GWAS for urinary sodium and potassium excretion highlights pathways shared with cardiovascular traits
13 August 2019
Pazoki R, Evangelou E, Mosen-Ansorena D, Pinto RC, Karaman I, Blakeley P, Gill D, Zuber V, Elliott P, Tzoulaki I, Dehghan A.
Nature Communications (2019) 10:3653
Main findings: We performed GWAS of urinary sodium and potassium excretion (from spot urine) using linear mixed model (LMM) association testing implemented in BOLT-LMM (v2.3) software (Fig. 1; Supplementary Fig. 1)14. We, included ~8.8M single-nucleotide polymorphisms (SNPs) imputed to the Haplotype Reference Consortium (HRC) panel at MAF<0.5% from European ancestry participants in UKB (genotyping and imputation [GRCh37] data release 2017). Characteristics of the population are presented in Supplementary Table 1. Of the 50 novel loci identified for urinary sodium and 13 for urinary potassium excretion, 4 overlapped between sodium and potassium excretion (Supplementary Datas 1–6). Conditional analysis revealed no secondary signal. SNP-based heritability was 6.4% for urinary sodium and 4% for potassium excretion. The strongest urinary sodium locus was in MLIP gene (rs838133) (P=1.9×10−25) followed by CYP1A1 (rs2472297) (P=6.7×10−23) and FTO (rs11642015) (P=6.7×10−23) loci. MLIP is a muscular lamin A/C interacting protein with protein binding transcription factor activity. CYP1A1 encodes a cytochrome P450 superfamily enzyme involved in drug metabolism and lipid synthesis. It is also known for association with habitual coffee intake15. FTO is associated with body mass index (BMI) and other anthropometric traits16. We, additionally, identified SNPs within neuronal sodium channel and potassium channel loci SCN2A and KCD13, as well as ten microRNA and long intergenic noncoding RNA genes. The strongest urinary potassium signal was at ADRA2C (an alpha-2-adrenergic receptor) followed by CYP1A1 and AHR loci; the latter two genes have previously shown association with coffee intake15. Our sensitivity analysis (n=262,531) excluding participants who suffered from renal diseases or participants who used medications which may affect sodium and potassium excretion showed that 17 sodium SNPs annotated to AHR, MLIP, CYP1A1, ADH1B, LINC01114, LINC02424, RARB, LOC105378330, DCDC1, PKHD1, NOVA1-AS1, MLXIPL, FTO, MIR642A, GCKR, HTR4, and SCN2A and 3 potassium SNPs annotated to ADRA2C, SLC4A7, CYP1A1 remained genome-wide significant despite the large reduction in sample size (Supplementary Datas 7 and 8). Our LMM-based results showed that 33 urinary sodium lead SNPs and 8 urinary potassium SNPs remained strongly (P<1×10−5) associated with urinary traits, the slightly larger P value is likely due to the decrease in the sample size as the effect estimates are quite small. Effect estimates of the lead SNPs were correlated (r=0.98) before and after exclusion for urinary sodium. The correlation was 0.97 for urinary potassium SNPs. Genome-wide correlation of effect estimates before and after exclusion was 0.79 for urinary sodium and potassium SNPs.