Common variant of BCAS3 is associated with gout risk in Japanese population: the first replication study after gout GWAS in Han Chinese

Background Gout is a common disease resulting from hyperuricemia which causes acute arthritis. A recent genome-wide association study (GWAS) of gout identified three new loci for gout in Han Chinese: regulatory factor X3 (RFX3), potassium voltage-gated channel subfamily Q member 1 (KCNQ1), and breast carcinoma amplified sequence 3 (BCAS3). The lack of any replication studies of these three loci using other population groups prompted us to perform a replication study with Japanese clinically defined gout cases and controls. Methods We genotyped the variants of RFX3 (rs12236871), KCNQ1 (rs179785) and BCAS3 (rs11653176) in 723 Japanese clinically defined gout cases and 913 controls by TaqMan method. rs179785 of KCNQ1 is also evaluated by direct sequencing because of difficulties of its genotyping by TaqMan method. Results Although the variants of RFX3 and BCAS3 were clearly genotyped by TaqMan method, rs179785 of KCNQ1 was not, because rs179785 (A/G) of KCNQ1 is located at the last nucleotide (“A”) of the 12-bp deletion variant (rs200562977) of KCNQ1. Therefore, rs179785 and rs200562977 of KCNQ1 were genotyped by direct sequencing in all samples. Moreover, by direct sequencing with the same primers, we were able to evaluate the genotypes of rs179784 of KCNQ1 which shows strong linkage disequilibrium with rs179785 (D’ = 1.0 and r2 = 0.99). rs11653176, a common variant of BCAS3, showed a significant association with gout (P = 1.66 × 10− 3; odds ratio [OR] = 0.80); the direction of effect was the same as that seen in the previous Han Chinese GWAS. Two variants of KCNQ1 (rs179785 and rs179784) had a nominally significant association (P = 0.043 and 0.044; OR = 0.85 and 0.86, respectively), but did not pass the significance threshold for multiple hypothesis testing using the Bonferroni correction. On the other hand, rs200562977 of KCNQ1 and rs12236871 of RFX3 did not show any significant association with gout. Conclusion BCAS3 is a coactivator of estrogen receptor alpha, and the influence of estrogen to serum uric acid level is well known. Our present replication study, as did the previous gout GWAS, demonstrated the common variant of BCAS3 to be associated with gout susceptibility.


Background
Gout, which can also cause acute arthritis, is a common disease resulting from hyperuricemia. An increasing number of patients nowadays suffer from gout. Although various investigations aiming to elucidate the pathogenesis of this common disease are being conducted worldwide, most of the common genetic causes of gout remain to be clarified. Previous function-based genetic studies [1][2][3] and genome-wide association studies (GWASs) [4][5][6] have revealed that gout is associated with several genes, such as ATP-binding cassette transporter, subfamily G, member 2 (ABCG2/BCRP) and glucose transporter 9 (GLUT9/SLC2A9). Especially, by performing a GWAS of clinically-ascertained gout, our Japanese report identified five gout loci including MYL2-CUX2 and cornichon family AMPA receptor auxiliary protein 2 (CNIH-2) [6]. Subsequent fine mapping analysis of the MYL2-CUX2 region found that rs671 of aldehyde dehydrogenase 2 (ALDH2) is a gout locus which is an Asian specific one [7]. Li et al. recently performed a GWAS of clinically-ascertained gout and identified the following three new loci for gout in Han Chinese: regulatory factor X3 (RFX3), potassium voltage-gated channel subfamily Q member 1 (KCNQ1) and breast carcinoma amplified sequence 3 (BCAS3) [8]. However, there is no replication study of these three loci using other population groups. We therefore performed a replication study using Japanese clinically-defined gout cases and controls.

Patients and controls
This study was approved by the institutions' Ethical Committees (National Defense Medical College and Nagoya University). All procedures were performed in accordance with the Declaration of Helsinki, with written informed consent obtained from each subject. The cases comprised 723 gout patients assigned from Japanese male outpatients at Ryougoku East Gate Clinic (Tokyo, Japan). All patients were clinically diagnosed with primary gout according to the criteria established by the American College of Rheumatology [9]. Patients with inherited metabolic disorders, including Lesch-Nyhan syndrome and phosphoribosylpyrophosphate synthetase I superactivity, were excluded. Hyperuricemia was defined as the serum uric acid (SUA) level that exceeds 7.0 mg/dl (= 416.36 mol/l) according to the guideline of the Japanese Society of Gout and Nucleic Acid Metabolism [10]. The control group comprised 913 Japanese males without hyperuricemia and gout history, recruited from the participants in the Daiko Study, part of the Japan Multi-Institutional Collaborative Cohort Study (J-MICC Study) [11]. The mean age (± SD) of case and control groups was 45.5 years (± 10.6) and 53.5 years (± 10. 3), respectively, and their mean body mass index was 25. 3 kg/m 2 (± 3.6) and 22.9 kg/m 2 (± 2.9), respectively.

Genotyping
Genomic DNA was extracted from whole peripheral blood cells [12]. Genotyping of the three single nucleotide polymorphisms (SNPs), rs12236871 of RFX3, rs179785 of KCNQ1 and rs11653176 of BCAS3, was Well-separated clusters are clearly visible, representing each genotype (T/T, T/C and C/C). c Representative plots of rs179785 of KCNQ1. Computer auto analysis divided these plots into three groups (G/G, A/G and A/A). However, the plots seemed to be clustered into four groups (labeled as Groups 1 to 4). We therefore employed direct sequencing to confirm the genotypes. As a result, a 12-bp deletion variant of KCNQ1 (rs200562977) was identified in several samples of all of four groups, and rs179785 of KCNQ1 is located at the last nucleotide of this deletion variant (also see Figs. 2 and 3). Therefore, because it is difficult to genotype rs179785 using the TaqMan method, we performed the subsequent genotyping of rs179785 and rs200562977 by direct sequencing performed using the TaqMan method (Thermo Fisher Scientific, Waltham, MA, USA) employing a LightCycler 480 (Roche Diagnostics, Mannheim, Germany) [12] with minor modifications. For genotyping KCNQ1variants, DNA sequencing analysis was performed with following primers: forward 5'-ACTTCCTGCCTCTGCTTTC-3′ and reverse 5'-TGAAGGAAGTGACCCCTG-3′. Direct sequencing was performed using a 3130xl Genetic Analyzer (Thermo Fisher Scientific) [12].

Data analysis
The software R version 3.1.1 (http://www.r-project.org/) [13] with the GenABEL package was used for all calculations in the statistical analysis. The association analyses were examined using the chi-square test. The pairwise linkage disequilibrium (LD) was calculated using data from the 1000 Genomes phase 3 JPT (Japanese in Tokyo) [14]. All P values were two-tailed and P values of < 0.05 were regarded as statistically significant.

Results
A representative plots of genotyping results by TaqMan method is shown in Fig. 1. Although allelic discrimination plots of rs12236871 of RFX3 (Fig. 1a) and rs11653176 of BCAS3 (Fig. 1b) are clearly divided into three groups for each genotype (major allele homozygote, heterozygote and minor allele homozygote), the plots of rs179785 of KCNQ1 are clustered into four groups, labeled as Groups 1, 2, 3 and 4 in Fig. 1c. Thus, to confirm the genotypes of samples of Groups 1 and 2, direct sequencing was performed to analyze the DNA sequence around rs179785 of KCNQ1. Although the genotypes of almost all the samples in Group 2 shown in Fig. 1c were heterozygous (A/G) for rs179785 (Fig. 2a), the heterozygous 12-bp deletion variant of KCNQ1, rs200562977 (Fig. 2b), was frequently found in Group 1 samples in Fig. 1c. Actually, rs179785 (A/G) is located at the last nucleotide ("A") of this 12-bp deletion variant (rs200562977), as shown in Fig. 3. Further direct sequencing analysis revealed that Groups 3 and 4 also include a  (Fig. 2c). These findings suggest that it is difficult to correctly genotype rs179785 of KCNQ1 using the TaqMan or DNA micro-array method. Therefore, in subsequent analyses, rs179785 and rs200562977 of KCNQ1 were genotyped by direct sequencing, not by the TaqMan method, in all samples. Moreover, by direct sequencing with the same primers, we were able to evaluate the genotypes of rs179784 of KCNQ1, which is located downstream from rs179785 by 305 bp (Fig. 3), and which shows strong LD with rs179785 (D' = 1.0 and r 2 = 0.99).
All samples were successfully genotyped for the three variants of KCNQ1 (rs179785, rs200562977 and rs179784) by direct sequencing. The call rates for the two SNPs (rs12236871 and rs11653176) by TaqMan method were more than 97.0%. All the variants were in Hardy-Weinberg equilibrium (P > 0.05). Table 1 shows the genotyping results of the three loci (RFX3, KCNQ1 and BCAS3) for 723 clinically-defined gout patients and 913 controls. The common variant of BCAS3, rs11653176, showed a significant association with gout (P = 1.66 × 10 − 3 ; odds ratio [OR] = 0.80; 95% confidence interval [CI]: 0. 70-0.92). The direction of effect was the same as observed in the previous gout GWAS [8]. rs179785 and rs179784 of KCNQ1 had a nominally significant association (P = 0.043 and 0.044; OR = 0.85 and 0.86; 95% CI: 0.73-0.99 and 0.75-1.00, respectively), but did not pass the significance threshold at P value < 0.017 (= 0.05/ 3) for multiple hypothesis testing using the Bonferroni correction. On the other hand, rs200562977 of KCNQ1 and rs12236871 of RFX3 did not show any significant association with gout.

Discussion
In this study, we were able, for the first time, to replicate the association between rs11653176 of BCAS3 and gout. rs2079742, another intronic SNP of BCAS3, was previously reported to have an association with SUA level at the genome-wide significance level; however, it was not replicated in the same report [5]. BCAS3 is a coactivator of estrogen receptor alpha (ER-α) and is overexpressed in breast cancer cells [15], in which it is associated with tumor grade and proliferation [16]. The influence of sex hormones on SUA level is well known [17]. Especially, estradiol is thought to affect SUA levels through mechanisms modulating renal urate reabsorption and Fig. 3 Location of three variants: rs179785, rs200562977 and rs179784 of KCNQ1. rs179785 (A/G) is located at the last nucleotide, "A", of the 12-bp deletion site on rs200562977 (g.2781508_2781519del12GGCGGGCAGGCA). rs179784 is located 305 bp downstream from rs179785 and shows strong linkage disequilibrium with rs179785 (D' = 1.0 and r 2 = 0.99). For direct sequencing of the three variants of KCNQ1, the primers were designed as follows: 5'-ACTTCCTGCCTCTGCTTTC-3′ (forward primer) and 5'-TGAAGGAAGTGACCCCTG-3′ (reverse primer), respectively secretion. Increased SUA levels in postmenopausal women could be caused by the loss of estradiol. In addition, SUA levels decrease in postmenopausal patients using postmenopausal hormone compared with patients not using it [18]. Our findings suggest that risk allele (C) of rs11653176 of BCAS3, may increase renal urate reabsorption which results in increase of SUA levels and gout risk. Thus, although additional genetic and/or functional analyses will be necessary, the common variant of BCAS3 might affect gout susceptibility in ways that are attributable to individual differences in responses to the effects of estrogen. Very recently, we have reported further GWAS of clinically-ascertained gout and identified 10 gout loci including HIST1H2BF-HIST1H4E, solute carrier family 17 member 1 (SLC17A1), solute carrier family 22 member 12 (SLC22A12), NIPA like domain containing 1 (NIPAL1) and family with sequence similarity 35 member A (FAM35A) [19]. Together with these gout loci, BCAS3, which is originally identified by the Chinese gout GWAS, will be very important for personalized genome medicine and/or prevention of gout.

Conclusions
In summary, our present replication study demonstrated, as did a previous gout GWAS [8], an association between gout and the common variant of BCAS3. These findings suggest that the BCAS3 locus is likely to have a common pathophysiological risk for gout.