Genome-wide aggregated trans-effects analysis for circulating proteins indicates a key role of immune checkpoints in type 1 diabetes

The “omnigenic” hypothesis postulates that polygenic effects of common variants on typical complex traits coalesce via trans-effects on the expression of a relatively sparse set of “core” effector genes and their encoded proteins in relevant tissues. The objective of this study was to identify core proteins for type 1 diabetes. We used summary statistics for SNP associations with plasma levels of 5130 proteins in three large cohorts to compute genome-wide aggregated trans-effects (GATE) scores for protein levels in two type 1 diabetes case-control studies (6828 cases, 416,000 controls). GATE scores for 27 proteins were associated with type 1 diabetes. Fourteen were replicated between datasets, 11 had support in Mendelian randomisation analysis and 9 had experimental support in mouse models of autoimmune diabetes. The strongest associations were for immune checkpoints (PDCD1, CD5, TIGIT, LAG3), chemokines and innate immune system proteins (NCR1, KLRB1). Whilst PDCD1 is a known cause of monogenic autoimmune diabetes, neither it nor most of the core proteins identified here were previously reported as GWAS hits for type 1 diabetes. These results identify possible drug targets with genetic support for causality and suggest that PD-1 agonists under development for other indications should be trialled for type 1 diabetes prevention.