BACKGROUND: Diabetic kidney disease (DKD) is a major complication of diabetes, but circulating proteins linking glycaemic stress to kidney injury remain unclear. We aimed to identify glycaemia-linked plasma proteins associated with DKD risk and assess their treatment responsiveness.
METHODS: We integrated Mendelian randomization, cis-protein quantitative trait locus-based summary-data-based Mendelian randomization (cis-pQTL-based SMR), deCODE protein genome-wide association study (GWAS) validation, prospective UK Biobank analysis, and semaglutide-associated proteomic assessment. We screened 2,923 plasma proteins for hemoglobin A1c (HbA1c)-driven effects, and evaluated incident DKD.
RESULTS: Fifty proteins were associated with genetically predicted HbA1c, and 10 were supported by independent HbA1c datasets. Cis-pQTL-based SMR analyses prioritized ITGBL1 and KLK13. In deCODE validation, KLK13 was the only candidate supported across both glycaemic trait-to-protein and protein-to-DKD analyses (OR 1.13, 95% CI 1.02-1.24). Among 9,936 UK Biobank participants without baseline DKD, 107 developed incident DKD. Higher KLK13 was associated with incident DKD (HR 1.52, 95% CI 1.13-2.06), whereas serine peptidase inhibitor Kazal type 6 (SPINK6) showed stronger association (HR 2.02, 95% CI 1.49-2.74). KLK13 decreased with semaglutide, but attenuated after weight and HbA1c adjustment.
CONCLUSIONS: KLK13 was prioritized as a glycaemia-linked circulating protein associated with DKD, implicating protease-inhibitor biology in DKD progression.