Fc Galactosylation Promotes Hexamerization of Human IgG1, Leading to Enhanced Classical Complement Activation

Abstract

Human IgG contains one evolutionarily conserved N-linked glycan in its Fc region at position 297. This glycan is crucial for Fc-mediated functions, including its induction of the classical complement cascade. This is induced after target recognition through the IgG-Fab regions, allowing neighboring IgG-Fc tails to associate through Fc:Fc interaction, ultimately leading to hexamer formation. This hexamerization seems crucial for IgG to enable efficient interaction with the globular heads of the first complement component C1q and subsequent complement activation. In this study, we show that galactose incorporated in the IgG1-Fc enhances C1q binding, C4, C3 deposition, and complement-dependent cellular cytotoxicity in human erythrocytes and Raji cells. IgG1-Fc sialylation slightly enhanced binding of C1q, but had little effect on downstream complement activation. Using various mutations that decrease or increase hexamerization capacity of IgG1, we show that IgG1-Fc galactosylation has no intrinsic effect on C1q binding to IgG1, but enhances IgG1 hexamerization potential and, thereby, complement activation. These data suggest that the therapeutic potential of Abs can be amplified without introducing immunogenic mutations, by relatively simple glycoengineering.

FIGURE 1.

IgG–Fc glycosylation. (A) Schematic representation of the IgG Y-shaped structure; the Fc-N297 glycan and Fab and Fc regions are indicated. (B) Composition of the N297-glycan biantennary structure with the distinct sugar groups and their respective locations. (C) Ribbon structure of dimeric IgG–Fc regions with N-glycans (1HZH); highlighted areas indicate the N297-glycosylation site (pink) or mutated amino acids involved in Fc interaction presented in this study (E345 [red], E430 [blue], K439 [green], and S440 [orange]).

FIGURE 2.

IgG glycoengineering for altered galactosylation and sialylation. (A) Schematic setup for the production process in HEK FreeStyle cells, with the addition of relevant substrates and constructs coding for enzymes prior/during transfection. (B and C) Fc glycosylation profiles of produced anti-biotin mAbs using different glycoengineering techniques to increase Fc galactosylation and sialylation, analyzed by mass spectrometry. The bar graphs represent the mean and SEM of five different mAbs. For the statistical analysis, an ordinary one-way ANOVA with Tukey multicomparison test was performed. (D–F) Relative levels of IgG binding of produced anti-biotin mAbs to 5× biotin/BSA are presented as relative value to the maximum response of the unmodified WT mAb, determined by ELISA (n = 3). Curve fitting was performed using nonlinear regression dose-response curves with log(agonist) versus response–variable slope (four parameters) in GraphPad Prism 8.0.2. No differences in opsonization were observed between glycovariants. **p ≤ 0.01, ****p ≤ 0.0001.

FIGURE 3.

Complement-activating properties of glycoengineered and mutated anti-biotin mAbs shown for (A) all unmodified anti-biotin variants and (B–G) glycoengineered variants. In (A1–A7), relative binding of C1q (n = 4) is shown. (B1–B7) Relative deposition of C4 (n = 4). (C1–C7) Relative deposition of C3 (n = 4). (D1–D7) Complement-mediated lysis of biotinylated RBCs (n = 4). All C1q binding and C4 and C3 deposition were determined by ELISA. Data represent the mean and SEM of four independent experiments; all values were presented as relative value to the maximum response of the unmodified WT mAb. Curve fitting was performed using nonlinear regression dose-response curves with log(agonist) versus response–variable slope (four parameters) in GraphPad Prism 8.0.2. The maximum response and EC₅₀ were calculated, which are presented in Fig. 4. Mutants RGY and E430G are more potent in activating the complement system compared with the WT, whereas E439K and S440K perform worse. Differences in complement activation between glycoengineered variants were only observed for the WT, E439K, and S440K, not for RGY and E430G.

FIGURE 4.

The maximum response and EC₅₀ of complement-activating properties of glycoengineered and mutated anti-biotin mAbs extracted from (Fig. 3. (A–D) Maximum response. (E–H) EC₅₀ values. Data represent the mean and SEM of four independent experiments; the maximum response is presented as relative value to the maximum response of the unmodified WT mAb. For the statistical analysis, an ordinary one-way ANOVA with Tukey multicomparison test was performed. Nonsignificant comparisons were not shown. Higher levels of Fc galactosylation and Fc sialylation significantly induce the maximum response and reduce the EC₅₀ of complement activation by WT, E439K, and S440K mAbs. These differences were not observed for RGY and E430G mAbs. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.

FIGURE 5.

Complement-mediated lysis of biotinylated Raji cells. (A) Flow cytometric gating strategy; cells were gated based on the forward-/side-scatter, and the percentage of dead cells was calculated using the LIVE/DEAD Fixable Near-IR Dead Cell Stain. (B) Complement-mediated lysis with unmodified anti-biotin variants. (C–H) Complement-mediated lysis using glycoengineered anti-biotin variants. Data represent the mean and SEM of three independent experiments; curve fitting was performed using nonlinear regression dose-response curves with log(agonist) versus response–variable slope (four parameters) in GraphPad Prism 8.0.2. (I) The maximum response and (J) EC₅₀ of complement-mediated lysis of glycoengineered and mutated anti-biotin mAbs extracted from (Fig. 5C–H. For the statistical analysis, an ordinary one-way ANOVA with Tukey multicomparison test was performed. Nonsignificant comparisons were not shown. *p ≤ 0.05, **p ≤ 0.01, ****p ≤ 0.0001.

FIGURE 6.

Hexamerization of RGY mAbs is facilitated by galactosylation. (A) The oligomerization of the RGY Abs was visualized by SEC-MALS and was observed in a concentration-dependent manner. Approximately 90% of the Ab was present as hexamer in solution at a concentration of 1 mg/ml. However, the lower the concentration the higher the abundancy of monomeric IgG. Absorbance was measured at 280 nm (OD), and normalization was performed using the sum of all values in the data set (area under the curve [AUC]), shown on the left y-axis. Molecular mass was measured using multiangle light scattering (MALS) and shown on the right y-axis. (B–E) Representative SEC curves of glycoengineered RGY mAbs at different concentrations. Normalization was performed based on the AUC. (F) Fc galactosylation significantly induces oligomerization of RGY mAbs at all concentrations. Data represent the mean and SEM of three to five independent samples. For the statistical analysis, an ordinary one-way ANOVA with Tukey multicomparison test was performed. *p ≤ 0.05, ****p ≤ 0.0001.