Identification of C21orf59 and ATG2A as novel determinants of renal function-related traits in Japanese by exome-wide association studies

INTRODUCTION

Chronic kidney disease (CKD) is an important risk factor for cardiovascular disease as well as end-stage renal disease [1–6]. Genome-wide association studies (GWASs) have identified various genes and loci for renal function-related traits in European ancestry [7–11], African-American [12, 13], or East Asian [14] populations. A recent meta-analysis of GWASs in European ancestry populations identified 53 loci that were significantly related to estimated glomerular filtration rate (eGFR) [15].

Hyperuricemia is a crucial risk factor for gout, a common inflammatory type of arthritis [16, 17], cardiovascular disease [18], and cancer [19]. Although multiple renal transporters contribute to the maintenance of normal circulating uric acid levels by mediating the excretion or reabsorption of uric acid in the proximal kidney tubules, the underlying mechanisms of such homeostasis have not been fully elucidated [20]. The heritability of the serum concentration of uric acid has been estimated to be 40% [21], suggesting that genetic variants contribute to regulation of this parameter by influencing uric acid synthesis, excretion, or reabsorption [21, 22]. Previous GWASs have identified single nucleotide polymorphisms (SNPs) significantly associated with the serum uric acid concentration or the prevalence of gout [23–29]. A large-scale GWAS in European ancestry populations identified 28 loci that influence the serum concentration of uric acid [30].

Most genetic variants identified in these various previous GWASs were common SNPs with a minor allele frequency (MAF) of >5% and a small individual effect size. Given that these common SNPs may explain a small fraction of the heritability of CKD or hyperuricemia, it is expected that low-frequency (0.5% ≤ MAF < 5%) or rare (MAF <0.5%) variants with larger effect sizes also contribute to the genetic architecture of these conditions [31]. Although several polymorphisms have been found to be significantly associated with renal function [14] or gout [32, 33] in Japanese, genetic variants—including low-frequency and rare variants—that influence renal function-related traits or contribute to predisposition to CKD or hyperuricemia in the Japanese population have not been identified definitively.

We have now performed exome-wide association studies (EWASs) with the use of exome array-based genotyping methods in order to identify SNPs—especially low-frequency or rare coding variants with moderate to large effect sizes—that influence renal function-related traits or confer susceptibility to CKD or hyperuricemia in Japanese. Given that most of the known low-frequency or rare variants were not included in arrays adopted in previous related GWASs, we used Illumina human exome arrays that provide coverage of functional SNPs in entire exons including low-frequency and rare variants.

RESULTS

EWASs for eGFR and serum concentrations of creatinine and uric acid

We examined the relation of genotypes for 41,352 SNPs that passed quality control to eGFR or the serum concentration of creatinine in 12,565 subjects by linear regression analysis. Manhattan plots of the EWASs for eGFR and the serum creatinine concentration are shown in Supplementary Figure 1. After Bonferroni’s correction, 25 and seven SNPs were significantly [P < 1.21 × 10^–6 (0.05/41,352)] associated with eGFR (Table 1) or the serum concentration of creatinine (Table 2), respectively. None of these SNPs was associated with both eGFR and serum creatinine concentration.

We next examined the relation of genotypes for 41,372 SNPs that passed quality control to the serum concentration of uric acid in 9934 subjects not taking uric acid-lowering medications by linear regression analysis. A Manhattan plot for this EWAS is also shown in Supplementary Figure 1. After Bonferroni’s correction, six SNPs were significantly [P < 1.21 × 10^–6 (0.05/41,372)] associated with the serum concentration of uric acid (Table  3).

EWASs for CKD and hyperuricemia

We performed an EWAS for CKD with 5161 subjects [3270 individuals with CKD (eGFR of <60 mL min^–1 1.73 m^–2), 1891 controls (eGFR of ≥90 mL min^–1 1.73 m^–2)] (Table 4). Age, the frequency of men, body mass index, and the prevalence of hypertension, diabetes mellitus, dyslipidemia, and hyperuricemia as well as systolic blood pressure, fasting plasma glucose level, blood glycosylated hemoglobin (hemoglobin A_1c) content, and serum concentrations of triglycerides and uric acid were greater, whereas the serum concentration of high density lipoprotein (HDL)–cholesterol and hemoglobin concentration were lower, in subjects with CKD than in controls.

We examined the relation of allele frequencies of 41,352 SNPs to CKD with Fisher’s exact test. A Manhattan plot for the EWAS of CKD is shown in Supplementary Figure 2. After Bonferroni’s correction, 49 SNPs were significantly [P < 1.21 × 10^–6 (0.05/41,352)] associated with CKD (Supplementary Table 1). The genotype distributions for these SNPs were in Hardy-Weinberg equilibrium (P > 0.001) among both subjects with CKD and controls (Supplementary Table 2).

The relation of the 49 SNPs identified by the EWAS to CKD was examined further by multivariable logistic regression analysis with adjustment for age, sex, and the prevalence of hypertension and diabetes mellitus (Supplementary Table 3). Four SNPs (rs707926 of VARS, rs76974938 of C21orf59, rs112311672 of HDAC10, rs41272317 of ACAD11) were related (P < 0.05 in at least one genetic model) to CKD, although there was no SNP significantly [P < 2.55 × 10^–4 (0.05/196)] associated with this condition (Table 5). The minor alleles of these SNPs were all risk factors for CKD. The rs76974938 [C/T (D67N)] SNP of C21orf59 was associated with both eGFR and CKD.

We next performed an EWAS for hyperuricemia with 11,686 subjects (2045 individuals with hyperuricemia, 9641 controls), the characteristics of whom are shown in Table 6. Age, the frequency of men, body mass index, and the prevalence of smoking, hypertension, diabetes mellitus, dyslipidemia, and CKD were greater in subjects with hyperuricemia than in controls.

We examined the relation of allele frequencies of 41,372 SNPs to hyperuricemia with Fisher’s exact test. A Manhattan plot for the EWAS is shown in Supplementary Figure 2. After Bonferroni’s correction, 35 SNPs were significantly [P < 1.21 × 10^–6 (0.05/41,372)] associated with hyperuricemia (Supplementary Table 4). The genotype distributions for these SNPs were in Hardy-Weinberg equilibrium (P > 0.001) among both subjects with hyperuricemia and controls (Supplementary Table 5).

The relation of the 35 SNPs identified by the EWAS to hyperuricemia was examined further by multivariable logistic regression analysis with adjustment for age and sex (Supplementary Table 6). Three SNPs (rs115445569 of ACOT11, rs116911833 of TRIM7, rs60854092 of NOTCH2) were related (P < 0.05 in at least one genetic model) to hyperuricemia, although there was no SNP significantly [P < 3.57 × 10^–4 (0.05/140)] associated with this condition (Table 7). The minor T and A alleles of rs115445569 and rs116911833, respectively, were risk factors for hyperuricemia, whereas the minor A allele of rs60854092 was protective against hyperuricemia.

Relation of identified SNPs to eGFR or serum concentrations of creatinine or uric acid

We examined the relations of genotypes for SNPs identified in the various EWASs to eGFR or serum concentrations of creatinine or uric acid by one-way analysis of variance (ANOVA). The 25 SNPs identified in the EWAS for eGFR were all significantly [P < 0.0018 (0.05/28)] associated with eGFR (Table 8). Among the four SNPs identified in the analysis of CKD, only rs76974938 was significantly related to eGFR. The seven SNPs identified in the EWAS for serum creatinine concentration were all significantly [P < 0.0045 (0.05/11)] related to the serum concentration of creatinine (Table 9). Among the four SNPs identified in the analysis of CKD, only rs76974938 was significantly related to the serum concentration of creatinine. The six SNPs identified in the EWAS for serum uric acid concentration were all significantly [P < 0.0056 (0.05/9)] related to the serum concentration of uric acid, but the three SNPs identified in the analysis of hyperuricemia were not (Table 10).

Relation of SNPs identified in the present study to phenotypes previously examined in GWASs

We examined the relation of genes, chromosomal loci, and SNPs identified in the present study to phenotypes previously probed in GWASs with data available in public databases [GWAS Catalog (http://www.ebi.ac.uk/gwas) and GWAS Central (http://www.gwascentral.org/browser)]. Chromosomal region 11p11.2 and SASHI were previously shown to be susceptibility loci for CKD [15] or diabetic nephropathy [34], respectively. In addition, MARCH1 and RFTN1 were previously shown to be related to urinary uromodulin [35] or serum uric acid [36] levels, respectively. The remaining 31 SNPs identified in the present study as being related to eGFR, the serum creatinine concentration, or CKD were not previously identified as genetic determinants of renal function-related traits or CKD (Supplementary Table 7). SLC22A12 [27, 37], CDC42BPG [29], and SLC2A9 [32, 37–39] were previously implicated as determinants of the serum concentration of uric acid or gout, whereas the remaining four SNPs identified in the present study as being related to the serum concentration of uric acid or hyperuricemia were not previously found to influence serum uric acid levels, hyperuricemia, or gout (Supplementary Table 8).

DISCUSSION

We have now shown that rs76974938 [C/T (D67N)] of C21orf59 was associated with eGFR and CKD. Twenty-four additional SNPs and seven SNPs were significantly related to eGFR or the serum concentration of creatinine, respectively. Among these genes and loci, 11p11.2 and SASH1 were previously identified as susceptibility loci for CKD [15] or diabetic nephropathy [34], respectively. MARCH1 and RFTN1 were also previously shown to be related to urinary uromodulin [35] or serum uric acid [36] levels, respectively. The remaining 21 and six SNPs identified in our study are thus potential novel loci related to eGFR or to the serum concentration of creatinine, respectively, in Japanese. We also identified three additional SNPs related to CKD, all of which are candidates for novel susceptibility loci for this condition.

We also found that rs121907892 [G/A (W258*)] and rs505802 (G/A) of SLC22A12, rs188780113 [G/A (R478C)] of ATG2A, rs55975541 [G/A (R1237W)] of CDC42BPG, and rs3775948 (G/C) and rs3733591 [T/C (H265R)] of SLC2A9 were significantly associated with the serum concentration of uric acid. In addition, rs115445569 [C/T (R64Q)] of ACOT11, rs116911833 [G/A (T80M)] of TRIM7, and rs60854092 [T/A (F1689I) of NOTCH2 were related to hyperuricemia, although the genotypes of these SNPs were not related to the serum concentration of uric acid. Among these genes, SLC22A12 [27, 37], CDC42BPG [29], and SLC2A9 [32, 37–39] were previously found to be related to serum uric acid levels or gout. The remaining four genes (ATG2A, ACOT11, TRIM7, and NOTCH2) may be novel loci that influence the serum concentration of uric acid or confer susceptibility to hyperuricemia.

SNPs associated with renal function

The chromosome 21 open reading frame 59 gene (C21orf59) is located at chromosomal region 21q22.11 (NCBI Gene, https://www.ncbi.nlm.nih.gov/gene) and is widely expressed including in the kidney (The Human Protein Atlas, http://www.proteinatlas.org). Studies in zebrafish and Xenopus have revealed that the protein encoded by C21orf59 activates the motility and polarization of cilia and thereby contributes to cilia-mediated processes such as the generation of fluid flow [40]. The primary cilium is a microtubule-based organelle. Impairment of ciliary function can result in polycystic kidney disease [41]. C21orf59 contributes to the assembly of dynein arms in motile cilia, with mutations in C21orf59 having been found to cause ciliary dyskinesia [42]. We have now shown that rs76974938 [C/T (D67N)] of C21orf59 was significantly associated with eGFR and CKD, with the minor T allele representing a risk factor for CKD. Given that C21orf59 is implicated in cilium function and that cilia serve as mechanosensors to detect fluid flow in the lumen of renal tubules, the association of C21orf59 with eGFR and CKD might reflect an effect of this gene on renal tubular function.

In a previous GWAS for individuals of European ancestry [9], the MAFs of CKD-associated SNPs were 13% to 27%, their effect sizes (percentage change in serum creatinine levels) were –1.0% to 1.1%, and the allele odds ratios (ORs) for CKD were 0.84 to 1.07. Another GWAS performed with subjects of European ancestry [10] identified eGFR-associated SNPs whose MAFs were 5% to 50% and ORs for CKD were 0.80 to 1.19. In a meta-analysis of GWASs with a total of 137,629 individuals of European ancestry, the MAFs of eGFR-associated SNPs were 10% to 47% [15], and in a similar analysis with a total of 71,149 East Asian individuals the MAFs of kidney function-associated SNPs were 11% to 43% [14].

In the present study, rs76974938 of C21orf59 had a MAF of 2.4% and an allele OR of 4.52 for CKD, and it showed a difference in eGFR of 6.0 mL min^–1 1.73 m^–2 or in the serum concentration of creatinine of 13.3 μmol/L among genotypes. This SNP was thus a low-frequency variant with a moderate effect size. For the remaining 24 SNPs [MAF, difference in eGFR (mL min^–1 1.73 m^–2) among genotypes] identified in the EWAS of eGFR, rs77804526 (0.1%, 27.8), rs180950245 (0.1%, 20.3), rs12822449 (0.2%, 19.9), and rs79637542 (0.2%, 14.0) were rare variants with a large effect size; rs200982668 (1.3%, 8.8), rs139537100 (1.2%, 8.8), rs188378669 (1.2%, 8.9), rs200787930 (1.2%, 8.8), rs61734696 (1.2%, 8.8), rs199921354 (1.2%, 8.8), rs115287176 (1.2%, 8.8), rs138406927 (2.1%, 6.9), rs146879198 (1.2%, 8.5), rs146092501 (1.2%, 8.4), rs192210727 (1.3%, 7.6), rs147284320 (2.0%, 5.2), rs78010183 (1.8%, 5.3), rs201459911 (0.7%, 7.3), and rs199576535 (1.0%, 6.1) were low-frequency variants with a moderate effect size; and rs115910467 (8.2%, 6.9), rs17115182 (7.0%, 6.5), rs1873059 (47.4%, 2.6), rs77473776 (30.6%, 3.9), and rs2886644 (11.0%, 5.0) were common variants with a low to moderate effect size (Supplementary Figure 3A). In the case of the seven SNPs [MAF, difference in serum creatinine concentration (μmol/L) among genotypes] identified in the EWAS for serum creatinine concentration, rs139421991 (0.3%, 53.2), rs35602605 (0.1%, 122.1), rs188929035 (0.4%, 20.6), rs199980930 (0.1%, 73.0), and rs35620248 (0.3%, 47.9) were rare variants with a moderate to large effect size; rs148658404 (0.7%, 32.1) was a low-frequency variant with a moderate effect size; and rs3922872 (5.5%, 10.5) was a common variant with a small effect size (Supplementary Figure 3B).

SNPs associated with the serum concentration of uric acid

The autophagy related 2A gene (ATG2A) is located at chromosomal region 11q13.1 (NCBI Gene) and is expressed in various organs including the kidney and gastrointestinal tract (The Human Protein Atlas). Autophagy is an intracellular process in which cytoplasmic material is enveloped by authophagosomes and delivered to lysosomes for degradation [43]. Mammalian ATG2A is localized to the autophagosome membrane as well as to lipid droplets. Knockdown of both ATG2A and ATG2B in adipocytes results in an increase in the size and number of lipid droplets and in their clustering. The mammalian proteins thus promote autophagosome formation [44]. Activation of neutrophils in gout promotes the formation of proinflammatory neutrophil extracellular traps, a process linked to autophagy-related signaling and interleukin-1β activity [45]. We have now shown that rs188780113 [G/A (R478C)] of ATG2A was significantly associated with the serum concentration of uric acid, with the minor A allele being related to decreased uric acid levels. Although the molecular mechanism underlying this association remains unclear, it might be attributable to the role of ATG2A in autophagy-related metabolic processes.

The acyl-CoA thioesterase 11 gene (ACOT11) is located at chromosomal region 1p32.3 (NCBI Gene) and is expressed in various organs including the kidney and gastrointestinal tract (The Human Protein Atlas). Mammalian acyl-CoA thioesterases catalyze the conversion of activated fatty acids to the corresponding nonesterified fatty acid and coenzyme A [46]. ACOT11 mRNA levels were increased in mouse brown adipose tissue by cold exposure and decreased by warm temperatures as well as were higher in this tissue of obesity-resistant mice than in obesity-prone mice [47]. ACOT11 functions to reduce energy consumption and conserve calories, also suggestive of a role in obesity [48]. In nutritional excess, the overproduction of free fatty acids by ACOT11 induces insulin resistance, leading to increases in inflammation and endoplasmic reticulum stress [48]. We have now shown that rs115445569 [C/T (R64Q)] of ACOT11 was related to hyperuricemia, with the minor T allele representing a risk factor this condition. The association of ACOT11 with hyperuricemia might reflect the fact that ACOT11 is implicated in adipose tissue metabolism and that obesity is a risk factor for hyperuricemia.

The tripartite motif containing 7 gene (TRIM7) is located at chromosomal region 5q35.3 (NCBI Gene) and is widely expressed including in the kidney and gastrointestinal tract (The Human Protein Atlas). A family of tripartite motif (TRIM) proteins comprises >70 members and contributes to various biological processes [49–54]. TRIM7 is a glycogenin-interacting protein that contributes to cellular glucose metabolism [55]. We have now shown that rs116911833 [G/A (T80M)] of TRIM7 was significantly associated with hyperuricemia, with the minor A allele representing a risk factor for this condition. Given the potential role of TRIM7 in diverse cellular processes, the molecular mechanism underlying this association remains unclear.

The NOTCH2 gene is located at chromosomal region 1p12 (NCBI Gene) and is ubiquitously expressed including in the kidney and gastrointestinal tract (The Human Protein Atlas). Notch2 functions as a receptor for membrane-bound ligands and is implicated in the development of vessels, kidney, and liver [56]. It also contributes to skeletal homeostasis, osteoblastogenesis, and osteoclastogenesis [57]. Mutations in NOTCH2 that result in the production of a stable truncated protein are responsible for Hajdu-Cheney syndrome, which is characterized by defects in craniofacial development, osteoporosis with fractures, acro-osteolysis, neurological complications, cardiovascular defects, and polycystic kidneys [58]. Mutations of NOTCH2 also cause Alagille syndrome [59]. NOTCH2 was shown to be a susceptibility locus for type 2 diabetes mellitus [60]. We have now shown that rs60854092 [T/A (F1689I)] of NOTCH2 was significantly associated with hyperuricemia, with the minor A allele being protective against this condition. Given that insulin resistance may be a common underlying mechanism contributing to type 2 diabetes mellitus and hyperuricemia [61], the association of NOTCH2 with hyperuricemia might reflect the effect of this gene on type 2 diabetes.

In a previous large-scale GWAS for serum uric acid concentration [30], the MAFs of related SNPs were 11% to 49% and their effect sizes ranged from –10.9 to 22.2 μmol/L (–0.184 to 0.373 mg/dL). In our study, among the six SNPs significantly related to the serum concentration of uric acid, the MAF (difference in serum uric acid level among genotypes) of rs121907892 or rs188780113 was 2.4% (283 μmol/L) and 3.3% (134 μmol/L), respectively. These SNPs were thus low-frequency variants with a large effect size. The MAF (difference in serum uric acid level among genotypes) of the remaining four SNPS (rs505802, rs55975541, rs3775948, and rs3733591) was 17.5% (43 μmol/L), 16.5% (26 μmol/L), 42.4% (21 μmol/L), and 28.7% (15 μmol/L), respectively. These SNPs were thus common variants with a small to moderate effect size (Supplementary Figure 4). The MAFs (allele ORs) of the three SNPs related to hyperuricemia (rs115445569, rs116911833, and rs60854092) were 1.1% (1.51), 2.0% (1.21), and 4.6% (0.85), respectively. These SNPs were thus low-frequency variants with a small to moderate effect size.

Study limitations

There are several limitations in the present study. (i) Although an eGFR is affected by body weight, especially skeletal muscle mass, the formula for eGFR used in the present study did not include body weight. (ii) We did not obtain information on microhematuria or microalbuminuria in controls. Information by detailed clinical examination including renal biopsy was not obtained in most subjects, given that such diagnostic procedures were not considered feasible for an epidemiological study whose subjects were recruited from the general population. (iii) Some patients were diagnosed for CKD with an eGFR <60 mL min^–1 1.73 m^–2 at one point of examination. Dehydration status in such subjects was not examined. Data for the rate of GFR fall in a long-term follow up were not available for the present study subjects. (iv) Given that our results were not replicated, they will require validation in other subject panels or ethnic groups. (v) It is possible that SNPs identified in the present study are in linkage disequilibrium with other polymorphisms in nearby genes that are actually responsible for the observed associations. (vi) Three SNPs associated with CKD were not related to eGFR or to the serum concentration of creatinine. In addition, the three SNPs associated with hyperuricemia were not related to the serum concentration of uric acid. It is possible that these discrepancies are attributable to the effects of medical treatment. (vii) The molecular mechanisms of the observed associations remain to be determined.

CONCLUSIONS

In conclusion, we have identified rs76974938 [C/T (D67N)] of C21orf59 as a novel genetic determinant of renal function and CKD. We identified an additional 21 and six SNPs as novel determinants of eGFR or the serum concentration of creatinine, respectively, as well as an additional three SNPs as novel candidate susceptibility loci for CKD. We also identified rs188780113 [G/A (R478C)] of ATG2A as a novel determinant of the serum concentration of uric acid as well as rs115445569 of ACOT11, rs116911833 of TRIM7, and rs60854092 of NOTCH2 as potential novel susceptibility loci for hyperuricemia. Analysis of these SNPs may be informative for assessment of the genetic risk for CKD or hyperuricemia in Japanese.

MATERIALS AND METHODS

Study subjects

For the EWASs of eGFR or the serum concentration of creatinine, a total of 12,565 Japanese was examined. A total of 9934 subjects who did not take uric acid-lowering medications was examined in the EWAS for the serum concentration of uric acid. In the EWASs for CKD or hyperuricemia, 5161 individuals (3270 subjects with CKD, 1891 controls) or 11,686 individuals (2045 subjects with hyperuricemia, 9641 controls) were examined, respectively. Study subjects were recruited as previously described [62].

Glomerular filtration rate was estimated with the use of the simplified prediction equation derived from the modified version of that described in the Modification of Diet in Renal Disease (MDRD) Study, as proposed by the Japanese Society of Nephrology [63]: eGFR (mL min^–1 1.73 m^–2) = 194 × [age (years)]^–0.287 × [serum creatinine (mg/dL)]^–1.094 × [0.739 if female]. The National Kidney Foundation–Kidney Disease Outcomes Quality Initiative guidelines recommend a diagnosis of CKD if eGFR is <60 mL min^–1 1.73 m^–2 [4]. Nonlinear relations between GFR and the risk of adverse events, such as death, cardiovascular events, and hospitalization, have been demonstrated, with an increased risk being associated with an eGFR of <60 mL min^–1 1.73 m^–2 and the risk markedly rising further when values fall below 45 mL min^–1 1.73 m^–2 [64]. We thus adopted the criterion of an eGFR of <60 mL min^–1 1.73 m^–2 (actual range, 2.5 to 59.9 mL min^–1 1.73 m^–2) for the diagnosis of CKD in the present study. The control individuals for the EWAS of CKD had an eGFR of ≥90 mL min^–1 1.73 m^–2 (actual range, 90 to 584.9 mL min^–1 1.73 m^–2) and did not appear to have functional or structural abnormalities of the kidneys or a history of renal disease.

Hyperuricemia was defined as a serum uric acid concentration of >416 μmol/L (actual range, 422 to 1172 μmol/L) or the taking of uric acid-lowering medication. Individuals taking drugs that may cause secondary hyperuricemia were excluded. The control individuals for the EWAS of hyperuricemia had a serum uric acid concentration of ≤416 μmol/L (actual range, 36 to 416 μmol/L) and had no history of hyperuricemia or gout or of taking uric acid-lowering medication. Autopsy cases were excluded from controls for the EWASs of both CKD and hyperuricemia.

The study protocol complied with the Declaration of Helsinki and was approved by the Committees on the Ethics of Human Research of Mie University Graduate School of Medicine, Hirosaki University Graduate School of Medicine, Tokyo Metropolitan Institute of Gerontology, and participating hospitals. Written informed consent was obtained from each participant or families of the deceased subjects.

EWASs

Methods for collection and extraction of genomic DNA samples were described previously [62]. All EWASs (Supplementary Figure 5) were performed with the use of the HumanExome-12 v1.1 or v1.2 DNA Analysis BeadChip or Infinium Exome-24 v1.0 BeadChip (Illumina, San Diego, CA, USA). Detailed information of these exome arrays and methods of quality control were described previously [62]. Totals of 41,352 SNPs (eGFR, serum creatinine concentration, CKD) or 41,372 SNPs (serum uric acid concentration, hyperuricemia) that passed quality control were analyzed.

Statistical analysis

The relation of genotypes of SNPs to eGFR or the serum concentrations of creatinine or uric acid in the EWAS was evaluated by linear regression analysis. In comparisons of characteristics between subjects with CKD or hyperuricemia and corresponding controls, quantitative data were compared between two groups by the Mann–Whitney U-test, given that variables showed skewed distribution (P <0.01 by the Kolmogorov-Smirnov and Lilliefors test). Categorical data were compared between two groups by the Pearson’s chi-square test. Allele frequencies were estimated by the gene counting method, and Fisher’s exact test was applied to identify departure from Hardy-Weinberg equilibrium. The relation of allele frequencies of SNPs to CKD or hyperuricemia in the EWAS was examined with Fisher’s exact test. To compensate for multiple comparisons of genotypes with eGFR or the serum concentrations of creatinine or uric acid, or of allele frequencies with CKD or hyperuricemia, we applied Bonferroni’s correction for statistical significance of association. Given that 41,352 or 41,372 SNPs were analyzed, the significance level was set at P < 1.21 × 10^–6 (0.05/41,352 or 0.05/41,372) for the EWASs. Quantile-quantile plots either for P values of genotypes in the EWASs for eGFR or the serum concentrations of creatinine or uric acid or for those of allele frequencies in the EWASs for CKD or hyperuricemia are shown in Supplementary Figures 6 and 7, respectively. The inflation factor (λ) was 1.05 for eGFR, 1.06 for the serum concentration of creatinine, 1.05 for the serum concentration of uric acid, 1.17 for CKD, and 1.32 for hyperuricemia. Multivariable logistic regression analysis was performed with CKD as a dependent variable and independent variables including age, sex (0, woman; 1, man), the prevalence of hypertension and diabetes mellitus (0, no history of these conditions; 1, positive history), and genotype of each SNP. Similar analysis was performed with hyperuricemia as a dependent variable and independent variables including age, sex, and genotype of each SNP. A detailed method of analysis was described previously [62]. The relation of genotypes of isolated SNPs to eGFR or the serum concentrations of creatinine or uric acid was examined by one-way ANOVA. Bonferroni’s correction was also applied to other statistical analysis as indicated. Statistical tests were performed with JMP Genomics version 6.0 software (SAS Institute, Cary, NC USA).

Authors’ contributions

Y. Yamada contributed to conception and design of the study; to acquisition, analysis, and interpretation of the data; and to drafting of the manuscript. J. Sakuma, I. Takeuchi, and Y. Yasukochi contributed to analysis and interpretation of the data as well as to revision of the manuscript. K. Kato, M. Oguri, T. Fujimaki, H. Horibe, M. Muramatsu, M. Sawabe, Y. Fujiwara, Y. Taniguchi, S. Obuchi, H. Kawai, S. Shinkai, S. Mori, and T. Arai each contributed to acquisition of the data and to revision of the manuscript. M. Tanaka contributed to acquisition, analysis, and interpretation of the data as well as to revision of the manuscript. All authors approved submission of the final version of the article for publication.

CONFLICTS OF INTEREST

The authors declare no conflicts of interest.

FUNDING

This work was supported by CREST (H25-H30), Japan Science and Technology Agency (to Y. Yamada, J. Sakuma, and I. Takeuchi) and by Japan Society for the Promotion of Science KAKENHI grants JP15H04772 (to Y. Yamada), JP25242062 (to M. Tanaka), and JP16H01872 (to M. Tanaka).

REFERENCES

1. Webster AC, Nagler EV, Morton RL, Masson P. Chronic kidney disease. Lancet. 2016 (published online).

2. Weiner DE, Tighiouart H, Amin MG, Stark PC, MacLeod B, Griffith JL, Salem DN, Levey AS, Sarnak MJ. Chronic kidney disease as a risk factor for cardiovascular disease and all-cause mortality: a pooled analysis of community-based studies. J Am Soc Nephrol. 2004; 15: 1307–1315.

3. Chronic Kidney Disease Prognosis Consortium, Matsushita K, van der Velde M, Astor BC, Woodward M, Levey AS, de Jong PE, Coresh J, Gansevoort RT. Association of estimated glomerular filtration rate and albuminuria with all-cause and cardiovascular mortality in general population cohorts: a collaborative meta-analysis. Lancet. 2010; 375: 2073–2081.

4. National Kidney Foundation. K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Am J Kidney Dis. 2002; 39: S1–S266.

5. James MT, Hemmelgarn BR, Tonelli M. Early recognition and prevention of chronic kidney disease. Lancet. 2010; 375: 1296–1309.

6. Yamagata K, Ishida K, Sairenchi T, Takahashi H, Ohba S, Shiigai T, Narita M, Koyama A. Risk factors for chronic kidney disease in a community-based population: a 10-year follow-up study. Kidney Int. 2007; 71: 159–166.

7. Köttgen A, Glazer NL, Dehghan A, Hwang SJ, Katz R, Li M, Yang Q, Gudnason V, Launer LJ, Harris TB, Smith AV, Arking DE, Astor BC, et al. Multiple loci associated with indices of renal function and chronic kidney disease. Nat Genet. 2009; 41: 712–717.

8. Chambers JC, Zhang W, Lord GM, van der Harst P, Lawlor DA, Sehmi JS, Gale DP, Wass MN, Ahmadi KR, Bakker SJ, Beckmann J, Bilo HJ, Bochud M, et al. Genetic loci influencing kidney function and chronic kidney disease. Nat Genet. 2010; 42: 373–375.

9. Köttgen A, Pattaro C, Böger CA, Fuchsberger C, Olden M, Glazer NL, Parsa A, Gao X, Yang Q, Smith AV, O’Connell JR, Li M, Schmidt H, et al. New loci associated with kidney function and chronic kidney disease. Nat Genet. 2010; 42: 376–384.

10. Böger CA, Gorski M, Li M, Hoffmann MM, Huang C, Yang Q, Teumer A, Krane V, O’Seaghdha CM, Kutalik Z, Wichmann HE, Haak T, Boes E, et al. Association of eGFR-related loci identified by GWAS with incident CKD and ESRD. PLoS Genet. 2011; 7: e1002292.

11. Pattaro C, Köttgen A, Teumer A, Garnaas M, Böger CA, Fuchsberger C, Olden M, Chen MH, Tin A, Taliun D, Li M, Gao X, Gorski M, et al. Genome-wide association and functional follow-up reveals new loci for kidney function. PLoS Genet. 2012; 8: e1002584.

12. Kao WH, Klag MJ, Meoni LA, Reich D, Berthier-Schaad Y, Li M, Coresh J, Patterson N, Tandon A, Powe NR, Fink NE, Sadler JH, Weir MR, et al. MYH9 is associated with nondiabetic end-stage renal disease in African Americans. Nat Genet. 2008; 40: 1185–1192.

13. Liu CT, Garnaas MK, Tin A, Kottgen A, Franceschini N, Peralta CA, de Boer IH, Lu X, Atkinson E, Ding J, Nalls M, Shriner D, Coresh J, et al. Genetic association for renal traits among participants of African ancestry reveals new loci for renal function. PLoS Genet. 2011; 7: e1002264.

14. Okada Y, Sim X, Go MJ, Wu JY, Gu D, Takeuchi F, Takahashi A, Maeda S, Tsunoda T, Chen P, Lim SC, Wong TY, Liu J, et al. Meta-analysis identifies multiple loci associated with kidney function-related traits in east Asian populations. Nat Genet. 2012; 44: 904–909.

15. Pattaro C, Teumer A, Gorski M, Chu AY, Li M, Mijatovic V, Garnaas M, Tin A, Sorice R, Li Y, Taliun D, Olden M, Foster M, et al. Genetic associations at 53 loci highlight cell types and biological pathways relevant for kidney function. Nat Commun. 2016; 7: 10023.

16. Choi HK, Mount DB, Reginato AM. Pathogenesis of gout. Ann Intern Med. 2005; 143: 499–516.

17. Perez-Ruiz F, Marimon E, Chinchilla SP. Hyperuricaemia with deposition: latest evidence and therapeutic approach. Ther Adv Musculoskelet Dis. 2015; 7: 225–233.

18. Feig DI, Kang DH, Johnson RJ. Uric acid and cardiovascular risk. N Engl J Med. 2008; 359: 1811–1821.

19. Fini MA, Elias A, Johnson RJ, Wright RM. Contribution of uric acid to cancer risk, recurrence, and mortality. Clin Trans Med. 2012; 1: 16.

20. Eraly SA, Vallon V, Rieg T, Gangoiti JA, Wikoff WR, Siuzdak G, Barshop BA, Nigam SK. Multiple organic anion transporters contribute to net renal excretion of uric acid. Physiol Genomics. 2008; 33: 180–192.

21. Reginato AM, Mount DB, Yang I, Choi HK. The genetics of hyperuricaemia and gout. Nat Rev Rheumatol. 2012; 8: 610–621.

22. Merriman TR. An update on the genetic architecture of hyperuricemia and gout. Arthritis Res Ther. 2015; 17: 98.

23. Wallace C, Newhouse SJ, Braund P, Zhang F, Tobin M, Falchi M, Ahmadi K, Dobson RJ, Marçano AC, Hajat C, Burton P, Deloukas P, Brown M, et al. Genome-wide association study identifies genes for biomarkers of cardiovascular disease: serum urate and dyslipidemia. Am J Hum Genet. 2008; 82: 139–149.

24. Dehghan A, Köttgen A, Yang Q, Hwang SJ, Kao WL, Rivadeneira F, Boerwinkle E, Levy D, Hofman A, Astor BC, Benjamin EJ, van Duijn CM, Witteman JC, et al. Association of three genetic loci with uric acid concentration and risk of gout: a genome-wide association study. Lancet. 2008; 372: 1953–1961.

25. Vitart V, Rudan I, Hayward C, Gray NK, Floyd J, Palmer CN, Knott SA, Kolcic I, Polasek O, Graessler J, Wilson JF, Marinaki A, Riches PL, et al. SLC2A9 is a newly identified urate transporter influencing serum urate concentration, urate excretion and gout. Nat Genet. 2008; 40: 437–442.

26. Döring A, Gieger C, Mehta D, Gohlke H, Prokisch H, Coassin S, Fischer G, Henke K, Klopp N, Kronenberg F, Paulweber B, Pfeufer A, Rosskopf D, et al. SLC2A9 influences uric acid concentrations with pronounced sex-specific effects. Nat Genet. 2008; 40: 430–436.

27. Kolz M, Johnson T, Sanna S, Teumer A, Vitart V, Perola M, Mangino M, Albrecht E, Wallace C, Farrall M, Johansson A, Nyholt DR, Aulchenko Y, et al. Meta-analysis of 28,141 individuals identifies common variants within five new loci that influence uric acid concentrations. PLoS Genet. 2009; 5: e1000504.

28. Yang Q, Köttgen A, Dehghan A, Smith AV, Glazer NL, Chen MH, Chasman DI, Aspelund T, Eiriksdottir G, Harris TB, Launer L, Nalls M, Hernandez D, et al. Multiple genetic loci influence serum urate levels and their relationship with gout and cardiovascular disease risk factors. Circ Cardiovasc Genet. 2010; 3: 523–530.

29. Tin A, Woodward OM, Kao WH, Liu CT, Lu X, Nalls MA, Shriner D, Semmo M, Akylbekova EL, Wyatt SB, Hwang SJ, Yang Q, Zonderman AB, et al. Genome-wide association study for serum urate concentrations and gout among African Americans identifies genomic risk loci and a novel URAT1 loss-of-function allele. Hum Mol Genet. 2011; 20: 4056–4068.

30. Köttgen A, Albrecht E, Teumer A, Vitart V, Krumsiek J, Hundertmark C, Pistis G, Ruggiero D, O’Seaghdha CM, Haller T, Yang Q, Tanaka T, Johnson AD, et al. Genome-wide association analyses identify 18 new loci associated with serum urate concentrations. Nat Genet. 2013; 45: 145–154.

31. Manolio TA, Collins FS, Cox NJ, Goldstein DB, Hindorff LA, Hunter DJ, McCarthy MI, Ramos EM, Cardon LR, Chakravarti A, Cho JH, Guttmacher AE, Kong A, et al. Finding the missing heritability of complex diseases. Nature. 2009; 461: 747–753.

32. Matsuo H, Yamamoto K, Nakaoka H, Nakayama A, Sakiyama M, Chiba T, Takahashi A, Nakamura T, Nakashima H, Takada Y, Danjoh I, Shimizu S, Abe J, et al. Genome-wide association study of clinically defined gout identifies multiple risk loci and its association with clinical subtypes. Ann Rheum Dis. 2016; 75: 652–659.

33. Nakayama A, Nakaoka H, Yamamoto K, Sakiyama M, Shaukat A, Toyoda Y, Okada Y, Kamatani Y, Nakamura T, Takada T, Inoue K, Yasujima T, Yuasa H, et al. GWAS of clinically defined gout and subtypes identifies multiple susceptibility loci that include urate transporter genes. Ann Rheum Dis. 2016 (published online).

34. McDonough CW, Palmer ND, Hicks PJ, Roh BH, An SS, Cooke JN, Hester JM, Wing MR, Bostrom MA, Rudock ME, Lewis JP, Talbert ME, Blevins RA, et al. A genome-wide association study for diabetic nephropathy genes in African Americans. Kidney Int. 2011; 79: 563–572.

35. Olden M, Corre T, Hayward C, Toniolo D, Ulivi S, Gasparini P, Pistis G, Hwang SJ, Bergmann S, Campbell H, Cocca M, Gandin I, Girotto G, et al. Common variants in UMOD associate with urinary uromodulin levels: a meta-analysis. J Am Soc Nephrol. 2014; 25: 1869–1882.

36. Huffman JE, Albrecht E, Teumer A, Mangino M, Kapur K, Johnson T, Kutalik Z, Pirastu N, Pistis G, Lopez LM, Haller T, Salo P, Goel A, et al. Modulation of genetic associations with serum urate levels by body-mass-index in humans. PLoS One. 2015; 10: e0119752.

37. Giri AK, Banerjee P, Chakraborty S, Kauser Y, Undru A, Roy S, Parekatt V, Ghosh S, Tandon N, Bharadwaj D. Genome wide association study of uric acid in Indian population and interaction of identified variants with Type 2 diabetes. Sci Rep. 2016; 6: 21440.

38. Charles BA, Shriner D, Doumatey A, Chen G, Zhou J, Huang H, Herbert A, Gerry NP, Christman MF, Adeyemo A, Rotimi CN. A genome-wide association study of serum uric acid in African Americans. BMC Med Genomics. 2011; 4: 17.

39. Li C, Li Z, Liu S, Wang C, Han L, Cui L, Zhou J, Zou H, Liu Z, Chen J, Cheng X, Zhou Z, Ding C, et al. Genome-wide association analysis identifies three new risk loci for gout arthritis in Han Chinese. Nat Commun. 2015; 6: 7041.

40. Jaffe KM, Grimes DT, Schottenfeld-Roames J, Werner ME, Ku TS, Kim SK, Pelliccia JL, Morante NF, Mitchell BJ, Burdine RD. c21orf59/kurly controls both cilia motility and polarization. Cell Rep. 2016; 14: 1841–1849.

41. Yoder BK. Role of primary cilia in the pathogenesis of polycystic kidney disease. J Am Soc Nephrol. 2007; 18: 1381–1388.

42. Austin-Tse C, Halbritter J, Zariwala MA, Gilberti RM, Gee HY, Hellman N, Pathak N, Liu Y, Panizzi JR, Patel-King RS, Tritschler D, Bower R, O’Toole E, et al. Zebrafish ciliopathy screen plus human mutational analysis identifies C21orf59 and CCDC65 defects as causing primary ciliary dyskinesia. Am J Hum Genet. 2013; 93: 672–686.

43. Behrends C, Sowa ME, Gygi SP, Harper JW. Network organization of the human autophagy system. Nature. 2010; 466: 68–76.

44. Velikkakath AK, Nishimura T, Oita E, Ishihara N, Mizushima N. Mammalian Atg2 proteins are essential for autophagosome formation and important for regulation of size and distribution of lipid droplets. Mol Biol Cell. 2012; 23: 896–909.

45. Mitroulis I, Kambas K, Chrysanthopoulou A, Skendros P, Apostolidou E, Kourtzelis I, Drosos GI, Boumpas DT, Ritis K. Neutrophil extracellular trap formation is associated with IL-1β and autophagy-related signaling in gout. PLoS One. 2011; 6: e29318.

46. Cohen DE. New players on the metabolic stage: How do you like Them Acots? Adipocyte. 2013; 2: 3–6.

47. Adams SH, Chui C, Schilbach SL, Yu XX, Goddard AD, Grimaldi JC, Lee J, Dowd P, Colman S, Lewin DA. BFIT, a unique acyl-CoA thioesterase induced in thermogenic brown adipose tissue: cloning, organization of the human gene and assessment of a potential link to obesity. Biochem J. 2001; 360: 135–142.

48. Zhang Y, Li Y, Niepel MW, Kawano Y, Han S, Liu S, Marsili A, Larsen PR, Lee CH, Cohen DE. Targeted deletion of thioesterase superfamily member 1 promotes energy expenditure and protects against obesity and insulin resistance. Proc Natl Acad Sci USA. 2012; 109: 5417–5422.

49. Chen L, Chen DT, Kurtyka C, Rawal B, Fulp WJ, Haura EB, Cress WD. Tripartite motif containing 28 (Trim28) can regulate cell proliferation by bridging HDAC1/E2F interactions. J Biol Chem. 2012; 287: 40106–40118.

50. Sato T, Okumura F, Ariga T, Hatakeyama S. TRIM6 interacts with Myc and maintains the pluripotency of mouse embryonic stem cells. J Cell Sci. 2012; 125: 1544–1555.

51. Tocchini C, Keusch JJ, Miller SB, Finger S, Gut H, Stadler MB, Ciosk R. The TRIM-NHL protein LIN-41 controls the onset of developmental plasticity in Caenorhabditis elegans. PLoS Genet. 2014; 10: e1004533.

52. Zaman MM, Nomura T, Takagi T, Okamura T, Jin W, Shinagawa T, Tanaka Y, Ishii S. Ubiquitination-deubiquitination by the TRIM27-USP7 complex regulates tumor necrosis factor alpha-induced apoptosis. Mol Cell Biol. 2013; 33: 4971–4984.

53. Eames HL, Saliba DG, Krausgruber T, Lanfrancotti A, Ryzhakov G, Udalova IA. KAP1/TRIM28: an inhibitor of IRF5 function in inflammatory macrophages. Immunobiology. 2012; 217: 1315–1324.

54. Versteeg GA, Benke S, Garcia-Sastre A, Rajsbaum R. In TRIMsic immunity: positive and negative regulation of immune signaling by tripartite motif proteins. Cytokine Growth Factor Rev. 2014; 25: 563–576.

55. Skurat AV, Dietrich AD, Zhai L, Roach PJ. GNIP, a novel protein that binds and activates glycogenin, the self-glucosylating initiator of glycogen biosynthesis. J Biol Chem. 2002; 277: 19331–19338.

56. Isidor B, Lindenbaum P, Pichon O, Bézieau S, Dina C, Jacquemont S, Martin-Coignard D, Thauvin-Robinet C, Le Merrer M, Mandel JL, David A, Faivre L, Cormier-Daire V, et al. Truncating mutations in the last exon of NOTCH2 cause a rare skeletal disorder with osteoporosis. Nat Genet. 2011; 43: 306–308.

57. Canalis E, Zanotti S. Hajdu-Cheney syndrome, a disease associated with NOTCH2 mutations. Curr Osteoporos Rep. 2016; 14: 126–131.

58. Simpson MA, Irving MD, Asilmaz E, Gray MJ, Dafou D, Elmslie FV, Mansour S, Holder SE, Brain CE, Burton BK, Kim KH, Pauli RM, Aftimos S, et al. Mutations in NOTCH2 cause Hajdu-Cheney syndrome, a disorder of severe and progressive bone loss. Nat Genet. 2011; 43: 303–305.

59. McDaniell R, Warthen DM, Sanchez-Lara PA, Pai A, Krantz ID, Piccoli DA, Spinner NB. NOTCH2 mutations cause Alagille syndrome, a heterogeneous disorder of the notch signaling pathway. Am J Hum Genet. 2006; 79: 169–173.

60. Zeggini E, Scott LJ, Saxena R, Voight BF, Marchini JL, Hu T, de Bakker PI, Abecasis GR, Almgren P, Andersen G, Ardlie K, Boström KB, Bergman RN, et al. Meta-analysis of genome-wide association data and large-scale replication identifies additional susceptibility loci for type 2 diabetes. Nat Genet. 2008; 40: 638–645.

61. Li C, Hsieh MC, Chang SJ. Metabolic syndrome, diabetes, and hyperuricemia. Curr Opin Rheumatol. 2013; 25: 210–216.

62. Yamada Y, Sakuma J, Takeuchi I, Yasukochi Y, Kato K, Oguri M, Fujimaki T, Horibe H, Muramatsu M, Sawabe M, Fujiwara Y, Taniguchi Y, Obuchi S, et al. Identification of STXBP2 as a novel susceptibility locus for myocardial infarction in Japanese individuals by an exome-wide association study. Oncotarget. 2017; 8:33527-33535. doi: 10.18632/oncotarget.16536.

63. Matsuo S, Imai E, Horio M, Yasuda Y, Tomita K, Nitta K, Yamagata K, Tomino Y, Yokoyama H, Hishida A; Collaborators developing the Japanese equation for estimated GFR. Revised equations for estimated GFR from serum creatinine in Japan. Am J Kidney Dis. 2009; 53: 982–992.

64. Go AS, Chertow GM, Fan D, McCulloch CE, Hsu CY. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med. 2004; 351: 1296–1305.