Selective Tryptophan Oxidation of Monoclonal Antibodies: Oxidative

Subscriber access provided by Iowa State University | Library

Article

Selective Tryptophan Oxidation of Monoclonal Antibodies: Oxidative Stress and Modeling Prediction Jorge Alexander Pavon, Li Xiao, Xiaojuan Li, Jia Zhao, Danielle Aldredge, Eugene Dank, Alex Fridman, and Yan-Hui Liu Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b04768 • Publication Date (Web): 04 Jan 2019 Downloaded from http://pubs.acs.org on January 8, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Selective Tryptophan Oxidation of Monoclonal Antibodies: Oxidative Stress and Modeling Prediction

Jorge Alexander Pavon*1, Li Xiao2, Xiaojuan Li1, Jia Zhao1, Danielle Aldredge1, Eugene Dank1, Alex Fridman1, Yan-Hui Liu1 1Process

Research & Development, Merck & Co., Inc., Kenilworth, NJ 07033, USA and Informatics, Merck & Co., Inc., Kenilworth, NJ 07033, USA.

2Modeling

* To whom correspondence should be addressed Jorge Alexander Pavon 2015 Galloping Hill Road Kenilworth, New Jersey 07033 United States [email protected] (908)740-6886 Keywords: Critical Quality Attribute, Tryptophan oxidation, Complementarity determining region Monoclonal antibody, Solvent-accessible surface area

Abbreviations used: CQA, critical quality attribute; mAb, monoclonal antibody; IgG, immunoglobulin G; HC, heavy chain; AAPH, 2,2’-azobis(2-amidinopropane) dihydrochloride; CDR, Complementarity Determining Region; PTMs, post-translational modifications; tBHP, tert-butyl hydroperoxide; Fv, variable fragment; EIC, extracted ion chromatogram; SASA, solvent-accessible surface area; CHO, Chinese hamster ovary; DTT, dithiothreitol; IAM, iodoacetamide; PDB, Protein Data Bank; MD, molecular dynamics; Fab, antigen-binding fragment.

1

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 21

ABSTRACT Oxidation of tryptophan not only generates heterogeneity of a therapeutic monoclonal antibody (mAb), but it could also be a potential critical quality attribute (CQA) of the product. In this study, monoclonal antibodies (mAbs) A-C of IgG1 and IgG4 isotypes with oxidized tryptophan (Trp) residues were selectively generated by incubating the mAbs with 2,2’-azobis(2-amidinopropane) dihydrochloride (AAPH) in formulations containing L-methionine. The site-specific oxidation of tryptophan residues were confirmed by liquid chromatography coupled with mass spectrometry (LC-MS) studies. The site of oxidation was identified to be a conserved tryptophan residue in the heavy chain Complementarity Determining Region 3 (CDR3) of mAbs A and B with no significant oxidation found on other tryptophan residues including those in close proximity to CDR3. For mAb C, all tryptophan residues including one in the heavy chain CDR1 and a tryptophan in close proximity to heavy chain CDR3 were not susceptible to oxidation. For all three mAbs, the structure and tryptophan oxidation relationship was further studied by computational modeling of the variable domain of the antibodies (Fv). The computational modeling provided a structural understanding at the molecular level to the tryptophan oxidation, where high solvent accessibility is a prerequisite for heavy chain CDR3 tryptophan oxidation. However, higher oxidation susceptibility of tryptophan in heavy chain CDR3 did not linearly correlate to higher solvent accessibility, suggesting that other factors including side-chain orientation and/or surrounding structure elements around the heavy chain CDR3 may also be involved. Through this study, we demonstrate that a selective oxidation system, together with computational modeling, can be an important tool to identify potential CQAs of a therapeutic mAb such as tryptophan oxidation liabilities during the mAb’s development.

2


Page 3 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

INTRODUCTION Oxidation is a common degradation pathway of proteins and peptides. For biopharmaceutical

products, this detrimental modification causes product heterogeneity and has been reported to occur during production, purification, formulation, and storage1-7.

Depending on its effect on safety and

efficacy, oxidation can be deemed a critical quality attribute (CQA) of a biopharmaceutical product1-7. Hence, it is important to control and monitor the amount of oxidized variants of drug product and drug substance as it can affect release specifications.

It is well-known that methionine (Met), tryptophan

(Trp), histidine (His), and tyrosine (Tyr) are the major amino acid residues susceptible to oxidation in proteins1, 8-10, even though tryptophan and methionine are rare amino acids in proteins39. For therapeutic mAbs, the amino acids of most concern for oxidation are tryptophan and methionine. Tryptophan oxidation in the CDRs of a mAb is known to affect afﬁnity and speciﬁcity for antigen binding11-13. Methionine in the CDRs is less commonly found than tryptophan30, but its effect on antigen affinity and specificity when oxidized are expected to be similar to tryptophan14, 27, 40-41. Methionine oxidation is more commonly seen in the Fc domain of antibodies15-17, 41. High Fc oxidation levels of methionine can have adverse effects on the binding and affinity for Fc receptors including FcRn and the Protein A ligand14-17. Methionine is known to be susceptible to oxidation due to the intrinsic reactivity of sulfur atoms against peroxides18, 19; whereas tryptophan is reactive due to the electrophilic nature of the indole ring toward various forms of reactive oxygen species including those with metal complexes18, 20-21, 31. Multiple studies have been done over the years in order to understand if undesired post-translational modifications (PTMs), such as oxidation, can lead to structural changes that could potentially impact biological efficacy, clearance, safety, and immunogenicity of therapeutic mAbs. For a biotherapeutic mAb CQA determination, as well as for analytical method development for CQA monitoring, forced degradation is usually applied to generate significant levels of product-related variants, such as oxidation variants3, 10, 22. The challenge with forced degradation studies is the lack of selectivity, since most of the common stress conditions such as light, oxidation using chemicals, extreme pH and thermal stresses generate a mixture of product-related variants, which lead to difficulty in assessing the criticality of each attribute independently. Multiple examples in the literature have shown that forced oxidative conditions such as treatment with hydrogen peroxide or tert-butyl hydroperoxide (tBHP) can effectively generate methionine oxidized mAb variants with minimal changes to other attributes3, 5, 16, 18, 23, 41. In the case of tryptophan, due to its photosensitivity, light irradiation (photo stress) is the preferred condition to generate oxidized tryptophan variants of mAbs9,

11, 13, 24.

However, light irradiation is not selective due to the radical involvement;

additional product-related variants with attributes other than oxidized tryptophan have been observed. 3



Such product-related variants from light irradiation include aggregates, fragments, histidine and methionine oxidized variants, and tryptophan side chain cleavage9, 24-25,31. Ozone has also been used to target CDR tryptophan residues, but sample preparation requires dialysis against ozonized buffers38, 11. A few studies have shown that AAPH can be used to target exposed tryptophan and methionine residues with higher selectivity for tryptophan18, 26, 27. As has been discussed by Ji et al. 18, AAPH itself is not an oxidant, but the product from the reaction with molecular oxygen generates alkyperoxyl radicals, which can be used to investigate oxidation pathways from radical species without light irradiation. Selective oxidation of tryptophan or methionine residues in the model protein parathyroid hormone was achieved with AAPH and with addition of free methionine or tryptophan18. Similar work was carried out on a therapeutic mAb26 In the present study, a selective oxidation system consisting of AAPH as oxidant with added Lmethionine (L-Met) to the formulations was used to generate CDR-tryptophan-oxidized mAb variants of IgG1 and IgG4 isotypes with minimal increases to other product-related variants. Three recombinant mAbs A-C were selected for this study, where mAbs A and B have a conserved tryptophan residue in their heavy chain CDR3, and mAb C has a tryptophan residue in its heavy chain CDR1. Furthermore, addition of free tryptophan was also investigated to eliminate or minimize CDR tryptophan oxidation. To understand the observed differences in reactivity or oxidation susceptibility of tryptophan and the correlation to solvent exposure of the different CDR tryptophan residues, computational analysis of the variable fragment (Fv) of the three different mAbs was performed. The results showed that a level of solvent-accessible surface area (SASA) and solvent exposure are essential for oxidation of heavy chain CDR3 tryptophan by AAPH. The versatility of this oxidation system can be an important tool for selective generation of CDR-tryptophan oxidized mAb variants for further characterization by in vivo or in vitro assays.

2

MATERIALS AND METHODS Monoclonal Antibodies (IgGs). MAbs A (IgG1), B (IgG4), and C (IgG1) were produced in

Chinese hamster ovary (CHO) cells at MSD, and purified by protein A chromatography followed by ionexchange chromatography. The mAbs were formulated at 25 or 50 mg/mL in identical formulations with slight differences in pH. L-methionine (10 mM) was added to mAb A and mAb B was formulated with and without L-methionine prior to the forced degradation studies. All mAbs as liquid formulations were kept at -80 C until experiments. Forced oxidation of mAbs with AAPH. The reactions were conducted by incubating mAbs A-C with AAPH at 40 °C for predetermined time points ranging from 6 to 24 hours. For mAbs A and C, the 4


Page 4 of 21

Page 5 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


reactions were carried out by adding 17 mM (~ 51-fold excess in molar ratio) of AAPH to 50 mg/mL of antibodies for time points up to 6 hours, with or without addition of 2.5 mM DL-tryptophan to the reaction for mAb A, or without 2.5 mM DL-tryptophan in the reaction mixture for mAb C. For mAb B, an antibody stock solution was prepared by diluting to 10 mg/mL in the respective formulation buffer with or without 2.5 mM DL-tryptophan (final concentrations). For mAb B, the reactions were carried out at 10 mg/mL based on its attributes. The oxidation reaction for mAb B was initiated by the addition of 1.2 mM AAPH (~17-fold excess in molar ratio) for 24 hours or 3.4 mM (~ 51-fold excess in molar ratio) for 6 hours at 40 °C. Control samples of antibody solutions without AAPH were prepared and incubated at 40 °C for the duration of the stress (6 or 24 hours). Upon completion of the reaction, the samples were buffer exchanged into their respective formulation buffer with the 2-mL Zeba spin desalting column. The column was first equilibrated with formulation buffer three times, and the entire sample (0.7- 1.5 mL) was loaded onto the column and centrifuged for four minutes at 1000 g. All stressed and control mAb samples were stored at -80 °C until analysis. Reduced peptide mapping by LC-MS. For mAbs A, B and C, 100 µg of each sample was diluted with 20 µL of water, and then mixed with 80 µL of denaturing buffer (6 M Guanidine HCl, 50 mM Tris buffer at pH 8.0, and 5 mM Ethylenediaminetetraacetic acid (EDTA) to a final volume of approximately 100 µL. Two microliters of 1 M dithiothreitol (DTT) was added to each sample and the sample was incubated at 37 ºC for 30 min in the dark. The sample was cooled at room temperature for 5 min and subsequently alkylated with 5 µL of 1 M Iodoacetamide (IAM) at 37 ºC for 30 min in dark. The alkylation reaction was terminated by adding 1 µL of 1 M DTT. The reduced and alkylated sample was diluted with 500 µL of digestion buffer containing 50 mM Tris buffer at pH 8.0. For mAbs A and B, the samples were digested with Lys-C and trypsin, each at an enzyme:substrate ratio of 1:10 (w/w) and incubated at 37 °C for 4 hrs. This reaction was terminated by adding 15 µL of 20% TFA. For mAb C the sample was digested with Lys-C at an enzyme:substrate ratio of 1:40 (w/w) and incubated at 37 °C for 1 hr. Then another aliquot of Lys-C was added and continued the incubation for 3hrs. The final enzyme:substrate ratio was 1:20 (w/w) and total incubation time was 4 hrs. The digestion reactions for mAbs A, B, and C were terminated by adding 15 µL of 20% TFA. HPLC conditions and data analysis is described in the supporting information Computational modeling. The homology models of Fv domain of each of mAbs A, B and C were generated using Antibody Modeler of MOE33, a computer modeling package from CCG, where the templates were identified with the X-ray structures of the most homologous antibodies from a collection of more than 4000 antibody structures in PDB (The Protein Data Bank37), including framework and CDR loops. When there was no template structure found through automated procedure in MOE Antibody 5



Page 6 of 21

Modeler, such as the H3 loop in mAb A; the MOE antibody database was analyzed to select the most closely related structure as the template based on both sequence and structure information. The sequence alignment and model building followed MOE Antibody Modeler steps. For individual homology model, the solvent-accessible surface area of tryptophan residue was calculated with Protein Properties calculator of MOE. To sample the conformations of tryptophan residues, the molecular dynamics (MD) simulations of mAb structures were performed using DESMOND from Schrodinger34, with explicit solvent and periodic boundary conditions, through an orthorhombic box with TIP3P waters and the OPLS-3 force field for each mAb. The system was set up with DESMOND default setting (T=300K, P=1 atm). Each simulation was run for 50 ns. The MD trajectories were analyzed using the scripts provided by Schrodinger34. These included: SASA around the tryptophan residue, SASA around the nitrogen atom of the tryptophan sidechain indole ring, and the 2-shell waters that were calculated as the number of water molecules within a radius “R” from the nitrogen atom of tryptophan sidechain indole ring, at each MD simulation time step.

3

RESULTS Identification of oxidation sites of AAPH stressed mAbs by LC-MS/MS analysis. MAbs A-C

were treated with or without AAPH for 6 and 24 hours.

The sites and levels of tryptophan and

methionine oxidation were determined by LC-MS reduced peptide mapping followed by LC-MS/MS for site-specific oxidation characterization. It is worth noting that the oxidation conditions for this study were optimized to yield minimal levels of aggregates (results not shown). Aggregation is a degradation pathway commonly observed from the extended exposure to AAPH under high concentrations1, 27. Both mAbs A and B have a tryptophan residue within their heavy chain CDR3 regions, with tryptophan 104 (Trp104) in mAb A and Trp102 in mAb B, while mAb C has Trp33 in its heavy chain CDR1. Analysis of LC-MS peptide mapping data of Lys-C or Lys-C/trypsin digested mAbs suggests that AAPH treatment in the presence of L-methionine results in tryptophan oxidation for mAbs A and B while only methionine oxidation is observed for mAb C without L-methionine in the reaction solution (Table 1). LC-MS/MS analysis was applied on the oxidized heavy chain CDR peptides of mAbs A and B to identify the specific sites of oxidation. For mAb A, MS/MS of the + 16-Da oxidized peptide GGPYGW104YFDVW109 GQGTTVTVSSASTK (HC 98-223), which covers heavy chain CDR3 sequence GGPYGW104YFDV, is shown in Figure 1. Figure 1 shows the fragmentation of the +16-Da oxidized peptide, with the singly(m/z 2219.0486, 1+) and doubly-charged y20 with a loss of NH3 (m/z 1110.0272, 2+) demonstrating the addition of the +16-Da oxidation modification on Trp104. Since this increase in mass is absent in y20 6


Page 7 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


fragment ion of the apo-form peptide (data not shown), and y15-y19 fragment ions of the +16-Da oxidized peptide do not have the addition of 16 Da, it provides additional evidence that the oxidation site is Trp104. Furthermore, the inset on Figure 1 shows low abundant N-terminal fragment ions of b5-b7 ions, where addition of 16 Da to the b6 and b7 ions, but not the b5 ion were observed for the +16-Da modified peptide as compared with those of the apo-form peptide (not shown); consistent with Trp104 as the oxidation site. Similar analysis of the +32-Da oxidation products of mAb A confirms Trp104 as the only oxidation site (supporting information). Similar to mAb A, mAb B contains Trp102 in heavy chain CDR3 and Trp109 flanking CDR3, NYRW102FGAMDHW109GQGTTVTVSSASTK (HC 99-123).

The MS/MS analysis of the oxidized

peptides corresponding to +16-Da is shown in Figure 2. Figure 2 shows the fragmentation pattern with b and y series ions, for the +16-Da oxidized peptide; the series (b4-b7) show the +16-Da increase in mass in comparison to the apo-form peptide’s fragments, suggesting the oxidation site is located at Trp102. Met106 is ruled out as the oxidation site with the detection of b8 – b10 ion series; Trp109 is ruled out as an oxidation site by noting the lack of additional increase in mass for the b11 and y15 ions. Identification of Trp102 as the oxidation site is reinforced by the lack of +16-Da increase for the b3 fragment ion flanking Trp104. Similar analysis for the +32-Da oxidation products of mAb B (supporting information) confirms Trp102 as the only oxidation site.

Overall, for mAbs A and B, the heavy chain CDR3 Tryptophan

residues are the only oxidation sites within their corresponding heavy chain CDR3-containing peptides. The LC-MS/MS analysis of the peptide containing heavy chain CDR1 Trp33 in mAb C following AAPH treatment did not show evidence of oxidation. The extracted ion chromatograms (EIC) of both peptides (apo-form and same peptide following incubation with AAPH for 6 hours did not show any differences (results not shown); this suggests that AAPH was not able to modify any residues within the CDR1 peptide. Determination of oxidation levels by LC-MS of reduced peptide mapping of AAPH stressed mAbs. Having confirmed the sites of oxidation of mAbs A and B to be the tryptophan residues within the heavy chain CDR3, subsequent analysis was carried out to determine the levels of tryptophan oxidation. Table 1 shows the site and the level of oxidation of tryptophan residues within the heavy chain CDR3 of mAbs A and B, as well as oxidation of methionine residues in Fc of all mAbs, by LC-MS reduced peptide mapping. The level of methionine in the formulation is expected to suppress methionine oxidation in Fc, as has been shown in previous studies18, 24, with similar results obtained in this study (Table 1). For mAb C, the increase in Met250 oxidation is consistent with the lack of L-methionine in the reaction (Table 1).

Since tryptophan oxidation was not observed for mAb C, probing a selective

oxidation system was not necessary. Figures 3A shows representative extracted ion chromatogram (EIC) 7



Page 8 of 21

of oxidation modified peptides for mAb B after treatment with AAPH for 24 hours. Figure 3B shows the representative EIC under the same reaction conditions as Figure 3A, but with 2.5 mM of DL-tryptophan added to the reaction solution prior to the addition of AAPH. Respective EIC traces for mAb A after treatment with AAPH ± 2.5 mM DL-tryptophan are shown in Figures S3 and S4 (supporting information). It is evident that oxidized peptides eluted earlier than their corresponding native peptides and addition of DL-tryptophan reduced the level of oxidation of tryptophan residues in heavy chain CDR3. The identification of multiple oxidized peptides from AAPH treated mAbs A and B is consistent with tryptophan oxidation products.

MS detected ions corresponding to oxidized peptides with

monoisotopic masses that correspond to +4-, +16- and +32-Da oxidation species (Figures 3A and S3). These oxidized peptides correspond to tryptophan oxidation products of kynurenine (+4 Da), hydroxytryptophan and/or oxindolylalanine (+16 Da), and N-formylkynurenine (+32 Da)18,

28.

The

oxidation products of tryptophan residues in heavy chain CDR3 were also observed in the light stressed mAbs29. Table 1 summarizes the level of oxidation for the tryptophan residues in CDRs and Met252 (EU numbering system) residues in mAbs A-C. It is worth noting that other methionine residues including Met428 are also oxidized without L-methionine present, but for this study, only Met252 was monitored since it is known to be the most sensitive to oxidation among the Fc methionines15, 16, 23, 41. The lack or minimal increases in Met252 oxidation (Table 1) illustrate that this selective oxidation system is applicable to multiple mAbs of different isotypes (IgG1 and IgG4). Ranking the oxidation susceptibility of tryptophan residues in heavy chain CDRs. The susceptibility of heavy chain CDR tryptophan oxidation of mAbs A-B is ranked in the following order: mAb B (CDR3 Trp102) > mAb A (CDR3 Trp104) and mAb C has no oxidation of CDR1 Trp33. The results in Table 1 show that the tryptophan residue in mAb B has the highest susceptibility to oxidation by AAPH.

Furthermore, tryptophan oxidation for mAb B increases (~6.0%) after 6 hours and more

significant after 24 hours (~17%), despite the addition of free tryptophan in the reaction solution (Table 1).

For mAb C, no tryptophan-oxidized peptide was identified by LC-MS (data not shown).

As

mentioned above, the tryptophan residue in mAb C is located in heavy chain CDR1 (Trp33), while for mAbs A and B, the tryptophan residues are all located within heavy chain CDR3. Computational predictions of tryptophan oxidation. Through Antibody Modeler of MOE, a homology model of the Fv domain of each mAb, A, B, and C, was constructed using the templates identified based on sequence homology between the query sequence and sequence template.

The

templates were selected independently on different regions of Fv, e.g. framework (FR), heavy chain loop 1(H1), H2, H3, light chain loop 1 (L1), L2, and L3. The SASA of each tryptophan residue in the Fv domain of the mAbs was calculated using the homology models and is shown in Figure 4. Table 2, lists 8


Page 9 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the SASA values of CDR tryptophan residues and those flanking the CDR in the Fv domain of mAbs AC, along with the percentage of solvent exposure, as defined by the ratio with the surface area of tryptophan in the Gly-Trp-Gly tri-peptide. The data in Table 2 show that for mAbs A and B, the tryptophan residue located in CDR3 (Trp104 or Trp102) have SASA values of 120 Å2 for Trp104 (mAb A), and 191 Å2 for Trp102 (mAb B). As shown in Table 1 and confirmed by MS/MS (Figures 1 and 2), these two tryptophan residues in mAbs A and B are highly sensitive to oxidation by AAPH. Trp33, located in CDR1 of mAb C, is not susceptible to oxidation by AAPH and shows a calculated SASA value of 54 Å2. To investigate the level of solvent exposure of individual tryptophan residue in each mAb and its correlation with tryptophan oxidation by AAPH; the SASA values of all tryptophan residues within the Fv domains of the three antibodies were analyzed in detail. MAb A contains five tryptophan residues in the Fv domain, spreading across both heavy and light chains, as shown in Figure 4A; one on the light chain and four on the heavy chain. Most of these tryptophan residues are buried with SASA values < 11 Å2, except Trp104 located in CDR3 H3 loop. As a comparison, Trp104 has a SASA value of 120 Å2 and solvent exposure of 38.6%, whereas Trp109 has a SASA value of 10.2 Å2 and solvent exposure of 3.3% (Table 2). This corresponds to only ~8.5% of the SASA and solvent exposure of Trp104. The significant level of solvent exposure of Trp104 is consistent with its susceptibility to AAPH oxidation, while the lack of Trp109 oxidation can be explained by its limited solvent exposure. Similar to mAb A, mAb B has five tryptophan residues in its Fv domain, with only Trp102 showing significant solvent exposure (Figure 4B); as mentioned above, its SASA value is 191 Å2 with a solvent exposure of 61.5% (Table 2). All other tryptophan residues of mAb B are partially buried with SASA values less than 33 Å2. Consistent with SASA profile, only Trp102 was found to be susceptible to oxidation by AAPH.

It is worth noting that while Trp109 show a level of exposure (32.5 Å2);

experimentally, no oxidation was detected for this tryptophan residue that flanks CDR3. MAb C contains five tryptophan residues in its Fv domain (Figure 4C), similarly to mAbs A and B. However, this mAb is unique in that it does not have a tryptophan residue in its heavy chain CDR3; rather it has Trp33 located within H1 CDR1. The calculations (Table 2) revealed that all tryptophan residues in mAb C are either partially or completely buried; Trp33 is the most exposed in this antibody with a SASA of 53.6 Å2 and solvent exposure of 17%. Within its framework, mAb C contains a Trp105, with limited SASA (28 Å2) (Table 2), a residue in the position flanking heavy chain CDR3, and aligned to Trp109 of mAbs A and B. As mentioned, experimentally, none of mAb C tryptophan residues were found to be susceptible to AAPH oxidation, consistent with their limited solvent exposure. To investigate the effect of structure flexibility on the solvent exposure of tryptophan residues in the mAbs, the conformations of tryptophan residues were sampled through a 50 ns MD simulation with a 9



Page 10 of 21

fixed backbone for the three antibodies. Table 3 lists the SASA and 2-shell model averages over MD simulation trajectory on the most exposed tryptophan residue of each antibody. For the 2-shell models, results on different radii were included as well. Comparing Table 2 and Table 3, it is clear that SASA values from MD sampling averages (dynamic model) are similar to those from a single homology model (static model), and following a similar trend among the three antibodies. For example, Trp104 (mAb A) displayed a 120 Å2 of SASA from the static model, whereas it showed a 112.0 Å2 of SASA of MD average. The number of water molecules around each tryptophan residue computed using a 2-shell model also aligned with solvent exposure as measured with SASA; increased exposure of a tryptophan residue resulted in a larger SASA value, and thus, more water molecules in the shell. Table 3 also shows tryptophan side-chain SASA from MD averages, MD_SASA_sc. These are similar to SASA of the whole tryptophan residues. These results indicate that the solvent exposure of a tryptophan residue in an antibody could be measured by both static and dynamic models with similar results. The effect from MD sampling of tryptophan conformations does not significantly alter the exposure of tryptophan residues in the three antibodies studied here.

4

DISCUSSION Oxidation of tryptophan and methionine residues in the heavy chain CDRs and Fc regions of

therapeutic antibodies have been investigated and reported in numerous publications and continue to be of great concern due the potential loss of antigen binding affinity for the former and potential compromise on serum half-life via loss of FcRn binding capabilities and/or instability of the Fc fragment for the latter. It has also been well established that the oxidation propensity under different stress conditions including accelerated stability studies is directly related to the type, solvent exposure, and location of the amino acid in the three-dimensional structure5, 6, 11-13, 15-17. In recent years, forced oxidative conditions along with the use selective scavengers/excipients have been utilized to study site-specific oxidation12,

18, 26.

In the

present study, a similar selective oxidation approach has been utilized, in which tryptophan oxidized mAb variants of IgG1 (mAbs A) and IgG4 (mAb B) isotypes were generated, with the goal to relate oxidation propensity of tryptophan residues, in or flanking heavy chain CDRs, to the range of solvent accessibility values calculated from structure models. The oxidative conditions implemented for this study clearly show that a combination of AAPH and the addition of free L-methionine to the formulations generated mAbs A and B variants with high levels of oxidized heavy chain tryptophan in CDR3, with minimal impact to the Fc methionine oxidation levels. This is reflected by the low increases in Met252 oxidation. Strong evidence for the selective tryptophan oxidation comes from the inhibition observed when free tryptophan is added to the oxidation reaction 10


Page 11 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(Table 1). While similar approaches have been implemented in other studies with therapeutic mAbs12, 26, and combining with unpublished data of our internal antibodies, the results presented here clearly show that this optimized oxidation system can be applied to IgG1 and IgG4 isotypes. Characterization of the oxidized mAbs by LC-MS and MS/MS of the enzyme digested mAbs A and B confirmed that the selective oxidation occurred exclusively on the tryptophan residue located within heavy chain CDR3 (Figures 1, 2, S1 and S2). Other tryptophan residues in the antigen-binding fragment (Fab) and those flanking heavy chain CDR3 were not oxidized. Similarly, Trp33 in heavy chain CDR1 of mAb C was not oxidized.

These results are not entirely surprising when considering that

tryptophan residues within the CDR regions have been shown to be highly solvent exposed; likely due to their role in antigen binding and specificity. For mAbs A and B the levels of heavy chain CDR3 tryptophan oxidation correlated with decreases in potency. Multiple publications have highlighted the connection between solvent accessibility and oxidation propensity11-13, 30. While it is not surprising to find that the different heavy chain CDR3 tryptophan residues in mAbs A and B were oxidized by AAPH, the lack of oxidation in Trp33 of mAb C located in heavy chain CDR1 was not initially obvious. Furthermore, tryptophan oxidation by AAPH was only localized within heavy chain CDR3, but not tryptophan residues flanking the CDR region. Tryptophan residues in the framework adjacent to the C-terminus of the heavy chain CDR3 such as Trp109 in mAb A and mAb B, and Trp105 in mAb C; do not have a level of solvent exposure required for tryptophan oxidation susceptibility as measured by SASA. Solvent exposure, as quantified by SASA, is a common approach to predict antibody development liability hotspots, including methionine oxidation, tryptophan oxidation, and antibody aggregation

35, 36.

In this study, a set of static and dynamic models were defined to evaluate the tryptophan solvent exposure and establish a correlation with the tryptophan oxidation experimental data. Similar models have been shown to be highly accurate for the prediction of methionine oxidation35, 41. The results presented here demonstrate that several models were very effective in predicting tryptophan oxidation by AAPH, as shown in Table 2-3 and Figure 5. The SASA from a static homology model structure of an antibody was accurate to predict tryptophan oxidation. Its simplicity makes it more valuable for a rapid assessment of tryptophan oxidation in mAbs. However, caution is needed when building the homology model of the antibody as the tryptophan conformation might be sensitive to the templates used. In this work, different templates were tested for the homology models and variations in the SASA values were observed, but not large enough to alter the trend (data not shown). For the dynamic models, the trajectories were stable for mAb B and mAb C when all atoms in the antibody were allowed to freely move without any restraint. However, in mAb A, without any restraint, large conformation changes across the structure, including framework and CDR loops, were observed, resulted in significant fluctuations in SASA of several 11



Page 12 of 21

tryptophan residues over 50 ns MD trajectory. Thus, here only the SASA of MD averages from the fixed backbone trajectories were reported. The static model results (Table 2) and the analysis using different dynamic models (Table 3) showed that the tryptophan residues in mAbs A and B prone to oxidation have significantly higher levels of SASA and solvent exposure in comparison to other tryptophan residues within the Fv region (Figure 4). The SASA values of CDR3 tryptophan residues in mAbs A and B from the static models are at least > 120 Å2, suggesting that there is a minimum SASA required for oxidation propensity. For mAb C, the analysis showed that Trp33 has an SASA of 54 Å2; experimentally no oxidized peptides were detected from the AAPH treatment. The computational evaluation suggests that a value greater than 54 Å2 and ≤ 120 Å2 of SASA is required for oxidation propensity. It is worth noting that on a related study, the authors found that Trp53 in a therapeutic mAb was identified to be the root cause for reduction of molecular oxygen to hydrogen peroxide13. Trp53 was found to be sensitive to photo-oxidation (source of electrons) due to its higher surface exposure (115 Å2). In contrast, other tryptophan residues within the Fv region were not photosensitive with surface exposure values ≤ 45 Å2, 13. The surface exposure area of Trp53 is in the same range as mAb A Trp104 (120 Å2). This strongly supports the notion that a finite minimum solvent exposure value is linked to oxidation propensity independent of the mechanism. It is evident when comparing the experimentally observed tryptophan oxidation levels (Table 1) to the calculated SASA values (Table 2 and 3) that a trend is observed, higher oxidation propensity with higher solvent exposure (Figure 5). The blue and green bars in Figure 5 indicate that Trp102 of mAb B, with a % MD_SASA value of 57.3 Å2 is the most susceptible to oxidation by AAPH. However, when we attempted to use %MD_SASA to predict directly the absolute level of oxidation, it was not successful, as shown with the green bars in Figure 5 (compare ratios of blue and green bars between mAbs A and B). Since mAb C Trp33 is not susceptible to oxidation by AAPH, a green bar is not applicable (Figure 5). The lack of linear relationship between mAb A and mAb B could be rationalized by the oxidized product distribution observed (Figures 3 and S3). For mAb A, the relative distribution and intensity of oxidized tryptophan peptides appear comparable (Figure S3). This suggests that the activated AAPH species is able to oxidize the indole ring of Trp104 of mAb A in multiple positions without formation of a predominant oxidized peptide. In other words, the distributions of the various tryptophan oxidation products, N-formylkynurenine, kynurenine, hydroxytryptophan, and/or oxindolylalanine appear comparable with no obvious selective enrichment. It is worth emphasizing that the experimental results presented here do not provide structural information on the possible different oxygen additions to the sixmembered benzene ring of indole or formation of oxindolylalanine. However, it is reasonable to assume that the multiple +16-Da peptides observed are the result of oxygen addition to different carbons. 12


Page 13 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


In the case of mAb B, the oxidized heavy chain CDR3 peptide distribution appears different from mAb A where the +4 Da oxidation product kynurenine appears to be predominantly formed, in addition to more oxidized peptides (Figures 3A, and S3). This suggests that the reactivity of the activated AAPH species against the indole ring in Trp102 of mAb B may be different from that of CDR3 tryptophan residue in mAb A. A conceivable explanation for the higher oxidation propensity in mAb B relative to its exposed surface as compared to mAb A (Figure 5) comes from the analysis using the MD trajectories. In mAb B, larger conformation changes within the 50 ns MD trajectory were observed, compared to mAb A, indicating more flexibility for the Trp102 sidechain in mAb B. Although the MD trajectories performed here would not be able to define the specific movements within the side chain (indole ring), what this may suggest is that side-chain movements may lead to different orientations allowing different reactivity between AAPH and carbons within the indole ring. A significant number of studies have shown that the first step in the non-enzymatic and enzymatic oxidation of tryptophan to form formylkynurenine and subsequently kynurenine is oxygen addition across the C2-C3 bond located in the five-membered nitrogen-ring24, 28, 31, 32, 38. The fact that the predominant oxidized CDR3 peptide in mAb B contains Trp102 in the form of kynurenine, suggests that the larger conformations may allow more accessibility of C2-C3 in mAb B as compared to mAb A. Additional work is required in order to support this hypothesis and address the root cause for the lack of linear correlation between mAb A and mAb B (Figure 5). Our work showed that a level of SASA and solvent exposure are essential for oxidation of heavy chain CDR3 tryptophan by AAPH. However, SASA and solvent exposure alone cannot completely explain the observed trends of heavy chain CDR3 tryptophan oxidation. On the other hand, the oxidation trends can be explained by proposing that side-chain orientation and/or surrounding structure elements around the CDR are also critical factors. Independent of the additional factors, computational analysis and experimental data does reveal that relatively high SASA and high solvent exposures are a requirement for heavy chain CDR3 tryptophan oxidation by AAPH.

5

CONCLUSIONS In this study, we have shown that tryptophan oxidation propensity of a mAb is related to a level of

SASA and solvent exposure in heavy chain CDR3 tryptophan residues. Computational modeling by static and dynamic models correlates well with the experimental results. The study can predict a solvent exposure range required for a tryptophan residue in mAb CDRs to be susceptible to oxidation by AAPH and likely other oxidants. The study also shows that a combination of computational modeling and experimental data can be utilized to predict oxidation propensity in mAbs. Future work will investigate the importance of indole orientation within the CDR region of mAbs to predict reactivity. With the 13



Page 14 of 21

experimental and computational modeling results presented here, future computational analysis on other mAbs alone can likely predict oxidation propensity.

AUTHOR INFORMATION Corresponding Author * Jorge Alexander Pavon 2015 Galloping Hill Road Kenilworth, New Jersey 07033 United States [email protected] (908)740-6886

Acknowledgement We want to thank Robert Chou, Edward Sherer, Veronica Juan and Peter Salmon for their support and input.

References (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17) (18) (19) (20) (21)

Torosantucci, R.; Schöneich, C.; Jiskoot, W. Pharm Res. 2014, 31, 541-553. Jenkins, N.; Murphy, L.; Tyther, R. Molecular Biotechnology 2008, 39, 113-118. Haberger, M.; Bomans, K.; Diepold, K.; Hook, H.; Bulau, P; et al. mAbs 2014, 6, 327-339. Eon-Duval, A.; Broly, H.; Gleixner, R. Biotechnol. Prog. 2012, 28, 608-622. Chumsae, C.; Gaza-Bulseco,G.; Sun, J.; Liu, H. Journal of Chromatography B 2007, 850, 285-294. X. M. Lam, W. G. Lai, E. K Chan, V. Ling, C. C. Hsu. . Pharm Res. 2011, 28, 2543–2555. Wong, C.; Strachan-Mills, C.; Burman, S. Journal of Chromatography A. 2012, 1270, 153-161. Berlett, B. S.; Stadtman, E. R. The Journal of Biological Chemistry 1997, 272, 20313-20316. B. A. Kerwin, R. L. Remmele Jr. Journal of Pharm. Sciences 2007, 96, 1468-1479. Luo, Q.; Joubert, M. K.;Stevenson, R.; Ketchem,R. R.; Narhi, L. O.; J. Wypych, J. Journal of Biological Chemistry 2011, 286, 25134-25144. Z. Wei, Z.; Feng, J.; Lin, H.; Mullapudi, S.; Bishop, E.; Tous, G.I.; Casas-Finet, J.; Hakki, F.; Strouse, R.; Schenerman, M.A. Anal. Chem. 2007, 79, 2797-2805. Boyd, D.; Kaschak, T.; Yan, B. Journal of Chromatography B 2011, 879, 955-960. Sreedhara, L.; Lau, K.; Li, C.; Hosken, B.; Macchi, F.; Zhan, D.; Shen, A.; Steinmann, D.; Schöneich, C.; Lentz, Y. Mol. Pharmaceutics 2013, 10, 278-288. Yan, Y.; Wei, H.; Fu, Y.; Jusuf, S.; Zeng, M.; Ludwig, R.; Krystek Jr., S. R.; Chen, G.; Tao, L.; Das, T. K. Anal. Chem. 2016, 88, 2041–2050. Wang, W.; Vlasak, J.; Li, Y.; Pristatsky, P.; Fang, Y.; Pittman, T.; Roman, J.; Wang, Y.; Prueksaritanont, T.; Ionescu, R. Mol. Immunol. 2011, 48, 860-866. Bertolotti-Ciarlet, A.; Wang, W.; Lownes, R.; Pristatsky, P.; Fang, Y.; McKelvey, T.; Li, Y.; Li, Y; Drummond, T. Prueksaritanont, J. Vlasak. Mol. Immunol. 2009, 46, 1878-1882. Pan, H.; Chen, K.; Chu, L.; Kinderman, F.; Apostol, I.; Huang. G. Protein Science 2009, 18, 424-433. Ji, J.A.; Zhang, B.; Cheng, W.; Wang, Y.J. J. Pharm. Sci. 2009, 98, 4485-4500. Chu, J.; Yin, J.; Wang, D.I.C.; Trout, B. L. Biochemistry 2004, 43, 14139-14148. Thomas, A.H.; Serrano, M. P.; Rahal, V.; Vicendo, P.; Claparols, C.; Oliveros, E.; Lorente, C. Free Radical Biology and Medicine 2013, 63, 467–475. Casbeer, E.M.; Sharma, V.K; Zajickova, Z.; Dionysiou, D. D. Environ. Sci. Technol. 2013, 47, 4572−4580.

14


Page 15 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(22) Hawe, A.; Wiggenhorn, M.; De Weert, M.V.; Garbe, J.H.O.; Mahler, H.C.; Jiskoot, W. Journal of Pharma. Sciences 2012, 101, 895-913. (23) Hui, A.; Lam, X. M.; Kuel, C.; Grauschopf, U.; Wang, Y. J. PDA J Pharm Sci and Tech 2015, 69, 511-525. (24) Li, Y.; Polozova, A.; Gruia, F.; Feng, J. Anal. Chem. 2014, 86, 6850−6857. (25) Qi, P.; Volkin, D.B.; Zhao, H.; Bond, M. D. Journal of Pharma. Sciences 2008, 98, 3117-3130. (26) Folzer, E.; Diepold, K.; Bomans, K.; Finkler, C.; Schmidt, R.; Bulau, P.; Huwyler, J.; Mahler, HC.; Koulov, A.V. J. Pharm Sci. 2015, 104, 2824-2831. (27) Pavon, J. A.; Li, X.; Chico, S.; Kishnani, U.; Soundararajan, S.; Cheung, J.; Li, H.; Richardson, D.; Shameem, M.; Yang, X. Journal of Chromatography A 2016, 1431, 154–165. (28) Perdivara, I.; Deterding, L. J.; Przybylski, M.; Tomer, K.B. J. Am. Soc. Mass Spectr. 2010, 21, 1114-1117. (29) Li, X.; Xu, W.; Wang, Y.; Zhao, J.; Liu, Yan-Hui; Richardson, D.; Li, H.; Shameem, M.; Yang, X. J. Chromatography A 2016, 1460, 51–60. (30) Jarasch, A.; Koll, H.K.; Regula, J.T.; Bader, M.; Papadimitriou, A.; Kettenberger, H. J. Pharm. Sci. 2015, 104, 1885-1898. (31) Schöneich, C. Journal of Pharmacy and Pharmacology 2018, 70, 6555-6565. (32) Basran, J.; Efimov, I.; Chauhan, N.; Thackray, S.J.; Krupa, J.L.; Eaton, G.; Griffith, G.A.; Mowat, C.G.; Handa, S.; Raven, E. L. J. Am. Chem. Soc. 2011, 133, 16251–16257. (33) Molecular Operating Environment (MOE), 2016.0802; Chemical Computing Group ULC, 1010 Sherbooke St. West, Suite #910, Montreal, QC, Canada, H3A 2R7, 2017. (34) Desmond Molecular Dynamics System, Schrödinger Release 2017-1, D. E. Shaw Research, New York, NY, 2017. MaestroDesmond Interoperability Tools, Schrödinger, New York, NY, 2017. (35) Chennamsetty, N. Journal of Pharmaceutical Sciences 2015, 104, 1246–1255. (36) Jain, T.; Boland, T.; Lilov, A.; Burnina, I.; Brown, M.; Xu, Y., Vasquez, M. Bioinformatics 2017, 33, 3758–3766. (37) Berman, H.M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T.N.; Weissig, H.; Shindyalov, I.N.; Bourne, P. The Protein Data Bank, Nucleic Acids Research 2000, 28, 235-242. (38) Ehrenshaft, M.; Deterding, L. J.; Mason, P. R. Free Radical Biology and Medicine 2015, 89, 220-228. (39) Lodish, H.; Berk, A.; Zipursky, S. L.; et al. Molecular Cell Biology. 4th edition. New York: W. H. Freeman; 2000. Section 3.1, Hierarchical Structure of Proteins. (40) Jain, T.; Sun, T.; Durand, S.; Hall, A.; Houston, R. N. et al. Proc. Natl. Acad. Sci. 2017, 114, 944–949. (41) Yang, R.; Jain, T.; Lynaugh, H.; Nobrega, R. P.; Lu, X.; Boland, T.; et al. mAbs 2017, 9, 646–653.

15



Page 16 of 21

Tables Table 1: Percent oxidation of CDR tryptophan and Fc methionine residues in mAbs A-C by AAPH stress and LC-MS peptide mapping1 (%) oxidation CDR tryptophan2

(%) oxidation Fc methionine3

mAb A control

2.7

7.2

mAb A AAPH 6 hrs

25.9

7.5

mAb A AAPH 6 hrs + tryptophan

2.8

7.5

mAb B control

3.2

3.1

mAb B AAPH 6 hrs

43.0

5.6

mAb B AAPH 6 hrs + tryptophan (no methionine)5

9.0

19.5

mAb B AAPH 24 hrs

74.6

7.9

mAb B AAPH 24 hrs + tryptophan

20.3

6.8

mAb C control

ND

1.5

mAb C AAPH 6 hrs4

ND

29.3

mAb C AAPH 6 hrs + tryptophan

NA

NA

mAb condition*

*For

details on the conditions see materials and methods and results section The level (%) of tryptophan and methionine oxidation of each mAb was calculated by: EIC of the oxidized peptide /(EIC of the oxidized + EIC of the unmodified peptides) x 100%. In the case of tryptophan oxidation, the level of oxidation was a sum of all oxidized species. 2tryptophan residues, CDR3: mAb A Trp , mAb B Trp , CDR1 mAb C Trp 104 102 33 3Methionine residues, Fc CH2: mAb A Met 254, mAb B Met251, mAb C Met250 4For mAb C reaction buffer did not contain L-methionine 5For mAb B, AAPH 6 hour time point + tryptophan was stressed without L-methionine ND: not detected NA: not available 1

16


Page 17 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Table 2: Solvent accessibility calculation of selected tryptophan residues in mAbs A-C. mAb Residue ID Type SASA(Å2) (%) A Trp 104 hyd 120.0 38.6 A Trp 109 hyd 10.2 3.3 B Trp 102 hyd 191.1 61.5 B Trp 109 hyd 32.5 10.5 C Trp 33 hyd 53.6 17.2 C Trp 105 hyd 28.0 9.0

Table 3: Tryptophan oxidation prediction models 2Shell model _8Å

2Shell model _7Å

2Shell model _6Å

2Shell model _5Å

MD_SASA _ave / %_MD_SA SA_ave

mAb

Trp

loc ati on

A

104

H3

34

22

14

8.1 112.0 / 39.0

112.0

2.1

B

102

H3

38

29

18

10.7 164.4 / 57.3

150.8

3.5

33

H1

30

19

10

5.1 64.3 / 22.4

58.7

2.2

C

MD_SA SA_sc

MD_SASA _Natom

MD_SASA_ave, MD-simulation-averaged SASA %_MD_SASA_ave, % of MD-simulation-averaged SASA relative to fully exposed Trp SASA MD_SASA_sc, MD-simulation-averaged SASA of Trp side-chain MD_SASA_Natom, MD-simulation-averaged SASA of N atom of Trp side-chain indole ring

Figure Legends Figure 1: MS/MS Spectra corresponding to HC99-123 +16-Da oxidized peptides of mAb A. Fragmentation shows predominantly y ions series; inset shows low abundance b ions series. Selective y fragment ions are highlighted in red and discussed in the results section. Figure 2: MS/MS Spectra corresponding to HC99-123 +16-Da oxidized peptides of mAb B. Fragmentation shows of predominantly b ions series. Selective b fragment ions are highlighted in red and discussed in the results section. Figure 3: (A) Representative extracted ion chromatogram (EIC) corresponding to peptides containing the tryptophan residue on CDR of mAb B following incubation with 1.2 mM AAPH at 40°C for 24 hours; (B) mAb B, 1.2 mM AAPH for 24 hours with 2.5 mM DL tryptophan. The numbers on the legend note the mass difference from the native peptide. 17



Page 18 of 21

Figure 4: The homology models of the Fv domains of mAbs. The heavy chain is shown in green, the light chain in cyan and tryptophan residues are shown in orange (A) mAb A with Trp104 noted; (B) mAb B with Trp102 noted; (C) mAb C with Trp33 noted.

Figure 5: Relationship between Solvent exposure (%) and total percent of heavy chain CDR tryptophan oxidation for mAbs A-C.

Figures Figure 1:

Insert

18


Page 19 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 2:

Figure 3:

A

B

19



Page 20 of 21

Figure 4

Figure 5 %_MD_SASA_ave

% Oxidation

80 60 40 20 0 mAb-C

mAb-A

mAb-B

20


Page 21 of 21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


AAPH

Trp102 Native peptide


Selective Tryptophan Oxidation of Monoclonal Antibodies: Oxidative

Recommend Documents