Objective The present study investigated whether single nucleotide polymorphisms (SNPs) in the human urate transporter 1 (hURAT1) gene are associated with primary hyperuricaemia (HUA) in Han Chinese people.
Methods A total of 538 subjects (215 cases and 323 control subjects) were recruited from Qingdao, China. SNPs in potentially functional regions of the gene were identified and genotypes determined by direct sequencing. Association analyses were conducted using Fisher's exact test and logistic regression assuming a genotype model.
Results By sequencing the promoter, 10 exons, and the exon-intron junctions of the hURAT1 gene, 14 SNPs were identified. Two of the SNPs identified were associated with susceptibility to HUA. The first was a rare intron 3 (11 G→A) SNP (p=0.0005), where carriers of the ‘A’ allele had a 3.4-fold (95% CI 1.67 to 6.93) increased risk of HUA. The second was a common exon 8 (T1309C) SNP (rs7932775), where carriers of one and two ‘C’ alleles had respective fold increased risks of 1.64 (95% CI 1.07 to 2.52) and 2.32 (95% CI 1.37 to 3.95). These SNPs had a joint additive effect of risk of HUA, with those individuals carrying at least one ‘A’ allele at the intron 3 SNP and two ‘C’ alleles at rs7932775 having a 5.88-fold (95% CI 1.25 to 15.57) increased risk of HUA in comparison to those with no risk alleles.
Conclusion In conjunction with other studies, our results suggest that there are multiple genetic variants within or near hURAT1 that are associated with susceptibility to HUA in Han Chinese, including a novel SNP located in intron 3.
- single nucleotide polymorphisms (SNPs)
- genetic screening
- molecular genetics
Statistics from Altmetric.com
- single nucleotide polymorphisms (SNPs)
- genetic screening
- molecular genetics
Primary hyperuricaemia (HUA), an excess of uric acid in the blood, remains a major public health problem. There has been a dramatic rise in the prevalence of HUA over the last 80 years.1 We recently reported that the prevalence of HUA was estimated to be 13.19% in the coastal area of Shandong province in China.2 Uric acid is the end-product of either dietary or endogenous purine metabolism in humans. Serum uric acid concentrations primarily reflect the net balance between urate production through either exogenous or endogenous sources of purines and its excretion via other routes, such as the gastrointestinal tract and the kidneys. The absence of uricase, combined with extensive reabsorption of filtered urate, results in higher urate concentrations in the human plasma than in other mammals.3 As urate concentration increases in physiologic fluids, the risk for supersaturation and crystal formation generally increases. In addition to the morbidity that is attributable to gout, HUA is associated with the insulin resistance syndrome, hypertension, and nephropathy.4 5
Since the 1950s, it has been known that the urate transport mechanisms in the mammalian kidney are complicated6 because urate is transported bi-directionally, which is in contrast to other organic solutes.7 This model, called the ‘4-component theory’, is based on micropancture studies using experimental animals and includes glomerular filtration, reabsorption, secretion, and postsecretory reabsorption.8 9 Net urate excretion is determined by tubular urate secretion and postsecretory urate reabsorption. HUA may occur due to increased uric acid production, decreased uric acid excretion, or a combination of the two. In the general population, 80–90% of gout patients are underexcreters.10 It is thought that transporters for urate reabsorption and secretion may be present throughout the proximal tubules. The direction of net urate flux would thus depend on the number or activity of resorptive and secretory transporters present in the luminal and basolateral membranes, respectively. Urate transporter 1 (URAT1) has been identified as one of the important urate transporters. URAT1 is located in the apical membrane of proximal tubular cells and transports urate from the lumen to proximal tubular cells in exchange for anions.11 This exchanger is essential for proximal tubular reabsorption.12 URAT1 is thought to be the major mechanism for regulating blood urate concentrations.13
The human URAT1 (hURAT1) gene belongs to the organic ion transporter family (SLC22) and consists of 555 amino acid residues and 12 predicted putative transmembrane domains. hURAT1 (encoded by SLC22A12) is exclusively expressed in the kidney and has been localised to the apical membrane of the proximal tubule.11 It has been established that URAT1 is mainly involved in urate reabsorption and mutations in the SLC22A12 gene encoding URAT1 cause renal hypouricaemia. Clinically, a wide variety of therapeutic drugs and pharmacological reagents interact with it in the treatment of gout and HUA. In general, uricosuric drugs directly inhibit hURAT1 from the apical side, whereas antiuricosuric drugs serve as the exchanging anion from inside tubule cells, thereby enhancing urate transport by hURAT1 through trans-stimulation.11 Consistent with this, inactivating mutations in the hURAT1 gene cause idiopathic renal hypouricaemia in Japanese 11 14 15 and Korean 16 patients. Several polymorphisms within or near the hURAT1 gene have been reported recently. Although the functional effects of these polymorphisms have not yet been fully elucidated, we hypothesised that the variants in hURAT1 may have an effect on hURAT1 expression or activity, which may be associated with genetic susceptibility to HUA. To test this hypothesis, we performed a case–control study to investigate the association between single nucleotide polymorphisms (SNPs) discovered through the direct resequencing of the hURAT1 gene and the risk of HUA in Han Chinese people.
A total of 215 primary HUA patients, who visited the Department of Endocrinology at the affiliated Hospital of the Qingdao University Medical College, were selected. HUA was defined as the concentration of serum uric acid >420 μmol/l in males and post-menopausal females, and serum uric acid >357 μmol/l in pre-menopausal females. In our study, most patients (>80%) were not on a regular treatment for HUA. A total of 323 healthy controls with no personal or family history of HUA were also recruited. Written informed consent was obtained from each subject. All patients and controls were of the same predominant genetic background, Han Chinese. Subjects with past histories or possible indications of cancer, hepatic disease, or renal disease were excluded.
The molecular analysis of the hURAT1 gene was performed using genomic DNA obtained from the peripheral blood by conventional methods. Mutations in the promoter, 10 exons, and exon–intron junctions of hURAT1 were screened by direct sequencing ( GenBank accession NG_008110) . Intragenic primers used for the polymerase chain reaction (PCR) were designed by Primer expression 5.0. The detailed list of primers and PCR conditions can be found in table 1. PCR was carried out in a reaction volume of 20 μl containing 50 ng of genomic DNA, 200 μM dNTP, 0.25 units of proTaq DNA polymerase (Promega Corporation, Wisconsin, USA), and 200 μM intragenic primers. The PCR products were purified using shrimp alkaline phosphatase (SAP) and exonuclease I, and then sequencing was performed on an Applied Biosystems model 3730 automated sequencer (Applied Biosystems Corporation, CA, USA). Sequence data were compared with the published sequence of SLC22A12 (NG_008110).
t tests were used to assess the differences in age and blood urate values between cases and controls, while a case–control difference in gender distribution was determined using a Fisher's exact test. For each SNP, Hardy–Weinberg equilibrium (HWE) was estimated within the control subjects using an exact test17 that has been implemented in the Haploview software package.18 Allele frequencies were determined for each of the 14 SNPs overall and in cases and controls separately using standard gene counting, and differences in allele (one degree of freedom tests) and genotype (model-free two degree of freedom tests) frequencies between cases and control subjects were determined using Fisher's exact tests. Within the control subjects, the linkage disequilibrium (LD) measure r2 was calculated for all pairs of the 14 SNPs via an expectation–maximisation algorithm also implemented in Haploview. Based on the high degree of LD between the SNPs, tagSNPs were selected to minimise the number of correlated tests performed using the greedy algorithm of Carlson et al,19 as implemented in Haploview. Using an r2 threshold of 0.8, six tagSNPs were selected to represent the remaining eight SNPs. Based on six tests and an initial type 1 error rate of 0.05, we used a Bonferroni corrected p value threshold of 0.008 to determine statistical significance and adjust for the multiple comparisons done. To estimate odd ratios (ORs), 95% confidence intervals (CIs), and adjust for the possible confounding factors of age and gender, we also used logistic regression models to assess association.
For single SNP results that were determined to be significant, we tested for a multiplicative interaction between them and their risk on HUA using a likelihood ratio test based on a logistic model. Using the six tagSNPs (inferences were the same using all 14 SNPs), haplotype analysis was also conducted under additive, dominant, and recessive genetic models using the haplotype scoring method of Schaid,20 as implemented in the ‘haplo.stats’ R package,21 to account for the haplotype phase ambiguity of the genotype data. Similar to the single SNP analyses, haplotype analyses were conducted both unadjusted as well as adjusted for both age and gender. For individual statistically significant haplotypes, ORs were estimated to assess haplotype effects using a haplotype based generalised linear model (haplo.glm), also implemented in the haplo.stats package. In both the single SNP and haplotype analyses, there was little difference between the unadjusted and age and sex adjusted OR. Therefore, we only present the more parsimonious unadjusted results. Except where noted, all analyses were conducted in the R statistical language (v2.7.2).
Clinical features of the study population
Our study sample consisted of 215 HUA cases and 323 controls. All samples were unrelated Han Chinese individuals recruited from the affiliated Hospital of the Qingdao Medical College, in the coastal city of Qingdoa in Shandong province, China. Demographic and clinical characteristics of the study population are shown in table 2. Serum uric acid concentrations were well characterised, and the values differed significantly between cases and controls in both men (p<0.001) and women (p<0.001). As shown in table 2, cases were on average approximately 5 years older (p<0.001) at diagnosis in comparison to the age of controls at enrolment, and there was also a higher proportion of males among the cases (87.0%) in comparison to controls (63.8%; p<0.001). As a result, all SNP association analyses were conducted both unadjusted as well as adjusted for both age and gender.
Identification of polymorphisms in the hURAT1 gene
By sequencing the promoter, 10 exons, and the exon–intron junctions of the hURAT1 gene, 14 SNPs were identified. As shown in table 3, we discovered SNPs in the following regions of hURAT1: five SNPs were located in the promoter region (rs11602903, rs524023, rs9734313, rs559946, and rs3825018); five SNPs were located at exon–intron junctions (intron 2 rs576076, intron 2 rs10792441, intron 2 rs537246, intron 3 11 G→A, and intron 7 rs7929627); and four SNPs were located within an exon (exon 1 rs3825017, exon 1 rs3825016, exon 2 rs11231825, exon 8 rs7932775). None of the exonic SNPs coded for non-synonymous changes in the amino acid sequence of the hURAT1 protein.
HWE and LD analysis
Using an exact test, all 14 SNPs were determined to be in HWE in the control subjects (minimum p value=0.02 for rs3825017) and were therefore considered for further analysis. As measured by r2, pairwise LD between all 14 SNPs in the hURAT1 gene is shown in figure 1. As can be seen, there were a substantial number of SNPs that were in high LD (r2≥0.8). In total, there were six LD bins (figure 1), where all SNPs within each bin were in high LD with at least one other SNP within each bin, and using the greedy algorithm of Carlson et al,19 there was a corresponding total of six SNPs needed to tag the remaining eight SNPs in the sample.
Allele, genotype, and haplotype association analyses
The minor allele frequencies for each SNP in cases and controls are listed in table 3. Using a Fisher's exact test of differential allele frequency distributions, there were two tagSNPs showing differences with p values less the Bonferonni threshold of 0.008 (adjusted for the six tagSNPs or relatively uncorrelated tests). The first was the variant discovered within intron 3 (provisional dbSNP accession number ss161109885), 11 base pairs away from the 3′ exon 3 (p=0.0005), and the second was an SNP (rs7932775) within exon 8 (p=0.003) coding for a synonymous change in the protein sequence. For both SNPs, the minor allele frequency was higher in cases in comparison to controls, indicating an increased risk of HUA associated with the ‘A’ allele of the intron 3 SNP and the ‘C’ allele of rs7932775. As shown in figure 1, it is noteworthy that these SNPs are in very low LD with each other (r2=0.01). The low LD between these two SNPs reflects the large allele frequency differences in controls, where the “A” allele of the intron 3 SNP is rare (control frequency=0.02) in comparison to the more frequent ‘C’ allele of rs7932775 (control frequency=0.42). Further, the intron 3 SNP is in very low LD with the remaining SNPs (maximum r2=0.03), and with the exception of the single SNP tagged (intron 7 rs7929627; r2=1.0), rs7932775 is also in low LD with the other SNPs identified (maximum r2=0.2).
Table 4 shows the single SNP genotype frequencies for both cases and controls and unadjusted tests of association between single SNPs and risk of HUA using a genotype model (two degrees of freedom) for the six tagSNPs. The unadjusted results for the tagSNPs are presented here for brevity, as the tagSNPs had very similar results to the non-tagSNPs that they were selected to represent and because there was little evidence of confounding by age and sex (ie, the adjusted OR and 95% CI differed little from the unadjusted results). For completeness, we present an additional table as supplementary data (supplemental table 1), which contains both the unadjusted and sex and age adjusted results for all 14 SNPs.
In accordance with the allelic associations, the genotype model results indicate that the rare ‘A’ allele of the novel intron 3 SNP (p value=0.0009) and the relatively common ‘C’ allele of rs7932775 (p value=0.006) are both associated with an increased risk of HUA. In comparison to non-carriers, carriers of at least one ‘A’ allele (dominant model) at the intron 3 SNP had an increased risk of HUA of 3.40 (95% CI 1.67 to 6.93). For rs7932775, heterozygous and homozygous carriers of the ‘C’ had respective increased HUA risks of 1.64 (95% CI 1.07 to 2.52) and 2.23 (95% CI 1.37 to 3.95) relative to subjects homozygous for the non-risk allele. While there was no evidence of a multiplicative interaction between these two SNP and risk of HUA (interaction p=0.98), the two SNPs did appear to have an additive effect (p=7.6e−5 for a model containing both SNPs) on their risk of HUA. The 21 subjects with at least one risk allele at the intron 3 SNP and one risk allele at rs7932775 were at a 4.41-fold (95% CI 1.39 to 14.00) increased risk in comparison to those subjects with no risk variants at either SNP. The 14 subjects with at least one risk allele at the intron 3 SNP and two risk allele at rs7932775 were at a 5.88-fold (95% CI 1.25 to 15.57) increased risk in comparison to those subjects with no risk variants at either SNP.
The results of the haplotype association analysis were essentially consistent with the single SNP results (table 5). Under a dominant model (inferences were the same under an additive model, data not shown), there was statistically significant evidence of an overall haplotype association with susceptibility to primary HUA (p value=0.0004). In particular, there were two haplotypes that were associated with an increased risk. The first haplotype (h11 in table 5) was inferred to include both the intron 3 and rs7932775 risk alleles (p value=0.002), and similar to the two SNP result, it was associated with an increased risk of 4.15 (95% CI 1.64 to 10.48). The second significant haplotype (h12 in table 5) contained the risk allele for rs7932775 alone (p=0.001) and was associated with an increased risk of 4.84 (95% CI 0.82 to 28.54). The overall consistency between haplotype and single SNP results are to be expected given the observed LD pattern described above.
In the present study, we resequenced the promoter, all exons, and all intron/exon boundaries of the hURAT1 gene in a Han Chinese sample of HUA cases and controls to identify genetic variation that may have a direct functional association with susceptibility to HUA. In addition to the 13 SNPs that were previously catalogued in dbSNP, we identified a novel SNP in intron 3 (provisional dbSNP accession number ‘ss161109885’), 11 base pairs from the 3′ exon 3. This SNP showed the strongest level of association with primary HUA. While the risk allele at the intron 3 SNP was rare, we also identified a second independent and more common risk allele for rs7932775, which encodes a synonymous change in the protein sequence and is genetically identical (r2=1.0) to an SNP in intron 7 (rs7929627). To our knowledge, our study is the first to demonstrate statistically significant associations between these SNPs and primary HUA. These finding suggest that there are multiple independent genetic variants, both rare and common, in or near the hURAT1 gene that are associated with increased susceptibility to primary HUA in the Han Chinese.
There have been other studies that have investigated the association between common polymorphisms in hURAT1 and the risk of HUA. While the primary focus of a study of Japanese by Iwai et al15 was inactivating mutations in hURAT1 for subjects with renal hypouricaemia, they also investigated the effects of common polymorphisms of the hURAT1 gene on serum uric acid concentrations. Similar to our study, they identified six tagSNPs which captured the remaining common variants identified in their sample, suggesting a common LD structure in both populations. After correction for multiple testing, none of their single SNP analyses were significant. However, a marginally significant increase in serum uric acid concentration was seen for an intron 7 (p=0.16) SNP that was in high LD (r2=0.94) with our exon 8 SNP (rs7932775). Further, through a haplotype analysis, they identified a rare haplotype which was associated with uric acid concentrations. Of note, while their sequencing did identify the same intron 3 SNP evaluated in our study (minor allele frequency in the Japanese sample was 1.1%), this SNP was not included in the association analysis for HUA, presumably due to its low frequency in their sample.
In a more recent Japanese study,22 a common SNP (rs893006) in intron 4 of hURAT1 was shown to be associated with serum uric acid values. In this study, the major allele (‘G’) at rs893006 was associated with increased serum uric acid concentrations in male subjects (p=0.025). This finding was corroborated by a study of gout in a Chinese population.23 In the HapMap, rs893006 is in perfect LD (r2=1.0) in the Han Chinese samples with one of the promoter SNPs genotyped in our study (rs9734313). Similarly, our results suggest that the major allele (‘C’) at rs9734313 is associated with an increased risk of HUA (p=0.028), providing further validation evidence for this result. Further, in another report from a Korean study, SNP rs1529909 6092T→C, also in intron 4, was found to be involved in renal handling and the concentration of serum uric acid in Korean male subjects.24 However, this SNP was genotyped in neither our study nor the HapMap, negating our ability to assess its association with the other associated SNPs in this region.
Consistent with these findings, in a study of hURAT1's influence of fractional excretion of uric acid and HUA in a German Caucasian population by Graessler et al,25 it was found that SNPs in the ‘N-terminal’ region of hURAT1 were associated with these phenotypes. In particular, their best evidence for association with reduced fractional excretion of uric acid and a concomitant increased risk of HUA was with the 426C→T SNP (rs11231825) in exon 2. Under an additive model, the ‘T’ allele was significantly associated with an increased risk of HUA (p=0.0002). While the genotype distribution differed for this SNP between German Caucasians (CC=48.7; CT=44.5; and TT=6.8) and our Han Chinese (CC=4.7; CT=29.0; and TT=66.3) controls, the allele associated with increased risk of HUA in both studies is the ‘T’ allele. Our best evidence for association with this SNP came under a dominant model (p=0.015), where carriers of the ‘T’ allele had an increased risk of 1.56 (95% CI 1.09 to 2.24) in comparison to non-carriers. Therefore, our finding for this SNP is an independent confirmation of the original finding by Graessler et al.25 Further, we were also able validate their finding for the promoter SNP rs11602903 (p=0.028). Similar to rs11231825, the minor allele was different in the two populations; however, the ‘A’ allele remained associated with increased risk in both studies. In our Han Chinese sample, rs11231825, rs11602903, and rs9734313 (tagging rs893006, the intron 4 SNP mentioned above) are all in high LD (minimum r2=0.79) with one another, suggesting these three SNPs map to a common N-terminal HUA associated locus for Han Chinese.
While not the focus of the paper, Graessler et al also detected a marginal statistically significant effect (p=0.102) with the same exon 8 SNP (rs7932775) that we found to be significant,25 which was the same SNP that indirectly (through LD) showed suggestive evidence in the study by Iwai et al, referenced above. Again, similar to our study, Graessler et al showed that carriers of one and two ‘C’ alleles had, respectively, increased risks of HUA of 1.27 (95% CI 0.95 to 1.69) and 1.61 (95% CI 0.91 to 2.85). Therefore, our result provides the first evidence of a ‘C-terminal’ locus within hURAT1 that is associated with HUA. While Graessler et al did not find the intron 3 SNP in their study (possibly suggesting its low frequency in Caucasians or specificity for east Asian populations), the consistency of results in the two studies suggest that there is variation throughout the hURAT1 gene that is associated with HUA.
While our study took a candidate gene approach to identify SNPs associated with HUA, a number of genome-wide association (GWA) studies have been conducted analysing uric acid concentration as a quantitative trait.26–30 Interestingly, these studies independently did not show genome-wide significant results for SNPs in hURAT1. However, the most comprehensive meta-analysis of GWA studies of uric acid concentration to date from Kolz M et al31 (which included data from the GWA studies of Li et al, Wallace et al, Vitart et al, and Doring et al) did identify a series of SNPs in high LD that implicated hURAT1. The most significant SNP in this region was rs505802 (p=2.0×10−9), which is in high LD with rs11231825 in both HapMap Caucasian (r2=1.0) and Han Chinese (r2=0.97) samples, adding support for an N-terminal locus of hURAT1 that is associated with HUA. Of note, neither of the two SNPs (the intron 3 SNP or rs7932775) identified as statistically significant by our study were directly genotyped or imputed (ie, not evaluated for association) in any of these GWA studies. Based on our Han Chinese sample, these two SNPs are in very low LD (r2≤0.2) with the other SNPs that we found in hURAT1. Similarly, rs7932775, which is currently in the phase III version of the HapMap, is in low LD with all other HapMap SNPs in hURAT1 in both Caucasian (r2<0.45) and Han Chinese (r2<0.20) samples. While the frequency of the intron 3 SNP is still unclear outside of Japan and China, this low level of pair-wise LD tagging could explain why the current GWA studies have not identified more independent variants in hURAT1 associated with uric acid concentrations.
In the present study, our best evidence for association is the intron 3 SNP, which does not directly alter the protein sequence. However, this SNP could alter the splicing of the mRNA. There is growing evidence that alternative splicing is common among renal transporters.32 In most transporter genes, the spliced isoforms have been shown to be functional, resulting in a variety of physiological consequences. Mechanisms that increase protein diversity in all metazoans include the use of multiple transcription start sites,33 alternative pre-mRNA splicing,34–37 polyadenylation,38 pre-mRNA editing,39 and post-translational protein modifications.40 Among these mechanisms, alternative pre-mRNA splicing is considered to be the most important source of protein diversity in vertebrates.34 37 The fundamental problem in pre-mRNA splicing is ‘exon recognition’, the process by which exons are distinguished from introns and intron–exon boundaries are accurately defined.41 42 In humans, introns are generally larger than exons, requiring the splicing machinery to recognise small exon sequences located within vast stretches of intronic mRNA. Moreover, 5′ and 3′ splice sites are poorly conserved, and introns contain large numbers of cryptic splice sites, which can be selected for splicing when normal splice sites are altered by mutation. Therefore, given its proximity to the exon 4 intron–exon boundary, it is possible that the intron 3 SNP may alter the function of hURAT1 via exonic splicing, directly altering risk of HUA, or this SNP could be in LD with another, as of yet unidentified causal genetic variant.
In summary, our results alone and in conjunction with other studies indicate that there are at least three independent genetic loci in or near hURAT1 that are associated with risk of primary HUA in the Han Chinese. First, we were able to validate previous SNP associations in the N-terminal region of hURAT1 either directly (rs11231825 and rs11602903) or indirectly (rs893006 via LD with rs9734313), which all appear to be in high LD with one another (ie, one signal) in Han Chinese. Of note, this locus has also been shown to be associated with uric acid concentration in the most comprehensive meta-analysis of GWA studies to date. Further, for the first time, we have also demonstrated statistical significance for an independent, common locus in the C-terminal region of hURAT1 (rs7932775). Most interestingly, we discovered a third locus that may be specific to a novel SNP located in intron 3 that is associated with an increased risk of HUA in the Han Chinese. Our new genetic findings in HUA should be further validated in larger/ethnically distinct studies to assess the generalisability of these association findings and assess the potential use of these SNPs as markers of genetic susceptibility to HUA.
This work was supported by grants from the National Science Foundation of China (NSFC: 30570890 and 30871192) and the Henry Ford Health System Research Fund.
CL, LH and AML contributed equally to this work.
Funding Other Funders: National Science Foundation of China and Henry Ford Health System Research Fund.
Competing interests None.
Ethics approval This study was conducted with the approval of the Qingdao University.
Patient consent Obtained.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.