Excipients differentially influence the conformational stability and pretransition dynamics of two IgG1 monoclonal antibodies

Article (PDF Available)inJournal of Pharmaceutical Sciences 101(9):3062-77 · September 2012with65 Reads
DOI: 10.1002/jps.23187 · Source: PubMed
Abstract
Since immunoglobulins are conformationally dynamic molecules in solution, we studied the effect of stabilizing and destabilizing excipients on the conformational stability and dynamics of two IgG1 monoclonal antibodies (mAbs; mAb-A and mAb-B) using a variety of biophysical approaches. Even though the two mAbs are of the same IgG1 subtype, the unfolding patterns, aggregation behavior, and pretransition dynamics of these two antibodies were strikingly different in response to external perturbations such as pH, temperature, and presence of excipients. Sucrose and arginine were identified as stabilizers and destabilizers, respectively, on the basis of their influence on conformational stability for both the IgG1 mAbs. The two excipients, however, had distinct effective concentrations and different effects on the conformational stability and pretransition dynamics of the two mAbs as measured by a combination of differential scanning calorimetry, high-resolution ultrasonic spectroscopy, and red-edge excitation shift fluorescence studies. Stabilizing concentrations of sucrose were found to decrease the internal motions of mAb-B, whereas arginine marginally increased its adiabatic compressibility in the pretransition region. Both sucrose and arginine did not influence the pretransition dynamics of mAb-A. The potential reasons for such differences in excipient effects between two IgG1 mAbs are discussed.
Excipients Differentially Influence the Conformational Stability
and Pretransition Dynamics of Two IgG1 Monoclonal Antibodies
SANTOSH V. THAKKAR,
1
SANGEETA B. JOSHI,
1
MATTHEW E. JONES,
1
HASIGE A. SATHISH,
2
STEVEN M. BISHOP,
2
DAVID B. VOLKIN,
1
C. RUSSELL MIDDAUGH
1
1
Department of Pharmaceutical Chemistry, Macromolecule and Vaccine Stabilization Center, University of Kansas, Lawrence,
Kansas 66047
2
Formulation Sciences, MedImmune, Gaithersburg, Maryland 20878
Received 27 February 2012; revised 11 April 2012; accepted 20 April 2012
Published online 11 May 2012 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.23187
ABSTRACT: Since immunoglobulins are conformationally dynamic molecules in solution, we
studied the effect of stabilizing and destabilizing excipients on the conformational stability and
dynamics of two IgG1 monoclonal antibodies (mAbs; mAb-A and mAb-B) using a variety of
biophysical approaches. Even t hough the two mAbs are of the same IgG1 subtype, the unfolding
patterns, aggregation behavior, and pretransition dynamics of these two antibodies were strik-
ingly different in response to external perturbations such as pH, temperature, and presence of
excipients. Sucrose and arginine were identified as stabilizers and destabilizers, respectively, on
the basis of their i nfluence on conformational stability for both the IgG1 mAbs. The two excipi-
ents, however, had distinct effective concentrations and different effects on the conformational
stability and pretransition dynamics of the two mAbs as measured by a combination of differen-
tial scanning calorimetry, high-resolution ultrasonic spectroscopy, and red-edge excitation shift
fluorescence studies. Stabilizing concentrations of sucrose were found to decrease the internal
motions of mAb-B, whereas arginine marginally increased its adiabatic compressibility in the
pretransition region. Both sucrose and arginine did not influence the pretransition dynamics
of mAb-A. The potential reasons for such differences in excipient effects between two IgG1
mAbs are discussed. © 2012 Wiley Periodicals, Inc. and the American Pharmacists Association
J Pharm Sci 101:3062–3077, 2012
Keywords: proteins; immunoglobulins; stability; pretransition dynamics; protein aggrega-
tion; excipients; protein formulation; biophysical techniques
INTRODUCTION
Proteins in solution are inherently conformationally
dynamic molecules composed of atoms that are in a
state of constant motion at ambient temperatures.
1
At equilibrium, the native form of a protein is be-
lieved to sample a statistical ensemble of intercon-
verting microstates that undergo continuous fluctu-
ations, resulting in protein motions on the spatial
scale of sub-nanometer to tens of nanometers and
a temporal scale of femtoseconds to hours.
2
Pro-
tein dynamics are known to influence a wide variety
of biological processes including folding,
1
enzymatic
Additional Supporting Information may be found in the online
version of this article. Supporting Information
Corresponding to: C. Russell Middaugh (Telephone: +785-864-
5813; Fax: +785-864-5814; E-mail: middaugh@ku.edu)
Journal of Pharmaceutical Sciences, Vol. 101, 3062–3077 (2012)
© 2012 Wiley Periodicals, Inc. and the American Pharmacists Association
activity,
3,4
signaling,
5
allostery,
6
ligand binding,
7–9
and stability.
10
Several studies suggest that exam-
ining the dynamics of proteins could play a role in
elucidating more complex correlations that may exist
between protein stability and function.
11–15
Changes in solution properties (e.g., pH, tem-
perature, ionic strength, and the presence of co-
solvents) as well as the structure of water itself
(predominantly in the hydration layer) may signif-
icantly influence the structure, stability, dynamics,
and function of biologically and pharmaceutically
important proteins.
16–21
A variety of biophysical
techniques such as X-ray crystallography,
22,23
nu-
clear magnetic resonance,
3,24–26
neutron scattering,
23
isotope exchange,
27
ultrasonic spectroscopy,
10,20,28–31
and pressure perturbation calorimetry
20,32–35
have
been employed to probe fluctuations in the inter-
nal motions of proteins and/or their surrounding sol-
vent. Numerous lower resolution techniques such as
3062 JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 9, SEPTEMBER 2012
EXCIPIENTS DIFFERENTIALLY INFLUENCE THE CONFORMATIONAL STABILITY 3063
ultraviolet (UV)-absorption, fluorescence, circular
dichroism (CD), and light scattering among oth-
ers have commonly been employed to characterize
higher order structures, hydrodynamic properties,
and conformational stability of proteins.
10,36–44
Data
from these multiple biophysical techniques can be
combined in a vector-based stress/response method
known as an empirical phase diagram
45–47
(EPD). An
EPD displays distinct colored regions, which repre-
sent different conformational states of proteins and
other macromolecular systems as a function of so-
lution conditions such as pH and temperature. In
a recent study,
47
an EPD was generated for an
IgG1 monoclonal antibody (mAb-B) based on tech-
niques sensitive to the dynamic properties of pro-
teins such as high-resolution ultrasonic spectroscopy
(HR-US), pressure perturbation calorimetry, red-edge
excitation shifts (REES), and time-resolved fluores-
cence spectroscopy. The results showed a more com-
plex pattern of apparent structural transitions at
lower temperatures in the pretransition region (be-
low any detectable unfolding event) compared with
an EPD generated from biophysical data using static
(time-averaged) measurements such as CD, steady-
state fluorescence spectroscopy, and light scattering.
The pretransition region is defined as a temperature
range over which the change in parameters tradition-
ally used to evaluate a protein’s secondary structure,
tertiary structure, and conformational stability does
not deviate from a continuous change with tempera-
ture, as studied by methods such as CD, fluorescence
spectroscopy, and differential scanning calorimetry
(DSC). A better understanding of any relationship
between conformational stability and dynamics, espe-
cially in the pretransition region, may be important
to our understanding of the development and formu-
lation of biopharmaceutical drugs such as monoclonal
antibodies (mAbs).
Monoclonal antibodies are an important class of
dynamic, Y-shaped proteins that are good models
for studying the interrelationships between confor-
mational stability and dynamics. The two Fab do-
mains of immunoglobulins are connected to the Fc
domain by a highly flexible proline-rich hinge region,
which is believed to affect the structure and dynam-
ics of immunoglobulins.
48,49
Various analytical tech-
niques have been used to study the flexibility and
dynamics of antibodies.
20,27,47–54
Different molecules
within an immunoglobulin subclass, despite their
overall similarity in structure and sequence homol-
ogy, may display significant differences in their con-
formational stability, flexibility, and dynamics. The
conformational stability of antibody drugs, formu-
lated at both low and high concentrations, is signifi-
cantly influenced by environmental and formulation
factors during manufacturing, long-term storage, and
administration.
55–57
The effect of these factors on pro-
tein dynamics, however, has not been examined to
any great extent. It is therefore important to not only
better understand any relationship between confor-
mational stability and dynamics for different mAbs,
but to also examine the effect of various environmen-
tal factors (e.g., pH, temperature, excipients, etc.) on
their conformational stability and dynamics.
In this study, the effect of stabilizing and destabi-
lizing excipients on conformational stability and in-
tramolecular protein dynamics of two different IgG1
mAbs (mAb-A and mAb-B) is compared to further
understand the relationships between stability and
dynamics.
EXPERIMENTAL
Materials
The IgG1 mAbs (mAb-A and mAb-B) were provided
by MedImmune (Gaithersburg, Maryland). The stock
protein solutions were stored as received at 2
C–8
C.
The dialysis of stock protein solutions was carried out
overnight (at 4
C) using a 10 kDa molecular weight
cutoff dialysis cassette (Pierce, Rockford, Illinois) into
20 mM citrate–phosphate buffer at pH values rang-
ing from 3 to 8 at one unit intervals, unless otherwise
noted. The final ionic strength of the buffer was ad-
justed to 0.1 using NaCl. All of the buffer components
and other chemicals were purchased from Sigma
(St. Louis, Missouri) and Fisher Scientific (Pitts-
burgh, Pennsylvania). The protein concentration was
measured at room temperature by absorbance mea-
surement at 280 nm using an extinction coefficient
1.45 mL/(mg cm) in an Agilent 8453 UV–visible
spectrophotometer (Palo Alto, California), and di-
luted to the final concentration as indicated in each
experiment.
Methods
Steady-State Intrinsic (tryptophan) and Extrinsic
(1-anilino naphthalene-8-sulphonate) Fluorescence
Spectroscopy, Far-UV CD, and OD
350 nm
Turbidity
Measurements
For each of these four techniques, the experimen-
tal method as applied to mAbs has been described
previously.
47
High-Resolution Ultrasonic Spectroscopy
Ultrasonic measurements
20,47,58–60
were performed
using an HR-US 102 Spectrometer (Ultrasonic Sci-
entific, Dublin, Ireland) with a frequency range of
2–18 MHz and a resolution of 0.2 mm/s for velocity
and 0.2% for attenuation measurements. The sam-
ple and reference cells contained 1 mL of protein and
corresponding buffer solution, respectively. The dif-
ferential velocity and attenuation were monitored at
12 MHz from 10
Cto85
C and pH 3–8 using 5 mg/mL
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 9, SEPTEMBER 2012
3064 THAKKAR ET AL.
of mAb-A. The temperature of the cells was con-
trolled by a Phoenix P2 water circulator (Thermo
Haake, Newington, New Hampshire). The sample and
reference solutions were thoroughly degassed before
each measurement. Appropriate amounts of sucrose
or arginine were added to both protein and buffer so-
lutions while evaluating excipient effects. Data were
analyzed using HRUS v4.50.27.25 software (Ultra-
sonic Scientific, Dublin, Ireland). The coefficient of
adiabatic compressibility (β
s
) was determined using
the following equations
20
:
β
s
=−
1
V
V
P
s =−
1
v
0

v
0
P
s
=
β
v
0
lim
c0
(β/β
0
) V
0
c
where
V
0
=
ρ c
ρ
0
;andv
0
= lim
c0
1 V
0
c
β and β
0
are the adiabatic compressibility of the solu-
tion and buffer, respectively; ρ and ρ
0
are the density
of the solution and the corresponding buffer; ν
0
is the
partial specific volume of the IgG; V
0
is apparent vol-
ume fraction of the buffer; and c is the protein concen-
tration. The adiabatic compressibility of the sample
and buffer are related to the density (ρ) and ultrasonic
velocity (μ) by the Laplace equation,
61
β = 1/ρμ
2
. The
effect of excipients on the compressibility of mAb-A
and mAb-B was studied similarly using solution con-
ditions described later in the text.
Density
The density of protein samples (5 mg/mL) and cor-
responding buffer solutions was measured using a
DMA-5000 high-precision densitometer (Anton Paar,
Graz, Austria) at a precision of 1 × 10
6
g/cm
3
and
0.001
C. The densities of degassed solutions were
measured from 5
Cto55
Cat2.5
C intervals. The
instrument was calibrated daily with dry air and de-
gassed water before analysis. For the excipient stud-
ies, both the protein sample and corresponding buffer
solution contained equal predetermined quantities of
each excipient.
Differential Scanning Calorimetry
The differential scanning calorimetric studies were
performed using a MicroCal VP-Capillary DSC
with an autosampler (MicroCal, Northampton, Mas-
sachusetts). The pH (pH 3–8 at unit intervals) ex-
periments for mAb-A and mAb-B were performed us-
ing 1 mg/mL of protein in 20 mM citrate–phosphate
buffer (I = 0.1 adjusted by the appropriate addition
of NaCl). The temperature ramp was programmed
from 10
Cto90
C at a scanning rate of 60
C/h and a
filtering period of 16 s. Protein thermograms were ob-
tained by subtracting the corresponding buffer blank
from the sample thermogram. The transition mid-
points were obtained by determining t he baseline us-
ing linear or cubic functions, normalizing it to protein
concentration, and fitting the processed thermogram
to a non-two-state unfolding model. The endothermic
peak maximum of the heat capacity was considered
to be the apparent transition midpoint (T
M
)forthe
individual peaks that could be deconvoluted from the
thermogram. The effect of varying concentrations of
excipients was studied similarly at pH 4 and pH 4.5
for mAb-A and mAb-B, respectively.
Empirical Phase Diagrams
Empirical phase diagrams are constructed to visu-
ally represent changes in the structural
45,47
and
dynamic
47
properties of proteins in the form of colored
diagrams as a function of solution variables such as
pH and temperature. The rationale and methodology
of EPD construction are described elsewhere.
45,46
Two
separate EPDs were constructed using mAb-A. For
the first EPD, experimental data as a function of pH
and temperature from the following static biophysical
techniques were used: intrinsic tryptophan (Trp) flu-
orescence intensity, Trp peak position shifts, static
light scattering intensity, 1-anilino naphthalene-8-
sulphonate (ANS) fluorescence intensity, CD at 217
nm, and optical density at 350 nm (OD
350 nm
) values.
The second EPD was constructed to include mea-
surements of dynamic properties of mAb-A by adding
compressibility data from HR-US. The latter reflects
changes in global dynamics as a function of pH and
temperature. These two EPDs for mAb-A are com-
pared with previously published results for mAb-B.
47
Different protein concentrations were used for dif-
ferent techniques (e.g., 0.1 mg/mL for fluorescence
measurements and 5 mg/mL for ultrasonic measure-
ments), and the concentration was chosen to obtain
higher signal-to-noise in individual measurements.
Since the concentrations used fall within the dilute so-
lution regime as determined in control (concentration-
dependent) experiments, the differences in protein
concentration do not alter the pretransition region
significantly and hence should not influence the con-
formational stability and dynamics of the proteins as
measured by these techniques.
Excipient Screening
Intrinsic Trp fluorescence and static light scatter-
ing were used to screen a generally-regarded-as-safe
(GRAS) library of excipients to identify compounds
that either increased or decreased the stability of
mAb-B, as determined by changes in the T
M
of mAb-B
unfolding. The concentrations of excipients used were
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 9, SEPTEMBER 2012 DOI 10.1002/jps
EXCIPIENTS DIFFERENTIALLY INFLUENCE THE CONFORMATIONAL STABILITY 3065
higher than those commonly used in protein formu-
lations to facilitate screening of excipients. The T
M
was determined from a sigmoidal fit of the first tran-
sition in intrinsic fluorescence intensity versus tem-
perature plots. The data were acquired between 10
C
and 90
C in increments of 2.5
C. Static light scatter-
ing was used to assess the propensity of the IgG1 an-
tibody to aggregate. Prior to experimentation, mAb-
B was dialyzed into 20 mM citrate–phosphate buffer
(containing NaCl to adjust I = 0.1) at pH 4.5. These
stress conditions were selected on the basis of the re-
duced stability of mAb-B at this pH,
47
which should
facilitate the identification of stabilizing compounds.
The protein concentration employed was 0.1 mg/mL.
A few selected candidate stabilizers and destabilizers
from this primary screen were used to further evalu-
ate their effect on the second IgG1 molecule (mAb-A),
either by the fluorescence method and/or DSC.
Red-Edge Excitation Spectroscopy
Red-edge excitation is a characteristic property of po-
lar fluorophores, which exhibit excitation wavelength
dependent emission spectra.
62,63
This phenomenon
depends upon motional restriction of the environ-
ment of fluorophores. The steady-state fluorescence
measurements of REES were performed using a PTI
Quanta Master Spectrophotometer (Photon Technol-
ogy International, Inc., Lawrenceville, New Jersey).
The excitation and emission slit widths were set
to 2.5 and 3 nm, respectively. The emission spectra
(300–400 nm) were collected using different excita-
tion wavelengths from 292 to 308 nm at 4 nm inter-
vals. Both mAb-A and mAb-B at 0.1 mg/mL were stud-
ied in the absence and presence of selected excipients
using a 1 cm pathlength quartz cuvette. A temper-
ature ramp from 10
Cto70
Cat2.5
C increments
was used with an equilibration time of 3 min. An ap-
propriate blank spectrum was subtracted from the
sample spectrum. The emission peak position (and in-
tensity) was determined by a mean spectral center of
mass (MSM) method, which increased reproducibil-
ity and signal-to-noise ratio and thus improved our
ability to measure the relative shifts in peak posi-
tion in the presence of excipients. The peak position
maxima obtained by the MSM method is approxi-
mately 8–10 nm higher than the actual peak position
obtained by derivative analysis.
RESULTS
Characterization of Higher Order Structure,
Conformational Stability, and Dynamics of mAb-A as a
Function of pH and Temperature
Static (Time-Averaged) Measurements
The results from a variety of biophysical characteriza-
tion measurements of mAb-A as a function of pH and
temperature are shown in Figures 1a–1f and Supple-
mentary Figures S1a–S1f. A well-defined structural
transition occurs in mAb-A at 65
C–75
C over the
pH range of 5–8 as detected by increases in Trp flu-
orescence intensity (Fig. 1a) and a red shift of emis-
sion peak maximum (Fig. 1b). The onset temperature
(T
onset
) of unfolding at pH 5 starts approximately at
60
C with a broad unfolding curve (Fig. 1b). The early
T
onset
at pH 5 is accompanied by an increase in ANS
fluorescence intensity (Fig. 1c), suggesting the expo-
sure of apolar sites. The presence of intermolecular
β-structure-rich structures at pH 5–8 is also detected
at approximately 63
C–78
C as suggested by a de-
crease in the CD signal (Fig. 1f). The unfolding event
at pH 5–8 leads to further aggregation, which is ap-
parent from an increase in light scattering intensity
(Fig. 1d) and OD
350 nm
(Fig. 1e) measurements. The
T
onset
for the unfolding transition (Fig. 1c) in the range
of pH 5–8, however, follows an opposite trend com-
pared with the T
onset
of aggregate formation (Figs. 1d
and 1e). For example, the T
onset
of unfolding is low-
est (60
C) at pH 5 (followed by pH 6 < pH 7 < pH
8), whereas the T
onset
of aggregation is the highest
(85
C)atpH5(followedbypH6> pH 7 > pH 8). In
addition, at pH 4, the overall structural changes ob-
served for mAb-A as a function of temperature were
similar (Figs. 1b, 1c, and 1f) to that at pH 5 up to ap-
proximately 70
C, although the transitions were ob-
served at lower temperatures. The CD signal at pH 4,
however, continues to decrease above 75
C, suggest-
ing that the intermolecular interactions continue to
increase with increases in temperature. The behavior
of mAb-A at pH 3 is significantly different than that at
other pH values below 65
C, where the protein man-
ifests multiple distinct structural transitions appar-
ent from ANS fluorescence intensity change (Fig. 1c),
Trp peak position shifts (Fig. 1), and CD (Fig. 1f).
No increase in light scattering intensity (Fig. 1d) or
OD (Fig. 1e) was observed at pH 3 and 4 up to 90
C.
Fluorescence emission spectra (intrinsic and ANS) for
mAb-A at 15
C, 35
C, and 60
C are shown as a func-
tion of pH in Supplementary Figures S1a–S1f. These
data s uggest that mAb-A retains its native-like struc-
ture (Supplementary Figs. S1a, S1b, S1d, and S1e)
at temperatures in the pretransition region (15
C
and 35
C) compared with results at higher temper-
atures (e.g., 60
C) that results in structural disrup-
tions (Supplementary Figs. S1c and S1f). Upon exci-
tation at 295 nm, the intrinsic Trp fluorescence emis-
sion represents the average emission signal from all
the Trp residues (22) present in the IgG1 molecules
used in this study.
High-Resolution Ultrasonic Spectroscopy
The global dynamics of mAb-A were studied by de-
termining compressibility (volume fluctuations with
changes in pressure) as a function of pH and
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 9, SEPTEMBER 2012
3066 THAKKAR ET AL.
Figure 1. Effect of pH and temperature on conformational stability of an IgG1 monoclonal
antibody (mAb-A) as measured by a variety of biophysical techniques: (a) intrinsic tryptophan
fluorescence intensity, (b) intrinsic tryptophan fluorescence peak position shifts, (c) ANS fluo-
rescence intensity, (d) static light scattering, (e) OD
350 nm
, and (f) CD signal at 217 nm.
temperature using HR-US (Fig. 2). The adiabatic com-
pressibility of mAb-A was calculated as a function
of pH and temperature by determining the relative
changes in the ultrasonic velocity between the sam-
ple and reference. The adiabatic compressibility of
mAb-A was found to increase as a function of temper-
ature at all pH values. Plots of adiabatic compressibil-
ity versus temperature, however, show a unique non-
linear increase in the pretransition regions (<45
C)
starting at pH 4 and above. These HR-US deviations
occur at lower temperatures than the respective T
onset
and T
M
1 values as measured by DSC (see next sec-
tion). This nonlinear dependence of adiabatic com-
pressibility may therefore reflect some form of change
in the global dynamics of mAb-A in the pretransition
region.
EPDs for mAb-A
The EPD constructed for mAb-A using the time-
averaged measurements is presented in Figure 3a.
A broad structural transition is apparent between
60
Cand75
C for pH 5–8. At pH 4, two distinct
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 9, SEPTEMBER 2012 DOI 10.1002/jps
EXCIPIENTS DIFFERENTIALLY INFLUENCE THE CONFORMATIONAL STABILITY 3067
Figure 2. Adiabatic compressibility of mAb-A as a function of pH and temperature as mea-
sured by HR-US. The straight lines along the data points are a visual aid for comparison of
pretransition regions. The T
onset
and T
M
values represent the initiation and peak maximum for
the first thermal transition as determined separately by DSC.
structural transitions occur at approximately 50
C
and approximately 70
C. The contributions from ANS
fluorescence intensity changes (Fig. 1c) and CD sig-
nals (Fig. 1f) may contribute the most to these appar-
ent transitions at pH 4. The transitions i n mAb-A at
pH 3 start at approximately 20
C with multiple sub-
sequent minor transitions observed every 10
C–15
C.
The EPD in Figure 3a has been divided into three
distinct phases, i.e., Phases I, II, and III, represent-
ing regions of stable, unstable, and aggregated form of
mAb-A, respectively. It has previously been reported
47
that a “dynamic” EPD constructed using another IgG1
(mAb-B) was able to detect an additional transition
region in the low-temperature (<45
C) region aris-
ing from contributions of the adiabatic compressibil-
ity measurements. An EPD using mAb-A was there-
fore constructed combining static and compressibility
measurements as a function of pH and temperature
(Fig. 3b). REES data were not included in the EPD
because of lower resolution of these data at higher
temperatures. This “dynamic” EPD for mAb-A shows
an additional transition region (Phase 1
)atpH
4 and at temperatures below 45
C compared to the
static EPD (Fig. 3a). This EPD with mAb-A, together
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 9, SEPTEMBER 2012
3068 THAKKAR ET AL.
Figure 3. Empirical phase diagrams for mAb-A as a function of pH and temperature using (a)
static (time-averaged) biophysical techniques alone, and (b) data from static techniques in (a)
in conjunction with HR-US data. Static biophysical techniques data from intrinsic tryptophan
fluorescence intensity and peak position shifts, static light scattering intensity, extrinsic ANS
fluorescence intensity, circular dichroism at 217 nm, and OD
350 nm
values were used. A continu-
ous color in the phase diagram represents a single structural state of the protein. Transitions
in the proteins’ structure are manifested by changes in color.
with the previously published
47
results with mAb-B,
suggests that ultrasonic measurements provide addi-
tional information about conformational fluctuations
and flexibility in IgG1 that are not apparent with t he
use of conventional biophysical techniques alone, es-
pecially in the pretransition regions. The presence of
additional structural effects in the pretransition re-
gions of the two IgG1 antibodies at approximately pH
5–8 emphasizes the need for developing a better un-
derstanding of the effect of formulation components
not only on equilibrium conformational stability, but
also on the dynamic properties of protein therapeutic
drugs in s olution.
Thermal Stability (T
M
) of mAb-A and mAb-B
Differential scanning calorimetry is routinely used
to study the thermal stability of antibodies
20,64
by
measuring the differential heat capacity to determine
midpoints of thermal unfolding events (T
M
s). DSC
was used to directly compare the conformational sta-
bility of mAb-A and mAb-B as function of solution pH
(Fig. 4; Supplementary Table 1). Three distinct con-
formational transitions (T
M
1, T
M
2, and T
M
3) were
apparent for mAb-A at pH 3–7, albeit at variable
temperatures. Only two major structural transitions,
however, were readily observable at pH 8 (Fig. 4b).
In contrast, mAb-B showed only two distinct transi-
tions in the pH range of 5–8 by DSC (T
M
2andT
M
3
in Figs. 4c and 4d). In the case of mAb-B at pH 3 and
4, an additional lower temperature transition ( T
M
1)
was detected.
Screening of a GRAS Library of Excipients
To better understand the effect of formulation ex-
cipients on conformational stability and global dy-
namics of these two IgG1 mAbs, a first set of ex-
periments screened a GRAS library of excipients to
identify potential stabilizing and destabilizing excip-
ients using mAb-B. The EPDs for mAb-A (Fig. 3) and
mAb-B (Ramsey et al.,
47
Fig. 3) indicate conforma-
tional instability in the range of 55
C–65
CatpH
approximately 4–4.5. The intrinsic Trp fluorescence
intensity method was utilized to identify stabilizing
excipients (T
M
) under these accelerated pH condi-
tions that altered the stability of mAb-B’s tertiary
structure (Table 1). This methodology allowed for
determination of the protein unfolding temperature
(T
M
) as well as an assessment of aggregation behav-
ior (Table 1). Sugars and polyols in general increased
the T
M
of mAb-B, whereas amino acids such as argi-
nine and histidine lowered the transition tempera-
ture. A few selected candidate stabilizers (sucrose,
dextrose, and mannitol) and destabilizers (arginine)
were then tested with mAb-A (using DSC) at pH 4
to identify common excipients that would either sta-
bilize or destabilize both of the IgG1 antibodies (data
not shown). On the basis of their effects on the T
M
val-
ues, as measured by fluorescence spectroscopy with
mAb-B and DSC with mAb-A, sucrose was selected as
a representative candidate stabilizer, whereas argi-
nine was used as a destabilizing excipient for both
proteins. All subsequent studies evaluating the effect
of excipients on conformational stability and pretran-
sition dynamics were performed at pH 4 and pH 4.5
for mAb-A and mAb-B, respectively. These solution
conditions were selected because the magnitude of
stabilizing and destabilizing effects of these two ex-
cipients under neutral pH conditions were smaller
than the chosen more acidic pH solution conditions.
The selection of lower pH, however, does not preclude
our ability to study the effect of excipients on the
pretransition conformational dynamics of the mAbs
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 9, SEPTEMBER 2012 DOI 10.1002/jps
EXCIPIENTS DIFFERENTIALLY INFLUENCE THE CONFORMATIONAL STABILITY 3069
Figure 4. Representative differential scanning calorimetric thermograms for mAb-A (a) and
mAb-B (c) over pH 3–8 and 10
C–100
C (shown only above 20
C); and plot of midpoint of thermal
unfolding values (T
M
1, T
M
2, T
M
3) for mAb-A (b) and mAb-B (d) as a function of pH. Error bars
cannot be seen in (b) and (d) because they are within the symbols. See Supplementary Table 1
for summary of T
M
results and their corresponding SD values.
because the T
onset
of unfolding for both the proteins
under these conditions is still 45
C.
Effect of Arginine and Sucrose on Conformational
Stability and Dynamics of IgG1 mAb-A and mAb-B
Effect on Conformational Stability
The effect of sucrose and arginine on the confor-
mational stability of mAb-A and mAb-B was stud-
ied using DSC and intrinsic Trp fluorescence spec-
troscopy (Figs. 5 and 6; Supplementary Table 2). DSC
and fluorescence measurements were used to deter-
mine the effect of excipients on the overall thermal
stability and tertiary structure stability, respectively.
Fig. 5a shows a representative DSC thermogram
for mAb-B at pH 4.5. Three distinct transitions
were detected for mAb-B at pH 4.5 with T
M
1 being
the first low-temperature transition, T
M
2 the sec-
ond transition, and T
M
3 the third transition seen
at the highest temperature. Similarly, three distinct
transitions were observed for mAb-A at pH 4 (data
not shown). Arginine (up to 300 mM) was found
to destabilize both mAb-A (Fig. 5c; Supplementary
Table 2a) and mAb-B (Fig. 5e; Supplementary
Table 2c) in a concentration-dependent manner. Su-
crose (up to 500 mM) showed a concentration-
dependent stabilization effect on mAb-A (Fig. 5d; Sup-
plementary Table 2b) and mAb-B (Fig. 5f; Supplemen-
tary Table 2d).
One goal of a protein formulation strategy would
be to identify excipients that stabilize the first con-
formational transition (which is typically the unfold-
ing of CH2 domain for IgG1) and inhibit subsequent
protein unfolding. Figure 5b shows the effect of argi-
nine and sucrose concentration on the T
M
1 values
for both mAb-A and mAb-B. Arginine was more po-
tent at destabilizing mAb-A than mAb-B. For exam-
ple, to achieve approximately 2.5
C destabilization,
a lesser amount of arginine (highlighted with a rect-
angle) was required for mAb-A compared to mAb-B.
Sucrose, however, was a more potent stabilizer for
mAb-B compared to mAb-A. As shown in Figure 5b,
to achieve approximately 2.5
C stabilization, a lesser
amount of sucrose (highlighted with a rectangle) was
required for mAb-B compared to mAb-A.
Figure 6 shows the effect of the two excipients on
the tertiary structure stability of mAb-A (Fig. 6a) and
mAb-B (Fig. 6b) as monitored by Trp peak position
shifts as a function of temperature. Arginine does
not influence the tertiary structure stability of mAb-
A throughout the temperature range examined. For
mAb-A in presence of sucrose, however, the protein
showed blue-shifted Trp peak positions above 25
C,
suggesting that aromatic residues are shielded from
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 9, SEPTEMBER 2012
3070 THAKKAR ET AL.
Table 1. Effect of Excipients on Thermal Unfolding Temperature (T
M
) of mAb-B as Monitored by Intrinsic Tryptophan Fluorescence
Spectroscopy Using 0.1 mg/mL P rotein at pH 4.5 in 20 mM Citrate–Phosphate Buffer (Containing NaCl, I = 0.1 and Indicated Level of
Excipients)
Category Name Concentration
a
T
M
(
C)
b
T
M
(
C)
b
Aggregation Inhibition (%)
c
Protein IgG1 (mAb-B) 0.1 mg/mL 59.9
Stabilizer Lactose 10% 62.5 2.6 75.7
Trehalose 10% 61.8 1.9 30.5
Dextrose 10% 63.1 3.2 63.8
Sucrose 10% 62.4 2.5 81.9
Mannitol 10% 62.2 2.3 NC
Sorbitol 10% 61.9 2.0 59.7
Malic acid 0.30 M 62.7 2.8 ND
Destabilizer "-Cyclodextrin 2.5% 57.2 –2.7 NC
2-Hydroxypropyl-β-cyclodextrin 10% 56.8 –3.1 NC
Aspartic acid 0.075 M 58.5 –1.4 NC
Lactic acid 0.15 M 58.2 –1.7 NC
Arginine 0.30 M 55.9 –4.0 61.7
Diethanolamine 0.30 M 57.5 –2.4 NC
Guanidine 0.30 M 57.9 –2.0 NC
Histidine 0.21 M 54.1 –5.8 NC
Pluronic F-68 0.1% 56.7 –3.2 ND
Neutral excipients Sodium citrate 0.1 M 58.9 –1.0 NC
Brij 35 0.1% 60.0 0.1 57.8
Tween 20 0.1% 59.7 –0.2 NC
Tween 80 0.1% 58.9 –1.0 NC
Glycine 0.30 M 61.1 1.2 62.5
Proline 0.30 M 59.6 –0.3 NC
Glycerol 10% 58.9 –1.0 64.7
Dextran T40 0.0075 mM 59.4 –0.5 53.1
2-Hydroxypropyl-(-cyclodextrin 10% 58.9 –1.0 NC
Glutamic acid 0.30 M 60.7 0.8 ND
Lysine 0.30 M 58.9 –1.0 ND
T
M
and aggregation inhibition (%) columns represent the change in midpoint of thermal unfolding and change in aggregation, respectively, for mAb-B in
presence of excipients.
a
Excipient concentrations are higher than those commonly used in formulations to facilitate excipient screening.
b
The T
M
measurements were made within the standard deviation of ±0.5
C.
c
The aggregation inhibition (%) represents a mean of three measurements with a standard deviation of ±2.5% for excipients that inhibited mAb-B
aggregation. NC is a group of excipients that resulted in no change (±10%) in the aggregation behavior of mAb-B. ND is a group of excipients that increased
mAb-B aggregation.
the s olvent. In the case of mAb-B, in the absence of
excipients, the protein shows a red shift in Trp peak
position (T
onset
45
C) upon thermal unfolding, indi-
cating exposure of aromatic residues to the solvent
(Fig. 6b) with increasing temperature. Sucrose was
found to stabilize, whereas arginine destabilized the
tertiary structure of mAb-B in terms of both T
onset
and T
M
.
Effect on Protein Dynamics
The effect of sucrose and arginine on the global dy-
namics of mAb-A and mAb-B was first examined by
the determination of compressibility values using HR-
US as shown in Figure 7. The measurements of pro-
tein compressibility are directly related to the fluctu-
ations in volume of the protein, thereby reflecting a
form of the dynamics and flexibility of proteins. Figure
7a shows that the adiabatic compressibility increases
with temperature for both mAb-A (pH 4) and mAb-
B (pH 4.5). An increase in compressibility suggests
that the relative difference in the ultrasonic velocity
between the sample and the reference is decreased,
whereas the absolute value of the velocity increases
as a function of temperature. It is well known that
mobile or less structured molecules possess a lower
elastic modulus compared to rigid, more structured
species. This results in a decrease in sound veloc-
ity through the unstructured or mobile material. A
higher compressibility value is therefore indicative
of a less structured and/or a more dynamic protein.
As seen in Figure 7a, the compressibility of mAb-A
is relatively greater than mAb-B in the temperature
range 10
C–50
C, that is, before any major detectable
conformational transitions.
The effect of sucrose and arginine on mAb-A com-
pressibility as a function of temperature is shown in
Figure 7b. Sucrose does not significantly influence the
compressibility of mAb-A in the pretransition region,
whereas arginine marginally decreases the compress-
ibility of mAb-A, especially in the pretransition re-
gion. Figure 7c represents the effect of sucrose and
arginine on the compressibility of mAb-B. The com-
pressibility of mAb-B in the presence of arginine was
found to be marginally increased in the pretransition
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 9, SEPTEMBER 2012 DOI 10.1002/jps
EXCIPIENTS DIFFERENTIALLY INFLUENCE THE CONFORMATIONAL STABILITY 3071
Figure 5. Effect of sucrose and arginine on conformation stability of mAb-A and mAb-B as
measured by DSC: (a) representative DSC thermograms for mAb-B (pH 4.5) with T
M
1, T
M
2,
and T
M
3; (b) effect of different concentrations of excipients on T
M
1 for mAb-A and mAb-B. The
box represents the effective concentration of excipient required to have approximately 2.5
Cof
effect. Plots of T
M
values for mAb-A (c and d) and mAb-B (e and f) in the presence of varying
concentrations of arginine (c and e) and sucrose (d and f). Error bars often cannot be seen
because they fall within the dimensions of the data points.
region. The significant increase in the compressibility
of mAb-B in the presence of arginine above 55
C can
be most likely be explained by the formation of highly
compressible, irreversible aggregates. The compress-
ibility of mAb-B in the presence of sucrose was sig-
nificantly reduced throughout the range of temper-
atures used in this study, but predominantly i n the
pretransition region. The lowering of compressibil-
ity suggests that the global dynamics of mAb-B are
dampened in presence of stabilizing concentrations of
sucrose.
Figure 8 shows the REES results for mAb-A and
mAb-B in the presence of sucrose and arginine.
The REES is a phenomenon in which there is a
shift in emission spectra maxima upon red-edge
excitation.
62,63
Such an effect is primarily seen when
the lifetime of solvent relaxation is equal to or larger
than the lifetime of the fluorophore of interest. The
solvent reorientation or relaxation around an excited
state fluorophore is influenced by dynamic motions
within proteins and solvent fluctuations around the
fluorophore’s environment. At 10
C, both mAb-A and
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 9, SEPTEMBER 2012
3072 THAKKAR ET AL.
Figure 6. Effect of sucrose and arginine on Trp peak po-
sition shifts for (a) mAb-A and (b) mAb-B as a function of
temperature using fluorescence spectroscopy.
mAb-B show red-edge shifts in the absence of excipi-
ents, suggesting that the fluorophores examined (Trp
in this case) are in an environment wherein the life-
time of solvent relaxation is either equal to or longer
than the lifetime of the fluorophore. Such a system is
suitable for studying the effects of excipients on the
conformational flexibility and dynamics of proteins by
monitoring their effect on the magnitude of any ob-
served red-edge shifts. Since the REES effect occurs
because the longer wavelength excitation results in
photoselection of fluorophores that are strongly inter-
acting with polar solvent molecules in their vicinity, a
less dynamic (or more rigid) fluorophore environment
will lead to a decrease in solvent relaxation of the flu-
orophore and thus the magnitude of red-edge shifts
will increase. As shown i n Figures 8a and 8b, neither
sucrose nor arginine altered the red-edge shifts ob-
served in mAb-A in the temperature range spanning
the pretransition region of the antibody. This result
suggests that these two excipients do not significantly
affect the global dynamics of mAb-A in which the local
environment around Trp residues was sampled. By
contrast, concentrations of sucrose t hat stabilized the
Figure 7. Effect of sucrose and arginine on adiabatic com-
pressibility of mAb-A and mAb-B as measured by HR-US:
(a) comparison of compressibility of mAb-A and mAb-B in
the pretransition region, (b) effect of arginine and sucrose
on adiabatic compressibility of mAb-A, and (c) effect of argi-
nine and sucrose on adiabatic compressibility of mAb-B.
tertiary structure of mAb-B were found to increase
the magnitude of red-edge shifts in the pretransition
region (Fig. 8d). Arginine did not alter the magni-
tude of the red-edge shifts in the pretransition region
(<55
C) of mAb-B (Fig. 8c).
DISCUSSION
Comparison of Higher Order Structure, Thermal Stability
Behavior, and EPDs Between mAb-A and mAb-B
47
The biophysical data suggest that mAb-A undergoes
multistep unfolding upon thermal unfolding with
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 9, SEPTEMBER 2012 DOI 10.1002/jps
EXCIPIENTS DIFFERENTIALLY INFLUENCE THE CONFORMATIONAL STABILITY 3073
Figure 8. Red-edge excitation shifts (REES) fluorescence measurements with mAb-A (a and
b) and mAb-B (c and d) in the presence and absence of sucrose (b and d) and arginine (a and c).
formation of intermolecular β-structure-rich
oligomeric structures at pH 4–8 at temperatures prior
to major unfolding/aggregation events. These inter-
molecular interactions appear to be accompanied
by shielding of aromatic residues from the solvent,
as suggested by the blue shift in Trp peak position
observed at pH 5–8 between approximately 45
Cand
65
C (Fig. 1b). The T
onset
of unfolding for mAb-A was
inversely related to the T
onset
of aggregation between
pH 5 and 8. For example, mAb-A at pH 5 had the
lowest T
onset
of unfolding but the highest T
onset
of
aggregation as measured by ANS fluorescence inten-
sity and static light scattering (or OD), respectively
(Figs. 1c, 1d, and 1e). An initial increase in ANS
intensity (Fig. 1c) at pH 5 reaches a plateau above
65
C. This result along with a continuous decrease in
the CD signal (Fig. 1f) at pH 5 above 65
C suggests
that no additional apolar residues are exposed in
this temperature range. There is, however, an in-
crease in β-structure-rich intermolecular oligomeric
structures. These results suggest that mAb-A at pH
5 forms partially altered structures that are either
stable in solution and/or form oligomeric species that
are resistant to the formation of larger aggregates
that can be detected by static light scattering and
OD measurements. Furthermore, a steep increase
in ANS fluorescence intensity and increase in CD
signal (Figs. 1c and 1f) at pH 6–8 above 78
C suggest
that additional aromatic residues are being exposed
and that the intermolecular β-structures begin to
dissociate. The T
onset
for such a dissociation event
above 78
C is found to be in the following order:
pH 8 < pH 7 < pH 6 (Fig. 1f). Once these structures
dissociate and additional aromatic residues are
exposed, the antibody may become more prone to
formation of irreversible aggregates. The detection of
larger aggregated species (Figs. 1d and 1e) follows
a similar trend as mentioned above (i.e., T
onset
for
aggregation pH 8 < pH 7 < pH 6), arguing that a
dissociation of relatively stable oligomeric species
may precede the formation of larger aggregates in
solution at higher pH. The sudden drop in light
scattering intensity and OD (after an initial increase
at temperatures above 83
C) indicates that the ag-
gregates eventually fall out of solution. At pH 3 and
4, mAb-A manifests blue-shifted Trp peak positions,
an increase in ANS intensity, and decreases in CD
signal which indicate conformational alterations at
much lower temperatures compared to higher pH
conditions. Nevertheless, the protein remains aggre-
gation resistant under these conditions. In contrast,
a single cooperative transition was observed in the
case of mAb-B
47
between 60
Cand70
C at pH values
ranging from 5 to 8 based on static measurements
(Fig. 1).
47
The unfolding event in mAb-B leads to
the exposure of aromatic residues and the protein
subsequently forms larger aggregates at pH 5–8
(Figs. 1e and 1f).
47
Furthermore, mAb-B at pH 5–8
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 9, SEPTEMBER 2012
3074 THAKKAR ET AL.
aggregates over a very narrow temperature range
as detected by static light scattering. The T
onset
of
unfolding had no correlation with the T
onset
of aggre-
gation in this pH range (Figs. 1e and 1f).
47
In this
same study, mAb-B at pH 4 showed a broad unfolding
transition and aggregated to a lesser extent than at
pH 5–8, whereas the protein at pH 3 was found to
be resistant to formation of detectable aggregates. In
summary, these results from a variety of biophysical
measurements clearly show that IgG1 mAb-A and
mAb-B have distinct patterns of conformational
alterations and aggregation behavior in response to
changes in pH and temperature.
Antibodies by virtue of their multidomain structure
routinely display multiple conformational transitions
due to environmental stresses such as changes in pH
and temperature, which can be detected by DSC. A
number of previous DSC studies have assigned these
different transitions to the unfolding of antigen bind-
ing (Fab) region, crystallizable (Fc) region, or individ-
ual domains within Fab and Fc regions.
65–67
The pres-
ence of three transitions detected by DSC for mAb-A
(Figs. 4a and 4b) over a wide range of pH (3–7) cor-
related with spectroscopic measurements, which also
showed multiple transitions (Figs. 1b, 1c, and 1f). The
highest T
onset
of unfolding for mAb-A (Figs. 1b and 1c)
and the lowest T
onset
of aggregation (Figs. 1d and 1e)
at pH 8 indicate that the protein aggregates rapidly
upon unfolding. Such a concerted phenomenon may
explain only two major transitions that are apparent
at pH 8. In contrast, mAb-B was found to have two
major transitions at higher pH except at pH 3 and 4
(Figs. 4c and 4d). Our previous spectroscopic results
47
using mAb-B have shown that it undergoes a single
cooperative unfolding transition at pH 5–8. There-
fore, the two main structural transitions (compared
with three transitions for mAb-A) detected by DSC for
mAb-B may suggest that the unfolding of one domain
leads to an immediate subsequent unfolding of other
domains within the mAb-B molecule. The additional
transition detected at pH 3 and 4 by DSC (Fig. 4d) was
also observed by other spectroscopic methods at sim-
ilar temperatures (Ramsey et al.,
47
Figs. 1b and 1f).
The thermal behavior and aggregation data for mAb-
B suggests that no detectable stable i ntermediates
are formed and that the protein undergoes concerted
thermal unfolding and subsequent aggregation under
these solution conditions.
In addition, comparison of structural features and
thermal stability profiles of the two mAbs can be com-
pared by their differences in the static and dynamic
EPDs, i.e., mAb-A (Fig. 3 in this work) and mAb-
B (Ramsey et al.,
47
Fig. 3). The static EPDs gener-
ated for mAb-A and mAb-B
47
were both able to detect
differences in conformational stability as a function
of temperature and pH. The dynamic EPD contain-
ing the additional compressibility results for mAb-A
(Fig. 3b) was able to detect changes in dynamic be-
havior in the pretransition region at a broader pH
range (pH 4–8) than mAb-B (Ramsey et al.,
47
Fig. 3),
wherein the dynamic EPD was found to contain addi-
tional regions in the pH range of 6–8. This difference
in the EPDs of mAb-A and mAb-B may potentially be
due to inherent differences between the two IgGs or
due to experimental differences such as: (1) the dy-
namic properties of mAb-B were studied above 20
C
rather than 10
C for mAb-A. Therefore, the confor-
mational fluctuations that may exist at lower tem-
peratures could be incompletely represented in the
EPD for mAb-B, and/or (2) ANS fluorescence results
for mAb-B (Ramsey et al.,
47
Fig. 1f) at lower pH val-
ues show a broad unfolding transition with a T
onset
of
approximately 37.5
C. The smaller magnitude of dy-
namic fluctuations may therefore be obscured during
the mathematical data processing used to construct
the EPD.
Effects of Arginine and Sucrose on the Conformational
Stability and Pretransition Dynamics of mAb-A and
mAb-B
Arginine was found to be a destabilizer, whereas su-
crose was a stabilizer for both mAb-A and mAb-B in
a concentration-dependent manner, albeit at differ-
ent effective concentrations as determined by DSC
measurements. Arginine was a more potent destabi-
lizer for mAb-A compared to mAb-B, whereas sucrose
was more potent at stabilizing mAb-B than mAb-A
(Fig. 5b; Supplementary Table 2a). Arginine, how-
ever, did not perturb the tertiary structure stability
of mAb-A. Potential reasons for the destabilizing ef-
fect of arginine on mAb-A could either be the sup-
pression of intermolecular β-structure-rich oligomer
formation and/or promotion of its dissociation.
68
The
destabilization of such β-structure-rich intermediate
structures, which potentially stabilize partially al-
tered structures of mAb-A, may increase the propen-
sity of mAb-A to form larger aggregates. Further-
more, the blue shift observed in Trp peak position
as a function of temperature for mAb-A in the pres-
ence of sucrose suggests that this sugar may be stabi-
lizing the intermolecular β-structure-rich structures,
thus shielding the aromatic residues from the solvent.
Both of these excipients, however, influenced the ter-
tiary structure stability of mAb-B. Thus, the effect
of arginine and sucrose on the global thermal stabil-
ity and tertiary structure suggest that both mAb-A
and mAb-B interact with the s ame excipients in a
different manner and at different effective concentra-
tions. This may in part be because of the inherent
differences in the physicochemical properties and the
unfolding processes between the two proteins, where
mAb-A appears to have a greater propensity to form
β-structure-rich structures that stabilize the partially
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 9, SEPTEMBER 2012 DOI 10.1002/jps
EXCIPIENTS DIFFERENTIALLY INFLUENCE THE CONFORMATIONAL STABILITY 3075
altered native structure before any global unfolding/
aggregation event. In contrast, mAb-B undergoes a
more cooperative unfolding process starting with dis-
ruption of its tertiary structure followed by an imme-
diate aggregation of the structurally altered protein.
It is generally accepted that experimentally deter-
mined adiabatic compressibility values are comprised
of positive contributions from the intrinsic compress-
ibility of a protein and a negative contribution from a
hydration component.
61
Depending upon the magni-
tude of the intrinsic compressibility and the hydration
contribution, the apparent adiabatic compressibility
values may either be positive or negative. The lower
(negative at lower temperatures) adiabatic compress-
ibility values in the pretransition region for mAb-B
compared to mAb-A (Fig. 7a) suggest a combined ef-
fect of lower intrinsic compressibility and/or higher
hydration contribution to mAb-B. Such a combined
effect may result in stronger coupling of mAb-B con-
formational fluctuations to the fluctuations in the
surrounding hydration water and/or the proteins’ en-
vironment compared to mAb-A. Sucrose did not af-
fect the compressibility of mAb-A in the pretransition
region. This result may be related to the lower po-
tency of sucrose as a stabilizer of mAb-A. Arginine
marginally lowered the compressibility of mAb-A re-
sulting in negative values in the pretransition region,
suggesting an increased contribution of hydration to
the apparent adiabatic compressibility values. In con-
trast, arginine marginally increased the compressibil-
ity of mAb-B in the pretransition region. These mAb-B
compressibility values go from negative to positive at
lower temperatures, suggesting either an increase in
intrinsic compressibility and/or decrease in the hydra-
tion contribution. These effects, however, may not be
mutually exclusive. Arginine was found to destabilize
mAb-B by influencing its tertiary structure stability.
Such a destabilization effect may require perturba-
tion of the structure of water in the protein’s hydra-
tion layer, potentially due to the chaotropic nature of
guanidinium group in arginine, resulting in a lower
hydration contribution to the apparent compressibil-
ity of mAb-B. This could explain the marginal in-
crease in compressibility in the presence of arginine at
lower temperatures. Sucrose, however, significantly
decreased the compressibility of mAb-B especially in
the pretransition region. The negative values of mAb-
B compressibility in the presence of sucrose indicate
that the hydration component has a greater contribu-
tion to the experimentally determined compressibility
values. This suggests that solvent fluctuations around
the protein’s surface may have a significant effect on
the dynamic behavior of mAb-B in the presence of su-
crose. Overall, these compressibility measurements
suggest that mAb-B exists in a less dynamic or a more
compact form due to the potent stabilizing effect of su-
crose, which may increase the ordering of water in the
hydration layer due to a preferential hydration mech-
anism. In summary, HR-US results show that effec-
tive concentrations of arginine and sucrose did not
significantly i nfluence the dynamic behavior of mAb-
A in the pretransition region. Similarly, arginine did
not appreciably affect the pretransition dynamics of
mAb-B. Sucrose, however, significantly reduced the
dynamic behavior of mAb-B as indicated by the lower
compressibility values in the pretransition region.
Such a reduction in mAb-B pretransition dynamics
was also observed by increases in the magnitude of the
REES in the presence of stabilizing concentrations of
sucrose. The increase in magnitude of the red-edge ef-
fect suggests that solvent relaxation contributions in
the environment around the Trp residues in mAb-B
were reduced. This in turn suggests that the envi-
ronment around the Trp residues is rigidified in the
presence of sucrose, especially at lower temperatures,
potentially due to a reduction in the internal dynam-
ics of mAb-B. The inability of both arginine and su-
crose to affect the magnitude of red-edge effects for
mAb-A at different temperatures suggests that excip-
ients did not alter the dynamics of mAb-A in the im-
mediate environment of aromatic residues. Since the
conformational fluctuations in mAb-A and mAb-B are
differently coupled to solvent fluctuations, it is possi-
ble that excipients that modulate solvation properties
can influence the pretransition dynamics of proteins
whose conformational fluctuations are strongly cou-
pled to the surrounding solvent.
CONCLUSIONS
The two IgG1 mAbs used in this study exhibited
notable differences in their conformational stability
and dynamic properties as a function of pH, tem-
perature, and presence of excipients. Using a vari-
ety of biophysical techniques, these differences could
be summarized using an EPD approach. Sucrose and
arginine were found to influence the conformational
stability of mAb-A and mAb-B to varying extents at
different effective concentrations. These two excipi-
ents did not significantly influence the pretransition
dynamics of mAb-A as determined by both HR-US and
REES studies. In contrast, for mAb-B, the effects of
stabilizing concentrations of sucrose on the compress-
ibility and the magnitude of red-edge effects suggest
that both the internal dynamics and the surrounding
solvent dynamics of mAb-B are influenced by sucrose
with this particular IgG1. These results show that
formulation components can have unique effects on
the dynamics of individual proteins within a single
IgG1 subclass, especially in the pretransition region.
Although the conformational stability and dynamics
results in this work highlight some of the differences
in physical behavior between two similar IgG mAbs,
the molecular origin of these differences remains to be
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 9, SEPTEMBER 2012
3076 THAKKAR ET AL.
elucidated. Such distinct effects of excipients on dy-
namics of therapeutic proteins may thus profoundly
influence the long-term storage stability and effi-
cacy of biopharmaceutical products and thus require
evaluation during preformulation and formulation
activities.
ACKNOWLEDGMENTS
The authors would like to t hank MedImmune for pro-
viding IgG1 molecules and financial support for this
work. Dr. Jae Hyun Kim is acknowledged for his
help in automating the data processing steps for
REES studies and for constructing EPDs. We thank
Dr. Hardeep Samra from Formulation Sciences De-
partment at MedImmune for helpful discussions and
critical review of this manuscript.
REFERENCES
1. Karplus M, McCammon JA. 1983. Dynamics of proteins:
Elements and function. Annu Rev Biochem 52:263–300.
2. Kleckner IR, Foster MP. 2011. An introduction to NMR-based
approaches for measuring protein dynamics. Biochim Biophys
Acta 1814(8):942–968.
3. Boehr DD, Dyson HJ, Wright PE. 2006. An NMR perspective
on enzyme dynamics. Chem Rev 106(8):3055–3079.
4. Loria JP, Berlow RB, Watt ED. 2008. Characterization of en-
zyme motions by solution NMR relaxation dispersion. Acc
Chem Res 41(2):214–221.
5. Smock RG, Gierasch LM. 2009. Sending signals dynamically.
Science 324(5924):198–203.
6. Gunasekaran K, Ma B, Nussinov R. 2004. Is allostery an in-
trinsic property of all dynamic proteins? Proteins 57(3):433–
443.
7. Rejto PA, Verkhivker GM. 1996. Unraveling principles of
lead discovery: From unfrustrated energy landscapes to novel
molecular anchors. Proc Natl Acad Sci U S A 93(17):8945–
8950.
8. Kumar S, Ma B, Tsai CJ, Sinha N, Nussinov R. 2000. Fold-
ing and binding cascades: Dynamic landscapes and population
shifts. Protein Sci 9(1):10–19.
9. Teilum K, Olsen JG, Kragelund BB. 2009. Functional as-
pects o f protein flexibility. Cell Mol Life Sci 66(14):2231–
2247.
10. Kamerzell TJ, Middaugh CR. 2008. The complex inter-
relationships between protein flexibility and stability. J Pharm
Sci 97(9):3494–3517.
11. Jaenicke R, Bohm G. 1998. The stability of proteins in extreme
environments. Curr Opin Struct Biol 8(6):738–748.
12. Jaenicke R. 2000. Do ultrastable proteins from hyperther-
mophiles have high or low conformational rigidity? Proc Natl
Acad Sci U S A 97(7):2962–2964.
13. Hernandez G, Jenney FE, Jr., Adams MW, LeMaster DM.
2000. Millisecond time scale conformational flexibility in a
hyperthermophile protein at ambient temperature. P roc Natl
Acad Sci U S A 97(7):3166–3170.
14. Yesylevskyy SO, Kharkyanen VN, Demchenko AP. 2006.
The change of protein intradomain mobility on ligand bind-
ing: Is it a commonly observed phenomenon? Biophys J
91(8):3002–3013.
15. Roca M, Liu H, Messer B, Warshel A. 2007. On the relation-
ship between thermal stability and catalytic power of enzymes.
Biochemistry 46(51):15076–15088.
16. Fenimore PW, Frauenfelder H, McMahon BH, Parak FG.
2002. Slaving: Solvent fluctuations dominate protein dynam-
ics and functions. Proc Natl Acad Sci U S A 99(25):16047–
16051.
17. Frauenfelder H, Fenimore PW, McMahon BH. 2002. Hydra-
tion, slaving and protein function. Biophys Chem 98(1–2):35–
48.
18. Fenimore PW, Frauenfelder H, McMahon BH, Young RD.
2004. Bulk-solvent and hydration-shell fluctuations, simi-
lar to alpha- and beta-fluctuations in glasses, control pro-
tein motions and functions. Proc Natl Acad Sci U S A
101(40):14408–14413.
19. Frauenfelder H, Fenimore PW, Chen G, McMahon BH. 2006.
Protein folding is slaved to solvent motions. P roc Natl Acad
Sci U S A 103(42):15469–15472.
20. Kamerzell TJ, Ramsey JD, Middaugh CR. 2008. Immunoglob-
ulin dynamics, conformational fluctuations, and nonlinear
elasticity and their effects on stability. J Phys Chem B
112(10):3240–3250.
21. Priev A, Almagor A, Yedgar S, Gavish B. 1996. Glycerol de-
creases the volume and compressibility of protein interior. Bio-
chemistry 35(7):2061–2066.
22. Jiang JS, Brunger AT. 1994. Protein hydration observed
by X-ray diffraction. Solvation properties of penicillopepsin
and neuraminidase crystal structures. J Mol Biol 243(1):100–
115.
23. Savage H, Wlodawer A. 1986. Determination of water struc-
ture around biomolecules using X-ray and neutron diffraction
methods. Methods Enzymol 127:162–183.
24. Belton PS. 1994. NMR studies of protein hydration. Prog Bio-
phys Mol Biol 61(1):61–79.
25. Denisov VP, Jonsson BH, Halle B. 1999. Hydration of
denatured and molten globule proteins. Nat Struct Biol
6(3):253–260.
26. Halle B, Andersson T, Forsen S, Lindman B. 1981. Protein
hydration from water oxygen-17 magnetic relaxation. J Am
Chem Soc 103(3):500–508.
27. Kamerzell TJ, Middaugh CR. 2007. Two-dimensional corre-
lation spectroscopy reveals coupled immunoglobulin regions
of differential flexibility that influence stability. Biochemistry
46(34):9762–9773.
28. Sarvazyan AP, Chalikian TV. 1991. Theoretical analysis of an
ultrasonic interferometer for precise measurements at high
pressures. Ultrasonics 29(2):119–124.
29. Chalikian TV, Plum GE, Sarvazyan AP, Breslauer KJ. 1994.
Influence of drug binding on DNA hydration: Acoustic and
densimetric characterizations of netropsin binding to the
poly(dAdT).poly(dAdT) and poly(dA).poly(dT) duplexes and
the poly(dT).poly(dA).poly(dT) triplex at 25 degrees C. Bio-
chemistry 33(29):8629–8640.
30. Buckin VA, Kankiya BI, Bulichov NV, Lebedev AV, Gukovsky
I, Chuprina VP, Sarvazyan AP, Williams AR. 1989. Measure-
ment of anomalously high hydration of (dA)n.(dT)n double he-
lices in dilute solution. Nature 340(6231):321–322.
31. Kharakoz DP, Sarvazyan AP. 1993. Hydrational and in-
trinsic compressibilities of globular proteins. Biopolymers
33(1):11–26.
32. Cameron DL, Jakus J, Pauleta SR, Pettigrew GW, Cooper
A. 2010. Pressure perturbation calorimetry and the thermo-
dynamics of noncovalent interactions in water: Comparison
of protein–protein, protein–ligand, and cyclodextrin–adaman-
tane complexes. J Phys Chem B 114(49):16228–16235.
33. Mitra L, Rouget JB, Garcia-Moreno B, Royer CA, Winter
R. 2008. Towards a quantitative understanding of pro-
tein hydration and volumetric properties. Chemphyschem
9(18):2715–2721.
34. Mitra L, Smolin N, Ravindra R, Royer C, Winter R. 2006.
Pressure perturbation calorimetric studies of the solvation
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 9, SEPTEMBER 2012 DOI 10.1002/jps
EXCIPIENTS DIFFERENTIALLY INFLUENCE THE CONFORMATIONAL STABILITY 3077
properties and the thermal unfolding of proteins in solution—
Experiments and theoretical interpretation. Phys Chem Chem
Phys 8(11):1249–1265.
35. Zhai Y, Okoro L, Cooper A, Winter R. 2011. Applications
of pressure perturbation calorimetry in biophysical studies.
Biophys Chem 156(1):13–23.
36. Mach H, Volkin DB, Burke CJ, Middaugh CR. 1995. Ultravio-
let absorption spectroscopy. Methods Mol Biol 40:91–114.
37. Wiethoff CM, Russell Middaugh C. 2001. Light-scattering
techniques for characterization of synthetic gene therapy
vectors. Methods Mol Med 65:349–376.
38. Esfandiary R, Hunjan JS, Lushington GH, Joshi SB,
Middaugh CR. 2009. Temperature dependent 2nd derivative
absorbance spectroscopy of aromatic amino acids as a probe of
protein dynamics. Protein Sci 18(12):2603–2614.
39. Lucas LH, Ersoy BA, Kueltzo L A, Joshi SB, Brandau DT,
Thyagarajapuram N, Peek LJ, Middaugh CR. 2006. Probing
protein structure and dynamics by second-derivative ultravio-
let absorption analysis of cation–{pi} interactions. Protein Sci
15(10):2228–2243.
40. Bond MD, Panek ME, Zhang Z, Wang D, Mehndiratta P, Zhao
H, Gunton K, Ni A, Nedved ML, Burman S, Volkin DB. 2010.
Evaluation of a dual-wavelength size exclusion HPLC method
with improved sensitivity to detect protein aggregates and its
use to better characterize degradation pathways of an IgG1
monoclonal antibody. J Pharm Sci 99(6):2582–2597.
41. Cardamone M, Puri NK. 1992. Spectrofluorimetric assess-
ment of the surface hydrophobicity of proteins. Biochem J 282
(Pt 2):589–593.
42. LeVine H, 3rd. 1993. Thioflavine T interaction with synthetic
Alzheimer’s disease beta-amyloid peptides: Detection of amy-
loid aggregation in solution. Protein Sci 2(3):404–410.
43. Sreerama N, Manning MC, Powers ME, Zhang JX, Goldenberg
DP, Woody RW. 1999. Tyrosine, phenylalanine, and disulfide
contributions to the circular dichroism of proteins: Circular
dichroism spectra of wild-type and mutant bovine pancreatic
trypsin inhibitor. Biochemistry 38(33):10814–10822.
44. Stryer L. 1965. The interaction of a naphthalene dye with
apomyoglobin and apohemoglobin. A fluorescent probe of non-
polar binding sites. J Mol Biol 13(2):482–495.
45. Kueltzo LA, Ersoy B, Ralston JP, Middaugh CR. 2003. Deriva-
tive absorbance spectroscopy and protein phase diagrams as
tools for comprehensive protein characterization: A bGCSF
case study. J Pharm Sci 92(9):1805–1820.
46. Maddux N R, Joshi SB, Volkin DB, Ralston JP, Middaugh
CR. 2011. Multidimensional methods for the formulation of
biopharmaceuticals and vaccines. J Pharm Sci.
47. Ramsey JD, Gill ML, Kamerzell TJ, Price ES, Joshi SB, Bishop
SM, Oliver CN, Middaugh CR. 2009. Using empirical phase
diagrams to understand the role of intramolecular dynamics
in immunoglobulin G stability. J Pharm Sci 98(7):2432–2447.
48. Hanson DC, Yguerabide J, Schumaker VN. 1981. Segmental
flexibility of immunoglobulin G antibody molecules in solution:
A new interpretation. Biochemistry 20(24):6842–6852.
49. Boehm MK, Woof JM, Kerr MA, Perkins SJ. 1999. The Fab and
Fc fragments of IgA1 exhibit a different arrangement from that
in IgG: A study by X-ray and neutron solution scattering and
homology modelling. J Mol Biol 286(5):1421–1447.
50. Dangl JL, Wensel TG, Morrison SL, Stryer L, Herzenberg LA,
Oi VT. 1988. Segmental flexibility and complement fixation
of genetically engineered chimeric human, rabbit and mouse
antibodies. EMBO J 7(7):1989–1994.
51. Hanson DC. 1985. Some theoretical considerations regarding
the effects of steric hindrance and intrinsic global coupling on
the flexibility of Fc-anchored immunoglobulins. Mol Immunol
22(3):245–250.
52. Kawata Y, Hamaguchi K. 1991. Use of fluorescence energy
transfer to characterize the compactness of the constant frag-
ment of an immunoglobulin light chain in the early stage of
folding. Biochemistry 30(18):4367–4373.
53. Kim H, Matsunaga C, Yoshino A, Kato K, Arata Y. 1994. Dy-
namical structure of the hinge region of immunoglobulin G as
studied by
13
C nuclear magnetic resonance spectroscopy. J Mol
Biol 236(1):300–309.
54. Takahashi H, Suzuki E, Shimada I, Arata Y. 1992. Dy-
namical structure of the antibody combining site as studied
by
1
H–
15
N shift correlation NMR spectroscopy. Biochemistry
31(9):2464–2468.
55. Kozlowski S, Swann P. 2006. Current and future issues in
the manufacturing and development of monoclonal antibodies.
Adv Drug Deliv Rev 58(5–6):707–722.
56. Shire SJ. 2009. Formulation and manufacturability of biolog-
ics. Curr Opin Biotechnol 20(6):708–714.
57. Wang W, Singh S, Zeng DL, King K, Nema S. 2007. An-
tibody structure, instability, and formulation. J Pharm Sci
96(1):1–26.
58. Kamerzell TJ, Unruh JR, Johnson CK, Middaugh CR. 2006.
Conformational flexibility, hydration and state parameter fluc-
tuations of fibroblast growth factor-10: Effects of ligand bind-
ing. Biochemistry 45(51):15288–15300.
59. Shiio H. 1958. Ultrasonic interferometer measurements of
the amount of bound water. Saccharides. J Am Chem Soc
80(1):70–73.
60. Tamura Y, Gekko K. 1995. Compactness of thermally and
chemically denatured ribonuclease A as revealed by volume
and compressibility. Biochemistry 34(6):1878–1884.
61. Chalikian TV, Sarvazyan AP, Breslauer KJ. 1994. Hydration
and partial compressibility of biological compounds. Biophys
Chem 51(2–3):89–107; discussion 107–109.
62. Demchenko AP. 1982. On the nanosecond mobility in pro-
teins. Edge excitation fluorescence red shift of protein-
bound 2-(p-toluidinylnaphthalene)-6-sulfonate. Biophys Chem
15(2):101–109.
63. Demchenko AP. 2002. The red-edge effects: 30 years of explo-
ration. Luminescence 17(1):19–42.
64. Harn N, Allan C, Oliver C, Middaugh CR. 2007. Highly concen-
trated monoclonal antibody solutions: Direct analysis of physi-
cal structure and thermal stability. J P harm Sci 96(3):532–546.
65. Garber E, Demarest SJ. 2007. A broad range of Fab stabili-
ties within a host of therapeutic IgGs. Biochem Biophys Res
Commun 355(3):751–757.
66. Ionescu RM, Vlasak J, Price C, Kirchmeier M. 2008. Contri-
bution of variable domains to the stability of humanized IgG1
monoclonal antibodies. J Pharm Sci 97(4):1414–1426.
67. Tischenko VM, Abramov VM, Zav’yalov VP. 1998. Investiga-
tion of the cooperative structure of Fc fragments from myeloma
immunoglobulin G. Biochemistry 37(16):5576–5581.
68. Baynes BM, Wang DI, Trout BL. 2005. Role of arginine in
the stabilization of proteins against aggregation. Biochemistry
44(12):4919–4925.
DOI 10.1002/jps JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 101, NO. 9, SEPTEMBER 2012
    • "Understanding protein–excipient interactions is important to optimise protein formulations. These interactions can change the surface properties of the proteins and may change their secondary and tertiary structure [1,2], which could be detrimental for protein activity. Since proteins are administered parentally, the number of excipients that are clinically approved is limited. "
    [Show abstract] [Hide abstract] ABSTRACT: Protein formulation development relies on the selection of excipients that inhibit protein-protein interactions preventing aggregation. Empirical strategies involve screening many excipient and buffer combinations using force degradation studies. Such methods do not readily provide information on intermolecular interactions responsible for the protective effects of excipients. This study describes a molecular docking approach to screen and rank interactions allowing for the identification of protein-excipient hotspots to aid in the selection of excipients to be experimentally screened. Previously published work with Drosophila Su(dx) was used to develop and validate the computational methodology, which was then used to determine the formulation hotspots for Fab A33. Commonly used excipients were examined and compared to the regions in Fab A33 prone to protein-protein interactions that could lead to aggregation. This approach could provide information on a molecular level about the protective interactions of excipients in protein formulations to aid the more rational development of future formulations.
    Article · Jun 2016
    • "Identity, heterogeneity, impurity content and activity of therapeutic protein products are thoroughly investigated by using a wide range of chromatographic methods, such as reversed-phase liquid chromatography, sizeexclusion chromatography and ion-exchange chromatography. SEC is commonly used to determine size-related heterogene- ity [17,47,52,53]. Components smaller than the intact protein are usually the results of enzymatic or non-enzymatic cleavage and incomplete formation of mismatched disulfide bridges. "
    [Show abstract] [Hide abstract] ABSTRACT: Various liquid chromatographic techniques are considered standard analytical methods in proteins characterization. These methods provide essential information for drug approval, for biological and life sciences. On the other hand, there are some issues and challenges which have to be taken into account when analyzing these biopharmaceuticals. The aim of this review to summarize the most recent knowledge relating to the following topics: i) sample stability and complexity ii) adsorption problems: instrument inertness iii) adsorption problems: recovery from the stationary phase and iv) challenges in method development. This information is supposed to help practicing chromatographers in the emerging field of therapeutic protein chromatography.
    Article · Apr 2016
    • "Above 100 mM it begins to destabilize all three proteins in a similar manner to the guanidinium, although not as strongly (Fig. 2c). Arginine's mechanism of action is thought to be complex, as demonstrated by varying effects on protein stability and concentration-dependent actions (Thakkar et al., 2012; Falconer et al., 2011). However, current suggestions of weak transient interactions at low concentrations (Lim et al., 2009) and preferential exclusion due to increase in surface tension and self-association or 'stacking' of arginine molecules at higher concentrations (Shukla and Trout 2011; Das et al., 2007; Kita et al., 1994; Vondrášek et al., 2009) cannot be used to explain the trends presented in this paper as this would result in a stronger stabilization at higher concentrations. "
    [Show abstract] [Hide abstract] ABSTRACT: Three distinct interactions between the amino acid arginine and a protein explain arginine's ability to modulate the thermal stability of proteins. Arginine's effect on the protein unfolding behaves like the sum of its constituent parts, glycine and the guanidinium ion. The authors propose that glycine can affect the thermal stability of a protein in two ways: (1) direct interaction with the charged side chains and/or the peptide backbone of the protein which is observed at low concentrations and (2) competition for water between the unfolding protein and the cosolute increasing the energy required to hydrate the unfolding protein. The guanidinium ion acts by (3) direct interaction with apolar regions exposed during unfolding reducing the energy required to hydrate the unfolding protein. Copyright © 2015. Published by Elsevier B.V.
    Article · Mar 2015
Show more