Gout is characterised by recurrent attacks of intra-articular crystal deposition of monosodium urate (MSU). Its clinical manifestations include recurrent painful attacks of acute inflammatory arthritis, tophi, uric acid urolithiasis, renal impairment and metabolic syndrome. The risk factors of gout have also been complicated by environmental factors and genetic components, including diet, particularly fatty meats rich in purines; alcohol consumption1 and fructose2; low-level exposure to lead3; age and male gender. Most genetic components are based on rare forms of Mendelian disorders—for example, hypoxanthine-guanine phosphoribosyltranserase gene for X-linked gout,4 5 autosomal dominant medullary cystic disease on chromosome 1q21,6 familial juvenile hyperuricaemic nephropathy on chromosome 167 and uric acid nephrolithiasis on chromosome 10.8 The tumour necrosis factor α gene on chromosome 6,9 SLC22A12 gene or human urate transporter 1 gene,10 or based on the susceptible regions, such as chromosome 1q2111 or 4q2512 have also been shown to be related to gout or hyperuricaemia. However, gout is not rare in the general population, with an incidence of acute gout of about 5% every year among hyperuricaemic patients with serum urate levels >9.0 mg/dl,13 and a prevalence of 0.3% in adult Fukien-Taiwanese.14 Thus, we suspect that some susceptible genes may contribute to the pathogenesis of the disease.

Hyperuricaemia is well known to be the most important and direct factor in gout, possibly occurring because of decreased excretion or increased production of uric acid, or a combination of these two mechanisms. Underexcretion of uric acid accounts for most cases of hyperuricaemia, whereas overproduction accounts for only a minority of patients presenting with hyperuricaemia.15 Another independent mechanism for the development of gout disease is the inflammation response to MSU. Monocytes and immature macrophages act to stimulate and amplify an acute attack of gout through the release of tumour necrosis factor α,16 17 interleukin 1 (IL1),18 IL6,19 IL820 and cyclo-oxygenase 2,21 whereas differentiated macrophages may have an anti-inflammatory role in terminating an acute attack and in preserving the asymptomatic state through production of transforming growth factor β.22

However, the relationships between the associated genes and gout are various, involving the pathogenesis of hyperuricaemia, the inflammation response to MSU, or other causal pathways. Hence, we designed a genome-wide scan method to screen for susceptible genetic markers related to gout and to confirm their association by polymorphism analysis in a case–control study.

METHODS

Participants in our study were divided into two groups: one group comprised a family with gout which had eight patients with gout and 10 gout-free subjects; the other, included 148 unrelated subjects composed of 74 male patients with gout and 74 male healthy controls. The 74 patients with gout were enrolled from Kaohsiung Chang-Gung Memorial Hospital and 74 controls were enrolled from a community clinic. Gout was diagnosed by a rheumatologist in the hospital according to the criteria of Wallace et al,23 and the controls were diagnosed by a clinical doctor.

All eight patients with gout of the gouty family were receiving treatment for the disease with benzbromarone and none of them was obese or being treated for hypertension. Hyperuricaemia was defined in those subjects with plasma uric acid ⩾7 mg/dl. All the participants were of Fukien-Taiwanese heritage and none was Taiwanese aborigine, a population in which the prevalence of gout is high.14 The family with gout was included to delineate the association with gout by a microsatellite genome-wide scan method.

For the 148 case–control subjects, we performed a matched design by age within 1 year for further confirmation after the results were obtained from the genome-wide scan method. The process and design of this study were approved by the human research ethics committee of Kaohsiung Medical University Hospital and Kaohsiung Chang-Gung Memorial Hospital. Informed consent was obtained from each participant.

Genotyping of the genome-wide scan was provided by the Mammalian Genotyping Service of Marshfield Laboratory. The DNA samples were genotyped using the screening set 13, which has more accurate allele calling, less error or mistyping as a consequence of higher quality of trinucleotide and tetranucleotide short tandem repeat polymorphisms and more accurate mapping and spacing.24 A total of 382 autosomal markers spread over 22 chromosomes were used in this study. The marker with the largest logarithm of the odds (LOD) score among the 382 autosomal markers was identified as located at chromosome 4q21 from the genome-wide scan for the gout family (D4S3243, at 81 289 553 bp). Then we chose 29 single nucleotide polymorphisms (SNPs) spread 10 cM around this marker D4S3243, each SNP representing one gene, to confirm their relationships with gout in the case–control subjects. The 29 polymorphisms were located in the regions between 77 235 867 bp and 87 230 704 bp on chromosome 4. A further four SNPs located on the target gene were chosen to make fine associations with gout.

The MassARRAY (Sequenom, San Diego, USA) was used as a high-throughput genotyping system to identify the aforementioned 33 SNPs, which were selected to locate in the novel susceptible region among the 148 case–control subjects. The primers and probes for the multiplex PCR were designed by the SpectroDESIGNER software (Sequenom).

Statistics

We applied the PedCheck program25 and Statistical Analysis for Genetic Epidemiology (SAGE) software (MARKINFO, PEDINFO program) to check for any inconsistent Mendelian inheritance, or any genotyping error for the gout family. For the gout trait, we first performed non-parametric linkage (NPL)26 using Genehunter software (v2.1_5 beta), then transferred the NPL score into the LOD score using the formula (LOD = NPL²/(2Ln10)).27

The Hardy–Weinberg equilibrium among the control group was analysed by the method suggested by Wigginton et al.28 The t test was used to test the mean differences of demographic data, the Monte Carlo for χ² test was used to test the associations in genotypes of the 33 polymorphisms, and the odds ratios (OR) and 95% confidence intervals (95% CIs) were used to assess the strength of relationships in the inherited models, genotype and allele distributions of polymorphisms between the patients with gout and controls. Haplotype frequencies were estimated in the two polymorphisms of rs1458046 and rs11736177 and we also used the ORs and 95% CIs to estimate the haplotype relative risk between patients with gout and controls. Additionally, a conditional logistic regression was used to analyse the associations between selected polymorphisms and gout for adjustment of hyperuricaemia. If the p value was <0.05, the difference was considered to be significant. The PHASE program (V2.1) was used for haplotype frequencies estimation and SAS software (V9.13) was used for the other statistical analysis.

RESULTS

The family with gout included 44% (8/18) patients with gout, with a mean (SD) age of 38.8 (14.9) years, which did not differ significantly from the mean age of the gout-free subjects in the family (43.1 (14.8), p = 0.521). None of the gouty family was an alcohol consumer and all the patients with gout were male and receiving treatment. We used a genome-wide scan with 382 microsatellite polymorphisms to screen for the gout susceptible genetic markers of the gouty family. The multipoint NPL method was used for the analysis using marker information in each chromosome to screen for linkage results of all 22 chromosomes. Figure 1 shows the LOD score obtained from all 22 chromosomes. The highest LOD score (LOD score = 5.13; NPL = 4.86, p = 0.004) was found to be located at 87.9 cM (marker D4S3243, physical distance: 81 289 553 bp) on chromosome 4q21 for gout trait.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Multipoint linkage analysis using a non-parametric linkage (NPL) method in the genome-wide scan of a Taiwanese family with gout. The x axis represents the chromosome location for the 22 autosomes and the y axis represents the logarithm of the odds (LOD) score, calculated from the NPL_all statistic. The best peak is at 87.9 cM (marker D4S3243, physical distance: 81 289 553 bp) on chromosome 4q21, with LOD score = 5.13 for gout trait.

Figure 2 shows the construction of the pedigree and haplotype structure near 87.9 cM on chromosome 4. In total, 18 people were recruited from the pedigree; the 18 members were genome-scanned and another two members (II-3, II-11) were reconstructed from their offspring. We found all male subjects who inherited D4S3243 allele 3 had gout disease and the one relative who had not inherited allele 3 was gout-free with normouricaemia (III-8).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Pedigree of family with gout trait in this study and haplotypes for markers near D4S3243 on chromosome 4 are shown. The haplotypes of II-3 and II-11 were reconstructed from their offspring. All male subjects who inherited D4S3243 allele 3 had gout and the one who had not inherited allele 3 was gout-free (III-8). The haplotype segregating the disease locus is indicated by a black bar. A filled square or circle indicates a man or woman with gout. The symbol # indicates a subject with hyperuricaemia. The numbers beside the square or circle are age.

After the marker D4S3243 had been identified as the susceptible marker for gout, we conducted an SNP analysis in the 148 case–control subjects to explore the associations between polymorphisms of the neighbourhood regions of D4S3243 (87.9 cM) and gout. We first chose 29 polymorphisms as the candidates from 77 235 867 bp to 87 230 704 bp to show the associations between patients with gout and healthy controls. Table 1 lists the associations between the 29 polymorphisms and gout; our results showed that two polymorphisms, rs1458046 and rs11736177, had a significant association with gout (p = 0.013 for rs1458046 and p = 0.007 for rs11736177, respectively). A further four SNPs located in the same gene (cGK II) with rs11736177 were chosen to show the association with gout. Table 1 lists the associations of 33 SNPs with gout, the p value estimated in Hardy–Weinberg equilibrium and the genetic position.

The means of age, body mass index and systolic blood pressure among patients with gout were not significantly different from those of healthy controls (all p>0.05; table 2). However, the means of uric acid, diastolic blood pressure and creatinine were higher in the patients with gout than controls (p<0.05; table 2). A total of 36.5% (27/74) of patients with gout and 13.5% (10/74) of controls were receiving treatment for hypertension (p<0.01). Table 2 lists the associations in the distribution of genotypes, allele frequency, dominant and recessive models of inheritance between gout disease and the polymorphisms rs1458046 and rs11736177. At polymorphism rs1458046, the genotype GG was 12.22 times more likely to be associated with the development of gout than the genotype AA (95% CI 1.50 to 99.42, p = 0.004). The recessive model also displayed the same significant result (genotype GG vs AA/GA, p = 0.006), but neither the allele frequency nor the dominant model were found to have a significant association at polymorphism rs1458046. For polymorphism rs11736177, those with the genotype CC had a 3.32 times greater risk of developing gout disease than those with genotype AA (95% CI 1.17 to 9.37, p = 0.021) and both the allele C frequency and the recessive model showed the same significant association between patients with gout and controls (all p<0.05), but the dominant model did not show this significant association (p = 0.257, table 2). The power estimation for polymorphisms rs1458046 and rs11736177 in the recessive model were 77% and 85%, respectively.

Table 3 shows the haplotype frequency at polymorphisms rs1458046 and rs11736177 and the haplotype relative risk analysis between patients with gout and controls. With those carrying haplotype A/A (rs1458046/rs11736177) as the referent group, those carrying haplotypes A/C and haplotype G/C showed significant associations between patients with gout and control group (OR = 1.80, p = 0.039; OR = 3.32, p = 0.002; respectively). However, the haplotype G/A did not show the same significant association (OR = 1.52, 95% CI 0.70 to 3.31, p = 0.289). In other words, those carrying allele C at rs11736177 were more likely to develop gout, but those carrying allele G at polymorphism rs1458046 had no increased risk.

Because the polymorphism rs11736177 is located on cyclic GMP-dependent protein kinase II (cGK II) gene, we chose four further polymorphisms (rs10033237, rs7688672, rs6837293 and rs17005080) on the cGK II gene to refine our analysis of the association between cGK II and gout. Table 4 lists the associations with gout. At polymorphism rs10033237 (in intron 1), the recessive model (genotype GG vs genotypes GA or AA) showed a significant association between patients with gout and healthy controls (OR = 2.68, 95% CI 1.29 to 5.58, p = 0.008), but the allele frequency, dominant model and genotypes did not show the same association (p>0.05). At polymorphisms rs7688672 (in intron 8) and rs6837293 (in intron 10), all the genotypes, allele frequency and recessive model showed a significant association with gout (p<0.05). Furthermore, after adjustment for hyperuricaemia, the results showed only the recessive model in the polymorphisms of rs7688672 and rs6837293 to be significantly associated with gout disease (p = 0.019 and p = 0.026, respectively). However, no genotype distribution, allele frequency, or dominant or recessive model at polymorphism rs17005080 (in exon 3) showed the same association with gout (all p>0.05). The power of polymorphisms rs7688672 and rs6837293 in the recessive model were estimated to be 89% and 85%, respectively.

DISCUSSION

Using a genome-wide scan method, we found that a susceptible marker (D4S3243) mapping at chromosome 4q21 showed an association with the development of gout. Furthermore, we performed a finer association using SNP analysis on neighbourhood regions of marker D4S3243 and found that the most likely candidate gene was cGK II after adjusting for hyperuricaemia in a recessive model. cGK II is a cGMP-dependent protein kinase and has been mapped to chromosome 4q13.1–q21.1. The cGK II function in kidney is as an inhibitor of renin, which cleaves the peptide bond between the leucine and valine residues on angiotensinogen to create angiotensin I. The angiotensin I is then converted to angiotensin II by the enzyme angiotensin-converting enzyme. Angiotensin II was shown to cause vasoconstriction, renal vascular resistance increase, a decreased renal plasma flow rate and increased serum uric acid levels.29 However, under the dysfunction of cGK II activity, it may increase renin activity and increase the blood pressure and serum uric acid levels since cGK II is a renin inhibitor. In the healthy subjects in our study, those carrying the recessive genotypes at rs7688672 and rs6837293 did not show significant differences either in mean systolic blood pressure or serum uric acid levels from the other genotypes (data not shown, p>0.05). The finding did not suggest an association between the effects of the two polymorphisms in a recessive model and the consequence of the activated renin angiotensin system. In addition, we found that significant associations between the two polymorphisms in cGK II and gout disease were independent of hyperuricaemia; thus, we do not favour the suggestion that the causal association between cGK II and gout disease occurs through the renin angiotensin pathway, since this association did not provide evidence of increased serum uric acid levels in those carrying the recessive genotype.

Most of the urate can be excreted from the kidney by the action of organic anion transporter 1 (OAT1), OAT3 and urate transporter 1 (URAT1).30^–32 Some studies have even demonstrated that cGK II affects chondrocytes in rats33 and endochondral ossification in mice,34 but no study has found such effects on the joints or soft tissue in humans. Furthermore, during the acute gout phase, the monocytes and immature macrophages have a key role, secreting tumour necrosis factor α, IL1β, IL6 and IL8 and promoting secondary neutrophil capture by endothelial cells under physiological flow in response to MSU crystal uptake.18 35 Meanwhile, while the monocytes differentiate to macrophages, the differentiated macrophages play an anti-inflammatory role in terminating an acute attack by ingesting MSU crystals without proinflammatory cytokine secretion or endothelial cell activation.35 36 Despite the cGK II abundance in various organs,34 37 no study has reported the way in which the functions of cGK II act on endothelial activation in humans and the association with gout. We postulate that the dysfunction of cGK II may follow a pathway independent from hyperuricaemia, causing gouty arthritis which does not terminate endothelial cell activation or proinflammatory cytokine secretions.

Many genes occur within the range of 72.5 to 93.4 cM on chromosome 4. For instance, the polycystic kidney disease 2 (PKD2) gene,38 has been associated with renal tubular function. Mutations in this gene have been associated with autosomal dominance,39 suggesting that the PKD2 gene cannot be used to explain the association with gout in our study which suggests a recessive model. Moreover, Wallace et al40 also discovered the common solute carrier family 2 (SLC2A9) variants on chromosome 4 that increase serum urate and that each copy of the common allele increased the risk for hyperuricaemia. However, the evidence is limited to the association with urate levels, so the association between SLC2A9 and gout should be further confirmed. Our study did not show associations of the genes located in the X chromosome with gout, even though all the patients with gout in the gouty family were male.

In summary, using a genome-wide scan method in a case–control study, we found that a susceptible genetic marker D4S3243 (4q21)was significantly related to gout and that the cGK II gene was most likely to have an association with the gout after adjustment for hyperuricaemia in a recessive model. Meanwhile, we postulate that the underlying pathway of the association of cGK II with gout may due to the non-termination of the proinflammatory cytokine secretion or endothelial cell activation.