Modulating IgG effector function by Fc glycan engineering

Abstract

IgG antibodies contain a conserved N-glycosylation site on the Fc domain to which a complex, biantennary glycan is attached. The fine structures of this glycan modulate antibody effector functions by affecting the binding affinity of the Fc to diverse Fc receptor family members. For example, core fucosylation significantly decreases antibody-dependent cellular cytotoxicity (ADCC), whereas terminal α2,6-sialylation plays a critical role in the anti-inflammatory activity of human i.v. immunoglobulin therapy. The effect of specific combinations of sugars in the glycan on ADCC remains to be further addressed, however. Therefore, we synthesized structurally well-defined homogeneous glycoforms of antibodies with different combinations of fucosylation and sialylation and performed side-by-side in vitro FcγR-binding analyses, cell-based ADCC assays, and in vivo IgG-mediated cellular depletion studies. We found that core fucosylation exerted a significant adverse effect on FcγRIIIA binding, in vitro ADCC, and in vivo IgG-mediated cellular depletion, regardless of sialylation status. In contrast, the effect of sialylation on ADCC was dependent on the status of core fucosylation. Sialylation in the context of core fucosylation significantly decreased ADCC in a cell-based assay and suppressed antibody-mediated cell killing in vivo. In contrast, in the absence of fucosylation, sialylation did not adversely impact ADCC.

Keywords: Fc receptor; IgG; glycosylation.

Fig. 1.

Synthesis of homogeneous antibody glycoforms via chemoenzymatic glycosylation remodeling of the Fc glycans. Structures of Fc glycans on the synthetic rituximab glycoforms used for the present study: S2G2F, sialylated/core-fucosylated glycoform; G2F, asialylated/core-fucosylated; S2G2, sialylated/afucosylated; G2, asialylated/afucosylated; and G_(0-2)F, the commercial rituximab with core fucosylation and varied galactosylation.

Fig. S1.

LC-MS analysis of the glycosylation remodeling of rituximab. (A) LC-MS (after deconvolution) of the heavy chain of the commercial rituximab. (B) LC-MS of the heavy chain of transglycosylation product 5 (S2G2F-rituximab). (C) LC-MS of the heavy chain of transglycosylation product 6 (G2F-rituximab). (D) LC-MS of the heavy chain of transglycosylation product 7 (S2G2-rituximab). (E) LC-MS of the heavy chain of transglycosylation product 8 (G2-rituximab).

Fig. S2.

LC-MS analysis of the glycosylation remodeling of anti-mCD4 hIgG1 (GK1.5). (A) LC-MS (after deconvolution) of the heavy chain of GK1.5 expressed in 293T cells. (B) LC-MS of the heavy chain of transglycosylation product S2G2F-GK1.5. (C) LC-MS of the heavy chain of transglycosylation product G2F-GK1.5. (D) LC-MS of the heavy chain of transglycosylation product S2G2-GK1.5. (E) LC-MS of the heavy chain of transglycosylation product G2-GK1.5.

Fig. 2.

Binding of rituximab glycovariants to human FcγRIIIA-V158 (A) and FcγRIIIA-F158 (B). In A, FcγRIIIA-V158 (0.5 µg/mL) was coated onto a 96-well plate overnight. Binding was tested with rituximab glycoforms in concentrations ranging from 250 nM to 0.016 nM (fivefold serial dilutions). In B, FcγRIIIA-F158 (5 µg/mL) was coated onto a 96-well plate overnight. For afucosylated rituximab glycoforms, binding was tested using concentrations ranging from 750 nM to 0.016 nM (fivefold serial dilutions). For fucosylated rituximab glycoforms, binding was tested using concentrations ranging from 4,000 nM to 0.98 nM (fourfold serial dilutions).

Fig. 3.

Effects of core fucosylation and sialylation on in vitro ADCC. Raji cells were coated with the indicated concentrations of each glycovariant, and the cells were washed and incubated at a 1:1 ratio with PBMCs from four different leukocyte donors for 3.5 h before cell surface CD107a expression was assessed on CD56⁺CD3⁻ NK cells by immunofluorescence staining with flow cytometry analysis. Values represent the mean ± SEM frequency of CD107a⁺ cells among NK cells from cultures with the indicated cells from four individual leukocyte donors. Significant differences between the indicated sample and the corresponding G2F glycovariant sample are shown. *P < 0.05; **P < 0.01.

Fig. 4.

Effects of core fucosylation and sialylation on in vivo killing of B cells in FcγR-humanized mice. huCD20/FcγR-humanized mice (n = 4 per group) were given 2 mg/kg of the indicated rituximab huIgG1 glycovariant 2 d before analysis of the blood. (A) Mean ± SEM frequencies of B220⁺ cells among live, single lymphocytes. (B) Mean ± SEM absolute number of B220⁺ cells. Significant differences between the indicated sample and the G2F glycovariant sample are shown. *P < 0.05; **P < 0.01.

Fig. 5.

Effects of core fucosylation and sialylation on in vivo killing of CD4⁺ T cells in FcγR-humanized mice. FcγR-humanized mice (n = 3 per group) were given 2 mg/kg of the indicated GK1.5 huIgG1 glycovariant 2 d before analysis of the blood. (A) Mean ± SEM frequencies of CD4⁺ cells among live, single lymphocytes from the indicated tissue. (B) Mean ± SEM absolute number of CD4⁺ cells. Significant differences between the indicated sample and the corresponding G2F glycovariant sample are shown. *P < 0.05; **P < 0.01.