Subtype-specific gout susceptibility loci and enrichment of selection pressure on ABCG2 and ALDH2 identified by subtype genome-wide meta-analyses of clinically defined gout patients

Objectives Genome-wide meta-analyses of clinically defined gout were performed to identify subtype-specific susceptibility loci. Evaluation using selection pressure analysis with these loci was also conducted to investigate genetic risks characteristic of the Japanese population over the last 2000–3000 years. Methods Two genome-wide association studies (GWASs) of 3053 clinically defined gout cases and 4554 controls from Japanese males were performed using the Japonica Array and Illumina Array platforms. About 7.2 million single-nucleotide polymorphisms were meta-analysed after imputation. Patients were then divided into four clinical subtypes (the renal underexcretion type, renal overload type, combined type and normal type), and meta-analyses were conducted in the same manner. Selection pressure analyses using singleton density score were also performed on each subtype. Results In addition to the eight loci we reported previously, two novel loci, PIBF1 and ACSM2B, were identified at a genome-wide significance level (p<5.0×10–8) from a GWAS meta-analysis of all gout patients, and other two novel intergenic loci, CD2-PTGFRN and SLC28A3-NTRK2, from normal type gout patients. Subtype-dependent patterns of Manhattan plots were observed with subtype GWASs of gout patients, indicating that these subtype-specific loci suggest differences in pathophysiology along patients’ gout subtypes. Selection pressure analysis revealed significant enrichment of selection pressure on ABCG2 in addition to ALDH2 loci for all subtypes except for normal type gout. Conclusions Our findings on subtype GWAS meta-analyses and selection pressure analysis of gout will assist elucidation of the subtype-dependent molecular targets and evolutionary involvement among genotype, phenotype and subtype-specific tailor-made medicine/prevention of gout and hyperuricaemia.

the secondarily significant SNP, rs56093838 of SLC22A12/URAT1 (C), revealing these signals to be detected from the same locus. Because URAT1 is a well-known urate transporter that markedly affects serum uric acid level, the true associated gene for combined type gout on this locus is unlikely to be CDC42BPG. These findings suggest that SLC22A12/URAT1 is the actual associated gene for combined type gout on this locus (chromosome 11q13.1). Due to the lack of frequency data in 1000 genomes phase 3 JPT, r 2 data were calculated using 1000 Genomes Project Phase_3:EAS samples 1 .

Study subjects and patients' involvement
We performed subtype genome-wide meta-analyses based on two case-control data sets for clinicallydefined gout that included the Japonica Array 3 and Illumina Array data sets. Patients with known clinical parameters were recruited from Japanese male outpatients at the gout clinics of Jikei University Hospital (Tokyo, Japan), Midorigaoka Hospital (Osaka, Japan), Kyoto Industrial Health Association (Kyoto, Japan), Ryougoku East Gate Clinic (Tokyo, Japan), Nagase Clinic (Tokyo, Japan), Tokorozawa Central Hospital (Tokorozawa, Japan) and Wakasa Clinic (Tokorozawa, Japan). All 3104 cases (1048 cases for the Japonica Array and 2056 for the Illumina Array) were clinically diagnosed as having primary gout according to the criteria established by the American College of Rheumatology, 4 and their subtypes were also diagnosed along with their clinical parameters as described previously [5][6][7][8] (Table 1 and Online Supplementary Figure S1). As controls, 6081 individuals (1179 and 4902 controls for the Japonica Array and Illumina Array, respectively) were assigned from Japanese male participants in the Japan Multi-Institutional Collaborative Cohort Study (J-MICC Study). 9,10 This research was done without patient involvement. Patients were not invited to comment on the study design and were not consulted to develop patient-relevant outcomes or interpret the results. Patients were not invited to contribute to the writing or editing of this document for readability or accuracy.

Genotyping and imputation for the Japonica Array data set
A total of 1048 male clinically-defined gout cases and 1179 male controls from the J-MICC Study 9,10 were genotyped with the use of a Japonica SNP Array. 3 The clustering plots were first classified by the Ps classification function in the SNPolisher package (version 1.5.2; Affymetrix). Single nucleotide polymorphisms (SNPs) that were assigned 'recommended' by the Ps classification were retained. Seven samples (two cases and five controls) with a genotype call rate of < 0.98 were excluded. Three controls with inconsistent sex information between questionnaires and the estimate from genotype were excluded. The identity-by-descent method implemented in PLINK 1.9 software 11

Genotyping and imputation for the Illumina Array data set
As case data, 2056 male clinically-defined gout cases subjects were genotyped with the use of HumanOmniExpress or HumanOmniExpressExome BeadChip Arrays (Illumina, San Diego, CA, USA).
One sample with a genotype call rate of < 0.98 was excluded. No samples showed a discrepancy between genetic and reported sex. The identity-by-descent method implemented in PLINK 1.9 software 11 detected 17 duplicates or closely-related pairs of samples (pi-hat > 0.1875), with one sample of each pair being excluded. PCA 12 with the 1000 Genomes Project reference panel (Phase 3) 1 detected six subjects who were estimated to have ancestries outside the Japanese population: these were excluded.
As control data, 4902 male controls with SUA data from the J-MICC study, 9,10 who had been previously genotyped at the RIKEN Center for Integrative Medicine Sciences using a HumanOmniExpressExome BeadChip Array (Illumina, San Diego, CA, USA) with sample QC of the control data, 15,16 were used. No samples were excluded for reasons of a genotype call rate of < 0.98 or a discrepancy between genetic and reported sex. The identity-by-descent method implemented in PLINK 1.9 software 11 detected one sample having a relationship with a gout case (pi-hat > 0.1875), which was excluded. PCA 12 with the 1000 Genomes Project reference panel (Phase 3) 1 detected no subjects estimated to have ancestries outside the Japanese population. After sample QC, the case and control datasets were merged. Of the SNPs that were consistently genotyped across the arrays, we excluded SNPs at the pseudo-autosomal regions of chromosome X as well as those with a genotype call rate of < 0.98, a Hardy-Weinberg equilibrium exact test P value of < 1 × 10 -6 or a MAF of < 0.01. This quality control filtering resulted in the selection of 2032 case subjects and 4901 control subjects as well as 553,321 SNPs. Post-imputation quality control was performed by excluding SNPs with an imputation quality score (r 2 ) of < 0.3, MAF < 0.01 and indels. 1062 controls with hyperuricemia (SUA > 7.0 mg/dl) or with a past history of gout were excluded. 244 samples (7 cases and 237 controls) that overlapped with the Japonica Arrays samples were