METHOD: We performed a meta-analysis of kidney failure genome-wide association studies, using kidney transplant recipients as cases (n = 6942) and donors as controls (n = 4788). Secondary disease-specific analyses were performed for kidney failure due to diabetes, IgA nephropathy, glomerulonephritis or polycystic kidney disease. Subsequently, we investigated genetic overlap with eGFR and urinary albumin-creatinine ratio (UACR) variability, based on publicly available CKDgen consortium meta-analysis data.
RESULTS: In the primary analysis, we found two suggestive hits for kidney failure: rs17046239 in GRM7 (P = 8.9 × 10−8), and rs9273431 in HLA-DQB1 (P = 5.3 × 10−8). In disease-specific analyses, we found three genome-wide (P < 5 × 10−8) significant hits for kidney failure due to diabetes: rs9273431 in HLA-DQB1 (P = 5.0 × 10−50), rs2476601 in PTPN22 (P = 2.9 × 10−13) and rs7110099 in INS-IGF2 (P = 4.4 × 10−9). Furthermore, we found suggestive hits for kidney failure due to glomerulonephritis (rs6531751, nearest gene: PDS5a, P = 9.3 × 10−8) or polycystic kidney disease (rs111857047 in PTPRD, P = 9.6 × 10−8). As follow-up analysis, we performed lookups for the identified SNPs in prior kidney function GWAS results of the CKDGen consortium. Our top hits for kidney failure due to diabetes were nominally significant (P < 0.05) in the creatinine-based estimated glomerular filtration rate (eGFRcrea) and/or the UACR results. The other hits were not significantly associated with either phenotype. Linkage disequilibrium score regression (LDSC) analysis did not reveal a significant genetic correlation (rG) between kidney failure and eGFRcrea or UACR. There was significant genetic overlap of type 2 diabetes with both kidney failure due to any cause (rG = 0.27; P = 0.0027) and kidney failure due to diabetes (rG = 0.57; P = 0.0002). In silico sequencing revealed that our diabetes-induced kidney failure hits had previously been associated with auto-immunity and type 1 diabetes; the other hits yielded no results. A UK Biobank phenome-wide association scan confirmed the associations of the three diabetes-specific kidney failure genetic hits with auto-immunity and diabetes phenotypes, and additionally linked them to thyroid function and blood cell counts.
CONCLUSION: We identified three genome-wide significant variants associated with kidney failure due to diabetes and four additional suggestive hits associated with cause-specific or overall kidney failure. The lack of strong genetic overlap with eGFR and UACR suggests that there is a separate genetic component that drives the risk of progression to kidney failure, independent of normal kidney function.
- Kidney failure
- Kidney function
- Diabetes Mellitus