dilated aneurysmal
vessels''. A new group of molecules, named angiopoietins, are the ligands for the Tie-2
receptor". Angiopoietin-l is produced by cells surrounding endothelium undergoing
angiogenesis, and binds to Tie-2. This initiates intracellular signaling and production of
chemokines and differentiating factors that attract mesenchymal cells to migrate towards

484
growing new vessels , and then to differentiate into either smooth muscle cells or
pericytes-l'. This sequence of events results in a mature, stabilized blood vessel.
Disruption ofthe gene for angiopoietin-I by homologous recombination techniques in
embryonie mice resulted in leaky, aneurysmal vessels with deficient smooth muscle and
peri-endothelial cells and with a poorly organized peri-vascular extracellular matrix. Thus,
the angiopoietin/Tie-Z system appears to be a newly identified mechanism involved in
regulating angiogenesis at a late stage of blood vessel growth. Tie-2/angiopoietin gene
therapy may therefore be useful to stabilize poorly formed vessels seen in ocular diseases,
such as diabetic retinopathy, and in neurological diseases, such as moya-moya.
Angiopoietin-2 is a recently discovered endogenous antagonist of angiopoietin-l ,
acting at the site of Tie-2 receptor tyrosine kinase 31. The role of angiopoietin-Z may be to
serve as a negative regulator of the vascular stabilizing actions of angiopoietin-I . Therefore,
future angiogenesis-based therapies aimed at vascular stabilization via angiopoietin -I may
be reversed by pharmacologic application of angiopoietin-2. Altematively, angiopoietin-2
may itselfbe therapeutically useful as an angiogenesis inhibitor.

3.10. Lesson #10: Nonscientific Forces

Several non-scientific entitites , including the business and regulatory commuruues.
playa defining role in angiogenesis drug development. All of the angiogenesis-based drugs
in U.S . c1inical trials discussed in this chapter are in the process of running the gauntlet
imposed by the U.S. Food and Drug Administration. The FDA requires
biopharmaceutical companies to furnish extensive preclinical data for evaluation regarding a
new drug candidate prior to initiation of c1inical trials in a Investigational New Drug (!ND)
application. Once Phase 1 c1inical trials have been initiated, the company and its c1inical
investigators must submit further, extensive data to proceed to the subsequent c1inical trial
phases 2 and 3. When all c1inical trials have been completed , a process that can take 5
years or more, a New Drug Application (NDA) must be filed and evaluated by the FDA
prior to final approval ofthe drug for large-scale manufacturing and marketing .
While recent trends suggest that regulatory agencies such as the FDA are moving
towards greater efficiency and speed in drug evaluation and approval , the overall statistics
within the pharmaceutical industry are daunting. For every 10,000 potential drug
candidates discovered in the pharmaceutical research laboratory , only 5,000 remain viable
candidates for further development after pre-clinical testing using in vivo and in vitro assay
systems . Ofthose 5,000 remaining drug candidates, for which INDs are submitted, only 5
obtain FDA approval to be studied in human c1inical trials . Ofthe five drug candidates that
enter human c1inical trials, only one survives the evaluation process to become an approved
drug available to the general public. The average cost ofthis entire process is enormous, in
excess of $220 million per drug. Furthermore, the average time requied to bring a molecule
from the laboratory bench to the marketplace is more than 11 years .
The development of angiogenesis-based drugs is c1early laden with tremendous
financial risk that is borne by the pharmaceutical industry . The overall risk is heightened in
light of the amount of new information that is being generated in the field of angiogenesis
research on an almost daily basis, and the difficulty experienced by companies in obtaining
and analyzing this new information in the context oftheir drug development program .

4. EARLY POSITIVE RESUL TS OF CLINICAL TRIALS

Most phase 1 and 2 c1inical trial results are presented in meeting abstracts, or not
reported at all. Investigators prefer to publish the results of large Phase 3 c1inical trial
results when the data and analyses are conclusive and likely to survive the peer-review
process . There is merit in describing several optimistic but still anecdotal findings from the

485
current c1inical trials of angiogenesis-based therapies for three main reasons: (i) it is likely
to generate enthusiasm within biopharmaceutical companies to enter drug development in
this field; (ii) basic researchers benefit from such "field reports" on the therapeutic effects
of compounds they are currently studying at the laboratory bench; (iii) such reports
generate increased scrutiny in the design and analyses of clinical trials at the early stages,
leading to earlier detection and avoidance of potential problems prior to substantial
investment on the part of companies and clinical investigators. The following section
provides brief descriptions of clinical results in current human studies.

4.1. Therapeutic Angiogenesis

4.1.1 Coronary Artery Disease
Patients with severe coronary disease and regions of myocardium impossible to
successfully bypass by surgery have received VEGF, or basic or acidic FGF, delivered to
their heart by cardiac catheterization or by surgical epicaridal implantation . After
treatment, there is evidence by magnetic resonance imaging for increased and persistent
coronary collateralization, presumably due to angiogenesis. In addition, there has been
improvement in myocardial perfusion with decreased ischemia, decreased angina and
increased ability by patients to sustain physical activity .

4.1.2 Accelerating Angiogenesis to Speed Diabetic Ulcer Wound Healing

Patients with chronic diabetic foot ulcers have received topical bercaplermin
(rhPDGF) and experienced a doubling in the rate of healing, 48% of treated wounds healed
compared to 25% ofuntreated wounds. Complete closure of wounds was observed after
19 weeks of treatment. Successful healing of these wounds results in limb salvage and
avoids the need for limb amputation.

4.2. Inhibition of Angiogenesis

4.2.1.Antiangiogenic Therapy for Cancer
4.2.2.1. Cervical Carcinoma
The angiogenesis inhibitor TNP-470 is being tested against a variety of human
tumors . A 50 year old woman with advanced squamous cell carcinoma of the cervix and
bilateral pulmonary metastases was entered into a dose-escalation study of TNP-470.
There was a complete response after 18 weeks oftherapy with radiographie disappearance
ofher lung metastases for more than 2.5 years.

4.2.2.2. Glioblastoma multiforme

Glioblastoma is one of the most lethai forms of brain tumor. A 54 year old woman
with glioblastoma multiforme experienced increased survival and prolonged quality of life
after receiving horne administered intravenous TNP-470 therapy . No significant toxicities
were described by the patient.
A small number of patients with recurrent high grade gliomas who received oral
thalidomide had radiographie evidence of decreased tumor mass by CT scan. At least three
patient have remained alive for more than 7 months after receiving thalidomide .

4.2.2.3. Prostate Cancer

Oncologists from the National Cancer Institute have reported that of 12 patients with
androgen-independent prostate cancer who received oral thalidomide in a Phase 2 study, 4
patients (33%) experienced a decline (by 20-37%) in their serum PSA levels, with the
longest dec1inelasting for 84 days.

486
4.2.3.Inhibition of Ocular Neovascularization
4.2.3.1. Diabetic Retinopathy
At the University ofFiorida, Tallahassee, four patients suffering from severe diabetic
retinopathy have been reported to have experienced stabilization of their eye disease, as
documented by fluorescein angiography, after receiving interferon alfa-2a .

4.2.3.2. Infantile Hemangiomas

Pediatric patients with massive hemangiomas of infancy that are life- or vision-
threatening have been receiving daily intravenous interferon alfa-2a, iftheir lesions prove to
be unresponsive to steroid therapy . The treatment is Iong-term, requiring weeks before
results are observed, and months to regress the vascular lesions . Urinary levels of basic
FGF have been utilized as a surrogate marker to follow the antiangiogenic effect. Once the
lesions have regressed, treatment can be discontinued and the therapeutic effect is long-
lasting .

4.2.3.3. Recurrent Pterygium

Pterygium is a disfiguring, fibrovascular outgrowth on the conjunct ival and corneal
surface that in the U.S . affects approximately 3 million patients, most who have had
extensive sun exposure. Vision impairment can occur when the pterygium grows over the
visual axis ofthe cornea. Aggressive recurrence ofthe pterygia after cryosurgical treatment
is a common problem . Patients have been successfully treated with the topical drug, ISV-
120, a form of the metalloproteinase inhibitor Batimastat, with prevention of pterygium
regrowth. No toxicities were noted in Phase 1 studies, and currently a randomized,
placebo-controlled double-masked Phase 2 clinical trial is underway .

5. TEN RECOMMENDAnONS FROM THE FOUNDAnON

As angiogenesis-based therapies become a reality, the information generated by early

clinical trial results is providing the basis for improved drug development in this field.
Preliminary evidence suggests that it is possible to regress tumors, stabilize diabetic
retinopathy , treat hemangiomas, and prevent recurrent pterygia by using antiangiogenic
drugs . Alternatively, by using angiogenic therapy, it is possible to grow neovessels in
ischemic human hearts and legs, and to accelerate skin ulcer healing in diabetic patients.
It is still too early to tell which of the angiogenic stimulators and inhibitors will
ultimately receive regulatory approval for widespread patient use . A number of factors
should be considered by all developers and advocates of angiogenesis modulators as a
therapeutic approach .

The Angiogenesis Foundation presents ten recommendations to minimize the risks

inherent in such efforts:

I. Selection ofthe proper disease likely to respond to a specific antiangiogenic agent is

complex and should involve consultation with clinical angiogenesis specialists.
(Note: interferon alfa-2a is an effective agent for treating hemangiomas of infancy,
and possibly for diabetic retinopathy, but has been found ineffective for the treatment of
age-related macular degeneration .)

487
2. A profile of known angiogenic factors occuning in each target disease should be
compiled, in order to identify conditions most likely to respond to rationally-designed
growth factor antagonist drugs as opposed to those conditions reguiring a multiple drug
regimen.
(Note: given the multiplicity of angiogertic factors expressed in human breast cancer,
it is unlikely that a drug antagonizing only a single angiogenic factor would be successful.)

3. In Phase 1 clinical trials, investigators should determine the "Optimal Biological

Dose" rather than the "Maximal Tolerated Dose" for angiogenesis-based therapies.
(Note : inhibitory effects of antiangiogenic agents are seen at relatively low doses in
animal studies, in contrast to conventional cytotoxic chemotherapeutic drugs that require
high dosages. Investigators should therefore seek the lower dose threshold for efficacy,
rather then seek the "maximally tolerated dose" (MTD) , as is the goal for conventional
Phase 1 testing . MTD, at the threshold oftoxicity is often considered the optimal starting
dose for Phase 2 trials. We propose for the antiangiogenic approach, that investigators
should find the "Optimal Biologie Dose" (OBD), and use the OBD as the starting dose in
Phase 2. The OBD is likely to be much lower than the MTD, and therefore less toxic) .

4. Clinical trial design must account for the delayed response of antiangiogenictherapy.
(Note : for the treatment of tumors, investigators should anticipate a delayed clinical
response to antiangiogenic compounds, compared to conventional chemotherapies.
Therefore, treatment should be continued even in the face of disease progression .)

5. Interdisciplinary teams of clinicians should be assembled to participate in c1inical

trial studies.
(Note : for clinical trials in cancer patients, an ophthalmologist, dermatologist,
rheumatologist, and a cardiologist should be recruited to the study team in order to assess
the effect of an antiangiogenic agent on the eye, the skin, the joint and the heart in
subgroups of patients. Similar teams should be assembled for angiogenic drugs being tested
in cardiac and vascular surgery patients . Data acquired should be relayed to the
sponsoring pharmaceutical company as weil as regulatory agencies.)

6. Conventional medications should be screened for their potential effects on

angiogenesis.
(Note : medications commonly prescribed and taken by diabetic, arthritic, cardiac and
cancer patients should be assessed in standard in vivo and in vitro assays for angiogenesis.
For example, it may be advisable for patients requiring antiangiogenic therapy to avoid
medications that stimulate angiogenesis.)

7. Drug delivery issues should be addressed at an early stage in the development of

angiogenesis-based therapies.
(Note: once a target tissue has been selected, the most effective mode of drug
delivery should be identified, including newer technologies such as liposomal, adenoviral,
and electroporation-assisted methods. Oral formulations should be developed for systemic
therapy whenever possible .)

8. Clinical trial protocols must define and standardize clinical technigues employed by
physician investigators in treating and examining patients.
(Note : this issue is particularly applicable for wound healing studies employing
angiogenesis-based drugs, because diverse c1inical practitioners, such as surgeons,
endocrinologists, emergency room physicians, and nurses have differing technical

488
approaches to wound care. Such disparities can potentially confound the results of a
clinical trial.)

9. Efforts should be made to develop vascular stabilization strategies.

(Note: some diseases, such as diabetic retinopathy, in which neovascularization is
thought to be driven by hypoxie conditions, may actually require new blood vessels in
order to sustain tissue viability. In such instances, therapies to stabilize new vessels may
facilitate tissue perfusion while simultaneously diminishing the risk of hemorrhage.
Furthermore, therapeutic strategies that induce angiogenesis - in the heart for example -
may require stabilization of desirable, newly generated blood vessels.)

10. Regulatory agencies should become partners in treatment discovery to increase

efficiency in angiogenesis-based drug development.
(Note: as demonstrated in the development of AIDS treatments, the participation of
the U.S . FDA with industry at an early stage of drug development decreased the cost and
length oftime required for a drug to be tested, approved and brought to market in the U.S .
Similar collaborative efforts in the development of angiogenesis-based therapies would
create a new environment of incentives for companies to join the race to develop new
drugs).

8. CONCLUSION AND FUTURE OUTLOOK

The preliminary findings from various c1inical trials of angiogenesis-based therapies

fuel the exciting promise that angiogenesis research will bring new hope for treating
incurable diseases. These early trials are rich with information to help guide drug
development for angiogenic diseases. Basic scientists, clinical investigators, corporate
leaders and industry regulators may benefit from the ten recommendations presented here
by the Angiogenesis Foundation. Angiogenesis-based therapies of the future will emerge
from a more shrewd approach to clinical trial design, a greater knowledge about the precise
role of angiogenesis in disease, and an adaptation of present paradigms in drug development
to the unique properties of new blood vessel growth and its contro!.

9. ACKNOWLEDGMENTS

We would like to thank the staff of the Angiogenesis Foundation for their assistance
with information acquisition and research, Dr. Michael Maragoudakis for inviting us to
present this work at the 1997 NATO Advanced Study Institute meeting on angiogenesis,
and Dr. Mark Abelson and Michelle George of Ophthalmie Research Associates for helpful
discussions on clinical development and regulatory issues.

10. REFERENCES

1. Adamis A., Shima D., Tolentino M ., Gragoudas E., Ferrara N ., Folkman J., D'Amore
P., Miller 1. Inhibition of VEGF prevents ocular neovascularization in a non-human
primate. Archives of Ophthalmology 196;114:66-7 1.
2. Asahara, T., Murohara, T., Sullivan, A., Isner 1. Isolation of putative progenitor
endothelial cells for angiogenesis. Science 1997; 275:964-67
3. Ausprunk, D ., Folkman, J. Migration and proliferation of endothelial cells in
preformed and newly formed blood vesseIs during tumor angiogenesis. Microvasc
Res 1977 ;14 :53-65

489
4. Breier, G., Damert, A., Plate, KH., Risau, W. Angiogenesis in embryos and ischemic
diseases . Thrombosis Haemostasis 1997;78:679-83
5. Burr L.P.H. Intussusceptive microvascular growth: a new mechanism of capillary
network formation in angiogenesis. In: R Steiner, PB Weiss, R Langer.
Angiogenesis: key principles ofscience, technology and medicine. Basel: Birkhauser
Verlag, 1992; 393-400
6. Casey, R. and Li W. Factors controlling ocular angiogenesis. American Journal of
Ophthalmology 1997; 124(4): 521-529.
7. Davis, S., Aldrich, T.H., Jones, P.F., Acheson, A., Compton, D.L., Jain, V., Ryan,
T.E., Bruno, 1., Radziejewski, c., Maisonpierre, P.C., Yancopoulos, G.D. Isolation
of angiopoietin-I , a ligand for the TIE2 receptor, by secretion-trap expression
cloning. CellI996;87:1161-69.
8. Engler, C.B., Sander, B., Koefoed, P., Larsen, M., Vinding, T., and Lund-Anderson ,
H. Interferon alpha-2a treatment of patients with subfoveal neovascular macular
degeneration. Acta Ophthalmologica 1993;71:2731
9. Ezekowitz R.BA, Mulliken 1.B., Folkman 1. Interferon alfa-2a therapy for life-
threatening hemangiomas of infancy. The New England Journal of Medicine 1992;
326:1456-63
10. Folkman J. Tumor angiogenesis: therapeutic implications. The New England Journal
ofMedicine 1971; 285:1182-6.
11. Folkman, 1. Angiogenesis in cancer, vascular, rheumatoid and other disease. Nature
Medicine 1995;1:27-31
12. Folkman 1. New perspectives in clinical oncology from angiogenesis research.
European Journal ofCancer 1996; 32A:2534-39
13. Folkman 1. Antiangiogenic therapy, in: Cancer: Principles & Practice of Oncology,
Fifth Edition, ed. VT DeVita Jr., S Hellman, SA Rosenberg. Lippincott-Raven
Publishers, Philadelphia, 1997: 3075-85.
14. Folkman 1., Szabo S., Stovroff M., McNeil P., Li w., Shing Y Duodenal ulcer:
discovery of a new mechanism and development of angiogenic therap y that
accelerates healing. Annals ofSurgery 1991 ; 214:414-27.
15. Fung, W.E. Interferon alpha 2a treatment for age-related macular degeneration.
American Journal ofOphthalmology 1991;112:349-350.
16. Gill P.S., Lunardi-Iskandar, Y , Louie, S., Tulpule, A , Zheng, T., Espina , B.M.,
Besnier, 1.M., Hermans, P., Levine, AM., Bryant, 1.L., and Gallo, R.C. The effects
of preparation of human chorionic gonadotropin on AIDS-related Kaposi's sarcoma.
The New England Journal ofMedicine 1996;335: 1261-69.
17. Guyer, D.R., Tiedeman, 1., Yannuzzi, LA, Slakter, J.1., Parker, D., Kelley, 1., Tang,
R.A, Marmor, M., Abrams, G., Miller, 1.W., Gragoudas, E.R. Interferon-associated
retinopathy . Archives of Ophthalmology 1993;111 :350-56.
18. Harada, K , Grossman, W., Friedman, M., Edelman, ER, Prasad, P.V., Keighley,
C.S., Manning, W.1., Sellke, F.W., Simons, M. Basic fibroblast growth factor
improves myocardial function in chronically ischemic porcine hearts. Journal of
Clinical Investigation 1994;94:623-30.
19. Hertig, AT. Angiogenesis in the early human chorion and in the primary placenta of
the macaque monkey. Contrib Embryol 1935;25:39-81.
20. Hirschi, KK, D'Amore, P.A. Pericytes in the microvasculature. Cardiovascular
Research 1996;32:687-98.
21. Holmgren, L., O'Reilly, M.S., Folkman ,1. Dormancy of micrometastases: balanced
proliferation and apoptosis in the presence of angiogenesis suppression. Nature
Medicine 1995; 1:149-53.
22. Holmgren, L. Antiangiogenesis restricted tumor dormancy. Cancer and Metastasis
Reviews 1996;15:241-245.

490
23. Isner , J.M ., Pieczek, A, Schainfeld, R., Blair, R., Haley, L., Asahara, T., Rosenfield,
K., Razvi, S., Walsh , K., Symes, J.F. Clinical evidence of angiogenesis after arterial
gene transfer of phVEGF(165) in patient with ischaemic limb. The Lancet 1996;
348(9024):370-374.
24 . Jain, R.K. Barriers to drug delivery in solid tumors. Scientific American
1994;271:58-65.
25. Kudelka, AP ., Levy , T , Verschraegen, C.F ., Edwards, CL, Piamsomboon, A ,
Tennrungruanglert ,W., Freedman, R.S., Kaplan , AL., Kieback, D.G ., Me yers, C.A .,
Jaeckle, K .A., Loyer, E., Steger, M., Mante , R., Mavligit, G., Killian, A., Tang , R.A .,
Guttennan, Ll.l., and Kavanagh, 1.1. A phase I study of TNP-470 administered to
patients with advanced squamous cell cancer of the cervix, C1inical Cancer Research
17; 3:1501-05 .
26. Li, Y.W., Folkerth, R.D ., Watanabe , H., Yu, c., Rupnick, M ., Barcus, P., Scott ,
R.M ., Black, P.McL., SalIan, S.R., Folkman, J. Basic fibroblast growth factor in the
cerebrospinal fluid of children with brain tumours - correlation with microvessel
count in the tumour. Lancet 1994;344 :82-86.
27. Li, W .W., Li, VW., Casey, R. Arterial gene therapy . Lancet 1996; 348 : 1381.
28 . Lopez, J.1. and Simons , M. Local extravascular growth factor delivery in myocardial
ischemia. Drug delivery 1996; 3:143-47.
29 . Loughnan, M ., EIder, J., Kemp , A. Treatment of a massive orbital-capillary
hemangioma with interferon alfa-2b : short-tenn results. Archives of Ophthalmology
1992;110:1366-67.
30 . Lupia, E., Montrucchio, G., Battaglia, E., Modena, V., Camussi , G. Role of tumor
necrosis factor alpha and platelet activating factor in neoangiogenesis induced b y
synovial fluid of patients with rheumatoid arthritis. European J Immunol 1996;
26:1690-94.
31. Maisonpierre, P.c., Suri, C., Jones , P.F., Bartunkova, S., Wiegand , S.1.,
Radziejewski , C., Compton, D., McClain, J., Aldrich , TH., Papadopoulos, N ., Dal y,
TJ. , Davis, S., Sato, TN., Yancopoulos, G.D. Angiopoietin--2, a natu ral antagoni st
for Tie2 that disrupts in vivo angiogenesis. Science 1997;277 :55-60 .-
32 . Miller ,J.W . Vascu1ar endothelial growth factor and ocular neovascularization.
American Journal ofPathology 1997;151:13-23 .
33 . Nyugen, M ., Watanabe, H., Budson, AE., Richie, J.P ., Folkman, J. Elevated levels
of the angiogenic peptide basic fibroblast growth factor in urine of bladder cancer
patients. Journal of the National Cancer Institute 1993; 85 :24 1.
34 . Oliver, SJ., Banquerigo, M .L., Brahn, E. Suppression of collagen-induced arthriti s
using an angiogenesis inhibitor, AGM-1470, and a microtubule stabilized, taxo!. Cell
Immunol1994; 157:291-99.
35. O'Reilly, M .S., Boehm , T , Shing, Y, Fukai , N., Vasios, G., Lane, WS, Flynn, E.,
Birkhead, J.R., Olsen, B.R., Folkman ,J. Endostatin : an endogenous inhibitor of
angiogenesis and tumor growth . Cell 1997; 88:277-85 .
36 . Pluda, J.M . Tumor-associated angiogenesis: mechanisms, c1inical implications and
therapeutic strategies. Seminars in Oncology 1997;24 :203-18 .
37 . Pluda, J.M ., Feigal, E., Yarchoan, R. Noncytotoxic approaches to the treatment of
HIV-associated Kaposi's sarcoma. Oncology 1993; 7:25-33 .
38. Pu, L-Q, Snidennan, AD. , Brassard, R., Lachapelle, K.1. Graham, A.M ., Lisbona, R.,
and Symes, J.F. Enhanced revascularization of the ischemic limb by angiogenic
therapy . Circulation 1993 ;88 :208-15 ..
39 . ReIf, M ., Lejeune, S., Scott, P.A.E ., Fox, S., Smith, K., Leek, R., Moghaddam , A.,
Whitehouse, R., BicknelI, R., and Harris, AL. Expression of the angiogenic factors
vascular endothelial cell growth factor , acidic and basic fibroblast growth facto r,
tumor growth factor b- I, platelet-derived endothel ial cell growth factor, placenta

491
 growth factor , and pleiotrophin in human primary breast cancer and its relation to
angiogenesis . Cancer Research 1997; 57:963-69.
40. Risau, W. Mechanisms ofangiogenesis. Nature 1997: 386: 671-74 .
41. Robson , M .C. Exogenous growth factor application effect on human wound healing.
Progress in Dennatology 1996; 30:1-7] [DL Steed and the Diabetic Ulcer Study
Group . Clinical evaluation of recombinant human platelet-derived growth factor for
the treatment of lower extremity diabetic ulcers. Journal of Vascular Surgery 1995;
21:71-81.
42. Schweigerer, L., and Fotsis, T. Angiogenesis and angiogenesis inhibitors in pediatric
diseases . European Journal ofPediatrics 1992;151:472-76.
43. Skowsky, W.R., Siddiqui, T., Hodgetts, D., Lambrou, F.H., Stewart, MY , and
Foster, M .T. A pilot study of chronic recombinant interferon-alfa 2a for diabetic
proliferative retinopathy: metabolie effects and ophthalmologie effects. Journal of
Diabetes and Its Complications 196;10:94-99.
44. Suri, c., Jones, P.F., Patan, S., Bartunkova, S., Maisonpierre, P.C., Davis, S., Sato,
T.N ., Yancopoulos, G.D. Requisite role of angiopoietin-l , a ligand for the TIE2
receptor, during mebryonic angiogenesis. CeH 1996;87:1171-80.
45. Tolentino, M.J., Mi 11er, J.W., Gragoudas, E.S., Chatzistefanou, K., Ferrara, N.,
Adamis, A.P . Vascular endothelial growth factor is sufficient to produce iris
neovascularization and neovascular glaucoma in a non-human primate. Archives of
Ophthalmology 1996;114;964-70 .
46 . Vogt , B. and Frey, F.J . Inhibition of angiogenesis in Kaposi's sarcoma by captopril.
TheLancet 1997; 349 :1149.
47. White, C.W., Sondheimer, H.M ., Crouch, E.C., Wilson, H., Fan ,L.L. Treatment of
pulmonary hemangiomatosis with recombinant interferon alfa-2a. The New England
Journal ofMedicine 189;320:1197-200.
48 . Yanagisawa-Miwa , A., Uchida, Y., Nakamura, F., Tomaru , T., Kido , H., Kamijo, T .,
Sugimoto, T., Kaji, K., Utsuyama, M., Kurashima , c., Ho, H. Salvage of infarcted
myocardium by angiogenic action of basic fibroblast growth factor. Science
1992;257 :1401-03.

492
Abstracts of Posters
SITE-DIRECTED MUTAGENESIS DISTINGUISHES DIFFERENT BFGF
ACTIVITIES ON ENDOTHELIAL CELLS IN VITRO

M. Bastaki, A. Gualandris and M. Presta

General Pathology & Immunology

University of Brescia, Italy

Basic FGF (bFGF) is involved in angiogenesis by stimulating endothelial cell

proliferation, migration and production of matrix degrading enzymes, such as urokinase-
type plasminogen activator (uPA)(1) . Deletion or substitution of basic amino acid residues
with neutral amino acids in different regions of bFGF has produced mutants (fig. 1) that
retain full mitogenic and receptor binding activity but have lost their uPA-inducing
capacity(2,3,4,5).
We tested various bFGF mutants for the capacity to cause tube formation in bovine
aortic endothelial (BAE) cell cultures on collagen gel and compared this activity to their
capacity to activate ERK-2 signalling and uPA upregulation .
The results of this study demonstrate that there is an "apparent" correlation between
the ability ofthe bFGF mutants to induce uPA upregulation and their capacity to promote
the formation of tube-like structures by BAE cells grown in three-dimensional collagen gel.
Nevertheless, uPA upregulation is not an absolute requirement for tube formation .
Activation ofERK-2 in FGFR-dependent signal transduction pathway is not sufficient to
induce uPA activity or tube formation by BAE cells in vitro.
In conclusion, the morphogenic activity of bFGF does not depend on its uPA-
inducing capacity but both activities depend on common structural properties of the
growth factor molecule, distinct from those required for ERK-2 activation and mitogenicity.

REFERENCES

I. Purification from a human hepatoma cellline of a basic fibroblast growth factor-like

molecule that stimulates capillary endothelial cell plasminogen activator production,
DNA synthesis and migration, MPresta et.al Mol. Cell. Biol. 1986, 6 :4060
2. A six-amino acid deletion in basic fibroblast growth factor dissociates its mitogenic
activity from its plasminogen activator-inducing activity, Alsacchi et.al Proc . Nat\.
Acad. Sci. USA 1991,88 :2628-26-32

495
3. Structure-function relationship of basic fibroblast growth factor : site-directed
mutagenesis of a putative heparin-binding and receptor-binding region, M'Presta et.al
Biochem. Biophys. Res. Com. 1992, 185(3) :1098-1107
4. Subcellular localisation and biologica1 activity of Mr 18,000 basic fibroblast growth
factor: site-directed mutagenesis of a putative nuclear translocation sequence
MiPresta et.al Growth Factors 1993, 9: 269-278
5. Diminished heparin binding of a basic fibroblast growth factor mutant is associated
with reduced receptor binding, mitogenesis, plasminogen activator induction and in
vitro angiogenesis Li'. Li et.al Biochemistry 1994,33 :10999-11007

496
EXPERIMENTAL ACTIVATION OF ENDOTHELIUM OF MOUSE CAROTIDE
ARTERY BY TNF" IN VIVO

M. Bastaki, E. Smith, Y. Gao and CC. Haudenschild

Exp. Pathology, JH. Holland lab. American Red Cross,

Rockville, MD, USA

Atherosclerosis is a complex pathological process which shares many aspects with

inflammation'l-". Besides SMC, the important cell components of the atherosclerotic
lesion are blood-bome T-Iymphocytes and macrophages, the main source of several
cytokines that affect SMC proliferation and ECM production". An early event in
atherosclerosis is the recruitment of circulating leukocytes into the vascular wall, which
occurs through the interaction of leukocytes with the endothelial surface". Under normal
conditions, the vascular endothelium is remarkably resistant to adhesion of circulating cells,
but for unclear reasons it can become activated and attract pro-inflammatory cells from the
circulation . Such reactive endothelium can be characterised as "dysfunctional" . As such, it
shares much with the microvascular endothelium when it is stimulated to undergo
angiogenesis. TNFa, a cytokine produced by macrophages is shown to affect important
functions of endothelial cells, such as regulation of adhesion molecule expression and
anticoagulant and fibrinolytic capacity and induction of apoptosisv". However, these data
were obtained in vitro. Considering the complexity of atherogenesis and the variety of cell
types involved, investigation of this early mechanism in vivo is desirable. We employed a
mouse in vivo experimental system with a local delivery of TNFa to a carotid artery and
restoration ofthe blood flow, an advandage to any in vitro system . We studied the effects
of TNFa on the endothelial lining of the vessel in terms of general morphology, induction
of apoptosis, expression of adhesion molecules and neointima formation . In addition, the
histological evaluation was critically improved by the en-face preparation of the vessels .
This model allows for studies of early events of post-receptor activation, such as protein
phosphorylation of signal transduction pathways, which occur within minutes of
exposure". The endothelium seems to be resistant to extensive injury from the treatment
(no apoptotic features observed) . Preliminary data indicate modest upregulation of at least
one adhesion molecule, ICAM-l, by TNFa. The irregular cellularity of the endothelial and
smooth muscle cells observed two weeks after surgery indicates a mild response to injury,
according to previous observations ofirregular pattern of cell density. The subtle changes

497
that render endothelium "dysfunctional " and hence possibly a target for monocytes are
still elusive. Using this system we believe we can address some of these questions, by
appropriately reproducing the conditions thought as associated with pro-atherogenic state .
The ability to use the mouse for such in vivo studies enjoys the great advantage of the
availability of several transgenic models, some of which are directly associated with
atherogenesis.

REFERENCES

1. Pathogenesis of atherosclerosis: state of the art. CC.Haudenschild. Cardiovascular

Drugs and Therapy 1990,4:993-1004
2. The pathogenesis of atherosclerosis: a perspective for the 1990s. RRoss. Nature
1993,362:801-809
3. Cellular mechanisms of atherogenesis. PE. DiCorleto. American Journal of
Hypertension 1993, 6(11 Pt 2): 314S-318S
4. Tumor necrosis factor-activates smooth muscle cell migration in culture and is
expressed in the balloon-injured rat aorta. S.Jovinge, A.Hultgardh-Nilsson,
J.Regnstrom and J.Nilsson. Arteiosclerosis, Thrombosis and Vascular Biology 1997,
17: 490-497
5. Endothelial cell inflammatory responses to tumor necrosis factor. Ceramide-
dependent and -independent mitogen-activated protein kinase cascades. V.Modur,
GA.Zimmerman, SMPrescotl and TMMclntyre . The Journal of Biological Chemistry
1996,271(2): 13094-13102

498
mE INFLUENCE OF THE FmRIN NElWORK STRUCTURE ON THE
FORMAnON OF CAPILLARY-LIKE TUBULAR STRUCTURES HUMAN
MICROVASCULAR ENDOTHELIAL CELLS

Annemie Collen, Pieter Koolwijk, and Victor W . M. van Hinsbergh

Gaubius Laboratory TNO-PG, Zernikedreef 9, 2333 CK Leiden,

The Netherlands

Fibrin is an important matrix involved in the healing of wounds . Changes in the

polymerization conditions before gelation ofthe clot have an important influence on the
structure of the fibrin formed and might as such influence the interactions with cells of
different types. In an in vitro angiogenesis model we tested whether changes in the fibrin
structure might have an influence on the formation of capillary-like tubular structures of
human microvascular endothelial cells. This in vitro model (1) is composed of a fibrin
matrix upon which endothelial cells are seeded and stimulated with a growth factor
(bFGF or VEGF) and a cytokine (TNFa) to invade the underl ying matrix and to form
tubular-like structures, while control cells remain as a monolayer upon the matrix.
Formation of these can be quantified by determining the length of the structures b y
image anal ysis and by measuring the accompanying 125I-fibrin degradation.
If the polymerization conditions are changed by altering the pH of the fibrinogen
solution different structures are formed . Polymerization at a pH of 7.0 results in a
fibrin pll7 .o- matrix with an high absorbancy indicating opaque, porous, rigid polyrners,
while fibrin pH7.8- matrices formed at a pH of 7.8 show a low absorbancy indicating
transparent, dense and malleable network. Matrices formed at a physiologicaJ pH
resemble more the opaque networks.
Endothelial cells stimulated to induce tubular-like structures formed quantitatively
and qualitatively different structures in the different matrices. On a fibrin fibrin pH7.0_
matrix an extensi ve network of long and connected structures are formed ; on a fibrin
fibrin pH7.4_matrix a network is also formed, although to a lesser extent, while on a fibrin
pH7.8- matrix much smaller and isolated structures are formed. The formation of capillary-
like tubular structures is u-PA-dependent since serine-protease inhibitors and u-PA
antibodies could completely inhibit the formation and the accompanying degradation .
To our surprise the stimulation of the endothelial cells with bFGF alone induced a
detachment and the accompanying degradation.

499
 These results show that different fibrin networks lead to a different ingrowth of
endothelial cells in a fibrin matrix and this ingrowth is u-PA-dependent. Secondly they
show that endothelial cells stimulated with bFGF detach from the matrix and induce
lysis of the underlying matrix. Environmental changes during polymerization of fibrin
may be crucial in determining the extent of capillary-like tubular structures as weil as
the rate offibrinolysis. This might playa determining role in the healing of wounds and
in the vascularization of fast growing tumours.

REFERENCES

P. Koolwijk et a!., 1996. Cooperative effect of TNFa, bFGF and VEGF on the
formation of tubular structures of human microvascular endothelial cells in a fibrin
matrix. Role ofurokinase activity. J. CeI!. Bio!. 1996,132: 1177-1188.

500
QUANTITATIVE EVALUATION OF THE VASCULAR ARCIllTECfURE OF TUE
CmCK CUORIOALLANTOIC MEMBRANE BY MEANS OF M1CROVASCULAR
CORROSION CASTING

Ch. Dimitropoulou'<", E. Fait l , S.-H. Gnoth', W. Malkusch- ,

M. E . Maragoudakis', M.A. Konerding'

'Institute of Anatomy , Johannes Gutenberg-University Mainz, D-55099

Mainz
2Kontron Elektronik GmbH, Oskar-van-Miller Str. 1, D-85385 Eching
3Department of Pharmacology, Medical School, University of Patras, Patras
G-2611O

INTRODUCTION AND AlM

Microvascular corrosion casting allows for detailed examination of the vascular

architecture ofnearly a11 organs and tissues (for review, see Lametschwandter et al., 1993;
Konerding, 1991; Konerding et al., 1995). This methodology can also be used for studies on
the CAM model for assessing physiological embry-onie angiogenesis (Burton et al., 1989).
The chorioallantoic membrane is a widely used experimental model for investigating
biological processes in vivo. In the present study we tried to establish microvascular
corrosion casting as a quantitative method for measuring changes in the CAM
vascularization during embryonie development in order to apply it in a next step for
experimentally induced angiogenesis.

MATERIALS & METUODS

The in vivo CAM angiogenesis assay as described by Folkman (1985) and modified
by Maragoudakis et a1. (1988) was used. Fertilised eggs of the domestic fowl (Gallus
domesticus, straight white Leghorn) were taken after incubation at the stages 33, 36, 38, 40,
43 and 53 H-H (corresponding to day 8, 10, 12, 14, 16 and 18). From each group ten
specimens were obtained . Vascular corrosion casts were made according to
Lametschwandter et a1 . (1990) . Half the specimens of each group were fixated with
glutaraldehyde to avoid excessive evasates.
Examination and photographic documentation of the microvascular corrosion casts
were carried out with a Stereoscan MK-250 scanning electron microscope (Cambridge,
England). For morphometric analysis from each specimen at least 10 stereo pairs were

501
 Table 1: Percentage of capillary densities in different days of incubation .

day 8 10 12 14 16 18
% or 84 8 92 865 8 88 8 7.8 899
cap. • • • • •
5 .7
•
3.7
dens. 5.2 7.7 5.3 7.3

taken using a tiIt angle difference of exactly 6°. I1ford FP4 films (I1ford, England) were used
for photographic demonstration. The images were stored in a PC-based image processing
system (KS 300, Kontron Elektronik, Eching, Germany), and measurements were
performed for the 3D calculation of lengths as recently described in detail (Malkusch et al.,
1995).
The percentage ofthe capillary density was calculated by the percentage of black and
white on the photos . In each specimen at least one cm! of the membrane was measured. By
using the centripetal ordering method described by Fenton and Zweifach (1981), the orders
of vessels were defined. Bach vessel (most distal precapillary or most proximal
postcapillary) that was in contact with the plexus was defined as order 1 vessel. Every
time two first order vessels converged, they give an order 2 vessel. When two second order
vessels converged, then they formed avessei of order 3. By the convergence of two vessels
unequal in order, the highest order was retained. The number of vessels per photo was
measured and the percentage of vessels per membrane was calculated. In each egg at least
60 randomly chosen vessel segments were measured using a slide gauge for the calculation
of the diameters.
The statistics and the graphic demonstration was done using SigmaStat and SigmaPlot.
All groups were verified on normality before statistic calculation.

120
CI first order
• second order
100 C Ihird order

80

60

40

20

o
8 10 12 14 16 18
Days

Fig. 1: Diameter of order 1, 2 and 3 vessels.

502
 70
13 firs t or der
• second order
60
o third order

50

40
ci
c
30

20

10

o
B 10 12 14 16 1B
Days

Fig. 2: Number of CAM-draining and supplying vessels in different days

RESULTS

Density of the plexus: The values of CAMs densities were between 84 ± 5.2 at day 8
and 89.9 ± 3.7 at day 18 (table 1). Statistics reveal significant differences between the
average densities of two continuous different days, but not when the densities of all days
were compared. Only the average densities of day 12 compared to that of day 14 was not

400
llJ first order

• second order
D third order
300

200

100

o
B 10 12 14 16 1B
Days

Fig. 3: Lengths offirst, second and third order vessels in different days.

503
significant different. Specimens with and without prefixation did not show significant
differences (below).
Diameter of the vessels : The comparison of the diameters of the different orders
showed significant differences in the eggs of the same day. Significantly different are also
the diameters ofthe same type ofvessels during incubation (fig. 1).
Number of vessels : The numbers of vessels per square unit was measured in all
specimens . Statistical analysis showed significant differences in eggs of the same day when
orders of vessels were compared (fig. 2). When comparing the different days, statistical
significant were only the numbers of first- and second order vessels between day 10 & day
12 with an augmentation of 140 and 175% respectively. The number ofvessels remained in
equallevels during the rest ofthe incubation.
Length of vessel: The mean lengths of different orders of vessels were always
significant different at eggs of the same day, when compared, except of order 2 to order 3
vessels of days 8 and 18 (fig. 3). No significant difference was found between the length of
the same vessel type (order 1,23) of different days.

CONCLUSIONS

Microvascular corrosion casting was used for the quantitative evaluation of the CAM
of eggs at different days of embryonie development. The measurements showed :
1. Capillary density does not change significantly when 2.5% cacodylate-buffered
glutaraldehyde is used as fixative, except at day 08.
2. Capillary density changes through the days, but not always significant.
3. Only the number of first and second order vessels increase significantly during
incubation . Comparing the eggs ofthe same day, then, the numbers of different orders
are always significant different.
4. The lengths of vessels of the same order are not significant different during
incubation . Different orders differ significantly to each other when eggs of the same
day are compared .
5. The mean diameter of chorioallantoic membrane vessels of an order does not alter
statistically significantly during development ofthe embryo.
The agreement ofmost ofthe results with those of others researchers (DeFouw et al.,
1989) allows the corrosion cast technique to be used as a quantitative method of calculating
changes in CAM during incubation.

ACKNOWLEDGEMENT

This study was supported by the European Community within the framework of the
Human Capital and Mobility Program "Mechanisms for the Regulation of Angiogenesis"
(ERB CHRXCT 940593) .

REFERENCES

Burton, G. 1., Palmer, M.E., 1989, The chorioallantoic capillary plexus of the chicken
embryo: a microvascular corrosion casting study. Scann. Microsc. 3: 549-558.

DeFouw, D.O., Rizzo, V. 1., Steinfeld,R., Feinberg, R. 1., 1989, Mapping of the
microcirculation in the chick chorioallantoic membrane during normal angiogenesis. Microv.
Res. 38: 136-147.

504
Fenton, B ., Zweifach, B. W ., 1981, Microcirculatory model relating geometrical variation to
changes in pressure and flow rate . Ann. Biomed. Eng . 9: 303-321 .

Folkman, 1., 1985, Tumor angiogenesis. Adv . Cancer Res . 43 : 172-203 .

Konerding, M . A., 1991, Scanning e1ectron microscopy of corrosion casts In medicine.

Scanning Microsc. 5 : 851-865 .

Konerding, M. A., Miodonski, A. 1., Lametschwandter, A., 1995, Microvascular corrosion

casting in the study oftumor vascularity: a review. Scanning Microsc. 9: 1233-1244.

Lametschwandtner, A., Lametschwandtner, U., Weiger, T., 1990, Scanning e1ectron

microscopy of vascular corrosion casts - technique and applications: updated review.
Scanning Microsc.4: 889-941 .

Malkusch, W., Konerding, M . A., Klapthor, B., Bruch, 1., 1995, A simple and accurate
method for 3-D measurements in microcorrosion casts illustrated with tumor
vascularization. A. Cell. Path . 9: 69-81 .

Maragoudakis, M. E., Sarmonika, M ., & Panoutsakopoulou, M ., 1988, Rate of basement

membrane biosynthesis as an index to angiogenesis. Tissue and Cell 20 : 531-539.

505
QUANTITATIVE EVALUATION OF ANGIOGENESIS INDUCED BY THROMBIN
IN THE CmCK CHORIOALLANTOIC MEMBRANE: SEM STUDIES ON
CORROSION CASTS

Ch. Dimitropoulou'<', E. Fait' , S.-H. Gnoth', W. Malkusch- ,

M. E. Maragoudakis', M.A. Konerding'

'Institute of Anatomy, Johannes Gutenberg-University Mainz, D-55099

Mainz
2Kontron Elektronik GmbH, Oskar-van-Miller Str. 1, D-85385 Eching
3Department of Pharmacology, Medical School, University of Patras , Patras
G-2611O

INTRODUCTION AND AlM OF THE STUDY

The chick chorioallantoic membrane (CAM) is a common model for studing biological
processes in vivo . Thrombin , in addition to its pivotal role in hemostasis, has a multitude
of effects that could influence the angiogenic cascade. It has already been shown that
thrombin promotes angiogenesis dose-dependent in CAM (Tsopanoglou et al., 1993). In the
present study we evaluated with microvascular corrosion casting the morphometrical
changes ofthe vascular architecture ofthe CAM after the application ofthrombin.

MATERIALS AND METHODS

On the CAMs of fertilised eggs of Gallus domesticus at stage 36 H-H, discs covered
with and without 1 unit (8.4 pmol/disc) thrombin and cortisone acetate (249 nmol/disc)
were placed (Tsopanoglou et al.,1993, Maragoudakis et al., 1988). The incubation went on,
and at stage 37 H-H vascular corrosion casts were made.
For morphometric analysis from each specimen at least 10 stereo pairs were taken
with a SEM using a tilt angle difference of exactly 6°. Images were stored in a PC-based
image processing system forthe 3D calculation of lengths (Malkusch et aI.,1995). At least
30 single randomly chosen photographs were taken ofthe plexus.
The percentage of the capiIIary density was caIculated by assessing the percentage of
black and white areas. For each investigated part of CAM, at least 80 mm! of membrane
were measured . The orders ofvessels were defined (Fenton and Zweifach , 1981) and the
number of vessels per area was measured and the percentage of vessels per cm- was
caIculated. In each egg at least 60 randomly chosen vessel segments were measured using a
slide gauge for the calculation of the diameters.

506
 n.s.d.

2 3
orders

Fig . 1: Length offirst, second and third order with and without application ofthrombin.

C conlrol
50
EI lest

40

30 , , ,, ,, ,,
,,,,,,, ,,,,,
,,,,, ,,,
,,,,,, , ,
ci
e ,,, , ,,, , n.s .d .
20 , , , ,, , , ,
,,,,
,,,,,, , ,, , ,,
,,,,, ,, ,
,,,,
,,, ,,,,,,, ,, n.s.d.
10 , , ,,,,, ,
, ,,,
,,,, ,,,,, , ,,
,,,, , , , ,
,,,,
, ,,,
0
2 3
order

Fig. 2: Number of CAM vessels of different orders with and without application of
thrombin.

507
 n.s .d.

Fig 3: Diameters of three orders of CAM vessels with and without application of
thrombin .

RESULTS

Length ofvessels: The mean lengths of the first order vessels are significantly shorter
in controls (89 ± 51.2~) than in tests (104.74 ± 74.8 um) . For the second order vessels
the control values are not significantly bigger than the test values. Also the mean lengths of
the third order vessels do not differ significantly between the controls and tests (fig. 1).
Vessel numbers : The number of vessels of the CAM pieces treated with thrombin is
significant higher for the first order vessels in comparison to those of the controls. The
mean values of treated CAMs are 43.5 ± 24.0 and 30.6 ± 14.6 for untreated CAMs . By
comparing the second and third order vessels between controls and tests, no significant
differences are seen (fig. 2).
Vessel diameters: The diameters oftests are always bigger than those of controls . For
the first order vessels the values are 19.4 ± 1.2 and 15.6 ± 1.0, for the second order 32.9 ±
1.4 and 28.9 ± 3.5, and for the third order 57.2 ± 11.0 and 50.8 ± 9.0 for tests and controls
respectively. Statistical analysis reveals significant differences for the first and second
orders ofvessels (fig. 3).
Density of plexus: In controls 78.1 ± 2.8 % of the measured area represent CAM
capillaries.In specimens treated with thrombin the average capillary density is increased by
40%.

CONCLUSIONS

Microvascular corrosion casting was used for the quantitative evaluation of the CAM
vessels after treatment with thrombin 1U at day 9 of embryonie development.
Measurements were made after two further days on day 11 and showed :
1. The capillary density increases after the application of thrombin significantly in the
majority of eggs.
2. The number of first order of vessels is significant increased after treatment with
thrombin, but not of second and third order vessels.

508
3. The first order vessels were longer in tests than in controls. The second and third
order vessels did not reveal any significant differences.
4. The mean diameters of the first and second order vessels were significantly bigger in
tests than in the controls, whereas the third order diameters were not.
Thus, this quantitative study showed unequivocally that the induction of angiogenesis
in CAM by vasoactive agents especially by thrombin modulates the vascular architecture
by enhancing the density of capillaries and the amount of pre- and postcapillary vessels.

ACKNOWLEDGEMENT

This study was supported by the European Community within the framework of the
Human Capital and Mobility Program "Mechanisms for the Regulation of Angiogenesis"
(ERB CHRXCT 940593).

REFERENCES
Tsopanoglou, N. E., Pipili-Synetos, E., Maragoudakis, M. E., 1993, Thrombin prornotes
angiogenesis by a mechanism independent of fibrin formation . Am . 1. Phys., 264 : 1302-
1307.

Maragoudakis, M. E., Sarmonika, M ., & Panoutsakopoulou, M ., 1988, Rate of basement

membrane biosynthesis as an index to angiogenesis. Tissue and Cell , 20 : 531-53 9.

Malkusch, W ., Konerding, M . A., Klapthor, B., Bruch, 1., 1995, A simple and accurate
method for 3-D measurements in microcorrosion casts illustrated with tumor
vascularization. A. Cell . Path ., 9: 69-81.

Fenton, B., Zweifach, B. W., 1981, Microcirculatory model relating geometrical variation to
changes in pressure and flow rate . Ann . Biomed. Eng ., 9: 303-321 .

509
ROLE OF VEGF ISOFORMS ON IN VITRO ANGIOGENESIS

s. Donnini, L. Morbidelli, A. Patenti, H. A. Weieh* and M. Ziehe

Dept. of Pharmacology, University of Florence , Viale Morgagni 65,
Florence, Ital y and
*GBF, Dept. of Gene Regulation and Differentiation, Braun schweig,
Germany.

The major subtype s of vascular endothelial growth factor (VEGF) in both normal
and tumor tissues are the 121-, 165-, and 189-amino acid-type mole cules. These
isoformsare biologicall y different based on their heparin binding affinity but no clear
information is so available on differences in the biological function s amon g the VEGF
subtypes in vivo and in vi tr o.
The aim of this study was to characterize the endotheiial cell funtion s revelant for
angiogenesis by diffemt isoforrns of VEGF. They were produced and purified using a
consistent preparation method in two related experimental models: adhesion assay
and chemotaxis assay. Homodimers for VEGF121 and VEGF165 together with the
heterodimer VEGF121/165 were tested and their effects were compared to basic
fibroblast growth factor (bFGF). Post-capillar y endotheiial cells from bovine
coronary venules (CVEC) were used in this study.
In vitto endotheiial cells mobilization and adhesion were promoted by VEG F121
homodimers and the heterodimer, but not by VEGFI65 . Heterodimer VEGF1 21/165
produced intermediate activities between the two homodimers. The effect of VEGF
peptides but not the one elicited by bFGF was completely blocked by the
preincubation of cells with anti-KDR antibody .
These results suggest that homogenous preparations of VEGF121 and VEGF165
can clearly display distinct biological effects probably mediated by the interaction with
the extracellular matrix, stability of diffusion rate.

510
CHARACTERIZAnON OF THE RECEPTOR TYROSINE KINASE, FIt-4 BY
GENE TARGETING

Daniel J. Dumont* +, Lotta Jussila! , Arja Kaipainen !, Li-Fong Seet+, Tuija

Mustonen* and Kari Alitalo!

*Dept. of Med . Biophys. , Univ . of Toronto, Ontario, Canada

! Molecular/Cancer Biol. Lab., The Haartman Institute, Univ . ofHelsinki, PL
2100014 Helsinki , Finland
+Amgen and Ontario Cancer Institutes, Toronto, Canada

The receptor tyrosine kinase, FIt-4, is expressed almost exclusively in celIs of the
endotheliallineage. Exp ression is first detected at Day 8.0 of development in angioblasts
and veins. In the older embryo and adult this expression is retained in the endothelial
lining ofthe Iympatics and high endothelial venules. Expression ofFIt-4 provides the first
molecular marker ofthe developing Iymphatics .

In order to determine the role ofFIt-4 within development we have engineered mice
lacking a functional FIt-4 gene. Mice heterozygous for this mutation have no obvious
phenotype, however mice homozygous for the mutation do not survive beyond Da y
11.0 of gestation. In order to follow the fate of FIt-4 null endothelial cells we designed
the targeting construct such that the E.coli LacZ gene was inserted into the Flt-4 locus
and its transcription was under control of the Flt-4 promoter. Preliminary results suggest
that although mice are dying around Day 11.0 of development there is no loss of the
Iympatics vessels. Currently a more detailed analysis ofthis mutant is in progress.

REFERENCES

Kaipainen, A., Korhonen, J., Mustonen , T., van Hinsbergh, V.W.M., Fang, G-H , Dumont
0 ., Breitman, M ., Alitalo K. Expression of the FLT4 Receptor tyros ine kinase becomes
restricted to Iympatic endothelium during development. Proc. Natl. Acad. Sei., USA 92:
3566-3570 ,1995 .

511
MACROPHAGE VEGF ANGIOGENIC POTENTIAL IS REVERSIBLY INHIBITED
BY ADP-RIBOSYLATION

lohn 1. Feng, Thomas K. Hunt, Heinz Scheuenstuhl, Gabriel Ledger,

Perveen Ghani, and Zamir Hussain

Departments of Surgery and Restorative Dentistry

University of California
San Francisco, Califomia, USA 94143-0522.

High lactate levels, a putative symbol of metabolic need, exist in wounds and
angiogenesis occurs in response to a "perceived need" for wound perfusion. Lactate
stimulates macrophage-derived angiogenic activity in culture, by stimulating expression of
vascular endothelial growth factor (VEGF). Lactate decreases macrophage NAD+ and
ADP-ribosylation, and increases VEGF mRNA and protein release. VEGF has been shown
to accept ADP-ribose. We now show (1) ADP-R reversibly inhibits VEGF angiogenic
potential , and (2) this potential is augmented by decreasing macrophage cytoplasmic ADP-
ribosylation .

METHODS

Bone marrow-derived macrophages were serum starved 24 hr for expts :

Expt 1: 15mM lactate was added for 24hr. The cell-conditioned media (CCM) and cell-free
control were centrifuged, and the supernatant dialyzed, lyophilized, and reconstituted in
buffer. ADP-ribosylation was performed with cholera toxin A subunit, NAD+, and DTT
in buffer at 37°C for 90 min. In controls, the above reaction was kept at 4°C. ADP-R was
cleaved by NH20H at 37°C for 6hr. Sampies were dialyzed, and concentrated to 20X
original volumes.
Expt 2: Serum-free cultures were divided into 2 groups : (1) I ug/ml, cyclohexamide was
added to inhibit protein synthesis; i(2)cyclohexamide and 15 mM lactate were added to
decrease ADP-ribosylation in the absence of protein synthesis . The CCM was centrifuged,
dialyzed, lyophilized, and reconstituted in H20 to 15X original volumes. Angiogenic
activity was assayed by the in vivo mouse Matrigel assay and expressed as mean score ±
standard error. Scoring system : 0 = No vessels, 0.5 = Incomplete vessel formation , I =
Well-formed vessels throughtout Matrigel plug.

512
RESULTS

Expt 1: ADP-ribosylation decreases CCM angiogenic activity and NH20H treatment

returns activ ity (ANOVA, P < 0.05, A vs. C and R). Expt 2: Angiogenic activity increases
with lactate in the absence of protein synthesis (unpaired Student's t-test, p<0.05, X vs .
Y). Cell-free media controls did not significantly contribute to the observed findings .

Expt 1 Groups n Score Expt 2 Groups n Score

Control (C) 1 0.68 ± 0.10 Cyclohexamide (X) 8 0.19 ± 0.09
1
ADP-R (A) I 0.25 ± Cyclohex+Lactate (Y) 8 0.63 ± 0.16*
2 0.10*
ADP-R, Cleave (R) 1 0.58 ± 0.14
2

CONCLUSIONS

I) Macrophages release angiogenic factor (VEGF) whose activity can be inhibited by

ADP-R. 2)This effect is reversible with treatment with hydroxylamine, which is specific
for cleaving ADP-ribose from arginine residues. 3) Even with protein synthesis inhibited,
existing angiogenic factor activity is enhanced by lactate. These results suggest that post-
translational modification of angiogenic factor by ADP-ribosylation can regulate
angiogenesis.

513
THERAPEUTIC APPROACHES TO TUMOR ANGIOGENESIS

Raffaella Giavazzi and Giulia Taraboletti

Mario Negri Institute for Phannacological Research,

Via Gavazzeni, 11,20125 Bergamo, Italy.

The complexity of the angiogenic process in tumor progression has allowed the
identification of numerous potential inhibitors which act at various stages of the
angiogenesis cascade. Several physiological factors have anti-angiogenic properties and
anti-angiogenic activity has been found in natural products, some of which are already in
clinical development. One of the critical steps in the angiogenesis process is the
remodelling and penetration of extracellular matrix by new capillaries. We show that
antineoplastic drugs, acting on endothelial cell motility and invasiveness, mayaiso
manifest anti-angiogenic proprieties. Matrix proteinase inhibitors represent an interesting
class of angiostatic compounds because they prevent the breakdown of the matrix proteins
and thereby maintain the integrity of the endothelium. Specifically, we have used the
hydroxamate batimastat (BB-94), as a representative compounds of this dass and
investigated its anti-angiogenesis and antineoplastic activity in a variety of model systems.
We have shown that BB-94 inhibited tumor-induced angiogenesis in vivo and blocked
endothelial cell invasiveness in vitro; treatment with BB-94 prevented metastasis formation
and tumor growth in experimental models. To study the potential use of this treatment for
use in therapeutic setting, BB-94 has been combined with conventional cytotoxic drugs .
We found that BB-94 markedly potentiates the anti-tumor activity of cisplatin in a model
of human ovarian carcinoma xenograft in atymic mice. From the numerous preclinical
studies that have been conducted, anti-angiogenic therapy appears to be able to contain
tumor growth , to be active on drug resistant tumors. The use of this class of drugs in
combination with conventional therapies offers interesting possibilities for the future
treatment of neoplastic diseases .

REFERENCES

1. Taraboletti G, Garofalo A, Belotti D, Drudis T, Borsotti P, Scanziani E, Brown PD ,

and Giavazzi R. Inhibition of angiogenesis and murine hemangioma growth by

514
 Batimastat, a synthetic inhibitor of matrix metalloproteinases. J Natl Cancer Inst
87:293-298,1995 .
2. Brown PD and Giavazzi R. Matrix Metalloproteinase inhibition: A review of anti-
tumor activity. Annals ofOncology 6:967-974, 1995
3. Belotti D, Vergani V, Drudis T, Borsotti P, Pitelli MR, Viale G, Giavazzi R and
Taraboletti G. The microtubule affecting drug paclitaxel has antiangiogenic activity.
CIin Cancer Res 2:1843-1850, 1996

515
INCREASED OXYGEN TENSION POTENTIATES ANGIOGENESIS

Jeffrey Gibson, Adam Angeles, Thomas K. Hunt

Department of Surgery,
University of Califomia,
San Francisco, Califomia, USA 94143-0522

PURPOSE

Lactate and Iocal oxygen tension enhance angiogenesis. Prior data shows that
endothelial cell response to a standardized VEGF stimulus is depressed by hypoxia . We
studied endothelial cell responses to VEGF in varying oxygen tensions in the Matrigel
angiogenesis model. We hypothesized that tissue hyperoxia would enhance
neovascularization.

METHODS

The control group consisted of Matrigel alone [Iml] which was injected
subcutaneously into female Swiss-Webster mice. VEGF [IOOng] or anti-VEGF antibody
[lug] were separately added to the Matrigel prior to injection for groups two and three.
Each group was divided into four sub-groups and exposed to a varying oxygen
environments : (1) continuous room air; (2) 100% oxygen at 1 atrnosphere ; (3) 100%
oxygen at 2 atmospheres ; and (4) 100% oxygen at 2.5-3.0 atmospheres, with subcutanous
oxygen tension measurements via Clark electrode in a11 sub-groups . Each test group was
treated for 90 minutes twice daily. At 7 days, the gel plugs were harvested and sections
were graded , in a double blind fashion, for the extent of neovascular formation . Grading: 0
= no vessels; 0.5 = scattered vessels; and 1 = maximal vessels in a11 quadrants . Results are
presented as mean score ± SEM.

516
RESULTS

Angiogenesis is significantly increased in unsupplemented Matrigel in all hyperoxic

groups compared to room air. Anti-VEGF antibody inhibited the angiogenic effect of
increased oxygen tension . VEGF promoted angiogenesis, but diminished the degree of
oxygen enhancement.

Condition n Score
21% oxygen 1 ATA (room air) 21 0.02±0.02
Matrige1 alone 100% oxygen 1 ATA 19 0.29±0 .07*
100% oxygen 2 ATA 21 0.45±0 .08*
100% oxygen 2.5 ATA 10 0.45±0.09*
100% oxygen 3.0 ATA 9 0.28±0.09*
21% oxygen 1 ATA (room air) 9 0.39±0.07
Matrigel + VEGF 100% oxygen 1 ATA 10 0.60±0.12
100% oxvgen 2 ATA 10 0.65±0.08
100% oxvzen 2.5 ATA 10 0.70±0.08*
21% oxygen 1 ATA (room air) 9 0.11±0.07
Matrigel + 100% oxygen 1 ATA 10 0.25±0.08
anti-VEGF ab 100% oxygen 2 ATA 10 0.10±0.07*
100% oxygen 2.5 ATA 10 0.20±0.08*

CONCLUSION

Angiogenic response to a standardized stimulus is proportional to ambient P02 over a

wide range. Oxygen is less enhancing in VEGF supplemented matrigel. Anti-VEGF
antibody inhibits the response to increased oxygen tension even below that seen in
nonsupplemented matrigel plugs. Therefore, oxygen enhances VEGF production and/or
assists endothelial cell response to VEGF.

517
REVASCULARIZATION OF ISCHEMIC MYOCARDIUM WITH THE LAMININ
ANGIOGENIC PEPTIDE SIKVAV

Derrick. S. Grant, R.W. Rose, M.G. Magno, J.D. Mannion, P .M .

Consigny, R.C. Morrison .
Thomas Jefferson University, Philadelphia, USA

A major cause of heart failure is a myocardial infarct due to coronary artery disease, and
insufficient collateralization. Basic fibroblast growth factor (bFGF) or vascular endothelial
cell grawth factor (VEGF) can assist in the revascularization ofthe cardiac tissue ; however,
the action and stability of these substances is unclear. Recently we have investigated the
local administration of a potent and specific angiogenic synthetic peptide derived from the
laminin alpha 1 chain, SIKVAV (Ser-Ile-Lys-Val-Ala-Val), abitity to enhance
collateralization of the myocardium in an ischemic rat model. Since SIKVAVis a stable
synthetic compound we compared it to bFGF or VEGF to revascularize ischemic
myocardium. The peptide was administercd directly to the epicardium by a slow release
pellet, and the resulting blood flow was evaluated quantitatively by angiography or silicon
rubber casts . Radiographs of hearts treated with SIKVAV peptide demonstrated greater
revascularization in the ischemic zone of the left ventricle, whereas the contral heart
showed a large ischemic zone. Measurement of the patent vessel number in LM
histological sections of the untreated rat' s left ventricle demonstrated very few vessels (10-
20/l00Ilm) in the ischemic zones whereas vessel numbers were approximately 100/100llm
in normal zones ofthe ventricle . SIKVAV was able to double the number of vessels in the
ischemic zones to 40/1001lm. VEGF showed the greatest increase in revascularization of
the ischemic myocardium (80/100/lffi). The data show that the SIKVAV peptide can
induce significant revascularization ofischemic myocardium and may have potential c1inical
application in cardiac pathologies .

518
THE ROLE OF NITRIC OXIDE IN THE ANTIANGIOGENIC AND ANTITUMOUR
EFFECTS OF X-RAY IRRADIATION IN VIVO.

O. Hatjicondi', D. Kardamakis', P. Ravazoula', 1. Dimopoulos', and M.E .

Maragoudaki S3

Departments of Radiology', Histopathology'', Pharmacology',

University of Patras Med ical School, 265 00 Rio, Patras , Greece

The involvement ofthe nitric oxide (NO) pathway in the effects of X-ray irradiation
on tumour vasculature was investigated in a transplanted tumour vasculature was
investigated in a transplanted tumour model. Groups of Winstar rats were inoculated
with sarcoma Walker 256 cells, and were treated with a)radiation (7.5 Gy given in a single
fraction), b)with the nitric oxide synthase inhibitor W -nitro-L-arginine methyl ester
hydrochloride (L-NAME) (30 mg per kgr weight for 3 and 14 days) , and c)with both
radiation and L-NAME. All animals were killed three days post-irradiation and
histology sections from the tumors were obtained and assessed for number of vessels
per optic field and for positively stained NO synthase cells, using immuno-
cytochemistry with the monoclonal antibody NOS-lI (i-NOS) . For measuring tumour
volume, tumour's maximum and minimum axes were measured daily for 14 days in a
separate group of animals, before and after treatment with radiation and/or L-NAME.
To NO synthase activity was meaured in homogenates ofirradiated tumours . We have
shown that the effects of radiation on tumour volume and number of vessels are
partially reversed by treatment with L-NAME. Although NO synthase activity is
present in both normal and tumour cells , radiation increases the number of macrophages
and leucocytes, which subsequently results in an increased localization of NO synthase
in vivo and production of NO in vitro, At the end of the experiment, at day 14, the
animals were sacrified and the tumour was dissected out and weighed. The weight of
irradiated tumours was decreased by 93% compared to the controls and of tumours
irradiated and treated with L-NAME showed a 30% decrease. These results indicate
that in this experimental system the radiation-mediated effects on tumour vasculature
may in part at least mediated via NO-production . These findings may provide an
explanation for the effects of ionizing radiation on tumours and may offer an alternative
therapeutic approach in pathological conditions, where angiogenesis is involved.

519
ANGIOGENIN INTERNALIZATION AND PROCESSING IN CULTURED
VASCULAR SMOOTH MUSCLE CELLS

Elissavet HATZI§, Yann BASSAGLIA, Christiane DOREY*, Josette

BADET*

CNRS URA 1813, Lab. CRRET, Universite Paris XIT-Val de Mame, 94010
Creteil, France .
§Present address : Laboratory of Biological Chemistry, University of Ioannina,
Medical School, 45110, Ioannina, Greece.
*Present address : INSERM U427, Universite PARIS V, Faculte des Seiences
Pharmaceutiques et Biologiques de Paris, 4 avenue de l'Observatoire, 75270
Paris cedex.

Angiogenin is a blood vessel inducing polypeptide of 14 kDa (1). Its primary

structure displays 35% identity to that of pancreatic ribonuclease A. Angiogenin has a
endonucleolytic activity toward 28S and 18S rRNAs as weil as tRNAs (2). A functional
enzymatic active site and a cell-binding domain are both required for its angiogenic
property to be expressed . The presence of angiogenin in normal plasma suggests that it
might be involved in endothelium homeostasis. Angiogenin specific receptors have been
first demonstrated on endothelial cells (3). In smooth muscle cells, angiogenin induces
activation ofphospholipase C and cholesterol esterification (4). We have previously shown
that l25I-angiogenin binds specifically to these cells, at 4°C, on two classes of binding
sites . The presence of angiogenin cell-surface receptors, the ribonucleolytic activity of the
molecule and the intracellular location of ribonuclease inhibitor, a tight binding inhibitor of
both ribonucleolytic and angiogenic activities of angiogenin (5), have brought us to study
the intemalization and the intracellular fate ofthe molecule in aortic smooth muscle cells.
Time-course studies using radiolabelIed angiogenin showed a continuous
intemalization of the molecule that stabilized at 24h. The intemalization was energy-
dependent since this process was afTected by treating the cells with sodium azide, an
inhibitor of oxidative phosphorylation.
Intemalized angiogenin appeared remarkably resistant to proteolysis for an
angiogenic polypeptide which reminds the stability of ribonucleases. However a limited
processing was observed by SDS-polyacrylamide geJ e1ectrophoresis of the cells and
autoradiography. Degradation pathway was investigated by treating the cells with
lysosomotropic agents such as methylamine, ammonium chloride and chloroquine.

520
 Intemalized angiogenin was detected by immunofluorescence in permeabilized cells
using specific polyclonal antibodies purified by angiogenin-affinity chromatography, and
localized in the cytoplasm .
These results showing angiogenin binding, intemalization and metabolism in vascular
smooth muscle cells support the hypothesis that angiogenin could contribute to vessel wall
homeostasis.

REFERENCES

1. Fett J.W. et al., Biochemistry , 1985,24: 5480

2. Riordan J.F. & Vallee B.L., Br. Cancer, 1988, 57: 587
3. BadetJ. et al., Proc. Natl. Acad. Sci. USA, 1989,86: 8427
4. Moore F. & Riordan J.F., Biochemistry 1990,29: 228
5. Lee, F.S. & VaHeeB .L., Prog. Nucl. Ac. Res. Mol. Biol., 1993,44:1

ACKNOWLEDGMENTS

We are grateful to Dr. P. D'Amore (Boston, MA) and Dr. J.K. McDougall (Seattle,
WA) for the gift of aortic smooth muscle cells. We express our gratitude to Dr. D .
Barritault, in whose laboratory we developed the angiogenin project . This work was
supported by the Association de la Recherche sur le Cancer (grant n06831), la Fondation de
France, and la Fondation pour la Recherche Medioale.

521
DISTRIBUTION OF FIBROBLAST GROWTH FACTOR AND RECEPTORS
DURING ANGIOGENESIS IN ADULT RAT MESENTERY

F. Hansen-Smith, L. Morris, M. Scheer

Department ofBiological Sciences,
Oakland University, Rochester, MI 48309-4401 USA

Although basic fibroblast growth factor (bFGF) is widely recognized as an angiogenic

factor, its in vivo functions within the microcirculation under physiological conditions
remain poorly understood. Most evidence for an in vivo angiogenic function of bFGF
comes from injured tissue . If bFGF is angiogenic for microvessels under physiological
conditions, then its receptors (bFG-R) should be located within the microvascular bed, and
microvessel growth should occur in intact microvessels in the presence of bFGF . The
mesenteric microcirculation of adult rats has an intrinsic angiogenic activity (Hansen-Smith ,
et al, 1996). Therefore, we used this microvascular bed as an assay system to determine
whether bFGF has an angiogenic function under physiological conditions . Extirpated
mesenteric windows from adult female rats were analyzed as whole mounts, using the
Griffoma simplicifolia I lectin as a fluorescent marker for all intact capillaries (independent
ofperfusion status) (Hansen-Smith , et al, 1988). Antibodies to bFGF and bFGF -Rs were
used to detect the co-distribution of the antigen with microvessels, using indirect
immunofluorescence. Mapping ofthe microvessels by confocal microscopy failed to reveal
bFGF or FGF-R in capillaries, including capillary sprout tips, but trace amounts were
present on the few large arterioies present. In contrast, mast cells were intensely
immunofluorescent for both bFGF and bFGF-R, whether or not they were closely
associated with the sprouting capillaries. Isolated windows were also organ-cultured for 24
hours in the presence or absence of serum to determine whether FGF-R's could be induced
in capillaries or sprout tips, but the microvessels remained negative for FGF-R.
Angiogenesis was evaluated by image analysis for mesenteric windows organ-cultured in
serum-free conditions in the presence or absence of bFGF . The density of nodes (branch
points) per field (I mm 2), an index of angiogenesis, revealed no difference between intact
controls (26.8 +2.9) and organ-cultured windows under serum-free conditions (23.4 +2.8),
but the addition of 1 ug/ml bFGF resulted in significantly fewer nodes (19.4 + 1.5, p<0.05).
Intemode distances and lengths of terminal sprouts were longer in the bFGF-treated
windows, possibly a result of loss of capillary branches and/or growth in length of existing
capillaries. The absence of immunocytochemically detectible bFGF-Rs associated with

522
capillaries in vivo and in organ culture suggests that vascular bFGF does not play a direct
role in physiological angiogenesis. However, an indirect role of bFGF in microvascular
remodelling, mediated by mast cells, cannot be excluded.

REFERENCES

1. Hansen-Smith, F. M., G. Joswiak and 1. Baustert. Regional differences in

spontaneously occurring angiogenesis in the adult rat mesentery. Microvasc. Res. 47: 369-
76, 1994.
2. Hansen-Smith, F.M., L. Watson, D.Y. Lu, and I. Goldstein . Griffoma simplicifolia 1:
Fluorescent tracer for microcirculatory vessels in non-perfused thin muscles and sectioned
muscle . Microvasc. Res . 36: 199-215, 1988..

523
DOES NEUTRAUZATION OF BFGF ALTER MICROVASCULAR NElWORK
PATIERNING IN SPONTANEOUSLY ANGIOGENIC MESENTERIC WINDOWS?

F. Hansen-Smith, 1. Marshick, L. Lynn, L. Morris

Department ofBiological Sciences,

Oakland University, Rochester, MI 48309-4401, USA

Extensive study in vitro has established that exogenously added bFGF is angiogenic,
stimulating both cellular proliferation and cellular migration. Studies of exogenously added
bFGF in vivo are conflicting with regard to their role in angiogenesis, but Norrby (1994)
reported a direct angiogenic effect of bFGF in mesentery. Although endothelial cells
synthesize endogenous stores of bFGF in vitro , it is not yet clear whether endogenous
bFGF has an angiogenic role in intact microvessels under physiological conditions .
Spontaneous angiogenesis occurs under physiological conditions in mesenteric "windows"
of adult rats, and we postulated that immunological neutralization of endogenous bFGF in
the mesenteric micocirculation could cause rarefaction and/or a diminution of this
angiogenesis in vivo . Female Sprague Dawley rats were given a single i.p.injection of
neutralizing antibody to bFGF (n-bFGF, 1 ug in 1 ml) or a PBS control.
Immunocytochemically , bFGF was indetectible for at least three days thereafter .
Angiogenesis in mesenteric windows was evaluated by confocal microscopy, using
rhodamine Griffonia simplicijolia I to delineate the microvasculature in whole mounts .
Microvessel density, evaluated by a grid-intersect technique, increased significantly--by
50%-- within three days, with areturn to normal after seven days. A similar change was
found in total vessel length per field. Image analysis of capillary branching patterns
indicated a transient angiogenesis after 3 d. with a significant reduction of inter-node
distance from 138.5 + 3 in controls to 124.6 + 2 in n-bFGF-treated . The number of nodes
per field increased significantly, from 22.4 + 2.6 to 33.4 + 2.8 nodes per mrrr'. All
parameters returned toward control value by 7 days post-treatment. bFGF in mesenteric
windows is associated with mast cells, not vessels, so the observed alterations in
microvascular patterns may be indirect (Scheer, et al, 1995). Mast cell degranulation was
found 1-3 d. after injection of n-bFGF, but not saline, and mast cell integrity was re-
established by 7 days. We postulate that mast cell -associated bFGF may function as a
local angiostasis system . Since feed arterioies are bFGF-positive and bFGF is a
vasodilator, hemodynamic consequences of n-bFGF mayaiso have contributed to the

524
altered microvascular density. We conclude that endogenous bFGF may be angiostatic or
angiogenic, but the former may be an indirect effect and a function of the locus of bFGF
expression and FGF receptor localization.

REFERENCES

1. Norrby, K. Basic fibroblast growth factor and de novo mammalian angiogenesis.

Microvasc. Res . 48: 96-113, 1994.
2. Scheer, M ., A Haas, F. Hansen-Smith. Distribution of FGF-R during angiogenesis in
adult rat mesentery. Microcirc . 2: 99, 1995

525
DO FmROBLAST GROWTH FACTORS REGULAlE POSTNATAL
ANGIOGENESIS IN ras RETINA?

F . Hansen-Smith, L. Morris, M. Whiteher

Oakland University, Roehester, MI 48309-4401, USA

The retina is a rieh souree offibroblast growth faetors, yet while retinal endothelial eells in
vitro are responsive to fibroblast growth faetors, a speeifie funetion ofbasie FGF (FGF-2)
and aeidieFGF (FGF-l) during developmental angiogenesis in the retina has yet to be been
documented .
Mierovessels in the rodent retina develop postnatally and are readily studied in flat mounts
(Connolly, et al, 1988), providing an ideal microvascular network for analysis of the role of
regulatory factors in angiogenesis. We postulated that if the fibroblast growth factor(s)
have a regulatory role in retinal angiogenesis during development, their distribution and that
of their receptors should correlate with the spatial and temporal changes occurring during
the early growth and remodelling of the microvessels in the postnatal retina . Flat mounts
and cryosections from retinas of newborn-adult rats were fluorescently double-Iabeled with
Griffonia simplicifolia I (to delineate microvessels) and antibodies to bFGF, aFGF , and
FGF-receptors (FGF-R, bek and jlg) and examined by confocal microscopy . The
distribution of bFGF and FGF-receptors was similar. Microvessels in adult eontrols
lacked immunocytochemically detectible FGF-R, but choroidal vessels were intensely
positive. b-FGF and FGF-R were extensively distributed within the neonatal retina,
accompanying the vessels developing radially from the optic disc. Examination at high
magnification showed that in most vessels the antigen was associated with GFAP-positive
supporting cells instead of endothelium. Presumptive angioblasts were negative. The
pattern ofFGF-R-postitive cells extending spatially beyond the leading edges of capillary
sprouts suggested a possible function in providing migration cues for endothelial cells
during early stages . At 6-days FGF-Rs were present on unidentified round cells (possibly
presumptive perieytes/smooth muscle cells) near and adhering to arterioles, which were
FGF-R positive. After the first week of development there was Iittle correlation between
the distribution of FGF-R in the neuronal retinal cells (positive) and that of microvessels
(negative) in the expanding mierovascular network. In contrast to bFGF, a-FGF was
extensively distributed throughout the neuronal and vascular retina, showing no correlation
with changes in angiogenesis or FGF-R's. These results suggest that bFGF may be

526
involved in regulating the growth of retinal arterioles. However, any effect on capillary
growth must be indirect, ie, via the perivascular cells or b-FGF-induced vasodilation of the
arterioles. Due to its non-specific distribution, aFGF is unlikely to be a selective
regulatory factor. In contrast to the intra-retinal vessels, the choroidal vessels appear to be
regulated by bFGF, based on the immunocytochemical evidence.

REFERENCES

Connolly, S., T. Hores , L. Smith, and P.D' Amore Characterization of vacularaa

development in the mouse retina. Microvasc . Res. 36: 275-90, 1988

527
CAPILIARY GROWTH IN DEVEWPING SKELETAL MUSCLE IS NOT
DEPENDENT ON b-FGF

F . Hansen-Smith, L. Monis, A. Dunning

Oakland University , Rochester, MI 48309-4401 USA

Basic fibroblast growth factor (bFGF) is widely accepted as an angiogenic factor.

However, under what conditions (ie, physiological vs. pathological) and in what vessels
(ie, macro- vs. micro-vessels) bFGF actually regulates angiogenesis in intact vessels remains
to be determined . This point is critical to all potential future c1inical applications of bFGF
since any treatments may have systemic effects. Rapid, physiological angiogenesis occurs
during the postnatal growth period. Postnatal capillary growth in rat skeletal muscle
(sternomastoid, SM) in rats shows a biphasic growth pattern which also differs between
the red and white areas (Hansen-Sm ith, et al, 1992). We predicted that if bFGF is a
physiological regulatory factor during this time, increased expression of bFGF would be
seen during the most rapid growth phase, and selectively in the highly oxidative "red"
regions. The distribution ofbFGF and FGF-receptors (FGF-R , ie,bek,jlg) was determined
by indirect immunocytochemistry in cryostat sections of SM muscles from newborn-30d.
postnatal rats. Griffoma simplicifolia I lectin was used to identify microvessels by double-
fluorescence labelling. The presence of bFGF in the tissue was also probed by Western
blot. In selected sampies the sites of mRNA expression for bFGF were localized by in situ
hybridization histochemistry. Capillaries at all stages of postnatal growth were negative
for bFGF and FGF-R. Although bFGF could be faintly detected in the tissues by Western
blot, this is probably within the muscle fibers. bFGF and its mRNA were, however,
localized to some-but not all--arterioles, particularly at 3-weeks, when rapid arteriolar
growth occurs (Hansen-Smith). In contrast to mesentery , most mast cells were negative for
bFGF. These results argue against a significant role for bFGF in regulating postnatal
capillary angiogenesis. To test for possible functional responsiveness to bFGF at levels
lower than detectible by immunocytochemistry, sheets of neonatal abdominal muscle were
organ-cultured in the presence or absence ofbFGF or neutralizing antibody to bFGF for 24
hrs. Image analysis of capillary branching patterns failed to reveal new sprouts in the
presence of bFGF , and neutralizing antibody had no significant effect. The expression of
bFGF or FGF-R was not affected by these in vitro treatments. These results further
support our conclusion that bFGF is not a physiological regulator of capillary angiogenesis

S28
during postnatal growth of skeletal muscle. bFGF may, however, regulate growth of
terminal arterioles.

REFERENCES

1. Hansen-Smith, F.M., L. Morris, G. Joswiak. Postnatal proliferation of microvessels

and the distribution of bFGF in rat sternomastoid muscle. FASEB 1. 6: A 1600,
1992.
2. Hansen-Smith, F.M., K. Banker, L. Morris, G. Joswiak. Alternative histochemical
markers for skeletal muscle capillaries. Microvasc. Res. 44: 112-16,1992.

529
HUMAN MAST CELL TRYPTASE IS PRü-ANGIOGENIC

David A. Johnson*, Cynthia B. Peterson!, and M. Sharon Stack+

*Dept. ofBiochem. & Mol. Biol., lH. QuilIen Collen ofMedicine, East
Tennessee State University, Johnson City, TN 37614-0581, USA
! Dept. ofBiochem. & Cell. & Mol. Biol., Univ. ofTennessee, Knoxville,
TN37996, USA
+Dept. ofObstetrics and Gynecology, Northwestem Univ. Med . Sch., Chicago,
11.60611, USA

Tryptase is a 31kDa, glycosylated, trypsin-Iike enzyme that is stored in and

reIeased from all human mast cells. Tryptase exists as a tetramer and binds heparin,
which stabilizes activity, but this protease is not inhibited by blood plasma proteinase
inhibitors, such as anti-thrombin III and u\-proteinase inhibitor. Tryptase functions in
vivo have not been established, but in vitro limited proteolytic c1eavages of high
molecular weight kininogen and fibrinogen, and activation of pro-urokinase have been
reported . Research is focusing on the hypothesis that mast cell tryptase is pro-
angiogenic via activation ofpro-urokinase and via direct c1eavage of extracellular matrix
components. Fibrin and fibrinogen, components of tumor associated stroma, are
specifically c1eaved at Arg-Sn, destroying the Arg-Gly-Asp cell binding site recognized
by cellular intergrins. A second c1eavage at Lys-21 appears to be responsible for
preventing thrombin initiated cl otting. Tryptase competes with thrombin for fibrinogen
at equal molar ratios of each enzyme. Tryptase also makes very specific c1eavages in
human vitronectin (VN), rapidly converting the nkDa monomer to a 58kDa product.
Tryptase c1eavage of VN greatly reduces the binding of meIanoma and ovarian cancer
cells. A later c1eavage site has been identified as Lys-88 . These findings indicate that
mast cell tryptase may play a significant role in extracellular matrix modifications
critical to angiogenesis.

AKNOWLEGMENT

Supported by the American Heart Association, Tennessee Affiliate and the East
Tennessee State University Cardiovascular Res. Institute.

530
THE ROLE OF MACROPHAGE RESPONSE DURING THE EARLY PHASE OF
TUMOR GROWTH

Kisseleva E., Fichtner I. *, Becker M.*, and Lemm M .*

Institute for Experimental Medicine

Department of Immunology, St.Petersburg, Russia
*Max-Delbruck-Center for Molecular Medicine,
Berlin-Buch, Germany.

Macrophages may play a central but rather dualistic role in determ ining the host
response to tumor both by preventing and promoting tumor growth . Activated
macrophages destroy their targets by release of cytokines: TNF and IL-l. From the other
hand, the presence of macrophages can support tumor growth . Tumor associated
macrophages may affect the induction of neovascularization by production of TNF-a [1].
Angiogenesis and inflammation play a key role during the early phase of the tumor
growth as the prevalance of tumor supporting or rejecting mechanisms defines further
nodule and stroma formation . The aim of the present stud y was to investigate the
participation of macrophages and cytokines during the early growth of a syngeneic
transplantable highly immunogenic murine leukemia P388/adria [2]. Functional activity of
peritoneal macrophages and the serum level of TNF a were studied in different groups of
mice. Mice from group I (control ) received saline. Mice from group 2 (tumor bearers)
with fast s.c. 100% tumor growth were compared with animal s from group 3 that were
previously twice immunized with lethally irradiated P388/adria cells and later on inoculated
with viable tumor cells. Tumors in group 3 grew only in 25% of animals and were
characterized by a significant delay in growth . Activity of peritoneal macrophages was
studied by NOrproduction and NBT-test. Both tests revealed an early high systemic
activation of macrophages in group 2. This coincided with the elevation of serum TNF-a
level. The N0 2-production by peritoneal macrophages correIated weil with the dynamics of
serum cytokine level while NBT -test did not. Studies in group 3 showed total abrogation of
macrophage and cytokine reactions. The production of inhibitory factors by macrophages
in previously immuni zed mice is suggested . The fact that the early acti vation of
macrophages and increase of serum TNF-a level occured in animals with fast growing
tumors and was decreased or absent in animals with tumor delay or rejections allo ws to
suppose that this reaction plays more supporting than a protecting role for tumor growth .
The present data are the first attempt to emphasize the role of earl y nonspecific
inflammatory response during tumor growth in an experimental model.

53 1
REFERENCES

1. Leibovich S.1., Polverini P.1., Shepard H.M., Wiseman D.M., Shively V., Nuseir N .
Macrophage-induced angiogenesis is mediated by tumor necrosis factor-a. Nature,
1987, v. 329, p. 630-632.
2. Kisseleva E., Becker M., Lemm M., Fichtner I. Involvement of macrophages and
cytokines into rejection mechanisms ofthe drug-resistant and immunogenic murine
lymphomaP388/adria. Anticancer Res. 1996, v. 16, p. 1971-1978.

532
OVEREXPRESSION OF VEGFNPF IN SKIN OF TRANSGENIC MICE INDUCES
ANGIOGENESIS, VASCULAR HYPERPERMEABILITY AND ACCELERATED
TrnMORDEVELOPMENT

F. Larcher, R. Murillas and J.L. Jorcano

Department of Cell and Molecular Biology, CIEMAT 28040 Madrid, Spain

Upregulation of keratinocyte-derived VEGF expression has been recently

established in both, neoplastic (1) and non-neoplastic processes of skin such as
psoriasis, wound healing and bullous diseases (2). To further characterize the effects of
VEGF in skin in vivo we have developed transgenic mice expressing the mouse
VEGF I20 under the control of keratin K6 regulatory sequences (a 2,4 kb 5' fragment) that
confer transgene inducibility upon hyperproliferative stimuli. As expected from the
inducible nature of the transgene, two of the three founder mice obtained (V27 and
V208), did not show an apparent phenotype. However the other founder (V2)
developed scattered red spots throughout the skin. This animal was shown to be
highly mosaic for transgene integration and its transgenic offspring developed a striking
phenotype characterized by swelling and erythema resulting in early postnatal
lethality. Histological examination of skin of these transgenics demonstrated highly
increased vascularisation, edema and disruption of skin architecture . Expression of the
transgene was silent in adult animals of lines derived from founders V27 and V208.
However, phorbol ester-induced hyperplasia resulted in transgene expression and
increased cutaneous vascularization. Skin carcinogenesis performed on hemizigous
crosses of V208 mice with activated H-ras-carrying TG-AC transgenic mice resulted in
accelerated papilloma development, increased tumor burden and progression towards a
malignancy prone, keratoacanthoma type lesion.

REFERENCES

1. Larcher F., Robles A., Duran H., Murilias R., Quintanilla M ., Conti C.J., & Jorcano J.L.
Upregulation of VEGFNPF in mouse skin carcinogenesis correlates with malignant
progression state and activated H-ras levels . Cancer Res. 56: 5391-5396, 1996.
2. Detmar M. Molecular regulation of angiogenesis in the skin. 1. Invest. Dermatol . 106,
207-208, 1996.

533
RATIONAL DRUG DESIGN ON ANGIOGENIN

D.D. Leonidas', R. Shapiro/", L.I. Irons', N. Russo'', and K.R . Acharya'

'School ofBiology and Biochemistry, Univ . ofBath, Claverton Down, Bath, UK

2Center for Biochemical and Biophysical Seiences and Medicine, and
3Department ofPathology, Harvard Medical School, Boston, MA 02115, USA

Human Angiogenin (Ang), is an atypical member of the pancreatic ribonuclease

(RNase) superfamily distinguised by its potent in vivo angiogenic activity and weak
ribonucleolytic activ ity . Like RNase A, Ang catalyzes the endonucleolytic cleavage of
RNA on the 3' side of pyrimidines and from structural studies (Acharya et al., 1994,
1995) it has been found that the ribonucleolytic center of Ang is identical to the
corresponding site of RNase A. The ribonucleolytic activity of Ang is essential for
angiogenicity : mutations that substantially decrease enzymatic activity also diminish the
capacity of Ang to induce blood vessel formation (Curran et al., 1993; Shapiro et al.,
1989; Shapiro and Vallee, 1989) . This strict correlation suggests that compounds
directed against the active site of Ang may be effective therapeutic agents since Ang
plays a critical role in the establishment and growth of some human tumours (Olson et
al., 1994, 1995) . We have initiated efforts to find potent nucleotide based inhibitors of
the enzymatic activity of this protein.
Two inhibitor molecules have been designed by mapping the ribonucleolytic center
of Ang using a combination of kinetic analysis with molecular modelling (Russo et al.,
1996). These inhibitors, 5'-diphosphoadenosine 3'-phosphate (ppA-3 '-p), and 5'-
diphosphoadenosine 2' -phosphate (ppA-2 ' -p), are the most potent nucleotide ligands
identified thus far, not only for Rnase A (K, values 240 nM and 520 nM, respectively)
but for Ang as well (K, values 300 nM and 110 nM, respectively). An initial
crystallographic study (at 1.7 Aresolution) has been performed (Leonidas et al., 1997) to
elucidate the molecular interactions ofthese ligands first to Rnase A before we move on
to Ang. Parallel study for Ang in the presence ofthese nucleotides is in progress.

REFERENCES

Acharya K.R. et al., Proc. Natl. Acad . Sei. USA 91, 2915-2919,1994.

534
Acharya K.R . et al., Proc . Natl. Acad. Sei USA 92,2949-2953, 1995.
Curran TP. et al., Biochemistry 32, 2307-2313, 1993.
Leonidas D.D. et al., Biochemistry in press.
Olson K.A. et al., Proc . Natl. Acad. Sei. USA 92, 442-446,1995 .
Olson K.A. et al., Cancer Res. 54, 4576-4579, 1994
Russo et al., Proc . Natl , Acad. Sei , USA 93, 1726- 1732, 1989.
Shapiro R. et al., Biochemistry 28, 1726-1732, 1989.
Shapiro R. & Vallee B.L., Biochemistry 28,7401-7402,1989.

535
TOWARDS MICE MODELS FOR HEREDITARY HAEMORHAGIC
TELANGIECTASIA TYPES 1 AND 2.

Dimitrios Liakos', Helen Arthur", Andrew Smith'", lohn Burn'', Austin G.

Diamond'.

'Department ofImmunology, Univers ity ofNewcastle upon Tyne

iiDepartment ofHuman Genetics University ofNewcastle upon Tyne
iiiCentre for Genome Research, University ofEdinburgh

Hereditary Haemorhagic Telangiectasia(HHT) is inherited in an autosomal dominant

fashion and results in abnormalities of vascular structure. Individuals with HHT present
with a wide diversity ofvascular abnormalities ofthe nose, skin, lung, gastrointestinal tract
and brain (1) . In the characteristic small telangiectases, arterioies and post capillary
venules display dilations and become connected, resulting in complete lack of capillaries.
Larger arteriovenous malformations, also lacking capillaries and consisting of direct
connections between arteries and veins, are found particularly in the brain and lungs of
HHT 1 patients.
HHT is genetically heterogeneous and is associated with mutations in two different
genes. HHT 1 patients carry mutations in endoglin, which maps to 9q33-q34 and encodes a
homodimeric integral membrane protein that is expressed in high levels on human vascular
tissues (2) . Mutations on HHT 2 patients map to 12q13 and have been shown to occur in
the serine threonine kinase receptor Alk-I (3) . Both genes are related to TGF beta
receptor, types III and I respectively. Mutation analysis has suggested that in HHT 1, the
disease results from the production oftruncated endoglin molecules, while in HHT 2, Alk-
1 receptors lack a functional kinase domain .
In order to study the disease in more detail, we are generating two different Oknock-
outO mice lacking expression of endoglin and Alk-I respectively.
The endoglin knock out mouse, presently under construction, will contain a
premature stop codon in exon 8, mimicking closely the type and location ofmutations seen
in HHT 1 patients. This mutation is expected to produce a truncated endoglin protein
which, in the mouse, will also be tagged with an HA epitope to permit immunological
detection. In addition, the lacZ coding region with an internal ribosome entry site has been
positioned immediately downstream of the endoglin mutation. Thus, expression from the
endoglin promoter can also be monitored using beta galactosidase activity .

536
 Construction of the Alk-l model is at a less advanced stage, but we intend to
interrupt the coding sequence upstream of the kinase domain to produce a cell surface
associated, but catal ytically inactive protein .

REFERENCES

I. Guttmacher, A.E ., D .A. Marchuk, and R.J . White, Hereditary hemorrhagic

telangiectasia [see comments]. [Review]. New England Journal of Medicine, 1995 .
333(14): p. 918-24.
2. McAllister, K.A., et al., Six novel mutations in the endoglin gene in hereditary
hemorrhagic teIangiectasia type 1 suggest a dominant-negative effect of receptor
function . Human Molecular Genetics, 1995. 4( 10): p. 1983-5
3. Johnson, D .W ., et al., Mutations in the activin receptor-like kinase 1 gene in
hereditary haemorrhagic teIangiectasia type 2. Nature Genetics, 1996. 13(2) : p. 189-
95 .

537
THE SULFONIC ACID POLYMERS PAMPS [pOLY(2-ACRYLAMIDO-2-METHYL-
I-PROPANESULFONIC ACID)] AND RELATED ANALOGUES ARE IDGHLY
POTENT INHffiITORS OF ANGIOGENESIS

S. Liekens, 1. Neyts, M. Vandeputte, P. Maudgal and E. De Clercq.

Rega Institute for Medical Research,

Katholieke Universiteit Leuven,
Leuven, Belgium.

The sulfonic acid polymers PAMPS [poly(2-acrylamido-2-methyl-l-

propanesulfonic acid)], PSS [poly(4-styrenesulfonic acid)] and PAS [poly(anetholesulfonic
acid)] proved to be highly potent inhibitors of angiogenesis in the chick chorioallantoic
membrane (CAM) assay. PAMPS was found to achieve a dose-dependent inhibition of
microvessel formation in the CAM assay, increasing from 57% ±16 inhibition at 10 ug/disc
to 72% ±15 at 150 ug/disc, Also PSS and PAS caused a strong inhibition of angiogenesis
(55% ± 19 and 48% ± 16, respectively at 50 ug/disc), whereas PVS [poly(vinylsulfonic
acid)] was found to be inactive at this dose. The compounds proved to be non-toxic for the
developing chick embryo at these doses. PAMPS, PAS and PSS, but not PVS, inhibited
microvessel formation in the rat-aorta-ring assay. In addition, the increased eH-
methyl]dThd uptake in endothelial cells in vitro upon stimulation with bFGF was inhibited
by PAMPS, PAS and PSS at 20 ug/ml . Astrang correlation (r = 0.95) was found between
the anti-angiogenic effect of the sulfonic acid polymers in the CAM assay and their
inhibition of the bFGF-induced mitogenic response, indicating that bFGF is the target for
these sulfonic acid polymers . PAMPS was found to be the most potent inhibitor of
angiogenesis and was therefore further evaluated in vivo. PAMPS inhibited
neovascularization in the cornea, induced by Hydron pellets containing sucralfate-stabilized
bFGF. In addition, when administered at a dose of 25 mglkg, 5 times a week for three
weeks, intratumorally, PAMPS caused complete regression of an established, highly
vascular, hemangiosarcoma in athymic nude mice. These results suggest that PAMPS is a
specific and potent inhibitor of angiogenesis with potential antitumour activity .

538
VEGF-C RECEPTOR BINDING AND PATTERN OF EXPRESSION WITH VEGFR-
3 SUGGESTS A ROLE IN LYMPHATIC VASCULAR DEVELOPMENT

Athina Lymboussaki, Eola Kukk, Suvi Taira, Arja Kaipainen, Michael

Jeltsch, V1adimirJoukov and Kari Alitalo

Molecular/Cancer Biology Laboratory

Haartman Institute PL21 (Haartmaninkatu 3) 00014
University of Helsinki, Finland

The vascular endothelial growth factor family has recently been expanded by the
isolation of two new VEGF-related factors, VEGF-B and VEGF-C. The physiological
functions of these factors are largely unknown . Here we report the cloning and
characterization of mouse VEGF-C, which is produced as a disulfide-linked dimer of 415
amino acid residue polypeptides, sharing an 85% identity with the human VEGF-C amino
acid sequence . The recombinant mouse VEGF-C protein was secreted from transfected
cells as VEGFR-3 (Flt4) binding polypeptides of 30-32x 10-3 Mr and 22-23x 10-3 M r
which preferentially stimulated the autophosphorylation of VEGFR-3 in comparison with
VEGFR-2 (KDR). In in situ hybridization, mouse VEGF-C mRNA expression was
detected in mesenchymal cells of postimplantation mouse embryos, particularly in the
regions where the Iymphatic vessels undergo sprouting from embryonie veins, such as the
perimetanephric, axillary and jugular regions. In addition, the developing mesenterium,
which is rich in lymphatic vessels , showed strong VEGF-C expression . VEGF-C was also
highly expressed in adult mouse lung, heart and kidney, where the expression level of
VEGFR-3 was also prominent. The pattern of expression of VEGF-C in relation to its
receptor VEGFR-3 during the sprouting of the lymphatic endothelium in embryos suggests
a paracrine mode of action and that one of the functions of VEGF-C may be in the
regulation of angiogenesis of the lymphatic vasculature.

539
THE EFFECTS OF 1,25-DIHYDROXYVITAMIN DJ ON ANGIOGENESIS IN
VITRO

D.J. Mantel!', E.B. Mawer-', NJ. Bundred/, and A.E. Canfield !

'Department ofMedicine
2Department of Surgery
University of Manchester, 2205 Stopford Building, Manchester, MI3 9PT, UK

Angiogenesis, the formation of new blood vessels from an existing vascular bed, is
essential for tumour growth and dissemination. It is a complex, multistage process
characterised by aseries of events, including alterations in phenotype, migration and
proliferation ofvascular endothelial cells. Angiogenesis is regulated by a large number
of factors including vascular endothelial growth factor (VEGF). Drugs targeted against
angiogenesis are being developed as a novel therapeutic approach for the treatment of
cancer. The purpose of this study was to investigate the possible antiangiogenic
activity of I,25-dihydroxyvitamin DJ (I ,25(OH)2DJ) using as in vitro model of
angiogenesis in which cells can be induced to change from a resting phenotype
(resembling cells lining a mature vessel) to an angiogenic phenotype (resembling cells in a
newly forming vessel) by the addition of VEGF. Using this model, we have examined
the effect of I,25(OHhDJ on various aspects of endothelial cell behaviour in vitro which
are relevant to angiogenesis namely: cell migration, proliferation and morphogenesis.
We have demonstrated that I,25(OH)2D3 (10-9 M) significantly inhibits VEGF-
induced endothelial cell proliferation (p<0.05). I,25(OH)2D3 also inhibited VEGF-
induced endothelial cell sprouting in a dose dependent manner. 10-9 M I,25(OH)2D3
resulted in an inhibition of VEGF-induced sprout formation by approximately 30%, 10-7
M I,25(OHhD3 inhibited sprouting by 50%. The addition of I,25(OHhD3 to VEGF
induced, preformed sprouting endothelial cells resulted in an accelerated regression to
resting phenotype. This was again a dose dependent effect. Using a wounded
monolayer assay, I,25(OHhD3 was shown to have no effect on endothelial cell
migration. Our results therefore demonstrate I,25(OH)2D3 that can inhibit the
angiogenic activity of VEGF. Future studies will be directed at elucidating the
mechanism by which this inhibition occurs.

540
EFFECT OF AICA-RIBOSIDE TREATMENT ON AORTIC ALTERATION IN
NEONA TAL STZ INDUCED DIABETIC RATS

Öztürk Ml.Tuncdernr Ml .Bdkent SI,Kaya F1,Yilmaza- Si, Akkan A.G2 ,

Özyazgan S2,Senses V2

'Department ofMedical Biology and

2Department of Pharmacology
Cerrahpasa Faculty of Medicine,
University of Istanbul, TURKEY

Diabetes mellitus (DM) is one ofthe risk factors of arterial disease and its complica-
tions. We examined the effect of 5-amino-4-imidazole carboxyamide (AICA)-riboside
which has insulinotropic'!' and antioxidant'" effects and inhibitor effect on g1uconeogene-
SiS(3) and is suggested to be used in DM treatment, on atherogenetic processes in neonatal
STZ-induced (neo-STZ) diabetic rats.
12 Wistar albino newborn rats were used for Neo-STZ diabetes model. A single dose
of 100 mg/kg STZ was injected i.p. to 2 days old newborn rats. The rats were divided into
two groups, then plasma glucose levels and body weights were measured beginning from 3
weeks. The first group (n=6) was sacrificed at the end ofthe 15th wk.To the second group
(n=6) AICA-riboside (10 mg/kg/day) was injected i.p. for one month beginning from the
16th wk . At the end experimental period the thoracic aortae of both groups were removed
and cut into two parts . One part of aorta was put in kreps ringer to observe contraction-
relaxation responses. The remaining part was prepared for electron microscopic examina-
tion .
The marked and widespread ultrastructural alterations were particularly observed in
intima layer of aortae of neo-STZ diabetic rat. There were discontinuity and abnormal ori-
entation in the endothelial lining. Occosional endothelial cells with nuclear changes indica-
tive of apoptosis were identified . Marked enlargement in subendothelial space (SES) and
accumulation amorphous material in SES were seen. In the AICA treatment group it was
seen that the intimal morphological alteration were partly prevented when compared to the
untreated diabetic group .
AICA-treatment prevented the change in the reactivity of the diabetic aortae to both
noradrenalin (NA) (contraction) and acetylcholine (AC) (relaxation). There were no signifi-
cant differences in PD 2 values for NA between AICA treatment diabetic (6.55±O.99) and

541
controls (7.20±0.70) while PD2 values of the aortic strips from untreated diabetic rats
(4.17±O.04) were markedly lower (p<O.OOOl) than the other groups . AC produced a dose-
dependent relaxation of the aortic strips precontracted with NA from controls (6.3±2.2)
and AICA-treated diabetic groups (5.0±1.08) while there was no response to AC of aortic
strips precontracted NA from the untreated diabetic rats (1.0±0.0). The blood glucose lev-
els were found to be reduced and the body weights to be increased by AICA-treatment
The results suggest that AICA-riboside may attribute to the restoration of endothe-
lial cell damages and improve the endothelial dysfunction in neo-STZ diabetic rats.

REFERENCES

1- AG Akkan and WJ Malaisse : Insulinotropic action of AICA-riboside . Insulin release

by isolated islets and the perfused pancreas. Diabetes Research (1994),25 : 13-23.
2- DA Bullough, et al: Acadesine (AICA-riboside) prevents oxidant-induced damage in
the isolated quinea pig heart. J Pharmacol Exp. Ther (1993) 226 (2): 666-672.
3- MF Vincent, et al. Inhibition by AICA riboside of gluconeogenesis in isolated rat
hepatocytes Diabetes (1991), 40: 1259-1266.

542
COMPARISON OF VEGF, VEGF-B AND VEGF-C AND Ang-l mRNA
REGULATION BY SERUM, GROWTH FACTORS, ONCOPROTEINS AND
HYPOXlA

Karri Paavonenl, B. Enholml , A. Ristimäki2, V. Kumarl, Y. Gunjil, 1.

Klefstroml , L. Kivinen3 , M. Laihoß, B. Olofsson4, V. Joukovl, U. Eriksson4
and Kari Alitalo 1

I Molecular/Cancer Bio!. Lab., 2 Dept. of Clin . Chemistry and Dept. of

Obstetrics and Gynecology, 3 Dept. of Virology, The Haartman Institute,
Univ . ofHelsinki, PL 2100014 Helsinki , Finland
4Ludwing Institute for Cancer Res., Stockholm , Sweden

The vascular endothelial growth factor (VEGF) family has recently been expanded
by the isolation of two additional growth factors, VEGF-B and VEGF-C. Here we
compare the regulation of steady state levels of VEGF, VEGF-B and VEGF-C mRNAs in
cultured cells by a variety of stimuli implicated in angiogenesis and endothelial cell
physiology. Hypoxia, Ras oncoprotein and mutant p53 tumor suppressor, which are
potent inducers of VEGF mRNA did not increase VEFG-B or VEGF-C mRNA levels .
Serum and its component growth factors , platelet-derived growth factor (PDGF) and
epidermal growth factor (EGF) as weil as transforming growth factor-ß (TGF- ß) and the
tumor promoter phorbol myristate 12,13-acetate (PMA) stimulated VEGF-C, but not
VEGF-B mRNA expression. Interestingly, these growth factors and hypoxia
simultaneously downregulated the mRNA of another endothelial cell specific ligand,
angiopoietin-l. Serum induction of VEGF-C mRNA occurred independently of protein
synthesis; with an increase ofthe mRNA half-life form 3.5 h to 5.5-6 h, whereas VEGF-
B mRNA was very stable (Tl/2>8h) . Our results reveal that the three VEGF genes are
regulated in a strikingly different manner, suggesting that they serve distinct , although
perhaps overlapping functions in vivo.

543
ROLE OF PLATELETS, NITRIC OXIDE AND CELL ADHESION MOLECULES IN
THE FORMATION OF TUBES BY ENDOTHELIAL CELLS IN VITRO

E. Papadimitriou, M. E. Maragoudakis and E. PipiIi-Synetos

Dept. ofPhannacology
School of Medicine
University of Patras
26500 Patras, Greece

Angiogenesis is a complex process, regulated by a number of factors and events, the

exact sequence ofwhich remains unclear. PIateIet activation seems to be an important event
in the angiogenic process' and thrombin, an important regulator of platelet activation, is
capable of stimulating angiogenesis', Nitric oxide (NO) is an endogenous mediator with a
wide repertoire of biological roles. The role of NO in angiogenesis as weil as in the
progression of tumor growth and metastasis is at the moment controversial and rather
poorly understood. It has been shown to be both a mediator of angiogenesis, by substances
which increase vascular permeability' , and an inhibitor of angiogenesis and tumor growth",
The role ofNO as a potent inhibitor of pIatelet aggregation and adhesiorr' makes it a prime
candidate as an antiangiogenic moIecuIe. It has recently been shown that NO decreases the
expression of vascular cell adhesion molecule 1 (VCAM-I) and intercellular adhesion
molecule 1 (ICAM-I) in human vascular endothelial cells".
In the present study, we investigated the effeet of platelets on the Matrigel tube
formation assay of angiogenesis in vitro' and the role of NO and cell adhesion molecules in
this phenomenon. When platelets were added to cultures of human umbilical vein
endotheIiaI cells (HUVECs) on Matrigel, the tube fonnation was increased by 90% over the
control (HUVECs alone). Thrombin further increased (in a statistically significant way)
tube fonnation by 20%. Sodium nitroprusside (SNP), as weIl as 8-br-cGMP, had no effect
on the increase oftube fonnation caused by platelets, while it abolished the increase caused
by.thrombin. FinaIly, antibodies to the extracellular domain of VE-cadherin and PECAM-I
(platelet endothelial cell adhesion molecule I) decreased tube fonnation in the presence or
absence of platelets .
These results suggest that platelets play an important role in tube fonnation on
Matrigel. This effect is only moderately increased by pIateIet activation by thrombin . NO
appears to effect this latter process, without affecting tube fonnation caused by

544
unstimulated platelets. We are currently investigating the hypothesis that NO reverses the
stimulatory effect of thrombin on tube formation by inhibiting the expression of cell
adhesion molecules, such as VE-cadherin andlor PECAM-1.

REFERENCES

1. D. R . Knighton, T. K. Hunt, K. K. Thakral and W. H. Goodson, Role of platelets and

fibrin in the healing sequence: an in vivo study of angiogenesis and collagen synthesis ,
Ann. Surg. 196: 179 (1982) .
2. N. E. Tsopanoglou, E. Pipili-Synetos and M. E. Maragoudakis, Thrombin prornotes
angiogenesis by a mechanism independent offibrin formation, Am. J. Physiol. 264:
C1302 (1993).
3. M. Ziehe, L. Morbidelli, E. Masini, S. Amerini, H. 1. Granger, C. A. Maggi , P.
Geppetti and F. Ledola, Nitric oxide mediates angiogenesis in vivo and endothelial cell
growth and migration in vitro promoted by substance P, J. Clin. Invest. 94: 2036
(1994) .
4. E. Pipili-Synetos, A. Papageorgiou, E. Sakkoula, G. Sotiropoulou, T. Fotsis, G.
Karakioulakis and M. E. Maragoudakis, Inhibition of angiogenesis , tumor growth and
metastasis by the NO-releasing vasodilators, isosorbide mononitrate and dinitrate, Er.
J. Pharmaeol. 116: 1829 (1995) .
5. S. Moncada, M. W. Radomski and R. M. Palmer, Endothelium derived relaxing
factor : identification as nitric oxide and role in the control of vascular tone and
platelet function, Bioehem. Pharmaeol. 37: 2495 (1988) .
6. B. V. Khan , D. G. Harrison, M. T. Olbrych, R. W. Alexander and R. M . Medford,
Nitric oxide regulates vascular cell adhesion molecule 1 gene expression and redox-
sensitive transcriptional events in human vascular endothelial cells, Proe. Natl. Aead.
Sei. U.S.A. 93: 9114 (1996) .
7. Y. Kubota, H. K. Kleinman, G. R. Martin and T. J. Lawley, Role oflaminin and
basement membrane in the morphological differentiation ofhuman endothelial cells
into capillary-like structures, 1. Cell Bio!. 107: 1589 (1988) .

545
 INVOLVEMENT OF NITRIC OXIDE IN 1HE INHIBITION OF ANGIOGENESIS
BY INTERLEUKlN-2

Eleni Sakkoula, Eva Pipili-Synetos & M.E. Maragoudakis

University ofPatras, Medical School,

Department of Pharmacology,
261 10, PATRAS, GREECE

Interleukin-2 is a weil described immunoregulatory cytokine with potent therapeutic

activity in a variety (I) and an inducer of nitric oxide (NO) synthase in mice and humans
(2) . Moreover NO has been shown to contribute to the IL-2-induced antitumour responses
(3, 4). We have shown that NO inhibits angiogenesis in the CAM as weil as tumour
growth and metastasis in the Lewis Lung carcinoma in mice (5). Since NO may mediate IL-
2 activity, the question was addressed as to whether the IL-2 antitumour effects may be
due to inhibition of angiogenesis via NO.
In order to examine the above hypothesis the activity of IL-2 was evaluated in the in
vivo and in vitro system ofthe chick chorioallantoic membrane (CAM). In this system , in
vivo, IL-2 caused a dose-dependent inhibition of angiogenesis. This result was fully
reversible by the competitive NO synthase inhibitor NG-nitro-L-arginine methyl ester (L-
NAME) . In the CAM in vitro, IL-2 caused an increase in NO synthase activity and this
effect was again fully reversed by L-NAME.
These results indicate that IL-2 may be an important antiangiogenic moleeule causing
its effect through NO synthesis. The antiangiogenic activity ofIL-2 may be at least partly
responsible for its antitumour properties.

REFERENCES

WHITTINGTON, R. & FAULDS, D. (1993) . Interleukin-2 :A Review . Drugs, 46(3),

446-514.

DENG, W ., TRIEL, B., TANNENBAUM, C.S., HAMILTON, T.A . & STUEHR, DJ.
(1993). Synergistic cooperation between Tcell Iymphokines for induction of the nitric
oxide synthase gene in murine peritoneal macrophages. 1. Immunol., 151,322-329.

546
HIBBS, 1.B., WESTENFELDER, C., TAINTOR, R., VAVRIN, Z., KABLITZ, c.,
BARANOWSKI, R.L., WARD, H., MENLOVE, L., McMURRRY, M.P., KUSHNER,
J.P. & SAMLOWSKI, W.E. (1992). Evidence for cytokine-inducible nitric oxide synthesis
from L-arginine in patients receiving interleukin-2 therapy. 1. Clin. Invest., 89, 867-877 .

YIM, C-Y., McGREGOR, 1.R., KWON, O.D., BASTIAN, N.R., REES, M., MORI, M .,
HlBBS, J.B. & SAMLOWSKI , W.E. (1995). Nitric oxide synthesis contributes to IL-2-
induced antitumour responses against intraperitoneal meth A tumour . 1. Immunol., 4382-
4390 .

PIPILI-SYNETOS, E., PAPAGEORGIOU, A., SAKKOULA, E., SOTIROPOULOU,

G., FOTSIS, T., KARAGIOULAKIS, G. & MARAGOUDAKIS, M.E. (1995) .
Inhibition of angiogenesis, tumour growth and metastasis by the NO-releasing
vasodilators, isosorbide mononitrate and dinitrate. Br. 1. Pharmacol. 116, 1829-1834

547
GONATOTROPIN CONTRIBUTION TO TUE ANGIOGENIC POTENTIAL OF
HUMAN EPITHELIAL OVARIAN CARCINOMA

Yael S. Schiffenbauer", Rinat Abramovitch", Gila Meir+, Nava Nevo",

Michael Holzingerf, Ahuva Itin*, Eli Kesler* and Michael Neeman".
+Dept. of Biological Regulation, The Weizmann Institute of Science,
Rehovot 76100, Israel.
itMeir Hospital, Kfar Saba Israel,
*Dept of Molecular Biology, Hebrew Univerity-Habassah Medical
School, Jerusalem. Israel.

Ovarian cancer presents a major challenge to oncological research. This cancer is

frequently diagnosed late and shows very poor prognosis and low survival despite the
apparent success of chemotherapy in achieving remission with no evidence of disease .
Recurrence is characterized by multifocal spread of tumors in the entire abdominal cavity .
The goal of this work was to determine whether gonadotropins (LH and FSH) act as
inducers of recurrence at this stage, by induction of neovascularization of small micro-
tumors .
The effects of gonatropins were tested using two difffent hormonal interventions:
ovariectomy and direct administration of gonatropins (9 units per animal every 48 hours) .
We found that gonatropins enhased to growth of three different human epithelial ovarian
cancer celllines implanted intraperitonealy as weil as subcutanous in female CD-l nude
mice.
In contrast with their pronounced effect in vivo, gonatropins did not affect the
proliferation and metabolism of human ovarian carcinoma in vitro . We therefore looked
more closely at the early days after implantation of MLS spheroids, and found that
primary neovascularization of the implanted human ovarian carcinoma spheroids was
enhanced in hypergonatropic mice as determined by the angiogenic contract from MR
imagesl .
The hormonal stimulation of angiogenic properties could also be reproduced in vitro.
Condiotioned medium of LH-stimulated MLS cells was mitogenic to bovine aortic as weil
as to capillary endothelial cells. DNA synthesis in the endothelial cells increased after
treatment with LH conditioned medium relative to non LH conditioned medium. No direct
effect ofLH on the endothelial cells was observed.
RT-PCR and in situ hybridization studies performed on MLS spheroids identified
the hormonally induced angiogenic growth factor as VEGF, showing a LH and FSH dose

548
dependent VEGF induction . In addition, the mitogenic effect of the LH-stimulated ovarian
cancer cells on the endothelial cells was almost completely canceled in the presence of anti
LH or anti VEGF neutralizing antibodies .
The results reported here suggest that gonatropins promote ovarian tumor growth by
induction of VEGF-mediated angiogenesis. Hormonal intervention may be effective in
prolonging dormancy of ovarian cancer micrometastases, but hormonal sensitivity will be
gradually lost along with hypoxie induction ofVEGF.

REFERENCES

1. R. Abramovitch, G. Meir and M. Neeman. Neovascularization induced growth of

implanted C6 g1ioma multicellular spheroids : magnetic microimaging Cancer Res .
55 : 1956-1962, 1995.

LH and FSH were provided by The National Hormone and Pituitary Program (NHPP).

549
VASCULAR MODULATlONS IN THE PREOVULATORY RAT FOLLICLE

C. Tempel, G. Meir and Neeman

Department ofBiological Regulation, Weizmann Institute ofScience,

Rehovot 76100, ISRAEL.

Ovulation and luteinization are key processes in mammalian reproduction . They are
also unique example of neovascularization and cell proliferation which are usually
characteristic of cancer and occuring in this case in weil regulated normal physiology. The
prevulatory follicle is 1 mm in diameter. It contains from the outside to the inside a
basement membrane, a rim of granulosa cells and the follicular fluid that surround s the
cumulus oocyte complex. The blood vessels surrounding each prevulatory follicle
proliferate without penetrating the basement membrane until ovulation . Elevated
expression ofVascular Endothelial Growth Factor (VEGF) was reported in the inner parts
of the maturing follicle (1). We applied magnetic resorance imaging for following the
changes in the vascular system during follicular development in the normal cycling rat
ovary.
Immature female rats were induced to ovulate by administration of Pregnant Mare
Serum Gonadotropins (pMSG) followed by administration of human Chorionic
Gonatropin (hCG) . This protocol results in synchronous development of up to 50 follicles
in each ovary . By application of magnetic resonance imaging (MRI) and spectroscopy
(MRS) we have followed the changes in the diffusion properties and energy metabolism of
explanted perfused follicles and measured the actual ovarian perfusion in vivo . Four hours
after the surge of LH, we found a significant decrease in the self-diffusion coefficient of
water in the follicular fluid, followed by an increase just before ovulation (2). Changes in
the metabolism were also observed by the same time: decrease amount of high energetic
metabolites, as ATP and Phosphocreatine, as weil as an increase rate of glucose metabolism
into lactate . These data suggested that metabolic stress in part ofthe normal physiology of
the developing preovulatory follicle after LH stimulation, and could perhaps contribute to
the induction ofVEGF expresion.
In vivo these effects may be modulated by changes in blood flow in the periphery of
the follicle . A specific MRI pulse sequence of measuring specific changes in ovarian blood
flow was developed . Our results showed that LH stimulation does not lead to long term
increased blood flow per volume, and thus angiogenesis exactly matches the increased
volume ofthe follicle during follicular maturation. In addition, inceased vessel permeability

550
was implicated by the spatial variance in the perfusion experiment, and was verified by
intravenous administration ofthe contact reagent Gd DTPA.
In conclusion, our data suggests that in the maturing follicle VEGF may be induced
by hypoxia and could contribute to the observed increased vessel permeability as weil as to
the angiogenic activity. Magnetic resonance imaging of ovarian blood flow provides a non
invasive method for monitoring the functional consequence of physiological angiogene sis.

REFERENCES

I. Tempel C., Schiffenbauer y'S., Meir G., Neeman M., Modulation of water diffusion
during gonatropin-induced ovulation : NMR microscopy of the ovarian follicle, Mag .
Res . Med ., 34, 213-218,1995 .
2. Phillips H.S., Hains J., Leung D. W., Ferrara, N. Yascular Endothelial Growth Factor
is expected in rat corpus luteum, Endocrinology , 127,965-967,1990.

551
OXYGEN REGULATES VEGF- INDUCED VASCULOGENESIS.

Alda Tufro McReddie and R. Ariel Gomez.

University of Virginia School of Medicine, Department of Pediatrics

CharlottesvilIe, VA 22908, USA

The determine whether low oxygen is a stimulus for endothelial cell

differentiation and vascular development in the Kidney, we examined the effect of low
oxygen on rat metanephric organ culture, a model known to rec apitulate
nephrogenesis in the absence of vessels. After 6 days in culture in standard (20% 02)
or low oxygen (l %-3% 02) conditions, metanephric Kidney growth and morphology
were assessed by light and electron microscopy, DNA measurement, and
morphometric analysis. Low oxygen induced proliferation of tubular epithelial celIs,
resulting in enhanced number of tubules of similar size. Endothelial cells forming
capillaries were localized in 3% 02 explants by light and electron microscopy and by
immunocytochemistry using endothelial cell markers. Flt-l , F1k-1 and ACE containing
cells were detected in 3% 02 explants whereas 20% 02 explants were virtually
negative. The determine the mechanism(s) of hypoxia-induced vasculogenesis, VEGF
expres sion was examined by northem blot analysis, VEGF mRNA levels were 10 fold
higner in 3% 02 explants than in 20% 02 explants. Addition of anti-VEGF antibodies
to 3% 02-treated explants prevented low oxygen-induced growth and endothelial cell
defferentiation and proliferation. The fate of developing endothelial cells was
examined by anti-Flk l and anti-PCNA antibodies and by labeling apoptotic cells using
the tunel method in 20% 02 and 3% 02 explants from Sprague-Dawley rats and
Flkl(+/-) transgenic mice expressing ßGal in endothelial cells. 3% explants showed
numerous Flkl-positive, PCNA-positive cells surrouding forming nephrons and
ureteric buds and few, randomly distributed apoptotic cells. In Flkl(+/-) mouse 3% 02
explants PCNA-positive cells co-Iocalized with ßGaI stained cells whereas 20%
explants showed fewer PCNA-positive cells, no Flkl-positive cells and numerous
apoptotic cells.
Our data indicate that low oxygen stimulates growth by cell proliferation and
induces tubulogene sis, endothelial cell differentiation and proliferation resulting in
vasculogenesis in metanephric Kidneys in culture. Upregulation of VEGF expression
by low oxygen and prevention of low oxygen-induced tubulogenesis and
vasculogenesis by anti-VEGF antibodies indicate that these changes were mediated by
VEGF. These data suggest that low oxygen isthe stimulus to initiate renal
vascularization.

552
REFERENCES
1) Tufro-McReddie et al. Oxygen regulates VEGF-induced vasculogenesis and
mephrogenesis. Oe. Biol (in press), 1997.
2) Tufro-McReddie et aI. VEGF induces nephrogenesis and endothelial cell
differentiation during Kidney development. Am. 1. Physiol: (in press), 1997.
3) Tufro-McReddie et al. Vasculogenesis in metanephric organ culture. J. Am. Soc.
Nephrol 7:1606,1996.

553
WOUND HEALING WITH EFG CONTAINING GELATIN SPONGES

K. U1ubayram*, C. Ertan#, P. Korkusuz+, N. Cakar+ and N. Hasirci*

*Middle East Technical University, Dept. Chemistry, 06531 Ankara

#Middle East Technical University, Medical Center, 06531 Ankara
+Hacettepe Univ., Fac. Med., Histology & Embryology , 06100 Ankara

There is an intense research on the synthesis of new materials which could be

candidates in biomedical applications . One important area is to cover the skin with a
compatible graft in case of severe bums in order to prevent humidity loss from the body
as weil as to prevent microbial attact to the damaged area. These kind of grafts generally
are elastic, water permeable ahd have biodegradible structure . Another area is the use of
growth factors to stimulate the tissue generation. These are various types of growth
factors and among them epidermal growth factor (EGF) is very important in wound
healing and epidermal regeneration.

In this study, a gelatinous skin graft was synthesized by freeze-dry technique and
certain amount of EGF was entrapped into the matrix. The prepared matrices were
tested on rabbits . For this purpose some wounds were created on the animals by
punching the skin. One punced area was left as is as control and the second one was
covered with gelatin while the third one was covered with EGF containing gelatin
sponge .

The skin was removed in every week and the tissue regeneration, blood vessel
formation , and tissue orientation was examined histologically. It was observed that
sponges which contained EGF were very effective in the tissue regeneration and wound
healing.

554
TUE REGULATIVE ACTIVITY OF DECORIN IN TUMOR MEDIATED
ANGIOGENESIS

Cigdem Yentsey l , Manoranjan Santra-, Renato V. Lozzo-, R.

Wesley Rose 1, Derric S. Grant 1.

IThomas Jefferson University, Department of Medicine, Careza

Foundation for Hematologic Research , 1015 Walnut Str., Curtis
BLDG., Room 703, Philadelphia, PA 19107-5099, USA.
2Thomas Jefferson University, Jefferson Medical College ,
Department of Pathology, Anatomy and Cell Biology and the
Kimmel Cancer , Jefferson Alumni Hall, 1020 Locust Str .,
Philadelphia, PA 19107, USA.

The progressive growth of most neoplasm is dependent on the induction and

formaton of nev blood vessels by factors secreted by tumor cells. We have examined
the in vitro and in vivo angiogenic ability of fibrosarcoma (HT1080), cervix carcinoma
(Hela) and, osteosarcoma (Saos2) wild-types or their decorin-expressing
counterparts, by evaluating the effect of their secretions on angiogenesis . This was
done in vitro by using human umbilical vein endothelial cells (HUVEC) to examine
four essential steps of angiogenesis cell proliferation, migration, attachment, and
differentiation (tube formation on Matrigel). The wild-type tumor secretions
enhanced HUVEC attachment, migration and differentiation, whereas their decorin-
transfected forms suppresed these processes . Matrigel induced tube formation by
HUVEC was greater in the wild-type cells as compared to their decorin tranfeced
forms. The same cells were evaluated in vivo angiogenesis model, in which tumor
cells were either mixed with Matrigel and injected subcutaneously in nude mice, or
sealed in millipore Chambers and then implanted subcutaneosuly in nude mice. The
wild-type tumor cells grew better in vivo then their decorin transfected counterparts
and stimulated more invasion of vessels into the surrounding tumor bed and the tissue
around the implanted chamber. These results indicate that decorin plays an important
role in the suppression of tumor-mediated angiogenesis,

555
ULTRASTRUCTURAL STUDY OF THREE lYPES OF PHYSIOLOGICAL
ANGIOGENESIS IN ADULT RAT SKELETAL MUSCLE

A-L Zhou, S. Egginton, and O. Hudlicka

Dept. ofPhysiology, University ofBirmingham, UK

Capillary growth in developing tissue or under pathological conditions occurs

either by sprouting (1, 2), or intussusceptive growtMongitudinal splitting (3, 4). Whild
growth factors are important in the latter, mechanical factors connected with increased
blood flow and/or growth of surrounding tissue may be important in the former. We
explored the capillary fine structure in 3 models of physiological angiogenesis in adult
skeletal muscles induced by various mechanical factors . Administration of the
vasodilator prazosin or extirpation of agonist led to growth of new capillaries by means
of increased blood flow (increased shear stress/wall tension) or by stretch of surrounding
tissue , respectively (5, 6). The luminal stimuli resulted in increased capillary growth
without disurbance ofthe basement membrane and abluminal sprouting, but accompanied
by a significant increase in luminal thin (filiform-like) or thick endothelial cytoplasmic
protrusions and formation of septa dividing the lume in two, possibl y followed by
subsequent longitudinal splitting of capillaries (7). The abnuminal stimuli (stretch)
resulted in abluminal sprouting and associated focal breakage ofthe basement membrane
and endothelial proliferation (8). Indirect electrical stimulation also induces capillary
growth , with stimuli acting both on the lumina! (due to increased blood flow in
contracting muscles) and ablumina! surfaces (alterating contraction and relaxation and
possible disturbance of the extracellular matrix, 9). These resulted in both lumina! and
abluminal growth patterns .

REFERENCES

1. Ciiff WJ, 1963. Observations on healing tissue : a combined light and electron
microscopic investigat ion. Phi!. Trans. Roy. .Soc. B 246: 305-325 .
2. Scoazec J-Y, Degott C., Reynes M., Benhamou JP and Feldmann G., 1989.
Mechanism of neovascularization: vascular sprouting can occur without
proliferation of endothelial cells. Lab. Invest. 51: 624-634.

556
3. Burri PH and Tarek MR, 1990 . A novel mechanism of capillary growth in the rat
pulmonary microcirculation. Anat. Rec . 228 :35-45 .
4. Van Groningen JP, Wenink ACG and Testers LHM, 1991. Myocardial capillaries:
increase in the number by splitting of existing vessels. Anat. Embryo\. 184 :65-70.
5. Dawson JM and Hudlicka 0 ., 1993. Can changes in microcirculation explain
capillary growth in skeletal musele? Int. J. Exp . Path. 74 : 65-71 .
6. Egginton S., and Hudlicka 0 ., 1992. Effect of long-term musele overload on
capillary supply, blood flow and performance in rat fast musele. 1. Physio\.
452 :9P .
7. Zhou A-L. , Egginton S. and Hudlicka 0., 1996. Ultrastructural evidence for a novel
mechanism of capillary growth in rat skeletal musele. 1. Physio\. 491 : 28P .
8. Zhou A-L. and Egginton S., 1997 . Capillary growth in overloaded rat skeletal
musele: an ultrastructural study . 1. Physiol., in press.
9. Hansen-Smith F .M ., Hudlicka O. and Egginton S. 1996 . In vivo angiogenesis in
adult rat skeletal musele : early changes in capillary network arch itecture and
ultrastructure. Cell Tissue Res . 286 : 123-136 .

557
VI
VI
00

Participants of the NATO Advanced Study Institute "Angiogenesis: Models, Modulators and C1inical Applications"
held at Ca psis Hotel, Rhodes, Greece, during 20-30 June, 1997
 1. Tziora, Georgia 15. DiPietro, Luisa 30. Zhou, Al-Ling 44. Rosen, Elliot 58. Tsopanoglou, Nick
2. Sgoutas, Maria 16. Dimitropoulou, Christiana 31. Weiss, Iaqueline 45. Liekens, Sandra 59. Presta, Marco
3. Sgoutas, Demelrios 17. Collen, Annernie 32. Hudlicka, Olga 46. Koolwijk, Pieter 60 . Sitaras, Nick
4. Kefalides, Nick 18. Sakkoula, Eleni 33. Babaoglou, Melih 47. Paavonen, Kari 61. Bicknell, Roy
5. Mannara, Anna 19. Tempel, Catherine 34. Borel, lacques 48. Navarro, Pilar 62. lohnson, David
6. Missirlis, Elephtheria 20. Benezra, David 35 . Borei, Mrs. 49. Hatji, Elissavet 63 . HOckei, Michael
7. Rose, Wesley 21. Schiffenbauer, Yael 36 . AchaIya, Ravi 50. West, David 64. Haudenschild, Christian
8. Ziehe, Maria 22. Neeman, Michal 37. Leonidas, Dimitrios 51. Krooo, Marioa 65 . Catravas, lohn
9. Maragoudakis,lane 23. Haljiconti, Olga 38 . Kisseleva, Ekaterina 52. BaneJjee,Dipak 66 . Hellerqvist, Cad
10. Maragoudakis, Michael 24. Malms, Helene 39. Hellerqvist, Charlotte 53. Fan, Tai-Ping 67 . Ribatti, Dornenico
11. Yenisey, Cidmen 25. Larcher, Femado 40 . Abramovich, Rinat 54. Schlenger, Kadeingcr 68 . Maragoudakis, Michael, Ir.
12. Kanthou, Chryso 26. Conti, Claudio. 41. Bastaki, Maria 55. Hassessian, Haroutioun
13. Kirernitei, Menemse .28. Papadimitriou, Evangelia 42. Lymboussaki, Athina 56. McLaughlin, Barry
14. Ozturk, Melek 29. Sotiropoulou, Georgia 43. lussila, Lotta 57. Dcchend, Ralf
VI
VI
-e
PARTICIPANTS

ABRAMOVICH, R. Weizmann Institute of Science, Department

of Biologica1 Regulation, 76 100 Rehovot,
ISRAEL

ACKARYA,R. University of Bath, School of Bio!. &

Biochemistry, Claverton Down, Bath BA2
7AY, UNITED KINGDOM

BABAOGLOU, M. Hacettepe University , Faculty of

Medicine, Department of Pharmacology,
06100 Sihhiye-Ankara, TURKEY

BANERJEE, D. University of Puerto Rico, School of

Medicine, Department of Biochemistry,
San Juan, PR 00936-5067, USA

BASTAKI, M. American Red Cross, The Jerome Holland

Laboratory, 15601 Crabbs Branch Way,
Rockville, MD 20855, USA

BELLAROSA, D. Menarini Ricerche, Department of

Pharmacology, V. Tito Speri 10, 00040
Pomezia, Roma, ITALY

BENEZRA, D. Department of Ophthalmology, Hadassah

Medical Organization Kiryat Hadassah,
P.O.B. 1200011-91120 JerUS.A.lem,
ISRAEL

BICKNELL, R. Molecular Angiogenesis Group Imperial

Cancer Research Fun, Institute of Molecular
Medicine, University of Oxford, John
Radcliffe Hospital, Oxford OX3 9DU ,
UNITED KINGDOM

BOREL, J. University of Reims, Faculty of Medicine,

51 Rue Cognacq Jay, 51095 Reims Cedex,
FRANCE

560
BROOKS,P. The Scripps Research Institute, 10550 N .
Torrey Pines Road, IMM24 , La Jolla, CA
92037, USA

CATRAVAS, J. Medical College of Georgia, Department of

Pharrnacol. & Toxicology, Augusta , GA
30912-2300, USA

COLLEN,A. TNO Gaubius Institute, P.O. Box 430, 2300

AK Leiden, NETHERLANDS

CONTI, C. Department of Carcinogenesis, M .D .

Anderson Cancer Center, Science Park-
Research Division, Smithville, Texas 78957,
U.SA

DECHERD,R. Max Delbrück Center, Robert Rössle Str.

101,31225 Berlin, GERMANY

DIMITROPOULOU, C. Institute of Anatomy, University of Mainz,

Johannes Gütenberg, Beckerweg 13, 55099
Mainz , GERMANY

DEMOPOULOS, 1. University of Patras Medical School,

Department of Radiology, 265 00 Rio,
Patras, GREECE

DIPIETRO, L. Loyola University Medical Center, Burn &

Shock Trauma Institute, Bldg. 110, Room
4251,2160 South First Avenue, Ma ywood ,
Illinois 60153, USA

DONNINI, S. Department of Pharrnacology, University of

Firenze, Viale de Morgagni, 6550134
Firenze, IT AL Y

FAN, T.P. University of Cambridge, Dept of

Pharrnacology, Tennis Court Road,
Cambridge CB2 lQJ, UNITED KINGDOM

FENG,1. Wound Healing Lab, UCSF, Department of

Surgery, 513 Pamassus Avenue HSW 1652,
San Francisco, CA 94143-0522, US.A.

FLORDELLIS, C. University of Patras Med . School,

Department of Pharrnacology, 261 10 Rio,
Patras, GREECE

GIA VAZZI, R. Instituto di Ricerche Farrnacologiche

MARIO NEGRI , Laboratori Negri Bergamo,
Via Gavazzeni 11, 24100 Bergamo, ITALY

561
GIBSON, 1. University of Califomia, Department of
Surgery, Pamassus Ave., Room HSE 839,
San Francisco, U.S.A.

HANSEN-SMITH, F. Department of Biological Sciences,

Oakland University, Rochester , MI 48309-
4401, USA

HASIRCI, N. Middle East Technical University,

Department of Chemistry, 06531 Ankara,
TURKEY

HASSESSIAN, H. University of Montreal, Faculty of

Medicine, Department of Ophthalmology,
Centre of Reseach Guy-Bemier (HMR),
5689 bou!. Rosemont, Pavillon Lavoisier-3,
Montreal (Quebec), HIT 2HI CANADA

HATJICONTI, O. University of Patras Med. School,

Department of Pharmacology, 261 10 Rio,
Patras, GREECE

HATZI, E. University of Ioannina Medical School,

Department of Biochemistry, Ioannina,
GREECE

HAUDENSCHILD, C. American Red Cross, The Jerome Holland

Laboratory, 15601 Crabbs Branch Way,
Rockville, MD 20855,USA

HELLERQVIST, C. Carbomed INC., 5115 Maryland Way,

Brentwood, TN 37027, U.S.A.

HELLERQVIST , CH. Vanderbilt University, School of Medicine,

Department of Biochemistry , 23rd Pierce,
Nashville, TN 37232-0146, U.S.A.

HOECKEL, M. University of Mainz Med. School,

Department of Obstretics/Gynecology,
Langenbeckerstr. I, P.O. Box 3960, 6500
Mainz, GERMANY

HUDLICKA, O. Department of Physiology , The University

of Birmingham, Medical School, Viricent
Drive, Birmingham B 152TJ, UNITED
KINGDOM

HUNT ,T. University of Califomia, Department of

Surgery, 513 Pamassus Avenue, Room HSE
839, San Francisco, CA 94143, USA

562
HUSSAIN,Z. University of Califomia, Department Of
Surgery, Parriassus Ave ., Room HSE 839 ,
San Francisco, U.S.A.

JOHNSON, D . East Tennessee State University, James H.

Quillen College of Med icine, Department
of Biochemistry & Mol. Biology, Box
70581 , Johnson City Tennessee 376 14-
0581 , U.SA

JUSSILA, L. University of Helsinki, Haartman Institute,

Department of Pathology, PO Box 21
(Haartmanisikatu 3), FIN 00014 Helsinki ,
FINLAND

KANTHOU,C. Thrombosis Res. Institute, Emanuel Kaye

Building, Manresa Road, Chelsea, London
SW3 6LR, UNITED KINGDOM

KEFALIDES, N . University ofPennsylvania, Department of

Medicine, University City Science Ctr,
3624 Market Street, Philadelphia,
Pennsylvania 19104, U.S.A.

KIREMITCI, M . Hacettepe University, Chem . Eng.

Department, 65320 Be ytepe-Ankara,
TURKEY

KISSELEVA, K. Institute of Experimental Medicine,

Academic Pavlou str. 12, St. Petersburg 197
376, RUSSIA

KLEINMAN, H . National Institute of Dental Res ., Section of

Dev . Biol. & Anomalies, Bldg . 30, R. 414 ,
Bethesda, MD 20892-4370, U.SA

KONERDING, M . Institute of Anatomy , University of Mainz,

Johannes Gütenberg, Beckerweg 13, 55099
Mainz, GERMANY

KONERDING-HERBST, C. Alfried-Krupp-Krankenhaus, Alfried-Krupp

str. 21, 45117 Essen, GERMANY

KOOLWIJK, P. TNO Gaubius Institute, P.O. Box 430, 2300

AK Lieden, THE NETHERLANDS

KROON, M . TNO Gaubius Institute, P.O. Box 430, 2300

AK Leiden , NETHERLANDS

LARCHER,F. Department of Cell & Molecular Biolog y,

CIEMAT, Avenida Complutense 22, 2804 0
Madrid, SPAIN

563
LELKES, P.I. University of Wisconsin, Lab of Cell
Biology, Department of Medicine, Sinai
Samaritan Medical Center, Mount Sinai
Campus, 950 North Twelth str., Box 342,
Milwaukee WI 53201-0342, USA

LEONIDAS, D. University of Bath School of Biol. &

Biochemistry, Claverton Down, Bath BA2
7AY, UNITED KINGDOM

LI, W. The Angiogenesis Foundation, Inc., P.O. Box

383011, Cambridge, MA 002238, USA

LIAKOS, D. Newcastle University, Medical Molecular

Biology Group, Floor 4, Catherine
Cookson Bldg., Framlington Place,
Newcastle upon Tyne NE2 4HH ,
UNITED KINGDOM

LIEKENS,S. K.U Leuven, Rega Institute , B-3000

Leuven, BELGIUM

LYMBOUSSAKI, A. University of Patras Medical School,

Department of Pharmacology, 261 10 Rio,
Patras, GREECE

MANOLOPOULOS, V. K.U LEUVEN Medical School,

Laboratory of Physiology , Gasthnisberg,
3000 Leuven, BELGIUM

MANTELL, D. University of Mancester School of Biological

Sciences, Biochemistry Research Division,
Manchester, M13 9PT, UNITED
KINGDOM

MARAGOUDAKIS, M. University of Patras Medical School,

Department of Pharmacology, 261 10 Rio,
Patras, GREECE

MARAGOUDAKIS, M.Jr. University of Patras Medical School,

Department of Pharmacology, 261 10 Rio,
Patras, GREECE

MCLAUGHLIN, B. Department of Rheumatology, Wolfson

Angiogenesis Unit, Clinical Science Bldg.,
Hope Hospital , Salford M6 8HD, UNITED
KINGDOM

MISSIRLIS, E. University of Patras Medical School,

Department of Pharmacology, 261 10 Rio,
Patras, GREECE

564
NAVARRO, P. Institute of Res. Phannacol. MARIO
NEGRI, Vascular Biology Lab., Via Eretria
62, 20157 Milano, ITALY

NEEMAN, M. Weizmann Institute of ScienceDepartment of

Biological Regulation, 76 100 Rehovot ,
ISRAEL

NICOSIA , R. Allegneny University of the Health

Sciences, Departmen of Pathology, New
Collge Bldg., MS #435, Broad and Vine
str., Philadelphia, PA 19102, U.S.A.

OZTURK, M. University of Istanbul, Cerrahpasa Faculty

of Medicine, Department of Medical
Biology, 34300 Cerrahpasa Istanbul ,
TURKEY

PAAVONEN,K. University of Helsinki, Haartman

Institute , Department of Pathology , P.O.
Box 21 (Haartmanisikatu 3), FIN 00014
Helsinki, FINLAND

PAPADIMITRIOU, E. University of Patras Medical School,

Department of Phannacology, 261 10 Rio,
Patras, GREECE

PIPILI-SYNETOS, E. University of Patras Medical School,

Department of Phannacology, 261 10 Rio,
Patras, GREECE

POLVERINI, P. University of Michigan, School of

Dentistry, Room 5217, Ann Arbor, MI
42109-1078, U.SA

PRESTA, M. University of Brescia, Department of

Science, Biomedicine and Biotechnology ,
Via Valsabbina 19, 1-25123 Brescia,
ITALY

RAVAZOULA, P. University of Patras Med . School,

Department of Pathology, 261 10 Rio,
Patras, GREECE

RIBATTI, D. University of Bari, Institute of Anatomy,

PIazza Giulio Cesare, 11 , Policlinico, 1-
70124 Bari, ITALY

ROSE, W. Thomas JetTerson Univ. Hospital , Cardeza

Foundation for Hematol. Res., 1015
Walnut Str., Curtis Bldg., R. 703,
Philadelphia, PA 19107-5099, U.S.A.

565
ROSEN,E. Long Island Jewish Medical Ctr.,
Department of Radiation Oneology, 270 05
76th Avenue, New Hyde Park, NY 11040,
U.S.A

SAKKOULA, E. University of Patras Medieal Sehool,

Department of Pharmaeology, 261 10 Rio,
Patras, GREECE

SCHIFFENBAUER, y. Weizmann Institute of Scienee,

Department of Biologieal Regulation, 76
100Rehovot, ISRAEL

SCHLENGER, K. University of Mainz, Department of

Gyneeology and Obstreeties,
Langebeekerstrabe 1, 55131 Mainz,
GERMANY

SGOUTAS, D. Clinieal Chemistry Laboratory, Emory

University Hospital , Room F-153C
Atlanta, Georgia 30322, U.SA

SITARAS, N. University of Athens, Medieal Sehool,

Deptartment of Pharmaeology, Goudi,
Athens, GREECE

SKOTSIMARA, P. University of Patras Med . Sehool,

Department of Toxieology, Gen. Hosp . of
Patras "Agios Andreas", Patras, GREECE

SOTIROPOULOU, G. University of Patras, Sehool of Pharmaey,

Department of Pharmaeognosis, 261 10
Rio, Patras, GREECE

TEMPEL , C. Weizmann Institute of Science,

Department of Biologieal Regulation, 76
100 Rehovot, ISRAEL

THOMPSON, D. University of Aberdeen, Department of

Pathology, Univ. Med. Bldgs., ForesterhilI
Aberdeen AB92ZD, UNITED KINGDOM

TSOPANOGLOU, N. University of Patras Medieal Sehool,

Department of Pharmaeology, 261 10 Rio,
Patras, GREECE

TUFRO-MCREDDIE, A. Division of Nephrology , Univ . Of Virginia

Sehool of Medieine, MR4 BldgIRoom
2006/2040 , Charlottesville VA 22908,
U.SA

566
TZIORA, G. University of Patras Medical School,
Department of Pharmacology, 261 10 Rio,
Patras, GREECE

WEIDNER, N . University of California, Department of

Pathology, San Francisco , U.S.A.

WEISS,1.B. Department of Rheumatology, Wolfson

Angiogenesis Uriit, Clinical Science Bldg.,
Hope Hospital , Salford M6 8HD, UNITED
KINGDOM

WEST,D. Department of Immunology, University of

Liverpool P.O. Box 147, Liverpool L69
3BX, UNlTED KINGDOM

WINKLER, J. SmithKline Beecham Pharmaceutical,

Department of Pharmacology, VW 2532,
709 Swedeland Rd, King of Prussia, PA
19406, USA.

YENISEY, C. Adnan Menderes University, Faculty of

Medicine, Department Of Biochemistry,
09100 Aydin, TURKEY

ZHOU, A-L. Department of Physiology, The University

of Birmingham, Medical School, Vincent
Drive, Birmingham B 152TJ, UNITED
KINGDOM

ZICHE, M. University of Firenze, Department of

Pharmacology, Viale Morgagni 65, 50134
Firenze, lTALY

567
I NDEX

Aden osine. 204 . 207. 209 . 210 Chick chorioallantoic membran e (C AM ), 35 37.
Ad here ns junetion. IX9. 197. 199 309- 3 12, 314
Ad hesion molecu lcs. 367. 369 . 373 Clinica l tria ls. 473, 4 75- 4 X6
Adrcna l rncdulla. 7. 9- c-Met, 4 15-420. 424
Adva ncc d canccr, 407 Co llage n. 285-287, 289- 294
Arnphornycin , 15 Co lonic carci nornas . 442, 444
Ang iogene sis. 7. 12. 14--- 16. 39.40.42.44. 4 7. Co ntras t enhancernent . 55, 57
4X. 5 ~52 . 55- 5X. 8 ~84 . 85. 87-89. 99 . Co rros io n ca st. 429 , 432. 439, 44 2
10 1. 102. 105. 106. 113-1 16. 118- 120. Cyc lic AMP, 209
121-126. 12X. 137. 138. 143- 14 7. Cyc lic GMP, 297, 298, 300
157-1 64. 165. 175, 225, 226, 22 8. Cy tokines,89. 113, 117. I 18, 349 , 352
23 ~232 , 233 , 238 , 240, 241 , 242. 248.
250 . 25 1, 297-299, 30 1-303 . 30 7- 309. Differe ntiat ion, 157, 159, 163, 164
3 1 I. 349 , 350, 352 . 353 , 389- 394 , Diffusion, 58
41 5-41 7, 420-422 ,424.425.473-487 Dol ic ho l-ca scade pa thwa y, 15
Angio genie switch. 45 8, 46 7 Dol-P-Man synthase, \5. 16
Angi og enin , 165-1 67, 175- 177
Anti-angi ogen esis, 476. 482 Endothelial cellts) , 47. 50, 5 1. 157-164. 187,
Anti-angiogenie cornpounds , 171 189-1 92, 194 , 197- 200, 241- 246, 248- 260,
As paragi ne- lin ked glycoproiein. 11. 16 285-288, 29~294. 367 , 369. 370, 372. 373,
Ate nolo l, 13, 16 377-379,384
Atherosc lerosis, 233 , 240 Endothelial-cell-stirnulating angiogenesis factor, 138.
146
Base me nt membrane(s) (BM) , 35, 36, 157, 159, Ende the lium. 100, 10 1. 106
1 62- 1 64 , 2 03 , 2 04 , 3 7 7, 3 7 8 , 3 8 ~3 8 5 Epitope, 233, 235, 240
Basic fibrobla st growth factor (bFGF), 126- 128. 436, Extracellular matrix, 47, 5 1, 225, 229 , 231
438 ,439
Beta-adrenocept or, 12, 13, 16 FGF-2 , 138-1 4 5
Biotechn ology,478 rece ptors , 143
Bladder ca ncer, 41 9, 420 Fibrin matrix, 250
Blood clotting Factor VIII:C, 12 Fibri noge n, 233
Brea st ca ncer, 41 7, 41 9, 421-424 Fibrino lysis, 233, 240
Fibrob lasu s), 47, 5~52 , 285 , 286 , 289- 291, 293 ,
c-AMP, 13, 15, 16 294
Ca d heri ns , 188-190, 195 Fibrob last growth facto rts), 99 ,107,1 10. 29 8-300,
Ca nce r, 473-480, 484, 486 302 , 303
Capillary endot helial ce lls, 12, 13, 15 Fibr oblasts, 47, 5~52
Capillary growth, 137, 138, 143-1 4 7
Ca p illary-like struc ture , 157, 164 Gene express ion, 349, 352
Ca pillary sprouts, 80, 82, 85 Gliornas, 420, 424
Ca pillary tube, 15, 18 Glucagon, 14
Catenins, 189, 190, 192, 195 Gonadotropins. 466
Ce ll synchro niza tio n, 16 Granule exocytosis, 2 10
Cerv ica l ca nce r, 40 7, 408 , 411-4 13 Granu loma, 113, 115--119

569
Growth and spread , 389, 391 Peptide f-meth-Ieu-phe (fMLP), 378--381, 383-3 85
Growth factors, 47, 86, 89, 285, 286, 288, 292,349, Perfusion, 56, 58
352 Pericytes, 47, 50, 51, 285, 287, 288, 29 1-294
Phage, 233--240
Heparin, 14, 18 Pharmaceutical, 473, 478, 483, 486
Hepatocyte growth factor, 415, 424 Phorbo1 myristate acetate, 14, 15
Histological, 40 Phospholipase A2, 113, 119
Hormones , 457 Platelet-activating factor, 113, 115, 117
Human, 24 1,242, 245, 247-250,252-254, 256-260 Platelet-derived endothelial cell growth factor, 213,
Hypoxia, 407, 408, 4 13, 462,464,466-468 216,2 19,220
Pleiotroph in, 213, 217, 2 19
ICI 118--551,16,13 PMN activation, 380, 38 1
Irnrnunolocalizat ion, 139 Polymorphonuclear leucocytes (PMN), 203, 210,
Inflarnrnation, 113-11 5, 119, 120,377,382, 384,385 211,377-384
Inhibitor design, 171 Populati on ecology, 407, 412, 413
Insulin like growth factor-I , 16 Postcapillary endothelial cell(s), 299-30 1
Insulin receptor, 12, 16 Prognosis, 65, 74, 390
Insulin, 12, 13-16 Prostagiandin EI, 15
Interleukin -I beta (ß), 120, 421, 422 Protamine, 113, 118
In vitro, 241-248, 250 Protein kinase C, 36, 37
Isosorbide dinitrate, 310-:-3 13 Purified growth factors, 40, 41
Isosorbide monon itrate, 310-313, 318
Rat aorta, 285, 287,288,290-294
Kaposi, 99, 102, 108 Rabbit cornea1assay(s), 39, 41, 42, 302
Rat cornea assay, 39, 40, 42
Lamin in, 159, 161, 164,225, 226, 228, 229 Regulation, 24 1,242, 246, 248, 249, 25 1
Lectin, 76, 80, 82, 85 Regulatory, 478, 483, 485, 487
Leukotrienes, I 13 Ribonuclease family, 167, 175-1 77
Lewis Lung carci noma, 3 12
Scatter factor, 415-425
Macrop hage(s), 121, 124, 128 Scanning electron microscopy, 429, 432, 439
Magnetic resonance imaging, 55, 59 Speroid, 458--466
Malignant progression, 407, 408, 4 10-4 13 Superoxide anion, 205, 210
MAPkinase, 113, 117, 119 Systemic treatment, 42, 43
Matrigel, 157-1 64, 225, 226, 228, 231, 367- 372
MCF-7, 2 13, 2 14-2 17, 219 Therapeutics,474
Medroxyproge sterone, 113, 11 5, 11 8 Thromb in, 14, 18, 225-230
Mesentery, 75, 76, 80-84 Thymidine phosphorylase, 2 13, 2 16, 220, 22 1
Metastasis, 307, 3 13--3 15, 39 1, 392-394 Thymosin beta 4, 157, 159-1 64
Method for measuring, 6 1, 66 Tight j unction, 189, 191, 197- 20 I
Microci rculation, 58 Transfection, 40
Microvascular endothelial cells, 7, 14 Tube formation, 367, 369, 370-374
Microvessel density, 6 1--{j5, 390, 392- 394 Tumor angiogenesis, 6 1, 62, 64, 66, 67, 69, 70, 73,
Midkine, 213, 214, 217, 2 19 74, 299, 301, 302,429
Migration, 157, 159, 161-164 Tumor growth, 307, 308, 312-314,
MMP, 242-245 Tumor necrosis factor, 114, 119,
Model, 241, 244, 250 Tumor tissue, 39
Monocytic cells, 373 Tumor vascular system, 429, 430, 432
Morphometry,438 Tumor vascularity, 389, 407, 413
MRI, 457,45 8,460, 461, 463-466,468 Tunicamycin, 15
mRNA, 137-1 39, 142, 143, 147 Type IV collagen, 226, 228, 229
Muscle stimulation, 138 (alpha 3 chain), 203, 204, 206, 207, 209
Type IV collagen peptides, 377, 382, 39 1
NCI domain of type IV collagen, 377-384
Neovasc ularization, 473, 474, 478, 479, 482, 485 u-PA, 241, 242, 244, 246-251
Nitric oxide, 297-303, 307, 308, 315
synthase,298--303 Vascular casting, 116
Vascular density, 36
Ovarian follicle, 466 Vascular endothelial growth factor, 2 13, 2 14, 219,
Overexpression, 40 221,298--300,302,303
Oxidized lipoprote ins, 367, 368 Vascular tumor, 102-1 06, 108

570
Vessel density, 40, 41 Wound, 465, 466, 468
Vessel length, 41 healing, 121, 124-126, 128,233,239,240
Vessel wall model, 378, 379, 381, 383
Von Wiliebrand's factor, 419 , 421, 422, 424 X-ray crystallography, 174

57 1