Blood vessel lining cells add anti-inflammatory sugar caps to antibodies

Why it matters: IgG antibodies carry a conserved sugar chain at Asn297 in the Fc region. A terminal sialic acid on this chain is thought to shift IgG toward anti-inflammatory activity, partly through the inhibitory receptor FcγRIIB. Epidemiological studies link higher IgG sialylation with pregnancy and disease remission, and lower sialylation with rheumatoid arthritis, tuberculosis and HIV. Yet how the body controls IgG sialylation has been unknown. The question matters more now that Fc-based drugs, including IVIg and Fc-fusion proteins, are widely used.

What was studied: The group's earlier work showed that deleting the sialyltransferase ST6Gal1 in B cells did not remove sialic acid from circulating IgG. B cells are inefficient at this, and sialylation does not happen in plasma or platelets. Here the researchers tested whether endothelial cells do it. These cells express FcRn, which pulls IgG out of the degradative endo/lysosomal pathway and returns it to circulation. They compared IgG from wild-type, FcRn-knockout and ST6Gal1-knockout mice using lectin ELISAs, with SNA detecting α2,6-linked sialic acid. They also incubated desialylated IgG with primary CD31+ mouse lung endothelial cells, the human endothelial line HMEC-1, and bone marrow-derived macrophages, then re-isolated the IgG from the medium.

Key results: Plasma IgG from FcRn-knockout mice was significantly less sialylated than wild-type IgG. It was not as low as in ST6Gal1-knockout mice, so FcRn greatly improves efficiency but is not strictly required. Other glycan features were largely unchanged. The exception was a higher ConA signal, which the authors suggest reflects better access to core mannose once sialic acid is lost. Bulk plasma glycoproteins and tissue sialylation were unaffected in FcRn-knockout mice, so the defect appears specific to IgG. Wild-type primary endothelial cells raised IgG sialylation about 13-fold over the desialylated input, while FcRn-knockout endothelial cells did not. Papain digestion placed most of the added sialic acid on the Fc glycan. HMEC-1 cells did the same, including with biotinylated IgG, which confirms that the originally added antibody was modified. Macrophages, with or without FcRn, did not sialylate IgG. The abstract and introduction also report that the activity depends on ST6Gal1 and occurs within FcRn recycling compartments. They report it is down-regulated by proinflammatory stimuli and up-regulated during gestation.

Implications: The findings propose that glycosylation of a secreted protein is not fixed when it leaves the producing cell, and can be remodeled by a distant cell type. The endothelium could be a regulator of humoral immunity. This offers a mechanism that fits decades of epidemiology on inflammation, pregnancy and IgG sialylation. It also suggests that the function of endogenous and therapeutic Fc-containing drugs could be altered by vascular state.

Caveats: Most data come from mice and cultured cells, with one human endothelial line. The assays are lectin-based, which indicates sialylation but does not give detailed glycan structures. Whether this process drives disease or clinical outcomes in humans remains untested. This summary draws on the abstract, introduction and early results sections available.