Amber M Luckett, Carolyn McGrail, Kathryn Murrall, Emily N Griffin, Robin N Beaumont, Gareth Hawkes, William A Hagopian, Stephen S Rich, Michael N Weedon, Sarah Richardson, Richard A Oram, EXE-T1D consortium, Kyle J Gaulton
We observed only moderate genetic correlation between DR4-T1D and DR3-T1D (rg=0.6), which was lower compared with stratifications based on age of onset and sex, and distinct patterns of genetic correlations with other autoimmune diseases. Among type 1 diabetes-associated loci, the IL2 locus had significantly larger effect on type 1 diabetes in DR4 while several other loci (TAGAP, KLRG1) had more nominal heterogeneity. There was stronger enrichment of DR4-T1D-associated variants in T cell cREs and T cell-related pathways, while DR3-T1D-associated variants were specifically enriched in mast cell cREs and secretion-related pathways. We finally prioritised specific loci annotated in mast cells with stronger effects on type 1 diabetes in DR3 individuals.
AIMS/HYPOTHESIS: There are substantial differences in the clinical presentation of type 1 diabetes depending on the first-developing autoantibody, although the underlying mechanisms are poorly understood. The DR3-DQ2 (DR3) and DR4-DQ8 (DR4) haplotypes at the MHC locus associate with GADA and IAA as the first-developing autoantibodies, respectively, and can therefore be used as proxies for first autoantibody development in large cohort studies of type 1 diabetes cases and control individuals.
METHODS: We performed the first genome-wide association study of type 1 diabetes stratified by DR3 and DR4 status using 9091 type 1 diabetes cases and 14,157 control individuals from multiple cohorts. We estimated heritability and genetic correlation between type 1 diabetes with DR3 and DR4 (DR3-T1D and DR4-T1D, respectively), and with other immune and glycaemic phenotypes. We assessed heterogeneity in effects on type 1 diabetes between DR3 and DR4 individuals at known type 1 diabetes loci. We determined enrichment of type 1 diabetes heritability in DR3 and DR4 among variants in cell-type-specific cis-regulatory elements (cREs) and biological pathways, and annotated risk variants in cREs.
RESULTS: We observed only moderate genetic correlation between DR4-T1D and DR3-T1D (rg=0.6), which was lower compared with stratifications based on age of onset and sex, and distinct patterns of genetic correlations with other autoimmune diseases. Among type 1 diabetes-associated loci, the IL2 locus had significantly larger effect on type 1 diabetes in DR4 while several other loci (TAGAP, KLRG1) had more nominal heterogeneity. There was stronger enrichment of DR4-T1D-associated variants in T cell cREs and T cell-related pathways, while DR3-T1D-associated variants were specifically enriched in mast cell cREs and secretion-related pathways. We finally prioritised specific loci annotated in mast cells with stronger effects on type 1 diabetes in DR3 individuals.
CONCLUSIONS/INTERPRETATION: We performed the first GWAS of type 1 diabetes stratified by DR3 and DR4 status, which revealed heterogeneity in genetic risk and biological mechanisms dependent on high-risk HLA background.