Professional Documents
Culture Documents
Hu 2017
Hu 2017
DOI 10.1002/btpr.2444
Zhilan Hu1#, Danming Tang1#, Shahram Misaghi1, Guoying Jiang2, Christopher Yu3,
Mandy Yim1, David Shaw1, Brad Snedecor1, Michael Laird1, and Amy Shen1*
1
Department of Early Stage Cell Culture; 2 Biological Technologies; 3 Protein Analytical
Chemistry, Genentech Inc., 1 DNA Way, South San Francisco, CA 94080, USA
*Correspondence to:
Key words:
CHO cell; Therapeutic monoclonal antibody; Charge variants; C-terminal lysine, Proline
amidation, Productivity
Running title: Lower productivity of C-terminal Lys deleted antibody expressing cell
lines.
This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process which may lead to
differences between this version and the Version of Record. Please cite this article as
doi: 10.1002/btpr.2444
© 2017 American Institute of Chemical Engineers Biotechnol Prog
Received: Dec 01, 2016; Revised: Jan 25, 2017; Accepted: Jan 27, 2017
Abstract
Monoclonal antibodies (mAbs) have been well established as potent therapeutic agents
and are used to treat many different diseases. During cell culture production, antibody
charge variants can be generated by cleavage of heavy chain (HC) C-terminal lysine and
antibodies with deletion of either HC C-terminal lysine (-K) or lysine and glycine (-GK).
Interestingly, clones that express antibodies lacking HC C-terminal lysine (-K) had
considerably lower specific productivities compared to clones that expressed either wild
type antibodies (WT) or antibodies lacking HC glycine and lysine (-GK). While no
secretion were observed, our analysis suggests that the lower specific productivity of
clones expressing antibody lacking HC C-terminal lysine was due to slower antibody HC
Introduction
human therapeutic agents for a wide range of human diseases, including many types of
cancer 1, 2. Process changes in production scale, media, and cell culture process
production batches 3-5. Monoclonal antibody products are heterogeneous and may
attributed to lack of process control 4, 5. One antibody product quality which requires
Therapeutic antibodies of IgG1, IgG2 and IgG4 isotypes have a basic lysine residue at
their C terminus (C-terminal lysine). However, the C-terminal lysine is often not detected
or detected at low levels in the final drug products because of peptidase action during
antibody isoforms bearing zero or one C-terminal lysine residues on each heavy chain
(HC). The HC of the human IgG1 antibody isotype ends with proline-glycine-lysine
amidating lyase (PAL), further catalyzes the hydroxylation of glycine and removal of the
glyoxylate from the glycine residue, leaving an amidated C-terminal proline (Figure 1A)
12-14
. Both lysine deletion and proline amidation of C-terminus of antibody HC can
In order to reduce the antibody charge variants heterogeneity and minimize variations in
the level of charge variants among production batches, we decided to generate antibodies
lacking HC C-terminal lysine (-K) or both lysine and glycine (-GK) (Figure 1B). The HC
C-terminal lysine of wild type (WT) antibodies is mostly removed during the mammalian
cell culture production phase and the remaining C-terminal lysine residues are removed
charge variant levels do not affect its potency, FcRn binding, or PK properties 6.
Therefore, antibody lacking HC C-terminal lysine should have the same potency as its
WT counterpart. Indeed, our prior evaluations did not reveal any differences in stability,
lacking either HC C-terminal lysine (-K) or both lysine and glycine (-GK)15. In this study,
and product quality using three different monoclonal antibody molecules (mAbs). We
observed that the antibodies lacking HC C-terminal lysine had increased basic variants
due to C-terminal proline amidation. Additionally, clones that express antibodies lacking
HC C-terminal lysine had lower specific productivities and titers compared to clones that
expressed WT antibodies. Interestingly, the clones that express antibodies lacking both
glycine and lysine had higher productivity and reduced levels of basic variants (by 90%)
antibody or antibody lacking HC lysine. Our results indicate that the lower specific
Results
Clones expressing antibodies lacking HC C-terminal lysine exhibit lower titers and
specific productivities
Plasmids for two different antibody mAb1 and mAb2 bearing HC C-terminal lysine or
glycine-lysine deletions, along with their WT versions, were transfected into CHO cells
and 216 clones from each expression construct were randomly picked and assayed for
antibody titers. Pools of the top 12 highest titer clones from each expression construct
were evaluated in a fed-batch production assay. Pools for both mAbs bearing HC lysine
deletions had lower specific productivities (Figure 1C) and titers (Figure 1D) relative to
(mAb3), we transfected CHO cells with plasmids of WT, HC C-terminal lysine deleted
and glycine-lysine deleted versions of mAb3. For each version, we analyzed antibody
expression for a total of 528 clones that were randomly picked per construct (Figure 2A)
and again the lysine deleted version showed lower titers relative to the WT or glycine-
lysine deleted versions. The top 96 clones were further assayed (data not shown) and
subsequently the top 20 individual clones of each expression construct were analyzed in a
14-day fed-batch production assay. The results confirmed that mAb3 clones expressing
lysine-deleted antibody indeed have overall lower titers due to lower specific
antibodies expressing clones were almost 2-folds higher than the clones expressing lysine
deleted antibodies (Figure 2B & C). We noted that all different versions of the mAb3
allow comparable cell growth as indicated by Day 14 integrated viable cell count (IVCC)
(Figure 2D). The data collected thus far for all three antibodies shows that, although the
same expression vector was used to express different antibodies, the clones expressing
that this phenomenon is not a clone or antibody specific effect and is indeed triggered by
We then used only mAb3 as a model molecule for further analysis. To eliminate clone-
specific variation in antibody production, we pooled the top 20 mAb3 expressing clones
of each version and performed a fed-batch production assay. Again, the pool of clones
expressing antibody lacking HC C-terminal lysine had lower titer throughout the
and lysine (Figure 3A). Although Day 14 IVCCs were similar among the three pools
(Figure 3C), the Day 14 specific productivity of the lysine-deleted antibody expressing
pool had decreased 32.3% and 50.2% compared to the WT and glycine-lysine-deleted
terminal lysine or glycine-lysine deletion lowers the level of antibody basic variants.
Interestingly, we observed that the antibodies bearing lysine deletion had 18.5% basic
variants, which was even higher than that for the WT antibodies (15.6%) (Figure 3E).
While levels of uncleaved C-terminal lysine in the lysine deleted version of antibody
were reduced to 0%, an increase in levels of proline amidation (16.9%) accounts for an
glycine-lysine double deleted antibodies with only 3.5% basic variants displayed the most
uniform HC C-terminus (-SLSLSP) (Figures 3D&E). Similar trends were observed when
%basic variants were compared for individual clones of each expression construct
(Figure 3F).
degradation
To determine the root cause of lower specific productivity of clones expressing antibody
lacking HC C-terminal lysine, we used the pool of the top 20 mAb3 clones for each of the
WT, lysine deleted, and glycine-lysine deleted versions of antibody and analyzed the
intracellular antibody levels by Western blot analysis. We observed that the intracellular
level of HC molecules in the cells expressing lysine deleted antibody was lower than that
comparable across all antibody versions (Figure 4A). Measuring HC and LC mRNA
levels revealed that the lysine deleted version had slightly higher HC mRNA levels
relative to the WT version, with the glycine-lysine deleted version having the highest HC
mRNA level among all three versions (Figure 4B, top). Levels of antibody LC mRNA
were comparable across all versions (Figure 4B, bottom). Therefore, altered mRNA
levels do not account for the observed low levels of intracellular HC bearing C-terminal
deleted, and glycine-lysine deleted versions. Considering that the lysine-deleted version
Western blot signal ratios revealed that HC bearing lysine deletion has ~2-fold higher
bearing glycine-lysine deletion (Figure 4C, compare lane 4 to lanes 2 and 6).
residing in the ER, their N-linked glycans are sensitive to EndoH digestion but once in
the Golgi, due to glycan modifications they become resistant to EndoH digestion.
trafficking patterns among the three versions of antibody, since comparable banding
patterns among all antibody versions were obtained upon EndoH or PNGaseF treatment
(Figure 4D). Lack of EndoH resistant antibody species shows that antibody molecules are
Clones expressing C-terminal lysine deleted antibody have lower antibody synthesis
rates
We also investigated rates of antibody synthesis for WT antibody and antibody lacking
C-terminal lysine or glycine-lysine. Cells were treated with cyclohexamide (CHX) to stop
protein synthesis and within 6 hours almost all the intracellular antibody molecules were
secreted from the cells into the media (Figure 5A). Removal of CHX restored protein
synthesis and levels of intracellular and secreted antibodies were monitored at 0 and 2
hours (Figure 5B). For the cells expressing antibody lacking glycine-lysine, maybe
because the intracellular HC levels were higher and overwhelmed secretion capacity,
there was some HC molecules left in the cells even after 6 hours of CHX treatment
(Figure 5A). Hence, there was also higher levels of HC molecules at the beginning of the
CHX washout step (Figure 5B). Within 2 hours, antibodies started to accumulate in the
cell and small amounts were secreted into the cell culture media. While levels of antibody
secretion for both WT and HC C-terminal lysine deleted versions were minimal (Figure
5B, lane 2 vs lane 4 and Figure 5C), intracellular accumulation of newly synthesized
lysine deleted antibody HC was only 42% of that of WT and was much lower than
suggest that for antibody bearing glycine-lysine double deletion, the HC translation,
translocation and folding is the fastest (Figure 5C). This may be partially due to its higher
mRNA level (Figure 4B). The faster production of folded antibody could explain the
higher titer and specific productivity of clones expressing the glycine-lysine deleted
version of mAb3. Our findings suggest that the HC translation/translocation (and folding)
rate is slower for HC lacking C-terminal lysine relative to WT, and HC lacking both
glycine and lysine. In sum, clones expressing antibody lacking HC C-terminal lysine
Discussion
In this study, we have shown that clones expressing antibodies with C-terminal lysine
deletion consistently have lower titers, a phenomenon that was observed in three different
test cases. After ruling out altered mRNA levels accounting for the observed protein level
differences, we found that the low productivity of these clones seems to be due to lower
antibody peptide synthesis, translocation and/or folding rates combined with a higher rate
terminal amino acid plays a role in regulating the release of a newly synthesized HC
molecule from the ribosome, or its translocation into the ER lumen and subsequent
proline (glycine-lysine deleted version) residue, can more easily leave the ribosome/ER
membrane interface and enter the aqueous ER lumen than a nascent HC polypeptide with
a glycine C-terminus (lysine deleted version). Once it folds and exits the ER, the lysine
deleted version of antibody is readily processed and transported out of the cell, as
indicated by EndoH treatment (Figure 4D) and the CHX washout experiment (Figure 5).
We don’t think the low specific productivity in –K construct is due to its poor binding to
the protein A/G column, because it has been shown via crystallographic refinement that
the primary binding site for protein A is on the Fc region, between the CH2 and CH3
domains 16.
We also observed that deletion of C-terminal lysine increased basic variant levels due to
more prevalent proline amidation (Figure 3E). In WT HC, proline amidation requires
removal of the C-terminal lysine by endogenous carboxypeptidase D11. However, for the
HC bearing a C-terminal lysine deletion this step has already been bypassed, leaving it
more prone to proline amidation. On the other hand, clones expressing antibody lacking
both HC C-terminal lysine and glycine have the lowest levels of C-terminal amino acid
noteworthy to keep in mind that all the antibodies tested here are of IgG1 isotype and for
now our findings with regards to –K and –GK versions of antibodies only apply to this
isotype.
Considering that each of the mAbs 1, 2, and 3 have different targets and hence different
CDR sequences, we are fairly confident that the observed higher titers of –GK mAbs is
not antibody specific. Based on this, we suggest that -GK version of an antibody could
heterogeneity and the need for monitoring this variable during manufacturing. Since
complement dependent cytotoxicity (CDC) occurs only when antibody is injected into the
bloodstream and the C-terminal lysine is removed, both WT (due to function of carboxyl
peptidases) and –GK versions of an antibody can equally accommodate this purpose. If
engineered 17. While our initial investigations suggested that deletion of C-terminal lysine
the glycine-lysine deletion in antibodies will need to be tested in order to ensure patient
Antibody-producing cell lines were generated at Genentech and derived from the
125 mL shake flask vessels shaking at 150 rpm, 37oC and 5% CO2. Cells were passaged
terminal deletion mutants of antibody heavy chain were designed similarly as WT only
Total RNA from cells was isolated using the RNeasy 96 kit following the manufacturer’s
protocol (Cat# 74181, Qiagen) and was treated with DNase (Cat#79254, RNase free
DNase kit, Qiagen) to remove any residual DNA that may be present in the isolated RNA
sample. qRT-PCR was performed using a universal qRT-PCR master mix according to
antibody heavy and light chains were normalized to mRNA levels for the housekeeping
gene β-microglobulin.
Heavy chain forward primer: TCA AGG ACT ACT TCC CCG AAC
Heavy chain reserve primer: TAG AGT CCT GAG GAC TGT AGG ACA GC
Heavy chain probe: FAM- ACG GTG TCG TGG AAC TCA GGC GC- TAMRA
Light chain forward primer: TGA CGC TGA GCA AAG CAG AC
Light chain reserve primer: CAG GCC CTG ATG GGT GAC
Light chain probe: FAM -ACG AGA AAC ACA AAG TCT ACG CCT GCG A-
TAMRA
CHO β-microglobulin forward primer: TCC TCT CAG TGG TCT GCT TGG
CHO β-microglobulin reserve primer: TGG CGT GTG TAG ACT TGC ACT T
CHO β-microglobulin probe: FAM - TGC CAT CCA GCG TCC CCC A - TAMRA
Western blot analysis was performed in order to determine the antibody heavy and light
chain expression levels. β-actin was used as the endogenous control protein. Cell pellets
were collected and lysed in lysis buffer (Cat# 9803, Cell Signaling Technology, Danvers,
MA). Total protein content of the lysed supernatants was quantified by Bradford assay
(Cat#1856210, Thermo Scientific, Rockford, IL). The samples were heated at 95°C for
3 min, and loaded onto SDS–PAGE gel and transferred to nitrocellulose membranes.
Membranes were then incubated with the indicated antibodies (anti-HC antibody,
Twenty-clone pool seed train samples from each construct were used for MG132
treatment studies. Seed train cells were seeded at 2 x 106 cells/mL for a total of 5 mL (10
x 106 cells) and treated with or without MG132 (Cat#474791, Sigma) at 10 µg/mL for 6
hours. At the end of the incubation, cells were spun down and analyzed by Western blot
Cells were lysed in lysis buffer (10 mM Tris, pH 8.0, 150 mM NaCl, 5 mM MgCl2, 0.5%
NP-40, protease inhibitor). Monoclonal antibodies were pulled down by protein A/G
the Endo H (Cat# P0702S) and PNGase F (Cat# P0704L) manufacturer (Bio Labs) for 10
min. 30 µL of the denatured antibody was kept as non-treated sample, the rest was
divided and transferred to two new tubes, incubated with either PNGase F or EndoH as
instructed at 37 oC for 1 hr. The samples were then analyzed by SDS-PAGE and Western
Charge variants were measured via capillary iso-electric focusing (cIEF). Liquid
chemically defined medium together with bolus feeds on days 3, 7, and 10. A
temperature shift from 37°C to 35°C was carried out on day 3. Day 14 titers were
and viable cell count was determined using Vi-Cell XR instrument (Beckman Coulter) 18.
Cells (20 x 106) from a pool of the top-20 clones expressing either WT, -K or –GK
antibody were washed with PBS, pelleted and resuspended into 30 mL seed train media
containing 20 µg/mL cyclohexamide (Cat# C7698, Sigma) and grown in shake flasks. At
0 hr, 3 hr and 6 hr time points, 2 x 106 cells (3 mL) were sampled and supernatants were
separated from the cells by centrifugation. After 6 hours, the remaining cells were
pelleted, washed with PBS and resuspended into 21 mL of seed train media in the
sampled and cell pellets and supernatants were separated. All collected cell pellets were
washed with PBS and lysed in 1 mL lysis buffer (20 mM Hepes, pH 7.5, 150 mM NaCl,
14,000 rpm (20,000 × g) for 10 min. Antibodies from both cell lysates and supernatants
were immunoprecipitated by protein A/G beads for further analysis including SDS-PAGE
Acknowledgements:
We would like to thank the Analytical Operations group at Genentech for their excellent
service and support, and Mike Laird and Steve Lang for their critical review of the
Figure legends
A-C. Titer (A), specific productivity (B), and IVCC (C) of the pools of the top 20 clones
expressing indicated versions of the antibody mAb3. Each bar represents the average
from two replicate production experiments. Each error bar denotes 1 standard deviation.
D. Percentage and type of basic variants of the pools of the top 20 clones characterized
by mass spectrometry (ND = Not Detectable).
E. Percent of basic variants of the pools of the top 20 clones expressing indicated
versions of the antibody mAb3.
F. Percent of basic variants of the single clones expressing indicated versions of the
antibody mAb3.
References
1. Reichert, J. M., Monoclonal antibodies in the clinic. Nat Biotechnol 2001, 19, (9),
819-22.
2. Adams, G. P.; Weiner, L. M., Monoclonal antibody therapy of cancer. Nat
Biotechnol 2005, 23, (9), 1147-57.
3. Carson, K. L., Flexibility--the guiding principle for antibody manufacturing. Nat
Biotechnol 2005, 23, (9), 1054-8.
4. Liu, H.; Gaza-Bulseco, G.; Faldu, D.; Chumsae, C.; Sun, J., Heterogeneity of
monoclonal antibodies. J Pharm Sci 2008, 97, (7), 2426-47.
5. Vlasak, J.; Ionescu, R., Heterogeneity of monoclonal antibodies revealed by
charge-sensitive methods. Curr Pharm Biotechnol 2008, 9, (6), 468-81.
6. Khawli, L. A.; Goswami, S.; Hutchinson, R.; Kwong, Z. W.; Yang, J.; Wang, X.;
Yao, Z.; Sreedhara, A.; Cano, T.; Tesar, D.; Nijem, I.; Allison, D. E.; Wong, P. Y.; Kao,
Y. H.; Quan, C.; Joshi, A.; Harris, R. J.; Motchnik, P., Charge variants in IgG1: Isolation,
characterization, in vitro binding properties and pharmacokinetics in rats. MAbs 2010, 2,
(6), 613-24.
7. Harris, R. J., Processing of C-terminal lysine and arginine residues of proteins
isolated from mammalian cell culture. J Chromatogr A 1995, 705, (1), 129-34.
8. Cai, B.; Pan, H.; Flynn, G. C., C-terminal lysine processing of human
immunoglobulin G2 heavy chain in vivo. Biotechnol Bioeng 2011, 108, (2), 404-12.
9. Dick, L. W., Jr.; Qiu, D.; Mahon, D.; Adamo, M.; Cheng, K. C., C-terminal lysine
variants in fully human monoclonal antibodies: investigation of test methods and possible
causes. Biotechnol Bioeng 2008, 100, (6), 1132-43.
10. Luo, J.; Zhang, J.; Ren, D.; Tsai, W. L.; Li, F.; Amanullah, A.; Hudson, T.,
Probing of C-terminal lysine variation in a recombinant monoclonal antibody production
using Chinese hamster ovary cells with chemically defined media. Biotechnol Bioeng
2012, 109, (9), 2306-15.
11. Hu, Z.; Zhang, H.; Haley, B.; Macchi, F.; Yang, F.; Misaghi, S.; Elich, J.; Yang,
R.; Tang, Y.; Joly, J. C.; Snedecor, B. R.; Shen, A., Carboxypeptidase D is the only
enzyme responsible for antibody C-terminal lysine cleavage in Chinese hamster ovary
(CHO) cells. Biotechnol Bioeng 2016.
12. Perkins, S. N.; Husten, E. J.; Eipper, B. A., The 108-kDA peptidylglycine alpha-
amidating monooxygenase precursor contains two separable enzymatic activities
involved in peptide amidation. Biochem Biophys Res Commun 1990, 171, (3), 926-32.
13. Kaschak, T.; Boyd, D.; Lu, F.; Derfus, G.; Kluck, B.; Nogal, B.; Emery, C.;
Summers, C.; Zheng, K.; Bayer, R.; Amanullah, A.; Yan, B., Characterization of the
basic charge variants of a human IgG1: effect of copper concentration in cell culture
media. MAbs 2011, 3, (6), 577-83.
14. Jefferis, R., Posttranslational Modifications and the Immunogenicity of
Biotherapeutics. J Immunol Res 2016, 2016, 5358272.
15. Jiang, G.; Yu, C.; Yadav, D. B.; Hu, Z.; Amurao, A.; Duenas, E.; Wong, M.;
Iverson, M.; Zheng, K.; Lam, X.; Chen, J.; Vega, R.; Ulufatu, S.; Leddy, C.; Davis, H.;
Shen, A.; Wong, P. Y.; Harris, R.; Wang, Y. J.; Li, D., Evaluation of Heavy-Chain C-
Terminal Deletion on Product Quality and Pharmacokinetics of Monoclonal Antibodies.
J Pharm Sci 2016.
16. Deisenhofer, J., Crystallographic refinement and atomic models of a human Fc
fragment and its complex with fragment B of protein A from Staphylococcus aureus at
2.9- and 2.8-A resolution. Biochemistry 1981, 20, (9), 2361-70.
17. van den Bremer, E. T.; Beurskens, F. J.; Voorhorst, M.; Engelberts, P. J.; de Jong,
R. N.; van der Boom, B. G.; Cook, E. M.; Lindorfer, M. A.; Taylor, R. P.; van Berkel, P.
H.; Parren, P. W., Human IgG is produced in a pro-form that requires clipping of C-
terminal lysines for maximal complement activation. MAbs 2015, 7, (4), 672-80.
18. Hu, Z.; Guo, D.; Yip, S. S.; Zhan, D.; Misaghi, S.; Joly, J. C.; Snedecor, B. R.;
Shen, A. Y., Chinese hamster ovary K1 host cell enables stable cell line development for
antibody molecules which are difficult to express in DUXB11-derived dihydrofolate
reductase deficient host cell. Biotechnol Prog 2013, 29, (4), 980-5.
Figure 1. C-terminal lysine deletion of antibody HC results in lower specific productivity and titer.
Figure 1
508x677mm (72 x 72 DPI)
Figure 2. Individual clones expressing antibody mAb3 bearing HC C-terminal lysine deletion have relatively
lower productivities.
Figure 2
508x677mm (72 x 72 DPI)
Figure 3. Pool of clones expressing antibody mAb3 bearing HC C-terminal lysine deletion has lower
productivity and higher basic variant levels.
Figure 3
508x677mm (72 x 72 DPI)
Figure 4. Pool of clones expressing antibody mAb3 lacking HC C-terminal lysine has similar antibody mRNA
level and trafficking pattern compared to WT antibody and antibody lacking glycine-lysine, but higher
degradation rates.
Figure 4
508x677mm (72 x 72 DPI)
Figure 5. Low productivity of mAb3 bearing HC C-terminal lysine deletion is accompanied by lower
intracellular HC expression/accumulation.
Figure 5
508x677mm (72 x 72 DPI)