Technical Note pubs.acs.org/ac
When Good Intentions Go Awry: Modification of a Recombinant Monoclonal Antibody in Chemically Defined Cell Culture by Xylosone, an Oxidative Product of Ascorbic Acid Chris Chumsae,*,†,§,⊥ Patrick Hossler,‡ Haly Raharimampionona,† Yu Zhou,†,∥ Sean McDermott,‡ Chris Racicot,‡ Czeslaw Radziejewski,*,† and Zhaohui Sunny Zhou*,§ †
Protein Analytics, Process Sciences Department, AbbVie Bioresearch Center, Worcester, Massachusetts 01605, United States Cell Culture, Process Sciences Department, AbbVie Bioresearch Center, Worcester, Massachusetts 01605, United States § Barnett Institute of Chemical and Biological Analysis, Department of Chemistry and Chemical Biology, Northeastern University, Boston, Massachusetts 02115, United States ‡
Downloaded by UNIV LAVAL on September 15, 2015 | http://pubs.acs.org Publication Date (Web): July 21, 2015 | doi: 10.1021/acs.analchem.5b00801
S Supporting Information *
ABSTRACT: With the advent of new initiatives to develop chemically defined media, cell culture scientists screen many additives to improve cell growth and productivity. However, the introduction or increase of supplements, typically considered beneficial or protective on their own, to the basal media or feed stream may cause unexpected detrimental consequences to product quality. For instance, because cultured cells are constantly under oxidative stress, ascorbic acid (vitamin C, a potent natural reducing agent) is a common additive to cell culture media. However, as reported herein, a recombinant monoclonal antibody (adalimumab) in cell culture was covalently modified by xylosone (molecular weight 148), an oxidative product of ascorbate. Containing reactive carbonyl groups, xylosone modifies various amines (e.g., the N-termini of the heavy and light chains and susceptible lysines), forming either hemiaminal (+148 Da) or Schiff base (imine, +130 Da) products. Our findings show, for the first time, that ascorbate-derived xylosone can contribute to an increase in molecular heterogeneity, such as acidic species. Our work serves as a reminder that additives to cell culture and their metabolites may become reactive and negatively impact the overall product quality and should be carefully monitored with any changes in cell culture conditions.
U
Pertinent to this study, media components or additives have been shown to affect product quality and protein modifications.23,24 The best known example is arguably glycation of proteins by glucose,25−27 an essential nutrient as the main energy source of cultured cells. While unavoidable, glycation nevertheless is predictable. On the other hand, it is challenging to predict and detect modifications by secondary metabolites or byproducts. For example, we previously reported that a recombinant monoclonal antibody was unexpectedly modified by an accumulation of methylglyoxal (MGO) during cell culture due to a change in media that was perceived as beneficial.28 Methylglyoxal is a dicarbonyl compound that is generated as a byproduct of glycolysis.28,29 Additionally, changes in the cell culture conditions, specifically the redox state, ultimately lead to increased amounts of this modification.30 Modifications from other reactive species include cysteinylation,31 glutathionylation,32,33 and N-homocysteintylation from homocysteine thiolactone.34 Therefore, changes to
ntil recently, cell culture scientists mostly focused on cell growth and protein expression level, but as demonstrated herein, the quality of the final products has recently emerged as another critical attribute to be considered. To this point, variations in product quality, or microheterogeniety, are mostly attributed to a myriad of post-translational modifications (PTMs).1−6 In order to reduce variability in product quality, cell growth, and expression levels, the current trend in recombinant monoclonal antibody production has been to move from complex undefined hydrolysate media to chemically defined media.7,8 Furthermore, it has been shown that modulation of the supplemental feed can impact the product quality of the protein drug.9−14 A common additive is the much-storied vitamin C (ascorbic acid). In addition to its function as a cofactor for collagen synthesis,15 it has been implicated as an antioxidant (biological reductant) and potent scavenger of reactive oxygen species (ROS).16−19 Large scale production of recombinant monoclonal antibodies require higher levels of oxygen to maintain higher cell densities.20 Such conditions may generate reactive oxygen species.21,22 Therefore, including antioxidants such as ascorbic acid is generally considered protective and desirable. © 2015 American Chemical Society
Received: February 16, 2015 Accepted: July 7, 2015 Published: July 7, 2015 7529
DOI: 10.1021/acs.analchem.5b00801 Anal. Chem. 2015, 87, 7529−7534
Technical Note
Analytical Chemistry
Additionally, the formation of xylosone from ascorbic acid and the subsequent modification to the antibody was confirmed using stable isotope labeled ascorbic acid. Details are described in the Supporting Information.
Downloaded by UNIV LAVAL on September 15, 2015 | http://pubs.acs.org Publication Date (Web): July 21, 2015 | doi: 10.1021/acs.analchem.5b00801
the cell culture medium may result in unexpected changes in product quality.35,36 As reported herein, we observed an increase in acidic species in a recombinant monoclonal antibody supplemented with ascorbic acid during cell culture. Additionally, we observed two unidentified masses in the reduced liquid chromatography/ mass spectrometry (LC/MS) analysis of acidic fractions from the weak cation exchange chromatography which exhibited molecular weight increases of 130 and 148, respectively. Detailed analyses revealed that these modifications occurred on the primary amines of the N-termini of heavy and light chains and also susceptible lysine residues. This was confirmed by in vitro incubation of native antibody with increasing concentrations of ascorbic acid. Given that the molecular weight of ascorbate is 176, it was hypothesized that metabolites or degration produts of this nutrient were the culprits. This was confirmed in more detailed mechanistic investigations using 13 C labeled ascorbate. As illustrated in Scheme 1, ascorbate is
■
RESULTS AND DISCUSSION Ascorbate Supplement in Cell Culture Induced New Acidic Variants. A recombinant monoclonal antibody was expressed in shake flasks using various supplementation and additives. Of specific interest were cell culture conditions under which the concentration of ascorbate was increased (0, 0.1, 1, and 3 mg/mL, respectively). When antibodies were analyzed by weak cation exchange chromatography, the formation of new acidic species directly correlated with ascorbate concentration (Figure S-1 in the Supporting Information). This observation led us to initiate a detailed structural analysis of the recombinant monoclonal antibody in order to determine the chemical nature and cause of the increase in charge heterogeneity. LC/MS Analysis of Reduced Antibody. Without separating antibody variants, reduced LC/MS analysis of the recombinant monoclonal antibody heavy chain and light chain from the ascorbate supplemented cultures did not show obvious differences in the mass spectra (Figure S-2 in the Supporting Information). Therefore, fractionation of the earlier eluting peaks and the Lys-0 peak from weak cation exchange (Figure S-1 in the Supporting Information) was employed. The analysis of the Lys-0 fraction by reduced LC/MS revealed a deconvoluted light chain mass spectra which was in good agreement with the expected light chain mass (Figure S-3A in the Supporting Information). Analysis of the mass spectra resulting from Peak A, however, exhibited discernible differences as shown in Figure S-3B in the Supporting Information. The deconvoluted light chain mass spectra showed the expected mass as well as several lower abundance peaks with higher molecular-weight. Two of the masses corresponded to mass increases of 148 and 130 Da. The levels of modification observed by the UV absorption (WCX) and mass spectrometry appear different. The difference may be attributed to several possibilities. First, Figure S-1 was based on UV absorption (WCX) that depends on the chromophores in each antibody variant, which were likely to be the same or similar; in comparison, Figure S-3 in the Supporting Information was based on ionization efficiencies of each species, which was likely to differ considerably given the noticeable changes in charge characteristics as evident from the large gap in elution time. Second, Figure S-3 in the Supporting Information accounts for the modification at the light chain only. The modification was also observed on heavy chain. In addition, each antibody has two light chains and two heavy chains, so the overall modification of the whole antibody should be more than individual light chain or heavy chain. In order to elucidate the nature of the additional species, a pure fraction of the Lys-0 species was treated with ascorbic acid in vitro. The deconvoluted spectrum (Figure S-3C in the Supporting Information) of the in vitro sample was highly similar to that from cell culture (Figure S-3B in the Supporting Information). First, the observed molecular weight of 23 408.5 represented the unmodified light chain from Lys-0 incubated with ascorbate and was in good agreement with the theoretical value of 23 408.1. Second, two other lower intensity peaks were also observed with masses of 23 538.3 and 23 556.6 Da which corresponded to mass increases of 130 and 148 Da, respectively
Scheme 1. Degradation of Ascorbic Acid to Xylosone and Subsequent Reaction with a Primary Amine to Form a Hemiaminal and Schiff Basea
a
The red dot denotes the carbon which is released as CO2.
first oxidized to dehydroascorbic acid. Subsequent decarboxylation generates xylosone.37,38 Xylosone is highly reactive and is capable of modifying susceptible primary amines, resulting in mass increases of 148 and 130 Da for the hemiaminal and Schiff base, respectively (Scheme 1). To the best of our knowledge, this is the first report of an ascorbate-originated xylosone modification of a recombinant monoclonal antibody in vitro and in cell culture.
■
MATERIALS AND METHODS Similar to our previous studies of antibody modifications by citric acid and methylglyoxal,28,39 the antibody variants were separated and detected by weak cation exchange chromatography (WCX), the molecular weights of the light and heavy chains from the reduced antibody were determined by mass spectrometry, and site of modifications were determined by peptide mapping (tryptic digestion followed by LC/MS). 7530
DOI: 10.1021/acs.analchem.5b00801 Anal. Chem. 2015, 87, 7529−7534
Technical Note
Downloaded by UNIV LAVAL on September 15, 2015 | http://pubs.acs.org Publication Date (Web): July 21, 2015 | doi: 10.1021/acs.analchem.5b00801
Analytical Chemistry
3.0 mg/mL) over time (3, 5, 11, and 14 days) confirmed our findings as shown in Figures S-12 and S-13 in the Supporting Information. Elucidation of the Modification Agent As Xylosone. The mass of ascorbate is 176 Da, which is 28 Da greater than the +148 modification observed, so it was unlikely that ascorbate itself modified the antibody but rather its degradation product(s) was the likely culprit(s). Another hint is the 18 Da difference (130 vs 148 Da) between the two observed modifications, suggesting a dehydration (elimination of a water molecule) following the initial reaction; this is reminiscent of modifications of lysine or arginine residues by methylglyoxal (MGO) or other carbonyl-containing molecules. All together, we postulated that the bona fide reactive species should be degraded from ascorbate and also contain reactive carbonyl group(s). Xylosone thus emerged as a likely candidate as it has been reported as a degradation product of ascorbate (see Scheme 1 for its formation pathway).37 Furthermore, xylosone has a mass of 148 Da and two carbonyl groups. The 130 Da adduct may be due to the loss of a water molecule following the initial addition reaction between the carbonyl group of xylosone and the amines in the protein (hemiaminal to Schiff base as shown in Scheme 1). Incubation of Antibody with 13C Regio-Labeled Ascorbate. Isotopic labeling and tracing often provides detailed mechanistic insights.45−47 In order to confirm that ascorbate was degraded to xylosone (losing the carbon atom at 1 position), which in turn modified the antibody, we used ascorbate regio-specifically labeled with13C (at C1, C2 or C3 carbon, respectively, as shown in Figure 1A). Following incubation with unlabeled ascorbate or one of the three regio-labeled ascorbate molecules, the antibody was analyzed by tryptic peptide mapping with mass spectrometry detection. The results (Figure 1B) lended conclusive evidence supporting the mechanism in Scheme 1: labeling at the C1 position (13C or 12 C) produced the same mass spectra; in contrast, ascorbate with 13C labeling at C2 or C3 shifted the peaks to 1 Da higher compared to 12C ascorbate (+149 vs +148 and 131 vs 130, respectively). In addition, MS/MS localized the modification to the light chain N-terminal residue including the heavy label for the C2 and C3 regio-labeled ascorbates as shown in Figure S-15 in the Supporting Information. Thus, the data have verified that the C1 carbon in ascorbate is lost but neither C2 nor C3. Therefore, our data confirmed that ascorbate was converted to xylosone with the concomitant loss of C1 carbon atom as illustrated in Scheme 1. Chemical Nature of the Modifications. When a carbonyl reacts with an amine, two products may form: the initial addition reaction leads to a hemiaminal (see Scheme 1) with a mass equals to the total of the masses of the two reactants (amine and carbonyl; for xylosone, 148 Da); a subsequent elimination of a water molecule (18 Da) leads to a Schiff base (imine, see Scheme 1) with a mass that is 18 Da less than the hemiaminal (148 − 18 = 130 Da). Hence the masses of the two adducts match with the observed masses. The underlying chemistry, including stereo-, regio-, and positional isomers of the adducts, also explains the multiple isobaric peaks observed (see Figures S-5 and S-7 in the Supporting Information). For instance, the formation of the hemiaminal can result in two stereoisomers (the chiral center is denoted with an ∗ in Scheme 1). Two other factors further complicate the situation: first, two carbonyls exist in xylosone; and second, xylosone may exist in both cyclic and acyclic forms
(Figure S-3C in the Supporting Information). Furthermore, these two masses are in good agreement with the two masses representing +130 Da and +148 Da from cell culture suggesting that they were generated due to the ascorbic acid supplementation. Tryptic Mapping and LC/MS/MS Detection. Tryptic mapping was performed on the recombinant monoclonal antibody produced in cell culture supplemented with up to 3 mg/mL of ascorbate and on the 0 Lys isoform incubated with ascorbate in vitro. It is important to note that the mass difference between these two species is 18 Da; therefore, they may be related and involve the loss of a water molecule, reminiscing modifications of lysine or arginine residues by methylglyoxal (MGO) or other carbonyl-containing molecules. Therefore, we searched for miscleaved trypic peptides and found none with an internal arginine. Then, we examined potential modifications of primary amines, specifically the Ntermini and miscleaved lysine containing peptides, which did indeed produce several possible sites of modification as shown in Table S-1 in the Supporting Information. Specifically, the light chain N-terminal peptide had two chromatographically resolved +148 Da peaks with close elution to the native peptide and a +130 peak that eluted later as shown in Figure S-5 in the Supporting Information. The MS/MS spectra were analyzed for the native and modified +148 Da and +130 Da peptides that all had y ion series which were in good agreement with the predicted amino acid sequence and covered the entire peptide except the first two N-terminal residues as shown in Figure S-6 in the Supporting Information. The b ion series was limited but had strong signal for the first three residues at the N-termini. In addition, a strong signal for the native and modified a1 and a2 ions was also present. Altogether, our data conclusively localized both modifications (+148 Da and +130 Da) to the N-terminal amine of the light chain (Asp1). The N-terminal peptides of the heavy chain also exhibited a parent peak in the native extracted mass chromatogram, two isobaric peaks in the extracted mass chromatogram corresponding to a mass increase of +148 Da and two isobaric peaks in the extracted mass chromatograms corresponding to a mass increases of +130 Da (Figure S-7 in the Supporting Information). Once again, the b ion series was used to definitively assign the +148 Da and +130 Da modification to the N-termini of the two sets of resolved modified peptides (see Figure S-8 in the Supporting Information). The analysis of other peptides from the recombinant monoclonal antibody showed the modification of lysine residues (see Figures S-9 and S-10 in the Supporting Information). However, lysine residues exhibited lower susceptibility to modification as compared to the N-terminal primary amines as shown in the analysis of the in vitro ascorbate incubation over time and at increasing concentrations of ascorbate (Figure S-11 in the Supporting Information). The observation is in good agreement with the lower pKa of the Nterminal amines (∼8) as compared to the pKa of the lysyl amine on the side chain (∼10);40,41 and of course, these modifications are also likely to be influenced by other factors such as the antibody structure and microenvironment.23,28,34,42−44 In addition, it is important to note that only a +148 Da species was seen for all modified lysine residues in the antibody, again suggesting local environment is likely to affect the nature and distribution of various chemical forms, as discussed in greater details later. Lastly, in vitro (0.1, 1.0, and 7531
DOI: 10.1021/acs.analchem.5b00801 Anal. Chem. 2015, 87, 7529−7534
Technical Note
Downloaded by UNIV LAVAL on September 15, 2015 | http://pubs.acs.org Publication Date (Web): July 21, 2015 | doi: 10.1021/acs.analchem.5b00801
Analytical Chemistry
Delta Mass database (www.abrf.org/index.cfm/dm.home) or the Unimod database (www.unimod.org). To the best of our knowledge, this is the first report of a xylosone modification of a recombinant monoclonal antibody. Cell Culture Media Additives. Cell culture scientists are under constant pressure to increase titers, enhance cell culture performance and improve product quality. The practice of modifying the cell culture medium is the predominant tool used to address these goals.10,49−54 However, it is difficult to predict whether a specific additive will address a specific biochemical need of the host cell with the desired outcome or whether these good intentions will go awry (see Table S-2 in the Supporting Information). Occasionally, a cell culture additive may have negative implications on product quality.55−58 In our study, a recombinant monoclonal antibody (adalimumab) exhibited a difference in product quality following a change to the cell culture conditions. Our initial observations were confounded by the fact that ascorbate unexpectedly degraded quite rapidly to xylosone that in turn exhibited reactivity with susceptible primary amines. It has been well established that increasing glucose levels in a cell culture will increase the potential for glycation to occur through well understood chemistry.25,26 However, the discovery that ascorbate supplementation to the cell culture of a recombinant monoclonal antibody induced a novel glycation-like modification by xylosone was quite surprising and is a reminder that product quality is another parameter that must be considered when making changes to the cell culture feeding strategies. Futhermore, it is worth noting that xylosone almost certainly modifies a myriad of proteins of the host cells, thereby directly affecting a broad range of biological activities ultimately impacting the culture viability. Employing strategies where scavengers are added to the media in order to protect the protein drug or limiting the addition of ascorbic acid to the cell culture may minimize the propensity of modifications by xylosone during cell culture.28
Figure 1. (A) Structures of the four different isotopic isoforms of ascorbate used to probe the structure of the +130 Da and +148 Da adducts. The red dot denotes carbon-13 labeling at a given position. (B) Mass spectra of the doubly charged modified light chain Nterminal trypic peptide of Antibody A incubated with 3 mg/mL of ascorbate. The pattern of mass shift indicates that the carbon atom at 1 position in ascorbate is cleaved off in the final adducts, consistent with the proposed mechanism of xylosone being the reactive intermediate (see Scheme 1).
■
CONCLUSIONS Cell culture supplementation with ascorbic acid caused an unexpected change to the product quality of a recombinant monoclonal antibody; therefore, the use of ascorbic acid as a supplement should be taken with caution. The recombinant monoclonal antibody was modified by xylosone, a highly reactive species generated as an oxidative degradation product of ascorbate but not directly introduced into the culture. In addition, xylosone almost certainly modifies a myriad of proteins of the host cells, thereby directly affecting a broad range of biological activities which may impact the culture viability. With the advancement in protein mass spectrometry and the increasing awareness of issues highlighted in this paper, similar modifications and mechanisms will likely be revealed in other systems and proteins. Altogether, better understanding and critical consideration of the latent reactivities of any addictives, and particularly, deleterious consequences, would be prudent to ensure the quality of protein products.
(see Schemes S-1 and S-2 in the Supporting Information), and each may lead to the hemiaminal and Schiff base forms described above. Furthermore, the Schiff bases can undergo further cyclization as well (Scheme S-2 in the Supporting Information). For this antibody, no modification of arginine was observed; at first, it was somewhat surprising to us as both xylosone and methylglyoxal contain two adjacent carbonyl groups. However, upon closer inspection, as shown in Scheme S-1 in the Supporting Information, several hydroxyl groups in xylosone can readily form stable cyclic hemiacetal or hemiketal with one of the carbonyl group, thereby leaving only one reactive carbonyl group for further reactions with amines that is similar to that of glycation of amines. Of course, protein structures and local environments can certainly affect the stability of the products described above.23 For example, for some other proteins, modification of arginine by xylosone was reported.48 Ascorbate oxidative degradants have been reported as a source of chemical modifications in human eye lens and were shown to modify lysine and arginine residues in model systems.37,38 However, neither a +148 Da or +130 Da mass deviation nor are listed as a xylosone modification in the ABRF
■
ASSOCIATED CONTENT
S Supporting Information *
Materials and additional experimental details, chromatograms, mass spectra, table of peptides identifies with modifications by xylosone, extracted ion chromatograms, tandem mass spectra, the relative susceptibilities of representative peptides modified by xylosone, scheme of possible reaction products between xylosone and a protein primary amine, reaction scheme of 7532
DOI: 10.1021/acs.analchem.5b00801 Anal. Chem. 2015, 87, 7529−7534
Technical Note
Analytical Chemistry
(14) Tachibana, H.; Taniguchi, K.; Ushio, Y.; Teruya, K.; Osada, K.; Murakami, H. Cytotechnology 1994, 16, 151−157. (15) Pinnell, S. R. Yale J. Biol. Med. 1985, 58, 553−559. (16) Lee, S.; Pagoria, D.; Raigrodski, A.; Geurtsen, W. J. Biomed. Mater. Res., Part B 2007, 83B, 391−399. (17) Fukumura, H.; Sato, M.; Kezuka, K.; Sato, I.; Feng, X.; Okumura, S.; Fujita, T.; Yokoyama, U.; Eguchi, H.; Ishikawa, Y.; Saito, T. J. Physiol. Sci. 2012, 62, 251−257. (18) Du, J.; Cullen, J. J.; Buettner, G. R. Biochim. Biophys. Acta, Rev. Cancer 2012, 1826, 443−457. (19) Levine, M.; Morita, K.; Heldman, E.; Pollard, H. B. J. Biol. Chem. 1985, 260, 15598−15603. (20) Xing, Z.; Kenty, B. M.; Li, Z. J.; Lee, S. S. Biotechnol. Bioeng. 2009, 103, 733−746. (21) Kang, X.; Li, N.; Wang, M.; Boontheung, P.; Sioutas, C.; Harkema, J. R.; Bramble, L. A.; Nel, A. E.; Loo, J. A. Proteomics 2010, 10, 520−531. (22) Murphy, M. P. Biochem. J. 2009, 417, 1−13. (23) Zhao, G.; Zhou, Z. S. Bioorg. Med. Chem. Lett. 2001, 11, 2331− 2335. (24) Biastoff, S.; Teuber, M.; Zhou, Z. S.; Dräger, B. Planta Med. 2006, 72, 1136−1141. (25) Quan, C.; Alcala, E.; Petkovska, I.; Matthews, D.; Canova-Davis, E.; Taticek, R.; Ma, S. Anal. Biochem. 2008, 373, 179−191. (26) Zhang, B.; Yang, Y.; Yuk, I.; Pai, R.; McKay, P.; Eigenbrot, C.; Dennis, M.; Katta, V.; Francissen, K. C. Anal. Chem. 2008, 80, 2379− 2390. (27) Yuk, I. H.; Zhang, B.; Yang, Y.; Dutina, G.; Leach, K. D.; Vijayasankaran, N.; Shen, A. Y.; Andersen, D. C.; Snedecor, B. R.; Joly, J. C. Biotechnol. Bioeng. 2011, 108, 2600−10. (28) Chumsae, C.; Gifford, K.; Lian, W.; Liu, H.; Radziejewski, C. H.; Zhou, Z. S. Anal. Chem. 2013, 85, 11401−11409. (29) Chaplen, F. W. R.; Fahl, W. E.; Cameron, D. C. Proc. Natl. Acad. Sci. U. S. A. 1998, 95, 5533−5538. (30) Williams, W. M.; Weinberg, A.; Smith, M. A. J. Amino Acids 2011, 2011, 461216. (31) Gadgil, H. S.; Bondarenko, P. V.; Pipes, G. D.; Dillon, T. M.; Banks, D.; Abel, J.; Kleemann, G. R.; Treuheit, M. J. Anal. Biochem. 2006, 355, 165−174. (32) Adachi, T.; Schöneich, C.; Cohen, R. A. Drug Discovery Today: Dis. Mech. 2005, 2, 39−46. (33) Sun, R.; Eriksson, S.; Wang, L. J. Biol. Chem. 2012, 287, 24304− 24312. (34) Zang, T.; Dai, S.; Chen, D.; Lee, B. W. K.; Liu, S.; Karger, B. L.; Zhou, Z. S. Anal. Chem. 2009, 81, 9065−9071. (35) Wang, Z.; Rejtar, T.; Zhou, Z. S.; Karger, B. L. Rapid Commun. Mass Spectrom. 2010, 24, 267−275. (36) Zhou, Z. S.; Smith, A. E.; Matthews, R. G. Bioorg. Med. Chem. Lett. 2000, 10, 2471−2475. (37) Nemet, I.; Monnier, V. M. J. Biol. Chem. 2011, 286, 37128− 37136. (38) Linetsky, M.; Shipova, E.; Cheng, R.; Ortwerth, B. J. Biochim. Biophys. Acta, Mol. Basis Dis. 2008, 1782, 22−34. (39) Chumsae, C.; Zhou, L. L.; Shen, Y.; Wohlgemuth, J.; Fung, E.; Burton, R.; Radziejewski, C.; Zhou, Z. S. Anal. Chem. 2014, 86, 8932− 8936. (40) Gilmore, J. M.; Scheck, R. A.; Esser-Kahn, A. P.; Joshi, N. S.; Francis, M. B. Angew. Chem., Int. Ed. 2006, 45, 5307−5311. (41) Witus, L. S.; Moore, T.; Thuronyi, B. W.; Esser-Kahn, A. P.; Scheck, R. A.; Iavarone, A. T.; Francis, M. B. J. Am. Chem. Soc. 2010, 132, 16812−16817. (42) Zhou, Z. S.; Flohr, A.; Hilvert, D. J. Org. Chem. 1999, 64, 8334− 8341. (43) Zhou, Z. S.; Jiang, N.; Hilvert, D. J. Am. Chem. Soc. 1997, 119, 3623−3624. (44) Gui, S.; Wooderchak-Donahue, W. L.; Zang, T.; Chen, D.; Daly, M. P.; Zhou, Z. S.; Hevel, J. M. Biochemistry 2013, 52, 199−209. (45) Zhang, J.; Zhang, T.; Jiang, L.; Hewitt, D.; Huang, Y.; Kao, Y.H.; Katta, V. Anal. Chem. 2012, 84, 2313−2320.
acyclic xylosone with a protein primary amine, list of relevant cell culture media compounds, and additional references. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b00801.
■
AUTHOR INFORMATION
Corresponding Authors
*E-mail: [email protected] *E-mail: [email protected] *E-mail: [email protected] Present Address ∥
Y.Z.: BioMarin, Novato, CA 94949-8703.
Downloaded by UNIV LAVAL on September 15, 2015 | http://pubs.acs.org Publication Date (Web): July 21, 2015 | doi: 10.1021/acs.analchem.5b00801
Notes
AbbVie Disclosure: The design, study conduct, and financial support for the study were provided by AbbVie (formerly the proprietary pharmaceutical division of Abbott Laboratories, now an independent biopharmaceutial company). AbbVie participated in the interpretation of data, review, and approval of the publication; all authors contributed to the development of the publication and maintained control over the final content. C.C., P.H., S.M., C. Rancicot, and C. Radziejewski are AbbVie employees. H.R. and Y.Z. are former AbbVie employees. Z.S.Z. serves as the doctoral advisor to C.C. and has received AbbVie support for this role. The authors declare no competing financial interest. ⊥ A portion of this work was conducted by C.C. as a Ph.D. student at the Department of Chemistry and Chemical Biology, Northeastern University, Boston, Massachusetts, USA.
■
ACKNOWLEDGMENTS The authors thank Dr. David Lee and Shanshan Liu for their critical review of the manuscript and Shanshan Liu for her help on the chemical scheme. This activity is partially supported by a grant from NIH NIGMS (Grant 1R01GM101396 to Z.S.Z.). This is contribution number 1056 from the Barnett Institute.
■
REFERENCES
(1) Zaia, J. Mass Spectrom. Rev. 2004, 23, 161−227. (2) Fenselau, C.; Vestling, M. M.; Cotter, R. J. Curr. Opin. Biotechnol. 1993, 4, 14−19. (3) Awdeh, Z. L.; Williamson, A. R.; Askonas, B. A. Biochem. J. 1970, 116, 241−248. (4) Liu, H.; Gaza-Bulseco, G.; Faldu, D.; Chumsae, C.; Sun, J. J. Pharm. Sci. 2008, 97, 2426−2447. (5) Manning, M.; Chou, D.; Murphy, B.; Payne, R.; Katayama, D. Pharm. Res. 2010, 27, 544−575. (6) Vlasak, J.; Ionescu, R. Curr. Pharm. Biotechnol. 2008, 9, 468−481. (7) van der Valk, J.; Brunner, D.; De Smet, K.; Fex Svenningsen, Å.; Honegger, P.; Knudsen, L. E.; Lindl, T.; Noraberg, J.; Price, A.; Scarino, M. L.; Gstraunthaler, G. Toxicol. In Vitro 2010, 24, 1053− 1063. (8) Wurm, F. M. Nat. Biotechnol. 2004, 22, 1393−1398. (9) Wong, D. C. F.; Wong, K. T. M.; Goh, L. T.; Heng, C. K.; Yap, M. G. S. Biotechnol. Bioeng. 2005, 89, 164−177. (10) Gramer, M. J.; Eckblad, J. J.; Donahue, R.; Brown, J.; Shultz, C.; Vickerman, K.; Priem, P.; van den Bremer, E. T.; Gerritsen, J.; van Berkel, P. H. Biotechnol. Bioeng. 2011, 108, 1591−602. (11) Jing, Y.; Borys, M.; Nayak, S.; Egan, S.; Qian, Y.; Pan, S.-H.; Li, Z. J. Process Biochem. 2012, 47, 69−75. (12) Crowell, C. K.; Grampp, G. E.; Rogers, G. N.; Miller, J.; Scheinman, R. I. Biotechnol. Bioeng. 2007, 96, 538−549. (13) Hossler, P.; McDermott, S.; Racicot, C.; Chumsae, C.; Raharimampionona, H.; Zhou, Y.; Ouellette, D.; Matuck, J.; Correia, I.; Fann, J.; Li, J. Biotechnol. Prog. 2014, 30, 1419−1431. 7533
DOI: 10.1021/acs.analchem.5b00801 Anal. Chem. 2015, 87, 7529−7534
Technical Note
Downloaded by UNIV LAVAL on September 15, 2015 | http://pubs.acs.org Publication Date (Web): July 21, 2015 | doi: 10.1021/acs.analchem.5b00801
Analytical Chemistry (46) Liu, M.; Zhang, Z.; Zang, T.; Spahr, C.; Cheetham, J.; Ren, D.; Zhou, Z. S. Anal. Chem. 2013, 85, 5900−5908. (47) Wan, W.; Zhao, G.; Al-Saad, K.; Siems, W. F.; Zhou, Z. S. Rapid Commun. Mass Spectrom. 2004, 18, 319−324. (48) Lee, K.-W.; Simpson, G.; Ortwerth, B. Biochim. Biophys. Acta, Mol. Basis Dis. 1999, 1453, 141−151. (49) Baker, K. N.; Rendall, M. H.; Hills, A. E.; Hoare, M.; Freedman, R. B.; James, D. C. Biotechnol. Bioeng. 2001, 73, 188−202. (50) Gu, X.; Wang, D. I. Biotechnol. Bioeng. 1998, 58, 642−8. (51) Qian, Y.; Jing, Y.; Li, Z. J. Biotechnol. Prog. 2010, 26, 1417−23. (52) Rodriguez, J.; Spearman, M.; Huzel, N.; Butler, M. Biotechnol. Prog. 2005, 21, 22−30. (53) Hossler, P.; Racicot, C.; McDermott, S. J. Glycobiol. 2014, 3, 108. (54) Chaderjian, W. B.; Chin, E. T.; Harris, R. J.; Etcheverry, T. M. Biotechnol. Prog. 2005, 21, 550−3. (55) Banks, D. D.; Hambly, D. M.; Scavezze, J. L.; Siska, C. C.; Stackhouse, N. L.; Gadgil, H. S. J. Pharm. Sci. 2009, 98, 4501−10. (56) Fischer, S.; Hoernschemeyer, J.; Mahler, H. C. Eur. J. Pharm. Biopharm. 2008, 70, 42−50. (57) Kim, K.; Rhee, S. G.; Stadtman, E. R. J. Biol. Chem. 1985, 260, 15394−15397. (58) Richheimer, A.; Robinson, A. B. J. Orthomol. Psychiatry 1977, 6, 290−299.
7534
DOI: 10.1021/acs.analchem.5b00801 Anal. Chem. 2015, 87, 7529−7534
When Good Intentions Go Awry: Modification of a Recombinant Monoclonal Antibody in Chemically Defined Cell Culture by Xylosone, an Oxidative Product of Ascorbic Acid Chris Chumsae,*,†,§,⊥ Patrick Hossler,‡ Haly Raharimampionona,† Yu Zhou,†,∥ Sean McDermott,‡ Chris Racicot,‡ Czeslaw Radziejewski,*,† and Zhaohui Sunny Zhou*,§ †
Protein Analytics, Process Sciences Department, AbbVie Bioresearch Center, Worcester, Massachusetts 01605, United States Cell Culture, Process Sciences Department, AbbVie Bioresearch Center, Worcester, Massachusetts 01605, United States § Barnett Institute of Chemical and Biological Analysis, Department of Chemistry and Chemical Biology, Northeastern University, Boston, Massachusetts 02115, United States ‡
Downloaded by UNIV LAVAL on September 15, 2015 | http://pubs.acs.org Publication Date (Web): July 21, 2015 | doi: 10.1021/acs.analchem.5b00801
S Supporting Information *
ABSTRACT: With the advent of new initiatives to develop chemically defined media, cell culture scientists screen many additives to improve cell growth and productivity. However, the introduction or increase of supplements, typically considered beneficial or protective on their own, to the basal media or feed stream may cause unexpected detrimental consequences to product quality. For instance, because cultured cells are constantly under oxidative stress, ascorbic acid (vitamin C, a potent natural reducing agent) is a common additive to cell culture media. However, as reported herein, a recombinant monoclonal antibody (adalimumab) in cell culture was covalently modified by xylosone (molecular weight 148), an oxidative product of ascorbate. Containing reactive carbonyl groups, xylosone modifies various amines (e.g., the N-termini of the heavy and light chains and susceptible lysines), forming either hemiaminal (+148 Da) or Schiff base (imine, +130 Da) products. Our findings show, for the first time, that ascorbate-derived xylosone can contribute to an increase in molecular heterogeneity, such as acidic species. Our work serves as a reminder that additives to cell culture and their metabolites may become reactive and negatively impact the overall product quality and should be carefully monitored with any changes in cell culture conditions.
U
Pertinent to this study, media components or additives have been shown to affect product quality and protein modifications.23,24 The best known example is arguably glycation of proteins by glucose,25−27 an essential nutrient as the main energy source of cultured cells. While unavoidable, glycation nevertheless is predictable. On the other hand, it is challenging to predict and detect modifications by secondary metabolites or byproducts. For example, we previously reported that a recombinant monoclonal antibody was unexpectedly modified by an accumulation of methylglyoxal (MGO) during cell culture due to a change in media that was perceived as beneficial.28 Methylglyoxal is a dicarbonyl compound that is generated as a byproduct of glycolysis.28,29 Additionally, changes in the cell culture conditions, specifically the redox state, ultimately lead to increased amounts of this modification.30 Modifications from other reactive species include cysteinylation,31 glutathionylation,32,33 and N-homocysteintylation from homocysteine thiolactone.34 Therefore, changes to
ntil recently, cell culture scientists mostly focused on cell growth and protein expression level, but as demonstrated herein, the quality of the final products has recently emerged as another critical attribute to be considered. To this point, variations in product quality, or microheterogeniety, are mostly attributed to a myriad of post-translational modifications (PTMs).1−6 In order to reduce variability in product quality, cell growth, and expression levels, the current trend in recombinant monoclonal antibody production has been to move from complex undefined hydrolysate media to chemically defined media.7,8 Furthermore, it has been shown that modulation of the supplemental feed can impact the product quality of the protein drug.9−14 A common additive is the much-storied vitamin C (ascorbic acid). In addition to its function as a cofactor for collagen synthesis,15 it has been implicated as an antioxidant (biological reductant) and potent scavenger of reactive oxygen species (ROS).16−19 Large scale production of recombinant monoclonal antibodies require higher levels of oxygen to maintain higher cell densities.20 Such conditions may generate reactive oxygen species.21,22 Therefore, including antioxidants such as ascorbic acid is generally considered protective and desirable. © 2015 American Chemical Society
Received: February 16, 2015 Accepted: July 7, 2015 Published: July 7, 2015 7529
DOI: 10.1021/acs.analchem.5b00801 Anal. Chem. 2015, 87, 7529−7534
Technical Note
Analytical Chemistry
Additionally, the formation of xylosone from ascorbic acid and the subsequent modification to the antibody was confirmed using stable isotope labeled ascorbic acid. Details are described in the Supporting Information.
Downloaded by UNIV LAVAL on September 15, 2015 | http://pubs.acs.org Publication Date (Web): July 21, 2015 | doi: 10.1021/acs.analchem.5b00801
the cell culture medium may result in unexpected changes in product quality.35,36 As reported herein, we observed an increase in acidic species in a recombinant monoclonal antibody supplemented with ascorbic acid during cell culture. Additionally, we observed two unidentified masses in the reduced liquid chromatography/ mass spectrometry (LC/MS) analysis of acidic fractions from the weak cation exchange chromatography which exhibited molecular weight increases of 130 and 148, respectively. Detailed analyses revealed that these modifications occurred on the primary amines of the N-termini of heavy and light chains and also susceptible lysine residues. This was confirmed by in vitro incubation of native antibody with increasing concentrations of ascorbic acid. Given that the molecular weight of ascorbate is 176, it was hypothesized that metabolites or degration produts of this nutrient were the culprits. This was confirmed in more detailed mechanistic investigations using 13 C labeled ascorbate. As illustrated in Scheme 1, ascorbate is
■
RESULTS AND DISCUSSION Ascorbate Supplement in Cell Culture Induced New Acidic Variants. A recombinant monoclonal antibody was expressed in shake flasks using various supplementation and additives. Of specific interest were cell culture conditions under which the concentration of ascorbate was increased (0, 0.1, 1, and 3 mg/mL, respectively). When antibodies were analyzed by weak cation exchange chromatography, the formation of new acidic species directly correlated with ascorbate concentration (Figure S-1 in the Supporting Information). This observation led us to initiate a detailed structural analysis of the recombinant monoclonal antibody in order to determine the chemical nature and cause of the increase in charge heterogeneity. LC/MS Analysis of Reduced Antibody. Without separating antibody variants, reduced LC/MS analysis of the recombinant monoclonal antibody heavy chain and light chain from the ascorbate supplemented cultures did not show obvious differences in the mass spectra (Figure S-2 in the Supporting Information). Therefore, fractionation of the earlier eluting peaks and the Lys-0 peak from weak cation exchange (Figure S-1 in the Supporting Information) was employed. The analysis of the Lys-0 fraction by reduced LC/MS revealed a deconvoluted light chain mass spectra which was in good agreement with the expected light chain mass (Figure S-3A in the Supporting Information). Analysis of the mass spectra resulting from Peak A, however, exhibited discernible differences as shown in Figure S-3B in the Supporting Information. The deconvoluted light chain mass spectra showed the expected mass as well as several lower abundance peaks with higher molecular-weight. Two of the masses corresponded to mass increases of 148 and 130 Da. The levels of modification observed by the UV absorption (WCX) and mass spectrometry appear different. The difference may be attributed to several possibilities. First, Figure S-1 was based on UV absorption (WCX) that depends on the chromophores in each antibody variant, which were likely to be the same or similar; in comparison, Figure S-3 in the Supporting Information was based on ionization efficiencies of each species, which was likely to differ considerably given the noticeable changes in charge characteristics as evident from the large gap in elution time. Second, Figure S-3 in the Supporting Information accounts for the modification at the light chain only. The modification was also observed on heavy chain. In addition, each antibody has two light chains and two heavy chains, so the overall modification of the whole antibody should be more than individual light chain or heavy chain. In order to elucidate the nature of the additional species, a pure fraction of the Lys-0 species was treated with ascorbic acid in vitro. The deconvoluted spectrum (Figure S-3C in the Supporting Information) of the in vitro sample was highly similar to that from cell culture (Figure S-3B in the Supporting Information). First, the observed molecular weight of 23 408.5 represented the unmodified light chain from Lys-0 incubated with ascorbate and was in good agreement with the theoretical value of 23 408.1. Second, two other lower intensity peaks were also observed with masses of 23 538.3 and 23 556.6 Da which corresponded to mass increases of 130 and 148 Da, respectively
Scheme 1. Degradation of Ascorbic Acid to Xylosone and Subsequent Reaction with a Primary Amine to Form a Hemiaminal and Schiff Basea
a
The red dot denotes the carbon which is released as CO2.
first oxidized to dehydroascorbic acid. Subsequent decarboxylation generates xylosone.37,38 Xylosone is highly reactive and is capable of modifying susceptible primary amines, resulting in mass increases of 148 and 130 Da for the hemiaminal and Schiff base, respectively (Scheme 1). To the best of our knowledge, this is the first report of an ascorbate-originated xylosone modification of a recombinant monoclonal antibody in vitro and in cell culture.
■
MATERIALS AND METHODS Similar to our previous studies of antibody modifications by citric acid and methylglyoxal,28,39 the antibody variants were separated and detected by weak cation exchange chromatography (WCX), the molecular weights of the light and heavy chains from the reduced antibody were determined by mass spectrometry, and site of modifications were determined by peptide mapping (tryptic digestion followed by LC/MS). 7530
DOI: 10.1021/acs.analchem.5b00801 Anal. Chem. 2015, 87, 7529−7534
Technical Note
Downloaded by UNIV LAVAL on September 15, 2015 | http://pubs.acs.org Publication Date (Web): July 21, 2015 | doi: 10.1021/acs.analchem.5b00801
Analytical Chemistry
3.0 mg/mL) over time (3, 5, 11, and 14 days) confirmed our findings as shown in Figures S-12 and S-13 in the Supporting Information. Elucidation of the Modification Agent As Xylosone. The mass of ascorbate is 176 Da, which is 28 Da greater than the +148 modification observed, so it was unlikely that ascorbate itself modified the antibody but rather its degradation product(s) was the likely culprit(s). Another hint is the 18 Da difference (130 vs 148 Da) between the two observed modifications, suggesting a dehydration (elimination of a water molecule) following the initial reaction; this is reminiscent of modifications of lysine or arginine residues by methylglyoxal (MGO) or other carbonyl-containing molecules. All together, we postulated that the bona fide reactive species should be degraded from ascorbate and also contain reactive carbonyl group(s). Xylosone thus emerged as a likely candidate as it has been reported as a degradation product of ascorbate (see Scheme 1 for its formation pathway).37 Furthermore, xylosone has a mass of 148 Da and two carbonyl groups. The 130 Da adduct may be due to the loss of a water molecule following the initial addition reaction between the carbonyl group of xylosone and the amines in the protein (hemiaminal to Schiff base as shown in Scheme 1). Incubation of Antibody with 13C Regio-Labeled Ascorbate. Isotopic labeling and tracing often provides detailed mechanistic insights.45−47 In order to confirm that ascorbate was degraded to xylosone (losing the carbon atom at 1 position), which in turn modified the antibody, we used ascorbate regio-specifically labeled with13C (at C1, C2 or C3 carbon, respectively, as shown in Figure 1A). Following incubation with unlabeled ascorbate or one of the three regio-labeled ascorbate molecules, the antibody was analyzed by tryptic peptide mapping with mass spectrometry detection. The results (Figure 1B) lended conclusive evidence supporting the mechanism in Scheme 1: labeling at the C1 position (13C or 12 C) produced the same mass spectra; in contrast, ascorbate with 13C labeling at C2 or C3 shifted the peaks to 1 Da higher compared to 12C ascorbate (+149 vs +148 and 131 vs 130, respectively). In addition, MS/MS localized the modification to the light chain N-terminal residue including the heavy label for the C2 and C3 regio-labeled ascorbates as shown in Figure S-15 in the Supporting Information. Thus, the data have verified that the C1 carbon in ascorbate is lost but neither C2 nor C3. Therefore, our data confirmed that ascorbate was converted to xylosone with the concomitant loss of C1 carbon atom as illustrated in Scheme 1. Chemical Nature of the Modifications. When a carbonyl reacts with an amine, two products may form: the initial addition reaction leads to a hemiaminal (see Scheme 1) with a mass equals to the total of the masses of the two reactants (amine and carbonyl; for xylosone, 148 Da); a subsequent elimination of a water molecule (18 Da) leads to a Schiff base (imine, see Scheme 1) with a mass that is 18 Da less than the hemiaminal (148 − 18 = 130 Da). Hence the masses of the two adducts match with the observed masses. The underlying chemistry, including stereo-, regio-, and positional isomers of the adducts, also explains the multiple isobaric peaks observed (see Figures S-5 and S-7 in the Supporting Information). For instance, the formation of the hemiaminal can result in two stereoisomers (the chiral center is denoted with an ∗ in Scheme 1). Two other factors further complicate the situation: first, two carbonyls exist in xylosone; and second, xylosone may exist in both cyclic and acyclic forms
(Figure S-3C in the Supporting Information). Furthermore, these two masses are in good agreement with the two masses representing +130 Da and +148 Da from cell culture suggesting that they were generated due to the ascorbic acid supplementation. Tryptic Mapping and LC/MS/MS Detection. Tryptic mapping was performed on the recombinant monoclonal antibody produced in cell culture supplemented with up to 3 mg/mL of ascorbate and on the 0 Lys isoform incubated with ascorbate in vitro. It is important to note that the mass difference between these two species is 18 Da; therefore, they may be related and involve the loss of a water molecule, reminiscing modifications of lysine or arginine residues by methylglyoxal (MGO) or other carbonyl-containing molecules. Therefore, we searched for miscleaved trypic peptides and found none with an internal arginine. Then, we examined potential modifications of primary amines, specifically the Ntermini and miscleaved lysine containing peptides, which did indeed produce several possible sites of modification as shown in Table S-1 in the Supporting Information. Specifically, the light chain N-terminal peptide had two chromatographically resolved +148 Da peaks with close elution to the native peptide and a +130 peak that eluted later as shown in Figure S-5 in the Supporting Information. The MS/MS spectra were analyzed for the native and modified +148 Da and +130 Da peptides that all had y ion series which were in good agreement with the predicted amino acid sequence and covered the entire peptide except the first two N-terminal residues as shown in Figure S-6 in the Supporting Information. The b ion series was limited but had strong signal for the first three residues at the N-termini. In addition, a strong signal for the native and modified a1 and a2 ions was also present. Altogether, our data conclusively localized both modifications (+148 Da and +130 Da) to the N-terminal amine of the light chain (Asp1). The N-terminal peptides of the heavy chain also exhibited a parent peak in the native extracted mass chromatogram, two isobaric peaks in the extracted mass chromatogram corresponding to a mass increase of +148 Da and two isobaric peaks in the extracted mass chromatograms corresponding to a mass increases of +130 Da (Figure S-7 in the Supporting Information). Once again, the b ion series was used to definitively assign the +148 Da and +130 Da modification to the N-termini of the two sets of resolved modified peptides (see Figure S-8 in the Supporting Information). The analysis of other peptides from the recombinant monoclonal antibody showed the modification of lysine residues (see Figures S-9 and S-10 in the Supporting Information). However, lysine residues exhibited lower susceptibility to modification as compared to the N-terminal primary amines as shown in the analysis of the in vitro ascorbate incubation over time and at increasing concentrations of ascorbate (Figure S-11 in the Supporting Information). The observation is in good agreement with the lower pKa of the Nterminal amines (∼8) as compared to the pKa of the lysyl amine on the side chain (∼10);40,41 and of course, these modifications are also likely to be influenced by other factors such as the antibody structure and microenvironment.23,28,34,42−44 In addition, it is important to note that only a +148 Da species was seen for all modified lysine residues in the antibody, again suggesting local environment is likely to affect the nature and distribution of various chemical forms, as discussed in greater details later. Lastly, in vitro (0.1, 1.0, and 7531
DOI: 10.1021/acs.analchem.5b00801 Anal. Chem. 2015, 87, 7529−7534
Technical Note
Downloaded by UNIV LAVAL on September 15, 2015 | http://pubs.acs.org Publication Date (Web): July 21, 2015 | doi: 10.1021/acs.analchem.5b00801
Analytical Chemistry
Delta Mass database (www.abrf.org/index.cfm/dm.home) or the Unimod database (www.unimod.org). To the best of our knowledge, this is the first report of a xylosone modification of a recombinant monoclonal antibody. Cell Culture Media Additives. Cell culture scientists are under constant pressure to increase titers, enhance cell culture performance and improve product quality. The practice of modifying the cell culture medium is the predominant tool used to address these goals.10,49−54 However, it is difficult to predict whether a specific additive will address a specific biochemical need of the host cell with the desired outcome or whether these good intentions will go awry (see Table S-2 in the Supporting Information). Occasionally, a cell culture additive may have negative implications on product quality.55−58 In our study, a recombinant monoclonal antibody (adalimumab) exhibited a difference in product quality following a change to the cell culture conditions. Our initial observations were confounded by the fact that ascorbate unexpectedly degraded quite rapidly to xylosone that in turn exhibited reactivity with susceptible primary amines. It has been well established that increasing glucose levels in a cell culture will increase the potential for glycation to occur through well understood chemistry.25,26 However, the discovery that ascorbate supplementation to the cell culture of a recombinant monoclonal antibody induced a novel glycation-like modification by xylosone was quite surprising and is a reminder that product quality is another parameter that must be considered when making changes to the cell culture feeding strategies. Futhermore, it is worth noting that xylosone almost certainly modifies a myriad of proteins of the host cells, thereby directly affecting a broad range of biological activities ultimately impacting the culture viability. Employing strategies where scavengers are added to the media in order to protect the protein drug or limiting the addition of ascorbic acid to the cell culture may minimize the propensity of modifications by xylosone during cell culture.28
Figure 1. (A) Structures of the four different isotopic isoforms of ascorbate used to probe the structure of the +130 Da and +148 Da adducts. The red dot denotes carbon-13 labeling at a given position. (B) Mass spectra of the doubly charged modified light chain Nterminal trypic peptide of Antibody A incubated with 3 mg/mL of ascorbate. The pattern of mass shift indicates that the carbon atom at 1 position in ascorbate is cleaved off in the final adducts, consistent with the proposed mechanism of xylosone being the reactive intermediate (see Scheme 1).
■
CONCLUSIONS Cell culture supplementation with ascorbic acid caused an unexpected change to the product quality of a recombinant monoclonal antibody; therefore, the use of ascorbic acid as a supplement should be taken with caution. The recombinant monoclonal antibody was modified by xylosone, a highly reactive species generated as an oxidative degradation product of ascorbate but not directly introduced into the culture. In addition, xylosone almost certainly modifies a myriad of proteins of the host cells, thereby directly affecting a broad range of biological activities which may impact the culture viability. With the advancement in protein mass spectrometry and the increasing awareness of issues highlighted in this paper, similar modifications and mechanisms will likely be revealed in other systems and proteins. Altogether, better understanding and critical consideration of the latent reactivities of any addictives, and particularly, deleterious consequences, would be prudent to ensure the quality of protein products.
(see Schemes S-1 and S-2 in the Supporting Information), and each may lead to the hemiaminal and Schiff base forms described above. Furthermore, the Schiff bases can undergo further cyclization as well (Scheme S-2 in the Supporting Information). For this antibody, no modification of arginine was observed; at first, it was somewhat surprising to us as both xylosone and methylglyoxal contain two adjacent carbonyl groups. However, upon closer inspection, as shown in Scheme S-1 in the Supporting Information, several hydroxyl groups in xylosone can readily form stable cyclic hemiacetal or hemiketal with one of the carbonyl group, thereby leaving only one reactive carbonyl group for further reactions with amines that is similar to that of glycation of amines. Of course, protein structures and local environments can certainly affect the stability of the products described above.23 For example, for some other proteins, modification of arginine by xylosone was reported.48 Ascorbate oxidative degradants have been reported as a source of chemical modifications in human eye lens and were shown to modify lysine and arginine residues in model systems.37,38 However, neither a +148 Da or +130 Da mass deviation nor are listed as a xylosone modification in the ABRF
■
ASSOCIATED CONTENT
S Supporting Information *
Materials and additional experimental details, chromatograms, mass spectra, table of peptides identifies with modifications by xylosone, extracted ion chromatograms, tandem mass spectra, the relative susceptibilities of representative peptides modified by xylosone, scheme of possible reaction products between xylosone and a protein primary amine, reaction scheme of 7532
DOI: 10.1021/acs.analchem.5b00801 Anal. Chem. 2015, 87, 7529−7534
Technical Note
Analytical Chemistry
(14) Tachibana, H.; Taniguchi, K.; Ushio, Y.; Teruya, K.; Osada, K.; Murakami, H. Cytotechnology 1994, 16, 151−157. (15) Pinnell, S. R. Yale J. Biol. Med. 1985, 58, 553−559. (16) Lee, S.; Pagoria, D.; Raigrodski, A.; Geurtsen, W. J. Biomed. Mater. Res., Part B 2007, 83B, 391−399. (17) Fukumura, H.; Sato, M.; Kezuka, K.; Sato, I.; Feng, X.; Okumura, S.; Fujita, T.; Yokoyama, U.; Eguchi, H.; Ishikawa, Y.; Saito, T. J. Physiol. Sci. 2012, 62, 251−257. (18) Du, J.; Cullen, J. J.; Buettner, G. R. Biochim. Biophys. Acta, Rev. Cancer 2012, 1826, 443−457. (19) Levine, M.; Morita, K.; Heldman, E.; Pollard, H. B. J. Biol. Chem. 1985, 260, 15598−15603. (20) Xing, Z.; Kenty, B. M.; Li, Z. J.; Lee, S. S. Biotechnol. Bioeng. 2009, 103, 733−746. (21) Kang, X.; Li, N.; Wang, M.; Boontheung, P.; Sioutas, C.; Harkema, J. R.; Bramble, L. A.; Nel, A. E.; Loo, J. A. Proteomics 2010, 10, 520−531. (22) Murphy, M. P. Biochem. J. 2009, 417, 1−13. (23) Zhao, G.; Zhou, Z. S. Bioorg. Med. Chem. Lett. 2001, 11, 2331− 2335. (24) Biastoff, S.; Teuber, M.; Zhou, Z. S.; Dräger, B. Planta Med. 2006, 72, 1136−1141. (25) Quan, C.; Alcala, E.; Petkovska, I.; Matthews, D.; Canova-Davis, E.; Taticek, R.; Ma, S. Anal. Biochem. 2008, 373, 179−191. (26) Zhang, B.; Yang, Y.; Yuk, I.; Pai, R.; McKay, P.; Eigenbrot, C.; Dennis, M.; Katta, V.; Francissen, K. C. Anal. Chem. 2008, 80, 2379− 2390. (27) Yuk, I. H.; Zhang, B.; Yang, Y.; Dutina, G.; Leach, K. D.; Vijayasankaran, N.; Shen, A. Y.; Andersen, D. C.; Snedecor, B. R.; Joly, J. C. Biotechnol. Bioeng. 2011, 108, 2600−10. (28) Chumsae, C.; Gifford, K.; Lian, W.; Liu, H.; Radziejewski, C. H.; Zhou, Z. S. Anal. Chem. 2013, 85, 11401−11409. (29) Chaplen, F. W. R.; Fahl, W. E.; Cameron, D. C. Proc. Natl. Acad. Sci. U. S. A. 1998, 95, 5533−5538. (30) Williams, W. M.; Weinberg, A.; Smith, M. A. J. Amino Acids 2011, 2011, 461216. (31) Gadgil, H. S.; Bondarenko, P. V.; Pipes, G. D.; Dillon, T. M.; Banks, D.; Abel, J.; Kleemann, G. R.; Treuheit, M. J. Anal. Biochem. 2006, 355, 165−174. (32) Adachi, T.; Schöneich, C.; Cohen, R. A. Drug Discovery Today: Dis. Mech. 2005, 2, 39−46. (33) Sun, R.; Eriksson, S.; Wang, L. J. Biol. Chem. 2012, 287, 24304− 24312. (34) Zang, T.; Dai, S.; Chen, D.; Lee, B. W. K.; Liu, S.; Karger, B. L.; Zhou, Z. S. Anal. Chem. 2009, 81, 9065−9071. (35) Wang, Z.; Rejtar, T.; Zhou, Z. S.; Karger, B. L. Rapid Commun. Mass Spectrom. 2010, 24, 267−275. (36) Zhou, Z. S.; Smith, A. E.; Matthews, R. G. Bioorg. Med. Chem. Lett. 2000, 10, 2471−2475. (37) Nemet, I.; Monnier, V. M. J. Biol. Chem. 2011, 286, 37128− 37136. (38) Linetsky, M.; Shipova, E.; Cheng, R.; Ortwerth, B. J. Biochim. Biophys. Acta, Mol. Basis Dis. 2008, 1782, 22−34. (39) Chumsae, C.; Zhou, L. L.; Shen, Y.; Wohlgemuth, J.; Fung, E.; Burton, R.; Radziejewski, C.; Zhou, Z. S. Anal. Chem. 2014, 86, 8932− 8936. (40) Gilmore, J. M.; Scheck, R. A.; Esser-Kahn, A. P.; Joshi, N. S.; Francis, M. B. Angew. Chem., Int. Ed. 2006, 45, 5307−5311. (41) Witus, L. S.; Moore, T.; Thuronyi, B. W.; Esser-Kahn, A. P.; Scheck, R. A.; Iavarone, A. T.; Francis, M. B. J. Am. Chem. Soc. 2010, 132, 16812−16817. (42) Zhou, Z. S.; Flohr, A.; Hilvert, D. J. Org. Chem. 1999, 64, 8334− 8341. (43) Zhou, Z. S.; Jiang, N.; Hilvert, D. J. Am. Chem. Soc. 1997, 119, 3623−3624. (44) Gui, S.; Wooderchak-Donahue, W. L.; Zang, T.; Chen, D.; Daly, M. P.; Zhou, Z. S.; Hevel, J. M. Biochemistry 2013, 52, 199−209. (45) Zhang, J.; Zhang, T.; Jiang, L.; Hewitt, D.; Huang, Y.; Kao, Y.H.; Katta, V. Anal. Chem. 2012, 84, 2313−2320.
acyclic xylosone with a protein primary amine, list of relevant cell culture media compounds, and additional references. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.5b00801.
■
AUTHOR INFORMATION
Corresponding Authors
*E-mail: [email protected] *E-mail: [email protected] *E-mail: [email protected] Present Address ∥
Y.Z.: BioMarin, Novato, CA 94949-8703.
Downloaded by UNIV LAVAL on September 15, 2015 | http://pubs.acs.org Publication Date (Web): July 21, 2015 | doi: 10.1021/acs.analchem.5b00801
Notes
AbbVie Disclosure: The design, study conduct, and financial support for the study were provided by AbbVie (formerly the proprietary pharmaceutical division of Abbott Laboratories, now an independent biopharmaceutial company). AbbVie participated in the interpretation of data, review, and approval of the publication; all authors contributed to the development of the publication and maintained control over the final content. C.C., P.H., S.M., C. Rancicot, and C. Radziejewski are AbbVie employees. H.R. and Y.Z. are former AbbVie employees. Z.S.Z. serves as the doctoral advisor to C.C. and has received AbbVie support for this role. The authors declare no competing financial interest. ⊥ A portion of this work was conducted by C.C. as a Ph.D. student at the Department of Chemistry and Chemical Biology, Northeastern University, Boston, Massachusetts, USA.
■
ACKNOWLEDGMENTS The authors thank Dr. David Lee and Shanshan Liu for their critical review of the manuscript and Shanshan Liu for her help on the chemical scheme. This activity is partially supported by a grant from NIH NIGMS (Grant 1R01GM101396 to Z.S.Z.). This is contribution number 1056 from the Barnett Institute.
■
REFERENCES
(1) Zaia, J. Mass Spectrom. Rev. 2004, 23, 161−227. (2) Fenselau, C.; Vestling, M. M.; Cotter, R. J. Curr. Opin. Biotechnol. 1993, 4, 14−19. (3) Awdeh, Z. L.; Williamson, A. R.; Askonas, B. A. Biochem. J. 1970, 116, 241−248. (4) Liu, H.; Gaza-Bulseco, G.; Faldu, D.; Chumsae, C.; Sun, J. J. Pharm. Sci. 2008, 97, 2426−2447. (5) Manning, M.; Chou, D.; Murphy, B.; Payne, R.; Katayama, D. Pharm. Res. 2010, 27, 544−575. (6) Vlasak, J.; Ionescu, R. Curr. Pharm. Biotechnol. 2008, 9, 468−481. (7) van der Valk, J.; Brunner, D.; De Smet, K.; Fex Svenningsen, Å.; Honegger, P.; Knudsen, L. E.; Lindl, T.; Noraberg, J.; Price, A.; Scarino, M. L.; Gstraunthaler, G. Toxicol. In Vitro 2010, 24, 1053− 1063. (8) Wurm, F. M. Nat. Biotechnol. 2004, 22, 1393−1398. (9) Wong, D. C. F.; Wong, K. T. M.; Goh, L. T.; Heng, C. K.; Yap, M. G. S. Biotechnol. Bioeng. 2005, 89, 164−177. (10) Gramer, M. J.; Eckblad, J. J.; Donahue, R.; Brown, J.; Shultz, C.; Vickerman, K.; Priem, P.; van den Bremer, E. T.; Gerritsen, J.; van Berkel, P. H. Biotechnol. Bioeng. 2011, 108, 1591−602. (11) Jing, Y.; Borys, M.; Nayak, S.; Egan, S.; Qian, Y.; Pan, S.-H.; Li, Z. J. Process Biochem. 2012, 47, 69−75. (12) Crowell, C. K.; Grampp, G. E.; Rogers, G. N.; Miller, J.; Scheinman, R. I. Biotechnol. Bioeng. 2007, 96, 538−549. (13) Hossler, P.; McDermott, S.; Racicot, C.; Chumsae, C.; Raharimampionona, H.; Zhou, Y.; Ouellette, D.; Matuck, J.; Correia, I.; Fann, J.; Li, J. Biotechnol. Prog. 2014, 30, 1419−1431. 7533
DOI: 10.1021/acs.analchem.5b00801 Anal. Chem. 2015, 87, 7529−7534
Technical Note
Downloaded by UNIV LAVAL on September 15, 2015 | http://pubs.acs.org Publication Date (Web): July 21, 2015 | doi: 10.1021/acs.analchem.5b00801
Analytical Chemistry (46) Liu, M.; Zhang, Z.; Zang, T.; Spahr, C.; Cheetham, J.; Ren, D.; Zhou, Z. S. Anal. Chem. 2013, 85, 5900−5908. (47) Wan, W.; Zhao, G.; Al-Saad, K.; Siems, W. F.; Zhou, Z. S. Rapid Commun. Mass Spectrom. 2004, 18, 319−324. (48) Lee, K.-W.; Simpson, G.; Ortwerth, B. Biochim. Biophys. Acta, Mol. Basis Dis. 1999, 1453, 141−151. (49) Baker, K. N.; Rendall, M. H.; Hills, A. E.; Hoare, M.; Freedman, R. B.; James, D. C. Biotechnol. Bioeng. 2001, 73, 188−202. (50) Gu, X.; Wang, D. I. Biotechnol. Bioeng. 1998, 58, 642−8. (51) Qian, Y.; Jing, Y.; Li, Z. J. Biotechnol. Prog. 2010, 26, 1417−23. (52) Rodriguez, J.; Spearman, M.; Huzel, N.; Butler, M. Biotechnol. Prog. 2005, 21, 22−30. (53) Hossler, P.; Racicot, C.; McDermott, S. J. Glycobiol. 2014, 3, 108. (54) Chaderjian, W. B.; Chin, E. T.; Harris, R. J.; Etcheverry, T. M. Biotechnol. Prog. 2005, 21, 550−3. (55) Banks, D. D.; Hambly, D. M.; Scavezze, J. L.; Siska, C. C.; Stackhouse, N. L.; Gadgil, H. S. J. Pharm. Sci. 2009, 98, 4501−10. (56) Fischer, S.; Hoernschemeyer, J.; Mahler, H. C. Eur. J. Pharm. Biopharm. 2008, 70, 42−50. (57) Kim, K.; Rhee, S. G.; Stadtman, E. R. J. Biol. Chem. 1985, 260, 15394−15397. (58) Richheimer, A.; Robinson, A. B. J. Orthomol. Psychiatry 1977, 6, 290−299.
7534
DOI: 10.1021/acs.analchem.5b00801 Anal. Chem. 2015, 87, 7529−7534
