Research Article

Rifampicin as an Oral Angiogenesis Inhibitor Targeting

Hepatic Cancers
1 1 3 2
Masayoshi Shichiri, Nozomi Fukai, Yutaka Kono, and Yujiro Tanaka
1
Medical Hospital and 2Department of General Medicine, Tokyo Medical and Dental University; and
3
Biofrontier Partners, Tokyo, Japan

Abstract combination antiangiogenesis protocols can be more effective

Angiogenesis is an important therapeutic target in cancer, and than single-agent therapies (3, 4). Furthermore, the combination
to fully exploit its therapeutic potential, combination chemo- of traditional chemotherapy with antiangiogenesis agents, which
therapeutic/antiangiogenic regimens should be optimized and targets the endothelial cells of the growing tumor vasculature, as

Downloaded from http://aacrjournals.org/cancerres/article-pdf/69/11/4760/2610231/4760.pdf by Charles University user on 06 October 2023

delivered earlier to more patients. Ideally, this could be well as the growing tumor cells themselves, has been shown to
done by a single potent oral agent with established safety. minimize the resistance and adverse effects of chemotherapeutic
Rifampicin, a semisynthetic antibiotic derived from the agents (4). However, successful clinical application of angiogen-
rifamycins, is one of the most commonly used pharmaceutical esis inhibitors, irrespective of monotherapy or combination
compounds worldwide in the treatment of tuberculosis. Here, therapy, requires long-term administration, but most of these
entering clinical trials are costly, large molecular weight proteins
we present the effects of oral rifampicin on human cancer
progression and its antiangiogenic properties, which were requiring parenteral administration (3, 5).
comparable to the angiogenesis inhibitor endostatin. Clini- Rifampicin is a semisynthetic antibiotic derived from the
cally, low-dose p.o. administration of rifampicin to six high- rifamycins, the common structure of which is a naphthohydroqui-
risk patients with hepatitis C virus–related liver cirrhosis none or naphthoquinone chromophore spanned by an aliphatic
resulted in a single occurrence of hepatocellular carcinoma chain. Rifampicin binds to and activates pregnane X receptor,
during the follow-up period of 97.3 F 29.1 (mean F SD) which affects cytochrome P450, glucuronosyltransferases, and
months. Experimentally, rifampicin rapidly and markedly p-glycoprotein activities (6). It is extensively used to treat
down-regulated the expression of a wide spectrum of tuberculosis and for controlling pruritus in patients with chronic
angiogenesis-associated genes in growing human microvas- cholestatic liver disease (7). Our encounter with a patient with
pulmonary tuberculosis combined with multiple hepatocellular
cular endothelial cells, thereby suppressing endothelial cell
proliferation and migration. Rifampicin, at higher concen- carcinomas (HCC), who seemed to show blunted tumor progres-
trations, also directly inhibited the growth of a variety of sion while receiving rifampicin, prompted us to study whether oral
human cancer cells. P.o. administration of rifampicin signif- rifampicin had a significant tumor-suppressive effect. Our data
icantly inhibited in vivo growth and metastases of subcutane- revealed involvement of both antiangiogenic properties and a
ous human cancer xenografts. Thus, the potent antiangiogenic direct tumor cell inhibitory effect of rifampicin in suppressing
properties of oral rifampicin therapy were effective in human cancer progression, and, considering its clinical pharma-
suppressing cancer progression. It provides a promising new cokinetics, oral rifampicin may be especially beneficial in targeting
addition to antiangiogenic strategies for designing human hepatobiliary tumors. Furthermore, we found that rifampicin exerts
cancer therapies. Considering the clinical pharmacokinetics of its antiangiogenic effect through inhibition of the expression of
rifampicin, which enters the enterohepatic circulation and angiogenesis-associated genes in a manner comparable to the
undergoes subsequent hepatic accumulation, it may be action of endostatin, an endogenous angiogenesis inhibitor known
especially beneficial as an antitumor agent targeting hepato- to down-regulate a variety of growth- and angiogenesis-related
genes in a wide range of endothelial lineage cells (8, 9).
biliary tumors. [Cancer Res 2009;69(11):4760–8]

Materials and Methods

Introduction
Clinical follow-up. Six patients were followed up monthly at Tokyo
Cancer progression and metastasis depend on recruitment of
Medical and Dental University Hospital, after a diagnosis of liver cirrhosis
new capillaries from preexisting vessels, a process known as by either liver biopsy and/or a combination of clinical features. During
angiogenesis (1). Because tumors that are unable to elicit monthly visits, they underwent physical examination, urinalysis, and
angiogenesis remain in a dormant state and fail to grow beyond complete blood count, with measurement of serum a-fetoprotein,
a few millimeters in size (2), the multistep process of prothrombin time, activated partial thromboplastin time, and serum
angiogenesis has fostered an intense search for agents that chemistry. All patients received a liver ultrasound, magnetic resonance
might inhibit this process (3). Although tumor growth can be imaging (MRI), and/or computed tomography (CT) at least every 6 mo. The
stunted by a variety of single angiogenesis inhibitors (2), protocol was approved by the Ethical Committee of Tokyo Medical and
Dental University.
In vivo human cancer xenograft experiments. CW-2 human colon
Note: Supplementary data for this article are available at Cancer Research Online cancer xenograft experiments were conducted in accordance with the
(http://cancerres.aacrjournals.org/). Guidelines for Animal Experimentation, and approved by the Animal
Requests for reprints: Masayoshi Shichiri, Tokyo Medical and Dental University Care and Use Committee of Tokyo Medical and Dental University. Male
Medical Hospital, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan. Phone/Fax: 81-3-
5803-4571; E-mail: mshichiri.cme@tmd.ac.jp.
athymic BALB/c Slc-nu mice (Japan SLC) were given free access to
I2009 American Association for Cancer Research. sterile rodent chow and water. CW-2 cells (1.2  106 per animal) were
doi:10.1158/0008-5472.CAN-08-3417 implanted s.c. in the hind flank region in 8-wk-old mice. When tumors

Cancer Res 2009; 69: (11). June 1, 2009 4760 www.aacrjournals.org

 Rifampicin as an Oral Angiogenesis Inhibitor

Table 1. Characteristics of patients receiving long-term, low-dose oral rifampicin

Patient no. Sex Age Platelet Serum Serum ALT a-Fetoprotein HCV Associated Total Rifampicin
(y) (104/mm3) albumin (IU/L; (ng/mL; diseases follow-up therapy
(mg/dL) normal normal period (mo) period (mo)
8–48) 0–10)
Genotype Copy no.
(kEq)

At the start of rifampicin

1 Male 56 5.9 3.2 82 64.0 Ib 1,500 MR due to 209 90

ruptured chordae
tendineae,
hypertension

Downloaded from http://aacrjournals.org/cancerres/article-pdf/69/11/4760/2610231/4760.pdf by Charles University user on 06 October 2023

2 Female 66 16.5 3.8 149 25.2 Ib 20,000 161 82
3 Female 64 13.4 3.8 139 93.3 Ib 1,600 Type 2 diabetes 121 103
4 Female 69 5.3 3.3 149 11.1 Ib 620 Hypertension 149 62
5 Female 60 8.6 4.0 129 3.3 Ib 5,500 187 149
6 Female 57 6.1 3.7 120 14.7 Ib (mutant) 110 Graves diabetes 135 98

Abbreviations: ALT, alanine aminotransferase; MR, mitral regurgitation.

had reached 200 mm3, animals were moved to a cage supplied with implanted into a subcutaneous dorsal air sac formed in a 9- or 10-wk-old
sterile drinking water containing either 0.2 mg/mL rifampicin/0.25% female Crlj:CD1 (ICR) mouse by injecting f8 mL of air. Each group of mice
DMSO or 0.25% DMSO alone. Tumor growth during the treatment period treated with tumor cells was given via a gastric tube with 0, 100, or 200 mg/
was monitored by measuring the tumor mass on the animals using kg/d rifampicin once a day or with 200 mg/kg/d silibinin twice daily as a
vernier calipers thrice a week. Tumor volume was calculated as the positive control for 5 d. Rifampicin was dissolved to 10 or 20 mg/mL using
product of length  (width)2  0.52. The well-established Lewis lung 0.5% carboxymethylcellulose sodium and silibinin to 10 mg/mL (0.9% NaCl,
carcinoma xenograft experiment was done by Shibayagi, Inc. LL/2 cells 3% ethanol, 1% Tween 80, and 6.6 mmol/L NaOH). The nontumor control
(106 per animal) were s.c. injected into the flank of 8-wk-old male group, which received chambers containing PBS in place of Sarcoma 180
athymic BALB/c AnNCrj-nu/nu mice (Japan Charles Liver). Animals were cells, was given the vehicle alone. On day 5, the implanted chambers were
started on 0.8 mg rifampicin, 3-formyl rifamycin SV, rifamycin SV, or removed from the subcutaneous fascia of the treated animals, and then a
double-distilled water (ddH2O) via a gastric tube once daily during black ring (Sanko Kagaku, Hiroshima) with the same inner diameter as the
weekdays when the tumor volume had reached 200 mm3. The tumor Millipore ring was inserted at the same site. The angiogenic response was
volume was calculated thrice a week as described above. Each group assessed under a dissecting microscope by counting newly formed blood
contained seven mice. On termination of the efficacy portion of the vessels >3 mm in length within the area encircled by the black ring. The
experiment, animals were euthanized; liver and lungs were excised, neovessels were morphologically distinguishable from the preexisting
sectioned, and photographed; and the ratio of metastatic to total section background vessels by their tortuous nature.
area was calculated. Tumor volume and metastasis data of treated Cell culture. Rat pulmonary arterial endothelial cells (rPEC) and rat
groups were compared with those of the control group using Wilcoxon’s aortic endothelial cells (rAEC) prepared from the pulmonary artery and
rank sum test. aorta of male Wistar rats, respectively, and diploid rat fibroblasts have been
In vivo transsplenic liver metastasis model. The growth of the A549 described previously (9, 12, 13). Human dermal microvascular endothelial
human lung cancer cells metastasized into liver was evaluated by cells (hMVEC) and human umbilical vein endothelial cells (hUVEC) were
Panapharm Laboratories, essentially as described previously (10). In brief, purchased from Kurabo and Sanko-Junyaku, and human retinal endothelial
1.5  106 cells in 0.05 mL PBS were injected into the medial splenic tip of cells (hREC) from Cell Systems Corporation. Human colon carcinoma cells,
the anesthetized BALB/cA Jcl-nu nu/nu mice (Japan Crea Laboratory) CW-2, human gastric cancer cells, MKN-1, and human liver cancer cells,
through a left lateral flank incision. No significant bleeding or extravasation HepG2, were purchased from RIKEN BioResource Center Cell Bank.
was encountered. After 10 min, the spleen was removed to avoid Endothelial cell migration, proliferation, and apoptosis assays.
intrasplenic tumor growth. Four days after injection, the treated mice were Anchorage-dependent cell migration was assessed using the monolayer-
divided into four groups and were started on 0, 50, or 100 mg/kg/d denuding protocol as described previously (14). Confluent cultures of
rifampicin or 30 mg/kg/d 5-fluorouracil (5-FU) via a gastric tube once daily endothelial cells in collagen-coated dishes were pretreated with or without
during weekdays. Rifampicin was dissolved to 0, 5, or 10 mg/mL using 0.5% rifampicin for 16 h, and then denuded with a surgical blade. Cells were
carboxymethylcellulose sodium solution, and 10 mL/kg/d was administered further incubated in 20% fetal bovine serum (FBS) supplemented with
at 10 a.m. after a 2-h fast. After 3 wk, the animals were anesthetized; blood growth factors, whereas phase-contrast photomicroscopy was done at the
was withdrawn for the measurement of CA15-3; and the animals were indicated times at four distinct sites, and the cell migration rate was
sacrificed to examine macroscopic metastasis as well as micrometastasis measured. DNA synthesis in confluent cultures was negligible during the
after fixation. monolayer wounding experiments, and repopulation within 24 h was
Mouse dorsal air sac assay. Malignant tumor cell–induced angiogenesis almost entirely to the result of cell migration (9). Exponentially proliferating
was assessed by Panapharm Laboratories (Kumamoto, Japan) as described cells plated in 24-well plates were incubated for 24, 48, or 72 h in the
(11). A Millipore chamber prepared by covering both sides with Millipore presence or absence of indicated concentrations of rifampicin, rifamycin SV,
filters (0.45-mm pore size) was filled with 5  106 Sarcoma 180 cells or 3-formyl rifamycin SV. Cells were then released from culture plates by
suspended in 0.15 mL PBS. The tumor cell–containing chamber was trypsinization, and cell numbers counted using a Sysmex CDA-500

www.aacrjournals.org 4761 Cancer Res 2009; 69: (11). June 1, 2009

Cancer Research

Autoanalyzer (Toa Medical Electronics) as described previously (9, 15, 16). Results
Proapoptotic effects were evaluated by adding rifampicin to growing cell
cultures and determining caspase-3 levels as described previously (17). Clinical benefit of low-dose, long-term oral rifampicin in six
Evaluation of antitumor activity using a human cancer cell line patients with hepatitis C virus–related liver cirrhosis. Since
panel coupled with a drug sensitivity database. A human cancer cell line 1994, six patients (one male and five females, 58.2 F 3.8 years at
panel, composed of 39 cell lines combined with a database for evaluating initial visit), confirmed to have HCV-related liver cirrhosis,
the cell growth inhibition profile, originally established according to the volunteered to receive low-dose oral rifampicin therapy (Table 1).
method of the National Cancer Institute (18, 19), was done by Takao Yamori They were at high risk for the development of HCC (24); therefore,
(Cancer Chemotherapy Center, Japanese Foundation for Cancer Research), they received monthly follow-up by a hepatologist (Y.T.) and 4- to
as previously described (20, 21). Antitumor activity was assessed by changes 6-monthly examinations using liver ultrasound, MRI, and/or CT. All
in total cellular protein after 48-h rifampicin treatment using a sulforhod- six patients had HCV genotype Ib with high HCV RNA levels (110–
amine B assay (20, 22).
20,000 kIU/mL), suggesting resistance to IFN-a therapy (25).
Gene inhibition profile determined by real-time quantitative
reverse transcription-PCR. Exponentially growing hMVECs or hUVECs Patient 1 received IFN-a treatment in 1991 and 1992 without a
supplemented with FBS and growth factors in 6-cm dishes (106 cells) were virological or sustained biochemical response. He had mitral
treated with the indicated concentrations of rifampicin, rifamycin SV, and 3- regurgitation as a result of ruptured chordae tendineae but did not

Downloaded from http://aacrjournals.org/cancerres/article-pdf/69/11/4760/2610231/4760.pdf by Charles University user on 06 October 2023

formyl rifamycin SV for the indicated time. Total RNA was extracted using receive cardiac surgery because of liver cirrhosis. He developed
QIAzol (Qiagen), and 1 Ag of RNA was reverse transcribed with the first- HCC in the S4 portion in May 1993 and another in the S8 portion in
strand cDNA synthesis kit (Qiagen). The expression of angiogenesis- November 1993, both of which were treated with transarterial
associated genes was quantified by a LightCycler (Roche Molecular embolization with associated chemotherapy. His serum a-fetopro-
Biochemicals)-based, real-time, quantitative reverse transcription-PCR tein, a commonly used tumor marker for HCC, started to increase,
(RT-PCR) protocol using Syber Green as described (9, 23). Quantification reaching 66 ng/mL in May 1994. He started treatment with
of c-myc, integrin-av, ID1, vascular endothelial growth factor (VEGF), and
rifampicin 300 mg daily, and within 4 months, his a-fetoprotein
hepatocyte growth factor was done by TaqMan fluorescence methods using
a Universal Probe Library (Roche Molecular Biochemicals) and Chromo 4 was reduced to 15 ng/mL. Because this decrease of a-fetoprotein
Four-Color Real-Time Detection System (Bio-Rad Laboratories). The did not necessarily indicate that rifampicin caused regression of
fluorescence data were quantitatively analyzed by using serial dilution of undetectably small HCC, his physician (Y.T.) hesitated to let him
control samples included in each reaction to produce a standard curve. continue with the antibiotic. They agreed to discontinue rifampicin
We have also quantified glyceraldehyde-3-phosphate dehydrogenase and in September 1995 when liver ultrasonography did not detect any
h-actin mRNA levels to confirm technical validity. abnormal mass. Three months later, however, HCC of f2.5-cm

Figure 1. Serial changes of serum

a-fetoprotein levels, HCC episodes, IFN-a
therapy, and duration of low-dose oral
rifampicin therapy in three patients with
HCV-related cirrhosis who showed
elevated serum a-fetoprotein levels.
Patient 1 (A ), patient 2 (B), and patient 3
(C ) received rifampicin, 150 or 300 mg
daily, during the period indicated RFP,
and serum a-fetoprotein levels were
determined monthly. IFN, conventional IFN
therapy; TAE, an episode of HCC treated
by thromboarterial embolization; TAE &
PEIT, an episode of HCC treated by
thromboarterial embolization and
percutaneous ethanol infusion therapy.

Cancer Res 2009; 69: (11). June 1, 2009 4762 www.aacrjournals.org

 Rifampicin as an Oral Angiogenesis Inhibitor

Figure 2. Inhibition of human cancer

progression by p.o. administration of
rifampicin. A, male immunodeficient mice
were implanted with CW-2 human colon

Downloaded from http://aacrjournals.org/cancerres/article-pdf/69/11/4760/2610231/4760.pdf by Charles University user on 06 October 2023

cells and oral rifampicin was begun when
tumors reached f200 mm3 (x). The tumor
volumes were compared with control mice
.
not treated with rifampicin ( ). Points,
mean for seven mice; bars, SE. B, male
immunodeficient mice were implanted with
LL/2 Lewis lung carcinoma cells, and oral
rifampicin (x), 3-formyl rifamycin SV (n),
.
rifamycin SV (E; 40 mg/kg/d), or water ( )
was administered via a gastric tube when
tumors reached f200 mm3. Points, mean
for seven mice; bars, SE. C and D, at the
end of LL/2 xenograft tumor studies, mice
were sacrificed, liver and lungs were
excised and sectioned, and the ratio of
metastatic to total section area was
calculated. Columns, mean for eight mice;
bars, SE. *, P < 0.05; **, P < 0.01;
***, P < 0.001.

diameter developed in the S8 portion, which was treated with (Fig. 1B). Patient 3 was referred to Y.T. in February 1998. IFN-a
transarterial embolization and percutaneous ethanol injection. therapy reduced her a-fetoprotein but was discontinued because of
Since then, he had just had rifampicin therapy, initially 300 mg and, adverse reactions. In August 1999, her a-fetoprotein increased to
after September 1996, 150 mg daily. A mass of 1.5-cm diameter was 129.2 ng/mL and she was started on rifampicin 150 mg daily, which
detected on the liver surface S4 area by MRI in July 1999, for which reduced her a-fetoprotein level thereafter (Fig. 1C). Patients 4 to 6
he received open liver biopsy followed by local microwave ablation started oral rifampicin 150 mg daily in the slim hope that it might
in October 1999. Resected liver specimens were histologically retard the development of HCC. Patients 5 and 6 received IFN-a
negative for HCC. Despite the very high expected recurrence rate before commencing rifampicin therapy, but soon discontinued
(26) and frequent episodes of hepatic encephalopathy, he was because of adverse effects. In June 2006, 10 years after the start of
totally free from HCC until he died suddenly in January 2002 rifampicin, patient 5 showed clear-cut evidence of HCCs in the S1
(Fig. 1A). Patient 2 has been followed since July 1992. Her a- and S8 portions, which were treated with transarterial embolization
fetoprotein level was normal until May 1998 when it started to followed by percutaneous radiofrequency catheter ablation.
increase sharply. She was prescribed oral rifampicin, 150 mg daily, Surprisingly, as of March 2008, this is the only HCC that developed
in October 1998, but her a-fetoprotein continued to increase, in the six patients during the entire rifampicin therapy period of
reaching a peak value of 113 ng/mL in August 1999. This increase 97.3 F 29.1 (mean F SD) months. Rifampicin induced neither an
led her physician (Y.T.) to increase the rifampicin dose to 300 mg adverse reaction nor HCV RNA negativity.
daily, which reduced her a-fetoprotein level to 26.4 ng/mL in Inhibition of in vivo cancer progression by oral rifampicin.
March 2000. Her a-fetoprotein never showed a substantial increase To explore the reasons for such an unexpectedly low occurrence of
thereafter, even after rifampicin was decreased to 150 mg daily HCC, which followed long-term, oral rifampicin therapy in the

www.aacrjournals.org 4763 Cancer Res 2009; 69: (11). June 1, 2009

Cancer Research

high-risk liver cirrhosis patients, we first studied the effect of oral cells. Rifampicin suppressed FBS-stimulated proliferation of
rifampicin on cancer progression in well-established mouse exponentially growing hMVECs (Supplementary Fig. S1A) and
xenograft models implanted s.c. with human colon cancer (CW-2) hRECs (Supplementary Fig. S1B). The effects were concentration
cells. BALB/c Slc-nu mice receiving rifampicin in their drinking and time dependent and also observed to a lesser degree in rPECs
water ingested 0.68 F 0.01 mg rifampicin/d (34.0 F 0.9 mg/kg/d) (Supplementary Fig. S1C). 3-Formyl rifamycin SV also exerted an
and showed a significantly reduced rate of tumor growth compared antiproliferative effect on hMVECs, hRECs, and rPECs (Supple-
with those receiving control 0.25% DMSO/ddH2O (Fig. 2A). Mice mentary Fig. S1D–F). However, rifamycin SV showed significant
receiving rifampicin for 4 weeks showed a mean tumor growth cytotoxic effects on the three types of endothelial cells (Supple-
volume 48% less than the control group (P < 0.05). We next mentary Fig. S1G–I). The inhibitory effects of rifampicin on VEGF-
implanted Lewis lung carcinoma (LL/2) cells subcutaneously in induced proliferation in quiescent hUVECs were also concentration
BALB/c AnNCrj-nu/nu mice (27). The mice that were orally given dependent (25–100 Ag/mL), but not as potent as in growing
either rifampicin, rifamycin SV, or 3-formyl rifamycin SV via a endothelial cells (IC50 f60 Ag/mL).
gastric tube (40 mg/kg/d) showed a significantly slower tumor We determined the antiproliferative effect of rifampicin on
growth rate (P < 0.05f0.01; Fig. 2B) and reduced metastatic areas human cancer-derived cells and compared their effects with

Downloaded from http://aacrjournals.org/cancerres/article-pdf/69/11/4760/2610231/4760.pdf by Charles University user on 06 October 2023

in the liver and lung in comparison with those receiving control endothelial cells. Rifampicin suppressed the proliferation of CW-2
0.25% DMSO/ddH2O (P < 0.05f0.01; Fig. 2C and D). We next and HepG2, but the effects were less potent than on endothelial
injected a human non–small cell lung cancer cell line (A549) into cells (Supplementary Fig. S2). To evaluate the antiproliferative
the spleen of BALB/cA Jcl-nu nu/nu mice and, 4 days after activity of rifampicin on human cancer-derived cells, a panel of 39
inoculation, started giving rifampicin via a gastric tube for 3 weeks. human cancer cell lines combined with a database was used (20).
Oral rifampicin reduced the degree of metastasis in hepatic This panel was made up of five breast cancer, six glioma, five
sections as observed photomicroscopically (data not shown) and colorectal cancer, seven lung cancer, one melanoma, five ovarian
dose-dependently inhibited the increase in serum CA15-3 concen- cancer, two renal cancer, six gastric cancer, and prostate cancer cell
tration, a specific tumor marker produced by metastasized A549 lines. Rifampicin at 10 4 mol/L (82 Ag/mL) significantly inhibited
cells (Fig. 3), indicating reduced growth of the hepatic metastatic the growth of the cancer cells (Supplementary Fig. S3). The
tumor. The effect of oral rifampicin was comparable to or exceeded sensitivity of each type of cancer cell to rifampicin, as assessed
that of 30 mg/kg/d 5-FU. These results showed that oral rifampicin by the concentration required for 50% growth inhibition (GI50),
potently suppresses the growth of both inoculated and metasta- was higher than 4.8  10 5 mol/L; the average GI50 value was
sized cancer independently of mouse strain and tumor origin, and
that two rifamycins also have a comparable suppressive effect on
tumor progression.
In vivo antiangiogenic effect of oral rifampicin. Sarcoma 180
cells induced a robust angiogenic response in the otherwise poorly
vascularized subcutaneous area when inserted in a Millipore
diffusion chamber and implanted in a dorsal air sac of Crlj:CD1
mice (28). Rifampicin (100 or 200 mg/kg/d) was given p.o. via a
gastric tube once daily for 5 days starting after implantation of the
Millipore chamber and was compared with those receiving silibinin
(200 mg/kg/d), a potent angiogenesis inhibitor (29). Control
animals were implanted with a Millipore chamber with or without
Sarcoma 180 cells and received ddH2O. At the end of the treatment,
animals were sacrificed, the chambers removed, and vasculariza-
tion in the skin in contact with the chamber was observed
microscopically. Oral rifampicin dose-dependently inhibited newly
formed blood microvessels without inducing any toxic effects
(Fig. 4). Administration of 200 mg/kg/d oral rifampicin resulted in
a 63.5% decrease in the number of neovessels (P < 0.0001) and
exceeded the effect of silibinin. To highlight the involvement of
angiogenesis in in vivo rifampicin-induced tumor suppression, its
direct antitumor effect was also assessed. CDF1 mice were
inoculated i.p. with 106 lymphocytic lymphoma cells (P388), and
100, 200, or 400 mg/kg/d rifampicin was administered i.p. on days 1
and 5. Rifampicin did not reduce the median survival of P388-
inoculated mice, suggesting no antitumor effect in a model where
angiogenesis is not involved. Taken together, the direct antitumor
effect of rifampicin can hardly explain the significant inhibition of
xenograft tumor progression in immunodeficient mice, but its
Figure 3. Inhibition of human lung cancer metastasis by p.o. administration
antiangiogenic effect can reasonably account for this inhibition. of rifampicin. A549 human non–small cell lung cancer cells were injected into
Effects of rifampicin on cell proliferation, migration, and the spleen of male immunodeficient mice and the indicated concentrations
apoptosis. To determine whether rifampicin and rifamycins of rifampicin or 5-FU were administered via a gastric tube for 3 wk. Hepatic
metastasis was evaluated by measuring the plasma level of CA15-3, a
induce in vitro antiangiogenic activities, we tested the antiprolifer- specific tumor marker produced by A549. Columns, mean for eight mice;
ative activity of rifampicin and rifamycins in vascular endothelial bars, SE. *, P < 0.05.

Cancer Res 2009; 69: (11). June 1, 2009 4764 www.aacrjournals.org

 Rifampicin as an Oral Angiogenesis Inhibitor

migration of hRECs (Supplementary Fig. S4C). However, rifampicin

did not acutely block VEGF-induced migration of hUVECs detected
by the Boyden chamber method. We next determined whether
rifampicin has significant proapoptotic activity. Rifampicin 10 to
40 Ag/mL did not induce apoptosis in rPECs, rAECs, hMVECs,
hRECs, or human cancer cells (CW-2, MKN-1, HepG2, U937) when
cultured under serum-supplemented conditions. The results
showed that rifampicin inhibits anchorage-dependent endothelial
cell migration without inducing any appreciable proapoptotic
activity in serum-supplemented conditions. Although these fea-
tures are characteristic of endostatin, the effects of 40 Ag/mL
rifampicin on endothelial proliferation seem to be more potent
than 10 6 mol/L endostatin (8).
We explored whether rifampicin possesses other features

Downloaded from http://aacrjournals.org/cancerres/article-pdf/69/11/4760/2610231/4760.pdf by Charles University user on 06 October 2023

characteristic of antiangiogenesis and anticancer reagents. We
measured cell-surface ATP synthase activity by quantifying the
nucleotides generated in culture media using the CellTiterGlo
luminescence assay (30) because potent angiogenesis inhibitors,
such as angiostatin, reduce ATP production by binding to the cell-
surface ATP synthase h-subunit (31). Rifampicin (10–40 Ag/mL) did
not suppress ATP production of rAECs incubated at 17% CO2 and 5%
CO2, suggesting no effect on ATP synthase activity. The inhibitory
effect of rifampicin on histone deacetylase-1 enzymatic activity was
measured by quantifying 7-amino-4-methyl coumarine after addition
of an acetylated fluorescent peptide to recombinant human histone
deacetylase 1 as a substrate. Rifampicin suppressed histone
deacetylase-1 activity only at very high concentrations (IC50 375
Ag/mL), showing its negligible effect. Rifampicin (0.1–100 Ag/mL) did
not suppress the infiltration of cultured HT1080 human fibrous
sarcoma cells in a Matrigel-based invasion assay (32). Rifampicin at
10 5 mol/L (8.2 Ag/mL) did not modify nerve growth factor–induced
neurite outgrowth of PC12 pheochromocytoma cells, the phenotype
associated with colchicine- and Taxol-induced microtubule assem-
bly. Rifampicin (1–100 Ag/mL) did not affect the activities of protein
kinase A, protein kinase C, epidermal growth factor receptor kinase,
or VEGF receptor (Flt-1) kinase (33), whereas it caused 50% to 80%
suppression of protein tyrosine kinase, calmodulin-dependent
protein kinase III (34), only at 100 Ag/mL. Taken together, these
results reveal that rifampicin shows some features characteristic of
other anticancer reagents but only at high concentrations in the
hepatobiliary system.
Down-regulation of angiogenesis-associated genes by
rifampicin. We next determined whether rifampicin rapidly
Figure 4. Inhibitory effect of rifampicin on angiogenic response by human down-regulated a variety of growth-, migration-, and angiogene-
sarcoma. Sarcoma 180 cells (A431) were implanted into the mouse dorsal air sis-associated genes in the endothelium because this phenomenon
sac and the indicated doses of rifampicin or silibinin were given via a gastric tube
for 5 d. A, tumor-induced angiogenesis in the skin was evaluated by counting
is induced by a potent angiogenesis inhibitor, endostatin (8, 9).
the number of neovessels microscopically. Columns, mean for eight mice; bars, Using a real-time quantitative RT-PCR, we quantified the mRNA
SE. *, P < 0.05; ***, P < 0.0001. B, morphologic observation of the effect of levels of angiogenic factors (VEGF and hepatocyte growth factor),
rifampicin on A431 cell–induced angiogenesis. The area circled with a black ring
corresponds to the exact portion of the air sac fascia in contact with a Millipore ID1, VEGF receptors (KDR and Flt-1), platelet/endothelial cell
chamber containing A431 cells. In comparison with nontumor control (i), adhesion molecule 1, focal adhesion kinase, endothelin 1,
the air sac fascia of a control mouse was densely capillarized (ii ), but very endothelin receptor type B, integrin-av, integrin-h3, and c-myc in
slightly capillarized in the fascia of a mouse treated with 100 mg/kg/d (iii ) and
200 mg/kg/d (iv ) rifampicin. hMVECs. Rifampicin rapidly down-regulated all of these genes in
a concentration-dependent manner (Fig. 5A); within 4 hours of
rifampicin treatment, mRNA levels of each gene treated with
f10 4 mol/L. Therefore, rifampicin directly suppressed the growth 40 Ag/mL rifampicin decreased to levels comparable to those in
of cancer cells, but the concentrations required for anticancer cell cells deprived of serum and growth supplements for 4 hours. The
proliferation were higher than those to inhibit endothelial cells. ability of rifampicin to down-regulate gene expression is either
Rifampicin and 3-formyl rifamycin SV concentration-dependent- absent or far less in nonendothelial cells, such as rat fibroblasts and
ly (0–40 Ag/mL) suppressed the progression rate of the denuded human cancer cell lines (HepG2, MKN-1, and CW-2). Rifamycin SV
edge of monolayer hMVECs, suggesting a migration inhibitory and 3-formyl rifamycin SV also reduced the mRNA levels of many
effect (Supplementary Fig. S4A and B). They also inhibited the genes, including integrin-av, integrin-h3, and c-myc (Fig. 5B); the

www.aacrjournals.org 4765 Cancer Res 2009; 69: (11). June 1, 2009

Cancer Research

spectrum of suppression was almost identical to that of rifampicin. never achieved sustained HCV negativity; and three patients
Similar results were obtained with rPECs, rAECs, and hUVECs. discontinued IFN due to adverse reactions. Second, they were
These results showed that rifampicin and two rifamycins potently relatively old, at 62.0 F 5.2 years, at the start of rifampicin therapy.
down-regulate proangiogenic gene expression in a variety of Third, they had high serum alanine aminotransferase levels before
endothelial cells, comparable to the effects of endostatin (8). rifampicin therapy (38). Fourth, patients 1 to 3 showed increasing
a-fetoprotein levels when rifampicin was started (39). Patient 1,
who had the highest risk for HCC because of relapsing episodes,
Discussion developed HCC shortly after discontinuation of rifampicin in 1995,
HCV-related liver cirrhosis is a major risk factor for development but HCC never recurred after restarting rifampicin. Because a-
of HCC, with an annual incidence reaching 7.9% in an untreated, fetoprotein can either be produced by HCC itself or by regenerating
biopsy-proven population (24, 35). Even after initial treatment, hepatic tissues after significant hepatic injury, its marked decrease
second HCCs develop at new foci in f30% of patients within the suggests either reduction of HCC, undetectable by ultrasound or
first year, leading to a poor prognosis (26). IFN treatment can MRI, or suppression of hepatic injury and subsequent regeneration.
inhibit HCC in patients with moderate liver fibrosis, but only Whatever the mechanisms, large-scale prospective trials should be

Downloaded from http://aacrjournals.org/cancerres/article-pdf/69/11/4760/2610231/4760.pdf by Charles University user on 06 October 2023

marginally in liver cirrhosis (35–37). Although the present clinical done to establish the overall clinical usefulness of low-dose oral
observation is restricted to only six patients, it should be rifampicin therapy.
emphasized that only a single HCC developed during the entire To gain insight into the mechanism behind the clinical benefit of
rifampicin therapy period of 97.3 F 29.1 months despite many low-dose, long-term rifampicin therapy, we tested its effects on
features indicating a high risk for HCC (24). First, none had any cultured endothelial and nonendothelial cells. Rifampicin and two
apparent benefit from IFN therapy; two did not receive IFN; four rifamycins similarly suppressed a variety of mitogenesis- and

Figure 5. Rifampicin and rifamycins down-regulate growth-, migration-, and angiogenesis-associated genes. hMVECs grown exponentially in 20% FBS and growth
supplements were incubated with the indicated concentrations of rifampicin (A), 3-formyl rifamycin SV (B), or rifamycin SV (B) for 4 h and the indicated mRNA
levels were determined by real-time quantitative RT-PCR. Columns, mean percentage of the level in untreated cells from four repeated experiments; bars, SE.

Cancer Res 2009; 69: (11). June 1, 2009 4766 www.aacrjournals.org

 Rifampicin as an Oral Angiogenesis Inhibitor

angiogenesis-related genes (40) in both human and rat endothelial administration (49). Peak plasma rifampicin after single oral doses
cells; the time course and potency, as well as the spectrum of gene of 150 and 300 mg are as high as 2 and 3.5 Ag/mL, respectively, in
suppression, are comparable to those previously observed with normal subjects, whereas hepatic bile and urine levels easily exceed
endostatin (9). There is general agreement that endostatin inhibits 100 Ag/mL (49). Even at 12 hours after an oral dose of 150 mg,
cultured endothelial cell migration, but it is less clear whether hepatic bile concentrations of rifampicin and its metabolite still
endostatin also affects endothelial cell apoptosis and proliferation remain at 100 Ag/mL in normal subjects. In liver cirrhosis, serum
(2, 9, 41–46). In our in vitro study using the same protocol as and biliary levels are further elevated and mean half-life is
our previous endostatin study (9), the antiproliferative effects of prolonged because its bile excretion reaches saturation, resulting
40 Ag/mL rifampicin and 3-formyl rifamycin SV on microvascular in very constant, high levels (50). Such clinical pharmacokinetics of
endothelial cells were clearly greater than those seen with the rifampicin explain the remarkable effect of low-dose, once daily,
treatment dose of endostatin. An antimigratory effect was also oral rifampicin on suppressing HCC in liver cirrhosis and provide a
evident, but this was not as marked as the antiproliferative effect. reasonable rationale for its use as an antiangiogenic agent targeting
These results show the antiangiogenic properties of rifampicin and the hepatobiliary system. Moreover, our results show that, at
3-formyl rifamycin SV and the potential utility of oral rifampicin as concentrations >100 Ag/mL, rifampicin could exert its direct

Downloaded from http://aacrjournals.org/cancerres/article-pdf/69/11/4760/2610231/4760.pdf by Charles University user on 06 October 2023

an angiogenesis inhibitor. Considering that its inhibition of growth-inhibitory effects on tumor cells, in addition to its
endothelial cell proliferation is more potent than that of endo- antiangiogenic activity, thus allowing an efficient combination of
statin, rifampicin could even be used in combination with chemotherapy and antiangiogenic therapy by a single oral agent.
endostatin, not to mention with other types of angiogenesis In summary, we have shown the effectiveness of oral rifampicin
inhibitor, or with chemotherapeutic agents. Rifampicin also in both clinical and experimental settings. Low-dose, long-term
showed direct growth inhibition of a variety of cancer cells, but oral rifampicin inhibited the development of HCC in HCV-related
only at high concentrations corresponding to bile and urine levels liver cirrhosis patients, whereas a comparatively larger dose of oral
after oral ingestion. The cell growth inhibition profile (20, 22) was rifampicin significantly suppressed the growth and metastases of
similar to that of IFN-a, TZT-1027, and vincristine. These results human cancer xenografts in mouse models. In addition, rifampicin
show that, although rifampicin and 3-formyl rifamycin SV markedly exerted antiangiogenic properties. These include down-regulation
inhibit the proliferation of endothelial cells, the growth-inhibiting of proangiogenic genes and inhibition of microvascular endothelial
effects of rifampicin on cancer cells are less marked, suggesting cell proliferation. Rifampicin at higher doses showed direct growth
that rifampicin has a more potent effect through antiangiogenesis inhibition of cancer cells. Considering its clinical pharmacokinetic
rather than a direct cancer-suppressive effect. profile, oral rifampicin could exert its direct tumor-suppressive
For in vivo experiments, we used well-established xenograft effect in the hepatobiliary system and, together with its potent
tumor models in mice to test the effectiveness of rifampicin as an antiangiogenic properties, may be especially beneficial in targeting
inhibitor of angiogenesis-dependent tumor growth. The growth of hepatobiliary tumors.
CW-2 colon primary tumors was significantly suppressed by oral
rifampicin. When rifampicin or rifamycins were administered to Disclosure of Potential Conflicts of Interest
Lewis lung carcinoma xenograft mice, lung and liver metastases, as
M. Shichiri and Y. Tanaka are co-inventors of a patent applied for by the Japan
well as tumor growth, were inhibited to a similar extent. Although Science and Technology Agency, and M. Shichiri and Y. Kono are co-inventors of a
the doses of rifampicin and rifamycins used in animal studies were patent applied for by the Tokyo Medical and Dental University. These patents are
relevant to this work and are therefore declared as competing financial interests.
>10-fold higher than the doses given to patients, the estimated N. Fukai disclosed no potential conflicts of interest.
plasma levels in these animals must be far lower than those
showing inhibitory effects on cancer cells. Furthermore, we also
performed the mouse dorsal air sac assay, which is well
Acknowledgments
documented to represent tumor-induced angiogenesis and its Received 9/16/08; revised 2/17/09; accepted 3/24/09; published OnlineFirst 5/19/09.
Grant support: Grants-in-Aid A and B for Scientific Research (M. Shichiri) and
inhibition in mouse xenografts (47, 48). Thus, significant inhibition Grants-in-Aid for Scientific Research on Priority Area (M. Shichiri) from the Ministry
of xenograft tumor progression in immunodeficient mice can of Education, Culture, Sports, Science and Technology of Japan.
barely be explained by the direct antitumor effect of rifampicin, but The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
can reasonably be accounted for by its antiangiogenic effect. with 18 U.S.C. Section 1734 solely to indicate this fact.
Despite the potential use of oral rifampicin as a general We thank Shinobu H. Yamaguchi for her technical assistance, and Chifumi Sato
M.D. for suggestions. The Screening Committee of New Anticancer Agents 2004,
angiogenesis inhibitor, its biliary excretion and enterohepatic supported by Grant-in-Aid for Scientific Research on Priority Area ‘‘Cancer’’ from the
circulation may be especially beneficial in accumulating this agent Ministry of Education, Culture, Sports, Science and Technology of Japan (chaired by
and its metabolites in the hepatobiliary system. Biliary and hepatic Takao Yamori, Ph.D.), determined tumor cell growth using a cancer cell line panel,
P388 lymphoma progression, VEGF-induced hUVEC migration, histone deacetylase 1
concentrations of rifampicin and its metabolites are known to be enzymatic activity, Matrigel-based invasion using HT1080, nerve growth factor–
far higher and more sustained than plasma levels after p.o. induced neurite outgrowth using PC12, and protein kinase inhibitory activity.

References 4. Kerbel RS, Kamen BA. The anti-angiogenic basis of signaling pathway for CYP3A induction. Proc Natl Acad
metronomic chemotherapy. Nat Rev Cancer 2004;4: Sci U S A 1998;95:12208–13.
1. Carmeliet P. Angiogenesis in health and disease. Nat 423–36. 7. Bachs L, Pares A, Elena M, Piera C, Rodes J. Comparison
Med 2003;9:653–60. 5. Folkman J. Antiangiogenesis in cancer therapy-endo- of rifampicin with phenobarbitone for treatment of
2. Kerbel R, Folkman J. Clinical translation of angiogen- statin and its mechanisms of action. Exp Cell Res 2006; pruritus in biliary cirrhosis. Lancet 1989;1:574–6.
esis inhibitors. Nat Rev Cancer 2002;2:727–39. 312:594–607. 8. Abdollahi A, Hahnfeldt P, Maercker C, et al. Endo-
3. Carmeliet P. Angiogenesis in life, disease and medi- 6. Bertilsson G, Heidrich J, Svensson K, et al. Identifica- statin’s antiangiogenic signaling network. Mol Cell 2004;
cine. Nature 2005;438:932–6. tion of a human nuclear receptor defines a new 13:649–63.

www.aacrjournals.org 4767 Cancer Res 2009; 69: (11). June 1, 2009

Cancer Research

9. Shichiri M, Hirata Y. Antiangiogenesis signals by 23. Shichiri M, Yoshinaga K, Hisatomi H, Sugihara K, 36. Niederau C, Lange S, Heintges T, et al. Prognosis of
endostatin. FASEB J 2001;15:1044–53. Hirata Y. Genetic and epigenetic inactivation of chronic hepatitis C: results of a large, prospective cohort
10. Kozlowski JM, Fidler IJ, Campbell D, Xu ZL, Kaighn mitotic checkpoint genes hBUB1 and hBUBR1 and study. Hepatology 1998;28:1687–95.
ME, Hart IR. Metastatic behavior of human tumor cell their relationship to survival. Cancer Res 2002;62: 37. Valla DC, Chevallier M, Marcellin P, et al. Treatment
lines grown in the nude mouse. Cancer Res 1984;44: 13–7. of hepatitis C virus-related cirrhosis: a randomized,
3522–9. 24. Yoshida H, Shiratori Y, Moriyama M, et al. controlled trial of interferon a-2b versus no treatment.
11. Oikawa T, Sasaki M, Inose M, et al. Effects of Interferon therapy reduces the risk for hepatocellular Hepatology 1999;29:1870–5.
cytogenin, a novel microbial product, on embryonic and carcinoma: national surveillance program of cirrhotic 38. Tarao K, Rino Y, Ohkawa S, et al. Association between
tumor cell-induced angiogenic responses in vivo . and noncirrhotic patients with chronic hepatitis C in high serum alanine aminotransferase levels and more
Anticancer Res 1997;17:1881–6. Japan. IHIT Study Group. Inhibition of hepatocarcino- rapid development and higher rate of incidence of
12. Prouty SM, Hanson KD, Boyle AL, et al. A cell culture genesis by interferon therapy. Ann Intern Med 1999; hepatocellular carcinoma in patients with hepatitis C
model system for genetic analyses of the cell cycle by 131:174–81. virus-associated cirrhosis. Cancer 1999;86:589–95.
targeted homologous recombination. Oncogene 1993;8: 25. Hayashi N, Takehara T. Antiviral therapy for chronic 39. Gebo KA, Chander G, Jenckes MW, et al. Screening
899–907. hepatitis C: past, present, and future. J Gastroenterol tests for hepatocellular carcinoma in patients with
13. Shichiri M, Ishimaru S, Ota T, Nishikawa T, Isogai T, 2006;41:17–27. chronic hepatitis C: a systematic review. Hepatology
Hirata Y. Salusins: newly identified bioactive peptides 26. Imamura H, Matsuyama Y, Tanaka E, et al. Risk 2002;36:S84–92.
with hemodynamic and mitogenic activities. Nat Med factors contributing to early and late phase intrahepatic 40. Brooks PC, Clark RA, Cheresh DA. Requirement of
2003;9:1166–72. recurrence of hepatocellular carcinoma after hepatec- vascular integrin avh3 for angiogenesis. Science 1994;
14. Fukai N, Shichiri M, Ozawa N, Matsushita M, Hirata tomy. J Hepatol 2003;38:200–7. 264:569–71.
Y. Coexpression of calcitonin receptor-like receptor and 27. O’Reilly MS, Boehm T, Shing Y, et al. Endostatin: an 41. Blezinger P, Wang J, Gondo M, et al. Systemic

Downloaded from http://aacrjournals.org/cancerres/article-pdf/69/11/4760/2610231/4760.pdf by Charles University user on 06 October 2023

receptor activity-modifying protein 2 or 3 mediates the endogenous inhibitor of angiogenesis and tumor inhibition of tumor growth and tumor metastases by
antimigratory effect of adrenomedullin. Endocrinology growth. Cell 1997;88:277–85. intramuscular administration of the endostatin gene.
2003;144:447–53. 28. Nakamura M, Katsuki Y, Shibutani Y, Oikawa T. Nat Biotechnol 1999;17:343–8.
15. Shichiri M, Fukai N, Ozawa N, Iwasaki H, Hirata Y. Dienogest, a synthetic steroid, suppresses both embry- 42. Dhanabal M, Ramchandran R, Waterman M, et al.
Adrenomedullin is an autocrine/paracrine growth factor onic and tumor-cell-induced angiogenesis. Eur J Phar- Endostatin induces endothelial cell apoptosis. J Biol
for rat vascular smooth muscle cells. Regul Pept 2003; macol 1999;386:33–40. Chem 1999;274:11721–6.
112:167–73. 29. Singh RP, Deep G, Chittezhath M, et al. Effect of 43. Sasaki T, Larsson H, Kreuger J, et al. Structural basis
16. Shichiri M, Kato H, Marumo F, Hirata Y. Endothelin-1 silibinin on the growth and progression of primary and potential role of heparin/heparan sulfate binding to
as an autocrine/paracrine apoptosis survival factor for lung tumors in mice. J Natl Cancer Inst 2006;98: the angiogenesis inhibitor endostatin. EMBO J 1999;18:
endothelial cells. Hypertension 1997;30:1198–203. 846–55. 6240–8.
17. Shichiri M, Yokokura M, Marumo F, Hirata Y. 30. Burwick NR, Wahl ML, Fang J, et al. An Inhibitor of 44. Yamaguchi N, Anand AB, Lee M, et al. Endostatin
Endothelin-1 inhibits apoptosis of vascular smooth the F1 subunit of ATP synthase (IF1) modulates the inhibits VEGF-induced endothelial cell migration and
muscle cells induced by nitric oxide and serum activity of angiostatin on the endothelial cell surface. tumor growth independently of zinc binding. EMBO J
deprivation via MAP kinase pathway. Arterioscler J Biol Chem 2005;280:1740–5. 1999;18:4414–23.
Thromb Vasc Biol 2000;20:989–97. 31. Moser TL, Stack MS, Asplin I, et al. Angiostatin binds 45. Dixelius J, Cross MJ, Matsumoto T, Claesson-Welsh L.
18. Monks A, Scudiero D, Skehan P, et al. Feasibility of a ATP synthase on the surface of human endothelial cells. Endostatin action and intracellular signaling: h-catenin
high-flux anticancer drug screen using a diverse panel of Proc Natl Acad Sci U S A 1999;96:2811–6. as a potential target? Cancer Lett 2003;19:1–12.
cultured human tumor cell lines. J Natl Cancer Inst 1991; 32. Ueda J, Kajita M, Suenaga N, Fujii K, Seiki M. 46. Skovseth DK, Veuger MJ, Sorensen DR, De Angelis
83:757–66. Sequence-specific silencing of MT1-MMP expression PM, Haraldsen G. Endostatin dramatically inhibits
19. Paull KD, Shoemaker RH, Hodes L, et al. Display and suppresses tumor cell migration and invasion: impor- endothelial cell migration, vascular morphogenesis,
analysis of patterns of differential activity of drugs tance of MT1-MMP as a therapeutic target for invasive and perivascular cell recruitment in vivo . Blood 2005;
against human tumor cell lines: development of mean tumors. Oncogene 2003;22:8716–22. 105:1044–51.
graph and COMPARE algorithm. J Natl Cancer Inst 1989; 33. Murakami Y, Fukazawa H, Mizuno S, Uehara Y. 47. Funahashi Y, Wakabayashi T, Semba T, Sonoda J,
81:1088–92. Conversion of epidermal growth factor (EGF) into a Kitoh K, Yoshimatsu K. Establishment of a quantitative
20. Yamori T, Matsunaga A, Sato S, et al. Potent stimulatory ligand for A431-cell growth by herbimycin A mouse dorsal air sac model and its application to
antitumor activity of MS-247, a novel DNA minor by decreasing the level of expression of EGF receptor. evaluate a new angiogenesis inhibitor. Oncol Res 1999;
groove binder, evaluated by an in vitro and in vivo Biochem J 1994;301:57–62. 11:319–29.
human cancer cell line panel. Cancer Res 1999;59: 34. Fukazawa H, Li PM, Mizuno S, Uehara Y. Method for 48. Tanaka NG, Sakamoto N, Inoue K, et al. Antitumor
4042–9. simultaneous detection of protein kinase A, protein effects of an antiangiogenic polysaccharide from an
21. Yamori T, Sato S, Chikazawa H, Kadota T. Anti-tumor kinase C, protein tyrosine kinase, and calmodulin- Arthrobacter species with or without a steroid. Cancer
efficacy of paclitaxel against human lung cancer dependent protein kinase activities. Anal Biochem Res 1989;49:6727–30.
xenografts. Jpn J Cancer Res 1997;88:1205–10. 1993;212:106–10. 49. Acocella G. Clinical pharmacokinetics of rifampicin.
22. Dan S, Tsunoda T, Kitahara O, et al. An integrated 35. Shiratori Y, Ito Y, Yokosuka O, et al. Antiviral therapy Clin Pharmacokinet 1978;3:108–27.
database of chemosensitivity to 55 anticancer drugs and for cirrhotic hepatitis C: association with reduced 50. McConnell JB, Powell-Jackson PR, Davis M, Williams
gene expression profiles of 39 human cancer cell lines. hepatocellular carcinoma development and improved R. Use of liver function tests as predictors of rifampicin
Cancer Res 2002;62:1139–47. survival. Ann Intern Med 2005;142:105–14. metabolism in cirrhosis. Q J Med 1981;50:77–82.

Cancer Res 2009; 69: (11). June 1, 2009 4768 www.aacrjournals.org