In this cohort study of patients with predialysis CKD of Chinese ethnicity, we found that the T allele of rs11864909 in the promotor of UMOD gene was associated with higher levels of eGFR and lower levels of serum uromodulin. The associations were consistent through different age, hypertension, eGFR, urinary ACR and causes of CKD groups. Furthermore, rs13333226, rs4293393 and rs6497476, also common variants of UMOD gene, were associated with the occurrence of all-cause mortality. When stratified by levels of 24-hour urinary sodium excretion, the association was only present among participants with the elevated excretion level of the urinary electrolyte.
Several common variants located in the UMOD gene promotor, such as rs12917707, rs13333226, rs6497476 and rs4293393, have been detected significantly associated with CKD-related traits in the recently published GWAS among populations of European descent (7, 9). Regarding the east Asians, another polymorphism in the promotor region of UMOD, rs11864909, was reported as the most significant genome-wide association signal, taking the place of rs12917707, owing to a very low minor allele frequency (< 0.01) of which among east Asians(10). In the current study of patients with CKD, we used a candidate gene strategy to select the above SNPs and confirmed the previous findings that T allele of rs11864909 was associated with elevated level of eGFR. Although we also detected either A allele of rs4293393/rs13333226 or T allele of rs6497476 was associated with higher concentration of eGFR, the differences failed to reach significant threshold due to extremely high frequency of the alleles (99.7% for both A allele of rs4293393/ rs13333226 and T allele of rs6497476). Besides eGFR, we measured serum uromodulin in the current study, which was positively correlated with the level of eGFR in previous studies(11, 12). We demonstrated that T allele of rs11864909 was associated with lower levels of the trait, which was consistent with the findings by Graciela and colleagues, where rs12917707 was associated with serum uromodulin based on a GWAS among participants of the Ludwigshafen Risk and Cardiovascular Health Study (2826 of the 3316 population with genotyping data). Likewise, genotypes of rs4293393, rs13333226 and rs6497476 were not significantly associated with the level of serum uromodulin in their study(4).
Besides CKD-related phenotypes, a recent GWAS has also reported the association between common variants of UMOD gene and hypertension(13). Another study by Gudbjartsson et al. additionally found that variants at UMOD was more strongly associated with CKD among older adults and those with multiple comorbidities, such as hypertension, diabetes and CVD, highlighting the role of the risk variants involved in the mechanisms relating to adaptation to aging(9). In fact, a basic medicine study, conducted by Trudu and colleagues, provided compelling evidence regarding the mechanisms linking the variants at UMOD and the development of hypertension and kidney lesions. They found overexpression of uromodulin due to the presence of UMOD risk variants both in vitro and in vivo, which could cause over-activation of TAL sodium–potassium–chloride co-transporter, leading to salt-sensitive hypertension and age-dependent kidney lesions. Blocking the pathway by diuretics could result in drop of BP(14). In the current study, we did not detect a significant relationship between all the studied SNPs and BP, which was consistent with the findings by Gudbjartsson and colleagues(9). However, we found that associations between rs11864909 and eGFR or serum uromodulin were more prominent among participants with higher urinary ACR (≥ 300 mg/g), which represents advanced kidney injury and is highly correlated with long-term hypertension and diabetes. However, we do not find advanced age or hypertension could modify the effect of the variant.
Regarding the relationship between variants of UMOD gene and adverse outcomes of CKD, the above-mentioned Ludwigshafen Risk and Cardiovascular Health Study also reported that T allele of rs12917707 was shown to be associated with reduced risk of mortality among those aged < 67 years after a median 9.9 years of follow-up(4). In our study, the majority of population was also aged < 67 years (89.53%) and we found consistent results that G allele of rs4293393/ rs13333226 and C allele of rs6497476 were associated with reduced risk of mortality. However, this association was in contrast to the serum uromodulin-lowering effect of the T allele of rs12917707, because the lower level of the serum biomarker was associated with increased risk of mortality as revealed in the above-mentioned study(4). In a previous publication, we also detected increased risk of ESKD associated with lower level of serum uromodulin based on the same population with the current study(5). However, there is also evidence supporting the protective role of the serum uromodulin-lowering allele of the variants of UMOD gene regarding risk of CVD. Based on the large-scale Malmo Diet and Cancer study with the exclusion of prior CVD events at baseline and with a follow-up of 12 years (n = 26654), Sandosh et al. reported each copy of the G allele of rs13333226 was associated with a 7.7% reduction for risk of CVD after adjusting for age, sex and body mass index, suggesting protective effect on risk of CVD of the serum uromodulin-lowering allele of the variants of UMOD gene. Only a tiny abbreviation was observed when SBP and/or DBP was added into the regression model(13).
In our study, we also detected that the association between G allele of rs4293393/ rs13333226 or C allele of rs6497476 and reduced risk of mortality was only present in those with larger-than median level of 24 h-urinary sodium excretion (≥ 135 mmol/24 h). However, important confounding factors were not adjusted due to the limited number of events in certain genotype groups. Further studies with large sample size, especially with longer follow-up time and plenty of events, are needed to validate the findings.
The C-STRIDE study recruited a large sample of patients with CKD around mainland China and evaluated their clinical characteristics comprehensively. Despite the advantages, some limitations that should be admitted. Baseline eGFR and serum uromodulin were measured only once and may be subject to instability. Follow-up time for progression of CKD and occurrence of outcomes was comparatively short, which may limit the study power due to the limited number of events recorded.