Gout results from deposition of monosodium urate (MSU) crystals. Elevated serum urate concentrations (hyperuricemia) are not sufficient for the development of disease. Genome-wide association studies (GWAS) have identified 28 loci controlling serum urate levels. The largest genetic effects are seen in genes involved in the renal excretion of uric acid, with others being involved in glycolysis. Whereas much is understood about the genetic control of serum urate levels, little is known about the genetic control of inflammatory responses to MSU crystals. Extending knowledge in this area depends on recruitment of large, clinically ascertained gout sample sets suitable for GWAS.
Key points
- •
Genome-wide association studies for serum urate have identified 28 loci influencing serum urate levels.
- •
The largest genetic effects on serum urate are within genes encoding transporters that excrete uric acid in the kidney and gut.
- •
Other genetic effects are within glycolysis genes.
- •
There are interactions between genes, and environmental influences on serum urate (diuretics, alcohol, sugar-sweetened beverages).
- •
Genome-wide association studies are required in gout using well-phenotyped cases to identify loci controlling progression from hyperuricemia to inflammatory gout.
Introduction
The central feature of gout is deposition of inflammatory monosodium urate (MSU) monohydrate microcrystals, which can lead to acute inflammatory arthritis, tendonitis, cartilage damage, and bone remodeling. Several checkpoints exist in the pathogenesis of gout (reviewed in Refs. ). Central to the development of gout is elevated tissue concentrations of urate, which in some individuals lead to formation of MSU crystals. Elevated serum urate levels (hyperuricemia) occur as a result of increased production of hepatic urate through the purine synthesis de novo and salvage pathways; however, renal underexcretion of uric acid is a dominant contributor, with reduced fractional excretion of uric acid in hyperuricemia and gout. Once formed, MSU crystals may induce an acute inflammatory response leading to acute gouty arthritis and/or a chronic granulomatous response with formation of tophi. Although hyperuricemia is present in virtually all people with gout, this biochemical abnormality is not sufficient for the development of clinically apparent joint disease, as most people with hyperuricemia do not develop gout.
Monogenic inborn errors of purine metabolism such as hypoxanthine-guanine phosphoribosyltransferase deficiency (Lesch-Nyhan syndrome) and 5-phosphoribosyl-1-pyrophosphate synthetase superactivity lead to rare pediatric syndromes of hyperuricemia, associated with neurodevelopmental disorders, early-onset gout, and kidney stones. In addition, familial juvenile hyperuricemic nephropathy is an autosomal dominant disorder of renal uric acid underexcretion caused by mutations in the uromodulin gene that leads to severe underexcretion-type hyperuricemia, early-onset gout, and chronic kidney disease. These rare monogenic disorders provide important insights into physiologic purine metabolism and uric acid excretion mechanisms, but do not account for the vast majority of hyperuricemia or gout observed in the general population. Renal uric handling of acid and hyperuricemia have a large heritable component (87% for fractional excretion of uric acid, 60% for serum urate). Consistent with these observations, genome-wide association studies (GWAS) have revealed that a polygenic component of common inherited variants contributes to the development of gout in the general population with, excepting the PRPSAP1 locus, little overlap with monogenic syndromes. This review focuses on recent insights into these common genetic variants that contribute to the development of gout, and their potential interaction with environmental risk factors.
Introduction
The central feature of gout is deposition of inflammatory monosodium urate (MSU) monohydrate microcrystals, which can lead to acute inflammatory arthritis, tendonitis, cartilage damage, and bone remodeling. Several checkpoints exist in the pathogenesis of gout (reviewed in Refs. ). Central to the development of gout is elevated tissue concentrations of urate, which in some individuals lead to formation of MSU crystals. Elevated serum urate levels (hyperuricemia) occur as a result of increased production of hepatic urate through the purine synthesis de novo and salvage pathways; however, renal underexcretion of uric acid is a dominant contributor, with reduced fractional excretion of uric acid in hyperuricemia and gout. Once formed, MSU crystals may induce an acute inflammatory response leading to acute gouty arthritis and/or a chronic granulomatous response with formation of tophi. Although hyperuricemia is present in virtually all people with gout, this biochemical abnormality is not sufficient for the development of clinically apparent joint disease, as most people with hyperuricemia do not develop gout.
Monogenic inborn errors of purine metabolism such as hypoxanthine-guanine phosphoribosyltransferase deficiency (Lesch-Nyhan syndrome) and 5-phosphoribosyl-1-pyrophosphate synthetase superactivity lead to rare pediatric syndromes of hyperuricemia, associated with neurodevelopmental disorders, early-onset gout, and kidney stones. In addition, familial juvenile hyperuricemic nephropathy is an autosomal dominant disorder of renal uric acid underexcretion caused by mutations in the uromodulin gene that leads to severe underexcretion-type hyperuricemia, early-onset gout, and chronic kidney disease. These rare monogenic disorders provide important insights into physiologic purine metabolism and uric acid excretion mechanisms, but do not account for the vast majority of hyperuricemia or gout observed in the general population. Renal uric handling of acid and hyperuricemia have a large heritable component (87% for fractional excretion of uric acid, 60% for serum urate). Consistent with these observations, genome-wide association studies (GWAS) have revealed that a polygenic component of common inherited variants contributes to the development of gout in the general population with, excepting the PRPSAP1 locus, little overlap with monogenic syndromes. This review focuses on recent insights into these common genetic variants that contribute to the development of gout, and their potential interaction with environmental risk factors.
Genome-wide association study findings for serum urate
Over the past 8 years, GWAS and subsequent meta-analyses have led to a considerable expansion in the knowledge of common genetic loci that are associated with hyperuricemia and gout in Europeans. Two meta-analyses, each involving more than 28,000 participants, found genome-wide genetic loci that are reproducibly associated with serum urate levels or gout ( Table 1 ). In 2009, a meta-analysis by Kolz and colleagues reported associations between 9 common genetic variant loci and serum urate concentrations: SLC2A9 , ABCG2 , SLC22A12 , SLC17A1 , SLC22A11 , SLC16A9 , GCKR , LRRC16A , and near PDZK1 . In 2010, a meta-analysis of the CHARGE consortium by Yang and colleagues reconfirmed 6 of these loci ( SLC2A9 , ABCG2 , SLC17A1 , SLC22A11 , GCKR , and PDZK1 ) and additionally identified the R3HDM2-INHBC region and RREB1 loci with genome-wide significance. The genetic urate risk score was strongly associated with the risk of gout. An Icelandic GWAS for serum urate, using whole genome sequence data in 15,506 individuals, identified 4 loci with a genome-wide level of significance at SLC2A9 , ABCG2 , PDZK1 , and ALDH16A1 . The variant underlying the ALDH16A1 association appears to be a previously unreported Icelandic-specific genetic variation present at a frequency of 1.8%, which encodes a proline to arginine amino acid change in the ALDH16A1 protein. It was estimated that the variant explained 0.5% of variance in serum urate in the Icelandic population.
Kolz et al, 2009 | Yang et al, 2010 | Tin et al, 2011 | Okada et al, 2012 | Köttgen et al, 2013 | |
---|---|---|---|---|---|
Sample size | 28,141 | 28,283 | 14,706 | 71,149 | >140,000 (primary data) |
Population | European | European | African American | East Asian | European (primary data), African American, Indian, Japanese |
Loci associated with serum urate with genome-wide significance |
|
|
|
|
|
Association with gout | Not reported | The genetic urate risk score was strongly and linearly associated with serum urate (multivariable adjusted P <4.5 × 10 −308 ) | Directionally consistent effects on serum urate and gout were observed ( P -binomial <.0001) | Not reported | Correlation between the effect on urate and the odds of gout for the replicated loci (Pearson correlation = 0.93); 17 of the replicated serum urate–associated SNPs reached statistical significance with gout ( P <.05) |
In other ancestral groups, a meta-analysis and a GWAS have demonstrated that 10 of the 11 loci that have been shown to influence serum urate levels in individuals with European ancestry were also significantly associated with serum urate or gout in African American sample sets. The GWAS in African Americans also identified a novel locus influencing serum urate near the SLC2A12 gene on chromosome 6. This gene is a good candidate, as it is a glucose transporter and a member of the same family as SLC2A9 , which has a very strong influence on serum urate across populations. Another GWAS among East Asians by Okada and colleagues in 2012 showed genome-wide significance of serum urate levels with SLC2A9 , ABCG2 , SLC22A12 , and MAF ; all of these loci overlap with those identified in Europeans (see below). A previous study in Japanese had identified SLC2A9 , ABCG2 , SLC22A12 , and LRP2 as being associated with serum urate at a genome-wide level of significance. LRP2 (lipoprotein receptor-related protein 2) has not been associated with urate in Europeans.
Most recently, by combining data from more than 140,000 individuals of European ancestry within the Global Urate Genetics Consortium (GUGC), Köttgen and colleagues have identified and replicated 28 genome-wide significant loci for serum urate. These loci included the 10 previously identified regions (as described above) and 18 new regions (in or near TRIM46 , INHBB , SFMBT1 , TMEM171 , VEGFA , BAZ1B , PRKAG2 , STC1 , HNF4G , A1CF , ATXN2 , UBE2Q2 , IGF1R , NFAT5 , MAF , HLF , ACVR1B-ACVRL1 , and B3GNT4 ) (see Table 1 ). Furthermore, the meta-analysis also studied 8340 individuals of Indian ancestry, 5820 African Americans, and 15,286 Japanese, and found that the serum urate effects were direction-consistent with similar magnitude for most single-nucleotide polymorphisms (SNPs), although allele frequencies at the index SNPs varied considerably across the groups.
In all gout samples combined (3151 cases and 68,350 controls) in the study by Köttgen and colleagues, 17 of 26 of the replicated urate concentration–associated SNPs showed nominal association with gout ( P <.05; see Table 1 ). Of note, these gout cases included 1036 cases of incident gout (meeting the American College of Rheumatology classification criteria ) over a period of up to 22 years in prospective cohort studies. The serum urate effects of these loci showed a positive linear correlation with the log odds of gout (Pearson correlation = 0.93), and genetic urate risk scores (created based on the number of risk alleles of serum urate–associated genes) were significantly associated with increased odds of prevalent gout (odds ratio [OR] = 1.11 per risk score unit increase) and incident gout over a period of up to 22 years (OR = 1.10, 95% confidence interval = 1.08–1.13). This study also evaluated the association with the fractional excretion of uric acid (FEUA, n = 6799) and found that SNPs at 10 replicated loci showed directionally consistent, significant association with FEUA (ie, SLC2A9 , GCKR , ABCG2 , RREB1 , SLC22A11 , NRXN2/SLC22A12 , UBE2Q2 , IGFIR , NFAT5 , and HLF ).
Notably in the same study, 2 novel regions of B3GNT4 and ACVR1B – ACVRL1 were discovered and replicated through a systematic functional association network analysis approach that incorporated previous knowledge on molecular interactions through which the gene products of implicated genes operate. Furthermore, the functional association network analysis also highlighted a specific subnetwork from the analysis around the inhibins-activins pathway. Finally, pathway analyses showed functional network associations with gene expression, cellular organization, carbohydrate metabolism, molecular transport, and endocrine system disorders (lowest P = 1 × 10 −28 ).
Most of the initially identified novel loci by GWAS studies and meta-analyses appear to encode proteins that are involved in the renal urate transport system (see Table 1 ). These findings appear to be consistent with reduced renal excretion of urate being the dominant cause of hyperuricemia and gout in most individuals. Nevertheless, new candidate genes for serum urate concentration, as reported by the meta-analysis by Köttgen and colleagues, highlight the importance of metabolic control of urate production (in addition to urate excretion) in the pathogenesis of gout.
Insights into mechanism
Apart from uric acid transporters, little is understood about how the other genes identified in the various GWAS regulate urate, largely because the causal genes have not been identified. Extensive correlation between genetic markers (linkage disequilibrium) encompassing more than 1 gene at each locus means that the gene cannot be inferred from the position of the associated marker. However, at the major urate and gout risk loci ( SLC2A9 and ABCG2 ), progress has been made into the functional role of the genetically associated variants (see later discussion). Data from other loci ( PDZK1 , SLC22A12 ) are specifically discussed here, illustrating an apparent exception to the expectation that serum urate–increasing genetic variants also increase the risk of gout. The GCKR gene and other glycolytic genes are also discussed. Finally, a locus ( PRPSAP1 ) in a pathway implicated in monogenic urate overproduction syndromes is mentioned.
SLC2A9
Genetic variation within introns 3 to 7 of SLC2A9 explains about 3.5% of variation in serum urate concentration in European Caucasians, an extremely large effect in the context of genetics of complex phenotypes. These variants are also strongly associated with gout, with effect sizes for the risk allele greater than OR 1.5 (eg, Refs. ). SLC2A9 is a voltage-dependent transporter believed to be responsible for reabsorption of uric acid into the circulation via the proximal renal tubule. The serum urate–associated intronic genetic variants are tightly correlated, which makes it very difficult to identify the causative variant. However, it is likely that the causative serum urate–raising variant increases the expression levels of an SLC2A9 isoform ( SLC2A9b [GLUT9S]) with a 28-residue portion missing from the N-terminus that is predominantly expressed on the apical (urine) membrane, presumably increasing reabsorption of uric acid from the filtered urine. A portion of the group of intronic variants associated with urate and gout in Europeans is essentially monomorphic in Asian populations (risk allele >99%), and has not been shown to influence urate levels or gout in Asian samples. However, other SNPs (eg, rs3775948 ) are strongly associated with serum urate in Asians and, importantly, these SNPs overlap with the associated intronic SNPs in Europeans. Thus the Asian data define a subset of associated SNPs and illustrate how transancestral mapping should allow fine-mapping of the etiologic variant at SLC2A9 .
ABCG2
The nonsynonymous Gln141Lys (Q141K) variant of ABCG2 explains about 0.5% of the variation in serum urate levels in Europeans, and is almost certainly the etiological variant, with the lysine allele associated with increased serum urate levels. Predictably the same allele increases the risk of gout in European, Chinese, Japanese, and New Zealand Pacific sample sets (OR ∼2) but, for unclear reasons, not in New Zealand Māori, despite an allele frequency similar to that of Europeans (reviewed in Ref. ).
ABCG2 , highly expressed in intestinal epithelial cells and also expressed in the apical membrane of the kidney proximal tubule, is an adenosine triphosphate (ATP)-dependent uric acid secretory molecule. The lysine allele encodes a molecule with approximately 50% reduced ability to transport uric acid, which results from instability in the nucleotide-binding domain and decreased protein expression. The defects in expression and function can be rescued by a histone deacetylase inhibitor approved by the Food and Drug Administration, revealing a possible new urate-lowering therapy for evaluation.
ABCG2 functions predominantly as a gut secretory uric acid transporter. With high expression of ABCG2 in extrarenal tissues (including the liver), this observation has led to the proposal that ABCG2 dysfunction encoded by the 141K allele and subsequent gut underexcretion of uric acid is a significant contributor to the overproduction of urate in the serum. This process results in the urate-increasing lysine allele, paradoxically, being associated with increased urinary excretion of uric acid as the renal uric acid excretion machinery physiologically adjusts to the lysine allele-mediated urate overproduction.
PDZK1 and SLC22A12
The PDZK1 gene encodes a PDZ domain–containing scaffold protein known to bind uric acid transporters and, presumably, arrange their cell-surface localization for optimal uric acid transport. PDZK1 is strongly associated with serum urate levels; however, the same variants had no effect on the risk of gout in a large well-powered European meta-analysis of cases nested within population-based cohorts (OR = 1.03). Why this is the case is unclear; however, it is notable that the same allele of the variant in PDZK1 that associates with increased serum urate is also associated with decreased blood pressure. The gout cases analyzed by Köttgen and colleagues would have included a significant proportion of cases secondary to diuretic medication (for hypertension). Thus, it is possible that any increased risk of primary gout (ie, owing to increased urate with no obvious secondary cause) mediated by the allele of PDZK1 via increased serum urate levels was negated by inclusion of cases with gout secondary to hypertension treatment, in which the other (serum urate–lowering) allele of PDZK1 was overrepresented. Of possible relevance for understanding the relationship between hyperuricemia and other metabolic conditions, PDZK1 also interacts with other molecules, including the high-density lipoprotein receptor known as scavenger receptor class B type 1, which is important in cholesterol metabolism.
The SLC22A12 gene encodes the canonical renal uric acid transporter URAT1. It is an example of a second locus with a strong genetic effect on serum urate, yet common variants are not associated with gout in Europeans, with inconsistent associations in other populations. For both PDZK1 and SLC22A12 the statistical power of the gout-association studies is unlikely to be an issue, as loci with a similar effect on serum urate are strongly associated with gout in the same samples (eg, GCKR , SLC22A11 , INHBC ).
Glycolytic Genes
The glucokinase regulatory protein gene ( GCKR ) highlights a serum urate–controlling pathway probably distinct from renal (and gut) excretion of uric acid, providing some clues to the etiologic links between gout and other associated metabolic conditions such as diabetes and dyslipidemia. It is strongly associated with serum urate in Europeans, and has been consistently associated with gout in Europeans and Chinese. Genetic variation in GCKR has also been associated with concentrations of triglyceride and fasting glucose, and the risk of type 2 diabetes. The association of GCKR with serum urate is weakened when triglyceride levels are accounted for and the same GCKR allele is also associated with triglyceride levels. The most plausible explanation for this observation is that GCKR affects both serum urate and triglyceride levels by a common unconfirmed mediator that could be glucose-6-phosphate. GCKR controls the hepatic production of glucose-6-phosphate, which is catabolized for triglyceride synthesis via glycolysis, pyruvate, and acetyl coenzyme A, while glucose-6-phosphate is also a precursor of de novo purine (uric acid) synthesis. Other loci encoding glycolysis genes are PKLR (encodes pyruvate kinase that catalyzes the final step of glycolysis, producing ATP and pyruvate), MLXIPL (encodes a glucose-responsive transcriptional factor that regulates PKLR expression), PRKAG2 (encodes the regulatory subunit γ2 of the adenosine monophosphate [AMP]-activated protein kinase, which senses cellular AMP:ATP ratio and activates glucose uptake and catabolism), NFAT5 (encodes a transcription factor that can influence glucose flux and the pentose phosphate pathway), and HNF4G (encodes a transcription factor responding to nutrient signals).
How the glucokinase regulatory protein and other glycolysis genes influence serum urate levels is unclear. It has been proposed that this could occur by altering the flux of glucose-6-phosphate through the pentose phosphate pathway that generates ribose-5-phosphate (a precursor of de novo purine synthesis and subsequent urate production) and/or altering the amount of lactate available physiologically, given that lactate influences excretion of renal uric acid, likely through a role as a cotransporter molecule for uric acid transporters. The observation that the GCKR and NFAT5 loci also associate with FEUA supports the latter possibility.
PRPSAP1
Of the 28 loci that influence serum urate levels, the phosphoribosylpyrophosphate synthetase–associated protein 1 gene ( PRPSAP1 ) has the most obvious role in modifying urate levels, as it is part of the pentose phosphate pathway and de novo synthesis of purines. Other genes in the pathway (phosphoribosylpyrophosphate synthetases and PRPS1 ) cause urate overproduction in purine-related Mendelian syndromes that include gout and neurodevelopmental abnormalities in their symptoms.