Molecular determinants of sialylated IgG anti-inflammatory activity

Abstract

IgG antibodies are the basis for many successful therapeutics. A single, N-linked glycan is present on the Fc on all IgGs, and the composition of that glycan exerts marked influence over effector functions of the IgG. We and others have shown terminal sialylation of the Fc glycan confers anti-inflammatory activity to the IgG1 subclass and is thought to be responsible for the anti-inflammatory activity of high-dose intravenous immunoglobulin. However, whether sialylation results in anti-inflammatory activity for other IgG subclasses is unknown. We found that IgG1 and IgG3, but not IgG2 nor IgG4, suppressed autoantibody-mediated inflammation in vivo when sialylated. This activity was dependent upon human DC-SIGN or its murine ortholog, SIGN-R1. Fc γ receptor-binding profiles for asialylated and sialylated IgG subclasses reinforced these similarities, with IgG1 and IgG3 sharing binding patterns. Amino acid sequence alignments identified two uniquely conserved amino acid residues at positions 234 and 327 of the CH2 domain of all human IgG1 and IgG3 allotypes that were distinct from IgG2 and IgG4. Indeed, molecular modeling of the shared residues in IgG1 and IgG3 revealed a hydrophobic-heavy interchain interaction that was not present in IgG2 nor IgG4. Introduction of those IgG1/3 residues into an IgG4-Fc resulted in anti-inflammatory activity in vivo when sialylated. Conversely, a reciprocal sialylated IgG1 mutant with IgG2/4 residues lacked anti-inflammatory activity. These results define the amino acid requirements of anti-inflammatory sialylated IgG and enable rational amino acid and glycan engineering across all therapeutic IgG subclasses.