https://orcid.org/0000-0002-1882-3182
ABSTRACT
Development of novel bispecific antibody (bsAb) platforms offers unprecedented opportunities for a wide variety of therapeutic applications. However, the expression and manufacturing of bsAbs with desired structures can be challenging. Owing to the uniqueness of each bsAb platform, more comprehensive and customized structural characterization is particularly important to understand the chemical or biological reactivity of bsAbs, as well as to guide process development, risk assessment, and manufacturing. In this work, we performed higher order structure characterization of the Regeneron bsAb platform with Fc site-specific substitutions through hydrogen deuterium exchange mass spectrometry (HDX-MS). Structural deprotection was identified at the CH2-CH3 interface in the Fc domain, owing to the site-specific substitutions. The structural deprotection was found to correlate with the decreased conformational stability of Fc domain. Under oxidative and thermal stress conditions, the Met residues located near the structurally deprotected region were identified to be susceptible to oxidation. In addition, the introduction of substitutions in the bsAb Fc resulted in a slight reduction of its binding affinity to the neonatal Fc receptor (FcRn). The detailed structural elucidation by HDX-MS improves understanding of the structure–property relationship of the Regeneron bsAb format, thus greatly aiding in process development, risk assessment, and manufacturing.
Introduction
Bispecific antibodies (bsAbs), an emerging therapeutic modality, have greatly evolved over the past several decades and garnered tremendous interest for the treatment of numerous diseases, such as cancer, inflammatory disorders, autoimmunity, and infectious diseases.Citation1–7 However, generating bsAbs is challenging, because two asymmetric antigen-binding regions must be assembled with the desired configuration within the immunoglobulin (IgG) architecture. Recent developments in biochemical and genetic engineering techniques have opened a path to using a wide range of recombinant bsAb formats with unique physicochemical and biological properties depending on the molecular design.Citation2,Citation8
One format involves chemically crosslinking two antibodies or antibody subunits to form a desired configuration with bispecificity.Citation9,Citation10 However, several limitations have been identified, including difficulties in manufacturing, propensity for aggregation, insufficient half-lives, and potential safety concerns. An alternative design introduces “knobs into holes” substitutions in the antibody Fc sequence; this scaffold enhances the desired heterodimerization rather than undesired mispairing.Citation11,Citation12 Because this format retains Fc-mediated effector function, the half-life can be prolonged beyond those of non-Fc-containing counterparts. Nevertheless, the unnatural “knobs into holes” substitutions may potentially induce immunogenicity or poor stability of the resulting bsAb, and therefore must be carefully evaluated during drug development and clinical studies. Furthermore, the removal of the two parental homodimers in the downstream purification process can be very difficult because their properties are highly similar to those of the heterodimer. Beyond approaches to promote heterodimerization, another strategy allows for selective purification of the heterodimers of interest from homodimer/heterodimer mixtures. In this format, mouse IgG2a and rat IgG2b antibodies are co-expressed within single cells.Citation13 Owing to the species specificity, each light chain (LC) can properly pair with its corresponding heavy chain (HC). Although two HCs can form dimers indiscriminately, heterodimers can be selectively purified from the mixtures through Protein A chromatography according to their differential binding affinities. However, the immunogenicity risk generated from the mouse IgG protein still poses challenges in the clinical application of this design.
To address the aforementioned limitations, we previously developed a unique format with site-specific substitutions in the Fc region, which allows for selective separation of the heterodimers of interest. In this design, the key 435His-Tyr436 residues on one of the HC of human IgG4 were substituted with Arg-Phe (resulting in an HC* chain with a human IgG3 sequence), and both HC and HC* contain different VH domains with differing specificity. To avoid the issue of HC-LC mispairing, we used an identical LC ().Citation14 This bsAb format was designed as a fully human antibody that can be purified downstream through differential Protein A affinity chromatography.Citation15 Because the substitutions reside at the antibody CH2-CH3 interface, a critical region that determines the Fc properties of antibodies, in this work, we characterized the higher order structure of bsAb through hydrogen deuterium exchange mass spectrometry (HDX-MS). By comparing the deuterium uptake profiles of peptides from the bsAb heterodimer and two other corresponding homodimers (HC/HC and HC*/HC*), we identified potential differences in folding structures and local dynamics of specific sequences at the region containing the site-specific substitutions. The effects of these structural differences were evaluated, including Fc domain thermostability, Met oxidation near the substitution sites, and the profiles of binding to FcRn. Subsequently, we confirmed our findings by using a series of Regeneron (REGN) bsAbs with the same site-specific substitutions. The detailed structural elucidation revealed by this work improves fundamental understanding of the Fc properties of the REGN bsAb format, thus providing guidance for process development, risk assessment, and manufacturing.
Figure 1. Schematic of the REGN bsAb format with site-specific substitutions at the Fc domain. The star represents the position of HY to RF substitutions in the three-dimensional structure of REGN bsAbs. EU numbering was used to label antibody sequences.
Results
Structural differences revealed by HDX-MS
To evaluate the effects of site-specific substitutions in the Fc domain, we performed HDX-MS on HC/HC* heterodimers of REGN bsAb-1 and the two parental HC/HC and HC*/HC* homodimers. After pepsin digestion, 98.7%, 99.6% and 100% sequence coverage was achieved for HC, HC*, and LC, respectively. Statistical testing was then performed to reveal the peptide sequences showing significantly different deuterium uptake profiles (Figure S1). As shown in the homology model in , three regions showed significantly different deuterium uptake at the Fc domain in the HC* containing site-specific substitutions (labeled with a side chain structure). The first region, with peptide sequence 241FLFPPKPKDTLM252, resides in the lower CH2 region, which is in close contact with either the original 435HY436 or mutated 435RF436 residues in the upper CH3 region. (EU numbering was used throughout this study.) The HDX kinetic plot of this peptide is shown in . At earlier time points, deuterium uptake showed a trend of HC*/HC* > HC/HC* > HC/HC, whereas the uptake in all three molecules converged at later time points (e.g., 14400 s). Of note, for the heterodimer containing HC and HC*, only half the peptide population was affected by the substitutions. Interestingly, we observed that the kinetics plot for the heterodimer was indeed intermediate between those of the two homodimers. Overall, these results indicated that the site-specific substitutions might have led to increased dynamics of the 241FLFPPKPKDTLM252 peptide, without affecting the solvent accessibility or folding of this peptide.
Figure 2. (a) Structural homology of the Fc domain of REGN bsAb-1. Regions showing statistically significant differences in deuterium uptake (HC* > HC) are in red. Regions with unchanged deuterium uptake after substitutions are in gray. Representative HDX-MS kinetic plots are shown for (b) CH peptide 241FLFPPKPKDTLM252, (c) CH peptide 427VMHEALHNHYTQK439 (for HC) and 427VMHEALHNRFTQK439 (for HC*) and (d) CH peptide 298STYRVVSVL306. The means and error bars (standard deviations) are based on triplicated experiments.
The second region showing significant uptake differences is located at the CH3 domain, which contains the mutated residues. Peptides with RF substitutions (red trace) showed higher deuterium uptake than the original peptide with HY residues (blue trace) at all time points (). These findings may suggest that this peptide folds differently because of the substitutions, thus increasing the structure’s solvent exposure. Interestingly, the deuterium uptake profiles of these two heterodimeric peptides perfectly aligned with those of their corresponding homodimeric peptides, thus further validating the differences in local structure between HC and HC* within this asymmetric bsAb format.
According to the literature,Citation16–21 the deuterium uptake of the aforementioned CH2-CH3 interface regions is sensitive to Fc glycosylation and Met oxidation. To confirm whether this structural difference was indeed caused by the substitutions, we evaluated the levels of glycosylation and Met oxidation of the three tested molecules through reduced peptide mapping. Similar oxidation levels were observed for both Met at the CH2-CH3 interface (Table S1), thus ruling out the possibility of oxidation. Meanwhile, comparable glycosylation profiles were also obtained across all tested molecules, except that slightly larger differences (~5%) were observed in the non-glycosylated form (Figure S2). Lower glycan occupancy usually results in higher deuterium uptake for certain peptides that interact with glycan. The HC*/HC* homodimer showed higher levels of aglycosylation, thus possibly explaining the slightly higher level of uptake for peptides 298STYRVVSVL306 next to the Asn297 glycosylation site (). However, because higher aglycosylation contributed to the greater uptake for peptides in , we performed HDX analysis for all three deglycosylated molecules, in which the deuterium exchange was independent of the effects of glycans. As shown in Figure S3, similar trends were observed in deglycosylated samples. Therefore, the structural deprotection identified in HC* appears to be caused primarily by the site-specific substitutions.
Effects on domain stability
The HDX analysis demonstrated that the CH2-CH3 interface of HC* exhibited structural deprotection, owing to the site-specific substitutions. Both the increased conformational dynamics of CH2 peptide 241FLFPPKPKDTLM252 and the higher solvent-accessibility of CH3 peptide 427VMHEALHNRFTQK439 (for HC*) could potentially influence the thermostability of their corresponding domains. To eliminate the potential impacts from Fab to the thermal stability of Fc, the Fc subunits of bsAb-1 HC/HC* heterodimers with the HC/HC and HC*/HC* homodimers were generated by FabRICATOR® digestion and purified by size exclusion chromatography (SEC). However, as the interactions between the two IgG4 CH3 domains are relatively weakerCitation22 and both chains in Fc lack hinge disulfide bonds to lock them in place, the resulting Fc of HC/HC* heterodimer undergoes chain exchange. This ultimately may lead to the formation of low-abundance HC/HC and HC*/HC* homodimers again, even after SEC purification.
Capillary differential scanning calorimetry (DSC) was implemented to examine the thermal unfolding of all three corresponding Fc-only samples, indicated by the melting onset temperature (Tonset) and the melting transitions Tm1 and Tm2 for CH2 and CH3 domains, respectively. The dotted traces in represent the domain-specific thermal transitions after cumulative fittings. Both the Tm of CH2 and CH3 domains are lower in HC*/HC* homodimer than in HC/HC homodimer. The heterodimer shows two thermal transitions that are between the two homodimers, respectively ( and Figure S4). The observed values of Tm1 (~65.0°C) and Tm2 (~68.0°C) are similar to the literature reported values for the Fc domain of the IgG4 subclass.Citation23 The same trend is also observed with intact HC/HC* heterodimers and two intact homodimers. The heterodimer shows two thermal transitions that can correspond to the Tm from single HC and single HC* respectively (Figure S5). Overall, the local structural deprotection caused by site-specific substitutions can positively correlate with the Fc domain destabilization. This correlation serves as another example demonstrating the use of HDX-MS to probe the local folding conformation and dynamics of antibodies to the corresponding global thermostability, which is consistent with other works using different antibody formats.Citation24–27
Figure 3. DSC thermograms of Fc subunit of (a) HC/HC homodimer, (b) HC/HC* heterodimer, (c) HC*/HC* homodimer of REGN bsAb-1 in 150 mM ammonium acetate, pH 6.0. Black trace represents the experimental data. The blue dotted traces represent the thermal transition fits and the red trace represents the cumulative fit data.
Table 1. Thermal stability of Fc subunit samples indicated by the melting onset temperature (Tonset) and the melting transitions (Tm and Tm) as measured by DSC. The mean and standard deviations (SD) are based on triplicate measurements.
Effects on Met oxidation
Due to the site-specific substitutions, the increased local solvent-accessibility and flexibility of CH2-CH3 interface region may also affect the post translational modifications (PTMs) of certain nearby residues. To investigate this further, we aimed to determine if structural deprotection could increase the susceptibility of two Met residues (HC Met252/HC* Met252 and HC Met428/HC* Met428) located at the CH2-CH3 interface to oxidation. We first subjected the heterodimer and two homodimers of bsAb-1 to oxidative stress using H2O2 at room temperature and quantified the relative abundance of oxidation by analyzing the peak areas of the extracted ion chromatograms (EICs) in reduced peptide mapping. To better compare the rate of oxidation across the three molecules, the levels of Met oxidation generated at different time points were normalized by subtracting the initial Met oxidation level for the same Met residue in each sample prior to incubation.
Oxidative stress induced by H2O2 allows for the assessment of a protein’s local conformation and dynamics by specifically measuring the oxidation level of Met side chain, while HDX measures the deuterium uptake on peptide backbones. Given that both assays are conducted at similar temperatures, we anticipate that they may provide similar information on protein conformation and dynamics. In , the rates of oxidation for HC Met252/HC* Met252 under H2O2 stress are illustrated, with comparable levels observed across all samples. These oxidation rates could possibly indicate the dynamics of the Met-containing peptide over time periods of 30, 60, and 120 minutes. These rates are consistent with the kinetics revealed by HDX at later time points (1800s and 14400s), as shown in . Notably, HDX can provide peptide dynamics information at shorter time periods more readily. In , a stronger correlation between deuterium uptakes and Met oxidation rates of HC Met428/HC* Met428 is demonstrated, with a higher oxidation rate in HC*/HC* homodimers compared to HC/HC. This suggests that the Met residue is more susceptible to oxidation in the structurally exposed HC* than in HC. Remarkably, the oxidation rate of Met in the heterodimer is precisely aligned with its corresponding Met in the homodimers, in agreement with what has been observed in HDX kinetic plot in . Overall, our findings demonstrate that local structural deprotection may increase the susceptibility of Met residues to oxidation, and that both Met oxidation by H2O2 and HDX may offer complementary information on the conformational and dynamic properties of proteins.
Figure 4. Quantification of (a) HC Met252/HC* Met252 and (b) HC Met428/HC* Met428 oxidation under H2O2 stress; (c) HC Met252/HC* Met252 and (d) HC Met428/HC* Met428 oxidation under thermal stress in bsAb-1 heterodimer and homodimers. The levels of Met oxidation generated in stressed samples at different time points were normalized by subtracting the initial Met oxidation level for the same Met residue in each sample prior to incubation. The means and error bars (standard deviations) are based on triplicated experiments.
In addition to investigating the impact of oxidative stress, it is also of interest to explore how accelerated stability conditions affect the oxidation of Met in bsAbs. These stability tests could potentially predict the bsAb’s stability and determine its shelf-life and storage conditions. To assess this effect, we incubated the heterodimer and two homodimers of bsAb-1 at 45°C for 7 days and 14 days, and monitored the oxidation rate of methionine. As compared to oxidative stress, the oxidation rate of methionine under thermal stress was significantly slower due to the limited presence of reactive oxygen species (ROS) in the incubation solutions. Moreover, the rate of Met oxidation under thermal stress could serve as an indicator of the conformation and dynamics of the CH2-CH3 interface region at the particular temperature of the thermal stress. shows the oxidation rates of HC Met252/HC* Met252 under thermal stress, which followed the pattern of HC*/HC* > HC/HC* > HC/HC. Interestingly, the oxidation trend was positively correlated with the trend in backbone dynamics of the peptide containing this Met residue at room temperature (), thus indicating that the Met within the more dynamic environment was possibly more prone to oxidation. demonstrates a similar correlation for the HC Met428/HC* Met428 at the CH2-CH3 interface, where the rate of oxidation for homodimers was HC/HC < HC*/HC×. This observation is consistent with the results obtained from oxidative stress experiments, which showed that the Met residue was more vulnerable to oxidation in structurally exposed HC* than in HC (). This finding is also in line with existing literature that highlights protein folding stabilities as crucial factors influencing the rates of methionine residue oxidation.Citation28,Citation29
Furthermore, we observed that the rate of HC Met428/HC* Met428 oxidation in heterodimers did not entirely correspond to the deuterium uptake for this Met-containing peptide. Although HDX and oxidative stress revealed that the HC from both homodimers and heterodimers possessed the same folding structure and backbone dynamics at room temperature (), HC Met428 remained more susceptible to oxidation in the HC/HC* heterodimer than the HC/HC homodimer under thermal stress. Similarly, HC* Met428 remained less susceptible to oxidation in the HC/HC* heterodimer than the HC*/HC* homodimer. Therefore, we speculated that the thermal stability of HC could induce a change in the thermal stability of HC* through an interplay between the two monomers within the heterodimer, and vice versa. This may ultimately lead to slightly different reactivity of the corresponding Met when each monomer is present in the heterodimer compared to the homodimer under thermal stress conditions.
To validate this finding, we conducted further experiments on several other REGN bsAbs with the same site-specific substitutions and comparable glycosylation profiles and Met oxidation at T0, as outlined in Table S2. To avoid any potential bias that may arise from Fab sequences, we selected other molecules with different Fab sequences. Compared to the regular IgG4 mAb 1–3 (lacking HY to RF substitutions), bsAbs 2–4 all exhibited significantly higher deuterium uptake for both the peptides identified in bsAb-1 (), while the uptake was relatively consistent for all other peptides in the Fc domain (Figure S6). Furthermore, we evaluated the rates of Met oxidation for this cohort of antibodies. In line with our previous observations, we found that bsAbs were more prone to the oxidation of the HC Met252/HC* Met252 residue when compared to regular mAbs (). Similarly, faster oxidation rates were observed for the HC Met428/HC* Met428 residue in structurally exposed HC* than in HC (red solid symbols vs. blue solid symbols in ). Additionally, slightly slower oxidation rates were observed for this HC Met428/HC* Met428 residue in regular mAbs when compared to HC from various bsAbs (blue empty symbols vs. blue solid symbols in ), which aligns with our previous findings in . It should be noted that, despite having similar conformation and dynamics, the rates of Met oxidation can still vary. Other factors, including N-glycosylation, Fab sequences, formulation, and others, can also impact Met oxidation. Overall, our findings demonstrate that Fc site-specific substitutions resulting in structural deprotection render Met more susceptible to oxidation at the CH2-CH3 interface. Therefore, careful characterization during bsAb development and production is crucial.
Figure 5. (a) Significant differences in deuterium uptake for the CH peptide 241FLFPPKPKDTLM252 between mAb 1–3 and bsAb 2–4. (b) Normalized oxidation level of HC Met252/HC* Met252 in thermal stressed samples. (c) Deuterium uptake of the CH peptide 427VMHEALHNHYTQK439 with or without HY to RF substitutions. (d) Normalized oxidation level of HC Met252/HC* Met252 in thermal stressed samples. The means and error bars (standard deviations) are based on triplicated experiments. Statistical testing was demonstrated in Figure S7.
Effects on FcRn binding
Numerous studies have demonstrated that the CH2-CH3 interface region participates in IgG Fc-FcRn binding, in which the strength of FcRn binding plays a crucial role in determining the serum half-life of therapeutic antibodies.Citation30–35 To investigate the impact of asymmetric structures at the CH2-CH3 interface region in HC and HC* on FcRn binding, we compared the FcRn binding profiles of HC/HC* heterodimers and two homodimers of REGN bsAb-1. Bio-layer interferometry (BLI) was used to determine the FcRn binding affinities, and to eliminate any potential interference from the Fab region, the Fc subunits of each sample were used. As depicted in and summarized in , the binding affinities (KD) and maximum binding signal (Rmax) values demonstrated a trend of binding strength for HC/HC > HC*/HC*, which is consistent with previous literature that IgG4 binds to FcRn stronger than IgG3.Citation36 According to previous studies, FcRn may independently bind both chains of IgG Fc,Citation37 which suggests that the appearance of HC* could negatively impact the binding strength. Indeed, we found that the KD of HC/HC* falls between the KD values of the two homodimers.
Figure 6. BLI sensorgrams of FcRn binding to Fc subunit of (a) HC/HC homodimer, (b) HC/HC* heterodimer, (c) HC*/HC* homodimer of bsAb-1.
Table 2. Steady state analysis of the bindings between human FcRn and REGN bsAb-1 Fc subunit samples.
Discussion
Here, we performed HDX-MS characterization of REGN bsAbs with Fc site-specific substitutions. By comparing the deuterium uptake plots between the HC/HC* heterodimer and the HC/HC and HC*/HC* homodimers, we identified structural deprotection at the CH2-CH3 interface region in HC* containing the substitutions. The structural deprotection was confirmed to be attributable primarily to the site-specific substitutions rather than to glycosylation and Met oxidation. Such amino acid substitutions have been reported to induce local conformational and dynamic changes in other engineered antibodies as well.Citation26,Citation35 Furthermore, structural changes resulting from site-specific substitutions and their subsequent effects on the properties of the Fc domain were carefully evaluated.
The more exposed folding structure and the increased local dynamics of specific sequences observed in HDX analysis could potentially be associated with a decrease in the global thermostability of the corresponding Fc domain. Our findings demonstrate that local structural differences identified through HDX prior to stress conditions play a significant role in determining oxidation rate of specific Met under both oxidative and thermal stress conditions. Of particular note, we observed that Met residues located at the structurally deprotected CH2-CH3 interface showed higher susceptibility to oxidation, as further confirmed in a collection of bispecific antibodies (bsAbs) with the same format. Moreover, our findings suggest the presence of interplay between the heavy chain (HC) and its variant (HC*) in our asymmetric bsAb format. This HC-HC* cross-talk also contributes to modulating the rate of Met oxidation, particularly under thermal stress conditions, where its impact becomes more pronounced. In addition, our findings suggest that the FcRn binding affinity of the bsAb Fc falls between that of the regular IgG4 Fc (HC/HC Fc) and IgG3 Fc (HC*/HC* Fc), indicating that site-specific substitutions can weaken the binding. The results from HDX-MS provide valuable insights into the structural and conformational dynamics of the REGN bsAb format, enhancing our understanding of its biological activity and offering guidance for the development of drug products.
Materials and methods
Materials
All reagents were commercially available and used as received unless stated otherwise. The chromatography solvents were of LC/MS grade and were obtained from Thermo Fisher Scientific (Waltham, MA). Monoclonal antibodies (bsAbs 1–4 and mAbs 1–3) were produced by Regeneron (Tarrytown, NY). Deuterium oxide (D2O), sodium phosphate, sodium chloride, guanidine hydrochloride, tris(2-carboxyethyl)-phosphine hydrochloride (TCEP-HCl), iodoacetamide and urea were purchased from Sigma-Aldrich (St. Louis, MO).
HDX-MS experiments
Labeling was performed on a LEAP PAL3 HDX automation system (Trajan Scientific and Medical, Morrisville, NC). Samples comprising 8 μL of 3 mg/mL mAb were labeled in 72 μL labeling buffer (1× phosphate-buffered saline (PBS) in D2O, pH = 7.4) three times for each indicated labeling time (30, 240, 1,800, and 14,400 s) at 25°C. After labeling, 50 μL of each labeled sample was quenched with 50 μL of quenching buffer (200 mM sodium phosphate, 4 M guanidine hydrochloride and 500 mM TCEP, pH = 2.3, in water) at 1°C. After quenching, 50 μL of the sample was injected into a chromatography system within a cold box connected to a Waters Acquity UPLC instrument (Waters, Milford, MA). The cold box and LC solvent precooler were maintained at 0°C for all experiments. Injected samples were passed over an immobilized pepsin column (2.1 × 30 mm, NovaBioAssays, Woburn, MA) at 100 μL/min for 120 s with mobile phase A (95% H2O, 5% acetonitrile and 0.5% formic acid). The resulting peptic peptides were captured and desalted on a SymmetryShield C8 trap column (3.9 × 20 mm, Waters, Mildford, MA). The digested peptides were separated on an Acquity BEH C18 column (2.1 × 50 mm, 1.7 μm, 130 Å, Waters, Milford, MA) over a 25 min linear gradient of mobile phase B (0.1% formic acid in acetonitrile) from 0.1% to 30.0%. A Thermo Q ExactiveTM Plus mass spectrometer was used to measure the mass of deuterium-labeled peptides.
HDX-MS data analysis
Raw files of both native and deuterated REGN antibody samples were processed in the Protein Metrics Byos® with HDX workflow followed by in-house data refinement software that can select high quality peptide with a redundancy value close to 4. The refined list of peptides was manually validated in Protein Metrics software, and corrections to chromatographic peak integration limits were applied if necessary. After data processing, a “volcano plot” statistical testingCitation38 was applied to identify peptides with significantly different deuterium uptake for each labeling time.
Preparation of Fc subunits of antibodies
REGN bsAb-1 samples were diluted to a concentration of 10 mg/mL in 1× PBS buffer at pH = 7.4. Each sample was then digested with Genovis FabRICATOR enzyme at a ratio of 1 unit/μg of substrate at a temperature of 37°C for 2 hours. After digestion, the resulting F(ab’)2 and Fc subunits were separated using a Waters ACQUITY UPLC BEH SEC column (1.7 µm, 4.6 mm × 300 mm) on a Waters ACQUITY UPLC system. 300 µg of the sample was injected per run, and multiple injections were performed to obtain sufficient amounts of the samples for other assays. The mobile phase consisted of 150 mM ammonium acetate at pH = 6.0, and the flow rate was set to 0.2 mL/minute.
Differential scanning calorimetry
All samples were diluted to 1 mg/mL in 150 mM ammonium acetate, pH = 6.0 and were performed in triplicates in the PEAQ-DSC instrument (MicroCal, Northampton, MA). All samples were subjected to a thermal ramp of 90°C/hr from 15°C to 110°C with a filtering period of 10 seconds in “no feedback” mode. Each thermogram was background subtracted, fitted with a linear baseline, and analyzed using the PEAQ-DSC software. Melting temperatures were averaged and standard deviations were calculated from triplicate measurements.
Thermal and H2O2 stress
All REGN bsAbs/mAbs drug substances were adjusted to 20 mg/mL with 10 mM histidine at pH = 6.0 for thermal stress. Samples were incubated at 45°C for 7 and 14 days. After thermal stress, each REGN antibody sample was buffer exchanged to 5 mM acetic acid. The REGN bsAbs-1 samples were adjusted to 4 mg/mL with 10 mM histidine at pH = 6.0 for H2O2 stress. Samples were incubated with 0.05% of H2O2 for 30, 60, and 120 minutes at room temperature. Following H2O2 stress, each sample was subject to buffer exchange into 5 mM acetic acid using Cytiva MicroSpinTM G-25 spin columns. During buffer exchange, excess amounts of H2O2 can be removed from the solution.
Tryptic digestion
Sample preparation for peptide mapping was implemented on a Beckman Coulter Biomek i5 automation system (Brea, CA). An aliquot of 100 µg of each sample was denatured and reduced in 40 μL solution containing 5 mM acetic acid and 5 mM TCEP-HCl by heating at 70°C for 10 minutes. After cooling to room temperature, samples were adjusted to pH = 8.0 by addition of 82 μL of H2O and 24 μL of 1 M Tris-HCl. Each sample was then alkylated with 4 mM iodoacetamide and digested with trypsin at an enzyme-to-substrate ratio of 1:20 (w/w) at 30°C for 140 minutes. After digestion, 56 μL of 8 M urea was added to the samples to maintain hydrophobic peptides in solution. Finally, 4.7 μL of 20% trifluoroacetic acid (TFA) was added to acidify the samples before MS analysis.
Peptide mapping and data analysis
For peptide mapping, peptide mixtures generated by trypsin digestion were separated with a Waters ACQUITY UPLC BEH C18 column (1.7 µm, 2.1 mm × 150 mm) on a Waters ACQUITY UPLC system at a flow rate of 0.25 mL/minute. The column was equilibrated with 99.9% mobile phase A (0.05% TFA in water) before sample injection and maintained at 40°C. Subsequently, 5 μg of each tryptic digest sample was injected onto the column. Peptides were separated with a gradient profile held at 0.1% mobile phase B (0.045% TFA in acetonitrile) for the first 5 minutes, then increased from 0.1% to 40% mobile phase B over the next 80 minutes. A Thermo Q Exactive mass spectrometer was used for peptide mass analyses, and higher-energy collisional dissociation was used for peptide fragmentation for MS/MS experiments. Byonic™ was used to aid in peptide identification. Skyline software was used for the extracted ion chromatogram peak area calculations. The relative abundance of a PTM was determined by calculating of the ratio of the extracted ion chromatogram (EIC) peak area of the peptide containing the PTM to the sum of the EIC peak areas of the corresponding native peptide and the peptide containing the PTM.
Bio-layer interferometry
FcRn binding affinities to antibodies were determined through bio-layer interferometry. Biotinylated human FcRn receptor was captured on streptavidin-coated biosensor surfaces, and the biosensor was stabilized with binding buffer (100 mM sodium phosphate, 150 mM NaCl, and 0.05% (v/v) surfactant PS20, pH = 6.0). Subsequently, biotinylated human FcRn captured biosensors were dipped into assay plates containing antibody Fc subunit samples prepared at concentrations ranging from 2.5 µM to 0.078 µM, then shaken for 2 minutes at 1,000 rpm. The dissociation of bound antibody Fc subunit samples from the biosensors was conducted in binding buffer and monitored for 1 minute.
The sensorgram of a biosensor dipped in binding buffer in the absence of each antibody Fc subunit sample was subtracted from the binding sensorgrams to remove binding signal changes due to the dissociation of the captured biotinylated human FcRn receptor from the biosensors. Biosensors with no biotinylated human FcRn receptor captured were dipped into antibody Fc subunit samples prepared at concentrations ranging from 2.5 µM to 0.078 µM as a control. The resultant sensorgrams were subtracted from the binding sensorgrams to remove binding signal changes due to nonspecific binding to the biosensors. Owing to the rapid association and dissociation of antibodies to FcRn, the sensorgrams could not be fitted for kinetic analysis. Therefore, the resultant sensorgrams were subjected to steady state analysis, and the affinity constant (KD) values were calculated.
Abbreviations
| bsAb | = | Bispecific antibody |
| BLI | = | Bio-layer interferometry |
| CH2 | = | Second constant Ig domain of the heavy chain |
| CH3 | = | Third constant Ig domain of the heavy chain |
| DSC | = | Differential scanning calorimetry |
| EIC | = | Extracted ion chromatogram |
| Fab | = | Fragment antigen-binding region |
| Fc | = | Fragment crystallizable region |
| FcRn | = | Neonatal Fc receptor |
| HC | = | Heavy chain |
| HDX-MS | = | Hydrogen deuterium exchange mass spectrometry |
| IgG | = | Immunoglobulin |
| KD | = | Binding affinity constant |
| LC | = | Light chain |
| mAb | = | Monoclonal antibody |
| PTMs | = | Post translational modifications |
| REGN | = | Regeneron |
| Rmax | = | Maximum binding signal |
| ROS | = | Reactive oxygen species |
| SD | = | Standard deviations |
| SEC | = | Size exclusion chromatography |
| Tm | = | Melting temperature |
| Tonset | = | Melting onset temperature |
Supplemental Material
Download PDF (671 KB)Disclosure statement
No potential conflict of interest was reported by the authors.
Supplementary material
Supplemental data for this article can be accessed online at https://doi.org/10.1080/19420862.2023.2228006
Correction Statement
This article has been corrected with minor changes. These changes do not impact the academic content of the article.
Additional information
Funding
References
- Brinkmann U, Kontermann RE. Bispecific antibodies: bispecific antibodies have emerged as molecules with a multitude of talents. Science. 2021;372(6545):916–10. doi:10.1126/science.abg1209.
- Labrijn AF, Janmaat ML, Reichert JM, Parren PW. Bispecific antibodies: a mechanistic review of the pipeline. Nat Rev Drug Discov. 2019;18(8):585–608. doi:10.1038/s41573-019-0028-1.
- Kufer P, Lutterbüse R, Baeuerle PA. A revival of bispecific antibodies. Trends Biotechnol. 2004;22(5):238–44. doi:10.1016/j.tibtech.2004.03.006.
- Bargou R, Leo E, Zugmaier G, Klinger M, Goebeler M, Knop S, Noppeney R, Viardot A, Hess G, Schuler M, et al. Tumor regression in cancer patients by very low doses of a T cell–engaging antibody. Science. 2008;321(5891):974–77. doi:10.1126/science.1158545.
- Van Der Neut Kolfschoten M, Schuurman J, Losen M, Bleeker WK, Martínez-Martínez P, Vermeulen E, Den Bleker TH, Wiegman L, Vink T, Aarden LA, et al. Anti-inflammatory activity of human IgG4 antibodies by dynamic Fab arm exchange. Science. 2007;317(5844):1554–57. doi:10.1126/science.1144603.
- Van Roy M, Ververken C, Beirnaert E, Hoefman S, Kolkman J, Vierboom M, Breedveld E, T Hart B, Poelmans S, Bontinck L, et al. The preclinical pharmacology of the high affinity anti-IL-6R Nanobody® ALX-0061 supports its clinical development in rheumatoid arthritis. Arthritis Res Ther. 2015;7(1):1–6. doi:10.1186/s13075-015-0651-0.
- Khan SN, Sok D, Tran K, Movsesyan A, Dubrovskaya V, Burton DR, Wyatt RT, Silvestri G. Targeting the HIV-1 spike and coreceptor with bi-and trispecific antibodies for single-component broad inhibition of entry. J Virol. 2018;92(18):e00384–18. doi:10.1128/JVI.00384-18.
- Brinkmann U, Kontermann RE. The making of bispecific antibodies. MAbs. 2017;9(2):182–212. doi:10.1080/19420862.2016.1268307.
- Ellerman D, Scheer JM. Generation of bispecific antibodies by chemical conjugation. In: Bispecific Antibodies. Berlin: Springer Berlin Heidelberg; 2011. pp. 47–63. doi:10.1007/978-3-642-20910-9_3.
- Doppalapudi VR, Huang J, Liu D, Jin P, Liu B, Li L, Desharnais J, Hagen C, Levin NJ, Shields MJ, et al. Chemical generation of bispecific antibodies. Proc Natl Acad Sci. 2010;107(52):22611–16. doi:10.1073/pnas.1016478108.
- Ridgway JB, Presta LG, Carter P. ‘Knobs-into-holes’ engineering of antibody CH3 domains for heavy chain heterodimerization. Protein Eng Des Sel. 1996;9(7):617–21. doi:10.1093/protein/9.7.617.
- Spiess C, Bevers J, Jackman J, Chiang N, Nakamura G, Dillon M, Liu H, Molina P, Elliott JM, Shatz W, et al. Development of a human IgG4 bispecific antibody for dual targeting of interleukin-4 (IL-4) and interleukin-13 (IL-13) cytokines. J Biol Chem. 2013;288(37):26583–93. doi:10.1074/jbc.M113.480483.
- Lindhofer H, Mocikat R, Steipe B, Thierfelder S. Preferential species-restricted heavy/light chain pairing in rat/mouse quadromas. Implications for a single-step purification of bispecific antibodies. J Immunol. 1995;155(1):219–25. doi:10.4049/jimmunol.155.1.219.
- Smith EJ, Olson K, Haber LJ, Varghese B, Duramad P, Tustian AD, Oyejide A, Kirshner JR, Canova L, Menon J, et al. A novel, native-format bispecific antibody triggering T-cell killing of B-cells is robustly active in mouse tumor models and cynomolgus monkeys. Sci Rep. 2015;5(1):1–2. doi:10.1038/srep17943.
- Tustian AD, Endicott C, Adams B, Mattila J, Bak H. Development of purification processes for fully human bispecific antibodies based upon modification of protein a binding avidity. MAbs. 2016;8(4):828–38. doi:10.1080/19420862.2016.1160192.
- Houde D, Peng Y, Berkowitz SA, Engen JR. Post-translational modifications differentially affect IgG1 conformation and receptor binding. Mol Cell Proteom. 2010;9(8):1716–28. doi:10.1074/mcp.M900540-MCP200.
- Burkitt W, Domann P, O’Connor G. Conformational changes in oxidatively stressed monoclonal antibodies studied by hydrogen exchange mass spectrometry. Protein Sci. 2010;19(4):826–35. doi:10.1002/pro.362.
- Zhang A, Hu P, MacGregor P, Xue Y, Fan H, Suchecki P, Olszewski L, Liu A. Understanding the conformational impact of chemical modifications on monoclonal antibodies with diverse sequence variation using hydrogen/deuterium exchange mass spectrometry and structural modeling. Anal Chem. 2014;86(7):3468–75. doi:10.1021/ac404130a.
- Wei B, Gao X, Cadang L, Izadi S, Liu P, Zhang HM, Hecht E, Shim J, Magill G, Pabon JR, et al. Fc galactosylation follows consecutive reaction kinetics and enhances immunoglobulin G hexamerization for complement activation. MAbs. 2021;13(1):1893427. doi:10.1080/19420862.2021.1893427.
- Kuhne F, Bonnington L, Malik S, Thomann M, Avenal C, Cymer F, Wegele H, Reusch D, Mormann M, Bulau P. The impact of immunoglobulin G1 Fc sialylation on backbone amide H/D exchange. Antibodies. 2019;8(4):49. doi:10.3390/antib8040049.
- Pan J, Zhang S, Chou A, Borchers CH. Higher-order structural interrogation of antibodies using middle-down hydrogen/deuterium exchange mass spectrometry. Chem Sci. 2016;7(2):1480–86. doi:10.1039/C5SC03420E.
- Rispens T, Davies AM, Ooijevaar-de Heer P, Absalah S, Bende O, Sutton BJ, Vidarsson G, Aalberse RC. Dynamics of inter-heavy chain interactions in human immunoglobulin G (IgG) subclasses studied by kinetic Fab arm exchange. J Biol Chem. 2014;289(9):6098–109. doi:10.1074/jbc.M113.541813.
- Kang H, Larson NR, White DR, Middaugh CR, Tolbert T, Schöneich C. Effects of glycan structure on the stability and receptor binding of an IgG4-Fc. J Pharm Sci. 2020;109(1):677–89. doi:10.1016/j.xphs.2019.10.036.
- Majumdar R, Manikwar P, Hickey JM, Samra HS, Sathish HA, Bishop SM, Middaugh CR, Volkin DB, Weis DD. Effects of salts from the Hofmeister series on the conformational stability, aggregation propensity, and local flexibility of an IgG1 monoclonal antibody. Biochemistry. 2013;52(19):3376–89. doi:10.1021/bi400232p.
- Manikwar P, Majumdar R, Hickey JM, Thakkar SV, Samra HS, Sathish HA, Bishop SM, Middaugh CR, Weis DD, Volkin DB. Correlating excipient effects on conformational and storage stability of an IgG1 monoclonal antibody with local dynamics as measured by hydrogen/deuterium-exchange mass spectrometry. J Pharm Sci. 2013;102(7):2136–51. doi:10.1002/jps.23543.
- Majumdar R, Esfandiary R, Bishop SM, Samra HS, Middaugh CR, Volkin DB, Weis DD. Correlations between changes in conformational dynamics and physical stability in a mutant IgG1 mAb engineered for extended serum half-life. MAbs. 2015;7(1):84–95. doi:10.4161/19420862.2014.985494.
- Toth IR, Pace SE, Mills BJ, Joshi SB, Esfandiary R, Middaugh CR, Weis DD, Volkin DB. Evaluation of hydrogen exchange mass spectrometry as a stability-indicating method for formulation excipient screening for an IgG4 monoclonal antibody. J Pharm Sci. 2018;107(4):1009–19. doi:10.1016/j.xphs.2017.12.009.
- Walker EJ, Bettinger JQ, Welle KA, Hryhorenko JR, Vargas AM, O’Connell MR, Ghaemmaghami S. Protein folding stabilities are a major determinant of oxidation rates for buried methionine residues. J Biol Chem. 2022;298(5):101872. doi:10.1016/j.jbc.2022.101872.
- Walker EJ, Bettinger JQ, Welle KA, Hryhorenko JR, Ghaemmaghami S. Global analysis of methionine oxidation provides a census of folding stabilities for the human proteome. Proc Natl Acad Sci. 2019;116(13):6081–90. doi:10.1073/pnas.1819851116.
- Roopenian DC, Akilesh S. FcRn: the neonatal Fc receptor comes of age. Nat Rev Immunol. 2007;7(9):715–25. doi:10.1038/nri2155.
- Martin WL, West AP Jr, Gan L, Bjorkman PJ. Crystal structure at 2.8 Å of an FcRn/heterodimeric Fc complex: mechanism of Ph-dependent binding. Mol Cell. 2001;7(4):867–77. doi:10.1016/S1097-2765(01)00230-1.
- Huang X, Zheng F, Zhan CG. Binding structures and energies of the human neonatal Fc receptor with human Fc and its mutants by molecular modeling and dynamics simulations. Mol Biosyst. 2013;9(12):3047–58. doi:10.1039/c3mb70231f.
- Bertolotti-Ciarlet A, Wang W, Lownes R, Pristatsky P, Fang Y, McKelvey T, Li Y, Li Y, Drummond J, Prueksaritanont T, et al. Impact of methionine oxidation on the binding of human IgG1 to FcRn and Fcγ receptors. Mol Immunol. 2009;46(8–9):1878–82. doi:10.1016/j.molimm.2009.02.002.
- Mo J, Yan Q, So CK, Soden T, Lewis MJ, Hu P. Understanding the impact of methionine oxidation on the biological functions of IgG1 antibodies using hydrogen/deuterium exchange mass spectrometry. Anal Chem. 2016;88(19):9495–502. doi:10.1021/acs.analchem.6b01958.
- Suzuki T, Hashii N, Tada M, Ishii-Watabe A. The influence of antibody engineering on Fc conformation and Fc receptor binding properties: analysis of FcRn-binding engineered antibodies and an Fc fusion protein. MAbs. 2021;13(1):1923366. doi:10.1080/19420862.2021.1923366.
- Vidarsson G, Dekkers G, Rispens T. IgG subclasses and allotypes: from structure to effector functions. Front Immunol. 2014;5:520. doi:10.3389/fimmu.2014.00520.
- Abdiche YN, Yeung YA, Chaparro-Riggers J, Barman I, Strop P, Chin SM, Pham A, Bolton G, McDonough D, Lindquist K, et al. The neonatal Fc receptor (FcRn) binds independently to both sites of the IgG homodimer with identical affinity. MAbs. 2015;7(2):331–43. doi:10.1080/19420862.2015.1008353.
- Hageman TS, Weis DD. Reliable identification of significant differences in differential hydrogen exchange-mass spectrometry measurements using a hybrid significance testing approach. Anal Chem. 2019;91(13):8008–16. doi:10.1021/acs.analchem.9b01325.