Identification of rs671, a common variant of ALDH2, as a gout susceptibility locus

Gout is a common disease resulting from hyperuricemia. Recently, a genome-wide association study identified an association between gout and a single nucleotide polymorphism (SNP) rs2188380, located on an intergenic region between MYL2 and CUX2 on chromosome 12. However, other genes around rs2188380 could possibly be gout susceptibility genes. Therefore, we performed a fine-mapping study of the MYL2-CUX2 region. From 8,595 SNPs in the MYL2-CUX2 region, 9 tag SNPs were selected, and genotyping of 1,048 male gout patients and 1,334 male controls was performed by TaqMan method. Eight SNPs showed significant associations with gout after Bonferroni correction. rs671 (Glu504Lys) of ALDH2 had the most significant association with gout (P = 1.7 × 10−18, odds ratio = 0.53). After adjustment for rs671, the other 8 SNPs no longer showed a significant association with gout, while the significant association of rs671 remained. rs671 has been reportedly associated with alcohol drinking behavior, and it is well-known that alcohol drinking elevates serum uric acid levels. These data suggest that rs671, a common functional SNP of ALDH2, is a genuine gout-associated SNP in the MYL2-CUX2 locus and that “A” allele (Lys) of rs671 plays a protective role in the development of gout.

Gout is a common disease resulting from hyperuricemia, and causes acute arthritis. Previous genetic and functional analyses revealed that ABCG2 dysfunctional variants caused gout 1-3 due to decreased urate excretion in gut 4 and kidney 5 . Genome-wide association studies (GWASs) of gout also showed genome-wide significant associations with ABCG2 and GLUT9 [6][7][8] . Recently, we revealed for the first time that the following 3 loci were associated with gout at the genome-wide significance level: rs1260326 of GCKR, rs4073582 of CNIH-2 and rs2188380 of MYL2-CUX2 8 . Among them, 2 SNPs are located in gene regions: rs1260326 is a nonsynonymous single nucleotide polymorphism (SNP) (Leu446Pro) of GCKR, and rs4073582 is an intronic SNP of CNIH-2. On the other hand, rs2188380 is located on an intergenic region between MYL2 and CUX2 8 . Additionally, we detected many SNPs showing significant associations with gout across the chromosome 12q24 region which were in strong linkage disequilibrium (LD) with rs2188380. MYL2 encodes a regulatory light chain associated with cardiac myosin β (or slow) heavy chain, and an association between MYL2 variant and high-density lipoprotein cholesterol metabolism was previously reported 9 . CUX2 regulates cell-cycle progression 10 and plays important roles in neural progenitor development in the central nervous system 10,11 . An association between CUX2 and type 1 diabetes has also been reported 12 . However, there is a possibility that the other genes around rs2188380 of MYL2-CUX2 can be gout susceptibility genes. Therefore, we performed fine-mapping of the MYL2-CUX2 region and a further association analysis of gout. From 8,595 SNPs in the MYL2-CUX2 region within 10 Mb across rs2188380, 45 SNPs in LD (r 2 ≥ 0.3) with rs2188380 were selected (Supplementary Figure S1). Among these 45 SNPs and rs2188380, 9 tag SNPs were selected for association analysis ( Fig. 1 and Supplementary Table S1). Genotyping results of the 9 tag SNPs for 1,048 gout patients and 1,334 controls were shown in Table 1. The call rates for the 9 SNPs were more than 95.0%. All the SNPs in the control group were in Hardy-Weinberg equilibrium (P > 0.05). Except for rs2555004, the other 8 SNPs showed significant associations at P < 5.6 × 10 −3 (= 0.05/9) with the Bonferroni correction, and rs671 (Glu504Lys) of aldehyde dehydrogenase 2 (ALDH2) had the most significant association with gout (P = 1.7 × 10 −18 ; odds ratio [OR] = 0.53; 95% confidence interval [CI]: 0.46-0.61, Table 1 and Supplementary Figure S2A).

Results
Next, the multivariate logistic regression analyses were performed to evaluate whether there was an additional association signal after the adjustment for the most significantly associated SNP rs671. We set the significance threshold as α = 6.3 × 10 −3 (= 0.05/8) with the Bonferroni correction. While the significant association of rs671 remained, the other 8 SNPs no longer showed a significant association with gout after the adjustment for rs671 (Table 2 and Supplementary Figure S2B). Among rs671 and 6 tagged SNPs (rs3782886, rs11066015, rs4646776, rs11066132, rs2074356 and rs11066280) shown in Supplementary Table S2, rs671 is the most promising functional SNP because only rs671 is a nonsynonymous variant (Glu504Lys). rs4766566 had a nominally significant association (P = 0.032), but did not pass the significant threshold for multiple hypothesis testing (Table 2). Additionally, the OR for rs4766566 became closer to 1.0 after the adjustment for rs671 (from 0.59 to 0.82; Tables 1 and 2). These data suggest that rs671 (Glu504Lys) of ALDH2 is a genuine gout-associated SNP in the MYL2-CUX2 locus. The 9 tag SNPs (rs7978484, rs16940688, rs2071629, rs2188380, rs11065783, rs3809297, rs4766566, rs671 and rs2555004), which are shown in bold and boxes, were selected for association analysis.
Scientific RepoRts | 6:25360 | DOI: 10.1038/srep25360 It is well-known that individuals with rs671 heterozygotes (A/G; Lys/Glu) have only 6.25% of enzyme activity of those with normal ALDH2 (G/G; Glu/Glu), and those with homozygotes (A/A; Lys/Lys) show almost no activity 13 . Therefore, it is expected that dominant model (G/G v.s. A/G or A/A) is the most likely genetic model for the association between rs671 and gout. Actually, the statistical significance of the association between rs671 and gout was improved by applying the dominant model (P = 2.9 × 10 −21 ; OR = 0.44; 95% CI: 0.37-0.52) compared with the result from the allelic model (P = 1.7 × 10 −18 ; OR = 0.53; 95% CI: 0.46-0.61; Table 1). These findings indicate that "A" allele (Lys) of rs671 plays a protective role in the development of gout.

Discussion
In this study, among 9 tag SNPs selected from 8,595 SNPs in the MYL2-CUX2 locus, only the association of rs671 of ALDH2 remained significant after the adjustment for each SNP with the Bonferroni correction ( Table 2). Together with the fact that rs671 (Glu504Lys) is a well-known dysfunctional SNP, we therefore concluded that rs671 is a genuine gout-associated SNP. Indeed, a previous Japanese study with 180 gout cases and 49 controls has indicated that the frequency of homozygotes (A/A; Lys/Lys) of rs671 was lower in gout patients than in controls 14 . The fine-mapping study for the associational signal on chromosome 12 identified by our GWAS reached the consistent result with the previous finding showing the association between ALDH2 and gout 14 .
ALDH2 is a crucial enzyme in the alcohol metabolism. Alcohol is oxidized to acetaldehyde by alcohol dehydrogenase, and acetaldehyde is further metabolized to acetate by aldehyde dehydrogenase 15 , which largely depends on ALDH2. rs671 (Glu504Lys), a common missense SNP of ALDH2 gene, severely decreases the activity of ALDH2 enzyme 13 . When acetate is metabolized to acetyl-coenzyme A, adenosine triphosphate (ATP) hydrolyzes to adenosine monophosphate (AMP) which is ultimately metabolized to uric acid. This alcohol metabolism is one of the reasons why alcohol drinking elevates serum uric acid (SUA) levels. Thus, the association between rs671 and gout is partly due to alcohol drinking.
The allele frequencies of rs671 of ALDH2 differ among populations: the 504Lys allele ("A" allele) is common in East Asians including Japanese, but quite rare in other populations such as European and African descendants 16 . Therefore, it is reasonable that this SNP has not been detected in the previous GWASs of gout in Europeans and African Americans due to its low frequency. We showed that rs671 of ALDH2 is an influential genetic factor for   Japanese as the other 4 previously reported loci 8 (ABCG2, SLC2A9, GCKR and CNIH-2) of gout, and further investigations in East Asian populations will be able to warrant these findings. In summary, Glu504Lys polymorphism (rs671), a common dysfunctional SNP of ALDH2, is identified as a genuine gout-associated polymorphism in the MYL2-CUX2 locus, and "A" allele (Lys) of rs671 plays a protective role in the development of gout.

Study participants. This study was approved by the institutional ethical committees (National Defense
Medical College and Nagoya University), and all procedures involved in this study were performed in accordance with the Declaration of Helsinki with written informed consent from each subject.
1,048 gout cases were assigned from the Japanese male outpatients at the gout clinics of Kyoto Industrial Health Association (Kyoto, Japan), or Ryougoku East Gate Clinic (Tokyo, Japan). All patients were clinically diagnosed as primary gout according to the criteria established by the American College of Rheumatology 17 . Patients with inherited metabolism disorders including Lesch-Nyhan syndrome were excluded. For the control group, 1,334 Japanese males with normal SUA (≤ 7.0 mg/dl) and without a history of gout were collected from the participants in the Shizuoka area in the Japan Multi-Institutional Collaborative Cohort Study (J-MICC Study) 18 Figure S1). After 8,550 SNPs in weak LD were excluded, the other 45 SNPs showing moderate to strong LD (r 2 ≥ 0.3) with rs2188380 remained. Next, we examined the LD between each pair of these 46 SNPs (Fig. 1), and searched for the SNPs that were tagging other SNPs with strong LD (r 2 ≥ 0.8).
Genetic analysis. Genomic DNA was extracted from whole peripheral blood cells 21 . Genotyping of the 8 SNPs was performed by the TaqMan method (Life Technologies Corporation, Carlsbad, CA USA) with a LightCycler 480 (Roche Diagnostics, Mannheim, Germany) 22 . To confirm their genotypes, DNA sequencing analyses were performed with the primers shown in Supplementary Table S4. Direct sequencing was performed with a 3130xl Genetic Analyzer (Life Technologies Corporation) 22 . The deviation from Hardy-Weinberg equilibrium in control samples was evaluated by chi-square test using the software R (version 3.1.1) (http://www.r-project. org/).

Statistical analyses.
The associations between SNPs and gout were examined with logistic regression analyses. For the robustness of the statistical test, random re-sampling methods with computer simulation are often applied 23,24 . In this study, the permutation test 24 was used for random re-sampling in a case-control study with replacement for 1,000,000 times, and the robustness of statistics was confirmed. All the logistic regression analyses and chi-square tests were performed with SPSS v.22.0J (IBM Japan Inc., Tokyo, Japan) and the software R (version 3.1.1) (http://www.r-project.org/). We examined the pairwise LD using PLINK v1.07 25 .