Introduction

Uric acid is the final product of purine oxidation in humans. Elevated serum urate, or hyperuricemia, has long been recognized as an independent risk factor for gout1,2. There is a renewed interest in hyperuricemia and its association with a number of other clinical disorders including hypertension, atherosclerosis, cardiovascular disease, chronic kidney diseases and abdominal obesity, glucose intolerance, insulin resistance and dyslipidemia, which are often subsumed under the term “metabolic syndrome”3.

Serum urate is balanced between uric acid production in the liver and its disposal via the kidney and gut4. The occurrence of hyperuricemia could be caused by disruptions in any part of this metabolic process. Both genetic and environmental factors, such as gender and body mass index (BMI), have a strong effect on the risk of hyperuricemia3. Among those factors, the attribution of genetic factors is estimated to be as high as 73%5. Recent genome-wide association studies (GWAS) have identified 28 loci associated with serum urate concentration6. However, only approximately 7% of the variation in serum urate concentration could be explained by those reported loci, suggesting the missing heritability remained to be explored6.

Human uncoupling proteins (UCPs) are mitochondrial transporters present in the inner membrane of mitochondria7. UCPs are capable of uncoupling ATP production from mitochondrial respiration by causing proton leak and preventing mitochondrial hyperpolarization and the formation of reactive oxygen species (ROS)8. Among the five identified UCPs, UCP2 is widely expressed in almost all mammalian tissues including white adipose tissue, liver, kidney, pancreatic islets, macrophages and retinal endothelial cells, indicating its involvement in a variety of physiologic or pathologic events9,10,11,12. Two of the most common polymorphisms of this gene, −866G/A (rs659366) in the promoter and Ala55Val (rs660339) in codon 55, were identified as being associated with different phenotypes7,12, including obesity, insulin resistance, type 2 diabetes mellitus (T2D), low-density lipoprotein (LDL) particle size, coronary incidence and other metabolic disorders9,10,13,14,15,16,17,18,19,20,21.

Given the involvement of UCP2 and hyperuricemia in a variety of metabolic disorders, we selected the two common loci −866G/A and Ala55Val to explore the association between genetic UCP2 variants and hyperuricemia in a Chinese population, offering a new diagnostic or therapeutic target for hyperuricemia.

Results

There was no association between SNPs and serum urate

The two loci were proven in Hardy-Weinberg equilibrium (−866G/A: P = 0.990; Ala55Val: P = 0.690). For −866G/A, AA, AG and GG genotypes accounted for 21.6%, 49.9% and 28.6% of hyperuricemic patients, respectively; in healthy controls, the distribution was 21.2%, 49.6% and 29.3%, respectively. As shown in Table 1, the −866G/A polymorphism was not found to be associated with serum urate (AA/GG: Beta = −0.008, P = 0.644; AG/GG: Beta = −0.012, P = 0.474) or with the risk of hyperuricemia (AA/GG: OR = 1.05, P = 0.603; AG/GG: OR = 1.03, P = 0.667). For Ala55Val, the TT, TC and CC genotype distribution was 21.5%, 50.5% and 28.0% in hyperuricemic patients, respectively and the distribution was 21.5%, 49.8% and 28.6% in healthy controls, respectively. No association was observed between Ala55Val polymorphism and serum urate (TT/CC: Beta = −0.013, P = 0.460; TC/CC: Beta = −0.017, P = 0.324). There was no difference in the distribution of the genotypes or alleles among hyperuricemic patients and healthy controls (TT/CC: OR = 1.02, P = 0.824; TC/CC: OR = 1.04, P = 0.652). Therefore, no statistically solid evidence supported the genetic effect of −866G/A and Ala55Val on serum urate or the risk of hyperuricemia in the total group of subjects.

Table 1 Association analysis of UCP2 variants with serum urate and hyperuricemia.
Full size table

UCP2 variants were associated with serum urate and hyperuricemia in female subgroups

As shown in Table 1, we stratified all subjects into male and female subgroups to further explore the gender-related genetic effects of the two polymorphisms. In the male subgroups, there were no significant associations between the two loci and serum urate or the risk of hyperuricemia (all P > 0.025). However, some nominal significant associations were found between −866G/A and the hyperuricemia risk (genotype AA: OR = 1.26, P = 0.038; allele A: OR = 1.12, P = 0.035), indicating a possible risky effect of the −866G/A variant on hyperuricemia incidence in males.

A significant association was found between SNPs and serum urate and hyperuricemia in the female subgroups. The −866G/A genotypes were associated with a lower serum urate (AA/GG: Beta = −0.078, P = 0.015; AG/GG: Beta = −0.104, P = 0.001) and a decreased risk of hyperuricemia (AG/GG: OR = 0.71, P = 0.025). The subjects carrying allele A had a lower serum urate and a decreased risk of hyperuricemia (A/G: Beta = −0.054, P = 0.006; OR = 0.80, P = 0.018). For Ala55Val, genotype TT carriers showed a lower serum urate (TT/CC: Beta = −0.075, P = 0.022) and a decreased risk of hyperuricemia (TT/CC: OR = 0.64, P = 0.020). Genotype TC carriers only had a lower serum urate (TC/CC: Beta = −0.082, P = 0.012) but no decreased risk of hyperuricemia (TC/CC: OR = 0.77, P = 0.093). Allele T was associated with a lower serum urate (T/C: Beta = −0.049, P = 0.016) and a decreased risk of hyperuricemia (T/C: OR = 0.79, P = 0.016).

Further analysis of association in females with different BMI levels

Further analysis was performed regarding the genetic effect of UCP2 variants on serum urate and the risk of hyperuricemia among females with different BMI levels (Table 2). The majority of the females enrolled were stratified into normal- or overweight group (Table 2). In the underweight subgroup, whose sample size was limited after stratification, no significant association was observed between the two loci and serum urate or hyperuricemia risk (all P > 0.025, Table 2). In the normal weight subgroup, −866G/A genotype AA + AG carriers were associated with a lower serum urate (AA + AG/GG: Beta = −0.095, P = 0.022) but not with a decreased risk of hyperuricemia (AA + AG/GG: OR = 0.65, P = 0.076). However, the Ala55Val genotypes or alleles showed no statistical association with serum urate (TT+TC/CC: Beta = −0.070, P = 0.091; T/C: Beta = −0.047, P = 0.106) or hyperuricemia (TT+TC/CC: OR = 0.72, P = 0.173; T/C: OR = 0.72, P = 0.051). In the overweight subgroup, the genotypes of both loci were associated a lower serum urate (AA+AG/GG: Beta = −0.138, P = 0.001; TT+TC/CC: Beta = −0.130, P = 0.003) and a significant, or at least marginal, decreased risk of hyperuricemia (AA + AG/GG: OR = 0.62, P = 0.015; TT+TC/CC: OR = 0.74, P = 0.027). However, the alleles of the loci were associated with a lower serum urate level (A/G: Beta = −0.072, P = 0.019; T/C: Beta = −0.072, P = 0.019) but not with a decreased risk of hyperuricemia (A/G: OR = 0.75, P = 0.036; T/C: OR = 0.74, P = 0.027). Our results suggested a stronger effect of UCP2 variants on overweight females (Table 2).

Table 2 Association of −866G/A and Ala55Val with serum urate and hyperuricemia in females with different BMI levels.
Full size table

Association between haplotypes and risk of hyperuricemia

As listed in Table 3, the haplotypes of the two loci were estimated in the total group of subjects and after stratification by gender. The −866G/A and Ala55Val variants were in strong linkage disequilibrium (D′ = 0.974, r2 = 0.936). The wild type haplotype G-C (−866G/A-Ala55Val) was applied as the reference one. Haplotype A-T made up for the most frequent one, while single mutation at −866G/A or Ala55Val each accounted for less than 1 percent (Table 3). In the total group of subjects, no haplotypes were correlated with susceptibility of hyperuricemia. In the female subgroups, haplotype A-T (−866G/A-Ala55Val) was associated with a decreased risk of hyperuricemia; however, this association was null in males. No further significant associations between hyperuricemia and other two rare haplotypes were found in our study, might partially due to the limited size of the rare haplotypes carriers (Table 3). These results correlated with the association between genotypes or alleles and hyperuricemia (Table 1).

Table 3 The OR and 95% CI between estimated haplotype frequencies.
Full size table

Discussion

Uncoupling protein 2 (UCP2) is present in the inner mitochondrial membrane and mainly decreases the ATP level and ROS produced by electron transport; therefore, UCP2 is involved in a board range of pathological processes. In the present study, we first focused on the relationship between UCP2 variants and serum urate and hyperuricemia, potentially examining the scope of the loci related to hyperuricemia.

The present study revealed no association between the two polymorphisms of UCP2 and serum urate or hyperuricemia in the total group of subjects. However, because serum urate is extensively influenced by gender differences, we stratified the total group of subjects and determined that −866G/A and Ala55Val were associated with serum urate and hyperuricemia in females22,23. Females with the −866G/A genotype AA+AG or allele A had lower serum urate and a decreased risk of hyperuricemia, indicating a protective role of −866G/A for hyperuricemia in females.

The −866G/A variant is a functional polymorphism located in the promoter region and putatively changes the transcription factor binding sites7. The wild type G allele in −866G/A was associated with lower UCP2 mRNA expression19,24. Increased UCP2 mRNA expression from the A allele was translated into an increased amount of UCP2 protein, with corresponding induced proton leak, decreased ATP/ADP ratio and enhanced elimination of ROS10,19. Hypermethylation in the promoter region could affect the binding of transcripation factors, causing aberrant gene expression. Consistent with our expectations, we found a typical CpG island in the UCP2 promoter region, which included the locus of the −866G/A variant, using information from the University of California–Santa Cruz (UCSC; Santa Cruz, CA, USA) database (http://genome.ucsc.edu/cgi-bin/hgGateway). We believe the UCP2 promoter variant −866G/A could shape this CpG island and protect the UCP2 promoter region from DNA methylation, uncovering a novel underlying mechanism that determines −866G/A increases UCP2 transcription.

Uric acid accumulation is caused by the acceleration of ATP degradation to AMP, a precursor of uric acid and UCP2 could decrease the ATP level and lower redundant AMP for uric acid formation7,25. Moreover, an elevation of serum urate concentration occurs as a physiologic response to increased oxidative stress26. Because the ROS level could be down-regulated by UCP2, a counter-regulatory increase of serum urate as an antioxidant defense is less urgent. Therefore, the −866G/A variant in the promoter region might serve as a protective factor through a higher UCP2 mRNA level and increased translation of the UCP2 protein, which might regulate ROS and modify the ATP/ADP ratio.

The other locus, Ala55Val, is a missense variant in exon 4 and is associated with an altered degree of uncoupling7. In our study, a protective effect for hyperuricemia and lower serum urate were observed in genotype TT and allele T in the female subgroups. However, the genetic effect of the Ala55Val variant was less clear. Several researchers identified an association of Ala55Val with the BMI level and type 2 diabetes mellitus (T2D), with controversial conclusions within cohorts and few functional studies were performed14,27,28. Similar to −866G/A, the protective role of the Ala55Val variant for hyperuricemia might also be attributed to altered UCP2 transcription.

In the male subgroups, a less statistically significant but possible effect of −866G/A and Ala55Val was observed for hyperuricemia risk and higher serum urate. Similar gender-associated genetic effects of UCP2 variants were more or less observed for diseases other than hyperuricemia7. For example, Heidema et al. suggested a genetic effect of UCP2 on weight gain was regulated through different mechanism in males and females29. Lee et al. demonstrated that the association between UCP2 variants and BMI was more apparent among female subjects30. Cheurfa et al. confirmed the association of UCP2 variants with coronary artery diseases in males but not females31. In the present study, we found UCP2 variants −866G/A and Ala55Val had a stronger effect on females with hyperuricemia. One possible explanation for the gender-associated genetic effects of UCP2 might be a regulation role of sex hormones such as estrogen. Estrogen was reported to repress UCP2 in a breast cancer cell line and papillary thyroid cancer cells32,33. Taken together, these results suggest the UCP2 protein level was down-regulated by estrogen in females but reversed by the variants of −866G/A and Ala55Val, providing a plausible explanation for the specific protective effects of UCP2 variants on females32.

Genetic effects on hyperuricemia and obesity have been widely recognized3. In the present study, we found that −866G/A and Ala55Val were associated with lower serum urate and a decreased risk of hyperuricemia in overweight females (Table 2). Among the normal weight females, the risk of hyperuricemia and serum urate level were associated with only −866G > A but not Ala55Val, while the association was observed with both the two loci in overweight group. These results indicated that obesity altered the genetic effects of the two loci on serum urate and overweight females might benefit more from the protective effects of UCP2 variants. Among the two loci, the functional −866G > A promoter variant displayed a stronger genetic effect. The association analysis in underweight females might have limited statistic power due to the small sample size (cases/controls: 5/16) and a larger cohort would be necessary for further analysis. The interactions between obesity, uric acid and UCP2 were complicated. BMI has long been viewed as an essential factor influencing uric acid3. UCP2 transcription was activated by fatty acids16. A recent meta-analysis revealed that UCP2 −866G/A and Ala55Val are associated with a risk of obesity27. Subtle intermediary obesity related phenotypes such as elevated triglycerides, total cholesterol concentrations, increased the risk of dyslipidemia and circulating leptin levels were also observed to be correlated with UCP2 variants34. Based on these results, we assumed lipid metabolism material such as fatty acids participated in and enhanced the genetic effect of UCP2 variants on serum urate regulation, explaining the stronger genetic effect of UCP2 variants on females with higher BMI levels observed in the present study.

The −866G/A and Ala55Val variants were in strong linkage disequilibrium (D′ = 0.974, r2 = 0.936). The haplotype frequency analysis revealed that variants of the two loci were more in co-variant haplotype A-T (−866G/A-Ala55Val) compared with the single variant forms of G-T or A-C (less than 1%, Table 3). Haplotype A-T was associated with a decreased risk of hyperuricemia only in females, which was consistent with the genotype or alleles results. However, the small sample size of the two rare haplotypes might limited the power of association analysis with hyperuricemia risk to a certain extent. But we did observed a possible correlation between prevalence of hyperuricemia and the rare haplotype A-C even in such small sample size. A future larger cohort might provide further solid evidence for correlation between the two rare haplotypes and hyperuricemia.

Conclusion

The present study identified a novel gene, UCP2, with two loci, −866G/A and Ala55Val; this gene influenced the serum urate concentrations and the risk of hyperuricemia in females. The associations of those loci were affected by gender and BMI. This study supported the potential involvement of this gene in the prevention, prediction and treatment of hyperuricemia.

Materials and Methods

Study design and approval statement

The study was approved by the Ethical Committees of the School of Life Science of Fudan University (approval number of 140). A total of 4332 subjects of similar lifestyle, diet, housing environment and no medical treatment on serum urate level were enrolled from the Taizhou Longitudinal Study, a prospective cohort initiated to explore the environmental and genetic risk factors for common non-communicable diseases35. The use of the cohort was aimed to minimize the heterogeneity among subjects and to get the maximum robustness of the present study36. The written informed consent approved by the Ethical Committees of the School of Life Science of Fudan University was collected from all participates. The associations of common UCP2 variants with serum urate and hyperuricemia were tested by linear regression and logistic regression with or without gender stratification, respectively. A body mass index (BMI) subgroup was also used for further analysis. All the experiment methods were performed in accordance with the approval document from Ethical Committees of the School of Life Science of Fudan University

Participants

All subjects were enrolled from Taizhou Longitudinal Study35, of which 1387 individuals had serum urate level over 7 mg/dl and were treated as hyperuricemic patients and 2945 individuals had normal serum urate (≤7 mg/dl) and were treated as healthy controls37. The subjects were divided into subgroups (underweight: BMI <18.5; normal weight: 18.50 ≤ BMI < 25; overweight: BMI ≥ 25) following the categories of the World Health Organization (WHO)38. The characteristics of the above individuals are shown in S.Table 1. Written informed consent was obtained from all subjects.

Genetic analysis

Genetic analysis was carried out in accordance with approval document offered by the Ethical Committees of the School of Life Science of Fudan University. For genetic analysis, peripheral blood was collected from all the individuals with written informed consent included in this study. Genomic DNA was extracted from whole blood using the QIAamp DNA Blood Mini kit (QIAGEN, Germany) and was stored at −20 °C. The DNA concentration and quality (including optical density (OD) 260/280 and 260/230 measurements) were determined using a Nanodrop Lite spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). Genotyping of −866G/A and Ala55Val in UCP2 were performed by SNPscan according to the manufacturer’s instructions.

Statistical analysis

The clinical characteristics were presented as the mean ± SD. Student’s t-test was used to test for a significant difference in the mean age, BMI and serum urate between hyperuricemia patients and healthy controls. The chi-square test was used to describe the gender distribution difference between hyperuricemia patients and healthy controls.

The chi-square test was used to test Hardy-Weinberg equilibrium (HWE) of the two loci. We conducted a logistic regression analysis to calculate adjusted odd ratio (OR) with 95% confidence interval (95% CI) and P-values to describe the distribution of −866G/A and Ala55Val adjusted for age and gender between hyperuricemia patients and healthy controls. A linear regression was performed to calculate Beta and P-values to estimate the effect on serum urate in different genotypes and alleles. Genotype GG, allele G of −866G/A and genotype CC, allele C of Ala55Val were used as references, respectively. Stratification into subgroups was performed on the basis of gender and different BMI values for further analysis. Haplotype frequencies between the hyperuricemic patients and controls were estimated by OR (95% CI) and chi-square test. The haplotype of the most frequent (−866G/A-Ala55Val, G-C) was used as the reference. A 2-sided P-value less than 0.025 was considered statistically significant after multiple correlation by Bonferroni method. The PHASE program (V2.1) was used for haplotype frequencies estimation and SPSS 19.0 was used for the statistical analysis.

Additional Information

How to cite this article: Yang, L. et al. Common UCP2 variants contribute to serum urate concentrations and the risk of hyperuricemia. Sci. Rep. 6, 27279; doi: 10.1038/srep27279 (2016).