- Email: [email protected]
Glycosylation analysis of an aggregated antibody produced by Chinese hamster ovary cells in bioreactor culture
Journal of Bioscience and Bioengineering VOL. xx No. xx, 1e6, 2013 www.elsevier.com/locate/jbiosc
Glycosylation analysis of an aggregated antibody produced by Chinese hamster ovary cells in bioreactor culture Masayoshi Onitsuka,1, * Akira Kawaguchi,2 Ryutaro Asano,3 Izumi Kumagai,3 Kohsuke Honda,2 Hisao Ohtake,2 and Takeshi Omasa1, 2 Institute of Technology and Science, The University of Tokushima, Minamijosanjima-cho 2-1, Tokushima 770-8506, Japan,1 Department of Biotechnology, Graduate School of Engineering, Osaka University, Yamadaoka 2-1, Suita, Osaka 565-0871, Japan,2 and Department of Biomolecular Engineering, Graduate School of Engineering, Tohoku University, 6-6-11-606 Aoba-yama, Aramaki, Aoba-ku, Sendai 980-8579, Japan3 Received 28 September 2013; accepted 1 November 2013 Available online xxx
N-Glycosylation of therapeutic antibodies contributes not only to their biological function, but also to their stability and tendency to aggregate. Here, we investigated the impact of the glycosylation status of an aggregated antibody that accumulated during the bioreactor culture of Chinese hamster ovary cells. High-performance liquid chromatography analysis showed that there was no apparent difference in the glycosylation patterns of monomeric, dimeric, and large aggregated forms of the antibody. In contrast, lectin binding assays, which enable the total amounts of specific sugar residues to be detected, showed that both galactose and fucose residues in dimers and large aggregates were reduced to 70e80% of the amount in monomers. These results strongly suggest that the lack of N-linked oligosaccharides, a result of deglycosylation or aglycosylation, occurred in a proportion of the dimeric and large aggregated components. The present study demonstrates that glycosylation heterogeneities are a potential cause of antibody aggregation in cell culture of Chinese hamster ovary cells, and that the lack of N-glycosylation promotes the formation of dimers and finally results in large aggregates. Ó 2013, The Society for Biotechnology, Japan. All rights reserved. [Key words: Chinese hamster ovary cell; Cell culture; Antibody production; Antibody aggregation; Glycosylation heterogeneity; Bispecific diabody]
Immunoglobulin G (IgG) antibody molecules consist of Fab and Fc regions, and the Fc region contains two N-glycosylation sites in the CH2 domains. N-Glycosylation enables the interaction of Fc receptors and C1q proteins, which is involved in effector functions such as antibody-dependent cellular cytotoxicity and complementdependent cytotoxicity. The N-linked oligosaccharides show marked heterogeneities in their terminal galactose and sialic acid and core fucose residues, which influence the expression of the effector functions (1,2). In addition to functional aspects, N-glycosylation of antibodies can influence their structural stability and aggregation. The trimming of N-linked oligosaccharides, deglycosylation, and aglycosylation reduce antibody structural stability, and make IgG molecules less resistant to aggregation (3e6). Aggregation of therapeutic antibodies can have a serious impact on their safety and quality (7,8). Aggregated antibodies lose their biological activity and exhibit reduced potency. In addition, it has been shown that the aggregation and misfolding of therapeutic proteins can induce a new and cryptic epitope presentation, which results in an unexpected immune response after administration (9,10). Chinese hamster ovary (CHO) cells are important for the industrial production of therapeutic glycoproteins because of its wide usage in GMP-certified recombinant protein production, the development of industrial serum-free media, and its high-level
* Corresponding author. Tel.: þ81 88 656 7408; fax: þ81 88 656 9148. E-mail address: [email protected] (M. Onitsuka).
production of recombinant proteins (11e16). In the manufacturing of therapeutic antibodies, antibody aggregation can occur at different stages, including cell culture fermentation, purification, formulation, and long-term storage. Although much is known regarding antibody aggregation during purification and formulation processes, it is less well understood how antibodies aggregate during cell culture fermentation. Media pH, temperature, oxidation status, and cultivation time may also possibly induce antibody aggregation during the cell culture process (7,8). It is well established that the glycosylation heterogeneity of recombinant glycoproteins is largely caused by environmental cell culture conditions (17e20). Here, we suggest the hypothesis that the heterogeneity of Nlinked oligosaccharides is a possible cause of antibody aggregation during the cell culture process. Various studies have investigated the contribution of N-glycosylation to the stability and aggregation of antibodies, as described above. Conversely, the difference in the N-glycosylation status between intact and aggregated antibodies has not been reported. In this study, we analyzed the glycosylation status of an aggregated antibody accumulated during the cell culture process to assess the relationship between glycosylation heterogeneity and aggregation. We used the humanized antiEGFR anti-CD3 bispecific diabody with an Fc portion, Ex3-scDbFc (Fig. 1A) (21,22), as the model antibody. Ex3-scDb-Fc retargets lymphokine-activated killer cells with a T-cell phenotype against epidermal growth factor receptor-positive cells and shows remarkable antitumor activity in vitro (21,22). We previously cultivated a CHO Top-H cell line expressing the bispecific diabody and analyzed the glycosylation pattern of Ex3-scDb-Fc (23). In
1389-1723/$ e see front matter Ó 2013, The Society for Biotechnology, Japan. All rights reserved. http://dx.doi.org/10.1016/j.jbiosc.2013.11.001
Please cite this article in press as: Onitsuka, M., et al., Glycosylation analysis of an aggregated antibody produced by Chinese hamster ovary cells in bioreactor culture, J. Biosci. Bioeng., (2013), http://dx.doi.org/10.1016/j.jbiosc.2013.11.001
2
ONITSUKA ET AL.
J. BIOSCI. BIOENG.,
A
N Ex3-scDb (Bispecific region)
N CH2
Fc
CH3 C
B
C
RCA I
LCA
FIG. 1. (A) Schematic diagram of Ex3-scDb-Fc (21,22). The black spheres in the CH2 domains represent the N-glycosylation sites focused on in this article. (B) The G2F glycoform of N-linked oligosaccharide. Closed square: N-acetylglucosamine; closed circle: mannose; open triangle: fucose; open circle: galactose. Ricinus communis agglutinin I (RCA I) and Lens culinaris agglutinin (LCA) used in this article preferentially recognize terminal galactose and core fucose residues, respectively.
addition, we used a-2,6 sialyltransferase derived from CHO cells (CHO ST6Gal I) for the a-2,6 sialylation of the bispecific antibody (24). Ex3-scDb-Fc is a promising candidate for a next-generation therapeutic antibody because of its dual functionalities. Despite its promising potency, the biophysical properties of the bispecific diabody remain to be elucidated. Factors affecting its aggregation in the manufacturing process need to be clarified and controlled. In this study, we investigated the glycosylation status of monomeric, dimeric, and large aggregated forms of the bispecific diabody in terms of their glycosylation patterns and their total amounts of sugar residues. MATERIALS AND METHODS Cell culture, antibody production, and purification The CHO Top-H cell line producing the Ex3-scDb-Fc bispecific diabody (23,24) was cultivated in suspension culture using serum-free ExCD medium [a mixture of ExCell 302 (SAFC Bioscience, St. Louis, MO, USA) and IS CHO-CD (Irvine Scientific, Santa Ana, CA, USA) supplemented with 1000 nM MTX and 1 mM G418]. The cells were cultivated in a 1-L glass bioreactor (Biott, Tokyo, Japan) containing 750 mL of the medium for approximately 2 weeks. The temperature was maintained at 37 C during cultivation. The agitation speed was 70 rpm and the headspace of the vessel was aerated with air supplied at a flow rate of 100 mL/min. The pH was controlled at 7.1. The dissolved oxygen (DO) concentration was measured with a DO sensor (InPro 6880; Mettler Toledo, Zurich, Switzerland) and was always kept above 40% air saturation. In operation, 25 mL of feeding solution containing 60 g/L glucose, 60 mM L-glutamine, and Yeastolate Ultrafiltrate (Life Technologies, Carlsbad, CA, USA) was added to the bioreactor every 3 days. The Ex3-scDb-Fc antibody was purified from the culture medium by HiTrap protein A affinity chromatography on an AKTA Prime Plus system (GE Healthcare, Buckingham, UK) with a single-step pH gradient from the equilibration buffer [25 mM Tris, 100 mM NaCl, 1 mM EDTA (pH 7.5)] to the elution buffer (1 M Arg-HCl, pH 4.2). Eluted fractions containing the antibody were dialyzed against the equilibration buffer [25 mM Tris, 100 mM NaCl, 1 mM EDTA (pH 7.5)], and further purified with a Sephacryl S-300 prepacked column (GE Healthcare). The purities of the fractionated monomeric, dimeric, and large aggregated forms of the antibody were checked by SDS-PAGE, and the antibody concentrations were determined using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific Inc., Rockford, IL, USA). A deglycosylated antibody was prepared by glycopeptidase F (GPF) digestion (Takara Bio, Otsu, Japan) of 0.5 mg of a monomeric bispecific antibody with 10 mU GPF in 100 mM TriseHCl buffer (pH 8.6), and the antibody was purified with Protein A Mag Sepharose Xtra (GE Healthcare). Deglycosylation was checked by SDS-PAGE.
Circular dichroism and fluorescence spectrum measurements Far-UV CD spectra were measured using a Jasco J-820 spectropolarimeter (Jasco, Tokyo, Japan) with a quartz cell of 1-mm path length. Antibody concentrations were prepared at 0.4 mg/mL in buffer comprising 25 mM Tris, 100 mM NaCl, and 1 mM EDTA (pH 7.5). The temperatures of the samples were kept at 20 C using a Peltier temperature controller (PYC-347WI; Jasco). Thermal denaturation experiments were performed by monitoring CD signals at 218 nm as the temperature increased. ANS binding fluorescence spectra were measured with a Hitachi F-7000 fluorescence spectrophotometer (Hitachi, Tokyo, Japan). 1-Anilinonaphthalene-8-sulfonate (ANS) was purchased from Sigma. The excitation wavelength was 365 nm, and each measurement sample contained 0.1 mg/mL antibody and 10 mM ANS in 25 mM Tris, 100 mM NaCl, and 1 mM EDTA (pH 7.5). The temperature of the sample was kept at 20 C by using circulating temperature-controlled water. High-performance liquid chromatography analysis of N-linked oligosaccharides N-linked oligosaccharides were released from 0.5 mg of monomeric, dimeric, and large aggregated forms of the antibody by glycopeptidase F digestion (Takara Bio) and purified with a cellulose cartridge Glycan preparation kit (Takara Bio). The reducing ends of the oligosaccharides were derivatized with 2aminopyridine-acetic acid. Pyridylaminated (PA) oligosaccharides were repurified with a Blot Glyco cleanup column (Sumitomo Bakelite, Tokyo, Japan). Prepared PA oligosaccharide was analyzed with an ODS C18 COSMOSIL 5C18-AR-II column (Nacalai Tesque) attached to a Shimadzu LC-20AD high-performance liquid chromatography (HPLC) system (Shimadzu Corporation, Kyoto, Japan). Variants were estimated by comparing their retention times with those of pyridylaminated standards (Masuda Chemical Industries, Takamatsu, Japan) and further identified by matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) mass spectroscopy measurements using an Autoflex mass spectrometer (Bruker Daltonics, Billerica, MA, USA). Lectin binding assay The total amounts of galactosylated and fucosylated oligosaccharides on the antibodies were estimated using an enzyme-linked lectin binding assay (24,25). The wells of a 96-well plate (Corning Inc., NY, USA) were coated with the monomeric, dimeric, and aggregated forms of the bispecific antibody (100 mL/well, 0.5e2.5 mg/mL). To prepare the denatured samples, each sample of bispecific antibody was heated at 60 C with 0.1% SDS and 5 mM DTT for 5 min. After immobilization and BSA blocking, 100 mL of 18 mg/mL biotinylated Ricinus communis agglutinin I (RCA I, RCA120) and biotinylated Lens culinaris agglutinin (LCA) solutions (Funakoshi, Tokyo, Japan) were added to each well. After 12 h incubation at 4 C, streptavidin-conjugated horseradish peroxidase (HRP) was used for detection. Absorbance at 405 nm was measured with an Infinite M200 microplate reader (Tecan, Grödig, Austria).
RESULTS AND DISCUSSION Characterization of the aggregated bispecific diabody that accumulated during the CHO cell culture process The CHO Top-H cell line producing Ex3-scDb-Fc was cultivated in a 1-L glass bioreactor. The cells were inoculated at 3 105 cells/mL from the preculture in the mid-exponential growth phase. The maximum cell density reached 50 105 cells/mL. The culture medium was harvested after 15 days when the cell viability reached 60%. The final concentration of Ex3-scDb-Fc was approximately 40 mg/L (Supplementary Fig. S1). Ex3-scDb-Fc was purified by protein A affinity chromatography. In the affinity purification, we used the 1M Arg-HCl solution (pH 4.2) as an eluent, because arginine solution can avoid the antibody aggregation in the purification process (26). The oligomeric state of the bispecific antibody was examined by size exclusion chromatography (Fig. 2A). The elution profile showed multiple peaks, and the peaks at elution volumes of 71, 58, and 42 mL corresponded to the monomeric, dimeric, and high-order aggregated forms, respectively. The homogeneity of the polypeptides was checked by SDS-PAGE in a reduced condition (Fig. 2B). Single bands at approximately 90 kDa, the size of a single-chained bispecific antibody, indicated that these oligomeric states were derived from the full length antibody. It should be noted that the dimeric and the high-order aggregated (large aggregate) forms were not formed in the purification process but in the cell culture process, because monomeric Ex3-scDb-Fc re-purified on a protein A column showed no tendency to form aggregates and retained the monomeric state (data not shown). The ratio of three oligomeric states relative to total Ex3-scDb-Fc was calculated based on the peak area of the size exclusion chromatography analysis (Fig. 2A).
Please cite this article in press as: Onitsuka, M., et al., Glycosylation analysis of an aggregated antibody produced by Chinese hamster ovary cells in bioreactor culture, J. Biosci. Bioeng., (2013), http://dx.doi.org/10.1016/j.jbiosc.2013.11.001
VOL. xx, 2013
GLYCOSYLATION ANALYSIS OF AGGREGATED ANTIBODY
A
B
300
A280 (mAU)
250
M
D
3
LA
120 -
M
91 -
200
62 -
150
LA
D
100
46 -
50
38 -
0 0
40
60
80
100
120
Retention volume (mL)
D 250
20
Fluorescence intensity (a.u.)
CD intensity (mdeg)
C
20
10 0 -10 -20 200
210
220
230
240
200 150 100 50 0 400
250
500
600
700
Wavelength (nm)
Wavelength (nm)
FIG. 2. Characterization of the oligomeric state of Ex3-scDb-Fc. (A) Size exclusion chromatography showing the elution profile of Ex3-scDb-Fc. The peaks at elution volumes of 71, 58, and 42 mL corresponded to the monomeric (M), dimeric (D), and large aggregated (LA) forms, respectively. (B) Reducing SDS-PAGE of monomeric (M), dimeric (D), and large aggregated (LA) forms of Ex3-scDb-Fc. (C) Circular dichroism spectra of monomeric (thick gray line), dimeric (thin black line), and large aggregated (thick black line) forms of Ex3scDb-Fc. (D) ANS fluorescence spectra of monomeric (thick gray line), dimeric (thin black line), and large aggregated (thick black line) forms of Ex3-scDb-Fc. The dashed line indicates ANS fluorescence without an antibody.
Analysis of the N-linked oligosaccharide structure To investigate the effects of the glycosylation status of Ex3-scDb-Fc on its aggregation, we first analyzed the glycosylation pattern of the monomeric, dimeric, and large aggregated forms of the antibody. The same amount of each form (0.5 mg) was treated with GPF and their oligosaccharides were purified and pyridylaminated. The N-linked oligosaccharide structures of each form were estimated by reverse-phase HPLC analysis on an ODS column. The elution profiles of the purified oligosaccharides are shown in Fig. 3A. Previously, our group reported the glycosylation pattern of Ex3-scDb-Fc (23), in which six peaks corresponding to N-oligosaccharides were detected. HPLC analysis in the present
A Fluorescence intensity (a.u.)
1,200,000
B
900,000
G1F
G0F
600,000
G2F
300,000 0 26
27
28
29
30
31
Retention time (min)
Percentage of glycoform (%)
The relative ratio of the monomeric, dimeric, and large aggregated forms were estimated to be 72%, 10% and 18%, respectively. Furthermore, we examined the solution structure of the monomeric, dimeric and large aggregated antibodies by circular dichroism (CD) and ANS fluorescence spectroscopy (Fig. 2C and D). CD spectra of the monomeric and dimeric forms exhibited the typical secondary structure of the antibody molecule, and the spectra for both states superposed very well. ANS binds to the hydrophobic area of the exposed side chains of the protein and shows large fluorescence intensity. The ANS florescence spectrum in the presence of monomeric Ex3-scDb-Fc was nearly identical to that of free ANS, indicating that the hydrophobic area of the monomer was completely buried in the antibody structure. The fluorescence intensity of the dimer was slightly increased. These data indicate that in the cell culture process, Ex3-scDb-Fc dimerizes without an accompanying structural change to the antibody. By contrast, the solution structure of large aggregates exhibited the formation of non-native b-strand structures and significantly increased the exposure of the hydrophobic regions to the solution (Fig. 2C and D). This series of results indicates that the large aggregated form of Ex3-scDb-Fc is in the misfolded state with a non-native b-strand structure and surface-exposed hydrophobic regions.
60 50 40 30 20 10 0
G0F
G1F
G2F
FIG. 3. (A) Reverse-phase HPLC profile of pyridylaminated derivatives of the N-linked oligosaccharides released from monomeric (thick gray line), dimeric (thin black line), and large aggregated (thick black line) forms of Ex3-scDb-Fc. (B) Relative fractions of the estimated oligosaccharide structures. Open, gray-filled, and black-filled bars represent the percentages in the monomeric, dimeric, and large aggregated forms, respectively. Error bars denote the standard deviation (n ¼ 3).
Please cite this article in press as: Onitsuka, M., et al., Glycosylation analysis of an aggregated antibody produced by Chinese hamster ovary cells in bioreactor culture, J. Biosci. Bioeng., (2013), http://dx.doi.org/10.1016/j.jbiosc.2013.11.001
ONITSUKA ET AL.
A
J. BIOSCI. BIOENG.,
B
1.2
0.8
Abs. at 405nm
Abs. at 405nm
1.0
C
1.0
Relative detection of sugar residue
4
0.8 0.6 0.4
0.6 0.4 0.2
0.2 0.0 0
50 100 150 200 250
[Ex3-scDb-Fc] (ng)
0.0
0
50 100 150 200 250
[Ex3-scDb-Fc] (ng)
1.2 1.0 0.8 0.6 0.4 0.2 0.0
Galactose (RCA I)
Fucose (LCA)
FIG. 4. Lectin binding assays (LBA) for the detection of specific sugar residues in Ex3-scDb-Fc. (A) RCA I assay for the detection of galactose residues. (B) LCA assay for the detection of fucose residues. (A, B) Open, gray-filled, and black-filled circles correspond to monomer, dimer, and large aggregates, respectively. Continuous lines represent the fits of the nonlinear regression analysis. (C) Relative detection of N-linked oligosaccharide residues estimated from the slope of the regression analysis. Open, gray-filled, and black-filled bars represent the detection of monomeric, dimeric, and large aggregated forms, respectively. Error indicates the standard error of regression analysis.
study corresponded well with the data from the previous study. Among the six peaks, three peaks with retention times of 27.5e30 min were dominant (Fig. 3A) relative to the other three peaks, and constituted approximately 90% of the total peak area. The retention time of the commercial oligosaccharide standards and MALDI-TOF mass spectroscopy analysis of the fractionated peaks ensured that the three peaks corresponded to the agalactosyl, mono-, and bi-galactosyl oligosaccharides with fucose (G0F, G1F, and G2F, respectively). We focused on the G0F, G1F, and G2F oligosaccharides in this study. All estimated Noligosaccharides, including minor components, are summarized in Supplementary data (Fig. S2). The molar ratio of these three oligosaccharides relative to total oligosaccharides was calculated based on the peak area of the HPLC profile. The observed glycosylation patterns are summarized in Fig. 3B. The overall distributions of all forms were similardthe profile exhibited a low level of fully galactosylated oligosaccharide with fucose (G2F), as reported in our previous study (23). The percentage of G2F in the large aggregates was slightly higher for the monomeric and dimeric forms. Both agalactosyl and monogalactosyl glycoforms accounted for over 30% of the total N-linked oligosaccharides. These two main components showed a similar molecular ratio among the three states, i.e., the monomer, dimer, and large aggregates. Overall, these results indicate that there is no apparent difference in the relative molar proportions of the three forms of Ex3-scDb-Fc antibody. However, a striking difference was shown in the absolute peak area in the HPLC profile (Fig. 3A). The profile of the dimeric and large aggregated forms exhibited lower peak heights as compared with monomeric Ex3-scDb-Fc, suggesting that the total amounts of N-oligosaccharide rather than the glycosylation pattern had dominant effects on aggregate formation. We next performed lectin binding assays to quantitatively estimate the total amounts of saccharide residues. The lectin binding assay enabled us to directly observe the degree of specific sugar residues in protein samples using enzyme-linked lectins (24,25). We used biotinylated R. communis agglutinin I (RCA I) and L. culinaris agglutinin (LCA), which are specific for terminal galactose residues and core fucose residues, respectively (Fig. 1B). The assay plate wells were directly coated with Ex3-scDb-Fc solution and the sugar residues were detected by an HRP-catalyzed enzymatic reaction. A schematic diagram of the lectin binding assay and example results are shown in Figs. S3 and S4. In the RCA I and LCA assays, the absorbance at 405 nm of the intact monomer increased proportionally up to 250 ng, whereas that of the deglycosylated form prepared by GPF treatment was almost identical throughout the range (Fig. S4). These data indicated that the assay system
works correctly and that the Ex3-scDb-Fc used in this study mainly contains N-linked oligosaccharides. We evaluated the relative levels of galactosylation and fucosylation from the slope following a non-linear regression analysis (Fig. 4A and B). The highest slope value was normalized to 1 (Fig. 4C). The assay was performed using denatured Ex3-scDb-Fc samples, which enabled us to detect the total amount of sugar residues because N-linked oligosaccharides, which are located between two CH2 domains, are surface exposed by complete unfolding. Both the RCA I and LCA assays showed that the relative levels of galactosylation and fucosylation in the dimeric and large aggregated forms were equally reduced to 70e80% of the level in monomeric Ex3-scDb-Fc (Fig. 4C). Again, there were no apparent differences in the proportions of the main glycoforms, G0F, G1F, and G2F, among the monomeric, dimeric, and large aggregated forms of the antibody (Fig. 3B). Therefore, we interpret the reduction in both galactose and fucose residues as indicating that the lack of N-linked oligosaccharides, that is deglycosylation or aglycosylation, occurred in a proportion of the dimeric and large aggregated forms. In addition to denaturing conditions, we performed the assay in non-denaturing conditions (Fig. S5), and the large aggregates showed the highest levels of galactose and fucose residues. Lectin binding assays in non-denaturing conditions were highly dependent on the antibody conformation, because the Nlinked oligosaccharides are located between the two CH2 domains of the Fc region (27). The highest levels of galactose and fucose residues detected demonstrated that the N-glycosylated region in large aggregates was surface exposed owing to its misfolded state. N-Glycosylation analysis revealed that the most striking difference in glycosylation status among the different forms of Ex3-scDb-Fc was the total amount of N-glycosylation rather than the glycosylation pattern. Effects of glycosylation on the structure and stability of Ex3scDb-Fc We examined the effects of N-linked glycosylation on the stability of, and aggregate formation by, the Ex3-scDb-Fc antibody. CD spectra of the monomeric form, including the deglycosylated monomer, showed that the lack of N-linked oligosaccharides had no effect on the overall secondary structure of the antibody (Fig. 5A). In contrast, thermal unfolding experiments revealed the contribution of N-glycosylation to the stability of Ex3-scDb-Fc (Fig. 5B). At temperatures above 45 C, the CD signal intensity at 218 nm showed two characteristic changesddecreased intensity up to about 60 C and then increased intensity above 60 C. The former and the latter signal changes are interpreted as structural unfolding and aggregation/ precipitation, respectively. In the deglycosylated form, the temperature of the aggregation/precipitation process (increased
Please cite this article in press as: Onitsuka, M., et al., Glycosylation analysis of an aggregated antibody produced by Chinese hamster ovary cells in bioreactor culture, J. Biosci. Bioeng., (2013), http://dx.doi.org/10.1016/j.jbiosc.2013.11.001
VOL. xx, 2013
CD intensity (mdeg)
A
GLYCOSYLATION ANALYSIS OF AGGREGATED ANTIBODY
10 5 0 -5 -10 200
Relative CD intensity
B
210
220
230
240
250
Wavelength (nm)
1 0 -1 -2 -3 30
40
50
60
70
80
Temperature (degree) FIG. 5. (A) Circular dichroism spectra of intact monomeric (thick gray line) and deglycosylated forms by GPF treatment (thin black line). (B) Thermal unfolding monitored by CD intensity at 218 nm. The thick gray line and the thin black line represent the curves for intact monomeric and deglycosylated forms, respectively.
CD intensity) was decreased by about 2 C, indicating that the resistance to aggregation of Ex3-scDb-Fc was actually reduced by deglycosylation. Molecular insights into how deglycosylation induces antibody aggregation in CHO cell culture The N-linked oligosaccharides on IgG interact with the polypeptide in the CH2 domain, and the carbohydrateeprotein interaction plays a crucial role in structural conformation and integrity for FcgRs binding (28,29). In addition to the functional aspect, the carbohydrateeprotein interaction contributes to the structural stability of IgG antibodies (3,4), and recent studies have shown that deglycosylation promotes antibody aggregation (5,6). However, the impact of glycosylation status on aggregation during the manufacturing of therapeutic antibodies has not been assessed. Here, we showed that a proportion (approximately 20e30%) of dimeric and large aggregated Ex3-scDb-Fc molecules was deglycosylated or aglycosylated, indicating the importance of glycosylation heterogeneity as a possible reason for antibody aggregation during CHO cell culture where the culture conditions influence the glycosylation heterogeneity. These aggregations impair antigen binding. Antibody dimer partially loses the antigen binding affinity (30), and large aggregates probably fully lost their binding activity due to misfolding (Fig. 2C). Deglycosylation and aglycosylation would have negative impacts on antibody structure and antigen binding. Protein aggregation is well represented by the nucleationdependent aggregation model, where small oligomers are initially organized as the aggregation nucleus and then subsequently polymerized into large aggregates (31,32). A recent review highlighted the role of monomeric and oligomeric forms with a nativelike structure as the nucleus for misfolded large aggregates under
5
non-denaturing conditions (33). Characterization of the aggregation nucleus with the native-like structure would provide clues to understanding protein aggregation. In the present study, we observed the dimeric form of Ex3-scDb-Fc with a native-like structure and reduced N-glycosylation. Molecular dynamics simulation predicts the aggregation prone motif in IgG antibodies, where the hydrophobic residues covered by N-linked oligosaccharides in the CH2 domain become the aggregation prone motif by deglycosylation (5,6). In addition, IgG antibodies have tendency to self-associate by a weak interaction among antibody molecules at Fc regions (34). A series of facts provide molecular insights into the mechanism of antibody aggregation in recombinant CHO cell culture. The exposure of hydrophobic residues by the lack of Nglycosylation converts monomeric Ex3-scDb-Fc into a dimer by hydrophobic stacking, and the dimer serves as an aggregation nucleus and promotes polymerization into large aggregates. This hypothetical model implies that deglycosylation is a possible reason for the formation of the aggregation nucleus, although the direct relationship between the dimer and the aggregation nucleus remains to be clarified. It is well established that N-glycosylation plays an important role in the folding and secretion of glycoproteins in the endoplasmic reticulum and Golgi apparatus (19,35,36). It remains unclear why a lack of N-linked oligosaccharides was present in the Ex3-scDb-Fc antibody during CHO cell culture. Endogenous Nglycosidases may leak from the CHO cells into the culture medium following cell death, and may deglycosylate secreted antibodies. Aglycosylated antibody molecules may be secreted following the depletion of intracellular nucleotide sugars as glycosylation precursors in the death phase of cell culture. To test these possibilities, the time-dependent N-glycosylation status of the antibody produced in CHO cell culture should be investigated in future work. In addition, supplementation of the inhibitor of N-glycosidases (37) and glycosylation precursors (38) would prevent the deglycosylation and aglycosylation, which should be examined in future work. Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jbiosc.2013.11.001. ACKNOWLEDGMENTS This study was supported by the Advanced Research for Medical Products Mining Programme of the National Institute of Biomedical Innovation (NIBIO). References 1. Arnold, J. N., Wormald, M. R., Sim, R. B., Rudd, P. M., and Dwek, R. A.: The impact of glycosylation on the biological function and structure of human immunoglobulins, Annu. Rev. Immunol., 25, 21e50 (2007). 2. Jefferis, R.: Glycosylation as a strategy to improve antibody-based therapeutics, Nat. Rev. Drug Discov., 8, 226e234 (2009). 3. Ghirlando, R., Lund, J., Goodall, M., and Jefferis, R.: Glycosylation of human IgG-Fc: influences on structure revealed by differential scanning micro-calorimetry, Immunol. Lett., 68, 47e52 (1999). 4. Mimura, Y., Church, S., Ghirlando, R., Ashton, P. R., Dong, S., Goodall, M., Lund, J., and Jefferis, R.: The influence of glycosylation on the thermal stability and effector function expression of human IgG1-Fc: properties of a series of truncated glycoforms, Mol. Immunol., 37, 697e706 (2000). 5. Voynov, V., Chennamsetty, N., Kayser, V., Helk, B., Forrer, K., Zhang, H., Fritsch, C., Heine, H., and Trout, B. L.: Dynamic fluctuations of protein-carbohydrate interactions promote protein aggregation, PLoS One, 4, e8425 (2009). 6. Kayser, V., Chennamsetty, N., Voynov, V., Forrer, K., Helk, B., and Trout, B. L.: Glycosylation influences on the aggregation propensity of therapeutic monoclonal antibodies, Biotechnol. J., 6, 38e44 (2011). 7. Cromwell, M. E., Hilario, E., and Jacobson, F.: Protein aggregation and bioprocessing, AAPS. J., 8, E572eE579 (2006). 8. Vázquez-Rey, M. and Lang, D. A.: Aggregates in monoclonal antibody manufacturing processes, Biotechnol. Bioeng., 108, 1494e1508 (2011).
Please cite this article in press as: Onitsuka, M., et al., Glycosylation analysis of an aggregated antibody produced by Chinese hamster ovary cells in bioreactor culture, J. Biosci. Bioeng., (2013), http://dx.doi.org/10.1016/j.jbiosc.2013.11.001
6
ONITSUKA ET AL.
9. Rosenberg, A. S.: Effects of protein aggregates: an immunologic perspective, AAPS J., 8, E501eE507 (2006). 10. Maas, C., Hermeling, S., Bouma, B., Jiskoot, W., and Gebbink, M. F.: A role for protein misfolding in immunogenicity of biopharmaceuticals, J. Biol. Chem., 282, 2229e2236 (2007). 11. Gatti, M., Wlaschin, K. F., Nissom, P. M., Yap, M., and Hu, W. S.: Comparative transcriptional analysis of mouse hybridoma and recombinant Chinese hamster ovary cells undergoing butyrate treatment, J. Biosci. Bioeng., 103, 82e91 (2007). 12. Liu, C. H. and Chen, L. H.: Promotion of recombinant macrophage colony stimulating factor production by dimethyl sulfoxide addition in Chinese hamster ovary cells, J. Biosci. Bioeng., 103, 45e49 (2007). 13. Fujiwara, M., Koizumi, S., and Takagi, M.: Effect of static pressure on intracellular pH of adhesive Chinese hamster ovary cells, J. Biosci. Bioeng., 104, 510e512 (2007). 14. Omasa, T., Takami, T., Ohya, T., Kiyama, E., Hayashi, T., Nishii, H., Miki, H., Kobayashi, K., Honda, K., and Ohtake, H.: Overexpression of GADD34 enhances production of recombinant human antithrombin III in Chinese hamster ovary cells, J. Biosci. Bioeng., 106, 568e573 (2008). 15. Kaneko, Y., Sato, R., and Aoyagi, H.: Changes in the quality of antibodies produced by Chinese hamster ovary cells during the death phase of cell culture, J. Biosci. Bioeng., 109, 281e287 (2010). 16. Omasa, T., Onitsuka, M., and Kim, W. D.: Cell engineering and cultivation of Chinese hamster ovary (CHO) cells, Curr. Pharm. Biotechnol., 11, 233e240 (2010). 17. Andersen, D. C. and Goochee, C. F.: The effect of cell-culture conditions on the oligosaccharide structures of secreted glycoproteins, Curr. Opin. Biotechnol., 5, 546e549 (1994). 18. Yang, M. and Butler, M.: Effects of ammonia and glucosamine on the heterogeneity of erythropoietin glycoforms, Biotechnol. Prog., 18, 129e138 (2002). 19. Hossler, P., Khattak, S. F., and Li, Z. J.: Optimal and consistent protein glycosylation in mammalian cell culture, Glycobiology, 19, 936e949 (2009). 20. Jefferis, R.: Recombinant antibody therapeutics: the impact of glycosylation on mechanisms of action, Trends Pharmacol. Sci., 30, 356e362 (2009). 21. Asano, R., Kawaguchi, H., Watanabe, Y., Nakanishi, T., Umetsu, M., Hayashi, H., Katayose, Y., Unno, M., Kudo, T., and Kumagai, I.: Diabody-based recombinant formats of humanized IgG-like bispecific antibody with effective retargeting of lymphocytes to tumor cells, J. Immunother., 31, 752e761 (2008). 22. Asano, R., Ikoma, K., Kawaguchi, H., Ishiyama, Y., Nakanishi, T., Umetsu, M., Hayashi, H., Katayose, Y., Unno, M., Kudo, T., and Kumagai, I.: Application of the Fc fusion format to generate tag-free bi-specific diabodies, FEBS J., 277, 477e487 (2010). 23. Kim, W. D., Tokunaga, M., Ozaki, H., Ishibashi, T., Honda, K., Kajiura, H., Fujiyama, K., Asano, R., Kumagai, I., Omasa, T., and Ohtake, H.: Glycosylation pattern of humanized IgG-like bispecific antibody produced by recombinant CHO cells, Appl. Microbiol. Biotechnol., 85, 535e542 (2010). 24. Onitsuka, M., Kim, W. D., Ozaki, H., Kawaguchi, A., Honda, K., Kajiura, H., Fujiyama, K., Asano, R., Kumagai, I., Ohtake, H., and Omasa, T.: Enhancement of sialylation on humanized IgG-like bispecific antibody by overexpression of
J. BIOSCI. BIOENG., a2,6-sialyltransferase derived from Chinese hamster ovary cells, Appl. Microbiol. Biotechnol., 94, 69e80 (2012). 25. Omasa, T., Tanaka, R., Doi, T., Ando, M., Kitamoto, Y., Honda, K., Kishimoto, M., and Ohtake, H.: Decrease in antithrombin III fucosylation by expressing GDP-fucose transporter siRNA in Chinese hamster ovary cells, J. Biosci. Bioeng., 106, 168e173 (2008). 26. Arakawa, T., Philo, J. S., Tsumoto, K., Yumioka, R., and Ejima, D.: Elution of antibodies from a Protein-A column by aqueous arginine solutions, Protein Expr. Purif., 36, 244e248 (2004). 27. Sondermann, P., Huber, R., Oosthuizen, V., and Jacob, U.: The 3.2-Å crystal structure of the human IgG1 Fc fragment-Fc gRIII complex, Nature, 406, 267e273 (2000). 28. Krapp, S., Mimura, Y., Jefferis, R., Huber, R., and Sondermann, P.: Structural analysis of human IgG-Fc glycoforms reveals a correlation between glycosylation and structural integrity, J. Mol. Biol., 325, 979e989 (2003). 29. Yamaguchi, Y., Nishimura, M., Nagano, M., Yagi, H., Sasakawa, H., Uchida, K., Shitara, K., and Kato, K.: Glycoform-dependent conformational alteration of the Fc region of human immunoglobulin G1 as revealed by NMR spectroscopy, Biochim. Biophys. Acta, 1760, 693e700 (2006). 30. Paul, R., Graff-Meyer, A., Stahlberg, H., Lauer, M. E., Rufer, A. C., Beck, H., Briguet, A., Schnaible, V., Buckel, T., and Boeckle, S.: Structure and function of purified monoclonal antibody dimers induced by different stress conditions, Pharm. Res., 29, 2047e2059 (2012). 31. Harper, J. D. and Lansbury, P. T., Jr.: Models of amyloid seeding in Alzheimer’s disease and scrapie: mechanistic truths and physiological consequences of the time-dependent solubility of amyloid proteins, Annu. Rev. Biochem., 66, 385e407 (1997). 32. Nilsson, M. R.: Techniques to study amyloid fibril formation in vitro, Methods, 34, 151e160 (2004). 33. Chiti, F. and Dobson, C. M.: Amyloid formation by globular proteins under native conditions, Nat. Chem. Biol., 5, 15e22 (2009). 34. Nishi, H., Miyajima, M., Wakiyama, N., Kubota, K., Hasegawa, J., Uchiyama, S., and Fukui, K.: Fc domain mediated self-association of an IgG1 monoclonal antibody under a low ionic strength condition, J. Biosci. Bioeng., 112, 326e332 (2011). 35. Fujiyama, K., Misaki, R., Sakai, Y., Omasa, T., and Seki, T.: Change in glycosylation pattern with extension of endoplasmic reticulum retention signal sequence of mouse antibody produced by suspension-cultured tobacco BY2 cells, J. Biosci. Bioeng., 107, 165e172 (2009). 36. Misaki, R., Sakai, Y., Omasa, T., Fujiyama, K., and Seki, T.: N-terminal vacuolar sorting signal at the mouse antibody alters the N-linked glycosylation pattern in suspension-cultured tobacco BY2 cells, J. Biosci. Bioeng., 112, 476e484 (2011). 37. Misaghi, S., Pacold, M. E., Blom, D., Ploegh, H. L., and Korbel, G. A.: Using a small molecule inhibitor of peptide: N-glycanase to probe its role in glycoprotein turnover, Chem. Biol., 11, 1677e1687 (2004). 38. Wong, N. S., Wati, L., Nissom, P. M., Feng, H. T., Lee, M. M., and Yap, M. G.: An investigation of intracellular glycosylation activities in CHO cells: effects of nucleotide sugar precursor feeding, Biotechnol. Bioeng., 107, 321e336 (2010).
Please cite this article in press as: Onitsuka, M., et al., Glycosylation analysis of an aggregated antibody produced by Chinese hamster ovary cells in bioreactor culture, J. Biosci. Bioeng., (2013), http://dx.doi.org/10.1016/j.jbiosc.2013.11.001
Glycosylation analysis of an aggregated antibody produced by Chinese hamster ovary cells in bioreactor culture Masayoshi Onitsuka,1, * Akira Kawaguchi,2 Ryutaro Asano,3 Izumi Kumagai,3 Kohsuke Honda,2 Hisao Ohtake,2 and Takeshi Omasa1, 2 Institute of Technology and Science, The University of Tokushima, Minamijosanjima-cho 2-1, Tokushima 770-8506, Japan,1 Department of Biotechnology, Graduate School of Engineering, Osaka University, Yamadaoka 2-1, Suita, Osaka 565-0871, Japan,2 and Department of Biomolecular Engineering, Graduate School of Engineering, Tohoku University, 6-6-11-606 Aoba-yama, Aramaki, Aoba-ku, Sendai 980-8579, Japan3 Received 28 September 2013; accepted 1 November 2013 Available online xxx
N-Glycosylation of therapeutic antibodies contributes not only to their biological function, but also to their stability and tendency to aggregate. Here, we investigated the impact of the glycosylation status of an aggregated antibody that accumulated during the bioreactor culture of Chinese hamster ovary cells. High-performance liquid chromatography analysis showed that there was no apparent difference in the glycosylation patterns of monomeric, dimeric, and large aggregated forms of the antibody. In contrast, lectin binding assays, which enable the total amounts of specific sugar residues to be detected, showed that both galactose and fucose residues in dimers and large aggregates were reduced to 70e80% of the amount in monomers. These results strongly suggest that the lack of N-linked oligosaccharides, a result of deglycosylation or aglycosylation, occurred in a proportion of the dimeric and large aggregated components. The present study demonstrates that glycosylation heterogeneities are a potential cause of antibody aggregation in cell culture of Chinese hamster ovary cells, and that the lack of N-glycosylation promotes the formation of dimers and finally results in large aggregates. Ó 2013, The Society for Biotechnology, Japan. All rights reserved. [Key words: Chinese hamster ovary cell; Cell culture; Antibody production; Antibody aggregation; Glycosylation heterogeneity; Bispecific diabody]
Immunoglobulin G (IgG) antibody molecules consist of Fab and Fc regions, and the Fc region contains two N-glycosylation sites in the CH2 domains. N-Glycosylation enables the interaction of Fc receptors and C1q proteins, which is involved in effector functions such as antibody-dependent cellular cytotoxicity and complementdependent cytotoxicity. The N-linked oligosaccharides show marked heterogeneities in their terminal galactose and sialic acid and core fucose residues, which influence the expression of the effector functions (1,2). In addition to functional aspects, N-glycosylation of antibodies can influence their structural stability and aggregation. The trimming of N-linked oligosaccharides, deglycosylation, and aglycosylation reduce antibody structural stability, and make IgG molecules less resistant to aggregation (3e6). Aggregation of therapeutic antibodies can have a serious impact on their safety and quality (7,8). Aggregated antibodies lose their biological activity and exhibit reduced potency. In addition, it has been shown that the aggregation and misfolding of therapeutic proteins can induce a new and cryptic epitope presentation, which results in an unexpected immune response after administration (9,10). Chinese hamster ovary (CHO) cells are important for the industrial production of therapeutic glycoproteins because of its wide usage in GMP-certified recombinant protein production, the development of industrial serum-free media, and its high-level
* Corresponding author. Tel.: þ81 88 656 7408; fax: þ81 88 656 9148. E-mail address: [email protected] (M. Onitsuka).
production of recombinant proteins (11e16). In the manufacturing of therapeutic antibodies, antibody aggregation can occur at different stages, including cell culture fermentation, purification, formulation, and long-term storage. Although much is known regarding antibody aggregation during purification and formulation processes, it is less well understood how antibodies aggregate during cell culture fermentation. Media pH, temperature, oxidation status, and cultivation time may also possibly induce antibody aggregation during the cell culture process (7,8). It is well established that the glycosylation heterogeneity of recombinant glycoproteins is largely caused by environmental cell culture conditions (17e20). Here, we suggest the hypothesis that the heterogeneity of Nlinked oligosaccharides is a possible cause of antibody aggregation during the cell culture process. Various studies have investigated the contribution of N-glycosylation to the stability and aggregation of antibodies, as described above. Conversely, the difference in the N-glycosylation status between intact and aggregated antibodies has not been reported. In this study, we analyzed the glycosylation status of an aggregated antibody accumulated during the cell culture process to assess the relationship between glycosylation heterogeneity and aggregation. We used the humanized antiEGFR anti-CD3 bispecific diabody with an Fc portion, Ex3-scDbFc (Fig. 1A) (21,22), as the model antibody. Ex3-scDb-Fc retargets lymphokine-activated killer cells with a T-cell phenotype against epidermal growth factor receptor-positive cells and shows remarkable antitumor activity in vitro (21,22). We previously cultivated a CHO Top-H cell line expressing the bispecific diabody and analyzed the glycosylation pattern of Ex3-scDb-Fc (23). In
1389-1723/$ e see front matter Ó 2013, The Society for Biotechnology, Japan. All rights reserved. http://dx.doi.org/10.1016/j.jbiosc.2013.11.001
Please cite this article in press as: Onitsuka, M., et al., Glycosylation analysis of an aggregated antibody produced by Chinese hamster ovary cells in bioreactor culture, J. Biosci. Bioeng., (2013), http://dx.doi.org/10.1016/j.jbiosc.2013.11.001
2
ONITSUKA ET AL.
J. BIOSCI. BIOENG.,
A
N Ex3-scDb (Bispecific region)
N CH2
Fc
CH3 C
B
C
RCA I
LCA
FIG. 1. (A) Schematic diagram of Ex3-scDb-Fc (21,22). The black spheres in the CH2 domains represent the N-glycosylation sites focused on in this article. (B) The G2F glycoform of N-linked oligosaccharide. Closed square: N-acetylglucosamine; closed circle: mannose; open triangle: fucose; open circle: galactose. Ricinus communis agglutinin I (RCA I) and Lens culinaris agglutinin (LCA) used in this article preferentially recognize terminal galactose and core fucose residues, respectively.
addition, we used a-2,6 sialyltransferase derived from CHO cells (CHO ST6Gal I) for the a-2,6 sialylation of the bispecific antibody (24). Ex3-scDb-Fc is a promising candidate for a next-generation therapeutic antibody because of its dual functionalities. Despite its promising potency, the biophysical properties of the bispecific diabody remain to be elucidated. Factors affecting its aggregation in the manufacturing process need to be clarified and controlled. In this study, we investigated the glycosylation status of monomeric, dimeric, and large aggregated forms of the bispecific diabody in terms of their glycosylation patterns and their total amounts of sugar residues. MATERIALS AND METHODS Cell culture, antibody production, and purification The CHO Top-H cell line producing the Ex3-scDb-Fc bispecific diabody (23,24) was cultivated in suspension culture using serum-free ExCD medium [a mixture of ExCell 302 (SAFC Bioscience, St. Louis, MO, USA) and IS CHO-CD (Irvine Scientific, Santa Ana, CA, USA) supplemented with 1000 nM MTX and 1 mM G418]. The cells were cultivated in a 1-L glass bioreactor (Biott, Tokyo, Japan) containing 750 mL of the medium for approximately 2 weeks. The temperature was maintained at 37 C during cultivation. The agitation speed was 70 rpm and the headspace of the vessel was aerated with air supplied at a flow rate of 100 mL/min. The pH was controlled at 7.1. The dissolved oxygen (DO) concentration was measured with a DO sensor (InPro 6880; Mettler Toledo, Zurich, Switzerland) and was always kept above 40% air saturation. In operation, 25 mL of feeding solution containing 60 g/L glucose, 60 mM L-glutamine, and Yeastolate Ultrafiltrate (Life Technologies, Carlsbad, CA, USA) was added to the bioreactor every 3 days. The Ex3-scDb-Fc antibody was purified from the culture medium by HiTrap protein A affinity chromatography on an AKTA Prime Plus system (GE Healthcare, Buckingham, UK) with a single-step pH gradient from the equilibration buffer [25 mM Tris, 100 mM NaCl, 1 mM EDTA (pH 7.5)] to the elution buffer (1 M Arg-HCl, pH 4.2). Eluted fractions containing the antibody were dialyzed against the equilibration buffer [25 mM Tris, 100 mM NaCl, 1 mM EDTA (pH 7.5)], and further purified with a Sephacryl S-300 prepacked column (GE Healthcare). The purities of the fractionated monomeric, dimeric, and large aggregated forms of the antibody were checked by SDS-PAGE, and the antibody concentrations were determined using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific Inc., Rockford, IL, USA). A deglycosylated antibody was prepared by glycopeptidase F (GPF) digestion (Takara Bio, Otsu, Japan) of 0.5 mg of a monomeric bispecific antibody with 10 mU GPF in 100 mM TriseHCl buffer (pH 8.6), and the antibody was purified with Protein A Mag Sepharose Xtra (GE Healthcare). Deglycosylation was checked by SDS-PAGE.
Circular dichroism and fluorescence spectrum measurements Far-UV CD spectra were measured using a Jasco J-820 spectropolarimeter (Jasco, Tokyo, Japan) with a quartz cell of 1-mm path length. Antibody concentrations were prepared at 0.4 mg/mL in buffer comprising 25 mM Tris, 100 mM NaCl, and 1 mM EDTA (pH 7.5). The temperatures of the samples were kept at 20 C using a Peltier temperature controller (PYC-347WI; Jasco). Thermal denaturation experiments were performed by monitoring CD signals at 218 nm as the temperature increased. ANS binding fluorescence spectra were measured with a Hitachi F-7000 fluorescence spectrophotometer (Hitachi, Tokyo, Japan). 1-Anilinonaphthalene-8-sulfonate (ANS) was purchased from Sigma. The excitation wavelength was 365 nm, and each measurement sample contained 0.1 mg/mL antibody and 10 mM ANS in 25 mM Tris, 100 mM NaCl, and 1 mM EDTA (pH 7.5). The temperature of the sample was kept at 20 C by using circulating temperature-controlled water. High-performance liquid chromatography analysis of N-linked oligosaccharides N-linked oligosaccharides were released from 0.5 mg of monomeric, dimeric, and large aggregated forms of the antibody by glycopeptidase F digestion (Takara Bio) and purified with a cellulose cartridge Glycan preparation kit (Takara Bio). The reducing ends of the oligosaccharides were derivatized with 2aminopyridine-acetic acid. Pyridylaminated (PA) oligosaccharides were repurified with a Blot Glyco cleanup column (Sumitomo Bakelite, Tokyo, Japan). Prepared PA oligosaccharide was analyzed with an ODS C18 COSMOSIL 5C18-AR-II column (Nacalai Tesque) attached to a Shimadzu LC-20AD high-performance liquid chromatography (HPLC) system (Shimadzu Corporation, Kyoto, Japan). Variants were estimated by comparing their retention times with those of pyridylaminated standards (Masuda Chemical Industries, Takamatsu, Japan) and further identified by matrix-assisted laser desorption/ionization-time of flight (MALDI-TOF) mass spectroscopy measurements using an Autoflex mass spectrometer (Bruker Daltonics, Billerica, MA, USA). Lectin binding assay The total amounts of galactosylated and fucosylated oligosaccharides on the antibodies were estimated using an enzyme-linked lectin binding assay (24,25). The wells of a 96-well plate (Corning Inc., NY, USA) were coated with the monomeric, dimeric, and aggregated forms of the bispecific antibody (100 mL/well, 0.5e2.5 mg/mL). To prepare the denatured samples, each sample of bispecific antibody was heated at 60 C with 0.1% SDS and 5 mM DTT for 5 min. After immobilization and BSA blocking, 100 mL of 18 mg/mL biotinylated Ricinus communis agglutinin I (RCA I, RCA120) and biotinylated Lens culinaris agglutinin (LCA) solutions (Funakoshi, Tokyo, Japan) were added to each well. After 12 h incubation at 4 C, streptavidin-conjugated horseradish peroxidase (HRP) was used for detection. Absorbance at 405 nm was measured with an Infinite M200 microplate reader (Tecan, Grödig, Austria).
RESULTS AND DISCUSSION Characterization of the aggregated bispecific diabody that accumulated during the CHO cell culture process The CHO Top-H cell line producing Ex3-scDb-Fc was cultivated in a 1-L glass bioreactor. The cells were inoculated at 3 105 cells/mL from the preculture in the mid-exponential growth phase. The maximum cell density reached 50 105 cells/mL. The culture medium was harvested after 15 days when the cell viability reached 60%. The final concentration of Ex3-scDb-Fc was approximately 40 mg/L (Supplementary Fig. S1). Ex3-scDb-Fc was purified by protein A affinity chromatography. In the affinity purification, we used the 1M Arg-HCl solution (pH 4.2) as an eluent, because arginine solution can avoid the antibody aggregation in the purification process (26). The oligomeric state of the bispecific antibody was examined by size exclusion chromatography (Fig. 2A). The elution profile showed multiple peaks, and the peaks at elution volumes of 71, 58, and 42 mL corresponded to the monomeric, dimeric, and high-order aggregated forms, respectively. The homogeneity of the polypeptides was checked by SDS-PAGE in a reduced condition (Fig. 2B). Single bands at approximately 90 kDa, the size of a single-chained bispecific antibody, indicated that these oligomeric states were derived from the full length antibody. It should be noted that the dimeric and the high-order aggregated (large aggregate) forms were not formed in the purification process but in the cell culture process, because monomeric Ex3-scDb-Fc re-purified on a protein A column showed no tendency to form aggregates and retained the monomeric state (data not shown). The ratio of three oligomeric states relative to total Ex3-scDb-Fc was calculated based on the peak area of the size exclusion chromatography analysis (Fig. 2A).
Please cite this article in press as: Onitsuka, M., et al., Glycosylation analysis of an aggregated antibody produced by Chinese hamster ovary cells in bioreactor culture, J. Biosci. Bioeng., (2013), http://dx.doi.org/10.1016/j.jbiosc.2013.11.001
VOL. xx, 2013
GLYCOSYLATION ANALYSIS OF AGGREGATED ANTIBODY
A
B
300
A280 (mAU)
250
M
D
3
LA
120 -
M
91 -
200
62 -
150
LA
D
100
46 -
50
38 -
0 0
40
60
80
100
120
Retention volume (mL)
D 250
20
Fluorescence intensity (a.u.)
CD intensity (mdeg)
C
20
10 0 -10 -20 200
210
220
230
240
200 150 100 50 0 400
250
500
600
700
Wavelength (nm)
Wavelength (nm)
FIG. 2. Characterization of the oligomeric state of Ex3-scDb-Fc. (A) Size exclusion chromatography showing the elution profile of Ex3-scDb-Fc. The peaks at elution volumes of 71, 58, and 42 mL corresponded to the monomeric (M), dimeric (D), and large aggregated (LA) forms, respectively. (B) Reducing SDS-PAGE of monomeric (M), dimeric (D), and large aggregated (LA) forms of Ex3-scDb-Fc. (C) Circular dichroism spectra of monomeric (thick gray line), dimeric (thin black line), and large aggregated (thick black line) forms of Ex3scDb-Fc. (D) ANS fluorescence spectra of monomeric (thick gray line), dimeric (thin black line), and large aggregated (thick black line) forms of Ex3-scDb-Fc. The dashed line indicates ANS fluorescence without an antibody.
Analysis of the N-linked oligosaccharide structure To investigate the effects of the glycosylation status of Ex3-scDb-Fc on its aggregation, we first analyzed the glycosylation pattern of the monomeric, dimeric, and large aggregated forms of the antibody. The same amount of each form (0.5 mg) was treated with GPF and their oligosaccharides were purified and pyridylaminated. The N-linked oligosaccharide structures of each form were estimated by reverse-phase HPLC analysis on an ODS column. The elution profiles of the purified oligosaccharides are shown in Fig. 3A. Previously, our group reported the glycosylation pattern of Ex3-scDb-Fc (23), in which six peaks corresponding to N-oligosaccharides were detected. HPLC analysis in the present
A Fluorescence intensity (a.u.)
1,200,000
B
900,000
G1F
G0F
600,000
G2F
300,000 0 26
27
28
29
30
31
Retention time (min)
Percentage of glycoform (%)
The relative ratio of the monomeric, dimeric, and large aggregated forms were estimated to be 72%, 10% and 18%, respectively. Furthermore, we examined the solution structure of the monomeric, dimeric and large aggregated antibodies by circular dichroism (CD) and ANS fluorescence spectroscopy (Fig. 2C and D). CD spectra of the monomeric and dimeric forms exhibited the typical secondary structure of the antibody molecule, and the spectra for both states superposed very well. ANS binds to the hydrophobic area of the exposed side chains of the protein and shows large fluorescence intensity. The ANS florescence spectrum in the presence of monomeric Ex3-scDb-Fc was nearly identical to that of free ANS, indicating that the hydrophobic area of the monomer was completely buried in the antibody structure. The fluorescence intensity of the dimer was slightly increased. These data indicate that in the cell culture process, Ex3-scDb-Fc dimerizes without an accompanying structural change to the antibody. By contrast, the solution structure of large aggregates exhibited the formation of non-native b-strand structures and significantly increased the exposure of the hydrophobic regions to the solution (Fig. 2C and D). This series of results indicates that the large aggregated form of Ex3-scDb-Fc is in the misfolded state with a non-native b-strand structure and surface-exposed hydrophobic regions.
60 50 40 30 20 10 0
G0F
G1F
G2F
FIG. 3. (A) Reverse-phase HPLC profile of pyridylaminated derivatives of the N-linked oligosaccharides released from monomeric (thick gray line), dimeric (thin black line), and large aggregated (thick black line) forms of Ex3-scDb-Fc. (B) Relative fractions of the estimated oligosaccharide structures. Open, gray-filled, and black-filled bars represent the percentages in the monomeric, dimeric, and large aggregated forms, respectively. Error bars denote the standard deviation (n ¼ 3).
Please cite this article in press as: Onitsuka, M., et al., Glycosylation analysis of an aggregated antibody produced by Chinese hamster ovary cells in bioreactor culture, J. Biosci. Bioeng., (2013), http://dx.doi.org/10.1016/j.jbiosc.2013.11.001
ONITSUKA ET AL.
A
J. BIOSCI. BIOENG.,
B
1.2
0.8
Abs. at 405nm
Abs. at 405nm
1.0
C
1.0
Relative detection of sugar residue
4
0.8 0.6 0.4
0.6 0.4 0.2
0.2 0.0 0
50 100 150 200 250
[Ex3-scDb-Fc] (ng)
0.0
0
50 100 150 200 250
[Ex3-scDb-Fc] (ng)
1.2 1.0 0.8 0.6 0.4 0.2 0.0
Galactose (RCA I)
Fucose (LCA)
FIG. 4. Lectin binding assays (LBA) for the detection of specific sugar residues in Ex3-scDb-Fc. (A) RCA I assay for the detection of galactose residues. (B) LCA assay for the detection of fucose residues. (A, B) Open, gray-filled, and black-filled circles correspond to monomer, dimer, and large aggregates, respectively. Continuous lines represent the fits of the nonlinear regression analysis. (C) Relative detection of N-linked oligosaccharide residues estimated from the slope of the regression analysis. Open, gray-filled, and black-filled bars represent the detection of monomeric, dimeric, and large aggregated forms, respectively. Error indicates the standard error of regression analysis.
study corresponded well with the data from the previous study. Among the six peaks, three peaks with retention times of 27.5e30 min were dominant (Fig. 3A) relative to the other three peaks, and constituted approximately 90% of the total peak area. The retention time of the commercial oligosaccharide standards and MALDI-TOF mass spectroscopy analysis of the fractionated peaks ensured that the three peaks corresponded to the agalactosyl, mono-, and bi-galactosyl oligosaccharides with fucose (G0F, G1F, and G2F, respectively). We focused on the G0F, G1F, and G2F oligosaccharides in this study. All estimated Noligosaccharides, including minor components, are summarized in Supplementary data (Fig. S2). The molar ratio of these three oligosaccharides relative to total oligosaccharides was calculated based on the peak area of the HPLC profile. The observed glycosylation patterns are summarized in Fig. 3B. The overall distributions of all forms were similardthe profile exhibited a low level of fully galactosylated oligosaccharide with fucose (G2F), as reported in our previous study (23). The percentage of G2F in the large aggregates was slightly higher for the monomeric and dimeric forms. Both agalactosyl and monogalactosyl glycoforms accounted for over 30% of the total N-linked oligosaccharides. These two main components showed a similar molecular ratio among the three states, i.e., the monomer, dimer, and large aggregates. Overall, these results indicate that there is no apparent difference in the relative molar proportions of the three forms of Ex3-scDb-Fc antibody. However, a striking difference was shown in the absolute peak area in the HPLC profile (Fig. 3A). The profile of the dimeric and large aggregated forms exhibited lower peak heights as compared with monomeric Ex3-scDb-Fc, suggesting that the total amounts of N-oligosaccharide rather than the glycosylation pattern had dominant effects on aggregate formation. We next performed lectin binding assays to quantitatively estimate the total amounts of saccharide residues. The lectin binding assay enabled us to directly observe the degree of specific sugar residues in protein samples using enzyme-linked lectins (24,25). We used biotinylated R. communis agglutinin I (RCA I) and L. culinaris agglutinin (LCA), which are specific for terminal galactose residues and core fucose residues, respectively (Fig. 1B). The assay plate wells were directly coated with Ex3-scDb-Fc solution and the sugar residues were detected by an HRP-catalyzed enzymatic reaction. A schematic diagram of the lectin binding assay and example results are shown in Figs. S3 and S4. In the RCA I and LCA assays, the absorbance at 405 nm of the intact monomer increased proportionally up to 250 ng, whereas that of the deglycosylated form prepared by GPF treatment was almost identical throughout the range (Fig. S4). These data indicated that the assay system
works correctly and that the Ex3-scDb-Fc used in this study mainly contains N-linked oligosaccharides. We evaluated the relative levels of galactosylation and fucosylation from the slope following a non-linear regression analysis (Fig. 4A and B). The highest slope value was normalized to 1 (Fig. 4C). The assay was performed using denatured Ex3-scDb-Fc samples, which enabled us to detect the total amount of sugar residues because N-linked oligosaccharides, which are located between two CH2 domains, are surface exposed by complete unfolding. Both the RCA I and LCA assays showed that the relative levels of galactosylation and fucosylation in the dimeric and large aggregated forms were equally reduced to 70e80% of the level in monomeric Ex3-scDb-Fc (Fig. 4C). Again, there were no apparent differences in the proportions of the main glycoforms, G0F, G1F, and G2F, among the monomeric, dimeric, and large aggregated forms of the antibody (Fig. 3B). Therefore, we interpret the reduction in both galactose and fucose residues as indicating that the lack of N-linked oligosaccharides, that is deglycosylation or aglycosylation, occurred in a proportion of the dimeric and large aggregated forms. In addition to denaturing conditions, we performed the assay in non-denaturing conditions (Fig. S5), and the large aggregates showed the highest levels of galactose and fucose residues. Lectin binding assays in non-denaturing conditions were highly dependent on the antibody conformation, because the Nlinked oligosaccharides are located between the two CH2 domains of the Fc region (27). The highest levels of galactose and fucose residues detected demonstrated that the N-glycosylated region in large aggregates was surface exposed owing to its misfolded state. N-Glycosylation analysis revealed that the most striking difference in glycosylation status among the different forms of Ex3-scDb-Fc was the total amount of N-glycosylation rather than the glycosylation pattern. Effects of glycosylation on the structure and stability of Ex3scDb-Fc We examined the effects of N-linked glycosylation on the stability of, and aggregate formation by, the Ex3-scDb-Fc antibody. CD spectra of the monomeric form, including the deglycosylated monomer, showed that the lack of N-linked oligosaccharides had no effect on the overall secondary structure of the antibody (Fig. 5A). In contrast, thermal unfolding experiments revealed the contribution of N-glycosylation to the stability of Ex3-scDb-Fc (Fig. 5B). At temperatures above 45 C, the CD signal intensity at 218 nm showed two characteristic changesddecreased intensity up to about 60 C and then increased intensity above 60 C. The former and the latter signal changes are interpreted as structural unfolding and aggregation/ precipitation, respectively. In the deglycosylated form, the temperature of the aggregation/precipitation process (increased
Please cite this article in press as: Onitsuka, M., et al., Glycosylation analysis of an aggregated antibody produced by Chinese hamster ovary cells in bioreactor culture, J. Biosci. Bioeng., (2013), http://dx.doi.org/10.1016/j.jbiosc.2013.11.001
VOL. xx, 2013
CD intensity (mdeg)
A
GLYCOSYLATION ANALYSIS OF AGGREGATED ANTIBODY
10 5 0 -5 -10 200
Relative CD intensity
B
210
220
230
240
250
Wavelength (nm)
1 0 -1 -2 -3 30
40
50
60
70
80
Temperature (degree) FIG. 5. (A) Circular dichroism spectra of intact monomeric (thick gray line) and deglycosylated forms by GPF treatment (thin black line). (B) Thermal unfolding monitored by CD intensity at 218 nm. The thick gray line and the thin black line represent the curves for intact monomeric and deglycosylated forms, respectively.
CD intensity) was decreased by about 2 C, indicating that the resistance to aggregation of Ex3-scDb-Fc was actually reduced by deglycosylation. Molecular insights into how deglycosylation induces antibody aggregation in CHO cell culture The N-linked oligosaccharides on IgG interact with the polypeptide in the CH2 domain, and the carbohydrateeprotein interaction plays a crucial role in structural conformation and integrity for FcgRs binding (28,29). In addition to the functional aspect, the carbohydrateeprotein interaction contributes to the structural stability of IgG antibodies (3,4), and recent studies have shown that deglycosylation promotes antibody aggregation (5,6). However, the impact of glycosylation status on aggregation during the manufacturing of therapeutic antibodies has not been assessed. Here, we showed that a proportion (approximately 20e30%) of dimeric and large aggregated Ex3-scDb-Fc molecules was deglycosylated or aglycosylated, indicating the importance of glycosylation heterogeneity as a possible reason for antibody aggregation during CHO cell culture where the culture conditions influence the glycosylation heterogeneity. These aggregations impair antigen binding. Antibody dimer partially loses the antigen binding affinity (30), and large aggregates probably fully lost their binding activity due to misfolding (Fig. 2C). Deglycosylation and aglycosylation would have negative impacts on antibody structure and antigen binding. Protein aggregation is well represented by the nucleationdependent aggregation model, where small oligomers are initially organized as the aggregation nucleus and then subsequently polymerized into large aggregates (31,32). A recent review highlighted the role of monomeric and oligomeric forms with a nativelike structure as the nucleus for misfolded large aggregates under
5
non-denaturing conditions (33). Characterization of the aggregation nucleus with the native-like structure would provide clues to understanding protein aggregation. In the present study, we observed the dimeric form of Ex3-scDb-Fc with a native-like structure and reduced N-glycosylation. Molecular dynamics simulation predicts the aggregation prone motif in IgG antibodies, where the hydrophobic residues covered by N-linked oligosaccharides in the CH2 domain become the aggregation prone motif by deglycosylation (5,6). In addition, IgG antibodies have tendency to self-associate by a weak interaction among antibody molecules at Fc regions (34). A series of facts provide molecular insights into the mechanism of antibody aggregation in recombinant CHO cell culture. The exposure of hydrophobic residues by the lack of Nglycosylation converts monomeric Ex3-scDb-Fc into a dimer by hydrophobic stacking, and the dimer serves as an aggregation nucleus and promotes polymerization into large aggregates. This hypothetical model implies that deglycosylation is a possible reason for the formation of the aggregation nucleus, although the direct relationship between the dimer and the aggregation nucleus remains to be clarified. It is well established that N-glycosylation plays an important role in the folding and secretion of glycoproteins in the endoplasmic reticulum and Golgi apparatus (19,35,36). It remains unclear why a lack of N-linked oligosaccharides was present in the Ex3-scDb-Fc antibody during CHO cell culture. Endogenous Nglycosidases may leak from the CHO cells into the culture medium following cell death, and may deglycosylate secreted antibodies. Aglycosylated antibody molecules may be secreted following the depletion of intracellular nucleotide sugars as glycosylation precursors in the death phase of cell culture. To test these possibilities, the time-dependent N-glycosylation status of the antibody produced in CHO cell culture should be investigated in future work. In addition, supplementation of the inhibitor of N-glycosidases (37) and glycosylation precursors (38) would prevent the deglycosylation and aglycosylation, which should be examined in future work. Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jbiosc.2013.11.001. ACKNOWLEDGMENTS This study was supported by the Advanced Research for Medical Products Mining Programme of the National Institute of Biomedical Innovation (NIBIO). References 1. Arnold, J. N., Wormald, M. R., Sim, R. B., Rudd, P. M., and Dwek, R. A.: The impact of glycosylation on the biological function and structure of human immunoglobulins, Annu. Rev. Immunol., 25, 21e50 (2007). 2. Jefferis, R.: Glycosylation as a strategy to improve antibody-based therapeutics, Nat. Rev. Drug Discov., 8, 226e234 (2009). 3. Ghirlando, R., Lund, J., Goodall, M., and Jefferis, R.: Glycosylation of human IgG-Fc: influences on structure revealed by differential scanning micro-calorimetry, Immunol. Lett., 68, 47e52 (1999). 4. Mimura, Y., Church, S., Ghirlando, R., Ashton, P. R., Dong, S., Goodall, M., Lund, J., and Jefferis, R.: The influence of glycosylation on the thermal stability and effector function expression of human IgG1-Fc: properties of a series of truncated glycoforms, Mol. Immunol., 37, 697e706 (2000). 5. Voynov, V., Chennamsetty, N., Kayser, V., Helk, B., Forrer, K., Zhang, H., Fritsch, C., Heine, H., and Trout, B. L.: Dynamic fluctuations of protein-carbohydrate interactions promote protein aggregation, PLoS One, 4, e8425 (2009). 6. Kayser, V., Chennamsetty, N., Voynov, V., Forrer, K., Helk, B., and Trout, B. L.: Glycosylation influences on the aggregation propensity of therapeutic monoclonal antibodies, Biotechnol. J., 6, 38e44 (2011). 7. Cromwell, M. E., Hilario, E., and Jacobson, F.: Protein aggregation and bioprocessing, AAPS. J., 8, E572eE579 (2006). 8. Vázquez-Rey, M. and Lang, D. A.: Aggregates in monoclonal antibody manufacturing processes, Biotechnol. Bioeng., 108, 1494e1508 (2011).
Please cite this article in press as: Onitsuka, M., et al., Glycosylation analysis of an aggregated antibody produced by Chinese hamster ovary cells in bioreactor culture, J. Biosci. Bioeng., (2013), http://dx.doi.org/10.1016/j.jbiosc.2013.11.001
6
ONITSUKA ET AL.
9. Rosenberg, A. S.: Effects of protein aggregates: an immunologic perspective, AAPS J., 8, E501eE507 (2006). 10. Maas, C., Hermeling, S., Bouma, B., Jiskoot, W., and Gebbink, M. F.: A role for protein misfolding in immunogenicity of biopharmaceuticals, J. Biol. Chem., 282, 2229e2236 (2007). 11. Gatti, M., Wlaschin, K. F., Nissom, P. M., Yap, M., and Hu, W. S.: Comparative transcriptional analysis of mouse hybridoma and recombinant Chinese hamster ovary cells undergoing butyrate treatment, J. Biosci. Bioeng., 103, 82e91 (2007). 12. Liu, C. H. and Chen, L. H.: Promotion of recombinant macrophage colony stimulating factor production by dimethyl sulfoxide addition in Chinese hamster ovary cells, J. Biosci. Bioeng., 103, 45e49 (2007). 13. Fujiwara, M., Koizumi, S., and Takagi, M.: Effect of static pressure on intracellular pH of adhesive Chinese hamster ovary cells, J. Biosci. Bioeng., 104, 510e512 (2007). 14. Omasa, T., Takami, T., Ohya, T., Kiyama, E., Hayashi, T., Nishii, H., Miki, H., Kobayashi, K., Honda, K., and Ohtake, H.: Overexpression of GADD34 enhances production of recombinant human antithrombin III in Chinese hamster ovary cells, J. Biosci. Bioeng., 106, 568e573 (2008). 15. Kaneko, Y., Sato, R., and Aoyagi, H.: Changes in the quality of antibodies produced by Chinese hamster ovary cells during the death phase of cell culture, J. Biosci. Bioeng., 109, 281e287 (2010). 16. Omasa, T., Onitsuka, M., and Kim, W. D.: Cell engineering and cultivation of Chinese hamster ovary (CHO) cells, Curr. Pharm. Biotechnol., 11, 233e240 (2010). 17. Andersen, D. C. and Goochee, C. F.: The effect of cell-culture conditions on the oligosaccharide structures of secreted glycoproteins, Curr. Opin. Biotechnol., 5, 546e549 (1994). 18. Yang, M. and Butler, M.: Effects of ammonia and glucosamine on the heterogeneity of erythropoietin glycoforms, Biotechnol. Prog., 18, 129e138 (2002). 19. Hossler, P., Khattak, S. F., and Li, Z. J.: Optimal and consistent protein glycosylation in mammalian cell culture, Glycobiology, 19, 936e949 (2009). 20. Jefferis, R.: Recombinant antibody therapeutics: the impact of glycosylation on mechanisms of action, Trends Pharmacol. Sci., 30, 356e362 (2009). 21. Asano, R., Kawaguchi, H., Watanabe, Y., Nakanishi, T., Umetsu, M., Hayashi, H., Katayose, Y., Unno, M., Kudo, T., and Kumagai, I.: Diabody-based recombinant formats of humanized IgG-like bispecific antibody with effective retargeting of lymphocytes to tumor cells, J. Immunother., 31, 752e761 (2008). 22. Asano, R., Ikoma, K., Kawaguchi, H., Ishiyama, Y., Nakanishi, T., Umetsu, M., Hayashi, H., Katayose, Y., Unno, M., Kudo, T., and Kumagai, I.: Application of the Fc fusion format to generate tag-free bi-specific diabodies, FEBS J., 277, 477e487 (2010). 23. Kim, W. D., Tokunaga, M., Ozaki, H., Ishibashi, T., Honda, K., Kajiura, H., Fujiyama, K., Asano, R., Kumagai, I., Omasa, T., and Ohtake, H.: Glycosylation pattern of humanized IgG-like bispecific antibody produced by recombinant CHO cells, Appl. Microbiol. Biotechnol., 85, 535e542 (2010). 24. Onitsuka, M., Kim, W. D., Ozaki, H., Kawaguchi, A., Honda, K., Kajiura, H., Fujiyama, K., Asano, R., Kumagai, I., Ohtake, H., and Omasa, T.: Enhancement of sialylation on humanized IgG-like bispecific antibody by overexpression of
J. BIOSCI. BIOENG., a2,6-sialyltransferase derived from Chinese hamster ovary cells, Appl. Microbiol. Biotechnol., 94, 69e80 (2012). 25. Omasa, T., Tanaka, R., Doi, T., Ando, M., Kitamoto, Y., Honda, K., Kishimoto, M., and Ohtake, H.: Decrease in antithrombin III fucosylation by expressing GDP-fucose transporter siRNA in Chinese hamster ovary cells, J. Biosci. Bioeng., 106, 168e173 (2008). 26. Arakawa, T., Philo, J. S., Tsumoto, K., Yumioka, R., and Ejima, D.: Elution of antibodies from a Protein-A column by aqueous arginine solutions, Protein Expr. Purif., 36, 244e248 (2004). 27. Sondermann, P., Huber, R., Oosthuizen, V., and Jacob, U.: The 3.2-Å crystal structure of the human IgG1 Fc fragment-Fc gRIII complex, Nature, 406, 267e273 (2000). 28. Krapp, S., Mimura, Y., Jefferis, R., Huber, R., and Sondermann, P.: Structural analysis of human IgG-Fc glycoforms reveals a correlation between glycosylation and structural integrity, J. Mol. Biol., 325, 979e989 (2003). 29. Yamaguchi, Y., Nishimura, M., Nagano, M., Yagi, H., Sasakawa, H., Uchida, K., Shitara, K., and Kato, K.: Glycoform-dependent conformational alteration of the Fc region of human immunoglobulin G1 as revealed by NMR spectroscopy, Biochim. Biophys. Acta, 1760, 693e700 (2006). 30. Paul, R., Graff-Meyer, A., Stahlberg, H., Lauer, M. E., Rufer, A. C., Beck, H., Briguet, A., Schnaible, V., Buckel, T., and Boeckle, S.: Structure and function of purified monoclonal antibody dimers induced by different stress conditions, Pharm. Res., 29, 2047e2059 (2012). 31. Harper, J. D. and Lansbury, P. T., Jr.: Models of amyloid seeding in Alzheimer’s disease and scrapie: mechanistic truths and physiological consequences of the time-dependent solubility of amyloid proteins, Annu. Rev. Biochem., 66, 385e407 (1997). 32. Nilsson, M. R.: Techniques to study amyloid fibril formation in vitro, Methods, 34, 151e160 (2004). 33. Chiti, F. and Dobson, C. M.: Amyloid formation by globular proteins under native conditions, Nat. Chem. Biol., 5, 15e22 (2009). 34. Nishi, H., Miyajima, M., Wakiyama, N., Kubota, K., Hasegawa, J., Uchiyama, S., and Fukui, K.: Fc domain mediated self-association of an IgG1 monoclonal antibody under a low ionic strength condition, J. Biosci. Bioeng., 112, 326e332 (2011). 35. Fujiyama, K., Misaki, R., Sakai, Y., Omasa, T., and Seki, T.: Change in glycosylation pattern with extension of endoplasmic reticulum retention signal sequence of mouse antibody produced by suspension-cultured tobacco BY2 cells, J. Biosci. Bioeng., 107, 165e172 (2009). 36. Misaki, R., Sakai, Y., Omasa, T., Fujiyama, K., and Seki, T.: N-terminal vacuolar sorting signal at the mouse antibody alters the N-linked glycosylation pattern in suspension-cultured tobacco BY2 cells, J. Biosci. Bioeng., 112, 476e484 (2011). 37. Misaghi, S., Pacold, M. E., Blom, D., Ploegh, H. L., and Korbel, G. A.: Using a small molecule inhibitor of peptide: N-glycanase to probe its role in glycoprotein turnover, Chem. Biol., 11, 1677e1687 (2004). 38. Wong, N. S., Wati, L., Nissom, P. M., Feng, H. T., Lee, M. M., and Yap, M. G.: An investigation of intracellular glycosylation activities in CHO cells: effects of nucleotide sugar precursor feeding, Biotechnol. Bioeng., 107, 321e336 (2010).
Please cite this article in press as: Onitsuka, M., et al., Glycosylation analysis of an aggregated antibody produced by Chinese hamster ovary cells in bioreactor culture, J. Biosci. Bioeng., (2013), http://dx.doi.org/10.1016/j.jbiosc.2013.11.001