Abstract

Uric acid (UA) is the final catabolic product of purine metabolism and elevated levels are associated with diabetes and cardiovascular disease. A recent meta-analysis of genome-wide association studies totalling 28 141 participants identified five novel loci associated with serum UA levels. In our population-based cohort of 7795 subjects, we replicated four of these five loci; PDZK1 (rs12129861, P = 1.07 × 10−3), glucokinase regulator protein (GCKR) (rs780094, P = 4.83 × 10−4), SLC16A9 (rs742132, P = 0.047) and SLC22A11 (rs17300741, P = 6.13 × 10−3), but not LRRC16A (rs742132, P = 0.645). Serum UA concentration is a complex trait, closely associated to renal UA handling (fractional UA excretion, P < 1 × 10−300), renal function (serum creatinine, P < 1 × 10−300) and the metabolic syndrome (including fasting insulin, P = 2.48 × 10−232; insulin resistance, P = 2.51 × 10−258; waist circumference, P < 1 × 10−300) and systolic blood pressure (P = 1.93 × 10−219). Together these factors explain 67% of the variance in UA levels. Therefore, we sought to determine the potential contribution of these factors to the association of these novel loci with UA levels, by including them as additional explanatory variables in our analyses, and by considering them as alternative response variables. The association with the GCKR locus is attenuated by serum triglycerides and fractional UA excretion. We also observed the GCKR locus to be associated with total cholesterol (P = 7.52 × 10−6), triglycerides (P = 2.65 × 10−9), fasting glucose (P = 0.011), fractional UA excretion (P = 3.36 × 10−5) and high-sensitive CRP (P = 1.18 × 10−3) also after adjusting for serum UA levels. We argue that GCKR locus affects serum UA levels through a factor that also affects triglycerides.

INTRODUCTION

Uric acid (UA) is the final catabolic product of purine metabolism. Serum UA level is a complex phenotype determined by the balance of urate production by purine catabolism and renal excretion. Serum UA levels cluster with various traits, including dietary factors, renal function, increased body mass index, blood pressure, insulin resistance and other components associated with the metabolic syndrome (1–5). As such, genetic associations with serum UA might originate from purine metabolism, renal function or just represent a proxy for another highly correlated trait.

A recent meta-analysis of genome-wide association studies from 14 studies totalling 28 141 participants of European descent identified five novel loci associated with serum UA at genome-wide significance levels (6). The novel loci included the glucokinase regulator protein (GCKR) of which the same SNP has previously been strongly associated with triglycerides, glucose and insulin levels (7,8). For this and the other loci, the mediating mechanism for the reported associations remains to be elucidated. Understanding mechanisms through which SNPs are associated with serum UA levels could help us in understanding the role in pathology and help identify new targets for intervention.

The reported genome-wide meta-analysis did not attempt to explore the nature of the association of identified SNPs with clinical biochemical correlates of renal function and the metabolic syndrome (6). Therefore, our aim was to determine the contribution of these potentially mediating factors for the association of the novel loci associated with serum UA levels. We genotyped the lead SNPs of the five novel loci in a cohort of 7795 subjects for which extensive phenotypic data are available in an attempt to determine the potential contribution of various metabolic and renal factors to the association of these loci with UA levels.

RESULTS

Replication of novel loci

Table 1 shows the baseline characteristics of the 7795 participating subjects from the Prevention of REnal and Vascular ENd stage Disease (PREVEND) study (genotyping data are shown in Supplementary Material, Table S2). Using models adjusted only for age and gender, we found evidence for association with serum UA concentration for four of the five novel reported loci: PDZK1 (rs12129861, P = 1.07 × 10−3), GCKR (rs780094, P = 4.83 × 10−4), SLC16A9 (rs742132, P = 0.047) and SLC22A11 (rs17300741, P = 6.13 × 10−3); but found no evidence for association for LRRC16A (rs742132, P = 0.645) (Table 2 and Fig. 1). Taking into account the previously reported genome-wide association data (6), the total evidence increased for the four replicated loci (P in the range 2.57 × 10−9 to 1.70 × 10−15) (Table 2). Plotting the mean UA concentration against the alleles suggests an additive effect of the PDZK1, SLC16A9 and SLC22A11 loci and a recessive effect of the T-allele in the GCKR locus (Fig. 1). We tested the independence of the loci by testing them all together in one age and gender adjusted model. We observed the four replicated loci to remain significant with similar effect sizes (Supplementary Material, Table S3). There is a linear dependence of UA concentrations on a genetic score obtained by counting the total number of risk alleles of the four replicated loci (P = 1.31 × 10−6 for data from PREVEND alone; Fig. 1).

Figure 1.

Unadjusted serum uric acid levels per genotype. Shown are the unadjusted serum uric acid levels ± standard error of mean per SNP and for the number of risk alleles.

Figure 1.

Unadjusted serum uric acid levels per genotype. Shown are the unadjusted serum uric acid levels ± standard error of mean per SNP and for the number of risk alleles.

Table 1.

Baseline characteristics association with serum UA

 Total (n = 7795) Serum UA
 
P-value 
  Correlation coefficient Variance explained  
Age (years) 49.6 ± 12.7 0.226 0.051 4.02 × 10−90 
Females 50.8% — 0.272 <1 × 10−300 
Serum uric acid (mg/dl) 5.15 ± 1.29 — — — 
Protein intake (g/day) 77 [54–112] 0.267 0.010 2.19 × 10−18 
Alcohol intake (g/day/kg body weight) 0.070 [0.007–0.166] 0.145 0.021 5.37 × 10−37 
Blood pressure 
 Systolic (mmHg) 129.2 ± 20.3 0.335 0.112 1.27 × 10−201 
 Diastolic (mmHg) 74.0 ± 9.8 0.348 0.122 1.93 × 10−219 
 Blood pressure lowering medication 15.8% 0.240 0.058 2.30 × 10−101 
Lipids 
 Total cholesterol (mg/dl) 217.4 ± 40.8 0.178 0.032 2.70 × 10−55 
 LDL cholesterol (mg/dl) 141.6 ± 36.8 0.192 0.037 4.89 × 10−62 
 HDL cholesterol (mg/dl) 50.8 ± 14.7 −0.366 0.134 6.68 × 10−241 
 Fasting triglycerides (mg/dl) 104.5 [77.3–151.8] 0.372 0.139 1.69 × 10−249 
 Lipid lowering medication 6.5% 0.124 0.015 6.66 × 10−28 
Renal function parameters 
 Absolute uric acid excretion (mg/kg/24 h) 0.16[0.11–0.23] −0.176 0.031 2.16 × 10−53 
 Fractional uric acid excretion (%) 41.8 ± 20.5 −0.504 0.255 <1 × 10−300 
 Serum creatinine (mg/dl) 0.94 ± 0.15 0.517 0.268 <1 × 10−300 
 Estimated glomerular filtration rate (ml/min/1.73 m293.9 ± 18.5 — 0.274 <1 × 10−300 
 Urinary albumin excretion (mg/24 h) 9.51 [6.37–17.94] 0.223 0.049 5.63 × 10−87 
 Cystatin-C (mg/l) 0.80 ± 0.17 0.351 0.123 1.30 × 10−211 
Metabolism 
 Fasting glucose (mg/dl) 85.5 [80.0–92.7] 0.270 0.073 7.19 × 10−120 
 Fasting insulin (μU/ml) 7.9 [5.5–11.7] 0.368 0.136 2.48 × 10−232 
 Homeostasis model assessment-estimated insulin resistance 1.64 [1.11–2.58] 0.388 0.150 2.51 × 10−258 
 Waist circumference (cm) 88.6 ± 12.6 0.518 0.269 <1 × 10−300 
 Body mass index (kg/m226.0 ± 4.0 0.368 0.136 8.36 × 10−228 
 Oral glucose lowering medication 1.7% 0.042 0.002 2.13 × 10−4 
Inflammation 
 High-sensitive CRP (mg/l) 1.45 [0.67–3.23] 0.172 0.030 2.16 × 10−46 
All above variables in one model   0.67  
 Total (n = 7795) Serum UA
 
P-value 
  Correlation coefficient Variance explained  
Age (years) 49.6 ± 12.7 0.226 0.051 4.02 × 10−90 
Females 50.8% — 0.272 <1 × 10−300 
Serum uric acid (mg/dl) 5.15 ± 1.29 — — — 
Protein intake (g/day) 77 [54–112] 0.267 0.010 2.19 × 10−18 
Alcohol intake (g/day/kg body weight) 0.070 [0.007–0.166] 0.145 0.021 5.37 × 10−37 
Blood pressure 
 Systolic (mmHg) 129.2 ± 20.3 0.335 0.112 1.27 × 10−201 
 Diastolic (mmHg) 74.0 ± 9.8 0.348 0.122 1.93 × 10−219 
 Blood pressure lowering medication 15.8% 0.240 0.058 2.30 × 10−101 
Lipids 
 Total cholesterol (mg/dl) 217.4 ± 40.8 0.178 0.032 2.70 × 10−55 
 LDL cholesterol (mg/dl) 141.6 ± 36.8 0.192 0.037 4.89 × 10−62 
 HDL cholesterol (mg/dl) 50.8 ± 14.7 −0.366 0.134 6.68 × 10−241 
 Fasting triglycerides (mg/dl) 104.5 [77.3–151.8] 0.372 0.139 1.69 × 10−249 
 Lipid lowering medication 6.5% 0.124 0.015 6.66 × 10−28 
Renal function parameters 
 Absolute uric acid excretion (mg/kg/24 h) 0.16[0.11–0.23] −0.176 0.031 2.16 × 10−53 
 Fractional uric acid excretion (%) 41.8 ± 20.5 −0.504 0.255 <1 × 10−300 
 Serum creatinine (mg/dl) 0.94 ± 0.15 0.517 0.268 <1 × 10−300 
 Estimated glomerular filtration rate (ml/min/1.73 m293.9 ± 18.5 — 0.274 <1 × 10−300 
 Urinary albumin excretion (mg/24 h) 9.51 [6.37–17.94] 0.223 0.049 5.63 × 10−87 
 Cystatin-C (mg/l) 0.80 ± 0.17 0.351 0.123 1.30 × 10−211 
Metabolism 
 Fasting glucose (mg/dl) 85.5 [80.0–92.7] 0.270 0.073 7.19 × 10−120 
 Fasting insulin (μU/ml) 7.9 [5.5–11.7] 0.368 0.136 2.48 × 10−232 
 Homeostasis model assessment-estimated insulin resistance 1.64 [1.11–2.58] 0.388 0.150 2.51 × 10−258 
 Waist circumference (cm) 88.6 ± 12.6 0.518 0.269 <1 × 10−300 
 Body mass index (kg/m226.0 ± 4.0 0.368 0.136 8.36 × 10−228 
 Oral glucose lowering medication 1.7% 0.042 0.002 2.13 × 10−4 
Inflammation 
 High-sensitive CRP (mg/l) 1.45 [0.67–3.23] 0.172 0.030 2.16 × 10−46 
All above variables in one model   0.67  

Shown are the baseline characteristics (mean ± SD, median [inter-quartile range] or % were appropriate) of the population and the correlation coefficient of each baseline characteristic with serum uric acid (UA) levels, the variance explained and the P-value for the correlation.

Table 2.

Five novel loci and association with serum uric acid concentrations

Locus SNP Effect allele (frequency) Beta SE 95% CI P-value Previous reported Beta* Previous reported SE* Previous reported P-value* Overall evidence Beta Overall SE Overall P-value 
PDZK1 rs12129861 A (48%) −0.0583 0.0178 [−0.0931; −0.0234] 1.07 × 10−3 −0.0623 0.0105 2.68 × 10−9 −0.0613 0.0090 1.25 × 10−11 
GCKR rs780094 T (35%) 0.0636 0.0182 [0.0279; 0.0992] 4.83 × 10−4 0.0515 0.0085 1.40 × 10−9 0.0537 0.0077 3.20 × 10−12 
LRRC16A rs742132 A (70%) 0.0087 0.0190 [−0.0286; 0.0461] 0.645 0.0538 0.0093 8.50 × 10−9 0.0451 0.0084 6.77 × 10−8 
SLC16A9 rs12356193 A (83%) 0.0468 0.0235 [0.0045; 0.0931] 0.047 0.0779 0.0136 1.07 × 10−8 0.0700 0.0118 2.57 × 10−9 
SLC22A11 rs17300741 A (48%) 0.0478 0.0175 [0.0136; 0.0821] 6.13 × 10−3 0.0616 0.0082 6.68 × 10−14 0.0591 0.0074 1.70 × 10−15 
Locus SNP Effect allele (frequency) Beta SE 95% CI P-value Previous reported Beta* Previous reported SE* Previous reported P-value* Overall evidence Beta Overall SE Overall P-value 
PDZK1 rs12129861 A (48%) −0.0583 0.0178 [−0.0931; −0.0234] 1.07 × 10−3 −0.0623 0.0105 2.68 × 10−9 −0.0613 0.0090 1.25 × 10−11 
GCKR rs780094 T (35%) 0.0636 0.0182 [0.0279; 0.0992] 4.83 × 10−4 0.0515 0.0085 1.40 × 10−9 0.0537 0.0077 3.20 × 10−12 
LRRC16A rs742132 A (70%) 0.0087 0.0190 [−0.0286; 0.0461] 0.645 0.0538 0.0093 8.50 × 10−9 0.0451 0.0084 6.77 × 10−8 
SLC16A9 rs12356193 A (83%) 0.0468 0.0235 [0.0045; 0.0931] 0.047 0.0779 0.0136 1.07 × 10−8 0.0700 0.0118 2.57 × 10−9 
SLC22A11 rs17300741 A (48%) 0.0478 0.0175 [0.0136; 0.0821] 6.13 × 10−3 0.0616 0.0082 6.68 × 10−14 0.0591 0.0074 1.70 × 10−15 

Shown are the five novel serum uric acid (UA) loci and the strength of the association (Beta) with serum UA in the PREVEND study. The previous reported effect size and standard error have been combined weighting by the inverse variance to calculate the overall evidence for association.

*Kolz et al. (6).

Association of UA concentrations with clinical and biochemical characteristics

Serum UA concentration is a complex trait and it is closely associated to protein intake, renal UA handling, renal function and components of the metabolic syndrome (2–5). In the PREVEND sample, we found these factors to be closely associated with serum UA (Table 1). The strongest bivariate correlations were found for gender (0.27; P < 1 × 10−300) Fractional UA excretion (r2 0.26; P < 1 × 10−300), estimated glomerular filtration rate (r2 0.27; P < 1 × 10−300) and waist circumference (r2 0.27; P < 1 × 10−300). Considering all non-genetic variables together, 67% of the variation in serum UA could be explained.

Potential mediators explaining the association of novel loci with UA concentrations

To study the nature of the observation between the five novel loci and serum UA, we determined whether the association could be driven by potential mediating factors representing UA clearance by the kidney, renal function and lipoprotein and metabolic pathways. In separate linear regression models that considered potential mediating factors, either one at a time or in groups, we tested the independence of the loci explaining UA levels (with all models adjusted for age and gender). For the PDZK1 (rs12129861), there was no substantial change in the genetic effect after adjustment for potential mediators. For GCKR (rs780094) locus, we noted that triglycerides and fractional UA excretion considerably attenuated the association with UA concentrations (Table 3). When both triglycerides and fractional UA excretion were considered together, the effect of GCKR locus was reduced to 0.010 ± 0.017 (P = 0.548). Excluding subjects using relevant medication known to influence a specific trait for the relevant association analysis did not change our findings (Supplementary Material, Tables S4–S6).

Table 3.

Age and gender adjusted association of novel loci for serum uric acid after adjustment for potential mediators/confounders

 PDZK1 (rs12129861)
 
GCKR (rs780094)
 
LRRC16A (rs742132)
 
 Beta ± SE P-value Beta ± SE P-value Beta ± SE P-value 
Basic model −0.0583 ± 0.0178 0.0011 0.0636 ± 0.0182 4.83E−4 0.0087 ± 0.0190 0.645 
Basic model plus each mediator/confounder, added one variable/group at a time 
 Protein intake −0.0585 ± 0.0181 0.0012 0.0638 ± 0.0185 5.70E−4 0.0033 ± 0.0194 0.864 
 Alcohol intake −0.0586 ± 0.0179 0.0011 0.0656 ± 0.0183 3.50E−4 0.0089 ± 0.0192 0.641 
Blood pressure 
 Systolic −0.0691 ± 0.0175 7.59E−5 0.0634 ± 0.0179 3.84E−4 0.0094 ± 0.0187 0.615 
 Diastolic −0.0630 ± 0.0176 3.53E−4 0.0641 ± 0.0180 3.72E−4 0.0069 ± 0.0188 0.711 
 Blood pressure lowering medication −0.0610 ± 0.0175 4.99E−4 0.0653 ± 0.0179 2.68E−4 0.0103 ± 0.0187 0.582 
Lipids 
 Total cholesterol −0.0597 ± 0.0177 7.58E−4 0.0529 ± 0.0182 0.0033 0.0027 ± 0.0190 0.888 
 LDL cholesterol −0.0638 ± 0.0180 4.04E−4 0.0570 ± 0.0185 0.0020 0.0088 ± 0.0193 0.650 
 HDL cholesterol −0.0605 ± 0.0176 5.79E−4 0.0653 ± 0.0179 2.83E−4 0.0110 ± 0.0188 0.560 
 Triglycerides −0.0667 ± 0.0174 1.22E−4 0.0315 ± 0.0178 0.0778 0.0087 ± 0.0187 0.639 
 Lipid lowering medication −0.0565 ± 0.0177 0.0015 0.06238 ± 0.0181 5.89E−5 0.0101 ± 0.0190 0.593 
Renal function parameters 
 Absolute uric acid excretion −0.0632 ± 0.0179 4.21E−4 0.0593 ± 0.0183 0.0012 0.0078 ± 0.0191 0.680 
 Fractional uric acid excretion −0.0535 ± 0.0162 9.81E−4 0.0297 ± 0.0167 0.0747 0.0081 ± 0.0174 0.643 
 Serum creatinine −0.0560 ± 0.0171 0.0011 0.0755 ± 0.0174 1.59E−5 0.0036 ± 0.0183 0.842 
 Estimated glomerular filtration rate −0.0563 ± 0.0171 0.0010 0.0735 ± 0.0175 2.76E−5 0.0021 ± 0.0183 0.908 
 Urinary albumin excretion −0.0600 ± 0.0177 7.08E−4 0.0619 ± 0.0181 6.38E−4 0.0109 ± 0.0189 0.566 
 Cystatin-C −0.0650 ± 0.0177 2.47E−4 0.0696 ± 0.0181 1.23E−4 0.0079 ± 0.0189 0.675 
 Oral glucose lowering medication −0.0589 ± 0.0178 9.64E−4 0.0644 ± 0.0182 4.19E−4 0.0093 ± 0.0190 0.626 
Metabolism 
 Fasting glucose −0.0663 ± 0.0181 2.56E−4 0.0735 ± 0.0186 7.87E−5 0.0005 ± 0.0195 0.981 
 Fasting insulin −0.0634 ± 0.0170 1.97E−4 0.0753 ± 0.0175 1.66E−5 −0.0020 ± 0.0183 0.915 
 Homeostasis model assessment-estimated insulin resistance −0.0651 ± 0.0170 1.37E−4 0.0780 ± 0.0175 8.44E−6 0.0024 ± 0.0183 0.895 
 Waist circumference −0.0642 ± 0.0168 1.34E−4 0.0673 ± 0.0172 9.48E−5 0.0160 ± 0.180 0.373 
 Body mass index −0.0640 ± 0.0167 1.29E−4 0.0588 ± 0.0171 5.82E−4 0.0167 ± 0.0179 0.351 
Inflammation 
 High-sensitive CRP −0.0653 ± 0.0186 4.54E−4 0.0393 ± 0.0190 0.0387 0.0118 ± 0.0198 0.552 
All above −0.0401 ± 0.0141 0.0045 0.0407 ± 0.0146 0.0054 −0.0034 ± 0.0152 0.821 

 
 SLC16A9 (rs12356193) SLC22A11 (rs17300741)   
 Beta ± SE P-value Beta ± SE P-value   

 
Basic model 0.0468 ± 0.0234 0.048 0.0485 ± 0.0183 0.0083   
Basic model plus each mediator/confounder, added one variable/group at a time   
 Protein intake 0.0454 ± 0.0240 0.059 0.0475 ± 0.0177 0.0075   
 Alcohol intake 0.0433 ± 0.0238 0.069 0.0514 ± 0.0176 0.0035   
Blood pressure   
 Systolic 0.0496 ± 0.0232 0.032 0.0429 ± 0.0171 0.012   
 Diastolic 0.0510 ± 0.0234 0.029 0.0427 ± 0.0173 0.013   
 Blood pressure lowering medication 0.0443 ± 0.0232 0.057 0.0462 ± 0.0171 0.0071   
Lipids   
 Total cholesterol 0.0424 ± 0.0236 0.072 0.0507 ± 0.0182 0.0055   
 LDL cholesterol 0.0471 ± 0.0240 0.049 0.0493 ± 0.0177 0.0055   
 HDL cholesterol 0.0469 ± 0.0234 0.045 0.0483 ± 0.0172 0.0051   
 Triglycerides 0.0493 ± 0.0231 0.033 0.0400 ± 0.0170 0.019   
 Lipid lowering medication 0.0162 ± 0.0260 0.050 0.0490 ± 0.0174 0.0048   
Renal function parameters   
 Absolute uric acid excretion 0.0427 ± 0.0238 0.073 0.0441 ± 0.0176 0.012   
 Fractional uric acid excretion 0.0406 ± 0.0216 0.060 0.0365 ± 0.0159 0.022   
 Serum creatinine 0.0419 ± 0.0227 0.065 0.0502 ± 0.0168 0.0028   
 Estimated glomerular filtration rate 0.0401 ± 0.0227 0.078 0.0529 ± 0.0168 0.0016   
 Urinary albumin excretion 0.0410 ± 0.0235 0.082 0.0500 ± 0.0174 0.0040   
 Cystatin-C 0.0379 ± 0.0234 0.107 0.0461 ± 0.017 0.0078   
Metabolism   
 Fasting glucose 0.0462 ± 0.0241 0.056 0.0448 ± 0.0178 0.012   
 Fasting insulin 0.0536 ± 0.0227 0.018 0.0388 ± 0.0167 0.020   
 Homeostasis model assessment-estimated insulin resistance 0.0545 ± 0.0227 0.017 0.0410 ± 0.0168 0.015   
 Waist circumference 0.0528 ± 0.0224 0.018 0.0700 ± 0.0165 2.33E−5   
 Body mass index 0.0541 ± 0.0222 0.015 0.0641 ± 0.0164 9.31E−5   
 Oral glucose lowering medication 0.0450 ± 0.0237 0.058 0.0478 ± 0.0175 0.0062   
Inflammation   
 High-sensitive CRP 0.0431 ± 0.0246 0.081 0.0464 ± 0.0182 0.011   
All above 0.0340 ± 0.0188 0.067 0.0113 ± 0.0140 0.417   
 PDZK1 (rs12129861)
 
GCKR (rs780094)
 
LRRC16A (rs742132)
 
 Beta ± SE P-value Beta ± SE P-value Beta ± SE P-value 
Basic model −0.0583 ± 0.0178 0.0011 0.0636 ± 0.0182 4.83E−4 0.0087 ± 0.0190 0.645 
Basic model plus each mediator/confounder, added one variable/group at a time 
 Protein intake −0.0585 ± 0.0181 0.0012 0.0638 ± 0.0185 5.70E−4 0.0033 ± 0.0194 0.864 
 Alcohol intake −0.0586 ± 0.0179 0.0011 0.0656 ± 0.0183 3.50E−4 0.0089 ± 0.0192 0.641 
Blood pressure 
 Systolic −0.0691 ± 0.0175 7.59E−5 0.0634 ± 0.0179 3.84E−4 0.0094 ± 0.0187 0.615 
 Diastolic −0.0630 ± 0.0176 3.53E−4 0.0641 ± 0.0180 3.72E−4 0.0069 ± 0.0188 0.711 
 Blood pressure lowering medication −0.0610 ± 0.0175 4.99E−4 0.0653 ± 0.0179 2.68E−4 0.0103 ± 0.0187 0.582 
Lipids 
 Total cholesterol −0.0597 ± 0.0177 7.58E−4 0.0529 ± 0.0182 0.0033 0.0027 ± 0.0190 0.888 
 LDL cholesterol −0.0638 ± 0.0180 4.04E−4 0.0570 ± 0.0185 0.0020 0.0088 ± 0.0193 0.650 
 HDL cholesterol −0.0605 ± 0.0176 5.79E−4 0.0653 ± 0.0179 2.83E−4 0.0110 ± 0.0188 0.560 
 Triglycerides −0.0667 ± 0.0174 1.22E−4 0.0315 ± 0.0178 0.0778 0.0087 ± 0.0187 0.639 
 Lipid lowering medication −0.0565 ± 0.0177 0.0015 0.06238 ± 0.0181 5.89E−5 0.0101 ± 0.0190 0.593 
Renal function parameters 
 Absolute uric acid excretion −0.0632 ± 0.0179 4.21E−4 0.0593 ± 0.0183 0.0012 0.0078 ± 0.0191 0.680 
 Fractional uric acid excretion −0.0535 ± 0.0162 9.81E−4 0.0297 ± 0.0167 0.0747 0.0081 ± 0.0174 0.643 
 Serum creatinine −0.0560 ± 0.0171 0.0011 0.0755 ± 0.0174 1.59E−5 0.0036 ± 0.0183 0.842 
 Estimated glomerular filtration rate −0.0563 ± 0.0171 0.0010 0.0735 ± 0.0175 2.76E−5 0.0021 ± 0.0183 0.908 
 Urinary albumin excretion −0.0600 ± 0.0177 7.08E−4 0.0619 ± 0.0181 6.38E−4 0.0109 ± 0.0189 0.566 
 Cystatin-C −0.0650 ± 0.0177 2.47E−4 0.0696 ± 0.0181 1.23E−4 0.0079 ± 0.0189 0.675 
 Oral glucose lowering medication −0.0589 ± 0.0178 9.64E−4 0.0644 ± 0.0182 4.19E−4 0.0093 ± 0.0190 0.626 
Metabolism 
 Fasting glucose −0.0663 ± 0.0181 2.56E−4 0.0735 ± 0.0186 7.87E−5 0.0005 ± 0.0195 0.981 
 Fasting insulin −0.0634 ± 0.0170 1.97E−4 0.0753 ± 0.0175 1.66E−5 −0.0020 ± 0.0183 0.915 
 Homeostasis model assessment-estimated insulin resistance −0.0651 ± 0.0170 1.37E−4 0.0780 ± 0.0175 8.44E−6 0.0024 ± 0.0183 0.895 
 Waist circumference −0.0642 ± 0.0168 1.34E−4 0.0673 ± 0.0172 9.48E−5 0.0160 ± 0.180 0.373 
 Body mass index −0.0640 ± 0.0167 1.29E−4 0.0588 ± 0.0171 5.82E−4 0.0167 ± 0.0179 0.351 
Inflammation 
 High-sensitive CRP −0.0653 ± 0.0186 4.54E−4 0.0393 ± 0.0190 0.0387 0.0118 ± 0.0198 0.552 
All above −0.0401 ± 0.0141 0.0045 0.0407 ± 0.0146 0.0054 −0.0034 ± 0.0152 0.821 

 
 SLC16A9 (rs12356193) SLC22A11 (rs17300741)   
 Beta ± SE P-value Beta ± SE P-value   

 
Basic model 0.0468 ± 0.0234 0.048 0.0485 ± 0.0183 0.0083   
Basic model plus each mediator/confounder, added one variable/group at a time   
 Protein intake 0.0454 ± 0.0240 0.059 0.0475 ± 0.0177 0.0075   
 Alcohol intake 0.0433 ± 0.0238 0.069 0.0514 ± 0.0176 0.0035   
Blood pressure   
 Systolic 0.0496 ± 0.0232 0.032 0.0429 ± 0.0171 0.012   
 Diastolic 0.0510 ± 0.0234 0.029 0.0427 ± 0.0173 0.013   
 Blood pressure lowering medication 0.0443 ± 0.0232 0.057 0.0462 ± 0.0171 0.0071   
Lipids   
 Total cholesterol 0.0424 ± 0.0236 0.072 0.0507 ± 0.0182 0.0055   
 LDL cholesterol 0.0471 ± 0.0240 0.049 0.0493 ± 0.0177 0.0055   
 HDL cholesterol 0.0469 ± 0.0234 0.045 0.0483 ± 0.0172 0.0051   
 Triglycerides 0.0493 ± 0.0231 0.033 0.0400 ± 0.0170 0.019   
 Lipid lowering medication 0.0162 ± 0.0260 0.050 0.0490 ± 0.0174 0.0048   
Renal function parameters   
 Absolute uric acid excretion 0.0427 ± 0.0238 0.073 0.0441 ± 0.0176 0.012   
 Fractional uric acid excretion 0.0406 ± 0.0216 0.060 0.0365 ± 0.0159 0.022   
 Serum creatinine 0.0419 ± 0.0227 0.065 0.0502 ± 0.0168 0.0028   
 Estimated glomerular filtration rate 0.0401 ± 0.0227 0.078 0.0529 ± 0.0168 0.0016   
 Urinary albumin excretion 0.0410 ± 0.0235 0.082 0.0500 ± 0.0174 0.0040   
 Cystatin-C 0.0379 ± 0.0234 0.107 0.0461 ± 0.017 0.0078   
Metabolism   
 Fasting glucose 0.0462 ± 0.0241 0.056 0.0448 ± 0.0178 0.012   
 Fasting insulin 0.0536 ± 0.0227 0.018 0.0388 ± 0.0167 0.020   
 Homeostasis model assessment-estimated insulin resistance 0.0545 ± 0.0227 0.017 0.0410 ± 0.0168 0.015   
 Waist circumference 0.0528 ± 0.0224 0.018 0.0700 ± 0.0165 2.33E−5   
 Body mass index 0.0541 ± 0.0222 0.015 0.0641 ± 0.0164 9.31E−5   
 Oral glucose lowering medication 0.0450 ± 0.0237 0.058 0.0478 ± 0.0175 0.0062   
Inflammation   
 High-sensitive CRP 0.0431 ± 0.0246 0.081 0.0464 ± 0.0182 0.011   
All above 0.0340 ± 0.0188 0.067 0.0113 ± 0.0140 0.417   

Shown are the effect size (Beta) ± standard error (SE) and P-value for the association of each locus with serum uric acid (UA) levels. First the basic model is presented only adjusted for age and gender below which each parameter is added to this basic model. The final row under each header considers all variables of that particular header in the basic model.

Association of novel loci with associated traits

Next we determined whether the five novel loci were associated with traits correlated with UA levels, considering one trait at a time. To determine whether such associations were potentially mediated by UA levels, we examined whether the strength of association changed when we adjusted for UA levels (with all models adjusted for age and gender) (Table 4; model 1). In models not adjusted for UA, we observed some association for the PDZK1 locus and systolic blood pressure (P = 2.90 × 10−3) and strong associations for the GCKR locus with total cholesterol (P = 7.52 × 10−6), fasting triglycerides (P = 5.65 × 10−9), renal function parameters (up to P = 3.36 × 10−5 for fractional UA excretion) and a weaker association with GCKR with C-reactive protein (P = 1.18 × 10−3). Finally we observed an association between SLC22A11 and waist circumference (P = 8.89 × 10−4), an effect that was present also when stratifying instead of adjusting for gender (data not shown). Some of these associations were attenuated when UA was included in the model (Table 4; model 2). The effect of the GCKR locus on triglyceride levels remained highly significant when both UA and fractional UA excretion were considered together in one model (P = 2.03 × 10−6). However, the effect of the GCKR locus on fractional UA excretion was markedly attenuated when both UA and triglycerides were considered together (P = 0.013).

Table 4.

Association of loci with potential mediators

 PDZK1 (rs12129861)
 
GCKR (rs780094)
 
LRRC16A (rs742132)
 
 Model 1
 
Model 2
 
Model 1
 
Model 2
 
Model 1
 
Model 2
 
 Beta ± SE P-value Beta ± SE P-value Beta ± SE P-value Beta ± SE P-value Beta ± SE P-value Beta ± SE P-value 
Protein intake 0.007 ± 0.008 0.372   0.006 ± 0.008 0.452   −0.003 ± 0.009 0.762   
Alcohol intake 0.003 ± 0.003 0.391   −0.003 ± 0.003 0.412   0.004 ± 0.003 0.284   
Blood pressure 
 Systolic 0.852 ± 0.286 0.0029 1.040 ± 0.282 2.25E−4 0.010 ± 0.293 0.972   −0.067 ± 0.307 0.828   
 Diastolic 0.209 ± 0.140 0.135   −0.029 ± 0.143 0.840   0.090 ± 0.150 0.550   
Lipids 
 Total cholesterol 0.154 ± 0.634 0.808   2.92 ± 0.651 7.52E−6 2.477 ± 0.647 1.30E−4 0.664 ± 0.681 0.329   
 LDL cholesterol 0.275 ± 0.600 0.647   1.648 ± 0.618 0.0076 1.395 ± 0.617 0.024 0.606 ± 0.647 0.349   
 HDL cholesterol −0.058 ± 0.219 0.791   0.0529 ± 0.225 0.814   0.011 ± 0.235 0.962   
 Fasting triglycerides 0.006 ± 0.008 0.447   0.049 ± 0.008 2.65E−9 0.039 ± 0.008 8.06E−7 0.000 ± 0.009 0.976   
Renal function parameters 
 Absolute uric acid excretion −0.001 ± 0.010 0.933   −0.002 ± 0.010 0.846   0.001 ± 0.010 0.943   
 Fractional uric acid excretion 0.116 ± 0.318 0.715   −1.351 ± 0.326 3.36E−5 −0.903 ± 0.296 2.32E−3 0.291 ± 0.340 0.392   
 Serum creatinine 0.000 ± 0.002 0.823   −0.005 ± 0.002 0.022 −0.008 ± 0.002 2.29E−4 0.003 ± 0.002 0.241   
 Estimated glomerular filtration rate 0.038 ± 0.230 0.870   0.426 ± 0.236 0.071   −0.226 ± 0.246 0.359   
 Urinary albumin excretion 0.002 ± 0.006 0.763   0.015 ± 0.006 0.020 0.011 ± 0.006 0.085 −0.004 ± 0.007 0.532   
 Cystatin-C 0.003 ± 0.003 0.227   −0.006 ± 0.003 0.017 −0.009 ± 0.003 5.97E−4 −0.001 ± 0.003 0.590   
Metabolism 
 Fasting glucose 0.12 ± 0.25 0.621   −0.64 ± 0.26 0.011 −0.81 ± 0.25 1.34E−3 0.28 ± 0.27 0.295   
 Fasting insulin 0.004 ± 0.009 0.653   −0.014 ± 0.009 0.154   0.002 ± 0.010 0.871   
 Homeostasis model assessment-estimated insulin resistance 0.009 ± 0.010 0.369   −0.020 ± 0.011 0.067   0.005 ± 0.011 0.657   
 Waist circumference 0.128 ± 0.175 0.466   −0.015 ± 0.179 0.932   −0.205 ± 0.188 0.274   
 Body mass index 0.083 ± 0.063 0.188   0.069 ± 0.064 0.284   −0.138 ± 0.068 0.041 −0.135 ± 0.063 0.034 
Inflammation 
 High-sensitive C-reactive protein 0.009 ± 0.018 0.606   0.059 ± 0.018 1.18E−3 0.051 ± 0.018 5.09E−3 0.018 ± 0.019 0.351   

 
 SLC16A9 (rs12356193) SLC22A11 (rs17300741)     
 Model 1 Model 2 Model 1 Model 2     
 Beta ± SE P-value Beta ± SE P-value Beta ± SE P-value Beta ± SE P-value     

 
Protein intake 0.001 ± 0.011 0.952   −0.002 ± 0.008 0.762       
Alcohol intake −0.004 ± 0.004 0.306   −0.002 ± 0.003 0.545       
Blood pressure 
 Systolic −0.250 ± 0.380 0.511   0.402 ± 0.281 0.153       
 Diastolic −0.225 ± 0.186 0.226   0.275 ± 0.138 0.046       
Lipids     
 Total cholesterol −0.040 ± 0.843 0.962   −0.350 ± 0.623 0.574       
 LDL cholesterol 0.215 ± 0.800 0.788   −0.465 ± 0.592 0.432       
 HDL cholesterol 0.071 ± 0.292 0.808   0.157 ± 0.215 0.465       
 Fasting triglycerides −0.012 ± 0.011 0.270   −0.004 ± 0.008 0.603       
Renal function parameters     
 Absolute uric acid excretion −0.005 ± 0.013 0.673   0.002 ± 0.009 0.809       
 Fractional uric acid excretion −0.653 ± 0.421 0.121   −0.488 ± 0.311 0.117       
 Serum creatinine 0.002 ± 0.002 0.364   −0.000 ± 0.002 0.912       
 Estimated glomerular filtration rate −0.204 ± 0.304 0.502   0.129 ± 0.245 0.568       
 Urinary albumin excretion 0.000 ± 0.008 0.970   −0.010 ± 0.006 0.117       
 Cystatin-C 0.002 ± 0.003 0.489   −0.002 ± 0.002 0.511       
Metabolism     
 Fasting glucose −0.423 ± 0.328 0.198   −0.254 ± 0.242 0.293       
 Fasting insulin −0.004 ± 0.012 0.716   −0.002 ± 0.009 0.782       
 Homeostasis model assessment-estimated insulin resistance −0.008 ± 0.0139 0.550   −0.009 ± 0.010 0.367       
 Waist circumference −0.241 ± 0.232 0.299   −0.571 ± 0.171 8.89E−4 −0.745 ± 0.162 4.44E−6     
 Body mass index −0.100 ± 0.084 0.234   −0.140 ± 0.062 0.024 −0.205 ± 0.058 4.20E−4     
Inflammation     
 High-sensitive C-reactive protein 0.005 ± 0.024 0.823   −0.010 ± 0.018 0.584       
 PDZK1 (rs12129861)
 
GCKR (rs780094)
 
LRRC16A (rs742132)
 
 Model 1
 
Model 2
 
Model 1
 
Model 2
 
Model 1
 
Model 2
 
 Beta ± SE P-value Beta ± SE P-value Beta ± SE P-value Beta ± SE P-value Beta ± SE P-value Beta ± SE P-value 
Protein intake 0.007 ± 0.008 0.372   0.006 ± 0.008 0.452   −0.003 ± 0.009 0.762   
Alcohol intake 0.003 ± 0.003 0.391   −0.003 ± 0.003 0.412   0.004 ± 0.003 0.284   
Blood pressure 
 Systolic 0.852 ± 0.286 0.0029 1.040 ± 0.282 2.25E−4 0.010 ± 0.293 0.972   −0.067 ± 0.307 0.828   
 Diastolic 0.209 ± 0.140 0.135   −0.029 ± 0.143 0.840   0.090 ± 0.150 0.550   
Lipids 
 Total cholesterol 0.154 ± 0.634 0.808   2.92 ± 0.651 7.52E−6 2.477 ± 0.647 1.30E−4 0.664 ± 0.681 0.329   
 LDL cholesterol 0.275 ± 0.600 0.647   1.648 ± 0.618 0.0076 1.395 ± 0.617 0.024 0.606 ± 0.647 0.349   
 HDL cholesterol −0.058 ± 0.219 0.791   0.0529 ± 0.225 0.814   0.011 ± 0.235 0.962   
 Fasting triglycerides 0.006 ± 0.008 0.447   0.049 ± 0.008 2.65E−9 0.039 ± 0.008 8.06E−7 0.000 ± 0.009 0.976   
Renal function parameters 
 Absolute uric acid excretion −0.001 ± 0.010 0.933   −0.002 ± 0.010 0.846   0.001 ± 0.010 0.943   
 Fractional uric acid excretion 0.116 ± 0.318 0.715   −1.351 ± 0.326 3.36E−5 −0.903 ± 0.296 2.32E−3 0.291 ± 0.340 0.392   
 Serum creatinine 0.000 ± 0.002 0.823   −0.005 ± 0.002 0.022 −0.008 ± 0.002 2.29E−4 0.003 ± 0.002 0.241   
 Estimated glomerular filtration rate 0.038 ± 0.230 0.870   0.426 ± 0.236 0.071   −0.226 ± 0.246 0.359   
 Urinary albumin excretion 0.002 ± 0.006 0.763   0.015 ± 0.006 0.020 0.011 ± 0.006 0.085 −0.004 ± 0.007 0.532   
 Cystatin-C 0.003 ± 0.003 0.227   −0.006 ± 0.003 0.017 −0.009 ± 0.003 5.97E−4 −0.001 ± 0.003 0.590   
Metabolism 
 Fasting glucose 0.12 ± 0.25 0.621   −0.64 ± 0.26 0.011 −0.81 ± 0.25 1.34E−3 0.28 ± 0.27 0.295   
 Fasting insulin 0.004 ± 0.009 0.653   −0.014 ± 0.009 0.154   0.002 ± 0.010 0.871   
 Homeostasis model assessment-estimated insulin resistance 0.009 ± 0.010 0.369   −0.020 ± 0.011 0.067   0.005 ± 0.011 0.657   
 Waist circumference 0.128 ± 0.175 0.466   −0.015 ± 0.179 0.932   −0.205 ± 0.188 0.274   
 Body mass index 0.083 ± 0.063 0.188   0.069 ± 0.064 0.284   −0.138 ± 0.068 0.041 −0.135 ± 0.063 0.034 
Inflammation 
 High-sensitive C-reactive protein 0.009 ± 0.018 0.606   0.059 ± 0.018 1.18E−3 0.051 ± 0.018 5.09E−3 0.018 ± 0.019 0.351   

 
 SLC16A9 (rs12356193) SLC22A11 (rs17300741)     
 Model 1 Model 2 Model 1 Model 2     
 Beta ± SE P-value Beta ± SE P-value Beta ± SE P-value Beta ± SE P-value     

 
Protein intake 0.001 ± 0.011 0.952   −0.002 ± 0.008 0.762       
Alcohol intake −0.004 ± 0.004 0.306   −0.002 ± 0.003 0.545       
Blood pressure 
 Systolic −0.250 ± 0.380 0.511   0.402 ± 0.281 0.153       
 Diastolic −0.225 ± 0.186 0.226   0.275 ± 0.138 0.046       
Lipids     
 Total cholesterol −0.040 ± 0.843 0.962   −0.350 ± 0.623 0.574       
 LDL cholesterol 0.215 ± 0.800 0.788   −0.465 ± 0.592 0.432       
 HDL cholesterol 0.071 ± 0.292 0.808   0.157 ± 0.215 0.465       
 Fasting triglycerides −0.012 ± 0.011 0.270   −0.004 ± 0.008 0.603       
Renal function parameters     
 Absolute uric acid excretion −0.005 ± 0.013 0.673   0.002 ± 0.009 0.809       
 Fractional uric acid excretion −0.653 ± 0.421 0.121   −0.488 ± 0.311 0.117       
 Serum creatinine 0.002 ± 0.002 0.364   −0.000 ± 0.002 0.912       
 Estimated glomerular filtration rate −0.204 ± 0.304 0.502   0.129 ± 0.245 0.568       
 Urinary albumin excretion 0.000 ± 0.008 0.970   −0.010 ± 0.006 0.117       
 Cystatin-C 0.002 ± 0.003 0.489   −0.002 ± 0.002 0.511       
Metabolism     
 Fasting glucose −0.423 ± 0.328 0.198   −0.254 ± 0.242 0.293       
 Fasting insulin −0.004 ± 0.012 0.716   −0.002 ± 0.009 0.782       
 Homeostasis model assessment-estimated insulin resistance −0.008 ± 0.0139 0.550   −0.009 ± 0.010 0.367       
 Waist circumference −0.241 ± 0.232 0.299   −0.571 ± 0.171 8.89E−4 −0.745 ± 0.162 4.44E−6     
 Body mass index −0.100 ± 0.084 0.234   −0.140 ± 0.062 0.024 −0.205 ± 0.058 4.20E−4     
Inflammation     
 High-sensitive C-reactive protein 0.005 ± 0.024 0.823   −0.010 ± 0.018 0.584       

Shown are the effect size (Beta) ± standard error (SE) and P-value for the association of each locus with the variables adjusted for age and gender (model 1) and for age, gender and serum uric acid (UA) levels (model 2).

DISCUSSION

A recent meta-analysis of genome-wide association studies including 28 141 subjects has identified five novel loci associated with UA levels. We are the first to attempt to independently replicate these loci and were able to replicate four of them in our cohort of 7795 individuals. We successfully replicated the PDZK1, GCKR, SLC16A9 and SLC22A11 loci, but failed to replicate the LRCC16A locus (rs742132). Although we might have experienced false-negative results, the alternative, that the initial finding was false-positive, can also be argued. Assuming the effect size reported by Kolz et al., our study had 94% power to replicate the LRCC16A association with P ≤ 0.05, and our power remains >90% if we integrate over the range of effect sizes implied by the estimate and standard error reported by Kolz et al. (6) (for details see Supplementary Material, Table S1). Furthermore, we note that of the 14 studies participating in the meta-analysis of genome-wide studies, the effect of the LRCC16A locus was mainly driven by the imputed SNP of a single study (TwinsUK; P = 1.02 × 10−4), excluding only that particular study would have reduced the overall P-value several orders of magnitude (approximately to 1 × 10−6), i.e. well below the threshold of genome-wide significance. Further replication efforts to determine whether this association is genuine are clearly warranted.

As has been well established, UA levels are closely associated with components of the metabolic syndrome and other cardiovascular risk factors representing a number of physiological pathways (2–5). In an attempt to discriminate the potential contribution of these various factors and gain further insights into the biological mechanisms underlying the association of the novel loci with UA concentrations we constructed multivariable models to adjust for these factors, one by one and per group of closely related factors. For the GCKR locus, we observed attenuating effects of both triglyceride levels and fractional UA excretion. Previously, the Diabetes Genetics Initiative genome-wide association study for type 2 diabetes has also found the GCKR (rs780094) T-allele to be strongly associated with serum triglycerides levels and also showed a trend toward association with decreased plasma glucose levels, higher insulin sensitivity and lower risk of type 2 diabetes (8). Plasma CRP levels also have been previously associated with the GCKR loci (9). Fine-mapping approaches revealed a common missense GCKR variant (rs1260326, Pro446Leu, r2 0.93 with rs780094) as the strongest association signal in the region (7), which is the same SNP that was independently identified by Kolz et al. (6) and replicated here. However, robust inference about causality cannot be made solely from the current observational data. For example, Figure 2 illustrates that several different causal diagrams are compatible with our observation that the association between the GCKR SNP and serum UA concentration is attenuated by triglyceride levels. However, the set of plausible causal diagrams can potentially be winnowed by looking for consistent patterns across multiple SNPs and by knowledge of the biological processes involved. Direct mediating effects (Fig. 2A and B) are argued against by the small overlap between SNPs associated with triglyceride levels (10,11) and SNPs associated with serum UA concentrations (although a lack of power cannot be excluded) (6). Therefore, it seems most plausible that the GCKR SNP is unique in affecting both serum UA concentrations and triglyceride levels via some common, unmeasured mediator (Fig. 2C). This can be further supported by the observed association between the GCKR SNP and triglyceride levels, when not adjusting for serum UA concentrations (P = 2.65 × 10−9 in PREVEND, Table 4). This also fits with current knowledge on the biological processes involved. GCKR is produced in hepatocytes, binds and moves glucokinase, whereby it controls both activity and intracellular location of the enzyme (12,13). Glucokinase in turn mediates phosphorylation of glucose to glucose-6-phosphate, which is both a precursor for liver glycogen synthesis and a precursor for de novo purine synthesis (14). Glycogen storage disease type 1 (also known as von Gierke disease) is caused by a deficiency in the activity of the enzyme glucose-6-phosphatase and is not only characterized by hypoglycaemia and glycogen storage, but also by hypertriglyceridaemia and hyperuricaemia (14). The strength of the association of the PDZK1 locus with UA levels was not affected by any of the tested potential mediators, suggesting greater likelihood of a direct effect. PDZK1 codes for the PDZ domain containing 1, a scaffolding protein reported to interact with proteins thought to relate to UA handling (URAT1, NPT1) (15,16). Interestingly, the T-allele causes lower UA levels, but was also associated with higher systolic blood pressure. Although this finding requires replication, it becomes interesting when PDZK1 is also related to inducible nitric oxide synthase activity and several ion exchange proteins. (17).

Figure 2.

Several different causal diagrams are compatible with our observation that the association between the GCKR SNP and serum UA concentration is attenuated by triglyceride levels. Direct mediating effects (A and B) are argued against by the small overlap between SNPs associated with triglyceride levels. Therefore, it seems most plausible that the GCKR SNP is unique in affecting both serum UA concentrations and triglyceride levels via some common, unmeasured mediator (C).

Figure 2.

Several different causal diagrams are compatible with our observation that the association between the GCKR SNP and serum UA concentration is attenuated by triglyceride levels. Direct mediating effects (A and B) are argued against by the small overlap between SNPs associated with triglyceride levels. Therefore, it seems most plausible that the GCKR SNP is unique in affecting both serum UA concentrations and triglyceride levels via some common, unmeasured mediator (C).

This study has several strengths, including the large number of subjects investigated and the detailed information on a wide range of traditional and novel cardiovascular risk factors associated with UA levels and renal UA excretion. However, a lack of detailed information on dietary factors that affect UA levels is a limitation. We have included alcohol and total protein intake. However, not all proteins contain the purine and can contribute to UA. Other dietary factors, including fructose and sodium intake, also have not been taken into account while they may be of importance (18). As an observational study, we cannot robustly make causal inference. When it is known a priori that a genetic variant causally affects only a single phenotype, Mendelian randomization can be applied, but the required causality assumption cannot be tested using only observational data (19). Observing that a genetic association is attenuated when another factor is included in the model can be consistent with a wide range of possible mechanisms.

In summary, we present the first data replicating four of the five recently identified loci for UA concentrations. We also present the first attempt to quantify the relative importance of potential underlying mechanism. Our data suggest a central role of the GCKR locus in UA metabolism and multiple components of the metabolic syndrome.

MATERIALS AND METHODS

An expanded description of the methods is provided in Supplementary Material, Supplementary Methods.

Study population

We studied 7795 Caucasian subjects of whom DNA was available and had data available on plasma UA participating in the PREVEND study, an ongoing prospective study investigating the natural course of increased levels of urinary albumin excretion and its relation to renal and cardiovascular disease in a large cohort drawn from the general population (20,21). The study was approved by the medical Ethics Committee of the University Medical Center Groningen and was conducted in accordance with the guidelines of the Declaration of Helsinki.

Uric acid

UA was measured in plasma and urine with the uricase PAP method as described previously (MEGA, Merck, Darmstadt, Germany). (22).

Genotyping

All genotyping was performed by KBiosciences (Mapple Park, Herts, UK; http://kbiosicence.co.uk). SNPs were genotyped using the KASPar chemistry, which is a competitive allele specific PCR SNP genotyping system using FRET quencher cassette oligos (http://www.kbioscience.co.uk/genotyping/genotyping_chemistry.htm). Blind duplicates, plate-identifying blanks and Hardy–Weinberg equilibrium tests were used as quality control test.

Statistical analysis

Statistical analysis was performed with STATA version 10.1 for Windows (StataCorp LP, College Station, TX, USA). All probability values were two tailed. Because of deviation from normal distribution, protein intake, fasting triglycerides, insulin, HOMA, absolute UA excretion and CRP were log transformed before analyses. To limit the undue influence of outliers in the regression analysis, for each trait we excluded the bottom and top 0.5% of the trait-level distribution in each study sample. All association analyses were age and gender adjusted. Genotype–phenotype associations were carried out under an additive model. To examine the extent to which various potential modifiers/confounders contributed to the association of serum UA levels, we initially considered each factor separately in a model that adjusted for age and gender. We considered the magnitude of change in Beta with and without adjustment for each risk factor a method previously described by Cook (23). A larger change in Beta toward the null implies a larger mediating effect of that factor on the SNP association with serum UA. Arbitrarlly we considered a >50% reduction in Beta meaningful. We also considered adjustment for groups of mediators of one system and considered all variables together.

Power calculations

Assuming the effect size reported by Kolz et al., our study has a power ranging from 93% for rs780094 up to 99% for rs12129861 (Supplementary Material, Table S1). Our power remains between 91 and 97% if we integrate over the range of effect sizes implied by the estimate and standard error reported by Kolz et al. (6) (Supplementary Material, Table S1).

Accession numbers

The GeneID (http://www.ncbi.nlm.nih.gov/sites/entrez) accession numbers for genes mentioned in this article are 5174 (PDZK1), 2646 (GCKR), 55604 (LRRC16A), 220963 (SLC16A9) and 55867 (SLC22A11).

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG online.

FUNDING

This work was further supported by the Innovational Research Incentives Scheme program of the Netherlands Organization for Scientific Research [NWO VENI, grant number 916.76.170 to P.v.d.H.], The Netherlands. P.v.d.H. is a research fellow of the Netherlands Heart Foundation [grant numbers 2006T003] and the Interuniversitair Cardiologisch Instituut Nederland (ICIN).

ACKNOWLEDGEMENTS

We thank the Dutch Kidney Foundation for their financial support for making the PREVEND study possible (E033). We also thank Dade Behring (Marbur, Germany) for supplying equipment (Behring Nephelometer II) and reagents for nephelometric measurements of hs-CRP, cystatin-c, and urinary albumin concentration.

Conflict of Interest statement. None declared.

REFERENCES

1
Feig
D.I.
Kang
D.H.
Johnson
R.J.
Uric acid and cardiovascular risk
N. Engl. J. Med.
 , 
2008
, vol. 
359
 (pg. 
1811
-
1821
)
2
Hayden
M.R.
Tyagi
S.C.
Uric acid: a new look at an old risk marker for cardiovascular disease, metabolic syndrome, and type 2 diabetes mellitus: the urate redox shuttle
Nutr. Metab (Lond)
 , 
2004
, vol. 
1
 pg. 
10
 
3
Nagahama
K.
Iseki
K.
Inoue
T.
Touma
T.
Ikemiya
Y.
Takishita
S.
Hyperuricemia and cardiovascular risk factor clustering in a screened cohort in Okinawa, Japan
Hypertens. Res.
 , 
2004
, vol. 
27
 (pg. 
227
-
233
)
4
Pacifico
L.
Cantisani
V.
Anania
C.
Bonaiuto
E.
Martino
F.
Pascone
R.
Chiesa
C.
Serum uric acid and its association with metabolic syndrome and carotid atherosclerosis in obese children
Eur. J. Endocrinol.
 , 
2009
, vol. 
160
 (pg. 
45
-
52
)
5
Schmidt
M.I.
Watson
R.L.
Duncan
B.B.
Metcalf
P.
Brancati
F.L.
Sharrett
A.R.
Davis
C.E.
Heiss
G.
Atherosclerosis Risk in Communities Study Investigators
Clustering of dyslipidemia, hyperuricemia, diabetes, and hypertension and its association with fasting insulin and central and overall obesity in a general population
Metabolism
 , 
1996
, vol. 
45
 (pg. 
699
-
706
)
6
Kolz
M.
Johnson
T.
Sanna
S.
Teumer
A.
Vitart
V.
Perola
M.
Mangino
M.
Albrecht
E.
Wallace
C.
Farrall
M.
, et al.  . 
Meta-analysis of 28,141 individuals identifies common variants within five new loci that influence uric acid concentrations
PLoS Genet.
 , 
2009
, vol. 
5
 pg. 
e1000504
 
7
Orho-Melander
M.
Melander
O.
Guiducci
C.
Perez-Martinez
P.
Corella
D.
Roos
C.
Tewhey
R.
Rieder
M.J.
Hall
J.
Abecasis
G.
, et al.  . 
Common missense variant in the glucokinase regulatory protein gene is associated with increased plasma triglyceride and C-reactive protein but lower fasting glucose concentrations
Diabetes
 , 
2008
, vol. 
57
 (pg. 
3112
-
3121
)
8
Saxena
R.
Voight
B.F.
Lyssenko
V.
Burtt
N.P.
de Bakker
P.I.
Chen
H.
Roix
J.J.
Kathiresan
S.
Hirschhorn
J.N.
Daly
M.J.
, et al.  . 
Genome-wide association analysis identifies loci for type 2 diabetes and triglyceride levels
Science
 , 
2007
, vol. 
316
 (pg. 
1331
-
1336
)
9
Ridker
P.M.
Pare
G.
Parker
A.
Zee
R.Y.
Danik
J.S.
Buring
J.E.
Kwiatkowski
D.
Cook
N.R.
Miletich
J.P.
Chasman
D.I.
Loci related to metabolic-syndrome pathways including LEPR,HNF1A, IL6R, and GCKR associate with plasma C-reactive protein: the Women's Genome Health Study
Am. J. Hum. Genet.
 , 
2008
, vol. 
82
 (pg. 
1185
-
1192
)
10
Aulchenko
Y.S.
Ripatti
S.
Lindqvist
I.
Boomsma
D.
Heid
I.M.
Pramstaller
P.P.
Penninx
B.W.
Janssens
A.C.
Wilson
J.F.
Spector
T.
, et al.  . 
Loci influencing lipid levels and coronary heart disease risk in 16 European population cohorts
Nat. Genet.
 , 
2009
, vol. 
41
 (pg. 
47
-
55
)
11
Kathiresan
S.
Willer
C.J.
Peloso
G.M.
Demissie
S.
Musunuru
K.
Schadt
E.E.
Kaplan
L.
Bennett
D.
Li
Y.
Tanaka
T.
, et al.  . 
Common variants at 30 loci contribute to polygenic dyslipidemia
Nat. Genet.
 , 
2009
, vol. 
41
 (pg. 
56
-
65
)
12
Van
S.E.
Short-term regulation of glucokinase
Diabetologia
 , 
1994
, vol. 
37
 
Suppl 2
(pg. 
S43
-
S47
)
13
de la Iglesia
N.
Veiga-da-Cunha
M.
Van Schaftingen
E.
Guinovart
J.J.
Ferrer
J.C.
Glucokinase regulatory protein is essential for the proper subcellular localisation of liver glucokinase
FEBS Lett.
 , 
1999
, vol. 
456
 (pg. 
332
-
338
)
14
Yang
C.J.
Mansfield
B.C.
Molecular genetics of type 1 glycogen storage diseases
Trends Endocrinol. Metab.
 , 
1999
, vol. 
10
 (pg. 
104
-
113
)
15
Anzai
N.
Miyazaki
H.
Noshiro
R.
Khamdang
S.
Chairoungdua
A.
Shin
H.J.
Enomoto
A.
Sakamoto
S.
Hirata
T.
Tomita
K.
, et al.  . 
The multivalent PDZ domain-containing protein PDZK1 regulates transport activity of renal urate-anion exchanger URAT1 via its C terminus
J. Biol. Chem.
 , 
2004
, vol. 
279
 (pg. 
45942
-
45950
)
16
Anzai
N.
Kanai
Y.
Endou
H.
New insights into renal transport of urate
Curr. Opin. Rheumatol.
 , 
2007
, vol. 
19
 (pg. 
151
-
157
)
17
Navarro-Lerida
I.
Martinez-Moreno
M.
Ventoso
I.
Varez-Barrientos
A.
Rodriguez-Crespo
I.
Binding of CAP70 to inducible nitric oxide synthase and implications for the vectorial release of nitric oxide in polarized cells
Mol. Biol. Cell
 , 
2007
, vol. 
18
 (pg. 
2768
-
2777
)
18
Johnson
R.J.
Perez-Pozo
S.E.
Sautin
Y.Y.
Manitius
J.
Sanchez-Lozada
L.G.
Feig
D.I.
Shafiu
M.
Segal
M.
Glassock
R.J.
Shimada
M.
, et al.  . 
Hypothesis: could excessive fructose intake and uric acid cause type 2 diabetes?
Endocr. Rev.
 , 
2009
, vol. 
30
 (pg. 
96
-
116
)
19
Sheehan
N.A.
Didelez
V.
Burton
P.R.
Tobin
M.D.
Mendelian randomisation and causal inference in observational epidemiology
PLoS Med.
 , 
2008
, vol. 
5
 pg. 
e177
 
20
Hillege
H.L.
Fidler
V.
Diercks
G.F.
van Gilst
W.H.
de
Z.D.
van Veldhuisen
D.J.
Gans
R.O.
Janssen
W.M.
Grobbee
D.E.
de Jong
P.E.
Urinary albumin excretion predicts cardiovascular and noncardiovascular mortality in general population
Circulation
 , 
2002
, vol. 
106
 (pg. 
1777
-
1782
)
21
Pinto-Sietsma
S.J.
Janssen
W.M.
Hillege
H.L.
Navis
G.
de
Z.D.
de Jong
P.E.
Urinary albumin excretion is associated with renal functional abnormalities in a nondiabetic population
J. Am. Soc. Nephrol.
 , 
2000
, vol. 
11
 (pg. 
1882
-
1888
)
22
Mostert
J.P.
Ramsaransing
G.S.
Heersema
D.J.
Heerings
M.
Wilczak
N.
De
K.J.
Serum uric acid levels and leukocyte nitric oxide production in multiple sclerosis patients outside relapses
J. Neurol. Sci.
 , 
2005
, vol. 
231
 (pg. 
41
-
44
)
23
Mora
S.
Cook
N.
Buring
J.E.
Ridker
P.M.
Lee
I.M.
Physical activity and reduced risk of cardiovascular events: potential mediating mechanisms
Circulation
 , 
2007
, vol. 
116
 (pg. 
2110
-
2118
)