https://orcid.org/0000-0001-9037-3668
ABSTRACT
IgG-based monoclonal antibody therapeutics, which are mainly IgG1, IgG2, and IgG4 subclasses or related variants, have dominated the biotherapeutics field for decades. Multiple laboratories have reported that the IgG subclasses possess different molecular characteristics that can affect their developability. For example, IgG1, the most popular IgG subclass for therapeutics, is known to have a characteristic degradation pathway related to its hinge fragility. However, there remains a paucity of studies that systematically evaluate the IgG subclasses on manufacturability and long-term stability. We thus conducted a systematic study of 12 mAbs derived from three sets of unrelated variable regions, each cloned into IgG1, an IgG1 variant with diminished effector functions, IgG2, and a stabilized IgG4 variant with further reduced FcγR interaction, to evaluate the impact of IgG subclass on manufacturability and high concentration stability in a common formulation buffer matrix. Our evaluation included Chinese hamster ovary cell productivity, host cell protein removal efficiency, N-linked glycan structure at the conserved N297 Fc position, solution appearance at high concentration, and aggregate growth, fragmentation, charge variant profile change, and post-translational modification upon thermal stress conditions or long-term storage at refrigerated temperature. Our results elucidated molecular attributes that are common to all IgG subclasses, as well as those that are unique to certain Fc domains, providing new insight into the effects of IgG subclass on antibody manufacturability and stability. These learnings can be used to enable a balanced decision on IgG subclass selection for therapeutic antibodies and aid in acceleration of their product development process.
Introduction
Monoclonal antibody (mAb) therapies have achieved tremendous success since the first antibody therapeutic (Orthoclone Okt3) was approved in 1986. Beyond the >100 mAb-related drugs currently on the U.S. market,Citation1 many more mAb-based drug candidates are in the clinical pipeline under extensive development to meet unmet medical needs in disease areas, such as oncology, immunology, cardiovascular, neurodegeneration, pain, and infectious diseases. Since 2014, this pipeline has supported the approval of innovative mAb-based therapeutic products in the United States or European Union (EU) at a rate of 6–13 products per year.Citation1
Among the approved mAb-related therapeutics, most belong to immunoglobin (Ig) G class, including IgG1, IgG2, and IgG4 subclasses.Citation1 IgG are multidomain and glycosylated heterodimers composed of two identical light chains (LC) and two identical heavy chains (HC) linked by multiple disulfide bonds. The LCs and HCs consist of variable regions and constant regions. Although there is a high level of sequence homology among the IgG subclasses, they distinguish from each other by small sequence differences in the constant region of the heavy chain, the length of the hinge that connects the antigen-binding (Fab) domain to the fragment crystallizable (Fc) domain, the number of disulfide bonds between the heavy chains within the hinge region, and the disulfide bond positions between the heavy and light chains. The hinge region of IgG1 or IgG4 antibodies consists of cysteine residues that form two interchain disulfide bonds, while IgG2 antibodies contain cysteine residues that form four interchain disulfide bonds. These differences impart a range of FcγR interactions and effector functions, such as antibody-dependent cell-mediated cytotoxicity (ADCC), antibody-dependent cellular phagocytosis (ADCP), and complement-dependent cytotoxicity (CDC), which have served as the primary consideration for the selection of IgG subclass for therapeutic applications and have also inspired numerous Fc engineering efforts to optimize the desirable FcγR interactions or effector functions.Citation2
Driven by the pressure to bring innovative medicines to patients faster and more cost-effectively, there has been a realization in recent years that the selection of IgG subclass can play a key role in improving Chemistry, Manufacturing, and Controls (CMC) properties of therapeutic antibodies and in optimizing the biological activity suitable to the therapeutic application. The appropriate choice of IgG subclass may also expedite a therapeutic mAb’s product development process. The manufacturing of mAb therapeutics is a sophisticated and expensive process that starts with fine-tuned cell culture using a dedicated clonal cell line to express the mAb, purification by a highly efficient downstream procedure to remove impurities, including host cell proteins (HCPs), followed by use of heavily regulated product manufacturing guidelines to produce parenteral material fit for human use. After product manufacturing, a long shelf life, typically 24–36 months under refrigerated conditions, is desirable to ensure consistent product performance and to reliably maintain ample supply for patients. Many of the diseases targeted by mAb therapies are chronic conditions and relatively high mAb concentrations with subcutaneous dosing for optimal efficacy and better patient compliance are frequently required. Such requirements impose a stringent demand for liquid formulation, thus placing a high premium on good molecular stability of the drug candidates. Therefore, the productivity of the cell culture process, the efficiency of impurity removal, and the long-term stability at high concentrations become key success factors for therapeutic mAb production. The distinct characteristics of the IgG subclasses can be taken into consideration to optimize the productivity and stability of the mAbs and to facilitate drug product development.
Despite the importance of the IgG subclass selection, there has been a paucity of the literature regarding the overall impact of IgG subclass on the manufacturing of mAb therapeutics. The acid-induced aggregation of the mAbs has been well documentedCitation3–5 and a strong dependence on the IgG subclass has been established, with IgG2 and IgG4 mAbs more prone to aggregation than IgG1 mAbs during low pH viral inactivation.Citation6–8 It is also well known that IgG1 mAbs are more susceptible to fragmentation,Citation9,Citation10 while IgG4 mAbs may be more prone to aggregation.Citation4,Citation11,Citation12 Using a high-throughput formulation screening method, Alekseychyk and coworkers studied a pair of IgG1 and IgG2 mAbs with the same variable region and reported an overall greater stability of the IgG1 mAb over the IgG2 mAb; principally, they noted that the IgG1 mAb has greater thermal stability at low pH and greater resistance to mechanical stress.Citation13 Tian et al.Citation12 conducted a systematic comparison of three series of mAbs in IgG1, IgG2, and IgG4 subclasses with identical variable region over a broad pH range and found that above pH 5 the aggregation propensity rank is IgG1 < IgG2 < IgG4 whereas the fragmentation propensity rank is IgG1 > IgG2 > IgG4. Neergaard and coworkers conducted a high concentration study on a pair of IgG1 and IgG4 mAb with identical variable region and observed a clear subclass effect on the stability of the mAbs.Citation14 Levy et al. analyzed residual HCP post-protein A affinity chromatography purification step in two unrelated IgG1 antibodies and two closely related IgG2 antibodies along with selective Fc and Fab domains from these mAbs.Citation15 Their results indicated a strong dependence of mAb sequences on the residual HCP levels, and that both the constant domain and the variable domain influence the amount and type of HCP association to the mAbs during protein A affinity chromatography process.
In our previous report,Citation16 we conducted a systematic molecular property investigation for 12 mAbs encompassing three independent variable regions across IgG1, an IgG1 variant, IgG2, and an IgG4 variant that revealed the differences in the properties of the IgG subclasses. Herein, we systematically evaluated the impact of IgG subclass on mAb therapeutic product development attributes, including Chinese hamster ovary (CHO) cell productivity, residual HCP level, Fc-glycosylation profile, physicochemical stability under stressed conditions, as well as upon long-term refrigerated storage conditions. We identified several attributes that are characteristic of each IgG subclass and attributes that are more influenced by the variable region under typical formulation conditions. These findings shed light on the common, as well as the unique, molecular attributes for each IgG subclass that can affect the development of mAbs as pharmaceutical products.
Results
Description of the antibodies and their bulk CHO stable cell line productivity
The 12 mAbs used in this study were described in a previous publication.Citation16 Briefly, three unrelated variable regions (Fv-A, Fv-B, and Fv-C) were each cloned into IgG1, IgG1 L234A, L235E, G237A, A330S, P331S (IgG1EN),Citation17 IgG2, and IgG4 S228P, L234A, L235A (IgG4PAA),Citation18–20 as listed in .
Table 1. List of the antibodies used in the study a.
A CHO bulk stable cell line was generated for each of the 12 mAbs and they were produced at 1 L scale in shake-flasks. The titer varied for each mAb and is listed in . Among the three series of mAbs, the Fv-A series demonstrated the highest titers ranging from 5.0 g/L to 7.1 g/L, while the Fv-B series and Fv-C series showed titers in the range of 3.4–4.9 g/L. With comparison across subclasses, the titer for IgG1 and IgG1EN mAbs trended moderately higher than that of the corresponding IgG2 and IgG4PAA mAbs.
Table 2. Summary of expression, purification, and formulation for the 12 antibodies investigated.
Purification and residual HCP
With a standard two-step purification process, i.e., protein A followed by cation exchange chromatography (CEX), the mAbs exhibited different levels of monomer percentage. For the mAbs with Fv-A, all mAbs yielded higher than 99% monomer regardless of IgG subclass. The monomer percentage varied in a small range (98.0–99.3%) for the mAbs with Fv-B, whereas all four mAbs with Fv-C exhibited purities of around 90%. In addition, aggregates were poorly resolved from monomeric species for all four mAbs with Fv-C by CEX.
The residual HCP in the antibody samples were measured by liquid chromatography mass spectrometry (LC-MS) methods. Following protein A/CEX purification, the residual HCP levels were highest for IgG4PAA within all three series, while IgG2 had HCP levels equivalent to or lower than IgG1 and IgG1EN. The Fv-A series exhibited the highest HCP levels within each subclass, while the levels were relatively similar between Fv-B and Fv-C mAbs ().
Upon the two-step purification, several mAbs, IgG1-A, IgG1EN-A, IgG4PAA-A, IgG4PAA-B, IgG1-C, IgG4PAA-C, displayed comparatively higher lipase and/or esterase HCP levels. Given the impact of lipase and esterase HCPs on polysorbate stability,Citation21,Citation22 these mAbs were further purified by hydrophobic interaction chromatography (HIC) to attain similar material quality for the subsequent studies and to minimize any potential impact of differential HCP levels on the stability of the mAbs. All final mAbs were >96% monomeric by size exclusion chromatography (SEC) except for IgG1EN-C and IgG2-C, which were 90% and 88% monomeric, respectively. Following protein A/CEX purification of IgG2-A mAb, significant precipitation of the CEX pool was noted; this mAb was not evaluated further. After purification, the mAbs were buffer exchanged by tangential flow filtration (TFF) into a matrix of 5 mM histidine, pH 6.
Following TFF, the HCP levels were measured again and all were found to be below 30 ppm with lipase and esterase levels both below level of detection ().
Fc-glycosylation profile
The structure of the conserved N-linked glycan attached to N297 (EU numbering) was assessed and found to be largely similar among the 11 purified mAbs. However, %G0F is elevated for IgG4 subclass, while %G1F trends lower for both IgG2 and IgG4 subclasses (see ). Relative to IgG1, IgG1EN has a 6–10% increase in G1F and a modest 2% increase in G2F with corresponding decreases in %G0 and %G0F.
Table 3. N-Glycan Profile at the Fc Site by LC-MS Peptide Mapping.
Physicochemical stability under stressed conditions and upon long-term refrigerated storage
To compare the physicochemical stability of these mAbs, all the IgG molecules except IgG2-A were concentrated to ~90 mg/mL in a common formulation buffer matrix consisting of 5 mM L-histidine at pH 6, 260 mM mannitol, and 0.03% (w/v) polysorbate 80. The samples were stored at 5°C, 25°C, and 35°C from 1 month to 30 months to assess their physicochemical stability under these conditions, including aggregation growth, fragmentation, charge variant profile change, and post-translational modifications (PTM). The results are described below. The protein concentration and solution pH were monitored and minimal change was observed over time.
Visual appearance
The concentrated samples were visually inspected immediately following concentrating and after 1-month incubation at 25°C and 35°C (). Given the same variable region, the IgG2 and IgG4PAA mAbs exhibited more opalescence than the IgG1 and IgG1EN mAbs. It is also evident that the variable region plays an important role, as the opalescence ranking is Fv-B series> Fv-C series > Fv-A series. These visual observations are consistent with the quantitative turbidity measurements reported previously.Citation16
Figure 1. Visual appearance of the mAb samples immediately after formulating, and after 1-month incubation at 25°C and 35°C, respectively.
Although differences in visual opalescence were noted between the subclasses within each series, this did not manifest into significant protein content loss. When stored at 5°C for 2.5 years, all samples maintained approximately 95% of their original protein concentrations.
Aggregate growth by SEC
The aggregate formation upon storage over time at 5°C, 25°C, and 35°C is illustrated in . After 1-month thermal stress at 25°C or 35°C, there is negligible subclass impact, as Fv-A, Fv-B, and Fv-C mAbs all demonstrated comparable levels of aggregate growth within their respective series. However, after 3 months at 25°C, IgG2-C displayed a higher aggregation level than the other mAbs in the Fv-C series.
Figure 2. Aggregate growth of the mAb samples upon incubation for (A) 30 months at 5°C, (B) 1 month (gray) and 3 months (black) at 25°C, and (C) 1 month at 35°C.
After 30 months of storage at 5°C, the Fv-A series displayed minimal aggregate growth, while the aggregate growth of IgG1-B and IgG1EN-B trended slightly higher among the Fv-B series and the aggregation level of IgG4PAA-B and IgG4PAA-C trended the lowest within the Fv-B and Fv-C series. Overall, all the mAbs in this study were stable under the conditions tested and there is no consistent trend in aggregate growth among the different IgG subclasses.
Fragmentation by SEC
SEC is widely utilized for monitoring mAb fragmentation. As illustrated in , mAbs with IgG1 and IgG1EN constant domains clearly demonstrated the highest fragmentation growth within each mAb series after incubation for 30 months at 5°C and for 3 months at 25°C, whereas IgG2 and IgG4PAA mAbs showed the lowest increase in fragmentation. Upon thermal stress at 25°C or 35°C for 1 month, the fragmentation growth exhibited a marked dependence on the variable domain, with IgG1-B and IgG1EN-B showing more pronounced increase over their counterparts in Fv-A or Fv-C series. After storage at 5°C for 30 months, however, the effects of variable domain on fragmentation are less prominent, with each IgG subclass displaying similar levels of fragmentation growth regardless of the variable domain.
Figure 3. Fragmentation growth of the mAb samples upon incubation for (A) 30 months at 5°C, (B) 1 month (gray) and 3 months (black) at 25°C, and (C) 1 month at 35°C.
Charge variants change by iCE
The charge variant change of all 11 mAbs was analyzed using imaged capillary electrophoresis (iCE). The changes in the charge variant content are summarized in .
Table 4. Summary of charge variant changes.
After 1-month thermal stress at 35°C, the IgG1EN and IgG4PAA mAbs demonstrated significant loss in the main peaks (10.4–15.2%) compared to the IgG1 and IgG2 mAbs (2–6.4%) across all three mAb series. For the IgG1EN mAbs, the decrease of the main peaks predominantly led to increases in their acidic peaks. Each IgG1EN mAb displayed a significant increase in the acidic peak relative to its respective IgG1 mAb within each series. The influence of IgG2 and IgG4 subclass on charge variant change is not universal across all series of mAbs. Although the charge variant change for IgG2-B is more muted, there is a significant increase in the acidic peak for IgG2-C. For IgG4PAA-A and IgG4PAA-C, the decrease of the main peak led to more increase in the acidic peak than in the basic peak, while the opposite was observed for IgG4PAA-B. The 25°C data largely follow a similar trend as the 35°C data.
After 30 months of incubation at 5°C, the charge variant differences among the IgG subclasses were smaller in general, but mimic the same trend as at the higher temperatures. While all Fv-A, all Fv-C, and most Fv-B mAbs showed significant increases in acidic species regardless of subclass, IgG4PAA-B exhibited a larger increase in basic species than in acidic species.
PTM hotspots on Fc by LC-MS and peptide mapping
While the PTM hotspots in the variable region are highly dependent on the sequences, some in the constant region are common across the different IgG subclasses. We focused our analysis on the Fc domains in the samples after 30 months storage at 5°C and 3 months thermal stress at 25°C to compare PTM changes across the series.
Six potential deamination sites, N315, N325, N361, N384, N389, and N434 (EU numbering), were identified in the Fc domains, with these sites conserved across the assessed subclasses. The deamidation-level changes at these sites, obtained by peptide mapping analysis, are depicted in . At 25°C after 3 months, the total deamidation level in the Fc domain increased between 4% and 10% for all the mAbs. Of the six potential deamidation sites, N325, N384, and N389 exhibited the most pronounced increase, while N315, N361, and N434 showed <0.5% deamidation-level increase in the 3-month 25°C stability samples. In all three series, the IgG1EN mAbs showed the highest increase in overall deamidation level in its Fc domain in the 3-month 25°C stability samples (7.1–9.9%), fueled predominantly by the large increase at N325. All the mAbs showed ~ 1–3% increase of total deamidation in the Fc domain upon refrigerated storage for 30 months, while a more prominent contribution from N325 is observed in the IgG1EN mAbs.
Figure 4. Deamidation level increase in Fc domains for A) Fv-A mAbs, B) Fv-B mAbs, and C) Fv-C mAbs.
While these IgG molecules share six common deamination sites in the Fc domains, they contain different numbers of potential oxidation sites in the respective Fc domains, consisting of Met, Trp, and His residues. Very low levels of oxidation at Trp and His residues were detected in all the stability samples and most oxidation occurred at the Met residues. Therefore, only the change in oxidation levels at the Met sites are shown in . The Fc domain of IgG1 contains two Met residues (M252 and M428; EU numbering), IgG1EN and IgG4 each contain three Met residues (M252, M358, and M428; EU numbering), whereas IgG2 contains five Met residues (M252, M282, M358, M397, and M428; EU numbering). Only M252 and M428 are shared by all the Fc domains. Within Fv-A and Fv-B series at 25°C for 3 months, the total oxidation-level changes for IgG1 samples (2.2–2.5%) were lower compared to IgG1EN (4.2–6.2%), IgG2 (5.7%), and IgG4PAA (3.3–8.8%), whereas Fv-C series showed marginal differences among IgG1, IgG1EN, IgG2, and IgG4PAA samples (2.6–3.8%). In the absence of a consistent trend in the total oxidation-level change among the subclasses, the instability at the common M252 is evident for all the mAbs after both 30-months incubation at 5°C and 3-months incubation at 25°C. Interestingly, IgG4PAA-B was particularly prone to oxidation at M252, suggesting a potential Fab contribution to the M252 oxidation. M428, which is also common to all the IgG molecules, showed smaller changes in oxidation level. M358, which is present in IgG1EN, IgG2, and IgG4PAA, is largely resistant to oxidation. Of the two additional Met residues in IgG2 that are not present in the other subclasses, M282 is susceptible to oxidation and exhibited sizable changes in stability samples under both conditions, while only a low oxidation level was observed at M397.
Figure 5. Oxidation level increase in Fc domains for A) Fv-A mAbs, B) Fv-B mAbs, and C) Fv-C mAbs.
Other PTMs, such as isomerization at Asp residues and succinimide formation at Asn or Asp residues, were also detected. The succinimide would likely convert to isoAsp or Asp during the sample preparation of the tryptic digestion. These PTMs were present at low levels (total changes ≤ 2% in stressed samples and <1% in 5°C long-term storage samples), and were not investigated further.
Discussion
Over the past few decades, mAbs have become an attractive class of human therapeutics to treat a variety of chronic and acute human diseases due to their specificity and safety. Given the number of naturally occurring IgG subclasses and the variety of Fc variants available that impart unique FcγR interactions and effector functions, selecting an appropriate IgG subclass or Fc variant offers an opportunity to greatly enhance therapeutic efficacy and improve safety for patients. At the same time, the choice of IgG subclass or Fc variant may lead to cost and timeline consequences for the drug development process. Understanding the molecular behavior or developability of the IgG subclasses and their variants has the potential to expedite the delivery of new medicines to patients.
Despite many structural similarities among mAbs, each therapeutic mAb often confers complex and unique challenges during the CMC development process. Topics relating to the development of therapeutic mAbs, ranging from cell culture productivity and HCP reduction, to physicochemical stabilities, have been of intense interest for years.Citation23–25 However, reports with a direct comparison of these characteristics used to dissect the IgG subclass differences are relatively sparse. While the complexity of mAb development warrants extensive investigations for each individual mAb drug candidate, we conducted a systematic study on three independent series of mAbs in different IgG subclasses under common formulation conditions to identify the attributes unique to each IgG subclass that can affect therapeutic product development.
CHO cells are widely used as an expression host for the manufacturing of biopharmaceutical drugs. It is crucial to have high-producing mAbs from CHO cells to efficiently generate large quantities of material to meet the demand of the global market. All the mAbs in this study demonstrated reasonably high titers from the CHO bulk stable cell lines. Since all mAbs within each series used identical codons for their variable domains, the impact of subclass could be directly compared. In this study, the IgG1 and IgG1EN mAbs on average provided approximately 30% increase in titer compared to the corresponding IgG2 and IgG4PAA mAbs. Similarly, as all mAbs of the same subclass used identical codons for their constant domains, it is evident that the variable region sequences had a large effect on the titer. We found that the Fv-A mAbs have higher titers compared with the Fv-B and Fv-C mAbs. It is worth noting that antibody expression levels are known to be influenced by multiple factors,Citation26 and the current data are derived from heterogeneous bulk stable cell lines without optimization for individual mAbs. While highly productive bulk stable cell lines are beneficial to generate large quantities of material during discovery and early development stages, the risk of lower titers from bulk may be mitigated by the clonal cell lines that are necessary for clinical development and commercialization.
Residual HCPs are process-related impurities in mAb products that are well known to present potential safety risks to patients or compromise product stability, and thus HCP levels must be closely monitored in drug products.Citation27–30 Following expression in CHO stable cell lines, the mAbs were purified through conventional protein A chromatography followed by cation exchange chromatography and residual HCPs were characterized. Within each series, the IgG4PAA mAbs displayed comparatively high levels of residual HCPs, whereas IgG2 mAbs had markedly lower residual HCPs. Similarly, the influence of the variable region on residual HCPs is apparent, as the Fv-A mAbs have higher HCP levels compared to their subclass counterparts within the Fv-B and Fv-C series. Our finding from the two-step purification is in agreement with a previous report analyzing one-step protein A purification that showed both the constant domain and the variable domain influence residual HCP levels.Citation15
N-linked glycosylation at N297 in the Fc domain is a common feature in all naturally occurring mAbs. While numerous reports have studied glycosylation differences among the subclasses for endogenous IgGs and the implications to the health of the plasma donors,Citation31,Citation32 scant literature exists detailing the influence of IgG subclass on glycan structure from recombinantly produced mAbs.Citation33 Given the importance of glycan structure, namely the impact of afucosylation and galactosylation, for effector functions, the literature on glycan content for recombinant mAbs has primarily focussed on IgG1.Citation34 The Fc mutation F241A in an IgG1 mAb has been shown to alter glycan structures, consequently modulating effector functions.Citation35,Citation36 In addition to revealing the impact of Fc-galactosylation on effector functions, Aoyama et al.Citation36 demonstrated that an G1F isomer, designated G1aF, and G2F have a stabilizing effect on the CH2 domain compared to G0F and the other G1F isomer, G1bF. In this work, we found an IgG subclass-dependent trend that the IgG4PAA mAbs tend to have higher level of G0F, whereas IgG1 and IgG1EN mAbs showed comparatively higher level of G1F across each series, without influence from the variable region. The potential implication of the Fc-glycosylation content difference on physicochemical stability was not further investigated.
To optimize therapeutic efficacy and ensure good patient compliance, subcutaneous administration of high concentration and low volume mAb therapeutics is often preferred. We evaluated the behavior of the mAbs at a moderately high concentration of 90 mg/mL in a common formulation buffer matrix. Upon concentration, the difference in solution opalescence between the mAbs of various IgG subclasses was immediately apparent. All mAbs of IgG4PAA and IgG2 subclasses displayed higher level of visual opalescence compared to IgG1 and IgG1EN mAbs. For all three series, the visual opalescence ranking is IgG4PAA> IgG2 (when available) > IgG1 ~ IgG1EN mAbs. The opalescence observation, however, did not lead to significant protein content loss in these samples. It is recognized that opalescence is relatively common for IgG2 and IgG4 mAbs at high concentrations,Citation37,Citation38 driven by self-association resulting from charge interaction or hydrophobicity or both. The opalescence observation, similar to viscosity, is significantly influenced by the variable region sequences, demonstrated by the different level of opalescence for Fv-A, Fv-B, and Fv-C mAbs as shown in . Fv-A mAbs showed the least opalescence yet the highest viscosity among the mAb (), consistent with the findings by Kingsbury et al.Citation39
Aggregation of therapeutic mAbs during storage represents a significant hurdle to product development. We evaluated the aggregate growth of the mAbs under thermal stress and long-term cold storage conditions. After 1-month thermal stress at 35°C, all mAbs showed aggregate increase of~0.6–1.4% (). The Fv-A and Fv-B mAbs each have a narrow range of aggregate growth among their respective series with little differentiation seen between the subclasses. For the Fv-C series, the IgG1 mAb demonstrated the highest increase of aggregates (around 1%), while the other mAbs showed ~ 0.6% increase. Such differences among mAbs are still considered to be minor. The aggregate growth data after 3-months storage at 25°C is congruent with the 1-month 35°C incubation data, except for the larger increase for IgG2-C, which was not observed after 1 month at 25°C. The 30-months 5°C storage data showed a similar pattern, with no distinguishing differences beyond experimental error among the IgG subclasses. The Fv-B IgG1 and IgG1EN mAbs display an increase in aggregate levels after 30 months at 5°C, but this trend was not present in Fv-A or Fv-C. Therefore, it can be concluded that in this study no significant IgG subclass-dependent difference or trend in aggregate growth is evident.
Interestingly, the above results demonstrated an indifference to thermal stability measured by differential scanning calorimetry, although it is perceived that protein unfolding at lower temperature is a likely driver of aggregation.Citation11,Citation40 Our 2021 publication on these mAbsCitation16 showed that the first thermal transition temperature for the IgG1EN mAbs is decreased to ~ 65°C, which is about 7°C lower compared with that of the IgG1 mAbs, and the Fab thermal transition temperature in IgG4PAA is slightly lower than those in the other IgG subclasses.Citation16,Citation41 However, these did not result in appreciable aggregation differences under the conditions tested. The early findings that IgG2 and IgG4 subclass aggregate more readily than IgG1 subclassCitation11,Citation42 were not substantiated in this study. Possible reasons for our findings are: (1) the lowered thermal transition temperatures are still well above the temperatures used in the thermal stress studies, (2) the first thermal transition temperature occurs in the CH2 domain, while unfolding of this domain has been shown to be reversible in the pH range of 4.5–7.5,Citation43 or (3) all the mAbs included in this study are overall very stable under the current conditions and their aggregation propensity may only be differentiated under alternative conditions.
IgG1 mAbs have been reported to be more prone to hinge region fragmentation compared to IgG2 and IgG4 mAbs under heat, trace metal, and pH-related stress conditions, which is one of their major degradation pathways.Citation4,Citation44,Citation45 We assessed fragmentation in the mAb samples after thermal stress and long-term cold storage. As expected, the IgG1 and IgG1EN mAbs displayed significantly increased fragmentation compared with the IgG2 and IgG4PAA mAbs after 30 months at 5°C. Interestingly, the fragmentation pattern across the IgG subclasses at low temperature is not necessarily reflected in the short thermal stress studies, as the 1-month stability data at 25°C or 35°C for Fv-A and Fv-C series showed only marginal differences among the IgG subclasses. The variable region influence in the thermal stressed data is evident and may suggest a temperature-dependent difference in the fragmentation mechanism in these mAbs. Although accelerated stability studies and forced degradation studies are widely used in the early stages of the mAb development process in order to identify formulation and product degradation risks, they are known to not fully represent the results upon long-term cold storage.Citation46
Protein charge heterogeneity is attributed to amino acid side-chain chemical modifications, protonation state of charged amino acids side chains, and charged glycans.Citation45,Citation47 While much effort has been spent to isolate, characterize, and evaluate the biological impact of mAb charge heterogeneity change, we are aware of few systematic reports that address the effect of IgG subclass on mAb charge heterogeneity changes. In this study, we found that IgG1EN and IgG4PAA mAbs are more susceptible to loss in the main peaks compared to IgG1 and IgG2 mAbs across all three series after 1-month 35°C or 3-months 25°C incubation. There is also a lack of uniform influence of IgG subclass on acidic or basic charge variant formation across each series, likely a reflection of the many residues in the variable regions that are also subject to modification.
Using the LC-MS peptide mapping method, we evaluated the impact of the IgG subclass on the common PTM sites in the Fc domain across the three mAb series. A consistent pattern emerged at the deamidation hotspots across the IgG subclasses upon 3-months 25°C incubation and 30-months 5°C incubation. Three of the six potential deamidation hotspots, N325, N384, and N389, accounted for nearly all the total deamidation changes in the Fc domains. N325 within the CH2 domain was found to be a primary site of deamidation in all the IgG molecules, especially upon thermal stress. N325 deamidation was previously shown to affect ADCC activity.Citation48 While the deamidation changes in the IgG1 mAbs and the IgG1EN mAbs are largely comparable at multiple sites, deamidation at N325 is exacerbated in the IgG1EN mAb in each series, likely driving the large acidic peak increases for this IgG1 variant. Interestingly, N325 is sequentially adjacent to the A330S/P331S mutations within the IgG1EN. Additionally, an IgG1 crystal structure (PBD: 1HZH) shows that N325 is spatially adjacent to L234, L235, G237 sites that are mutated in the IgG1EN. Within the IgG1 structure, N325 is structurally buried with low surface exposure, which may suggest the destabilization of the CH2 domain in IgG1EN as reported by TangCitation16 makes N325 more solvent-accessible and susceptible to deamidation. This finding is consistent with the report from Yan et al.Citation49 that deamidation risk at N325 is elevated under heat stress and is subject to local conformation and solvent accessibility. Taken together, thermal stability of the CH2 domain appears to directly affect N325 deamidation. It will be interesting to find if deamidation at N325 is also accelerated in other Fc variants with destabilized CH2 domains.Citation50–52
The assessment of the IgG subclass impact on the oxidation hotspots in the Fc domain is complicated by the varying number of susceptible Met residues present within the constant domains, which ranges from 2 to 5. Nevertheless, M252, which is common in all the Fc domains, was identified as a major contributor to the total oxidation-level change in all the samples. For both the Fv-A and Fv-B series, the oxidation-level change at M252 is exacerbated in the IgG1EN mAb compared with the IgG1 mAb, analogous to deamidation level change at N325, whereas the M252 oxidation level in IgG1-C and IgG1EN-C remains similar. In addition, there is a significantly larger change at M252 in IgG4PAA-B than in IgG4PAA-A or IgG4PAA-C, potentially suggesting a variable domain influence.
The physicochemical stability of an IgG mAb is complex and heavily subject to the nature of stress, while confounded with liabilities from both the constant and variable domains. To obtain a wholistic picture of the physicochemical instability challenges of different IgG subclasses, we normalized several important physicochemical changes from the 30-months cold storage data to a risk score of 1–3 based on the extent of the changes and method variation (see legend). The Fc domain deamidation risk score is based on the 3-months 25°C thermal stress data. For a more complete picture, we also assigned risk scores for previously reported turbidity and viscosity data () and included them in . As shown in , the IgG1 and IgG1EN mAbs exhibited a higher fragmentation risk, while the IgG1EN mAbs also showed increased charge variant change and deamidation risk. The IgG1 and IgG1EN mAbs have the lowest turbidity risk, whereas the risk is higher for IgG2 and IgG4PAA. While viscosity tends to be higher for IgG4PAA mAbs, the data point to an interplay between the variable domain and the IgG subclass. All the mAbs regardless of IgG subclass exhibited low aggregation risk in the current study.
Figure 6. Radar plots visualizing Fv-A (Blue), Fv-B (Orange), and Fv-C (Green) risk levels for indicated attributes of IgG1 (A), IgG1EN (B), IgG2 (C), and IgG4PAA (D): low risk (1), medium risk (2), and high risk (3). Risk levels were defined for each parameter. Aggregates and fragmentation: <1% = low risk, 1–2% = medium risk, >2% = high risk. Charge variant change and Fc deamidation: <5% = low risk, 5–10% = medium risk, >10% = high risk. Turbidity: <15 NTU = low risk, 15–20 NTU = medium risk, >20 NTU = high risk. Viscosity: <12 cP = low risk, 12–20 cP = medium risk, >20 cP = high risk.
In conclusion, our investigations led to a deeper understanding of IgG subclass impact on mAb product development. We compared several important CMC-related attributes for each subclass, including stable CHO bulk cell line titer, HCP removal, high concentration visual opalescence, fragmentation, charge variant profile change, and PTM hot spots susceptibility in the constant domains, while observing no consistent trend in aggregate growth or in oxidation hot spots in the constant domains under these conditions. The findings herein on manufacturability and stability should be leveraged in addition to intended biology during IgG subclass selection of therapeutic mAb candidates to facilitate risk mitigation strategies for the development of mAb drug products.
Materials and methods
Antibody expression
The antibodies used in this study were described by Tang et al.Citation16 Within each mAb series, variable domain codon utilization was matched across all subclasses; likewise, codon utilization within the constant domains is matched for each IgG subclass.
CHO bulk stable cell lines were generated for each of the 12 mAbs and the mAbs were produced in shake-flasks at 1 L scale. Expression titers were measured at harvest by ForteBio Octet RED utilizing Protein-A tips. A standard curve was established using a subclass-matched control.
Antibody purification
All antibodies were purified from conditioned media through a two-step, protein A followed by cation exchange chromatography purification process. Proteins were first purified by mAbSelect SuRe protein A (Cytiva) utilizing conventional methods. Elution mainstreams were neutralized to pH 5 and filtered prior to polishing with cation exchange chromatography.
The cation exchange chromatography utilized a POROS 50HS cation exchange column. After sample loading, the column was washed with 20 mM sodium acetate, pH 5.0 to remove non-bound contaminants. Proteins were eluted with a sodium chloride gradient and the main peak fractions were pooled and filtered.
Samples identified with high levels of HCPs were further purified through hydrophobic interaction chromatography using Phenyl HP (Cytiva) and a sodium sulfate reverse gradient at pH 8.0.
Host-cell protein quantitation
Residual host-cell protein identification and quantification were conducted according to the published method.Citation53 Briefly, an aliquot containing 1.0 mg protein of each sample solution was mixed with 5 μL of 1 M tris-HCl buffer, pH 8, and addition of water to 195 μL. Each solution was treated with 5 μL of trypsin (recombinant bovine trypsin, Lilly, Indianapolis, USA) and protein standard mixture at 37°C overnight. Each digest was reduced with dithiothreitol (DTT) and heated to approximately 80°C for 5–10 minutes. After centrifugation to remove precipitate, supernatant was acidified with 10% formic acid in water. The digests were characterized on an Orbitrap Fusion Lumos or HF-X mass spectrometer after separation on an Ultimate 3000 nano LC with Acclaim PepMap RSLC, 75 mm × 15 cm nano Viper C18 2 mm 100 Å (Thermo) RP column with a pre-column 30 mm × 5 mm, nano Viper capture column. Separations were achieved utilizing a water/acetonitrile gradient with 0.10% formic acid.
LC/MS/MS run was analyzed by Proteome Discoverer 1.4 against CHO-K1 2014 protein database with the addition of a control antibody and spiked protein sequences. The quantitative data of each HCP was calculated based on the area of top three peptides for each protein vs spiked proteins and their molecular weight.
Stability study protocol
A Cogent µScale TFF system equipped with a Pellicon 3, 30 kDa C-screen membrane was used to buffer exchange and concentrate the mAbs of interest to an intermediate concentration of 40 to 80 mg/mL in 5 mM histidine, pH 6.0. The mAbs were further concentrated to 90 mg/mL using Amicon Ultra-15 centrifugal filter concentrators (30 kDa NMWL). The mAb concentrations were confirmed by UV-Vis measurements and the final pH values were within the range of ±0.2 units. Each antibody solution was spiked with mannitol and polysorbate 80 stock solutions for a final matrix composition of 5 mM histidine, 260 mM mannitol, 0.03% polysorbate 80, pH 6.0 followed by sterile filtering. Final formulations were filled into cleaned and sterilized 5 mL (13 mm neck) glass vials with corresponding rubber septums. The stability samples were incubated for 1–3 months at 25°C and 35°C, and for 30 months at 5°C.
Size-exclusion chromatography for aggregation and fragmentation
mAb aggregates were measured by SEC-HPLC equipped with a UV detector (214 nm). Samples were analyzed utilizing a TSKgel G3000SWXL, 7.8 mm x 30 cm, 5 μm size-exclusion column (Tosoh) with isocratic elution of 50 mM sodium phosphate, 300 mM NaCl, pH 7.0 followed by integration for aggregates, main peak, and fragments.
Change of charge variants by iCE
The changes in charge variant profiles were evaluated using ProteinSimple iCE3 equipped with an autosampler and an FC-coated cIEF cartridge. Samples were diluted to 0.25 mg/mL in 3 M urea/0.35% methylcellulose with Pharmalytes 3–10 (Sigma) prior to loading into the instrument. Samples were focused at 1500 V for 1 minute followed by a second focusing period of 3000 V for 8 minutes.
Peptide mapping
Each antibody solution was diluted with water to 10 mg/mL and 2 mL aliquot of each solution was mixed with 5 mL of 6 M guanidine HCl, 0.25 M Tris-HCl buffer, pH 8, and 0.5 mL of 50 mg/mL DTT solution. Each solution was incubated at 37°C for 30 minutes, mixed with 0.75 mL of 100 mg/mL iodoacetamide solution and 90 mL of deionized (DI) water, and then treated with 1.5 mL of 0.2 mg/mL Lys-C solution (FUJIFILM Wako Pure Chemical Corporation, Prod No. 125–05061. Japan) and 1 mL of 0.5 mg/mL recombinant bovine trypsin (Lilly) at 37°C for 3 hours. Each digest was mixed with 1 mL of 10% trifluoroacetic acid (TFA) in DI water before LC/MS/MS analysis. The digests were characterized on an Orbitrap Fusion Lumos mass spectrometer after separation on a Vanquish UPLC with Waters Acquity UPLC BEH 300 C18, 2.1 × 150 mm, 1.7 mm particle size column. Separations were achieved using a water/acetonitrile gradient with 0.10% TFA in 60 minutes.
List of Abbreviations
| CEX | = | cation exchange chromatography |
| CHO | = | Chinese hamster ovary |
| CMC | = | Chemistry, Manufacturing and Controls |
| Fab | = | antigen-binding domain |
| Fc | = | crystallizable domain |
| FcγR | = | Fc gamma receptor |
| Fv | = | variable domain |
| HC | = | heavy chain |
| HCP | = | host cell protein |
| HIC | = | hydrophobic interaction chromatography |
| iCE | = | imaged capillary electrophoresis |
| IgG1EN | = | human IgG1 variant with |
| IgG4PAA | = | human IgG4 variant with S228P/L234A/L235A mutations |
| LC | = | light chain |
| LC-MS | = | liquid chromatography mass spectrometry |
| mAb | = | monoclonal antibody |
| PTM | = | post-translational modification |
| SEC | = | size exclusion chromatography |
| TFF | = | tangential flow filtration |
Acknowledgments
The authors kindly acknowledge Maria Hougland, Robert Peery, Ben Shanehsaz, and Richard Chen for assisting in antibody production and purification, and Dr. Michael De Felippis for encouragement and helpful discussions.
Disclosure statement
All authors in this report except Yu Tang are current employees of Eli Lilly and Company. All the data in this report were generated at Eli Lilly and Company, Indianapolis, IN. The authors do not have any conflict of interest or financial disclosure to report.
Additional information
Funding
References
- Kaplon H, Chenoweth A, Crescioli S, Reichert JM. Antibodies to watch in 2022. mAbs. 2022;14:2014296. doi:10.1080/19420862.2021.2014296. PMID: 35030985.
- Delidakis G, Kim JE, George K, Georgiou G. Improving antibody therapeutics by manipulating the Fc domain: immunological and structural considerations. Annu Rev Biomed Eng. 2022;24:249–12. doi:10.1146/annurev-bioeng-082721-024500. PMID: 35363537.
- Hari SB, Lau H, Razinkov VI, Chen S, Latypov RF. Acid-induced aggregation of human monoclonal IgG1 and IgG2: molecular mechanism and the effect of solution composition. Biochemistry. 2010;49:9328–38. doi:10.1021/bi100841u. PMID: 20843079.
- Ishikawa T, Ito T, Endo R, Nakagawa K, Sawa E, Wakamatsu K. Influence of pH on heat-induced aggregation and degradation of therapeutic monoclonal antibodies. Biol Pharm Bull. 2010;33:1413–17. doi:10.1248/bpb.33.1413. PMID: 20686240.
- Skamris T, Tian X, Thorolfsson M, Karkov HS, Rasmussen HB, Langkilde AE, Vestergaard B. Monoclonal antibodies follow distinct aggregation pathways during production-relevant acidic incubation and neutralization. Pharm Res. 2016;33:716–28. doi:10.1007/s11095-015-1821-0. PMID: 26563206.
- Saito S, Namisaki H, Hiraishi K, Takahashi N, Iida S. Engineering a human IgG2 antibody stable at low pH. Protein Sci. 2020;29:1186–95. doi:10.1002/pro.3852. PMID: 32142185.
- Mazzer AR, Perraud X, Halley J, O’hara J, Bracewell DG. Protein a chromatography increases monoclonal antibody aggregation rate during subsequent low pH virus inactivation hold. J Chromatogr A. 2015;1415:83–90. PMID: 26346187. doi:10.1016/j.chroma.2015.08.068.
- Jin W, Xing Z, Song Y, Huang C, Xu X, Ghose S, Li ZJ. Protein aggregation and mitigation strategy in low pH viral inactivation for monoclonal antibody purification. mAbs. 2019;11:1479–91. doi:10.1080/19420862.2019.1658493. PMID: 31441367.
- Brezski RJ, Oberholtzer A, Strake B, Jordan RE. The in vitro resistance of IgG2 to proteolytic attack concurs with a comparative paucity of autoantibodies against peptide analogs of the IgG2 hinge. mAbs. 2011;3:558–67. doi:10.4161/mabs.3.6.18119. PMID: 22123056.
- Suzuki S, Annaka H, Konno S, Kumagai I, Asano R. Engineering the hinge region of human IgG1 Fc-fused bispecific antibodies to improve fragmentation resistance. Sci Rep. 2018;8:17253. doi:10.1038/s41598-018-35489-y. PMID: 30467410.
- Ito T, Tsumoto K. Effects of subclass change on the structural stability of chimeric, humanized, and human antibodies under thermal stress. Protein Sci. 2013;22:1542–51. doi:10.1002/pro.2340. PMID: 23963869.
- Tian X, Langkilde AE, Thorolfsson M, Rasmussen HB, Vestergaard B. Small-angle x-ray scattering screening complements conventional biophysical analysis: comparative structural and biophysical analysis of monoclonal antibodies IgG1, IgG2, and IgG4. J Pharm Sci. 2014;103:1701–10. doi:10.1002/jps.23964. PMID: 24700358.
- Alekseychyk L, Su C, Becker GW, Treuheit MJ, Razinkov VI., Razinkov VI. High-throughput screening and stability optimization of anti-streptavidin IgG1 and IgG2 formulations. J Biomol Screen. 2014;19:1290–301. PMID: 25023322. doi:10.1177/1087057114542431.
- Neergaard MS, Nielsen AD, Parshad H, Van De Weert M. Stability of monoclonal antibodies at high-concentration: head-to-head comparison of the IgG1 and IgG4 subclass. J Pharm Sci. 2014;103:115–27. doi:10.1002/jps.23788. PMID: 24282022.
- Levy NE, Valente KN, Choe LH, Lee KH, Lenhoff AM. Identification and characterization of host cell protein product-associated impurities in monoclonal antibody bioprocessing. Biotechnol Bioeng. 2014;111:904–12. doi:10.1002/bit.25158. PMID: 24254318.
- Tang Y, Cain P, Anguiano V, Shih JJ, Chai Q, Feng Y. Impact of IgG subclass on molecular properties of monoclonal antibodies. mAbs. 2021;13:1993768. doi:10.1080/19420862.2021.1993768. PMID: 34763607.
- Gross JA, Dillon SR, Mudri S, Johnston J, Littau A, Roque R, Rixon M, Schou O, Foley KP, Haugen H, et al. TACI-Ig neutralizes molecules critical for B cell development and autoimmune disease. impaired B cell maturation in mice lacking BLyS. Immunity. 2001;15:289–302. PMID: 11520463. doi:10.1016/s1074-7613(01)00183-2.
- Angal S, King DJ, Bodmer MW, Turner A, Lawson AD, Roberts G, Pedley B, Adair JR. A single amino acid substitution abolishes the heterogeneity of chimeric mouse/human (IgG4) antibody. Mol Immunol. 1993;30:105–08. doi:10.1016/0161-5890(93)90432-b. PMID: 8417368.
- Bloom JW, Madanat MS, Marriott D, Wong T, Chan SY. Intrachain disulfide bond in the core hinge region of human IgG4. Protein science: a publication of the protein society. Protein Sci. 1997;6:407–15. PMID: 9041643. doi:10.1002/pro.5560060217.
- Schuurman J, Perdok GJ, Gorter AD, Aalberse RC. The inter-heavy chain disulfide bonds of IgG4 are in equilibrium with intra-chain disulfide bonds. Mol Immunol. 2001;38:1–8. doi:10.1016/s0161-5890(01)00050-5. PMID: 11483205.
- Li X, Wang F, Li H, Richardson DD, Roush DJ. The measurement and control of high-risk host cell proteins for polysorbate degradation in biologics formulation. Antib Ther. 2022;5:42–54. doi:10.1093/abt/tbac002. PMID: 35155990.
- Zhang S, Riccardi C, Kamen D, Reilly J, Mattila J, Bak H, Xiao H, Li N. Identification of the specific causes of polysorbate 20 degradation in monoclonal antibody formulations containing multiple lipases. Pharm Res. 2022;39:75–87. doi:10.1007/s11095-021-03160-3. PMID: 34981317.
- Szkodny AC, Lee KH. Biopharmaceutical manufacturing: historical perspectives and future directions. Annual review of chemical and biomolecular engineering. Annu Rev Chem Biomol Eng. 2022;13:141–65. PMID: 35300518. doi:10.1146/annurev-chembioeng-092220-125832.
- Kunert R, Reinhart D. Advances in recombinant antibody manufacturing. Appl Microbiol Biotechnol. 2016;100:3451–61. doi:10.1007/s00253-016-7388-9. PMID: 26936774.
- Le Basle Y, Chennell P, Tokhadze N, Astier A, Sautou V. Physicochemical stability of monoclonal antibodies: a review. J Pharm Sci. 2020;109:169–90. doi:10.1016/j.xphs.2019.08.009. PMID: 31465737.
- Ayyar BV, Arora S, Ravi SS. Optimizing antibody expression: the nuts and bolts. Meth (San Diego Calif). 2017;116:51–62. doi:10.1016/j.ymeth.2017.01.009. PMID: 28163103.
- Fischer SK, Cheu M, Peng K, Lowe J, Araujo J, Murray E, McClintock D, Matthews J, Siguenza P, Song A. Specific immune response to phospholipase B-Like 2 protein, a host cell impurity in Lebrikizumab clinical material. Aaps J. 2017;19:254–63. doi:10.1208/s12248-016-9998-7. PMID: 27739010.
- Gao SX, Zhang Y, Stansberry-Perkins K, Buko A, Bai S, Nguyen V, Brader ML. Fragmentation of a highly purified monoclonal antibody attributed to residual CHO cell protease activity. Biotechnol Bioeng. 2011;108:977–82. doi:10.1002/bit.22982. PMID: 21404269.
- Chiu J, Valente KN, Levy NE, Min L, Lenhoff AM, Lee KH. Knockout of a difficult-to-remove CHO host cell protein, lipoprotein lipase, for improved polysorbate stability in monoclonal antibody formulations. Biotechnol Bioeng. 2017;114:1006–15. doi:10.1002/bit.26237. PMID: 27943242.
- Hall T, Sandefur SL, Frye CC, Tuley TL, Huang L. Polysorbates 20 and 80 degradation by group XV lysosomal phospholipase A2 isomer X1 in monoclonal antibody formulations. J Pharm Sci. 2016;105:1633–42. doi:10.1016/j.xphs.2016.02.022. PMID: 27056628.
- Plomp R, Ruhaak LR, Uh HW, Reiding KR, Selman M, Houwing-Duistermaat JJ, Slagboom PE, Beekman M, Wuhrer M. Subclass-specific IgG glycosylation is associated with markers of inflammation and metabolic health. Sci Rep. 2017;7:12325. doi:10.1038/s41598-017-12495-0. PMID: 28951559.
- de Haan N, Reiding KR, Krištić J, Hipgrave Ederveen AL, Lauc G, Wuhrer M. The N-Glycosylation of Mouse Immunoglobulin G (IgG)-fragment crystallizable differs between IgG subclasses and strains. Front Immunol. 2017;8:608. PMID: 28620376. doi:10.3389/fimmu.2017.00608.
- Lund J, Takahashi N, Nakagawa H, Goodall M, Bentley T, Hindley SA, Tyler R, Jefferis R. Control of IgG/Fc glycosylation: a comparison of oligosaccharides from chimeric human/mouse and mouse subclass immunoglobulin Gs. Mol Immunol. 1993;30:741–48. doi:10.1016/0161-5890(93)90145-2. PMID: 8502242.
- Zhang Q, Joubert MK, Polozova A, De Guzman R, Lakamsani K, Kinderman F, Xiang D, Shami A, Miscalichi N, Flynn GC, et al. Glycan engineering reveals interrelated effects of terminal galactose and core fucose on antibody-dependent cell-mediated cytotoxicity. Biotechnol Prog. 2020;36:e3045. PMID: 32627435. doi:10.1002/btpr.3045.
- Thomann M, Reckermann K, Reusch D, Prasser J, Tejada ML. Fc-galactosylation modulates antibody-dependent cellular cytotoxicity of therapeutic antibodies. Mol Immunol. 2016;73:69–75. PMID: 27058641. doi:10.1016/j.molimm.2016.03.002.
- Aoyama M, Hashii N, Tsukimura W, Osumi K, Harazono A, Tada M, Kiyoshi M, Matsuda A, Ishii-Watabe A. Effects of terminal galactose residues in mannose α1-6 arm of Fc-glycan on the effector functions of therapeutic monoclonal antibodies. mAbs. 2019;11:826–36. doi:10.1080/19420862.2019.1608143. PMID: 30990348.
- Chow CK, Allan BW, Chai Q, Atwell S, Lu J. Therapeutic antibody engineering to improve viscosity and phase separation guided by crystal structure. Mol Pharm. 2016;13:915–23. doi:10.1021/acs.molpharmaceut.5b00817. PMID: 26849155.
- Mason BD, Zhang L, Remmele RL Jr., Zhang J. Opalescence of an IgG2 monoclonal antibody solution as it relates to liquid-liquid phase separation. J Pharm Sci. 2011;100:4587–96. doi:10.1002/jps.22650. PMID: 21638285.
- Kingsbury JS, Saini A, Auclair SM, Fu L, Lantz MM, Halloran KT, Calero-Rubio C, Schwenger W, Airiau CY, Zhang J, et al. A single molecular descriptor to predict solution behavior of therapeutic antibodies. Sci Adv. 2020;6:eabb0372. PMID: 32923611. doi:10.1126/sciadv.abb0372.
- Garber E, Demarest SJ. A broad range of Fab stabilities within a host of therapeutic IgGs. Biochem Biophys Res Commun. 2007;355:751–57. doi:10.1016/j.bbrc.2007.02.042. PMID: 17321501.
- Heads JT, Adams R, D’hooghe LE, Page MJ, Humphreys DP, Popplewell AG, Lawson AD, Henry AJ. Relative stabilities of IgG1 and IgG4 Fab domains: influence of the light-heavy interchain disulfide bond architecture. Protein Sci. 2012;21:1315–22. doi:10.1002/pro.2118. PMID: 22761163.
- Franey H, Brych SR, Kolvenbach CG, Rajan RS. Increased aggregation propensity of IgG2 subclass over IgG1: role of conformational changes and covalent character in isolated aggregates. Protein Sci. 2010;19:1601–15. doi:10.1002/pro.434. PMID: 20556807.
- Schön A, Freire E. Reversibility and irreversibility in the temperature denaturation of monoclonal antibodies. Anal Biochem. 2021;626:114240. PMID: 33964250. doi:10.1016/j.ab.2021.114240.
- Vlasak J, Ionescu R. Fragmentation of monoclonal antibodies. mAbs. 2011;3:253–63. doi:10.4161/mabs.3.3.15608. PMID: 21487244.
- Valliere-Douglass J, Jones L, Shpektor D, Kodama P, Wallace A, Balland A, Bailey R, Zhang Y. Separation and characterization of an IgG2 antibody containing a cyclic imide in CDR1 of light chain by hydrophobic interaction chromatography and mass spectrometry. Anal Chem. 2008;80:3168–74. doi:10.1021/ac702245c. PMID: 18355059.
- Nowak C, Kc J, Md S, Katiyar A, Bhat R, Sun J, Ponniah G, Neill A, Mason B, Beck A, et al. Forced degradation of recombinant monoclonal antibodies: a practical guide. mAbs. 2017;9:1217–30. PMID: 28853987. doi:10.1080/19420862.2017.1368602.
- Vlasak J, Ionescu R. Heterogeneity of monoclonal antibodies revealed by charge-sensitive methods. Curr Pharm Biotechnol. 2008;9:468–81. doi:10.2174/138920108786786402. PMID: 19075686.
- Lu X, Machiesky LA, De Mel N, Du Q, Xu W, Washabaugh M, Jiang XR, Wang J. Characterization of IgG1 Fc deamidation at asparagine 325 and its impact on antibody-dependent cell-mediated cytotoxicity and FcγRIIIa binding. Sci Rep. 2020;10:383. doi:10.1038/s41598-019-57184-2. PMID: 31941950.
- Yan Q, Huang M, Lewis MJ, Hu P. Structure based prediction of asparagine deamidation propensity in monoclonal antibodies. mAbs. 2018;10:901–12. doi:10.1080/19420862.2018.1478646. PMID: 29958069.
- Mimura Y, Church S, Ghirlando R, Ashton PR, Dong S, Goodall M, Lund J, Jefferis R. The influence of glycosylation on the thermal stability and effector function expression of human IgG1-Fc: properties of a series of truncated glycoforms. Mol Immunol. 2000;37:697–706. doi:10.1016/s0161-5890(00)00105-x. PMID: 11275255.
- Borrok MJ, Mody N, Lu X, Kuhn ML, Wu H, Dall’acqua WF, Tsui P. An “Fc-Silenced” IgG1 format with extended half-life designed for improved stability. J Pharm Sci. 2017;106:1008–17. doi:10.1016/j.xphs.2016.12.023. PMID: 28057542.
- Tam SH, McCarthy SG, Armstrong AA, Somani S, Wu SJ, Liu X, Gervais A, Ernst R, Saro D, Decker R, et al. Functional. Biophysical, and structural characterization of human IgG1 and IgG4 Fc variants with ablated immune functionality. Antibodies (Basel, Switzerland), 2017; Vol. 6. 10.3390/antib6030012. PMID: 31548527
- Huang L, Wang N, Mitchell CE, Brownlee T, Maple SR, De Felippis MR. A novel sample preparation for shotgun proteomics characterization of HCPs in antibodies. Anal Chem. 2017;89:5436–44. doi:10.1021/acs.analchem.7b00304. PMID: 28414239.