Engineering a human IgG2 antibody stable at low pH

Abstract

IgG2 subclass antibodies have unique properties that include low effector function and a rigid hinge region. Although some IgG2 subclasses have been clinically tested and approved for therapeutic use, they have a higher propensity than IgG1 for aggregation, which can curtail or abolish their biological activity and enhance their immunogenicity. In this regard, acid‐induced aggregation of monoclonal antibodies during purification and virus inactivation must be prevented. In the present study, we replaced the constant domain of IgG2 with that of IgG1, using anti‐2,4‐dinitrophenol (DNP) IgG2 as a model antibody, and investigated whether that would confer greater stability. While the anti‐DNP IgG2 antibody showed significant aggregation at low pH, this was reduced for the IgG2 antibody containing the IgG1 CH2 domain. Substituting three amino acids within the CH2 domain—namely, F300Y, V309L, and T339A (IgG2_YLA)—reduced aggregation at low pH and increased CH2 transition temperature, as determined by differential scanning calorimetric analysis. IgG2_YLA exhibited similar antigen‐binding capacity to IgG2, low affinity for FcγRIIIa, and low binding ability to C1q. The same YLA substitution also reduced the aggregation of panitumumab, another IgG2 antibody, at low pH. Our engineered human IgG2 antibody showed reduced aggregation during bioprocessing and provides a basis for designing improved IgG2 antibodies for therapeutic applications.

Abbreviations

ADCC

antibody‐dependent cellular cytotoxicity

CDC

complement‐dependent cytotoxicity

CH

heavy chain constant region

DSC

differential scanning calorimetry

HMWS

high molecular weight species

SDS‐PAGE

sodium dodecyl sulfate‐polyacrylamide gel electrophoresis

UHP‐SEC

ultrahigh‐pressure size‐exclusion chromatography

VH

heavy chain variable region

VL

light chain variable region

1. INTRODUCTION

Therapeutic antibodies have been approved for clinical use for a variety of diseases, including cancer, chronic and autoimmune diseases, and a number of novel antibodies are currently under development by several pharmaceutical companies.1 When developing IgG therapeutic antibodies, IgG subclass should be selected by carefully considering the requirement of an effector function. In general, IgG2 and IgG4 backbones are chosen only when a blocking function is required; an effector function is not desirable because these two subclasses have a more limited ability to activate antibody‐dependent cell‐mediated cytotoxicity (ADCC) and complement‐dependent cytotoxicity (CDC) than IgG1 and IgG3.2 IgG2 and IgG4 have unique physicochemical characteristics. IgG2 antibodies have different disulfide linkage isoforms IgG2A, IgG2B, and IgG2A/B, which display different functional potencies.3, 4 Unmodified IgG4 antibodies are functionally monovalent because of their ability to form Fab‐arm exchange in vivo and produce a hybrid molecule with bispecific antigen‐binding.5 Although some IgG2 and IgG4 antibody formats have been applied for therapy, some studies reported that IgG2 and IgG4 are prone to aggregation under low pH and heat‐induced stress, unlike IgG1.6, 7 Acid‐induced aggregation of individual IgG subclasses was reportedly ranked in the order of IgG1 < IgG2 < IgG4.8 Several therapeutic IgG1 antibody studies have suggested that aggregation exhibit either reduced or no biological activity and more importantly, may cause immunogenicity.9 Aggregate formation occurs at various steps of the manufacturing process, such as production, purification, formulation, and storage.10 In addition, a variety of conditions, including freeze‐thawing, agitation, thermal stress, and buffer formulation, contribute to aggregate formation.11 Therefore, it is critical to prevent the aggregation of therapeutic protein products during manufacturing, particularly at low pH, which is required for purification (e.g., protein A chromatography and virus inactivation) and has been shown to promote IgG aggregation.12, 13

With respect to IgG2 subclass aggregation, a previous study has reported that the major aggregation pathway at low pH 4.0 led to the formation both of dimers and high‐molecular weight (HMW) aggregates, suggesting that unfolding of the CH2 domain may drive the aggregation at low pH.14 Other studies have also suggested the importance of CH2 domain in reducing acid‐induced aggregate formation.6 Although the contribution is minor, the CH3 domain contribution to unfolding under low pH (under pH 3.1) has been reported.15

In this study, we attempted to reduce acid‐induced aggregation of the IgG2 subclass antibody via domain exchange of CH1, CH2, and CH3 as well as the hinge region in the IgG1 subclass. We observed that IgG2 antibodies in which the CH2 domain was replaced by that of IgG1 resisted aggregation at low pH. We also identified essential amino acids within the IgG2 CH2 domain, which are critical to minimize aggregate formation. Additionally, we evaluated the effect of substituting these amino acids on antibody bioactivity. Our engineered IgG2 antibody would contribute to the development of therapeutic IgG2 antibodies with low effector function and decreased aggregate formation under low pH treatment.

2. RESULTS

2.1. Design and generation of IgG1/IgG2 domain‐exchanged antibodies

We used anti‐DNP IgG1 and IgG2 as model antibodies. Heavy chain constant region (CH) and light chain constant region were expressed with identical heavy and light chain variable regions (VH and VL, respectively). To identify the critical domain for aggregate formation at low pH, three CH domains (CH1, CH2, and CH3) and the hinge region were exchanged between IgG1 and IgG2 (Figure 1). We constructed four domain‐exchanged antibodies, namely 1,122, 2,211, 2,212, and 2,221, with the four digits in each tag indicating the antibody subclass for CH1, hinge, CH2, and CH3 regions, respectively.

Figure 1.

Summary of the domain‐exchanged antibodies constructed in this study. Closed and open squares represent domains derived from human IgG1 or IgG2, respectively. All antibodies shared the variable regions of the anti‐DNP antibody and the light chain constant region of the Ck isotype

2.2. Aggregate formation by IgG2/IgG1 domain‐exchanged antibodies at low pH

The stability of the IgG1, IgG2, and domain‐exchanged IgG2 antibodies at low pH was investigated by ultrahigh‐pressure size‐exclusion chromatography (UHP‐SEC). Representative UHP‐SEC chromatograms of IgG2 are shown in Figure S1. The proportion of HMWS of IgG1 under low pH stress was <2% (Figure 2). IgG2 and IgG1122 showed around 14% HMWS formation after incubation at pH 3.5 for 60 min. In contrast, IgG2211, with both the CH2 and CH3 domains exchanged, showed around 2% HMWS formation. To determine which among CH2 and CH3 is critical for reducing aggregate formation, we investigated low pH stress of individual CH2 and CH3 domain‐exchanged antibodies (IgG2212 and IgG2221). The CH3 domain‐exchanged IgG2221 showed around 14% HMWS formation under low pH stress. In contrast, IgG2212, with only the CH2 domain‐exchanged, showed around 2% HMWS. These results revealed that CH2 is the most critical domain for IgG2 aggregate formation under low pH stress, although replacing both CH2 and CH3 domains of IgG2 with those of IgG1 had the most significant effect.

Figure 2.

HMWS of domain‐exchanged antibodies under conditions of low pH stress, as determined by UHP‐SEC. Antibodies were incubated at pH 3.5 for 10 or 60 min at 37°C. Data are presented as mean ± SD, and all experiments were performed in triplicate

2.3. Amino acid residues in the CH2 domain of IgG2 that are critical for aggregation at low pH

We next compared the amino acid sequences of the IgG1 and IgG2 CH2 domains to identify the amino acids within the CH2 domain that are critical for stabilization against low pH stress (Figure 3a). Six amino acids differ between IgG1 and IgG2 CH2 domains (positions 274, 296, 300, 309, 327, and 339 using European Union [EU] index numbering). We produced six antibodies, each with one of these six amino acids swapped between IgG2 and IgG1 and examined their behavior under low pH stress (Figure 3b). We intentionally excluded the CH2 N‐terminal residues, also called the lower hinge region, from the amino acid exchange between IgG2 and IgG1 to maintain the unique characteristic of IgG2.16, 17 A hinge region is important for agonistic activity and Fab‐arm flexibility and low effector function.17, 18 The lower hinge region of IgG2 lacking one of the two glycine residues at positions 235 and 236 is responsible for the shortest hinge of all IgG classes that related to the IgG2 rigid structure. Notably, a shorter and stiffer hinge region mediates resistance to microbial proteases.19, 20

Figure 3.

(a) Amino acid alignment of the heavy chain CH2 domains of human IgG1 and IgG2. (b) Summary of amino acid substitution antibodies constructed in this study. Gray bars indicate substitution sites, and open squares indicate human IgG2 domains. All antibodies shared the variable region of the anti‐DNP antibody and the light chain constant region of the Ck isotype. Substitutions and their positions are indicated

Among the six amino acid mutations, substituting F300Y, V309L, and T339A in IgG2 decreased HMWS formation relative to wild‐type IgG2 (Figure 4a). To determine the effect of the mutations in combination, we examined the effect of double substitutions—i.e., F300Y and V309L (F300Y_V309L), F300Y, and T339A (F300Y_T339A), and V309L and T339A (V309L_T339A) (Figure 3b). V309L_T339A, F300Y_T339A, and V309L_T339A combined substitution enhanced the effect of every single substitution and decreased the HMWS formation at pH 3.5 for 60 min compared to that by IgG2 (Δ10%). Finally, we generated an IgG2 variant with F300Y, V309L, and T339A triple substitution (IgG2_YLA), and it showed <5% HMWS even at extremely low pH (pH 3.5 for 60 min) (Figure 4a). We compared the low pH stability data of both IgG2212 and IgG2_YLA (Figure 4a). IgG2212 is a chimeric antibody in which all IgG2 CH2 amino acids are exchanged with IgG1 CH2 regions, which include the CH2 N‐terminal region. HMWS formation by IgG2212 was 2.5% under low pH stress (pH 3.5 for 60 min), and that by IgG2_YLA was 2.4%, indicating that IgG2_YLA has almost the same stability as IgG2212. These data suggest that IgG2_YLA achieved almost the same low pH stability as IgG2212 without exchanging amino acids in the CH2 lower hinge region.

Figure 4.

(a) Proportion of HMWS in preparations of anti‐DNP antibodies with multiple substitutions under low pH stress (IgG1 and IgG2 data are the same as in Figure 2). (b) The proportion of HMWS in preparations of pH‐stressed panitumumab, as determined by UHP‐SEC. Antibodies were incubated at pH 3.5 for 10 or 60 min at 37°C. Data are presented as mean ± SD, and all experiments were performed in triplicate

2.4. Generalizability of low pH stabilization mutation

We also introduced the most prominent stabilizing mutation (IgG2_YLA) into panitumumab, another IgG2 subclass antibody that targets the epidermal growth factor receptor (EGFR) and is routinely administered for colorectal cancer therapy. Panitumumab showed around 25% HMWS formation at pH 3.5 for 60 min. In contrast, YLA mutation‐stabilized panitumumab's HMWS was around 5% (Figure 4b).

2.5. Biological activity of IgG2 antibody variants

We examined the antigen‐binding activity of stabilized anti‐DNP IgG2 variants (IgG2_YLA) and found that it was comparable to that of wild‐type anti‐DNP IgG2 (IgG2) (Figure 5), indicating that the YLA mutation within the CH2 domain does not adversely impact antigen‐binding.

Figure 5.

ELISA analysis of anti‐DNP IgG2 and IgG2_YLA antibodies binding to DNP. Antibody concentrations of 0.6, 2.4, 9.8, 39, 156, 625, and 2,500 ng/ml were tested, and the binding of each antibody was evaluated using a horseradish peroxidase‐conjugated anti‐human IgG (H + L) secondary antibody by measuring the optical density (OD) at 450–570 nm

The C1q binding activity of anti‐DNP IgG2 variants (IgG2_YLA) was also evaluated. C1q is the first molecule that triggers CDC complement activation.21 IgG2 showed lower binding activity for C1q than IgG1, as previously reported (Figure 6a).19 The stabilized IgG2_YLA variant also showed a lower binding activity for C1q, similar to IgG2.

Figure 6.

(a) Effect of IgG2 CH2 domain amino acid substitutions on C1q binding, as determined by ELISA. Briefly, 96‐well microplates were coated with varying concentrations of IgG1, IgG2, and IgG2_YLA. After immobilization of the antibodies, microplates were blocked with 0.1% gelatin, including ELISA diluent buffer. Next, 2 μg/ml of human C1q was added and incubated. C1q binding was assessed using an anti‐C1q peroxidase‐conjugated secondary antibody by measuring the optical density (OD) at 450 nm. Data are presented as mean ± SD, and all experiments were performed in triplicate. (b) Effect of IgG2 CH2 domain amino acid substitutions on hFcγRIIIa (hFcγRIIIa‐158V, hFcγRIIIa‐158F) and FcRn binding, as determined by surface plasmon resonance measurement using a T100 biosensor instrument and CM5 sensor chip Biacore. Assays were performed with anti‐tetra‐His antibody‐immobilized CM5 sensor chips using an Amine Coupling Kit (BIAcore). The hexa‐His‐tagged hFcγRIIIa antibodies were captured by the immobilized anti‐tetra‐His antibodies. The data obtained by the injection of antibodies were corrected for the blank control prior to data analysis. The dissociation constant (KD) for hFcγRIIIa was calculated by steady‐state analysis using BIAcore T100 kinetic evaluation software (BIAcore)

Further, the binding profiles of the anti‐DNP IgG2 variants to human FcγRIIIa‐158V (high affinity) and FcγRIIIa‐158F (low affinity) allotypes that mediate ADCC activity, and a neonatal receptor (FcRn) for IgG, which is a key receptor for in vivo pharmacokinetics, were analyzed using the Biacore system (Figure 6b).22, 23 IgG2_YLA displayed no binding to FcγRIIIa‐158V and FcγRIIIa‐158F, similar to IgG2. Both IgG2 and IgG2_YLA showed a similar affinity for FcRn (pH 6.0 and 7.4) and a similar binding to FcγRIIIa and FcRn.

2.6. Thermodynamic stability of IgG2 antibody variants

To investigate the mechanistic basis for the stability of IgG2 variants against low pH stress, we carried out differential scanning calorimetry (DSC) analysis of IgG2_YLA in phosphate‐buffered saline (PBS) buffer (pH 7.4) and found that the onset temperature of unfolding (T _onset) of IgG2_YLA slightly exceeded that of IgG2 (63°C vs. 62°C) due to the YLA mutation within the CH2 domain (Figure 7a,b). On the other hand, the second unfolding transition temperature (i.e., T _m2) values were almost identical (75°C), likely because the antibodies have the same variable region. We also examined IgG2 and IgG2_YLA by DSC analysis at different pH values (pH 4.0, 5.0, and 6.0). As expected, the T_m1 value (CH2 domain) of IgG2_YLA was higher than that of IgG2 at each pH (4.0, 5.0, and 6.0). Especially, the difference in the Tm1 value between IgG2 and IgG2_YLA at low pH was higher than that at neutral pH (T _m1 value difference between IgG2 and IgG2_YLA at pH 4.0 was 5°C and that at pH 6.0 was 3°C) (Figure 7a,b). These data suggested that the YLA mutation increases the CH2 domain T _m1 value under low pH, thereby enhancing resistance to low pH stress.

Figure 7.

DSC analysis of the thermodynamic stability of antibody domains in various pH buffers. (a) DSC charts of IgG2 and IgG2_YLA antibodies at different pH buffers (4.0/5.0/6.0/7.4). (b) Melting transition (T _m, T _m1, T _m2, T _m3) and onset temperature (T _onset) of IgG2 variant antibodies was determined from the DSC analysis

3. DISCUSSION

Human IgG2 antibodies have unique physicochemical characteristics, such as the ability to form disulfide bond‐mediated structural isoforms that offer different biological activities and to assembly into covalent dimers in vivo.4, 24 IgG2 antibodies have a low effector function among the IgG subclasses, making them ideal for immunotherapeutic applications in which the effector function is not required.19 One of the most critical issues hindering the successful application of therapeutic IgG2 antibodies is aggregate formation during production, which can affect their efficacy and safety by eliciting unwanted immunogenic responses. In this study, we engineered an IgG2 antibody with a reduced propensity for aggregation at low pH. F300Y substitution within IgG2 CH2 domain markedly increased low pH stability. An X‐ray crystallography analysis of IgG1 revealed that the Y300 side chain is packed against H268, R292, and E294.6 It was previously suggested that at pH ~5, H268 and E294 engage in charge–charge interactions that stabilize the CH2 domain.6 Under more acidic conditions, this interaction may be abolished due to E294 protonation (theoretical pKa = 4.07); under such conditions, Y300, unlike F300, may form a hydrogen bond with H268 and thus preserve the folded structure.

V309L and T339A substitutions also showed increased stability relative to that of IgG2 at low pH. Val309 of IgG2 is located near the CH2–CH3 interface and is critical for aggregate formation. X‐ray crystallography analysis of IgG2 revealed that Leu309 of IgG1 and Val309 of IgG2 are hydrophobic amino acids with side chains oriented in the same direction.25 Sidechain differences may alter stability at low pH. Thr339 of IgG2 is also located near the CH2–CH3 interface. IgG1 Fc X‐ray crystallography revealed that Ala339 of IgG1 in the A chain interacts with CH3 D376 through two bridging water molecules (water 90 and 120); substituting Thr339 of IgG2 with Ala339 of IgG1 could enhance the CH2–CH3 interaction and increase low pH stability. Thus, the substitution of both V309L and T339A may alter the CH2 pivot and increase the resistance of IgG2 to low pH through an epistatic effect (Figure 8).

Figure 8.

Structure of the Fc region of human IgG1 and IgG2. The model was built from published structural data of human IgG1 (Protein Data Bank accession code: http://bioinformatics.org/firstglance/fgij//fg.htm?mol=3AVE) and IgG2 (Protein Data Bank accession code: http://bioinformatics.org/firstglance/fgij//fg.htm?mol=4HAF). Amino acids of human IgG1 and IgG2 described in the discussion section are highlighted in a different color

Compared to that of the IgG2 CH2 domain, substituting the CH3 domain of IgG1 with that of IgG2 did not significantly improve low pH stability. It was reported that the only sequence difference between the IgG1 and IgG2 CH3 domains is at residue 397; substituting IgG2 of M397 with V397 destabilized the CH3–CH3 interaction.26 Data for aggregate formation study under low pH stress showed that IgG2 CH3 domain‐exchanged antibody, IgG2221, showed a little improvement in low pH stability compared to that of IgG2; however, IgG2211 with both CH2 and CH3 domains of IgG2 replaced with those of IgG1 had the most significant effect in improving low pH stability, especially under the most severe conditions (pH 3.5, 60 min) (Figure 2). We only tested stability at pH value 3.5, which is sufficient for elution of protein A and virus inactivation during antibody purification.27, 28 Nuclear magnetic resonance data have shown that the unfolding of the CH3 domain occurs at pH values ranging from 2.5 and 3.1.15 Therefore, at pH values below 3.5, it may be necessary to further stabilize the CH3 domain to improve antibody resistance to low pH stress. Besides, some excipients reportedly affect the IgG2 antibody thermal unfolding temperature (T _m) at low pH.29 This indicates that an appropriate formulation may stabilize the IgG2_YLA antibody and decrease the aggregation during storage.

Regarding the biological activity of IgG2_YLA, a triple mutation within the CH2 domain did not abolish antigen‐binding compared with wild‐type IgG2. In relation to C1q binding, the IgG2_YLA antibody with the Y300 mutation showed reduced C1q binding (Figure 6a). Compared with IgG2, the Y300 residue of IgG1 does not directly bind to C1q but is positioned near the C1q binding site.30 P331S amino acid exchange is known to reduce binding for C1q.31 Thus, introducing the P331S mutation with IgG2_YLA may decrease C1q binding, even at a high concentration of 10 μg/ml. For ADCC activity, a strong correlation has been reported between ADCC activity and FcγRIIIa binding. Previous studies of the IgG1 Fc and FcγRIIIa complex structure revealed that all the three mutation sites, Y300, L309, and A339, do not belong to the binding site.32 Thus, these substitutions might not change the antibody's affinity for FcγRIIIa of IgG2_YLA. In relation to FcRn binding, it has been reported that residue IgG1 L309 is included in the key contact positions on the Fc region with FcRn.33 However, compared to that of IgG2, IgG2_YLA has almost the same KD with FcRn. Therefore, it is expected that IgG2_YLA and IgG2 have almost the same PK profile. However, further studies are needed to understand the PK profiles of IgG2_YLA fully.

When considering appropriate therapeutic applications, identifying the optimal antibody isotype is critical for therapeutic success. One study reported that rituximab formatted with IgG2 triggers programmed cell death more effectively than rituximab formatted with the standard IgG1 version.34 The heavy chain's hinge region and CH1 domain contain the molecular determinants for these effects. Another study demonstrated that the EGFR antibody of the human IgG2 isotype, panitumumab, mediates ADCC; however, in contrast to IgG1, the mechanism of panitumumab is limited to recruiting myeloid effector cells against EGFR‐positive tumor cells.35 The IgG2 subclass is predominantly selected for neutralizing antigens or inhibiting receptor‐ligand interactions.36 When IgG with low effector function but high resistance to acidic conditions is needed, IgG1 LALA or IgG1 N297V may be a better choice, as the disulfide isoforms of IgG2 limit the extent to which quality control is feasible.37, 38 Different structures and functions of the human IgG subclass may be advantageous for specific indications, depending on the targeted antigen or epitope, desired mode of action, pharmacokinetic properties, and biopharmaceutical considerations. Our engineered antibody IgG2_YLA contains only three amino acid substitutions within the constant CH2 domain outside of the hinge region and upper CH1 domain, which are integral to the unique character of IgG2. Considering that we introduced the least number of mutations in the wild‐type IgG2 sequence, our engineered antibody IgG2_YLA retained native IgG2 properties while its stability against low pH conditions was considerably boosted.

4. CONCLUSION

We engineered an IgG2 antibody with three amino acid substitutions (F300Y, V309L, and T339A) within the CH2 domain. This antibody was stable and resisted aggregation at low pH, which expectedly reduces immunogenicity and improves purification yield. Moreover, the antibody has negligible effector function and the same antigen‐binding capacity as wild‐type IgG2, making it ideal for therapeutic applications.

5. MATERIALS AND METHODS

5.1. Materials

All antibodies used in this study were expressed and purified by Kyowa Kirin (Tokyo, Japan). Briefly, antibodies expressed by Expi293F cells (Invitrogen, Carlsbad, CA) and secreted into the culture medium were purified with a protein A‐conjugated sepharose column (GE Healthcare, Chicago, IL). The buffer was replaced with Dulbecco's PBS; Gibco, Gaithersburg, MD) in the NAP25 desalting column (GE Healthcare). Antibody concentrations were adjusted to 1.0 mg/ml, and their purity was verified by sodium dodecyl sulfate‐polyacrylamide gel electrophoresis (SDS‐PAGE). The anti‐DNP antibodies in which the CH domain was replaced with that of IgG1 or containing amino acid substitutions within the CH2 domain were constructed: IgG1122, IgG2211, IgG2212, IgG2221, IgG2_Glu256Asp(E256D), IgG2_Met258Leu(M258L), IgG2_Asn292Lys(N292K), IgG2_Ile422Val(I422V), and IgG2_Arg425His(R425H) (according to EU numbering). The anti‐EGFR antibodies (Panitumumab) and amino acid substitutions within the CH2 domain were also constructed (Panitumumab_YLA). The carbohydrate structure of all antibodies was analyzed by mass spectrometry, and no significant difference was observed.

5.2. UHP‐SEC

The proportion of HMWS of all antibodies was analyzed by UHP‐SEC using an ACQUITY UPLC BEH200 SEC column (200 Å, 1.7 μm, 4.6 × 150 mm; Waters, Milford, MA). The mobile phase consisted of 20 mM sodium phosphate (pH 7.0) and 500 mM sodium chloride. The experimental conditions were as follows: flow rate, 0.5 ml/min; detection wavelength, 215 nm; and analysis time, 5.5 min. A total of 5 μg of each antibody was injected into the column. The peaks of monomers, HMWS, and low molecular weight species were analyzed using Empower 2 chromatography data software (Waters).

5.3. Evaluation of low pH stress stability

The pH of the antibody buffer solution was adjusted to 3.5 using 0.1 M citric acid buffer (pH 2.7). The antibodies were incubated at 37°C and pH 3.5 for 10 or 60 min in a PCR Thermocycler (Applied Systems, Foster City, CA) before neutralization with 500 mM phosphate buffer (pH 8.0). The samples were maintained at 4°C until the aggregation analysis. The degree of aggregation was analyzed by UHP‐SEC, as previously described.

5.4. Measurement of DNP‐binding activity

The DNP‐binding of each antibody was evaluated by ELISA. Briefly, 50 μL of 10 μg/ml bovine serum albumin conjugate of DNP (Molecular Probes) was coated on a microplate (NUNC) for 12 hr at 4°C. Diluted series of antibodies were incubated with the DNP‐coated surface for 2 hr at room temperature. The bound antibodies were detected with goat anti‐human kappa‐HRP (Southern Biotech) as a second antibody, and TMB substrate was used for measurement at 450 nm using a plate reader.

5.5. Preparation of hexa‐his‐tagged recombinant human FcγRIIIa and FcRn

Human FcγRIIIa (hFcγRIIIa‐158V, hFcγRIIIa‐158F) and hFcRn were prepared as previously described.39 Briefly, vectors encoding the receptors with hexa‐His‐tag substitution of transmembrane and intracellular domains were constructed and electroporated into Expi293F cells. The receptors were purified from cell culture supernatant using a nickel‐nitrilotriacetic acid Superflow column (Qiagen, Valencia, CA). Protein purity was confirmed by SDS‐PAGE.

5.6. Human FcγRIIIa and FcRn binding assay

Anti‐DNP antibody‐binding affinity for recombinant hFcγRIIIa (hFcγRIIIa‐158V, hFcγRIIIa‐158F) and hFcRn (pH 6.0 and pH 7.4) was tested with Biacore (GE Healthcare), as previously described, using human anti‐IgG2 isotype control antibody.40

5.7. C1q binding assay

Antibody‐binding capacity to human C1q (Quidel, San Diego, CA) was measured by ELISA, as previously described.21

5.8. DSC analysis

The thermal stability of individual antibody domains was evaluated by DSC. Measurements were performed with 0.5 mg/ml IgG in different pH buffer (pH 4.0/5.0/6.0/7.4) using a Micro Cal VP‐Capillary DSC system (Malvern Instruments, Malvern, UK). Temperature scans were performed from 25°C to 100°C at a scan rate of 1°C/min. A buffer‐buffer reference scan was subtracted from each sample scan before normalizing the concentration. Baselines were established using Origin7.0 (OriginLab, Northampton, MA) and cubic interpolation of the pre‐ and post‐transition baselines.

Supporting information

Supplementary Figure S1 UHP‐SEC analysis of IgG2.

Saito S, Namisaki H, Hiraishi K, Takahashi N, Iida S. Engineering a human IgG2 antibody stable at low pH. Protein Science. 2020;29:1186–1195. 10.1002/pro.3852