Mapping of the gene network that regulates glycan clock of ageing

Abstract

Glycans are an essential structural component of immunoglobulin G (IgG) that modulate its structure and function. However, regulatory mechanisms behind this complex posttranslational modification are not well known. Previous genome-wide association studies (GWAS) identified 29 genomic regions involved in regulation of IgG glycosylation, but only a few were functionally validated. One of the key functional features of IgG glycosylation is the addition of galactose (galactosylation), a trait which was shown to be associated with ageing. We performed GWAS of IgG galactosylation (N=13,705) and identified 16 significantly associated loci, indicating that IgG galactosylation is regulated by a complex network of genes that extends beyond the galactosyltransferase enzyme that adds galactose to IgG glycans. Gene prioritization identified 37 candidate genes. Using a recently developed CRISPR/dCas9 system we manipulated gene expression of candidate genes in the in vitro IgG expression system. Upregulation of three genes, EEF1A1, MANBA and TNFRSF13B, changed the IgG glycome composition, which confirmed that these three genes are involved in IgG galactosylation in this in vitro expression system.

Keywords: CRISPR/dCas9; genome-wide association study; glycan clock; glycosylation; immunoglobulin G.

Figure 1

Genome-wide significant associations with galactosylation phenotypes. (A) Top SNP in identified genomic regions for each associated trait, (B) Venn diagram showing the number of genes mapped by positional mapping, chromatin interaction mapping, eQTL mapping and genome-wide gene-based association analysis (MAGMA), (C) Manhattan plot of genome-wide significant associations in IgG galactosylation GWAS with prioritized genes in each locus. Plot shows -log10(p-values) of association on y-axis and SNPs ordered by chromosomal location on x-axis. Red line indicates the genome-wide significance threshold (2.5 ×10^-8). Orange gene names indicate novel loci associated with IgG glycosylation.

Figure 2

Changes in IgG glycan composition upon targeting of selected GWAS loci associated with IgG galactosylation (HIVEP2, MANBA, TNFRSF13B and EEF1A1) in dCas9-VPR monoclonal cell lines. Samples containing non-targeting gRNAs served as controls. Changes in transcript levels are given as fold change values and changes in IgG phenotype are given as a relative change compared to control samples. (A) Manipulation of the HIVEP2 gene did not result in a statistically significant change in HIVEP2 transcript level, however, did induce a significant increase of digalactosylated structures (G2) with a concomitant decrease of agalactosylated IgG glycan structures (G0). (B) Targeting of MANBA by dCas9-VPR elevated transcription level of this gene which resulted in decrease of monogalactosylated IgG glycan structures (G1) (C) Targeting TNFRSF13B by dCas9-VPR resulted in ~ 400-fold increase of transcript levels and significant change of IgG agalactosylated IgG glycans (G0). (D) Successful upregulation of the EEF1A1 locus was followed by an increase of agalactosylated IgG glycans (G0) with a concomitant decrease of monogalactosylated IgG glycans (G1). Nominal p-value: *<0.05; **<0.01; ***<0.001; ns, not significant.

Figure 3

A graphical review of up-to-date functional validation of GWAS hits associated with IgG glycosylation using HEK293FS transient expression system with stably integrated dCas9-VPR or dCas9-KRAB fusions. When specific gRNA targeted dCas9-VPR to two estrogen associated genes, RUNX3 and SPINK4 (Mijakovac et al. 2021), as well as to the TNFRSF13B, MANBA and EEF1A1 loci, transcription was upregulated that resulted in decreased levels of galactosylated (red arrow within box), increased levels of agalactosylated IgG glycan structures (green arrow within box) or both. Downregulated transcription of SPPL3 resulted in an increase in galactosylated structures with a concomitant decrease in agalactosylated IgG glycan structures. Protein structures do not depict true protein structures in humans, generic protein shapes are chosen for easier visualization. Figure was created with BioRender.com. Accessed on 16 November 2021.