Focus on aggregation: types, causes,

characterization, and impact

John Philo
V.P. & Director of Biophysical Chemistry

2007 Alliance Protein Laboratories; all rights reserved

 Outline

Why do we care about aggregates in

biopharmaceuticals?
Review some basic facts about aggregate
sizes and types
Aggregation mechanisms
Utility of sedimentation velocity for analysis
of long-lived aggregates
A few words about field-flow fractionation
(FFF)
 Time wont permit talking about light
scattering techniques today, but

Background and examples can be found on the APL

web site, www.ap-lab.com
Many articles, talks, and posters on aggregation and
comparability studies can be downloaded from our
Further Reading page
 Protein aggregates: What is all the
fuss about?

Aggregates (both large and small) often are

a major degradation product
Hence they often are a major factor limiting
shelf life
Aggregates in the product may affect its:
1. manufacturability
clogged columns or diafiltration membranes
2. bioactivity (potency)
3. serum half-life or absorption rate
4. immunogenicity
 Why is there heightened concern about
immunogenicity?

In Europe about 200 patients taking Eprex (one

brand of recombinant erythropoietin, EPO) developed
antibodies that cross-reacted and neutralized their
own internally-produced EPO
Consequently those patients made no new red blood
cells and require regular transfusions (Pure Red-Cell
Aplasia)
While the manufacturer has published evidence that
this immunogenicity was not due to aggregates, this
incident has raised alarm bells about immunogenicity
 Some known cases where aggregates cause
immunogenicity

1. Early versions of intravenous immunoglobulin

from donor blood (IVIG) had high aggregate
levels and caused anaphylaxis
similar experience for human serum albumin
2. Aggregate levels in human growth hormone
(hGH) correlated with persistence of anti-hGH
antibody in patient serum
3. A recent study using interferon- and
transgenic mice confirmed that
immunogenicity depends on the type and size
of the aggregates
S. Hermeling et al. (2006) J. Pharm. Sci. 95, 1084-1096.
 AAPS Protein Aggregation and
Immunogenicity Focus Group

Sponsored a workshop September 2006 in Colorado to

summarize current state-of-the-art and remaining
challenges
talks, posters, summaries available at
http://www.aapspharmaceutica.com/inside/Focus_Groups/ProteinAgg/in
dex.asp
Goal is to form an industry consortium to sponsor (pay
for) new studies to better define how immunogenicity
varies with aggregate type and size
Please join us!
 The word aggregate covers a wide spectrum
of types and sizes of associated states

1. rapidly-reversible non-covalent small oligomers

(dimer, trimer, tetramer)
2. irreversible non-covalent oligomers
3. covalent oligomers (e.g. disulfides)
4. large aggregates (> 10-mer)
+ could be reversible if non-covalent
5. very large aggregates (diameter ~50 nm to 3
m)
+ could be reversible if non-covalent
soluble
6. visible particulates
+ probably irreversible insoluble
Reversible vs. irreversible aggregates

reversible irreversible
 Whether aggregates are irreversible or
reversible depends on the context

solvent components
salts, sugars, other excipients
organic modifiers (alcohols, acetonitrile)

pH
temperature
how long you wait
 Aggregates have a spectrum of lifetimes

rates of non-covalent association and dissociation (half-

times) can vary from milliseconds to days
metastable oligomers with dissociation rates of hours to
days occur fairly frequently
+ for an antibody example see J.M.R. Moore et al. (1999)
Biochemistry 38: 13960-13967
+ see also Philo, J.S. (2006) AAPS Journal 8(3): E564-E571
many common analytical methods will detect only the
longer-lived species
it may take hours to days for a protein to re-equilibrate
its association after a change in concentration, solvent
conditions or temperature
 Aggregation mechanisms (1): reversible
association of native protein

or

Native Reversible
protein oligomerization

Higher oligomers
+ sucrose (possibly
irreversible)
 Aggregation mechanisms (2): oligomerization
following conformational change

Conformational Oligomerization
Native
change or partial of non-native
protein
unfolding protein

Higher oligomers
+ sucrose (probably
irreversible)
 Aggregation mechanisms (3): oligomerization
driven by covalent modification

Modified
protein Oligomerization
Native
(oxidation, of modified
protein
deamidation, protein
etc.)
Higher oligomers
+ sucrose (possibly
irreversible)
 Aggregation mechanisms (4): nucleation
controlled aggregation (seeding)

Critical nucleus
Addition of protein
(aggregate of
monomers onto
Native native or
surface of nucleus
protein modified
(often with partial
protein, or a Visible
unfolding)
contaminant) particulates or
precipitation
+ sucrose
 Aggregation mechanisms (5): surface-induced
aggregation

Adsorption of
Container
Native protein monomers Aggregation of
surfaces and
protein onto surfaces altered protein (as
air-liquid
promotes partial in mechanism 2)
interfaces
unfolding
+ sucrose
+ detergent
 Our analytical challenge

1. Any protein sample may contain aggregates with a

wide range of sizes, types, and lifetimes
2. Any one analysis method may not detect all the
aggregate sizes or types that are present
3. The measurement itself may perturb the aggregate
distribution that was initially present
 The measurement itself may create or destroy
aggregates

dissociation or loss of aggregates can be caused by: SEC SV FFF

dilution +++ + +++

change of solvent conditions +++ - ++

adsorption to surfaces +++ + ++

physical filtration (e.g. column frit) +++ - -

physical disruption (e.g. shear forces) ++ - -

creation of new aggregates can be caused by:

change of solvent conditions +++ - ++

surface or shear-induced denaturation ++ - +

concentration on surface - - ++
 Regulatory concerns about analytical methods
for aggregates

Although SEC is usually the primary method of aggregate

analysis the regulatory agencies are well aware that SEC
columns can act as filters and that the SEC mobile phase
can change the distribution of non-covalent aggregates
but SEC is often the only qualified method that can be validated for
lot release
Therefore by phase 3 (and sometimes earlier) they will now
nearly always ask for cross-validation of SEC methods by
orthogonal methods
 Methods typically used to cross-check SEC

analytical ultracentrifugation (AUC)

sedimentation velocity (primarily)
sedimentation equilibrium (occasionally)
light scattering
flow mode classical scattering used after SEC (SEC-
MALLS) has been validated
dynamic light scattering (DLS)
batch mode classical scattering has been validated
field-flow fractionation (FFF)
usually used with MALLS to measure true MW
Sedimentation velocity
 The fundamentals of sedimentation velocity

1.5
1.4
1.3 meniscus cell base
1.2 The sedimentation coefficient
1.1 is determined from the
1.0 centrifugal diffusion boundary motion over time. It
0.9 force depends on both molecular
friction
Absorbance

0.8
weight and molecular shape.
1.6

0.7
region of solute

1.4

0.6 1.2

0.5 1.0
depletion

Absorbance

0.8
0.4
0.6
boundary

0.3
0.4
0.2
0.2

0.1 0.0

0.0 6.0 6.1 6.2 6.3 6.4 6.5 6.6

Radius (cm)
6.7 6.8 6.9 7.0

-0.1
6.0 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.0
Radius (cm)
High resolution analysis of a highly stressed antibody
sample resolves 6 aggregate peaks plus 2 fragments

1.0 main peak (monomer), 45.5%

c(s), normalized (total area = 1)

0.8

dimer 30.6%
0.6
? HL half molecule, 0.8%
? free light chain, 1.4%

trimer 14.6%
0.4

heptamer, 0.1%
tetramer 5.3%

hexamer, 0.4%
pentamer 1.4%
0.2

0.0
0 2 4 6 8 10 12 14 16 18 20 22 24

sedimentation coefficient (Svedbergs)

 This interferon- sample is 13.7% non-covalent
aggregate; by the standard SEC method it would
be pure monomer
7
IFN- in 5 mM glycine, pH 3, 86.3% main peak
6 0.35

0.30
5
0.25

4 0.20
c(s)

0.15
3

0.10
2
0.05

1 0.00
0 2 4 6 8 10

0
0 2 4 6 8 10

sedimentation coefficient (Svedbergs)

 Adding NaCl to interferon- formulations leads to a broad
distribution of non-covalent aggregates out to ~100-mers

3
0.15 no salt
20X expanded
2 0.10
c(s)

0.05
1
0.00
0 2 4 6 8 10
0
1.5

0.15 +50 mM NaCl

1.0 20X expanded
0.10
c(s)

0.05
0.5
0.00
0 2 4 6 8 10
0.0
0.07
0.06
+150 mM NaCl
0.05
0.04
c(s)

0.03
0.02
0.01
0.00

0 8 16 24 32 40
sedimentation coefficient (Svedbergs)
 SV is very useful for comparability studies, giving
comparability of conformation as well as aggregation

2
0.02
X 100

lot 1
lot 2
normalized c(s)

0.95%
0.01
1 0.42%
0.05% 0.03%
0.30%
0.07%

0.10%
0.00
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30

0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
sedimentation coefficient (Svedbergs)
 The peril: c(s) distributions are often misunderstood

1. the effective resolution goes down as the fraction of

minor peaks goes down
2. the resolution you can achieve for a 150 kDa antibody
is much greater than for a 20 kDa cytokine
3. in general it is not possible to uniquely assign a
stoichiometry to each aggregate peak
4. the nature of the noise (variability) is very different
than in chromatography
5. for reversibly associating proteins the peaks probably
do not represent individual molecular species
Field-flow fractionation (FFF)
Principles of cross-flow FFF

figure courtesy Wyatt Technology

 FFF of acid-exposed IgG (2 hr at pH 2.9, 5 C)
(courtesy K. Tsumoto and D. Ejima)
Chromatograms 070508-120min_01
0.10

0.08

0.06

LS, AUX (volts)

FFF using 0.1 M 0.04

citrate, pH 2.9
0.02

0.00

-0.02
0 5 10 15 20
Time (min)
Chromatograms CF15 XF23 FF20-3_01
0.05

0.04

FFF after titration 0.03

to neutral pH, elute

LS, AUX (volts)

using 0.1 M 0.02

phosphate, pH 6.8 0.01

0.00

-0.01
0 5 10 15 20
Time (min)
 Advantages & drawbacks of FFF

main advantages
1. much less surface area for absorption of sticky
aggregates than SEC columns
2. can separate a much wider range of aggregate
sizes than SEC
drawbacks
1. some proteins stick to all the available membranes
2. many parameters need to be optimized during
method development
3. high dilution may dissociate reversible aggregates
 Summary

1. Aggregation is a complex phenomenon!

2. No single analytical method is optimal for all types and
sizes of aggregates
3. Sedimentation velocity has many advantageous
properties
it is the primary tool we use at APL to cross-check SEC methods
(and help improve them)
it suffers from low throughput and requires a very highly trained
operator
4. Our ability to characterize aggregates unfortunately far
exceeds our knowledge of how specific aggregate types
affect product safety or efficacy