eBook Series

2015
Most Read
TopArticles
5
FORMULATION, FILL-FINISH
Evaluating Freeze-Thaw Processes in Biopharmaceutical Development

SINGLE-USE PRODUCTION
A Novel Seed-Train Process

MONOCLONAL ANTIBODIES
Hamster Phospholipase B-Like 2 (PLBL2)

BIOSIMILARS
Rapid Development and Scale-Up of Biosimilar Trastuzumab

QUALITY CONTROL
Fundamental Strategies for Viral Clearance
B i o P r o c e s s TECHNICAL
Evaluating Freeze–Thaw Processes
in Biopharmaceutical Development
Small-Scale Study Designs
Manasi Puri, Sorina Morar-Mitrica, George Crotts, and Douglas Nesta
R
egulations mandate that
biopharmaceutical product
quality be controlled
throughout manufacturing,
storage, transportation, and delivery
to patients (1). Operations often
include freezing and thawing of a
bulk drug substance, dilution of
that purified substance to a target
concentration, filtration, filling into
a selected container–closure system,
additional processing (e.g.,
lyophilization), inspection,
packaging, storage, transport, and
delivery (2).
Freezing is a common processing
step used to maintain stability and
quality of a drug substance during
development and production of
biopharmaceutical products. It is
generally agreed upon that freezing
drug substance maximizes
productivity and reduces overall
production costs by decoupling bulk
solution manufacture and storage
steps from final product
Product Focus:  Monoclonal
antibodies
Process Focus:  Downstream
processing, formulation, fill–finish
Who Should Read:  Process/product
development, manufacturing, and
analytical personnel
Keywords:  Drug product, drug
substance, stability, freezing, thawing,
single-use technology
Level:  Intermediate
BioProcess International 13(1) January 2015
manufacture. Freezing provides
f lexibility and cost savings by
enabling batch processing: Large
volumes of an expensive biological
drug substance can be frozen in
batches to allow the drug product
to be manufactured based on real-
time commercial or clinical
demands. Freezing a drug substance
before drug product manufacture
also reduces microbial growth risk
and eliminates mechanical
(agitation) stresses during
transportation.
Most important, this allows for a
longer shelf life. Freezing decelerates
chemical degradation (because
reaction kinetics depend directly on
temperature) and physical
degradation (immobilizing drug
substance in a frozen matrix limits
protein–protein interactions) (3–13).
Furthermore, application of freeze–
thaw processes extends to storage of
process intermediates, allowing for
longer hold steps between
manufacturing operations. Although
freezing is considered to be a
conservative approach, frozen
storage is one of the most efficient
and reliable methods to minimize
protein interactions with container–
closures and extend the shelf life of
biological products in solution-based
formulations (13).
Drug substance (typically stored
in plastic bottles or single-use plastic
bags) is exposed to various interfaces
on contact with container–closure
systems. Likewise, drug product
often is filled into multiple-
component container–closure
systems, such as vials or prefilled
syringes with rubber stopper
closures (2). Interactions with
container surfaces (e.g., glass or
polymers) and other constituents
(e.g., silicone oil) can significantly
inf luence drug substance and drug
product stability. Such interactions
may be further exacerbated with
mechanical stresses such as agitation
during transportation. Freezing can
limit exposure of a protein to those
interfaces and prevent shipping- and
storage-related solution instabilities.
Overall, freezing can be leveraged
to maximize protein stability
throughout a supply chain. However,
additional controls may be needed to
limit the number of freeze–thaw
cycles that occur at clinics,
pharmacies, or depot sites to that
supported by available stability data.
Doing so ensures that product
quality is not compromised.
Freeze –Thaw R elated
Process Challenges
Freezing and thawing a
biopharmaceutical product may
change the chemical and physical
properties of the product solution.
That in turn can stress proteins,
may irreversibly denature complex
macromolecular structure, and
could alter their stability (7, 13).
Because freeze–thaw process
parameters have been identified to
play a significant role in product
quality through alteration of a
protein’s microenvironment, gaining
product- and process-specific
knowledge is critical to using
freeze–thaw applications (11, 14, 15).
Freezing and thawing can occur
at varying rates, especially when
passively performed (at an
uncontrolled rate). Those rates are
known to affect protein stability.
No standard and agreed definition
of freezing or thawing rates has yet
been adopted across the industry
(11, 14). Depending on the
equipment used, freezing rates may
be slow (<1 °C/min), intermediate
(1–10 °C/min), or rapid (10–900 °C/
min).
For example, fast freezing rates
are achieved by immersing samples
in subzero solvents or baths
consisting of alcohols or liquid
nitrogen (80–900 °C/min).
Thawing is also performed at
various rates. Like freezing,
thawing can be slow (1–5 °C/min),
or intermediate
(>5 °C/min). It is not always feasible
to monitor temperature profiles
during freezing and thawing.
Therefore, in design of a systematic
small-scale study, it is important to
incorporate an evaluation of
different freezing and thawing rates
(16, 17).
Slow freezing rates can result in
cryoconcentration, in which
proteins and excipients form
concentration gradients near the
freeze front and get excluded from
the ice–liquid interface (2). This
can further lead to pH shifts and
phase separation among those
components, resulting in protein
structural damage (2, 14). Exposure
to concentrated solutes — due to
water crystallization — also can
result in a loss of a protein's
thermodynamic stability, leading to
unfolding events and eventually
causing aggregation.
Fast freezing rates lead to smaller
ice-crystal formation, which
exposes proteins to a large ice–
liquid interface. Concentration and
adsorption of proteins at the surface
of ice crystals along the ice–liquid
interface can result in their partial
unfolding, increased aggregation,
and decreased biological activity (17,
18). One control strategy is the use
of formulation excipients that act as
stabilizers (e.g., surfactants and
sugar cryoprotectants) to reduce the
level of protein denaturation
induced upon exposure to ice-liquid
interfaces (18).
Of further concern is the fact
that fast freezing can entrap air in
the ice. When released during
thawing, the entrapped air can
denature proteins as air–liquid
interfaces form (19, 20).
Correspondingly, slow thawing
rates can result in ice
recrystallization, and associated
shear stress can damage proteins
(21). Ideally, rapid freezing and
thawing are usually preferred for
protein stability (22).
Other Concerns: Published
literature offers no consistent
terminology or parameters to
describe and execute various rates of
freezing and thawing. Freeze–thaw
studies in traditional stability
programs use passive freezing and
thawing (19). Such procedures
emulate a freezing process in which
a drug substance or drug product is
simply placed on a shelf in a freezer
storage chamber and permitted to
cool and freeze at an uncontrolled
rate. Temperatures can vary among
different commercial freezer units
used for storage, partly due to
differences in their size, cooling
capacity, and loads. Even within the
same freezer chamber, the freezing
rate experienced by a protein
solution could vary depending on
the freezer load and the container’s
position within the chamber.
Passive freezing has been shown to
result in protein aggregation (18).
It has been reported that active
control of freezing rates can reduce
stresses that cause protein structural
and functional losses. Freezing rates
can be highly controlled at
predetermined levels; however, such
an approach requires highly
specialized and expensive
equipment (including freeze dryers
or commercially available controlled
freeze–thaw systems). Actively
controlled thawing also involves
specialized equipment such as
programmable water baths and
freeze dryers. In contrast, passive
thawing can be achieved, for
example, by placing a frozen
container at room temperature or at
2–8 °C. Thus, the need to fully
control freezing and thawing rates
should be evaluated carefully.
Freeze–thaw studies are the most
direct approaches to evaluating the
impact of freezing and thawing
rates and to defining process
conditions (14).
S mall-S cale D esigns
of Freeze –Thaw Studies
Early in development, material
availability is usually limited. This
creates a need for reliable small-
scale models for various freeze–thaw
processes that mimic, as closely as
possible, the processes and
conditions that will occur at large
scale (2). Small-scale models must
be designed or selected that are
predictive of potential large-scale
freeze–thaw issues.
Figure 1 illustrates a systematic
approach to designing such studies
for evaluation of freezing and
thawing parameters as well as
protein stability following freeze–
thaw stress. In general, study design
starts with selection of a
representative formulation and
small-scale container–closure, the
number of freeze–thaw cycles
needed for large-scale manufacture,
freezing and thawing temperatures
and rates, and equipment and
protocols. Study design is followed
by execution of a protocol. Finally,
long-term effects of freeze–thaw
January 2015 13(1) BioProcess International
stress are evaluated through a stability study evaluating
protein attributes under different storage conditions.
Design of small-scale studies is based on the most
representative freeze–thaw conditions to be used at a
manufacturing site (8). Product stability depends on
selected formulation and container–closure materials.
Thus, freeze–thaw evaluation first must include
assessment of a product’s compatibility with the contact
material of the container–closure system during
freezing and thawing. Second, special attention must
Figure 1:  Small-scale study design considerations for evaluating various
freezing and thawing process parameters
Drug Product/Substance
Select representative small-scale container
(materials of construction and fill volume).
Select relevant number of freeze−thaw cycles
aligned with production scale.
Select representative freezing and thawing
temperatures and rates based on
process knowledge.
Select laboratory equipment to execute study.
Execute study with appropriate stress protocol.
Perform active Perform passive
(controlled rates) (uncontrolled rates)
freeze−thaw freeze−thaw
protocol. protocol.
Test samples immediately after stress and/or
place on long-term stability testing
according to stability protocol.
Perform analytical and biophysical testing.
Evaluate effects of freezing and thawing rates
on product quality.
BioProcess International 13(1) January 2015
be paid to selecting appropriate scaled-down container–
closure systems or configurations.
For laboratory-scale studies, it is important to
choose an appropriate fill volume with the
representative head space or surface-area-to-volume
exposure. Scale-down is based on an attempt to
maintain the ratio of surface area to volume similar to
that at large scale. However, because of significantly
different path lengths, cooling and heating rates
encountered during freezing and thawing are generally
much slower at large scale. Thus, process times for
large-scale operations are longer than for small-scale
studies, particularly with passive freezing and thawing
processes. As a result, large-scale cryoconcentration
becomes an important factor governing product quality.
It can occur at small-scale too but is magnified by bulk
freezing in which the extent of freeze concentration
can be directly correlated to solution volume (3).
Moreover, it is not always feasible to monitor
temperature profiles during freezing and thawing. As
stated above, in designing a systematic small-scale
study, it is important to evaluate different freezing and
thawing rates under both active-control and passive
conditions, then determine whether those rates
constitute a critical process parameter (CPP) for
product quality (14, 23).
Our focus herein resides primarily on the impact of
freezing and thawing rates on drug substance and drug
product stability in representative small-scale
container–closure systems. In addition, our small-scale
case studies were designed with an emphasis on
adopting a robust approach to understanding
interactions between formulations and process
conditions. Overall, these studies systematically and
comprehensively evaluate the impact of relevant
processing parameters (freezing and thawing rates),
formulation parameters (MAb concentration, pH,
excipient levels), and container–closure systems on
short- and long-term MAb storage stability. In these
four case studies, we assessed the impact of varying
freezing and thawing rates on the stability of
• a low-concentration MAb-A drug product in glass
vials
• a high-concentration MAb-B drug product in glass
vials and drug substance in plastic bottles
• a MAb-C drug substance in single-use bags.
Here we show that understanding of a product’s
robustness to freeze–thaw stress can be achieved using
carefully designed scale-down models for freezing and
thawing.
C ase Study 1
A Controlled Freeze–Thaw Study of MAb-A at a Low
Concentration (0.2 mg/mL): MAb-A was formulated as
a refrigerated liquid at low concentration (0.2 mg/mL)
for intravenous (IV) administration. Early clinical
in-use stability studies had revealed significant
adsorption onto IV bags and giving sets upon addition

of the drug product to diluent in
the bags due to the resulting low
dosing concentration (0.002 mg/
mL). To overcome significant
protein loss to such binding, a
surfactant (polysorbate 80, PS80)
was introduced to the product
formulation at a concentration of
0.1% (w/v). Although the addition
of PS80 to the formulation
overcame the adsorption challenges
and provided protection from
mechanical stress, it also accelerated
chemical degradation of the MAb,
in particular by oxidation.
Literature shows that degradation
products generated through the
oxidation-induced degradation
pathway lead to accelerated MAb
deamidation, fragmentation, and
aggregation (24). Similar behavior
was observed here. Therefore,
oxidation of MAb-A needed to be
controlled to maintain product
quality.
This oxidation challenge was
identified shortly before
commencement of phase 3 clinical
trials. A multicomponent control
strategy was devised to overcome
the identified surfactant-induced
chemical degradation and allow
on-time start of clinical trials and
submission of a licensing
application. A key element of the
strategy entailed changing the
proposed long-term storage
conditions for this drug product
from refrigerated to frozen, which
could sufficiently inhibit the
observed chemical degradation.
However, before phase 3 supplies
were manufactured, it was also
imperative to assess the impact of
process parameters (e.g., freeze–
thaw rates) on drug product quality
and stability.
The experimental plan for a
small-scale, freeze–thaw study of
MAb-A consisted of a statistical
combination of two factors at two
levels each. Factor 1 was the type of
stress (freeze or thaw), and factor 2
was the stress rate (slow or fast).
Figure 2 shows the overall study
design, which consisted of four
combinations of stress and stress
rates: the fast-freeze, fast-thaw,
Figure 2 (Case Study 1):  Low-
Fast Thaw:
concentration MAb-A solutions in
glass vials were subjected to two
Thermocirculating
Slow Freeze: From water bath set at 25 °C
freeze–thaw cycles. Samples were
25 °C to −40 °C
either fast-frozen in a dry ice and (freeze dryer)
acetone slurry (active fast freeze
protocol) or slow-frozen at a rate of
0.1 °C/min (active slow freeze
protocol) using a laboratory freeze-
dryer. Frozen samples were
subsequently thawed in a water bath
at ambient temperature (active fast
thaw protocol) or thawed at the rate Fast Freeze:
1/1 dry ice and
of 0.1 °C/min (active slow thaw
protocol) in a laboratory freeze dryer. acetone slurry Slow Thaw: From
−40 °C to 25 °C
(freeze
Active Freezing Active
Rate Thawing Rate dryer)
Fast freeze Fast thaw
Slow freeze Slow thaw
Fast freeze Slow thaw
Slow freeze Fast thaw
Figure 3 (Case Study 1):  Summary results of in solution observed by size-
%oxidation at methionine (Met) residues exclusion chromatography (SEC) or
following two actively controlled freeze–thaw ProteinSimple microf low imaging
cycles for MAb-A; refer to Figure 2 for study
setup and sample description.
(MFI) analyses, respectively. No
significant changes were observed
10
9
in general appearance (GA), pH, or
% MAb A Oxidation
Met 254
8 absorbance at 280 nm (A280)
at Met Residues
Met 430
Met 48
7 analyses for all samples.
6
5 This small-scale study design
4 permitted rapid evaluation of the
3 impact of extreme freezing and
2
1 thawing rates on MAb-A quality. It
0
Initial Fast Fast Slow Slow supported implementation of
Fast Slow Fast Slow freezing and thawing processes
Freeze−Thaw during manufacture and storage of
this product in the selected
container–closure system. Study
slow-freeze, and slow-thaw results verified that control of two
protocols described. All solutions significant process parameters —
were tested in glass vials (the drug freezing and thawing rates — would
product container–closure) and not be required during handling of
subjected to two freeze–thaw cycles. this product.
As Figure 3 shows, MAb-A
samples subjected to two controlled C ase Study 2
freeze–thaw cycles experienced no Evaluation of MAb-B Stability at 200
increase in oxidation over the initial mg/mL After Short-Term Exposure to
set of samples. Deamidation Subzero Temperatures Above Tg ’:
(previously noted as a common Protein concentration and container
degradation pathway for this MAb) type can significantly affect MAb
also was not detected. Fluorescence behavior in a frozen state. So can
intensities did not change, storage temperature in relation to
indicating maintenance of the the glass transition temperature
overall tertiary structure and (Tg′), potentially causing irreversible
packing of the MAb when frozen damage such as loss of structure
and thawed at different rates. No through partial unfolding and
significant changes to the aggregate adsorption to a container through
or fragment profile were observed, exposure of hydrophobic residues (7,
nor was any increase of particulates 11). Loss of structural integrity can
January 2015 13(1) BioProcess International
Table 1 (Case Study 2):  Summary of a small-scale study designed to expose MAb-B drug
substance and drug product in their representative container–closure systems to various subzero placebo control were filled into
temperatures under active control small-scale high-density
Container 72 Three Six One polyethylene (HDPE) bottles and
Freeze–Thaw Stress MAb-B Sample Type Hours Months Months Year glass vials that were representative
Perform one cycle Drug substance HDPE T T T NT of the proposed commercial
of controlled 200 mg/mL (100 mL)
freezing at 0 °C,
configuration. Active and placebo
Formulation buffer T T T NT
–5 °C, or –15 °C. solutions were cooled or frozen to
Drug product Glass Vial T T T T
Move to long-term
200 mg/mL (3 mL)
0 °C, –5 °C, and –15 °C by placing
2–8 °C storage.
Formulation buffer T T T T
filled containers in a laboratory
bath containing a heat-transfer
T = tested; NT = not tested
f luid. After about 72 hours of
exposure to the target temperature,
Figure 4 (Case Study 2):  Summary results of %monomer and %HMW (high–molecular-weight
samples were moved to long-term
species) by SEC for MAb-B drug product at 200 mg/mL; following one cycle of actively controlled
freeze cycle (72-hour storage at 0 °C, –5 °C, or –15 °C), the product was stored at 2–8 °C for ≤12 storage at 2–8 °C for up to 12
months. Refer to Table 1 for study setup and sample description. months.
100 Stability of the MAb-B solutions
after 72 hours at 0 °C was assessed immediately after
after 72 hours at −5 °C subzero temperature treatment and
after 72 hours at −15 °C again after long-term storage by
98
general appearance, testing pH and
A280, and using SEC, capillary
Monomer (%) MAb-B

isoelectric focusing (cIEF),

96 ProteinSimple MFI, reverse-phase
chromatography (RP-HPLC), and
100
surface plasmon resonance (SPR)
Monomer (%) MAb-B

94 98 potency analysis. Structural

92
90
directly result in aggregate
formation through increased
interaction of partially unfolded
molecules with each other or with
other species upon thawing. That
phenomenon can be enhanced if the
frozen storage temperature is near
or above Tg′ of the formulation (25).
Stability studies revealed that
MAb-B failed acceptance criteria
for aggregates when frozen at
–20 °C (>Tg′) for a month. A
significant increase in subvisible
particles also was observed. Not
only were MAb-B drug substance
and product both formulated at very
high concentration (200 mg/mL),
but also they used a buffer system
with no cryoprotectant.
Temperature excursions below 2 °C
thus were strictly prohibited for
early phase clinical supplies.
96
94
92
90
0 3 6 9 12
Storage Time at 2−8 °C (months)
0 3 6 9 12
Storage Time at 2−8 °C (months)
However, those restrictions added
significant supply-chain risk
because the project progressed into
later-phase clinical development.
For example, according to internal
shipping temperature logs (data not
shown), temperature excursions
outside 2–8 °C were typically in the
range of –1 °C to –11 °C and tended
to be on the lower end of that
range. Thus, we needed to justify a
more practical temperature
excursion range.
A small-scale freeze–thaw
procedure — together with a long-
term stability study — was
implemented to discharge shipment
and storage excursion risks of short-
term exposure to subzero
temperatures >Tg′. Table 1 displays
the details of the study design.
MAb-B and a corresponding
integrity of the MAb was also
determined by biophysical testing:
f luorescence spectroscopy and
circular dichroism. Figure 4
summarizes SEC stability results.
MAb-B drug product at a
concentration of 200 mg/mL
exhibited no significant changes in
monomer concentration or
aggregate profile when stored for up
to 12 months (one year) at 2–8 °C
after 72-hour incubation at the
three subzero temperatures.
The results from the analytical
and biophysical testing showed that
MAb-B quality and stability in the
selected container–closure system
were not affected by short-term
exposure to temperatures between
–15 °C and 0 °C. These data were
used to justify expansion of the
excursion limits for this project,
which reduced the risk of stock-outs
and clinical delays.
C ase Study 3
Drug Substance Robustness, Freeze–
Thaw, and Long-Term Stability in
Small-Scale System Configuration
for a 100-mg/mL MAb-C:
Commercial-scale production of
drug substance is no longer limited

BioProcess International 13(1) January 2015

 Figure 5 (Case Study 3):  Summary results of Table 2 (Case Study 3):  Summary of a small-scale, passive freeze–thaw study for MAb-C using
%monomer by SEC for MAb-C drug substance disposable 5-mL EVA bags filled at 40% volume with samples representative of formulation robustness
after five cycles of passive freeze–thaw cycles Freeze–Thaw MAb-C MAb-C
from ≤–60 °C to 2–8 °C, followed by six-month Stress Sample Conditions Concentration pH
storage at 2–8 °C; refer to Table 2 for study Perform five a Low MAb concentration, low pH 80 mg/mL 5.8
setup and sample description. cycles of passive b Low MAb concentration, high pH 80 mg/mL 6.2
freezing and
thawing from c High MAb concentration, low pH 120 mg/mL 5.8
100
≤–60 °C to 2–8°C d High MAb concentration, high pH 120 mg/mL 6.2
Monomer (%) MAb-C

98 e Center point (target formulation) 100 mg/mL 6.0

96
Figure 6 (Case Study 4):  A small-scale, controlled freeze–thaw study using single-use 30-mL
Sample A
94 Sample B Celsius bags for MAb-C; solutions were aliquoted into each bag at approximately 66% fill volume.
Sample C The 120-mg/mL samples were subjected to three freeze–thaw cycles in which the rates of freezing
Sample D
92 Sample E and thawing were actively controlled (table below). Samples were either frozen in liquid nitrogen
(active fast freeze protocol, graph below) and then thawed in a water bath at ambient temperature
90
−1 0 1 2 3 4 5 6 7 (active fast thaw protocol), or frozen from 25 °C to −70 °C at a controlled rate of 0.1 °C/min (active
Storage Time (months) slow freeze protocol) and thawed from −70 °C to 25 °C at a controlled rate of 0.1 °C/min (active slow
thaw protocol) by using Celsius S3 system (photo below). After three freeze–thaw cycles, the bags
to rigid geometric containers such were placed at 2–8 °C and 25 °C for long-term storage up to one, three, and six months.
as bottles and carboys. Single-use Sample Set Freezing Rate Thawing Rate
bag systems made of f lexible f Controlled Fast Controlled Fast
polymeric materials offer an g Controlled Slow Controlled Slow
alternative. One such example is the h Controlled Fast Controlled Slow
Celsius FFT (f lexible freeze–thaw) i Controlled Slow Controlled Fast
bag system from Sartorius Stedim j Uncontrolled Uncontrolled
Biotech. The FFT bag uses Celsius
film, of which the product-contact (graph)  Controlled slow freeze-thaw temperature profile
component is ethylene vinyl acetate at 0.1 °C/min using the Celsius® S3 System
(EVA). 30
(photo)  A 30-mL
20
The Celsius bag was identified as
10
Celsius bag
Temperature (°C)

the preferred container for phase 3

0
and commercial supply of MAb-C −10
−20
drug substance formulated at 100
−30
mg/mL. Upon information review
−40
and consultation with the vendor, −50
−60
5-mL Flexboy bags from Sartorius −70
Stedim Biotech were determined to
0 50 100
Time (hours)
150
represent a suitable scale-down
model of commercial sizes in the
Celsius bag system. 2–8 °C storage chamber (passive
The goal of this study was to thawing protocol), each at about
assess robustness of the drug- 12 hours duration.
substance formulation to freeze– MAb-C stability was assessed
thaw stress and subsequent storage after bags were subjected to five
at various conditions in the 5-mL passive freeze–thaw cycles from
bags. Table 2 lists the two ≤–60 °C to 2–8 °C and after storage
formulation robustness factors at ≤–60 °C, –40 °C, –20 °C,
incorporated into this study design: 2–8 °C, and 25 °C for one, three,
MAb concentration (±20% target) and six months. Samples were
and pH range (±0.2 units from analyzed for general appearance,
target). Solutions of MAb-C were pH, and freezing-point osmolality,
aliquoted into each bag at about and by microf luidic chip
40% fill-volume capacity. Samples electrophoresis (Caliper Life
were subjected to five freeze–thaw Sciences), SEC, cIEF, and SPR.
cycles in which the rates of freezing Figure 5 displays SEC %monomer
and thawing were passively results, showing no significant
controlled. Samples were frozen by change in the purity profile for
placing them horizontally in MAb-C drug substance after five
≤–60 °C storage chambers (passive passive freeze–thaw cycles from
freezing protocol) and thawed in a ≤–60 °C to 2–8 °C followed by six-
containing
MAb-C BDS at
120 mg/mL
after being
subjected to
controlled
fast-freezing
protocol using
liquid nitrogen
month storage at 2–8 °C regardless
of sample type.
Results from this study confirm
long-term stability of the MAb-C
product within specified
formulation robustness ranges
following five passive freeze–thaw
cycles. Consequently, no active
freeze–thaw control was
implemented at the drug-substance
manufacturing site. Instead, filled
Celsius bags were frozen by placing
them in a standard ≤–60 °C freezer,
and the substance was thawed at
2–8 °C in a cold room.
C ase Study 4
Active and Passive Freeze–Thaw of
MAb-C at 120 mg/mL in 30-mL
Celsius Bags: Combining actively
controlled freeze–thaw technology

January 2015 13(1) BioProcess International

Table 3 (Case Study 4):  SEC and cIEF results for MAb-C drug substance subjected to three cycles
of active freezing and thawing from –70 °C to 25 °C and stored up to six months at 2–8 °C; refer to
Figure 6 for study setup and sample description.
Size-Exclusion Chromatography
MAb-C
Sample Monomer HMW LMW
Initial 97.5% 1.3% 1.2%
f 96.9% 1.9% 1.3%
g 96.9% 1.9% 1.3%
h 96.7% 2.0% 1.3%
i 96.8% 1.9% 1.3%
j 96.8% 1.9% 1.3%
with single-use bags can greatly
reduce overall cost of goods and
enhance operational f lexibility (26).
The Celsius S3 system from
Sartorius Stedim Biotech is
designed to reproduce active control
of large-scale freezing and thawing
conditions in small-scale EVA bags
(of 30 mL and 100 mL total
volume) (8, 26). This study was
designed to assess whether an
actively controlled freezing and
thawing process using the Celsius
S3 system would provide an
advantage over passive freezing and
thawing for MAb-C. Resulting
data were used to determine
whether to purchase a controlled
freeze–thaw system for the
manufacturing site.
Figure 6 summarizes the study
design and procedure. Single-use
30-mL Celsius bags were used.
Samples were subjected to three
freeze–thaw cycles with different
rates of freezing and thawing
achieved using Celsius S3 system, a
liquid nitrogen bath, and a water
bath. GA, pH, freezing-point
osmolality, microf luidic chip
electrophoresis, SEC, cIEF, SPR,
and PS80 testing were used to test
MAb-C stability under stresses of
extreme freeze–thaw rates followed
by long-term storage (at 2–8 °C and
25 °C for up to six months) in this
small-scale system configuration.
As Table 3 shows, no significant
changes were observed in either the
size (measured using SEC) or
charge (measured using cIEF)
profiles of MAb-C after three
freeze–thaw cycles. A similar
observation was made for cycled
BioProcess International 13(1) January 2015
Capillary Isoelectric Focusing
Main Acidic Basic
57.6% 34.8% 7.6%
57.1% 35.2% 7.7%
57.2% 35.4% 7.5%
56.2% 35.3% 8.5%
56.7% 35.9% 7.5%
57.6% 35.0% 7.6%
samples stored at 2–8 °C for six
months.
The results of this study —
together with those described in
case study 3 above — indicate that
active control of freezing and
thawing rates during manufacture
was not necessary for MAb-C.
Although such results may not be
observed for all MAbs in
development, they were not
unexpected here because the
formulation was specifically
designed to withstand stresses that
can be induced by freezing and
thawing.
D emonstrating D ue D iligence
Freeze–thaw processes are routinely
used to improve storage and
shipping stability in production and
storage of biopharmaceutical
solutions. However, freezing and/or
thawing processes can adversely
affect protein product stability and
quality. For this reason, relevant
studies are needed to assess
formulation robustness to freeze–
thaw stress and container–closure
compatibility with as well as to
define freezing and thawing
parameters for large-scale
production.
Designing freeze–thaw studies at
conditions representative of stresses
expected at large-scale production
is a significant challenge requiring
consideration of multiple factors
such as container–closure material
and size, fill volume, and freezing
and thawing rates. Although small-
and large-scale contact materials
can be matched with relative ease,
the significantly different path
lengths make freezing and thawing
rates extremely difficult to match.
One approach we adopted at small
scale was to show that freezing and
thawing rates do not significantly
inf luence product quality in the
selected formulations. Neither
freezing nor thawing rates (both
significant process parameters)
needed to be controlled during
product handling.
In all four case studies, freeze–
thaw cycles tested at small scale
mimicked the product-specific
system and process for potential
application at large-scale
production. For active control of
freezing and thawing rates, we used
laboratory approaches such as a
freeze dryer, liquid nitrogen, dry-
ice acetone slurry, and a water bath
(in addition to specialized
technologies such as the Celsius S3
freeze–thaw system). These studies
exemplified how small-scale freeze–
thaw approaches can be
implemented in the
biopharmaceutical industry to
confirm container–closure selection,
verify formulation robustness,
support temperature excursion
limits, set process parameters, and
(most important) justify the need
for either active or passive freeze–
thaw procedures and associated
equipment before expensive
manufacturing campaigns begin.
Although our intent in these studies
was not to evaluate concentration
effects, their results illustrate the
success of such designs across a
wide range of concentrations
(0.2 mg/mL MAb-A in case study 1
and 200 mg/mL MAb-B in case
study 2).
Understanding the product
robustness to freeze–thaw stress can
be achieved using carefully designed
scale-down models for freezing and
thawing. Also, we propose that
systematic small-scale designs
similar to those discussed here
could be adopted as predictive,
robust approaches and applied
across the biopharmaceutical
industry to assess the impact of
freeze–thaw process on product
quality and to support
manufacturing, storage, and
handling decisions.
R eferences
1 Ammann C. Stability Studies Needed
to Define the Handling and Transport
Conditions of Sensitive Pharmaceutical or
Biotechnological Products. A APS Pharm.
Sci. Tech. 12(4) 2011: 1264–1275.
2 Rathore N, Rajan RS. Current
Perspectives on Stability of Protein Drug
Products During Formulation, Fill and
Finish Operations. Biotechnol. Prog. 24(3)
2008: 504–514.
3 Kolhe P, et al. Large-Scale Freezing
of Biologics: Understanding Protein and
Solute Concentration Changes in a
Cryovessel — Part 2. BioPharm Int. 23(7)
2010: 40–49.
4 Kolhe P, et al. Large-Scale Freezing
of Biologics: Understanding Protein and
Solute Concentration Changes in a
Cryovessel — Part 1. BioPharm Int. 23(6)
2010: 53–60.
5 Kolhe P, Amend E, Singh SK.
Impact of Freezing on pH of Buffered
Solutions and Consequences for Monoclonal
Antibody Aggregation. Biotechnol. Prog.
26(3) 2010: 727–733.
6 Kolhe P, Badkar A. Protein and
Solute Distribution in Drug Substance
Containers During Frozen Storage and
Post-Thawing: A Tool to Understand and
Define Freezing-Thawing Parameters in
Biotechnology Process Development.
Biotechnol. Prog. 27(2) 2011: 494–504.
7 Kueltzo LA, et al. Effects of Solution
Conditions, Processing Parameters, and
Container Materials on Aggregation of a
Monoclonal Antibody During Freeze–
Thawing. J. Pharm. Sci. 97(5) 2008: 1801–
1812.
8 Le Saout X, et al. Safe Freeze–Thaw
of Protein Drug Products: A QbD
Approach. BioPharm Int. 23(12) 2010:
28–43.
9 Miller MA, et al. Frozen-State
Storage Stability of a Monoclonal Antibody:
Aggregation Is Impacted By Freezing Rate
and Solute Distribution. J. Pharm. Sci.
102(4) 2013: 1194–1208.
10 Rodrigues MA, et al. Effect of
Freezing Rate and Dendritic Ice Formation
on Concentration Profiles of Proteins
Frozen in Cylindrical Vessels. J. Pharm. Sci.
100(4) 2011: 1316–1329.
11 Singh SK, et al. Large-Scale Freezing
of Biologics. BioProcess Int. 7(9) 2009:
32–44.
12 Singh SK, et al. Frozen State Storage
Instability of a Monoclonal Antibody:
Aggregation As a Consequence of Trehalose
Crystallization and Protein Unfolding.
Pharm. Res. 28(4) 2011: 873–885.
13 Webb SD, et al. Freezing Bulk-Scale
Biopharmaceuticals Using Common
Techniques — and the Magnitude of
Freeze-Concentration. BioPharm Int. 15(5) Reichert JM. Antibody-Based
2002: 22–34. Therapeutics to Watch in 2011. MAbs 3(1)
14 Bhatnagar BS, Bogner RH, Pikal MJ. 2011: 76–99.
Protein Stability During Freezing: Karow AR, Bahrenburg S, Garidel P.
Separation of Stresses and Mechanisms of Buffer Capacity of Biologics — from Buffer
Prog. 29(2) 2013: 480–492. •
Protein Stabilization. Pharm. Dev. Technol. Salts to Buffering By Antibodies. Biotechnol.
12(5) 2007: 505–523.
15 Singh SK, et al. Large-Scale Freezing
of Biologics: A Practitioner’s Review, Part 2:
Practical Advice. BioProcess Int. 7(10) 2009: Corresponding author Manasi Puri is a
34–42. scientist; Sorina Morar-Mitrica is a
16 Kelley B. Industrialization of MAb senior scientific investigator; George
Production Technology: The Bioprocessing Crotts is manager; and Douglas Nesta is
Industry at a Crossroads. MAbs 1(5) 2009: director of biopharmaceutical product
440–449. sciences at GlaxoSmithKline, 709
17 Chang BS, Kendrick BS, Carpenter Swedeland Road, King of Prussia, PA
JF. Surface-Induced Denaturation of Proteins 19406; 1-610-270-5900; manasi.5.puri@
During Freezing and Its Inhibition By gsk.com.
Surfactants. J. Pharm. Sci. 85(12) 1996: 1325–
1330.
18 Strambini GB, Gabellieri E. Proteins For electronic or printed reprints, contact
in Frozen Solutions: Evidence of Ice-Induced Rhonda Brown of Foster Printing Service,
Partial Unfolding. Biophysical. J. 70(2I) 1996: rhondab@fosterprinting.com, 1-866-879-9144
971–976. x194. Download personal-use–only PDFs
19 Lashmar UT, Vanderburgh M, Little online at www.bioprocessintl.com.
SJ. Bulk Freeze–Thawing of Macromolecules
Effects of Cryoconcentration on Their
Formulation and Stability. BioProcess Int. 5(6)
2007, 44–54.
20 Miller R, et al. Dynamics of Protein
and Mixed Protein/Surfactant Adsorption
Layers at the Water–Fluid Interface. Adv.
Colloid Interface Sci. 86(1) 2000: 39–82.
21 Cao E, et al. Effect of Freezing and
Thawing Rates on Denaturation of Proteins
in Aqueous Solutions. Biotechnol. Bioeng.
82(6) 2003: 684–690.
22 Yan L, Zhu Z. Development of
Antibody-Based Therapeutics for Oncology
Indications. Drug Dev. Res. 67(9) 2006: 699–
728.
23 Heller MC, Carpenter JF, Randolph
TW. Protein Formulation and Lyophilization
Cycle Design: Prevention of Damage Due to
Freeze-Concentration Induced Phase
Separation. Biotechnol. Bioeng. 63(2) 1999:
166–174.
24 Luo Q , et al. Chemical Modifications
in Therapeutic Protein Aggregates Generated
Under Different Stress Conditions. J. Biol.
Chem. 286(28) 2011: 25134–25144.
25 Singh SK, et al. Best Practices for the
Formulation and Manufacturing of Biotech
Drug Products. BioPharm Int. 22(6) 2009:
32–48.
26 Voute N et al. Disposable Technology
for Controlled Freeze–thaw of
Biopharmaceuticals at Manufacturing Scale.
BioProcess Int. 2(9) 2004: S40–S43.
Further R eading
Elvin JC, Couston RG, van der Walle
CF. Therapeutic Antibodies: Market
Considerations, Disease Targets and
Bioprocessing. Int. J. Pharm. 440(1) 2013:
83–98.
January 2015 13(1) BioProcess International
S i n g l e -U s e PRODUCTION
A Novel Seed-Train Process
Using High-Density Cell Banking, a Disposable Bioreactor,
and Perfusion Technologies
by Benjamin Wright, Mike Bruninghaus, Mike Vrabel, Jason Walther, Neha Shah,
Seul-A Bae, Timothy Johnson, Jin Yin, Weichang Zhou, and Konstantin Konstantinov
A
typical cell culture process
begins with thawing of a
cryopreserved cell-bank vial,
followed by successive
expansions into larger culture vessels
such as shake flasks, spinners, Wave
bags, and stirred bioreactors (1). When
culture volume and cell density meet
predetermined criteria, the culture is
transferred to a production bioreactor
in which cells continue to grow and
express product. This approach
presents several challenges.
Shake flasks or spinners used in the
initial stages require manual
manipulations inside a laminar flow
hood, making them vulnerable to
contamination and operator error.
Also, because a conventional
cryopreserved cell-bank vial contains
low cell numbers, the seed-train
process is time consuming. It may take
even longer and become more complex
when production bioreactors are very
large because additional large-volume
culture vessels are required to support
scale up to a production bioreactor.
Finally, production bioreactors are
usually inoculated at a density <0.5 ×
106 viable cells/mL, so a necessary
growth phase of 5–10 days offers little
practical value because cell
concentrations (and thus initial product
yield) during that period will be low.
Previous efforts have shown that
high-density (HD) cell banking can be
an effective means to reduce the
number of seed-train steps required
and also improve operational success in
seed-train processes. Tao et al. reported
using HD cell banks containing
Figure 1:  Comparing a conventional with an improved seed-train process for inoculation of a
500-L perfusion bioreactor; the new process uses high-density (HD) cell banking, single-use
technology, and high-density perfusion at the n – 1 stage to allow for high-density inoculation in
the production bioreactor. HD cell banking and disposables improve operational success by
reducing seed-train complexity and contamination potential. HD perfusion at the n – 1 stage allows
for HD inoculation of the production bioreactor, reducing the time to steady-state cell density by
4–5 days and increasing productivity by 10% for a 50-day run.
Conventional Seed Train
Standard WCB
1 mL @ 2 × 107 cells/mL
50-L SS batch 500-L SS batch
125-mL to 10-L glass spinners (n − 1) + ATF production
Improved Seed Train
HD WCB
4.5 mL @ 10 × 107 cells/mL
50-L SUB + ATF 500-L SUB + ATF
2-L and 20-L Wave bioreactors
(n − 1) production
450 million viable cells/vial to directly facility flexibility, time savings in
inoculate a 20-L Wave bag (2). That setup, cleaning, and sterilization, and
study revealed that HD cell-bank use reduced capital investment (4–6).
could eliminate several intermediate Disposable Wave bioreactors from GE
shake-flask expansion steps and reduce Healthcare Life Sciences are routinely
process times by up to nine days. used in seed-train expansion processes
Heidemann et al. combined HD cell because they decrease contamination
banks with a single seed bioreactor risk (typically requiring no
capable of operating at several working manipulations in laminar-flow hoods).
volumes to reduce seed-train expansion Also, because such bioreactors are
duration by 60–70% (3). disposable, no resources are needed for
Single-use technology offers assembly, cleaning, or sterilization (7).
numerous advantages over stainless Industry interest in perfusion cell-
steel systems, including enhanced culture processes has been increasing
M arch 2015 13(3)s BioProcess International
Figure 2:  Viable cell density (VCD) profiles for a complete three-stage Figure 3:  Viable cell density (VCD) as a function of capacitance for n – 1
seed train bioreactor run
120 120
Viable Cell Density (106 cells/mL)

Viable Cell Density (106 cells/mL)

100 100

80 80

60 60

40 40

20 20
Actual
Predicted
0 0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 0 50 100 150 200
Culture Time (days) Capacitance (pF)

(8). Several companies have allowing for direct inoculation into a a 20-L Wave CellBag bioreactor at a
successfully used perfusion bioreactors 2-L Wave bioreactor. Replacing 7.5-L working volume. A model 20/50
in manufacturing commercial spinners (125 mL to 10 L in the figure) EHTD Wave bioreactor system (GE
therapeutics (9). Perfusion processes with Wave bioreactors (2 L and 20 L) Healthcare Life Sciences) served as the
can reach high cell densities in significantly reduces the number of rocking platform for both the 2-L and
relatively small bioreactor volumes (10) required manipulations in a laminar 20-L bags, maintaining cultures at a
and remove unstable product from flow hood. That improves operational temperature of 37 °C, a 16-rpm
cultures more quickly (11) than batch success by allowing operation under a rocking speed, and a 7° rocking angle.
and fed-batch processes. Because of “closed” system. Use of ATF perfusion A 20% O2 and 5% CO2 gas mixture
those and other advantages, perfusion technology at the n – 1 stage (50-L flowed into the head space at a rate of
is often considered in the context of bioreactor with ATF perfusion device 0.25 slpm.
production bioreactors rather than as in the figure) also enables cell densities Seed-Train n – 1 Perfusion Culture:
an element of seed expansion. to be ≥50 × 106 viable cells/mL. That We used a 15-L glass bioreactor
Whitaker (12) and Heidemann (3) allows for an inoculation density of 5 × (Broadley-James Corporation)
briefly mentioned using perfusion in 106 viable cells/mL in the 500-L equipped with an ATF4 perfusion
seed expansion, but they did not discuss production bioreactor rather than 0.5 × device (Refine Technology) to mimic
its potential to improve terminal seed- 106 viable cells/mL in the conventional the n – 1 seed-train stage (50-L
train stages, to reduce bioreactor size or cell culture process. Higher seeding bioreactor in Figure 1). A 20-µm
steps, or to reduce the growth-phase density helps reduce the 500-L sintered sparger added oxygen to
duration for production bioreactors. production bioreactor growth phase by control dissolved oxygen (DO) at 40%,
Pohlscheidt (13) used a 3,000-L n – 1 about five days, saving manufacturing and a 1-mm drilled hole sparger added
perfusion bioreactor with an inclined plant time for a 500-L bioreactor nitrogren to control the dissolved CO2
cell settler to achieve n – 1 cell densities operating for 50 days. level <120-mmHg pressure. As
up to 15.8 × 106 viable cells/mL, needed, we added 0.5 M sodium
decreasing cultivation times by 20% in Materials and Methods carbonate (Na 2CO3) to the culture for
a fed-batch production process. Cell Line and Media: All experiments maintaining pH >6.85 and a 10%
However, higher cell densities are hard used a commercially available FoamAway antifoam solution (Life
to obtain with inclined settler devices chemically defined medium (Life Technologies) to control the foam level.
because of perfusion rate limitations Technologies) supplemented with We seeded a 15-L bioreactor at 10-L
and long residence times for cells in 4 mM glutamine (Sigma-Aldrich). The working volume with 0.5 × 106 viable
suboptimal conditions (10). Chinese hamster ovary (CHO) cell line cells/mL and operated the culture in
Here we describe a novel seed-train is proprietary to Genzyme and produces batch mode until day 2, when perfusion
process using HD cell banking, a recombinant human enzyme. began. Initially, an online capacitance
disposable Wave bioreactors, and Batch Seed-Train Culture in 2-L and sensor (Aber Instruments) and a DeltaV
Refine Technology’s Alternating 20-L Bags: We thawed an HD cell- programmable logic controller
Tangential Flow (ATF) perfusion cell bank vial (10 × 107 viable cells/mL, (Emerson Process Management)
culture technologies. Figure 1 compares 4.5 mL/vial) into a 2-L Wave CellBag controlled the cell-specific perfusion
an example of this process for a 500-L bioreactor (GE Healthcare Life rate (CSPR) at 0.2 nL/cell/day. We
production perfusion bioreactor with a Sciences) at a 1-L working volume. capped the perfusion rate at four
conventional seed-train process. Using When viable cell density (VCD) in the bioreactor working volumes per day
the HD cell bank approach eliminates 2-L Wave bag reached 3.0 × 106 viable (RV/day) after the VCD reached 20 ×
two intermediate spinner steps by cells/mL, we expanded the culture into 106 viable cells/mL.

BioProcess International 13(3)s M arch 2015

 50-mL Spin Tube Batch Refeed Model for Inoculation
Density Evaluation: We removed aliquots of the culture
from the n – 1 seed-train bioreactor when its cell density
reached 25 × 106, 50 × 106, and 100 × 106 viable cells/mL
and subsequently inoculated those samples into 50-mL spin
tubes (TPP Techno Plastic Products) at three different
inoculation densities in triplicate: 0.5 × 106, 2.5 × 106, and
5.0 × 106 viable cells/mL in a working volume of 10 mL.
Refeeds were performed once daily because continuous
perfusion could not be performed at this small scale.
Starting on day 1 we refed the cultures by removing the
spin tubes from their incubator, spinning cells down at
1,100 rpm for five minutes, removing the supernatant, and
then adding fresh media to resuspend the cells. That refeed
strategy is designed to provide the same CSPR across
different inoculation density conditions. A Multitron II
shaking incubator (HT Infors) maintained all spin tubes at
a temperature of 37 °C with a rocking rate of 160 rpm, a
rocking angle of 45°, relative humidity of 80%, and CO2
concentration of 5%.
9, which is the density required to inoculate a 500-L bioreactor
at 5 × 106 viable cells/mL (from a 50-L n – 1 bioreactor).
Initiated on day 2, n – 1 bioreactor perfusion was
controlled to maintain a CSPR of 0.2 nL/cell/day by an
online capacitance probe to automatically increase the feed
rate as cell density rose. Success of such a perfusion-rate
control strategy depends on the capacitance probe’s ability to
accurately estimate VCD. Here, we used the strategy only
until day 7 because the perfusion rate was capped at
4 RV/day, but the probe was able to accurately estimate VCD
throughout the run to day 12, when 110 × 106 viable cells/mL
was reached (Figure 3). Using control strategy during the cell
Figure 4:  Viable cell density (VCD, solid line) and viability (dashed line) for
spin tubes inoculated at (a) 0.5 × 106 viable cells/mL, (b) 2.5 × 106 viable
cells/mL, and (c) 5.0 × 106 viable cells/mL inoculation cell density; n – 1
stage final cell density is 25 × 106 (blue trends), 50 × 106 viable cells/mL
(green trends), 100 × 106 cells/mL (red trends), respectively. Shaded areas
represent ±1 standard deviation (n = 3).
120 100

Viable Cell Density (106 cells/mL)

Production Bioreactor: To mimic the 500-L production 100 A 80
bioreactor in Figure 1, we operated 15-L bioreactors
(Broadley-James Corporation) with 10-L working volumes 80

Viability (%)
in perfusion mode using an ATF4 perfusion device (Refine 60
Technology) with 0.2-µm polyethersulfone filters. A 60
20-µm sintered sparger added oxygen to control DO at 40
40
40%, and a 1-mm drilled hole sparger added nitrogen to
maintain dissolved CO2 level <120-mmHg pressure. We 20 20
maintained pH >6.85 by adding 0.5 M Na 2CO3 and
controlled foam levels by adding 10% FoamAway antifoam 0 0
solution through a peristaltic pump. 0 1 2 3 4 5 6 7 8
One set of production bioreactors was inoculated at a low Culture Time (days)
cell density (0.5 × 106 viable cells/mL) from a seed vessel 120 100
Viable Cell Density (106 cells/mL)

operated in batch mode. Perfusion started on day 1 at

100
0.5 RV/day and increased daily by 0.5 RV/day until reaching B 80
a 2 RV/day perfusion rate. Another set of production reactors 80

Viability (%)
was inoculated at a high cell density (5.0 × 106 viable 60
cells/mL) from an n – 1 15-L perfusion seed vessel when its 60
density reached 50 × 106 viable cells/mL. Because of that 40
high inoculation density, perfusion started at 2 RV/day 40
immediately after inoculation. For all production bioreactors,
20 20
an online capacitance sensor and bleed pump controlled the
VCD at 40 × 106 viable cells/mL. All production bioreactors 0
operated for 50 days. 0
0 1 2 3 4 5
Analytical Methods: We determined VCD using a Culture Time (days)
ViCell XR cell viability analyzer (Beckman Coulter), 120 100
Viable Cell Density (106 cells/mL)

measured pH and pCO2 off-line using a RAPIDLab blood

gas analyzer (Siemens), and quantified protein productivity 100
C 80
with a proprietary photometric enzymatic activity assay.
80
Viability (%)

60
Results and Discussion 60
HD Cell Culture Using ATF Perfusion in the n – 1 Bioreactor:
40
Figure 2 shows the VCD growth profile of our seed-train 40
process including the two Wave bioreactor stages and the
n – 1 stage. The 2-L and 20-L Wave stages lasted eight and 20 20
five days, respectively. The n – 1 bioreactor operated for
0
12 days, reaching a final VCD of 110 × 106 viable cells/mL 0 1 2 3 4 5
0
and >98% cell viability. VCD reached 50 × 106 cells/mL on day Culture Time (days)

M arch 2015 13(3)s BioProcess International

Figure 5:  Viable cell density (VCD) profiles for 10-L production bioreactors
inoculated at 0.5 × 106 viable cells/mL from an n – 1 bioreactor at 2
.5 × 106 viable cells/mL (n = 2) compared with 10-L production bioreactors
inoculated at 5.0 × 106 viable cells/mL from an n – 1 bioreactor at
50 × 106 viable cells/mL (n = 2); dashed line represents the target viable cell
density for steady-state operation, and shaded areas represent the range.
Figure 6:  Cumulative product titer as a function of integral viable cell
concentration for 10-L production bioreactors inoculated at
0.5 × 106 viable cells/mL from an n – 1 bioreactor at
2.5 × 106 viable cells/mL (red points) compared with 10-L production
bioreactors inoculated at 5.0 × 106 viable cells/mL from an n – 1 bioreactor
at 50 × 106 viable cells/mL (green points); error bars represent the range.
3,000,000

Cumulative Product Acitivity (U/L)

70 .
Viable Cell Density (106 cells/mL)

Low-density n − 1 + inoculation
60 2,500,000 High-density n − 1 + inoculation
(n = 2)
50 2,000,000
40
1,500,000
30
1,000,000
20 (n = 2)
Low-density n − 1 + inoculation 500,000
10 High-density n − 1 + inoculation
Target steady-state cell density
0 0
0 10 20 30 40 50 0 500 1,000 1,500 2,000
Culture Time (days) IVCC (106 cells/day/mL)

growth phase allows much more cultures only up to 20 × 106 viable phase duration by four to five days,
accurate control of CSPR, which can be cells/mL to match their CSPR to that increasing productivity by 10% for a
advantageous. Previous reports have of the perfusion bioreactor model. We 50-day run. Additionally, we have
shown that growth and productivity observed no growth differences among shown the n – 1 bioreactor could reach
can be affected by a CSPR difference of the n – 1 density conditions for all 100 × 106 viable cells/mL, theoretically
0.05–0.10 nL/cell/day (14). Our strategy three inoculation densities tested. But inoculating a 500-L bioreactor at 10 ×
also reduces the risk of incorrect or the 100 × 106 viable cells/mL condition 106 viable cells/mL, which would
missed set-point changes because had a noticeably lower starting viability reduce the growth phase duration by
perfusion rate is automatically increased at all inoculation densities than at the an additional five to six days.
as the culture grows. In a stepwise other n – 1 densities. Alternatively, our seed-train process
approach, by contrast, operators must To study the effects of n – 1 density can inoculate a production bioreactor
adjust the rate manually (15). and inoculation density on production operated in either batch or fed-batch
Evaluating the Effects on Cell Growth bioreactor cell growth and productivity, mode. For a 21-day fed-batch cell
of n – 1 Split Density and Production we inoculated bioreactors at 5.0 × culture process using a 500-L bioreactor,
Bioreactor Inoculation Density: Using a 106 viable cells/mL from an n – 1 an inoculation density of 5.0 × 106 viable
small-scale spin tube model, we bioreactor at 50 × 106 viable cells/mL cells/mL could reduce the production
determined how n – 1 cell density and compared those with bioreactors bioreactor duration by 25%. Overall,
affected cell growth in the production inoculated at 0.5 × 106 viable cells/mL that could allow for an additional five to
bioreactor. We removed culture samples from an n – 1 bioreactor at 2.5 × six batches per year, increasing
from an n – 1 bioreactor when cell 106 viable cells/mL. Figure 5 illustrates manufacturing productivity by 25–30%.
densities were 25 × 106 viable cells/mL, comparable cell growth for both In most cases, batch and fed-batch
50 × 106 viable cells/mL, and 100 × conditions. Cell viability remained processes use production bioreactors that
106 viable cells/mL, respectively. Then >90% for both inoculation conditions are much larger than those used for
we split those into three aliquots diluted throughout the 50-day run. Figure 6 perfusion processes, which requires
to 0.5 × 106, 2.5 × 106, or 5 × 106 viable plots cumulative productivity against several seed-train stages in large-scale
cells/mL with fresh media for integrated VCD and indicates a similar stainless steel bioreactors. Figure 7
inoculation in triplicate into spin tubes. specific production rate between compares a conventional seed-train
Daily refeeds began on day 1 at rates conditions. Most notably, the steady- process used to inoculate a 10,000-L
adjusted based on the inoculation state cell density of 40 × bioreactor with another process using a
density of the spin tubes. That allowed 106 viable cells/mL (Figure 5) and the 50-L n – 1 bioreactor using an ATF
each inoculation density condition to required titer for purification were perfusion device. The 50-L n – 1
experience the same CSPR at reached four to five days sooner for the bioreactor at 50–100 × 106 viable
corresponding cell densities. highest inoculation density condition. cells/mL can inoculate a 10,000-L
Figure 4 shows cell count and Economic and Operational bioreactor at 0.25-0.50 × 106 viable
viability profiles plotted to compare the Advantages: Our seed-train process cells/mL, thus eliminating two
n – 1 density conditions at each has both economic and operational intermediate seed-train stages. The
inoculation density. Because the advantages. For example, inoculating a benefit here is an increase in facility
maximum perfusion rate in the spin- 500-L perfusion bioreactor at 5.0 × capacity use because no space and
tube model is 1 RV/day, we grew these 106 viable cells/mL reduces the growth utilities are required for 250-L n – 2 and

BioProcess International 13(3)s M arch 2015

Figure 7:  Comparing a conventional seed-train process to the future seed-train process for
inoculation of a 10,000-L batch or fed-batch bioreactor; the future seed-train process uses high-
density cell banking, disposables, and high-density perfusion at the n – 1 stage to eliminate two
large-volume seed-train stages.
Conventional Seed Train
Standard WCB
1 mL @ 2 × 107 cells/mL
50-L SS batch 250-L batch
125-mL to 10-L
(n − 3)
glass spinners
Improved Seed Train
HD WCB
4.5 mL @ 10 × 107 cells/mL
2-L and 20-L Wave bioreactors
2,500-L n – 1 bioreactors. Additional
large savings could be realized in
up-front capital expenditures.
A Simplified Seed Train
We have successfully demonstrated a
novel seed-train cell culture process
involving HD cell banking, single-use
technology, and perfusion at the n – 1
bioreactor stage. This new procedure
reduces the seed-train process
complexity by decreasing the number
of small-scale expansion stages and
eliminating the need for labor-intensive
operations such as cleaning, assembly,
and sterilization. Manufacturing
operations can rely on the same general
seed-train process with a 50-L SUB at
the n – 1 stage and a 500-L SUB at the
production bioreactor stage.
Use of ATF perfusion at the n – 1
stage can help cultures achieve viable
cell densities ≤100 × 10 6 viable cells/
mL in 12 days without
compromising culture growth
characteristics after further
expansion steps. Based on our work,
a 500-L production bioreactor can
be inoculated at a cell density of
5 × 10 6 viable cells/mL to reduce the
time to steady-state cell density by
four to five days and provide a 10%
increase in the overall productivity
of a 50-day run. This n – 1
perfusion technology can be adapted
10,000-L SS batch
2,500-L batch
(n − 1) production
(n − 2)
50-L SUB + ATF 500-L SUB + ATF
(n − 1) production
in several ways: An inoculation
density of 10 × 10 6 viable cells/mL
could be obtained in a 500-L
bioreactor. For batch or fed-batch
processes that require larger
production bioreactors (e.g., 10,000-
L), one or two stages can be
eliminated from the expansion
process by using a single 50-L
perfusion bioreactor at the n – 1
stage.
R eferences
1 Li F, et al. Current Therapeutic
Antibody Production and Process
Optimization. Bioprocessing J. September–
October 2005: 1–8.
2 Tao Y, et al. Development and
Implementation of a Perfusion-Based High
Cell Density Cell Banking Process.
Biotechnol. Progr. 27(3) 2011: 824–829.
3 Heidemann R, et al. A New Seed-
Train Expansion Method for Recombinant
Mammalian Cell Lines. Cytotechnol. 38(1–3)
2002: 99–108.
4 Smelko JP, et al. Performance of High
Intensity Fed-Batch Mammalian Cell
Cultures in Disposable Bioreactor Systems.
Biotechnol. Progr. 27(5) 2011: 1358–1364.
5 Papavasileiou V, Siletti C, Petrides D.
Systematic Evaluation of Single-Use Systems
Using Process Simulation Tools. BioPharm
Int. 21, March–April 2011: S16–S29.
6 Monge M, Sinclair A. Disposables
Cost Contributions: A Sensitivity Analysis.
BioPharm Int. 22, April 2009: 24–29.
7 Singh V. Disposable Bioreactor for
Cell Culture Using Wave Induced Agitation.
Cytotechnol. 30, 1999: 149–158.
8 Langer ES. Trends in Perfusion
Bioreactors. BioProcess Int. 9(10) 2011: 18–22.
9 Chu L, Robinson DK. Industrial
Choices for Protein Production By Large-
Scale Cell Culture. Curr. Opin. Biotechnol.
12, 2001: 180–187.
10 Voisard D, et al. Potential of Cell
Retention Techniques for Large-Scale High-
Density Perfusion Culture of Suspended
Mammalian Cells. Biotechnol. Bioeng. 82(7) 2003:
751–765.
11 Kadouri A, Spier RE. Some Myths and
Message Concerning the Batch and Continuous
Culture of Animal Cells. Cytotechnol. 24, 1997:
89–98.
12 Whitaker SC, Francis R, Siegel RC.
Validation of a Continuously Perfused Cell
Culture Process for Production of Monoclonal
Antibody. ACS Symp Ser. 698, 1998: 28–43.
13 Pohlscheidt M, et al. Optimizing Capacity
Utilization By Large Scale 3,000-L Perfusion in
Seed Train Bioreactors. Biotechnol. Progr. 29(1)
2013: 222–229.
14 Dowd JE, et al. Optimization and Control
of Perfusion Cultures Using a Viable Cell Probe
and Cell Specific Perfusion Rates. Cytotechnol.
42(1) 2003: 35–45.
15 Kurkela R, Fraune E, Vihko P. Pilot-Scale
Production of Murine Monoclonal Antibodies in
Agitated, Ceramic-Matrix or Hollow-Fiber Cell
Culture Systems. BioTechniques 15, 1993: 674–683.
Further Reading
Abu-Absi SF, et al. Defining Process Design
Space for Monoclonal Antibody Cell Culture.
Biotechnol. Bioeng. 106(6) 2010: 894–905.
Konstantinov KB, et al. The ‘‘Push-to-Low’’
Approach for Optimization of High-Density
Perfusion Cultures of Animal Cells. Adv. Biochem.
Eng. Biotechnol. 101, 2006: 75–98.
Woodside SM, Bowen BD, Piret JM.
Mammalian Cell Retention Devices for Stirred
175. •
Perfusion Bioreactors. Cytotechnol. 28(1) 1998: 163–
Corresponding author Benjamin Wright is a
process engineer, Neha Shah is a staff scientist
I, Jin Yin is associate director, Mike Vrabel is a
process engineer, Jason Walther is a staff
scientist II, Seul-A Bae is a process engineer,
and Konstantin Konstantinov is vice
president, all in late-stage process
development; Mike Bruninghaus is a
manufacturing specialist supervisor in
upstream purification operations; and
Timothy Johnson is director of special projects
and strategic support at Genzyme, a Sanofi
Company, 200 Staples Drive, Framingham, MA
01702; 1-508-271-2080; benjamin.wright@
genzyme.com. Weichang Zhou is vice
president of biologics process development at
WuXi AppTec.
M arch 2015 13(3)s BioProcess International
B i o P r o c e s s TECHNICAL

Hamster Phospholipase
B-Like 2 (PLBL2)
A Host-Cell Protein Impurity in Therapeutic Monoclonal
Antibodies Derived from Chinese Hamster Ovary Cells

Martin Vanderlaan, Wendy Sandoval, Peter Liu, Julie Nishihara, George Tsui, Margaret Lin,
Feny Gunawan, Sara Parker, Robert Ming Wong, Justin Low, Xiangdan Wang, Jihong Yang,
Karthik Veeravalli, Patrick McKay, Chris Yu, Lori O’Connell, Benjamin Tran, Rajesh Vij, Chris
Fong, Kathleen Francissen, Judith Zhu-Shimoni, Valerie Quarmby, and Denise Krawitz

A
ll recombinant protein
biotherapeutics must be tested for
the presence of residual host-cell
protein (HCP) impurities (1–3).
The most common analytical method for
doing so is a polyclonal sandwich
immunoassay. Polyclonal anti-HCP
antibodies are selected to recognize the
broadest population of HCPs possible.
The immunogen and analytical standard
are produced from a blank-run
fermentation that mimics the production
run but lacks the specific biotherapeutic
protein. Because of the large number of
impurities present in harvested cell-
culture fluid (HCCF) that might
copurify with the protein of interest, an
HCP immunoassay must detect a broad
spectrum of proteins. Thus, a limited
Product Focus:  Monoclonal
antibodies (CHO derived)
Process Focus:  Production and
downstream processing
Who Should Read:  Process/
product development, analytical, and
manufacturing personnel
Keywords:  HCP ELISAs, copurifying
host impurities, chromatography,
in-process pools, assay nonlinearity
Level:  Advanced
number of anti-HCP antibodies in the
immunoassay reagents will be directed
against each particular protein impurity.
Moreover, the assay results are
immunologically weighted: Highly
immunogenic HCPs will be detected
more readily than non- or weakly
immunogenic HCPs.
During biopharmaceutical production,
a single HCP can copurify with a
product through the downstream
process. By contrast, the HCP
immunogen and assay standard contain
the thousands of host proteins that could PLBL2 models from PSI, the online protein
be final-product impurities. Antibodies model portal WWW.PROTEINMODELPORTAL.ORG
used in such assays often are affinity
purified against the same diverse mixture it. Likewise, if a copurifying impurity is
of proteins as those used in the highly immunogenic, the assay could
immunogen. So the relevance of the overquantify its value. For these reasons,
antibody population and assay standard orthogonal analytical methods often are
to a particular product-copurifying used to ensure product purity (3).
impurity might not be optimal. That As noted in the earliest publications
disconnect makes enzyme-linked on this topic (4), one property of HCP
immunosorbent assays (ELISAs) only ELISAs is that a particular test sample
semiquantitative for HCPs. may not dilute in parallel with the
Typically, a single HCP numerical analytical HCP standard used. In that
value is reported for such assay results, paper, nonlinear sample dilution was
representing the ratio of HCPs (ng) to ascribed to “antigen excess”: a single or
product (mg). That value can reflect a limited number of copurifying HCP
single protein or a collection of several impurities saturating available antibodies
such impurities. Furthermore, if a to that particular protein. As a sample is
potentially significant copurifying diluted, the measured dilution-corrected
impurity is nonimmunogenic (or weakly impurity level rises. As the authors note,
immunogenic), then the assay could miss for cases in which dilution-corrected

BioProcess International 13(4) A pril 2015

values eventually plateau, the HCP
antigen has been diluted sufficiently to
fall within the linear range of the assay
for that particular HCP, so the plateau
value is reported as an estimate of that
impurity level. When no plateau is
reached, the value from the most dilute
sample is reported, which might be only
a minimum estimate of the HCP level.
In theory, if different products show
nonlinear dilution behaviors in an HCP
ELISA, each product might have its own
unique copurifying HCP antigen in
excess. However, here we report that the
same HCP — phospholipase B-like 2
(PLBL2) — copurifies with multiple
Chinese hamster ovary (CHO)-produced
antibody therapeutics. In all cases, this
impurity manifests as a nonlinear
dilution in our CHO protein (CHOP)
ELISA.
Materials and Methods
In-Process Pool Samples: All in-process
pool samples came from therapeutic
humanized monoclonal antibodies
(MAbs) currently or formerly in
development at Genentech (a member of
the Roche group). We named our
samples accordingly: MAb 1 served as a
negative control, having no detectable
CHOP as measured by multiple
analytical methods; MAb 2 had extreme
dilution-nonlinearity in the platform
CHOP ELISA and was investigated
further to identify the cause of that
nonlinearity (5). Other MAbs had
varying levels of nonlinear CHOP
dilution.
Development of Antibodies and
ELISAs for CHOP and PLBL2: For the
CHOP assay, we concentrated HCCF
from a culture of nontransfected CHO
cells by ultrafiltration, then diafiltered it
against phosphate-buffered saline (PBS).
We used the resulting mixture of CHO
proteins as an immunogen to generate
anti-CHOP antibodies and an analytical
standard for our assay. Antisera from
immunized goats were purified over an
affinity column made by immobilizing
the immunogen on activated glycerol-
controlled pore glass (glyceryl-CPG)
from EMD Millipore.
For the PLBL2 MAb-based assay, we
immunized mice with CHO-derived
PLBL2 (see below) and made
hybridomas from their spleen and lymph
node cells (6). For further development,
we selected a pair of MAbs that were
suitable for generating a sandwich
ELISA. Transient transfection cultures
of CHO cells produced antibodies with
cloned mouse immunoglobulin (IgG)
genes from the hybridomas, and we
affinity purified those antibodies using
protein A chromatography. We used one
member of the antibody pair as a coating
antibody and biotinylated the other for
use as a detection antibody.
For the PLBL2 polyclonal antibody-
based assay, we immunized rabbits with
the CHO-derived PLBL2. The
resulting antibodies were affinity
purified over a PLBL2 column made by
coupling PLBL2 to activated glycerol-
CPG. We either used those purified
antibodies to coat microplates or
conjugated them to horseradish
peroxidase (HRP) for use as detection
antibodies in the sandwich ELISA.
Using methods that are well-
described in literature (7), we formatted a
platform CHOP assay as a sandwich
ELISA. It’s based on direct conjugation
of HRP to the detection antibody, with
detection using tetramethylbenzidine
(TMB, formulated as SureBlue Reserve
from KPL) as a substrate. For this assay,
the biotinylated second antibody was
detected by streptavidin-HRP followed
by washing and TMB substrate.
Purification and Identification of
PLBL2 As an Impurity in MAb Products:
We purified PLBL2 from final MAb 2
product by first loading a 1-mg sample of
product in PBS with 0.05% Tween 20
mobile phase onto a 1.6- to 2.0-mL
CHT-1 ceramic hydroxyapatite column
from Bio-Rad Laboratories (8, 9).
Continued washing with PBS-Tween
preceded a step elution with 0.5 M
sodium phosphate to remove the
antibody from the column. We collected
fractions throughout the chromatography
process and assayed them with the
CHOP ELISA, pooling CHOP-positive
fractions for further analysis.
Because the pooled fractions had low
concentrations of a potentially sticky
protein, we took care to minimize losses
on surfaces and filters. Pools were
concentrated to 20 μL using Amicon
spin filters prewet with sodium-dodecyle
sulfate (SDS). Then we added Laemmli
sample buffer containing 2% SDS
directly to the filters and mixed with a
vortex mixer before sample recovery. The
entire retentate was transferred to a
single lane on an SDS polyacrylamide
gel electrophoresis (SDS-PAGE) gel.
After resolving that lane by
electrophoresis, we stained the results
either with SYPRO Ruby fluorescent
protein dye or SimplyBlue Safe stain
(both from Life Technologies) and rinsed
with water before imaging.
As described elsewhere (10), we
excised gel bands covering either the
entire 30–70 kDa region or discrete,
visible bands within that range, then
washed those in 25 mM ammonium
bicarbonate (Sigma) containing 5%
acetonitrile (Burdick and Jackson),
followed by 25 mM ammonium
bicarbonate in 100 μL of 50:50
acetonitrile:water for 20 minutes before
100% acetonitrile. The in-gel trypsin
digestion (0.2 μg, Promega) in 25 mM
ammonium bicarbonate at pH 8 was
incubated overnight at 37 °C, then
arrested by addition of 5 μL of 1% formic
acid. We extracted peptides from the gel
slices using 0.1% formic acid containing
5% acetonitrile (50 μL) and finally with
50 μL of pure acetonitrile. Extractions
were pooled, dried completely, and
reconstituted in a volume of 10 μL with
0.1% formic acid.
We injected half the total sample
volumes (5 µL) with an autosampler onto
a 75-μm × 150-mm Waters column at a
1-μL/min flow rate using a Waters
nanoAcquity ultraperformance liquid
chromatography (UPLC) system. It
applied a gradient from 98% Solvent A
(water + 0.1% formic acid) to 80%
Solvent B (acetonitrile + 0.08% formic
acid) over a 40-minute period. We
analyzed the samples on-line using
nanospray ionization into a hybrid LTQ-
Orbitrap XL mass spectrometer
(Thermo Scientific). Data were collected
in data-dependent mode with the parent
ion analyzed by Fourier-transform mass
spectrometry (FTMS), and we selected
the eight most abundant ions for
fragmentation and analysis with the
LTQ instrument. Using the Mascot
search algorithm (Matrix Sciences), we
analyzed tandem mass spectrometric
data against the Uniprot database
including peptide sequences from all
mammals.
A pril 2015 13(4) BioProcess International
Gene Sequence Primers
pRK-F 5′-GGTGACACTATAGAATAACATCCACTTTGCC-3′
pRK-R 5′-GTGAAATTTGTGATGCTATTGCTTTATTTGTAACC-3′
PLBL2-mid-F 5′-CGCATCATCAAGAAGTACCAGCTGCAG-3′
PLBL2-mid-R 5′-GCACGTACTTCCACAGGGCTGG-3′
Cloning, Expression, and Purification
of Hamster PLBL2: Technicians at Blue
Heron Biotechnology synthesized and
cloned the hamster PLBL2 gene
(G3I6T1_CRIGR) with a C-terminal
hexahistidine-tag into a Hoffmann-La
Roche proprietary vector. The sequence
of the PLBL2 gene was confirmed using
the following primers: pRK-F, pRK-R,
PLBL2-mid-F, and PLBL2-mid-R (see
the “Gene Sequence Primers” box above).
Using standard methods, we
transformed One Shot TOP10
chemically competent Escherichia coli cells
(Life Technologies) with the pRK5sm–
PBLB2 plasmid using standard methods,
then selected transformants on Luria
broth (LB) agar plates containing 50 µg/
mL carbenicillin antibiotic. From an
overnight culture of TOP10 cells, we
purified the plasmids using a QIAprep
spin maxiprep kit (Qiagen). Transient
transfections at the 35-L scale were
expressed PLBL2 as described elsewhere
(11).
We ultrafiltered HCCF tenfold by
tangential-flow filtration (TFF) using
10-kDa molecular weight cutoff
(MWCO) membranes. The ultrafiltered
HCCF was diafiltered against 10
volumes of a PBS-NaCl buffer. Then we
applied the resulting material to a
Ni-NTA immobilized metal-affinity
chromatography column (Qiagen) and
eluted with an increasing imidazole
gradient. After conditioning the
Ni-NTA pool with a buffer containing
sodium sulfate, we applied it to an Octyl
Sepharose hydrophobic-interaction
chromatography (HIC) column from GE
Healthcare. The HIC column was eluted
with a decreasing sodium sulfate
gradient.
We rechromatographed the resulting
pool on a Ni-NTA column step-eluted
with a high concentration of imidazole,
then concentrated the resulting pool
using centrifugal filtration units equipped
with 10-kDa MWCO membranes
(EMD Millipore). The concentrated
Ni-NTA rechromatographed pool was
BioProcess International 13(4) A pril 2015
formulated on a Superdex 200 size-
exclusion chromatography (SEC) column
from GE Healthcare. We collected
fractions from that column and analyzed
them with SDS-PAGE, high-pressure
liquid chromatography (HPLC), and
standard Western blotting techniques
(using antibodies to detect CHOPs and
phospholipase B2) to determine purity.
Surface Plasmon Resonance (SPR): To
evaluate binding interactions between
CHO PLBL2 and intact antibodies or
fragments, we used SPR technology on a
Biacore T200 instrument from GE
Healthcare. Using two different formats
ensured that molecules for comparison
were kept in the mobile phase as analytes
instead of being immobilized as ligands
onto sensor chips. All SPR experiments
were carried out at 25 °C and corrected
against signals from the in-line reference
cell and buffer-blank.
With the first format, we evaluated
interactions of PLBL2 with various
rhuMAbs. Here, anti-hu IgG Fc (GE
Healthcare) was covalently immobilized
to all four flow cells (FCs) of a Series S
CM5 sensor chip (GE Healthcare) to a
level of ~6,500 response units (RU). We
then used immobilized anti-hu IgG Fc to
capture rhuMAbs diluted into HBS-EP+
buffer — 10 mM 4-(2-hydroxyethyl)-1-
piperazineethanesulfonic acid, 150 mM
sodium chloride, 3 mM
ethylenediaminetetraacetic acid, 0.05%
polysorbate 20, pH 7.4 — for FC2, FC3,
and FC4 to achieve an immobilization
level of ~900 RU. FC1 served as an
in-line reference cell. With CHO
PLBL2 diluted to 1.0 µM in HBS-EP+
buffer subsequently injected into all four
FCs at a 50-µL/min flow rate for 60
seconds, dissociation proceeded for three
minutes. At the end of each association/
dissociation cycle, we regenerated the
sensor chip by injecting 3 M magnesium
chloride into all four FCs for 30 seconds.
Using the second format, we
compared PLBL2 binding interactions
with full-length, F(ab′)2, and Fc
fragments generated from Genovis IdeS
enzymatic cleavage of three rhuMAbs in
PBS at an enzyme:antibody ratio of
1:200. We incubated the mixture
overnight at ambient temperature before
affinity purification across MAbSelect
Sure protein A and Capto-L protein L
resins (both from GE Healthcare) to
remove IdeS and undigested MAbs
before beginning our SPR studies. Here,
CHO PLBL2 was immobilized onto
two of the four flow cells (FC2 and FC4)
of a Series S CM5 sensor chip using a
standard amine coupling and blocking
procedure recommended by the
manufacturer and ~4,000-RU
immobilization levels. We used the two
remaining FCs (FC1 and FC3) as in-line
references for FC2 and FC4 and treated
them similarly (without immobilized
proteins) before blocking. We diluted
intact antibodies and fragments to 1 μM
in HBS-EP+ buffer and ran them over
FCs coated with immobilized CHO
PLBL2 for five minutes at a 5-μL/min
flow rate. In-line reference-corrected
responses were reported five seconds
before the end of each sample injection.
And we regenerated the sensor chip by
injecting 10 mM Glycine (pH 2.0) over
the immobilized CHO PLBL2 at a
30-μL/min flow rate for 10 seconds.
CHOP ELISA Blocking with Anti-PLBL2
Antibodies: To determine whether
PLBL2 was the only CHOP impurity
causing nonlinear MAb 2 sample
dilution, we assayed in-process pool
samples potentially containing both
PLBL2 and other CHOP in our PLBL2
ELISA and in our CHOP ELISA.
Samples were assayed either neat or after
preincubation with rabbit polyclonal anti-
PLBL2.. Sample dilutions were assayed
by CHOP ELISA either neat (undiluted)
or after preincubation with rabbit
polyclonal anti-PLBL2. In particular, we
noted the impact of anti-PLBL2
antibodies on the CHOP ELISA and
whether the sample diluted nonlinearly.
Results
We used a platform CHOP ELISA to
detect CHOP impurities in
biopharmaceuticals produced by a
relatively harmonized upstream cell
culture process. Using a common assay
for multiple products makes it possible to
identify atypical results in specific clinical
candidates. In one such case, we observed
an atypical, profoundly nonlinear dilution
of samples late in purification. Figure 1
shows the CHOP ratios (ng CHOP/mg
product) of twofold serially diluted
samples from the initial HCCF and
several intermediate purification pools.
Those early in the downstream process
dilute in parallel with the assay standard
(giving dilution-independent CHOP
ratio values); later pools show increasing
values as samples are diluted.
We calculated the coefficient of
variation (CV) for assays at each of those
sample dilutions. If the CV was ≤20%,
we concluded that differences across
dilutions are within acceptable levels of
variability for ELISA sample replicates,
and the sample was reported as
exhibiting no dilution nonlinearity. If the
CV is >20% and inspection shows an
increasing CHOP level with increasing
sample dilution, then a sample was
reported to have dilution-dependent
behavior. As illustrated in Figure 1, the
CV of dilutions for later pools was 155%,
well above the cut-off value.
One explanation for nonlinear sample
dilution is that a single CHOP is present
in excess of available anti-CHOP
antibodies (3, 4), so it exceeds the
ELISA’s antibody binding capacity.
Under such an interpretation, sample
dilution to the assay limit gives the most
reasonable estimate of the CHOP ratio.
For the Figure 1 preparation, we
estimated the CHOP ratio in the final
product to be ~300 ng/mg.
Because that sample contained an
abnormally high level of a potential
CHOP impurity in the final purification
pool, we investigated it further using an
approach described elsewhere (2). This
involves capturing the MAb on a
chromatographic support, where it is
tightly bound, and then collecting flow-
through and washes in an effort to isolate
the dissociated impurity from the
product. The approach is based on the
assumption that the reason for impurity
copurification is an interaction between it
and the MAb product, which must be
disrupted for impurity dissociation.
For that reason, we captured our
product on a 1-mL CHT column that
was washed with PBS-Tween (Figure 2).
The UV chromatogram trace shows
product elution late in the chromatogram
at >150 mM phosphate concentrations.
Collected fractions were assayed in the
CHOP ELISA (Figure 2a’s trace with
diamond marks), with immunoreactive
material (eluted early in the
chromatogram) well-separated from the
product. A small amount of such material
coeluted with the product and was not
pooled with other immunopositive
fractions for subsequent study.
As described above, the CHOP-
positive fractions (fractions 8–20) were
pooled, concentrated, and analyzed by
SDS-PAGE. Sypro Ruby-stained bands
were excised, digested with trypsin, and
the resulting gel-extracted peptides were
sequenced by mass spectrometry. After
comparing the peptide sequences with
databases of product, CHO proteins, and
other mammalian proteins, we identified
phospholipase B-like 2 (PLBL2) as the
main impurity.
We observed several dye-stained
bands (Figure 2b). Although some minor
bands were product-related, we identified
the three major bands as having
homology to hamster PLBL2. Literature
on PLBL2 (12, 13) and information from
its UniProt entry (Q8NHP8) describes
the protein as a mannose-6-phosphate
glycosylated lysosomal molecule that is
synthesized as a 66-kDa preproenzyme.
Under mildly acidic conditions, this
enzyme is clipped into two chains that
remain strongly associated in native
conditions; however, in the presence of
SDS and when heated, those chains
separate into N- and C-terminal
fragments. That explains the three bands
on the gel: those two clipped fragments
and the full-length protein.
Having identified our putative
CHOP impurity, we had the PLBL2
gene cloned for transient expression in
CHO cells, then purified the results and
used them to raise both mouse
monoclonal and rabbit polyclonal
antibodies. We developed PLBL2-
specific ELISAs using those reagents.
Both ELISAs gave comparable results
across multiple products and intermediate
pools. We used the MAb-based ELISA
to assess PLBL2 in all our MAb
portfolio biotherapeutics, including both
commercial products and those in clinical
development.
Figure 3 shows an immunoblot using
the rabbit polyclonal antisera. In the gel,
the PLBL2 standard is visible down to
Figure 1:  Nonlinear dilution of in-process pool
samples for a MAb product in clinical
development; twofold dilution series were
plotted according to CHOP ratios (ng CHOP/mg
product) for samples taken from the initial
harvested cell culture fluid (HCCF) and several
intermediate downstream purification pools.
Pools early in the purification process dilute in
parallel to the assay standard (giving dilution-
independent CHOP ratio values); later pools
show increasing CHOP ratio values as samples
are diluted. The only processing between the
third step and final pools is an ultrafiltration/
diafiltration (UF/DF) concentration and buffer
exchange step.
10 6
HCCF
105
CHOP Ratio (ng/mg)
104
First step pool
103
Second step pool
102 Third step pool
101
Final pool
0
0.00001 0.0001 0.10 10.0
Product Concentration (mg/mL)
Black diamonds = HCCF X = third step pool
black squares = first step pool * = final pool
open squares = second step pool
≥2 ng/lane and shows staining of both
the full-length protein and the smaller
clipped fragment. The CHOP ELISA
standard detected four bands. No anti-
PLBL2 reactive bands were detected in a
MAb 1 sample, which served as a
negative control throughout the study
(ELISA PLBL2 ratio in MAb 1 is below
the assay’s limit of detection, 0.3 ng/mg).
However, the same two bands detected
in the PLBL2 standard were detected in
two MAb 2 manufacturing runs: PLBL2
ratios of 120 ng/mg for run 2, and 300
ng/mg for run 3. Run 3 corresponds to
that shown in Figure 1, which had
pronounced nonlinear dilution in the
total CHOP ELISA.
We revised the MAb 2
manufacturing process to reduce PLBL2
levels and observed no anti-PLBL2
reactive bands in final-product samples
from the resulting updated process.
Corresponding ELISA data for those
lots showed PLBL2 levels <1 ng/mg. We
have found that purification processes
can be modified successfully to reduce
PLBL2 levels to <1 ng/mg.
Our survey of Genentech commercial
biologics showed that PLBL2 is
undetectable in Herceptin, Avastin,
Perjeta, Rituxan, and Xolair products.
A pril 2015 13(4) BioProcess International
Figure 2a:  Chromatographic separation of CHOP impurities from the product; (a) UV trace shows Figure 2b:  Chromatographic separation of
product elution late in the chromatogram at concentrations of phosphate >150 mM. Collected CHOP impurities from the product; CHOP-
fractions were assayed in the CHOP ELISA (diamonds) and show immunopositive material eluted positive fractions 8–20 were pooled,
early in the chromatogram (well-separated from product). A small amount of immunoreactive concentrated, and analyzed on a sodium-
material coeluted with product and was not included in subsequent studies. dodecyl sulfate polyacrylamide gel
500 electrophoresis (SDS-PAGE) gel.

400 CHOP ELISA of Fractions

CHOP (ng/mL)

300 UV Trace of IgG Product

200 250
C2 C1
100
150 A1
0 A2
0 5 10 15 20 25 30 35 40 A3 IgG
100
−100
Fraction
75 A4
A5
A6
Table 1:  SYPRO Ruby stained bands were excised and digested with trypsin; the resulting peptides 50
were analyzed by mass spectrometry. The coverage plot maps identified peptides of PLBL2 to the
highest database hit. Coverage: 195/585 (33%)
37 A7 PLBL2

MA A PMD R S P D G R A V R A L R L A L A L A S L T E V L L N C P A G A L P T QG P G R R R QN L A8
25 A9
D P P V S R V R S V L L D A A S GQ L R L V D G I H P Y A V AWA N L T N A I R E T GWA Y L D L G
20
IgG
A10
T N G S Y N D S L Q A Y A A G V V E A S V S E E L I Y M H WM N T M V N Y C G P F E Y E V G Y C E K
L K S F L E I N L E WM Q R E M E L S Q D S P Y W H Q V R L T L L Q L K G L E D S Y E G R L T F P T 15
G R F T I K P L G F L L L Q I A G D L E D L E Q A L N K T S T K L S L G S G S C S A I I K L L P G A 10
R D L L V A H N TWN S Y QNM L R I I K K Y Q L Q F R QG P Q E A Y P L I A G N N L V F S S Y P G
T I F S G DD F Y I L G S G L V T L E T T I G N K N P A LWK Y V Q P QG C V L EW I R N I V A N R
L A L D G A T W A D I F K Q F N S G T Y N N QWM I V D Y K A F I P N G P S P G S R V L T I L E Q I
lane 1 = reference material
lane 2 = ELISA positive fractions
P GMV V V A D K T E D L Y K T T YWA S Y N I P F F E I V F N A S G L QD L V A Q Y G DWF S Y T
K N P R A Q I F Q R DQ S L V E DMN S MV R L I R Y NN F L HD P L S L C E A C I P K P N A E N A
I S A R S D L N P A N G S Y P F Q A L Y Q R P H G G I D V K V T S F S L A K RM S M L A A S G P TW
D Q L P P F QW S L S P F R S M L HMG Q P D L W T F S P I S V PWD with antibodies), PLBL2 binding was
completely blocked in the PLBL2
Minimum PSMs: 0 1 2 5 Modified
ELISA. In the CHOP ELISA, samples
diluted linearly with a low CHOP value
However, we did find it as an impurity in samples. Additionally, although PLBL2 (<1 ng/mg). That result showed that all
several products that are still in clinical is suspected to have enzymatic activity, the previously observed nonlinearity
development. Nonlinear dilution no associated substrates have been could be explained by PLBL2 in the
behavior of samples in the CHOP published (14). If the protein is sample.
ELISA often evidenced PLBL2 as the enzymatically active, then its substrate Using the PLBL2-specific ELISA,
CHOP present in antigen excess. In remains unidentified. we assessed PLBL2 levels in HCCF
some cases, the CHOP ELISA showed We spiked purified PLBL2 into the from a number of production cell cultures
low CHOP ratio values in the final CHOP ELISA (Figure 4). At low and blank-run cultures. Levels varied
product (<2 ng/mg). But PLBL2 was concentrations (<10 ng/mL), the CHOP widely among production runs (Figure
present at similar levels based on assay accurately quantifies PLBL2 spikes. 5). For many MAbs, we measured several
PLBL2-specific ELISA results, Increasing concentration of PLBL2 HCCF samples for different production
suggesting that PLBL2 could be the decreases spike recovery consistent with cultures, and the bar graphs show their
predominant CHOP species even in saturation of the anti-PLBL2 antibodies minimum (solid bar) and maximum
products with low total CHOP. We in the overall CHOP ELISA. The most (hashed bar) values. With a median
implemented purification improvements accurate PLBL2 estimate (using only the PLBL2 level of 2.5 μg/mL, values
to reduce PLBL2 levels in all clinical CHOP ELISA) occurs when a sample is ranged 0.76–7.7 μg/mL. If expressed
products for which it was detected. diluted to near an assay’s quantitation relative to the total CHOP content of the
Reduction of PLBL2 from MAb 2 pools limit. That supports the “antigen-excess” HCCF, PLBL2 content ranged 0.16%–
(Figure 3, Runs 5–8) illustrates the hypothesis as the cause of sample dilution 1.2% of the CHOP, with a median value
success of those improvement efforts. nonlinearity in the CHOP ELISA. of 0.37%. Levels range over about an
We also evaluated fluorogenic To further establish a connection order of magnitude across cultures.
synthetic phospholipase substrates (Life between nonlinear dilution of samples Previous studies on MAb
Technologies) to determine to determine and PLBL2 as the causative impurity, we biotherapeutics have shown that the
whether the enzyme was active but found incubated in-process pools with a tenfold biggest influence on CHOP level in
no enzymatic activity with those molar excess of rabbit anti-PLBL2 before protein A pools is the antibody rather
substrates for either our purified PLBL2 ELISA testing. With that pretreatment than nonspecific binding to the
or MAb biotherapeutic in-process (designed to coat PLBL2 in the samples chromatography media (15). So

BioProcess International 13(4) A pril 2015

Figure 3:  Immunoblot using rabbit polyclonal antisera to PLBL2; (left to right) the lanes correspond Figure 4:  PLBL2 assayed in the platform
to MW standards, a PLBL2 standard at three different protein concentrations, the CHOP standard, enzyme-linked immunosorbent assay (ELISA)
MAb 1, and six lots of MAb 2. The first two of those six represent MAb 2 prepared before the for CHOP; at low levels, PLBL2 recovery is near
protocol was optimized, and the last four came after changes made to reduce PLBL2 levels. 100%. As PLBL2 levels increase, the assay
response does not similarly increase, which is
consistent with the interpretation that only a
limited number of antibodies is available to
250 bind PLBL2. Here, “antigen excess” is expected
150 to lead to nonlinear dilution in samples
100 containing >10 ng/mL of PLBL2.
75 1.5

Optical Density
50 1.0
37
0.5

25 0.0
20 1 10 100
Protein Concentration (ng/mL)
W

2- PL 2
PL 2
B2

rd

Ab b 1

M 2 n3
M 2R n2
M 2R n5
M 2R n6
2 n7
8
-n LB
ng B

n
M

da

Ab Ru
Ab Ru
Ab u
Ab u
Ab u
Ru
A
10 g P

M
an

the first paper that described an ELISA

g

M 2
-n

St

for HCPs— in that case for E. coli

20

OP

proteins in recombinant human growth

CH

potentially small differences in antibody
sequence might greatly influence CHOP
interactions with the antibodies. In an
effort to understand why CHO-derived
PLBL2 copurifies with only some
MAbs, we conducted two more studies.
First, we used SPR to look for
PLBL2–antibody (or antibody fragment)
interactions (Figure 6). This
demonstrated interactions between
immobilized MAbs and PLBL2 and vice
versa. However, measuring such
interactions required micromolar levels
of PLBL2 in the mobile phase. The
binding was readily reversible, with a
return to baseline in minutes. These
observations suggest that the binding
interaction is relatively weak, and only
high MAb concentrations during
purification processes allow these
impurities to be carried along.
Further SPR studies showed that
PLBL2 was interacting with the F(ab′)2
portion of the MAb rather than its Fc
region (Figure 7). MAb 1 served as the
negative control, and MAb 2 was the
positive antibody. These results show that
most of the SPR signal is derived from
the F(ab′)2 fragment. MAb 6 had a very
weakly positive SPR signal to both the
intact molecule and its F(ab′)2 fragment.
Finally, we assessed whether other
CHOP ELISA antibody preparations —
both those from commercial kits and
developed in-house — were responsive to
PLBL2. In some cases, PLBL2 detection
was low or nonexistent with those
antibody preparations (data not shown).
Thus, some CHOP ELISAs might be
insensitive to this particular impurity.If
the ELISA reagents are insensitive to
PLBL2, then samples shown to have low
CHOP values might yet contain high
levels of this impurity.
Discussion
PLBL2 may be a relatively common and
potentially overlooked HCP impurity in
therapeutic MAbs produced by CHO
cultures. In our case, HCP ELISA
dilution nonlinearity was not an assay
artifact; it was a consequence of antigen
excess and reflected the need to dilute
samples “into the range of the assay” so
there would be sufficient antibody to
capture and detect the HCP impurity. It
is important, therefore, to test samples for
HCP impurities at multiple dilutions.
Nonlinear responses should be
investigated because they may indicate
the presence of a single or limited
number of HCP species that can copurify
with product and might be present in
large amounts.
Identification of sample dilution
nonlinearity thus is an important signal
that an HCP may be present in antigen
excess. Our description of the antigen-
excess phenomenon and associated
nonlinearity is not new. It was detailed in
hormone (4).
Although the nonlinearity observed in
our CHOP ELISA for multiple different
products could have resulted from
multiple, different CHOP species, our
results show that a single CHOP had
independently copurified with multiple
antibodies. Several teams have put forth
the “hitchhiker impurities” idea that
major HCP impurities in final
therapeutic products are there because
they bind to the product (15–20). Sisodiya
et al. found that antibody–CHOP
interactions contributed significantly to
carrying CHOP along in the recovery
process and observed significant
differences between antibodies in their
interactions with CHOP (15).
Those observations have been
confirmed with PLBL2 as a specific
CHOP. It appears to copurify based on
an association with the Fab′2 portion of
particular antibodies. Also, the
probability that PLBL2 will copurify
with a given antibody may depend on the
particular clone selected for production.
CHO cells are heterogeneous in PLBL2
expression levels, as illustrated by the
production cultures whose HCCF
PLBL2 levels differ by over tenfold.
Similarly, PLBL2 levels in blank runs
vary, so anti-CHOP antibody
preparations produced in response to
those blank-run immunogens also could
differ in their ability to detect PLBL2.
Initially, we evaluated a commercial
ELISA test kit for hamster PLBL2 from

A pril 2015 13(4) BioProcess International

 Figure 5:  Levels of PLBL2 in HCCF samples displayed as μg/mL (left) and percentage of the total CHOP (right) from 13 different MAb products; for cases in
which multiple independent production cultures were measured for the same MAb, the solid bar shows the lowest observed value, and the stippled bar
shows the highest observed value. Wide variation among cultures in PLBL2 levels is shown by absolute μg/mL and as relative percentage of the total CHOP.
9 1.4
8 1.2
PLBL2 in HCCF (µg/mL)

PLBL2 as % of CHOP
1.0
6
5 0.8
4 0.6
3
0.4
2
1 0.2

0 0.0
1 2 3 4 5 6 7 8 9 10 11 12 13 1 2 3 4 5 6 7 8 9 10 11 12 13
Monoclonal Antibody ID Number Monoclonal Antibody ID Number

Cusio Bio in China. However, the kit did with recombinant therapeutics. We have hamster PLBL2.
not react with our purified PLBL2 observed no cases in which only the One primary concern about HCP
standard, making it unsuitable for use in N-terminal or C-terminal peptides were impurities is their potential to be
establishing product impurity levels. We present, nor any case in which one or the immunogenic. As foreign proteins, they
are unaware of any other commercially other fragment was in excess. might elicit an immune response in
available kits or reagents for detection of Phospholipases exist widely in nature exposed patients (23, 24). Hamster
hamster PLBL2, so we developed our and can be found in bee, spider, and PLBL2 has ~80% amino acid sequence
own immunoassays. snake venoms as well as in bacteria. identity with its human ortholog, and
PLBL2 was originally described in a However, nonmammalian species have many of their differences are surface-
study that screened for proteins distinct genes and share low sequence exposed residues. Given the multiple
containing mannose-6-phosphate. In homology with mammalian PLBL2. In amino-acid differences between human
both published literature and our own particular, although a phospholipase has and hamster PLBL2, it is reasonable to
experience, the two PLBL2 constituent been found to be a virulence factor in assume that PLBL2 would be
peptides remain closely associated after fungi, and infected humans have immunogenic in humans. So for patient
acid hydrolysis. Denaturing conditions generated antibodies to this enzyme (21, safety, PLBL2 levels should be reduced
are required to dissociate those 22), that molecule is unrelated to PLBL2. as much as practically possible in
fragments. Consistency between our two The natural substrate for PLBL2 is as yet recombinant biotherapeutics (25).
PLBL2-specific ELISAs (one based on a unreported, but it may be an arylamidase We have no reason to think that our
pair of mouse MAbs and one based on (14). Synthetic substrates used to measure observations are unique to Genentech
rabbit polyclonal antibodies) supports the bacterial phospholipases were not CHO cells or the Genentech production
idea that the whole protein copurifies hydrolyzed by our purified recombinant process. It is possible that PLBL2 is a

Figure 6:  Representative surface plasmon resonance (SPR) sensograms showing reversible binding between 1 μM PLBL2 and immobilized MAb
biotherapeutics; (left) MAb 1, with which PLBL2 tends to copurify, shows positive binding. Multiple traces in the MAb 1 panel show measurement
reproducibility with independent MAb 1 preparations (lots), each of which was made with the new process and is free of copurifying PLBL2. But the
positive interaction of PLBL2 and the antibody remains a property of the antibody. (right) MAb 1, with which PLBL2 does not copurify, does not respond
in the sensogram, showing that this antibody does not exhibit the PLBL2 binding properties of MAb 2. (* sample start)
20 20

17 17
MAb 1
14 14
MAb 2
Response Units

Response Units

11 11

8 8
Injection Injection
5 5 start stop

2 2

−1 −1
−10 10 30 50 70 90 110 130 150 −10 10 30 50 70 90 110 130 150
Time (seconds*) Time (seconds*)

BioProcess International 13(4) A pril 2015


frequent impurity in CHO-derived MAb biotherapeutics across
the industry. It is also possible that widely used commercial
CHOP ELISAs do not detect this impurity, or they could
significantly underreport its presence.
We have instituted process changes to reduce PLBL2 levels
in all our MAb products. As illustrated herein for one product,
it is possible to reduce PLBL2 to levels near or <1 ng/mg. We
reduced patient risk with a combination of a sensitive assay for
PLBL2 and implementation of PLBL2-reducing process
changes.
References
1 Tscheliessnig AL, et al. Host Cell Protein Analysis in Therapeutic
Protein Bioprocessing: Methods and Applications. Biotechnol. J. 8(6) 2013:
655–670.
2 Wang X, Hunter AK, Mozier NM. Host Cell Proteins in Biologics
Development: Identification, Quantitation, and Risk Assessment. Biotechnol.
Bioeng. 103(3) 2009: 446–458.
3 Zhu-Shimoni J, et al. Host Cell Protein Testing By ELISAs and the
Use of Orthogonal Methods. Biotechnol. Bioeng. 111(12) 2014: 2367–2379.
4 Anicetti VR, et al. Immunoassay for the Detection of E. coli Proteins
in Recombinant DNA Derived Human Growth Hormone.
J. Immunol. Meth. 91(2) 1986: 213–224.
5 Corren, J, et al. Lebrikizumab Treatment in Adults with Asthma.
New England J. Med. 365(12) 2011: 1088–1098.
6 McCullogh KC, Spier RE. Monoclonal Antibodies in Biotechnology.
Cambridge University Press: New York, NY, 2009.
7 Wild D. The Immunoassay Handbook: Theory and Applications of Ligand
Binding, ELISA, and Related Techniques, Fourth Edition. Elsevier Science:
Oxford, UK, 2013.
8 Gagnon P, et al. A Ceramic Hydroxyapatite-Based Purification
Platform: Simultaneous Removal of Leached Protein A, Aggregates, DNA,
and Endotoxins from MAbs. BioProcess Int. 4(2) 2006: 50–60.
9 Stanker LH, Vanderlaan M, Juarez-Salinas H. One-Step Purification
of Mouse Monoclonal Antibodies from Ascites Fluid By Hydroxylapatite
Chromatography. J. Immunol. Meth. 76(1) 1985: 157–169.
10 Champion KM, et al. Comparison of the Escherichia coli Proteomes for
Recombinant Human Growth Hormone Producing and Nonproducing
Fermentations. Proteomics 3(7) 2003: 1365–1373.
11 Wong AW, Baginski TK, Reilly DE. Enhancement of DNA Uptake
in FUT8-Deleted CHO Cells for Transient Production of Afucosylated
Antibodies. Biotechnol. Bioeng. 106(5) 2010: 751–763.
12 Deuschl F, et al. Molecular Characterization of the Hypothetical 66.3-
kDa Protein in Mouse: Lysosomal Targeting, Glycosylation, Processing, and
Tissue Distribution. FEBS Lett. 580(24) 2006: 5747–5752.
13 Lakomek K, et l. Initial Insight into the Function of the Lysosomal
66.3 Kda Protein from Mouse By Means of X-Ray Crystallography. BMC
Struct. Biol. 9, 2009: 56.
14 Repo H, et al. Is the Bovine Lysosomal Phospholipase B-Like Protein
an Amidase? Proteins 82(2) 2014: 300–311.
15 Sisodiya VN, et al. Studying Host Cell Protein Interactions with
Monoclonal Antibodies Using High-Throughput Protein A Chromatography.
Biotechnol. J. 7(10) 2012: 1233–1241.
16 Hunter AK, et al. Separation of Product Associating E. coli Host Cell
Proteins OppA and DppA from Recombinant Apolipoprotein A-I(Milano) in
an Industrial HIC Unit Operation. Biotechnol. Progr. 25(2) 2009: 446–453.
17 Levy NE, et al. Identification and Characterization of Host Cell
Protein Product-Associated Impurities in Monoclonal Antibody
Bioprocessing. Biotechnol. Bioeng. 111(5) 2014: 904–912.
18 Nogal B, Chhiba K, Emery JC. Select Host Cell Proteins Coelute
with Monoclonal Antibodies in Protein A Chromatography. Biotechnol. Progr.
28(2) 2012: 454–458.
19 Wilson MR, Easterbrook-Smith SB. Clusterin Binds By a Multivalent
Mechanism to the Fc and Fab Regions of IgG. Biochim. Biophys. Acta. 1159(3)
Figure 7:  Average SPR binding levels for 1 μM of antibody and antibody
fragments to immobilized CHO PLBL2 — the first three bars show the
binding response values of intact MAbs; the next three bars show
binding response values of the F(ab’)2 fragments of those MAbs; the final
three bars show binding response values of their Fc portions. Reference
flow-cell subtracted binding response values of samples to immobilized
CHO PLBL2 were reported as the response five seconds before the end of
each injection. Data are reported as average values, and error bars
represent standard deviations of four experimental replicates.
1,800
1,600
1,400
1,200
Response Units
1,000
800
1,459
600
400
766
200 85
−3
31 −2 9 −1 −1 −4
0
2 1 6 2 1 6 2 1 6 Buffer
−200
Monoclonal Antibody
1992: 319–326.
20 Zhang Q , et al. Comprehensive Tracking of Host Cell Proteins
During Monoclonal Antibody Purifications Using Mass Spectrometry. MAbs
6(3) 2014:659–670.
21 Chen SC, et al. Purification and Characterization of Secretory
Phospholipase B, Lysophospholipase, and Lysophospholipase/Transacylase
from a Virulent Strain of the Pathogenic Fungus Cryptococcus neoformans.
Biochem. J. 347(2) 2000: 431–439.
22 Santangelo RT, et al. Detection of Antibodies to Phospholipase B in
Patients Infected with Cryptococcus Neoformans By Enzyme-Linked
Immunosorbent Assay (ELISA). Med. Mycol. 43(4) 2005: 335–341.
23 Bailey-Kellogg C, et al. CHOPPI: A Web Tool for the Analysis of
Immunogenicity Risk from Host Cell Proteins in CHO-Based Protein
Production. Biotechnol. Bioeng. 111(11) 2014: 2170–2182.
24 Rosenberg AS, Worobec A. Immunogenicity Concerns of Therapeutic
Protein Products, Part 2: Considering Host-Specific and Product-Specific
Factors Impacting Immunogenicity. BioPharm Int. December 2004.
25 Koren E, et al. Recommendations on Risk-Based Strategies for
Products. J. Immunol. Meth. 333(1–2) 2008: 1–9. •
Detection and Characterization of Antibodies Against Biotechnology
Corresponding author Martin Vanderlaan is in analytical operations at
Genentech (a Member of the Roche Group), 1 DNA Way, South San
Francisco, CA 94080; 1-650-225-4439, fax 1-650-225-5695; vdl@gene.
com. Julie Nishihara, George Tsui, Margaret Lin, Feny Gunawan,
Sara Parker, Robert Ming Wong, Karthik Veeravalli, Patrick
McKay, Chris Yu, Lori O’Connell, Benjamin Tran, Chris Fong, Judith
Zhu-Shimoni, and Denise Krawitz are in technical development;
Wendy Sandoval, Peter Liu, and Rajesh Vij are in research; Justin
Low, Xiangdan Wang, Jihong Yang, and Valerie Quarmby are in
development sciences; and Kathleen Francissen is in CMC regulatory
— all at Genentech, as well.
For electronic or printed reprints, contact Rhonda Brown of Foster
Printing Service, rhondab@fosterprinting.com, 1-866-879-9144 x194.
Download personal-use–only PDFs online at www.bioprocessintl.
com.
A pril 2015 13(4) BioProcess International
B i o P r o c e s s TECHNICAL

Rapid Development and Scale-Up

of Biosimilar Trastuzumab
A Case Study of Integrated Cell Line and Process Development

Kumar Dhanasekharan, Claudia Berdugo-Davis, Xiaoming Liu, Carl Richey,

Zaneer Segu, David Calabrese, Valerie LeFourn, Pierre-Alain Girod, and Victor Vinci

C
ompared with that for new Figure 1a:  Cell-line development process used herein, with clone selection in 96-well and
drugs, biosimilar expansion to 6-well plates performed manually.
development faces Limiting dilution
Screening 1 Fed-batch
significantly condensed Vector
construction
Transfection of clones,
Retransfection
of lead clones
and screening 2
of clones, then
evaluation in
Cryo-
preservation
timelines from cell line to first-in- then expansion
expansion
shake flasks

human (FIH) trials. A biosimilar

development program needs to
accelerate quickly toward preclinical Limiting ELISA and ELISA and ELISA, SDS-PAGE,
and phase 1 studies; phase 2 studies dilution, expansion to expansion to and expansion
96-well plating 24-well plates 6-well plates to shake flask
typically are not required because
dose response and other patient-
treatment concepts are already
Figure 1b:  Analysis of isolated clones in
established by the original, European Medicines Agency 96-well plates
comparator medicine. Phase 3 (EMA), the US Food and Drug
140
studies typically are limited to Administration (FDA), and the Screening 1 Screening 2
fewer patients, which ultimately World Health Organization 120
shortens overall timelines and costs. (WHO) (1–3). This often requires 100
Titer (mg/L)

The key challenge remains: multiple iterations during early 80

demonstrating comparability and
high similarity based on in-depth
analytics as outlined in a number of
guidance documents from the
Product Focus:  Biosimilars
(monoclonal antibodies)
Process Focus:  Production and
Downstream Processing
Who Should Read:  Process and
product development, manufacturing,
and analytical personnel
Keywords:  Genetic engineering, DoE,
cell line and process optimization,
scale-up, and process control
Level:  Intermediate
phase development compared to 60
that for an originator product (4). 40
An integrated approach to cell line 20
development and early stage 0
screening and process development 1 21 41 61 81
is therefore imperative for making Ranked Clones
appropriate decisions in pursuing
the right candidate for preclinical protein. Although recombinant
and clinical manufacturing. Here, technologies are extremely powerful
we present such an approach for a tools for expression of
biosimilar trastuzumab, from cell- biotherapeutics, manufacturing
line through early phase success depends on identifying cell
development with manufacturing- lines that stably produce high
ready titers and clinically relevant protein titers. A critical part of
product quality attributes. biopharmaceutical development is
Our platform begins with choosing the right cells for meeting
development of a high-producing the objective of confirming cell
Chinese hamster ovary (CHO) cell productivities, cell-line stability,
line expressing the desired target and final-product quality. The most

BioProcess International 13(4) A pril 2015

 Figure 2:  Cell culture process design for 2-L and 20-L studies
common gene expression systems
are based on dihydrofolate reductase
(DHFR) and glutamine synthetase
(GS). P2
P0 P1
However, identifying suitable
cell lines is traditionally a time-
25 20 L 20 L
consuming, labor-intensive process mL
100
3L
mL 10 L
because their productivity and
stability can vary enormously.
Several thousands of clones must be 2L 2L
screened to find those with both
the highest yield and the best
Figure 3:  Downstream process design for 2-L and 20-L studies
product quality traits. To improve
cell-line selection and reduce Protein A with Low-pH
CEX with Polishing AEX Viral
timelines, we developed a stable cell MAbSelect viral
POROS XS with POROS HQ filtration
UF/DF
Sure inactivation
line using the SUREtechnology
Platform system from Selexis SA. It
is based on DNA-based elements
(Selexis Genetic Elements, SGEs) operations, with comparable process commercial chemically defined
that control the dynamic and product attributes. Every step (CD) product from Irvine
organization of chromatin within in this process was supported by Scientific. To keep the environment
all mammalian cells. They allow for detailed analyses: e.g., N-glycan of the cells unchanged throughout
unwinding a transgene’s genomic analysis to demonstrate the whole procedure, we used the
environment — a prerequisite for comparability and quality, and size- same medium as a basal medium for
efficient gene expression. The SGE exclusion chromatography (SEC) transfection, single-cell cloning,
approach has been shown to prevent and capillary isoelectric focusing and production. We incorporated
transgene silencing and provide (cIEF) to confirm quality based on genes encoding the Hc and Lc and
high transcription rates for all size variants and charge variants. selection markers into SUREtech
copies of the transgene, thus Finally, we have developed a vectors containing SGEs to
enabling higher and more stable process ready for transfer to maintain transcription at the
expression of recombinant proteins. manufacturing using chemically maximum level.
Subsequently, we performed defined medium under an Figure 1a is a schematic of the
design-of-experiment (DoE) studies aggressive timeline, with a product development process we used to
using shake f lasks to identify base titer in the range of 3–4 g/L, while establish the cell line for use in this
media and feed media strategies, meeting product quality attributes study. DNA was delivered by
confirming the results at 2-L scale. comparable to those of the electroporation to improve
We further optimized the 2-L innovator molecule. transfection efficiencies. We
process to meet desired product applied an antibiotic selection to
titers and quality. A three-column M aterials and M ethods create the stable pools, picking
downstream process was developed We adapted a current good about 100 clones for manual
to yield high monomer content. We manufacturing practice (CGMP) expansion from 96-well to 24- and
successfully scaled up our process to compliant CHO-K1 ATCC six-well plates. Based on
a pilot-scale setup for both derivative host cell line to grow in immunoglobulin G (IgG) titers, we
upstream and downstream BalanCD Growth A media, a transferred selected clones to shake

Figure 4:  Media screening with temperature shift (left); media screening with no temperature shift (right)
3.0
Viable Cell Density (107 cells/mL)

100 3.0 100

Viable Cell Density (107 cells/mL)

90 90
2.5 2.5
80 Flask 6 80
2.0 70 Flask 3 70
Viability (%)

2.0
Viability (%)

60 60
1.5 Flask 10 50 1.5 50
40 40
Flask 11 Flask 4
1.0 30 1.0
Flask 12 Flask 8 Flask 1 Flask 7
30
Flask 13 Flask 14 20 Flask 2 20
0.5 0.5
10 Flask 5 10
Flask 9
0.0 0 0.0 0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
Time (days) Time (days)

A pril 2015 13(4) BioProcess International

Table 1:  Typical analytical control strategy for early stage development (Char. = characterization) Analytical Abbreviations
Analytical Control Strategy
CD: Circular dichroism
Critical Quality Attribute Analytical Method In-Process Release Stability Char.
CE: Capillary electrophoresis
IgG Protein A HPLC × ×
Protein concentration UV-A280 × × × CE-SDS: Sodium dodecyl sulfate
Protein identification LC-MS peptide × ×
capillary electrophoresis
mapping cIEF: Capillary isoelectric focusing
Physicochemical LC-MS peptide × × DSC: Differential scanning calorimetry
characterization mapping
Biophysical analysis Far-UV CD × ×
ELISA: Enzyme-linked immunosorbent
assay
DSC thermal × ×
analysis FL-HPLC: High-performance liquid
Differential × × chromatography with fluorescence
fluorimetry detection
Functional analysis Binding ELISA × GC: Gas chromatography
Cell-based assays ×
HPLC: High-performance liquid
Residual DNA PCR × chromatography
Residual protein A Protein A ELISA × ×
iCE: Immunoaffinity capillary
Residual host-cell protein (HCP) HCP ELISA × ×
electrophoresis
Bioburden Plate-count × ×
methods ICP-MS: Inductively coupled plasma
Endotoxin LAL assay × × mass spectrometry
Charge variants (deamidation, IEX-HPLC × × × × IEX-HPLC: Ion-exchange high-
N-terminal pyroglutamates, cIEF or iCE × × × performance liquid chromatography
sialylation)
IgG: Immunoglobulin G
Size variants (aggregation, SEC-HPLC × × × ×
dimers, fragmentation) SDS-PAGE × × × ×
LAL: Limulus amebocyte lysate
CE-SDS × × × × LC-MS: Liquid chromatography–mass
Process impurities (Triton RP-HPLC × × spectrometry
X-100, residual solvents, GC × × PCR: Polymerase chain reaction
phosphenes, formulation
components) ICP-MS × × RP-HPLC: Reverse-phase high-
Glycan profile, intact protein FL-HPLC, LC-MS × × performance liquid chromatography
mass SDS-PAGE: Sodium dodecyl sulfate
Nonglycosylated product CE-SDS × × polyacrylamide gel electrophoresis
SEC-HPLC: Size-exclusion high-
f lasks, then further examined the Strategy Development: To evaluate a performance liquid chromatography
best performers during a 14-day set of commercially available cell UV-A280: 280-nm absorbance
fed-batch process. culture media with different feed ultraviolet spectroscopy
We accomplished those multiple options, we performed a basal media
steps of expansion without any screening study. To evaluate the
antibiotic selection, carefully media types for both basal and feed and confirm process parameters
characterizing and selecting clones media, total feed amount, feed such as temperature shift, pH, and
for attributes that would be duration, and temperature shift, we glucose levels, we ran two sets of
compatible with our production performed 14 shake-f lask studies four 2-L scale bioreactors,
requirements. Finally, we banked with 100-mL working volume (WV). evaluating two different
and tested 10 qualified clones for And we subsequently inoculated an temperature shift conditions (34 °C
cell identity and for endogenous and additional 12 shake f lasks to further and 32 °C). Glucose was fed up to 4
adventitious microbial contaminants optimize feeding strategy. g/L when it fell below 3 g/L; pH
(bacteria, fungi, and mycoplasma). To monitor those shake f lasks for was controlled using a sodium
To analyze their expression cell growth and viability, we used a bicarbonate–carbon dioxide system.
constructs at the nucleic-acid level Vi-Cell counter from Beckman The system controlled dissolved
(genetic stability), we assessed the Coulter; to monitor for metabolites, oxygen (DO) at a set point of 40%
constructs by quantitative we used a BioProfile Analyzer with a fixed air-f low rate and O2 on
polymerase chain reaction (qPCR). instrument from Nova Biomedical. demand. We defined our feed
Finally, we selected a single clone We took titer samples on day 3 and strategy, total feed amount, and
for further process development. then daily after day 5, finally feed duration based on results from
harvesting the f lasks when the our shake-f lask studies.
Process Development cultures reached ~60% viability. For the seed train, we thawed
and Scale-Up Cell Culture Process and passaged cells about every three
Basal Media Screening and Feed Development: To further evaluate days. For the last seed-train passage

BioProcess International 13(4) A pril 2015

Figure 5:  Effect of feeding timing and frequency on titer Figure 7:  Cell culture process scale-up; (a) viable cell density and viability,
(b) titer, (c) glucose profile, and (d) lactate profile
4.0
Flask 10
3.5 Flask 8
Flask 9 A 100
3.0 90
Flask 5 2.5

Viable Cell Density (107 cells/mL)

Flask 13 20-L (1), 1% feed D3−D18
2.5
Titer (g/L)

Flask 3 80
Flask 11 20-L (2), 1% feed D3−D18
2.0 Flask 4 2.0 70
Flask 2

Viability (%)
Flask 7
1.5 Flask 6 60
Flask 1 1.5
1.0 50
2-L golden batch
0.5 40
1.0
0.0 30
0 2 4 6 8 10 12 14 16 18 20 22 24
0.5 20
Time (days) 20-L (2), 1% feed D3−D12, 2% feed D13−18
10
20-L (1), 1% feed D3−D12, 2% feed D13−18
0.0 0
Figure 6:  Comparing the effects of temperature shift (TS) to 32 °C and 34 °C 0 2 4 6 8 10 12 14 16 18 20
from 37 °C on viable cell density, cell viability, and expression titer Time (days)
3.0 100 3.0
20-L (1), 1% feed D3−D18 B
90
Viable Cell Density (107 cells/mL)

20-L (2), 2% feed D3−D12, 1% feed D13−D18

2.5 2.5
20-L (1), 2% feed D3−D12, 1% feed D13−D18
80
20-L (2), 1% feed D3−D18
70 2.0
2.0

Titer (g/L)
Viability (%)
2-L TS 34 °C (1)
60
1.5
1.5 50
2-L TS 32 °C (1) 2-L TS 32 °C (2) 40 1.0
1.0
30
2-L TS 34 °C (2) 2-L golden batch
0.5
0.5 20
10
2-L Baseline 0.0
0.0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
0 2 4 6 8 10 12 14 16 18 20 22 24 26 Time (Days)
Time (days) 9
C
8 20-L (1), 1% feed D3−D18
(n – 1) before bioreactor inoculation, we expanded the 20-L (2), 1% feed D3−D18
7
cells in a Wave bioreactor from GE Healthcare 20-L (2), 2% feed D3−D12, 1% feed D13−D18
20-L (1), 2% feed D3−D12, 1% feed D13−D18
equipped with temperature control. Figure 2 shows this 6
Glucose (g/L)

2-L golden batch

process. 5
We sampled the culture daily to measure cell
4
density, viability, and metabolite levels. After day 3, we
sampled daily for analytical purposes (including titer 3
and quality attributes). Bioreactors were harvested 2
when cultures reached ~60% viability. Using a sterile
1
process, we collected the cell culture contents and
centrifuged them before filtering the supernatant 0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
through a 0.2-µm Pall Acropak 500 filter. Filtrate was Time (Days)
collected and stored in a –80 °C freezer for downstream 6
processing. 2-L golden batch
D
Cell Culture Process Scale-Up to Pilot Scale: To 5
demonstrate scalability, we scaled the 2-L process up to
20 L. We performed several 20-L bioreactor runs using 4
Lactate (g/L)

a Sartorius-Stedim BB30 stainless-steel bioreactor. The

system maintained scale-independent parameters (pH, 3
pCO2 , feed strategy, temperature shift, and so on) as 20-L (1), 2% feed D3−D12, 1% feed D13−D18
defined in the 2-L bioreactor cell cultures. Scale- 2
dependent parameters were scaled appropriately. We 20-L (1), 1% feed D3−D18
designed the air f low rate to a constant volume per 1
20-L (1), 1% feed D3−D18
volume per minute (vvm) that translated to 0.15 20-L (2), 2% feed D3−D12, 1% feed D13−D18
standard liters per minute (SLPM) with a ring sparger. 0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Agitation rate scale-up was based on constant power Time (Days)

A pril 2015 13(4) BioProcess International

per unit volume of f luid (P/V), which translated to 100 eluting the protein with 50 mM sodium acetate at pH
rpm. 3.4.
Downstream Development and Scale-Up: We After a viral inactivation step, we titrated the pool
developed and scaled up a three-column purification to pH 5.0. Next, cation exchange (CEX)
process to meet the requirements for our 20-L scale chromatography in bind–elute mode used POROS XS
bioreactor run. The downstream process starts with media from Life Technologies, with equilibration using
harvest clarification using a POD depth filtration 50 mM sodium acetate followed by a 4-CV wash at pH
system from EMD Millipore, followed by column 5.0. We used a gradient elution with sodium chloride
operations on a GE Healthcare ÄKTA Pilot system. from 75 to 270 mM over 20 CV. Then a polishing
The protein A capture step used MabSelect SuRe resin anion-exchange (AEX) chromatography used POROS
from GE Healthcare with phosphate-buffered saline HQ resin (also from Life Technologies) in f low-
(PBS) equilibration followed by loading and washing through mode, with a 3-CV chase at pH 7.8–8.0.
with PBS. After a secondary wash with PBS and 1 M Finally, we processed the eluate through a Virosart
sodium chloride, we ran a final wash with PBS before CPV virus filter followed by an ultrafiltration/

Figure 8:  Cell culture process optimized scale-up; (a) VCD and viability, (b) titer, (c) glucose profile, and (d) lactate profile

100 4.0
Viable Cell Density (×107 cells/mL)

A 90
B
2.5 3.5
20-L (2), 1% D3−18 2-L optimized: bolus feed
80
2-L golden batch 3.0
2.0 70 Viability (%) 2-L optimized: slow feed
20-L optimized: bolus feed
60 2.5
1.5 50 Titer (g/L) 2.0
20L (2), 1% D3−18
40
1.0 1.5
2-L optimized: bolus feed 30 20L-optimized: slow feed
2-L optimized: slow feed 1.0
5.0 20
20-L optimized, slow feed
20-L optimized, bolus feed 10 0.5
2-L golden batch
0.0 0 0.0
0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22 24
Time (Days) Time (Days)
8.0 6
20L-optimized, slow feed
2L-optimized: bolus feed C D
7.0
20L-optimized, bolus feed 5
20L-02 1% D3-18 2-L golden batch
6.0 2L-optimized: slow feed
4
Glucose (g/L)

Lactate (g/L)

2L golden batch
5.0 20-L (2) 1% feed D3−18
20-L optimized: slow feed
4.0 3
3.0
2
2-L optimized: bolus feed 20-L optimized: bolus feed
2.0
1 2L optimized: slow feed
1.0
0.0 0
0 2 4 6 8 10 12 14 16 18 20 22 0 2 4 6 8 10 12 14 16 18 20 22
Time (Days) Time (Days)

Figure 9:  Purity measured by size-exclusion high-performance liquid chromatography (SEC-HPLC) and charge variants measured by capillary isoelectric
focusing (cIEF) from Cook Pharmica’s 20-L run compared with the innovator molecule
140 140
Main Peak
120 Monomer 120
RS-Innovator
100
Wavelength (AU)

100
Wavelength (AU)

2-L Run

80 80
60 Acidic
60 Peaks
40 40 Basic
20 Peaks
20 Innovator
0 0
−20 20-L Run
−20
8 9 10 11 12 13 14 15 16 17 18
Time (minutes) 8 9 10 11 12 13 14 15 16 17 18
Time (minutes)

BioProcess International 13(4) A pril 2015

 Figure 10:  Comparing N-glycans released from the innovator product (black) and Cook Pharmica’s
diafiltration (UF/DF) operation in 20-L run (blue)
an ÄKTA Crossf low system from
GE Healthcare. 25.0
Analytical Development: Table 1

G0F
22.5
shows a typical analytical strategy
20.0
during early development. For the

Measurement? (units?)
current proof-of-concept study, we 17.5
considered a subset of those 15.0

G1Fa
methods as presented below. 12.5
10.0
R esults

G0F-GlcNAc
G0-GlcNAc
We made research cell banks 7.5

G1Fb

G2f+2SA
G2f+SA
(RCBs) from the single clone we

G2f
5.0

G0

M5

G1

G2
had selected. Figure 4 shows our 20-L Run
2.5
initial shake-f lask media screening Innovator
0.0
results. Those results helped us 8 9 10 11 12 13 14 15 16 17 18
identify basal media (Media A) and Time (minutes)
feed media (Feed B) with a Fucose Mannose Galactose N-acetyl-D-glucosamine Sialic acid

temperature shift (TS) to provide

the best performance in terms of
Figure 11:  Relative glycoform content from two different 20-L bioreactor runs; afucosylated and
viable cell density (VCD) and fucosylated glycoforms are comparable for both innovator and 20-L bioreactor runs.
viability over the duration of the
88.79 89.91 89.73
runs. Innovator
Our feed-strategy optimization 77.8 75.88
20 L (1)
results (Figure 5) indicate a 20-L (2)
nonlinear relationship between total 54.52
feed percentage and titer. Higher
total feed percentages seem to be 34.86
beneficial up to 16–20%; beyond
that point, however, additional feed 14.52 15.93
did not increase titer significantly. 5.65
8.33 7.27 7.22
Based on statistical analysis using 1.57 1.76

JMP software from SAS Institute, G0 G1 G2 Fucosylated Afucosylated

daily feeds from days 3 to 18 gave
the best titer results.
A temperature-shift study using
2-L bioreactors showed that the 34
°C condition helped cultures reach
higher VCDs, but that viability
declines more sharply and earlier
than it does at 32 °C. Results are
consistent between duplicates,
although a shift in growth rate was
observed in one run (Figure 6).
Based on protein A HPLC titer
measurements, cell culture
performance was slightly higher for
a TS to 32 °C than it was for a TS
to 34 °C.
We studied the combined effects
of pH dead-band and glucose
control to establish a baseline 2-L
process that we labeled as the
“golden batch.” Maintaining pH
above 6.8 led to higher cell growth
rates and higher peak VCDs, but
viability dropped to 60% or less
sooner. By contrast, maintaining
pH >6.6 results in cell cultures with whereas viability of the 2-L process
lower peak VCD but extended declines slower, so titer continues to
longevity of viability >60%. To go beyond 2.5 g/L.
achieve higher cell densities while We hypothesized that increased
ensuring extended longevity, we set mixing speed (P/V ) and slower
the pH dead band to 6.8–7.2 and additions of feed media and glucose
based our TS criteria on reaching a might be required to reduce
cell density of 1 × 10 6 cells/mL. variability in control of pH, feed,
And to minimize base consumption and glucose control. To test our
for pH control, we increased the hypothesis, we tried slow and bolus
post-TS pH dead band to 6.6–7.2. feed media and glucose additions
Cultures controlled at lower coupled with increased agitation
glucose levels (around 1 g/L) had rate. As part of scale-up
lower VCDs than those topped at 4 confirmation, we ran additional 2-L
g/L under the same pH conditions. cultures in parallel with the same
So we decided to maintain the increased agitation to correspond to
glucose levels at 4 g/L. a constant P/V and similar feed
Process Scale-Up Results: As addition schemes.
Figure 7 shows, metabolic profiles As the data show in Figure 8, in
for the process at 20-L scale all cases the peak cell density was
compare favorably with those at 2-L higher than it had been in previous
scale, although 20-L expression runs, and titer in one of the 20-L
titers f latten out at about 2 g/L runs reached 3.35 g/L. Although

A pril 2015 13(4) BioProcess International

the second 20-L run continued on a
similar track, a viability drop led us
to harvest the reactor early in that
case. One key noticeable feature in
these runs is the dramatic inf luence
on lactate profile of the controlled-
feeding strategy in later stages of
culture, in which the metabolism
shifted to lactate consumption.
We evaluated product quality
using size-exclusion high-
performance liquid chromatography
(SEC-HPLC) for size variants,
capillary isoelectric focusing (cIEF)
for charge variants, and N-glycan
analysis for glycosylation profiles.
Although purity was 99.8% after
three column steps, at the end of
the UF/DF step an overall purity of
99.1% monomer and host-cell
protein (HCP) <0.05 ng/mg were
reached with an overall product
yield of 77%.
As measured by SEC-HPLC, the
biosimilar purity is comparable with
that of the innovator molecule.
Additional product attributes —
such as the cIEF charge-variant
profiles — are also comparable
(Figure 9).
Because our monoclonal antibody
(MAb) molecule is a biosimilar
product, its N-glycan profile is an
important critical quality attribute
(CQA) for the design space
according to the quality by design
(QbD) paradigm. Figure 10 shows
the results of our comparison of
N-glycans released from the
innovator product (black) and Cook
Pharmica’s 20-L run product.
Clearly, the glycoforms present in
the innovator’s molecule are also
present in the material we
produced. That glycosylation profile
is reproducible and consistent across
multiple runs.
Relative afucosylation content in
both products is also comparable
(Figure 11). This is another
important CQA for a MAb such as
trastuzumab, which has bioactivity
based on its binding affinity toward
Fcγ receptors on effector cells and
clearance of target cells through
antibody-dependent cell-mediated
cytotoxicity (ADCC) activity.
However, fewer end-capped
BioProcess International 13(4) A pril 2015
galactosylated glycoforms appear to Scientific, and Patient Safety Perspectives.
•
J. Natl. Compr. Canc. Netw. 9, 2011: S1–S22.
be present in the biosimilar, and G0
appears to be relatively more
abundant than it is in the innovator.
Further iterative developments are Corresponding author Kumar
ongoing to improve comparability Dhanasekharan (Kumar.
and similarity through the use of Dhanasekharan@cookpharmica.com) is
different culture media director of process development;
supplements. Additional Claudia Berdugo-Davis is a senior
scientist, and Xiaoming Liu is an
characterization of the innovator
associate scientist in cell culture; Carl
molecule is also required to help us
Richey and Zaneer Segu are senior
understand lot-to-lot variability of scientists in biologics process
our biosimilar MAb. development and innovation; and Victor
Vinci is chief scientific officer and vice
A S uccessful Platform president at Cook Pharmica, 1300 South
We have demonstrated an Patterson Drive, Bloomington, IN 47402.
integrated cell-line development David Calabrese is director of cell line
effort with process development and services, Valerie LeFourn is head of cell
scale-up as a successful model for culture, and Pierre-Alain Girod is chief
expediting early phase development scientific officer at Selexis SA, 18 Chemin
des Aulx, 1228 Plan-les-Ouates,
for biosimilar antibodies. Our
Switzerland.
approach did not compromise on
the scientifically systematic
DoE-based approach, but it took a For electronic or printed reprints, contact
targeted set of experimental studies Rhonda Brown of Foster Printing Service,
to meet the requirements of product rhondab@fosterprinting.com, 1-866-879-9144
quality as well as process scale-up x194. Download personal-use–only PDFs
online at www.bioprocessintl.com.
needs. This approach was
successfully demonstrated using a
biosimilar Trastuzumab as a model
molecule, with scale-up from 2-L
bench scale through a 20-L pilot
scale process that will be scalable to
larger GMP operations for clinical
production.
R eferences
1 CBER/CDER. Guidance for Industry:
Scientific Considerations in Demonstrating
Biosimilarity to a Reference Product. US Food
and Drug Administration: Rockville, MD,
February 2012; www.fda.gov/downloads/
drugs/
guidancecomplianceregulatoryinformation/
guidances/UCM291128.pdf.
2 EMEA/CHMP/BMWP/42832/2005.
Guideline on Similar Biological Medicinal
Products Containing Biotechnology-Derived
Proteins As Active Substance: Non-clinical and
Clinical Issues. European Medicines Agency:
London, UK, 2006; www.ema.europa.eu/
docs/en_GB/document_library/Scientific_
guideline/2009/09/WC500003920.pdf.
3 Guidelines on Evaluation of Similar
Biotherapeutic Products. World Health
Organization: Geneva, Switzerland,
October 2009; www.who.int/biologicals/
publications/trs/areas/biological_
therapeutics/BS2110Dft_guidelines_Final_
HK_IK_29July_09.pdf.
4 Zelenetz AD, et al. NCCN
Biosimilars White Paper: Regulatory,
FOCUS  ON... QUALITY
Fundamental Strategies
for Viral Clearance
Part 1: Exploring the Regulatory Implications
by Kathryn Martin Remington
O
ver the past several decades,
biologics such as monoclonal
antibodies (MAbs) and
recombinant proteins have
provided therapeutic benefits and efficacy
for the treatment of human disease.
Completion of the human genome
project (launched in 1990) produced a
draft of the genome in 2001. A full
sequence was published on the 50th
anniversary (2003) of the initial
publication of Watson and Crick’s papers
on the double-helical structure of DNA
(1). That large volume of genetic
information has been translated into
usable scientific knowledge, helping
researchers in their hunt for better
therapeutic approaches and establishing
the functional relationships between
genes and clinical disorders. Ongoing
research has led to new understandings,
which have paved the way for many new
designs of recombinant biologics.
Today, most biopharmaceutical
products are proteins expressed in
recombinant hosts. Microbial systems are
attractive platforms because of their low
cost, high productivity, and rapid
implementation. And bacterial expression
systems present no concerns about
adventitious mammalian viruses.
However, many biopharmaceutical
molecules are simply too big or too
complex to be made by bacteria. For
example, properly folded proteins
(including MAbs with appropriate
glycosylation) require posttranscriptional
metabolic machinery that is available
only in mammalian cells. And because
such cells can add relevant human-like
H1N1 influenza virus
(WWW.VISUALSCIENCE.RU.COM)
posttranslational modifications, they have
become the preferred host for production
of complex protein therapeutics.
It has been more than a decade after
the completion of the human genome
sequence and more than 30 years after
regulatory approval of the first genetically
engineered drug. On 28 October 1982,
the FDA approved the first recombinant
medicine produced using genetic
engineering: recombinant human insulin,
Humulin, developed by Eli Lilly and
Company and Genentech. The drug was
expressed in bacteria. Today, such
biologics represent a major portion of the
pharmaceutical market, with most
approved in the United States and
Europe as glycosylated proteins produced
by mammalian cell culture.
Patents are expiring soon for more
than 20 first-generation blockbuster
biologics. As a result, an increasing
number of biosimilars (which can be
similar but not identical to innovator
biologics) are expected to be submitted
for approval. In addition, biobetters,
(considered improved versions of
innovator biologics) are gaining
popularity.
Pharmaceutical companies developing
biosimilars and biobetters must confirm
that the efficacy, safety, and purity of
their products are similar to those of
original innovator-approved drugs (2). In
this process, viral safety is of upmost
importance.
Because production of all
biopharmaceutical products carries an
inherent risk of potentially introducing
adventitious viruses, viral safety is an
essential consideration in the
development and manufacture of these
products. Viruses possessing diverse
characteristics can potentially
contaminate therapeutic agents, so
different methods of virus inactivation
and removal are required to guarantee a
safe drug supply.
Although viral safety is a focal point
for regulatory authorities, the potential
for transmission of diseases —
specifically viral — through any biologic
that uses a mammalian-derived
component in its manufacturing process
is a real risk.
Current state of the art precludes
claiming absolute absence of viral
presence in biologics. The practical
approach applied by most regulatory
bodies is evaluating safety case by case.
Such a strategy considers current, real,
and perceived virus contamination
concerns on the basis of source materials
used for manufacturing each product as
January 2015 13(1) BioProcess International
well as availability of adequate
technologies for detection and clearance
of viral agents.
Through international collaborations,
regulatory guidelines stress a holistic
approach that includes appropriate
sourcing of materials; demonstrating the
capability of a manufacturing process for
viral clearance; and testing in process,
among others. Critical use of an effective
viral safety program involves careful
selection and testing of cell lines and
original raw materials for adventitious
viruses. In addition, in-process testing
can identify contaminants such as
endogenous retroviral particles, which are
typically associated with mouse and
hamster cells. Viral testing is
complemented by viral clearance studies,
which demonstrate the capacity of
manufacturing steps to inactivate or
remove potential viral contaminants.
A Market in Transition
Over the past decades, the development
and production of medicinal products for
human use has undergone a major shift.
Among factors affecting the global
market for pharmaceutical manufacturing
is a paradigm shift in the need for
medicinal products. Demands and
patient needs are shifting from
blockbuster drugs to specialized, even
personalized medicine (3). Because many
innovator drugs are going off patent in
the next few years, the industry is moving
toward developing biosimilars and
biobetters.
A September 2013 report published
by IMS Health suggests that the global
market for biologics is expected to grow
from US$169 billion in 2012 to $221
billion in 2017 (4). The authors of that
report expect that biologics will continue
to outpace the growth of overall
pharmaceutical spending and expect it to
be driven by increased demand for MAb
innovator drugs and recombinant human
insulin. The authors also predict that by
2016, the market for biosimilars will be
worth about $5 billion, with expected
exponential growth afterward.
Complexity
The manufacturing process of biologics is
challenging and complex because such
products are derived from plants,
animals, or mammalian cells or
supplemented with reagents derived from
living systems or cell lines. The process
The quality, safety, and
efficacy of biologics
depend on careful
CONTROL of process
inputs and operating
conditions.
requires a high level of reproducibility in
commercial development and
manufacture to guarantee high-quality,
consistent biological products. The level of
complexity is generally influenced by a
host of factors, including individual cell
characteristics, growing environment, and
nutrients provided during manufacturing
(5–8).
Whereas synthetic small-molecule
pharmaceuticals rely on a chemical
analysis of components to assess purity,
the quality, safety, and efficacy of complex
biologics depend on careful control of
process inputs and operating conditions.
Processing involves many manufacturing
stages and parameters, as well as stringent
control mechanisms and process control
(9).
Although biopharmaceutical
production may be challenging,
guaranteeing virological safety is even
more complex. Aided by advanced
detection, donor selection, raw materials
screening, virus removal, and inactivation,
biomanufacturers have greatly increased
the safety profile of their products. Such
therapies include tissue-, blood- and
plasma-derived products, MAbs,
recombinant proteins, vaccines, and
animal products carrying an inherent risk
of transmitting viruses based on the
source material, manufacturing processes,
and routes of administration.
But despite stringent efforts on the
part of manufacturers, companies
continue to face threats of new and
emerging viruses. Unfortunately,
endogenous and adventitious viruses can
escape detection. That might be the result
of limited assay sensitivity or limitations
in detection methods, which require
increased efforts of viral clearance. As
part of such work, viral clearance studies
are designed to help demonstrate virus
reduction, achieved by a number of steps
in the manufacturing process in case
process intermediates are contaminated.
Vulnerability
Sourcing of raw materials is one of the
greatest concerns for regulators and
biopharmaceutical manufacturers. Most
incidents of viral contamination result
from using poorly characterized
materials. This risk is intensified by the
sourcing of raw materials from multiple
vendors. And because large volumes of
process media and gases are used to
produce biologics, viral contamination is
not necessarily a rare occurrence.
Evaluation of raw materials has
become a critical component of any risk
mitigation strategy. Such a strategy can
require evaluating source materials
beyond what can generally be expected or
is reported about the source. For example,
many processes now limit or entirely
move away from animal‑derived
materials for producing recombinant cell
lines or media formulations. But there is
still a risk that drug products have been
exposed to animal-derived materials at
one or multiple points in the supply
chain. Furthermore, it is not always
possible to move away from animal-
derived materials, which presents a risk of
introducing virus contamination into a
manufacturing system.
Rigorous evaluation and mitigation
are required to eliminate potential risks
of contamination. Some manufacturers
use virus inactivation methods such as
UV-C inactivation, high-temperature
short-time (HTST) treatment, or
removal (virus reduction filtration)
procedures for culture media or media
components. The viral reduction
potential of such risk‑mitigation
procedures can be evaluated in a viral
clearance study. Similarly, the viral
reduction potential of downstream
manufacturing process steps (designed to
complement rigorous upstream screening
and mitigation steps) also can be
evaluated in a viral clearance study.
Lessons learned from past incidents
show the importance of thoroughly
understanding the complexity of (and
possible risks in) the supply chain.
Regulatory authorities can use such
understanding — as well as the
knowledge of the extent to (and the
nature in which) biologics include
animal-derived components — to assess
how much viral clearance is required.
Mammalian Cells
With its intrinsic complexity, the
production of recombinant
biopharmaceuticals poses a series of
unique manufacturing challenges. Viral
safety is particularly challenging when
mammalian cells are used. Although
iatrogenic incidents have occurred in the
past decade as a result of biologics
contamination, incomplete inactivation of
viruses, or contamination of excipients in
a nonrecombinant product, only a limited
number of known and reported events
have taken place. (To date there has been
no transmission of an infectious virus to a
patient from a recombinant biotech
product. However, such transmissions
have occurred with plasma-derived
products, and some have resulted in
disease. Furthermore, although
contamination events for products
derived from mammalian cells have been
reported, they were detected before the
drugs were administered.)
In 2010, Kerr and Nims reported that
during a specific 10-year period, biologics
derived from a number of mammalian
cell culture processes were evaluated
using in vitro virus screening assays.
Reported results showed that only
reovirus type 2 and Cache Valley virus (a
bunyavirus contaminant) were detected
in biologics manufactured in Chinese
hamster ovary (CHO) cells. Other
reports showed that the murine
parvovirus mouse minute virus and
contamination with vesivirus 2117 had
also been detected in biologics
manufactured using CHO cells (10).
Myth and Solution
As a result of stringent regulatory
oversight, biologics manufacturers are
required to show that their products are
safe for human use. And although
biopharmaceutical scientists and
regulatory authorities understand that
absolute freedom of viral contamination
is impossible, much can be done to
ensure viral safety. This is reflected in the
regulatory landscape.
The International Conference on
Harmonisation of Technical
Requirements for Registration of
Pharmaceuticals for Human Use (ICH)
is a unique collaboration of regulatory
authorities and the pharmaceutical
industries of Europe, Japan, and the
United States. The organization has
developed guidelines requiring
manufacturers of biologics to demonstrate
their ability to remove or inactivate viral
contaminants (11–14).
The European Medicines Agency
(EMA), US Food and Drug
Administration (FDA), and other
authorities have adopted guidelines that
focus on testing and evaluating viral
safety of biotechnology products derived
from characterized cell lines of human or
animal origin. Those guidelines provide a
general framework for testing production
cell banks and bulk harvests for viral
contaminants and designing viral
clearance evaluation studies. The EMA
also has adopted guidelines concerning
clinical trials and submissions based on
an investigational medicinal product
dossier that includes viral safety
evaluation (15).
Investigational Products
The EMA has provided a regulatory
guideline governing virus safety for
investigational medicinal biotechnology
products (15). This document, which
became effective in 2009, applies to
clinical trial material derived from well-
characterized cell lines and harmonizes
the regulatory expectations throughout
the European Union. This guideline
acknowledges that, for clinical trial
materials, the manufacturing process and
full product characterization is still in
development. It offers reduced viral safety
testing and viral clearance study scope
compared with requirements for
marketing authorization provided in
ICH Q5A.
The viral safety of products intended
for clinical trials is a priority. However,
the EMA guideline acknowledges that
working cell banks may not yet be
established, and it lifts the requirement
for testing cells at the limit of in vitro
age. For viral clearance studies, a reduced
virus panel can be used (a retrovirus and
a parvovirus). Robustness studies are not
required for products at this stage of
development, again acknowledging that
process parameter ranges may not be
established yet. If a product used for a
clinical trial has been manufactured
using relatively new chromatography
resins, then clearance studies on aged
resins and resin sanitization studies are
not expected.These guidelines were
developed with input from industry and
other regulatory agencies that worked
together to assure viral safety.
Risk Management
A risk-based approach to viral safety is
consistent with the approach that is now
used to ensure quality in
biopharmaceuticals and is outlined in
several guidance documents. Such
documents include ICH Q8, which
includes process analytical technology
and quality by design; ICH Q9, which
includes quality risk management; and
ICH Q10, which presents guidelines on
pharmaceutical quality system (12–14).
Use of a risk-based approach to viral
safety is consistent with the viral safety
“tripod.” Selection and characterization
of source materials will define potential
contaminants that may be present. New
technologies such as massively parallel
sequencing are powerful tools for
detecting and identifying contaminants
that may be present, especially for novel
or complex source materials. The
resulting information can provide a
framework for more directed routine
testing.
Testing provides key information
about the design of a viral clearance
study. Understanding the potential
contaminants will guide selection of a
virus panel for study so that models of
potential contaminants are included.
Testing data also will direct the level of
clearance that is expected from
manufacturing processes. For example,
retroviral-like particles are quantitated in
bulk harvests from CHO cell lines. Viral
clearance for a process should be high
enough that the risk of that contaminant
in a final dose of product is less than one
in a million.
For many years, the focus of viral
clearance has been on downstream
manufacturing. Although downstream
viral clearance steps are still essential,
many manufacturers are looking at
upstream viral clearance steps. Including
a viral inactivation or removal step for cell
culture media or media components is
another way to mitigate the risk of viral
contamination. Steps such as UV-C
inactivation, HTST, and even virus
reduction filtration can provide yet
another layer of viral safety for a
manufacturing process.
Facility Design
One often-underestimated approach to
viral safety is the design of a
January 2015 13(1) BioProcess International
manufacturing facility. To comply with
regulatory directions, it is important to
consider where viral clearance takes place
in the production and manufacturing
process and which steps provide effective
viral inactivation or removal. Well-
designed facilities ensure adequate
segregation of process intermediates that
have been through an effective viral
reduction step from those intermediates
that have not. Also, dedicated viral
clearance must be planned in the
manufacturing process to alleviate safety
concerns.
As demonstrated by viral incidences
with vesivirus 2117, the possible
introduction of undetected viruses from
unknown sources is a great risk. Once a
viral contaminant enters the
manufacturing process, the only effective
control involves robust viral inactivation
or removal steps.
Lessons Learned
The most important lessons to be learned
from regulatory guidelines —and a
general principle that can be applied to
the manufacture of all
biopharmaceuticals — is that viral
clearance studies must be reviewed case
by case. Each study involves product-
specific risk factors. If a modular
approach to viral clearance is used for
similar products and processes, you must
demonstrate the absence of product-
specific influences on viral clearance.
That may involve verification of
processing parameters, analytical testing
of process intermediates, and/or some
evaluation of viral clearance. Testing
source materials will help you identify
potential contaminants and the level of
virus reduction to be achieved in
manufacturing.
Past incidents have shown that the
correct approach ultimately helps prevent
costly mistakes. Such mistakes can delay
regulatory approval of new biologics and
halt current production and
manufacturing of already approved
drugs.
Importance of Viral Clearance
Effective viral clearance is required to
guarantee the availability of a safe supply
of drug products for human use. Because
adventitious viruses present in original
materials can contaminate biologics
during processing, manufacturers must
BioProcess International 13(1) January 2015
follow regulatory requirements to ensure
adequate virus removal. During the past
decade, viral clearance studies have
received increased regulatory scrutiny and
become important factors to consider in
the design of manufacturing processes.
Viral clearance in downstream
purification can involve a number of
methods to inactivate or remove viruses
from a final therapeutic product. Because
contamination incidents also can occur in
bioreactors (adding associated regulatory
implications and costs associated with
decontamination), some manufacturers
have chosen to also include upstream
barriers in their processes.
Biomanufacturing processes are varied
and can include a number of upstream
and downstream steps. Virus-removing
(nano-) filters (designed for the removal
of almost all virus types through a size-
exclusion mechanism and
chromatographic resins in column or
membrane configurations) are typically
included downstream. Upstream steps
may include UV-C inactivation and
HTST. Exposure to low pH or solvent/
detergent treatment steps provides
inactivation of enveloped viruses
downstream.
In addition to viral clearance, a typical
manufacturing process also contains steps
to purify a biopharmaceutical product.
Athough such steps are not necessarily
designed to provide virus reduction, they
may have the potential to do so.
Chromatography is a good example of a
purification operation that can help
achieve viral clearance. However, because
a number of operating parameters can
influence viral reduction,
chromatography is generally less robust
than viral reduction by other methods
such as inactivation.
In the conclusion of this two-part
series, I discuss effective planning of viral
clearance studies as well as the
implementation of viral clearance
strategies.
References
1 Collins FS, et al. A Vision for the Future
of Genomics Research. Nature 422, 2003: 835–
847.
2 Islam R. Bioanalytical Challenges of 1-866-879-9144 x194. Download personal-
Biosimilars. Bioanalysis 6(3) 2014: 349–356.
3 Weiler A. A New Dawn for Western
CMOs. Pharm. Technol. 1 February 2014: S14–S17.
4 The Global Use of Medicines: Outlook Through
2017. IMS Institute for Healthcare Informatics:
Danbury, CT, September 2013.
5 Vyas VV, et al. Clinical Manufacturing of
Recombinant Human Interleukin 15: Production
Cell Line Development and Protein Expression in
E. coli with Stop Codon Optimization. Biotechnol.
Prog. 28(2) 2013: 497–507.
6 Hofland P. An Introduction to Recombinant
Biologics and Immunotherapy. TVCC, Almere, The
Netherlands, January 2003.
7 Schellekens H. Biosimilar Therapeutics —
What Do We Need to Consider? NDT Plus 2(sup
1) 2009: i27–i36.
8 Zhang J. Mammalian Cell Culture for
Biopharmaceutical Production. Manual of
Industrial Microbiology and Biotechnology, Third
Edition. Baltz RH, Davies JE, Demain AL, Eds.
American Society of Microbiology: Washington,
DC, 2010; 157–178.
9 Lee JF, Litten JB, Grampp G.
Comparability and Biosimilarity: Considerations
for the Healthcare Provider. Curr. Med. Res. Opin.
28(6) 2012: 1053–1058.
10 Kerr A, Nims R. Adventitious Viruses
Detected in Biopharmaceutical Bulk Harvest
Samples Over a 10-Year Period. PDA J. Pharm. Sci.
Technol. 64(5) 2012: 481–485.
11 ICH 5A: Viral Safety Evaluation of
Biotechnology Products Derived from Cell Lines of
Human or Animal Origin. The International
Conference on Harmonisation of Technical
Requirements for Registration of Pharmaceuticals
for Human Use: Geneva, Switzerland, October
1997.
12 ICH Q8: Pharmaceutical Development. The
International Conference on Harmonisation of
Technical Requirements for Registration of
Pharmaceuticals for Human Use: Geneva,
Switzerland, August 2009.
13 ICH Q9: Quality Risk Management. The
International Conference on Harmonisation of
Technical Requirements for Registration of
Pharmaceuticals for Human Use: Geneva,
Switzerland, November 2005.
14 ICH Q10: Pharmaceutical Quality System.
The International Conference on Harmonisation
of Technical Requirements for Registration of
Pharmaceuticals for Human Use: Geneva,
Switzerland, June 2008.
15 Guideline on Virus Safety Evaluation of
Biotechnological Investigational Medicinal Products.
UK, July 2008. •
Evaluation of Medicines for Human Use: London,
Kathryn Martin Remington, PhD, is a
principal scientist, development services,
clearance, at BioReliance; kathy.remington@
bioreliance.com.
For reprints, contact Rhonda Brown of Foster
Printing Service, rhondab@fosterprinting.com,
use–only PDFs at www.bioprocessintl.com.