What lies behind serum urate concentration? Insights from genetic and genomic studies

Many factors, including genetic components and acquired factors such as obesity and alcohol consumption, influence serum uric acid (urate) concentrations. Since serum urate concentrations are determined by the balance between renal urate excretion and the volume of urate produced via purine metabolism, urate transporter genes as well as genes coding for enzymes involved in purine metabolism affect serum urate concentrations. URAT1 was the first transporter affecting serum urate concentrations to be identified. Using the characterization of this transporter as an indicator, several transporters have been shown to transport urate, allowing the construction of a synoptic renal urate transport model. Notable re-absorptive urate transporters are URAT1 at apical membranes and GLUT9 at basolateral membranes, while ABCG2, MRP4 (multidrug resistance protein 4) and NPT1 are secretive transporters at apical membranes. Recent genome-wide association studies have led to validation of the in vitro model constructed from each functional analysis of urate transporters, and identification of novel candidate genes related to urate metabolism and transport proteins, such as glucokinase regulatory protein (GKRP), PDZK1 and MCT9. However, the function and physiologic roles of several candidates, as well as the influence of acquired factors such as obesity, foods, or alcoholic beverages, remain unclear.


Introduction
Hyperuricemia induces or facilitates gout, kidney stones, metabolic syndrome, hypertension and renal and cardio vascular disease, while exerciseinduced acute renal failure is a significant complication of renal hypouricemia [13]. Although hyperuricemia has been more closely associated with gout and kidney stones, it has been recently recog nized to be independently associated with components of metabolic syndrome, insulin resistance, hypertension, dyslipidemia and obesity. Metabolic syndrome is a clustering of cardiovascular disease risk factors and its prevalence is increasing. Several mechanisms for the asso ciation between hyperuricemia and metabolic syndrome have been proposed; insulin resistance leads to renal under excretion of uric acid (urate); increased lactate in obesity accelerates renal urate reabsorption via urate transporter 1 (URAT1); fatty acid synthesis accelerates de novo purine synthesis via the pentose phosphate pathway, and so on [4]. Recent studies have shown that hyper uricemia independently causes atherosclerosis through uratemediated inflammation and endothelial dysfunction, in addition to metabolic syndrome [5,6]. Thus, monitoring of serum urate concentrations in patients with hyper uricemia, kidney stones, metabolic syndrome, or renal or cardiovascular disease has been recommended, at least after a certain age.
Urate is the end product of human purine metabolism and is mainly excreted in urine. Serum urate concentrations are determined by the volume of urate produced via purine metabolism and by renal urate excretion. Many factors, including genetic components and acquired factors such as obesity and alcohol consumption, influence serum urate concentrations. Genetic links to serum urate concen trations have been identified, mainly from earlier studies of monogenic disorders, but have also been recently analyzed using genomewide association approaches. Mono genic disorders such as hypoxanthineguanine phos phoribosyl transferase deficiency (LeschNyhan syndrome, MIM 300322), phosphoribosyl pyrophosphate synthetase overactivity, familial juvenile hyperuricemic nephropathy (MIM 162000) and glycogen storage disease are well known to induce hyperuricemia, while molybdenum co factor deficiency (MIM 252150), xanthinuria (MIM 278300 and 603592), and renal hypouricemia (MIM 220150 and 612076) induce hypouricemia [715]. These diseases, with the exception of the renal disorders familial juvenile hyperuricemic nephropathy and renal hypouricemia, are classified as enzymatic deficiencies and have contributed to our understanding of purine metabolism. Uromodulin, also known as TammHorsfall glycoprotein, was recently shown to cause the allelic disorders familial juvenile hyperuricemic nephropathy and medullary cystic kidney

Kimiyoshi Ichida
Address: Tokyo University of Pharmacy and Life Science, 14321, Horinouchi Hachioji, Tokyo, 1920392 Japan. Email: ichida@toyaku.ac.jp ABCG2, ATP-binding cassette, sub-family G, member 2; α-KG, α-ketoglutarate; CsA, cyclosporine A; ERM, ezrin-radixin-moesin; EST, expressed sequence tag; GAPDH, glyceraldehyde3phosphate dehydrogenase; GKRP, glucokinase regulatory protein; GLUT, glucose trans porter; NHERF, Na + /H + exchanger regulatory factor; LRRC16A, leucinerich repeatcontaining protein 16A; MCT9, monocarboxylate trans porter 9; MRP4, multidrug resistance protein 4; NPT1, sodiumdependent phosphate transport protein 1; OAT, organic anion transporter; PAH, p-aminohippurate; PPAR-α, peroxisome proliferator-activated receptor-α; SMCT, sodium-coupled monocarboxylate transporter; SNP, single nucleotide polymorphism; URAT1, urate transporter 1. disease 2, based on genetic linkage analyses, and the pathophysiology of these diseases has been clarified [16,17]. URAT1 was also recently identified as a homolog of organic anion transporter 1 and a transporter responsible for renal hypouricemia [18]. Using this finding as an indicator, several transporters have been shown to transport urate, leading to a better understanding of urate handling in the kidney. In recent years, genetic factors affecting serum urate concentrations have been identified by genomewide association studies. Most genes indicated in these studies have been linked to urate transport; this is because most individuals tend to maintain renal urate excretion ability to some extent, which compensates for serum urate concentrations even under conditions of urate overproduction. This article reviews the major genes known to influence serum urate concentrations.

Urate handling in the kidney
In healthy males, the urate pool averages about 1,200 mg, with a mean turnover rate of 700 mg/day. Under normal circumstances, twothirds to threequarters of daily urate disposal is excreted by the kidney. Urate reabsorption dominates over secretion in the kidney, resulting in the excretion of approximately 10% of its filtered load at the glomerulus. Nonproteinbound urate is freely filtered at the glomerulus. Urate is mainly reabsorbed in the proximal tubule in the kidney. The process of reabsorption of urate through proximal tubular cells is achieved via uni directional transcellular transport, involving the uptake of urate into the cells from the proximal tubular fluid across the apical membrane, followed by excretion into the blood across the basolateral membrane. Secretion of urate through proximal tubular cells is achieved by the opposite route.
Until the last decade, renal handling of urate had been explained by a 'fourcomponent model', which separated renal urate transport into glomerular filtration, pre secre tory reabsorption, secretion and postsecretory reabsorp tion. The concept of presecretory and postsecretory reabsorption was based on the hypothesis that the anti uricosuric effect of pyrazinamide was due to inhibition of urate secretion by pyrazinoate, the active metabolite of pyrazinamide. However, some reports using membrane vesicles have indicated that the antiuricosuric effect of pyrazinamide results from enhanced urate reabsorption [19,20].

Recent genome-wide association studies for serum urate concentrations
Candidategene association studies depend on current knowledge of a phenotype's suspected pathology to select single nucleotide polymorphisms (SNPs) to test for asso ciations, while genomewide association studies are essen tially screening studies without prior biological hypotheses. Genomewide association studies have the power to identify multiple new associations, although these require external validation and large sample sizes to achieve extremely small Pvalues. Recently, genomewide associa tion studies for serum urate concentrations have been performed [2128]. In all of the studies, SNPs in SLC2A9, the glucose transporter gene family, were unexpectedly associated with serum urate concentrations. Similarly, several genes, including ABCG2, encoding a multidrug resistance protein, have been identified.

Major genes influence serum urate concentrations
URAT1, a member of the organic anion transporter (OAT) family, is encoded by SLC22A12, and is expressed in the kidney. URAT1 was identified as a transporter for urate reabsorp tion in exchange for lactate at the apical membrane of the renal proximal tubular cell [18] (Figure 1). URAT1 mediates the exchange of urate for several organic anions and inorganic anions, such as lactate, pyrazine carboxylic acid and chloride, and is cisinhibited by probenecid, benzbromarone, losartan, and lactate [18]. Enomoto et al. [18] demonstrated that URAT1 regulates serum urate concentrations by showing that three patients with renal hypouricemia had defects in SLC22A12. Patients with this disorder demonstrate extremely low serum urate concentrations, mostly under 1.0 mg/dl, because of increased urate excretion. This fact indicates that URAT1 is a major urate reabsorptive transporter and is a therapeutic target for the treatment of hyperuricaemia.
About 90% of Japanese patients with renal hypouricemia exhibit a defect in URAT1 [3]. Renal hypouricemia is common in Japanese populations and possibly also in the nonAshkenazi Jewish ethnic group. The high incidence of renal hypouricemia is a reflection of the high allele fre quency (2.30 to 2.37%) of the G774A mutation in SLC22A12 among Japanese [29,30]. This mutation was originally brought by immigrants from the Asian continent, and thereafter expanded in the Japanese population by founder effects [31].

SLC2A9 (GLUT9)
The putative function of glucose transporter (GLUT)9 had been obscure, although GLUT9 (SLC2A9) was cloned as a member of the facilitated glucose transporter family, based on sequence similarity to GLUTs [32]. However, several genomewide association studies have demonstrated a clear association of SNPs in SLC2A9 with serum urate concentrations [21,22,2427,33].
GLUT9 is highly expressed in the kidney and liver. GLUT9L (long isoform) is localized to basolateral mem branes in proximal tubule epithelial cells, while the splice variant GLUT9S (short isoform) localizes to apical mem branes [34] (Figure 1). Vitart et al. [26] showed that GLUT9 transports urate and fructose, using a Xenopus oocyte expression system. Anzai et al. [35] characterized GLUT9 in detail and reported that it did not stimulate any significant uptake of organic anionic substrates, such as paraaminohippurate (PAH), estrone sulfate or salicylate, or of substrates known to interact with URAT1, such as lactate, nicotinate, β-hydroxybutyrate, or salicylate, there by suggesting a narrower substrate specificity than that of URAT1. Kinetic analysis indicates that GLUT9 is a highcapacity urate transporter and that extracellular glucose can accelerate urate efflux by GLUT9 [36]. The fact that GLUT9 deficiency resulted in renal hypouricemia shows GLUT9 to be an efflux transporter of intracellular urate from the tubular cell to the interstitium/blood space [37]. Efflux transport of urate at basolateral membranes appears to depend principally on GLUT9L. On the other hand, URAT1 mainly acts as an influx transporter for urate at Urate transporters at the proximal tubule. Transporters responsible for urate reabsorption and secretion are illustrated [18,21,43,48,52,7780]. Sodiumanion cotransporters SMCT1 and 2 at the apical membrane are included in this figure because SMCT transports lactate, the counterpart of urate. URAT1 is a main transporter for urate reabsorption in exchange for lactate at the apical membrane. URAT1 mediates the exchange of urate for several organic anions and inorganic anions. The long isoform of GLUT9 (GLUT9L) is localized to basolateral membranes in proximal tubule epithelial cells, while the short splice variant (GLUT9S) localizes to apical membranes. GLUT9L is a primary efflux transporter of intracellular urate to the interstitium/blood space. Abbreviations: ABCG2, ATPbinding cassette, sub-family G, member 2; α-KG, α-ketoglutarate; CsA, cyclosporine A; ES, estrone sulfate; GLUT, glucose transporter; MRP4, multidrug resistance protein 4; NPT1, sodiumdependent phosphate transport protein 1; OAT, organic anion transporter; PAH, paraaminohippurate; SMCT, sodiumcoupled monocarboxylate transporter; URAT1: urate transporter 1. Mice with systemic knockout of Glut9 display moderate hyperuricemia, massive hyperuricosuria, and an early onset obstructive nephropathy [38]. In contrast, liver specific inactivation of the Glut9 gene in mice leads to severe hyperuricemia and hyperuricosuria in the absence of urate nephropathy. Fractional excretion of urate in liver specific Glut9 knockout mice was lower than that in systemic Glut9 knockout mice. These data and the absence of urate nephropathy in liverspecific Glut9 knockout mice suggest that the urate transport direction via Glut9 at apical membranes is reabsorptive in mice. The hyper uricemia in systemic and liverspecific Glut9 knockout mice indicates that urate cannot be converted to allantoin in the absence of liver Glut9; this demonstrates that hepatocytes take up urate via Glut9 at basolateral membranes. However, the physiological role for GLUT9S at renal apical membranes and for GLUT9 in hepatocytes in humans has not yet been defined.
Furthermore, GLUT9 is thought to act as a common transporter mediating urate metabolism and glucose and fructose metabolism, since both diabetes mellitus and high fructose intake influence serum urate concentrations. Further research into this relationship and the role of GLUT9 in diabetes and metabolic syndrome may lead to the development of effective prevention and treatment of these diseases.

ABCG2
ATPbinding cassette, subfamily G, member 2 (ABCG2) is a halftransporter and most likely functions as a homo dimer. ABCG2 was cloned in a project to characterize all human ATPbinding cassette superfamily genes [39]. First identified as a multidrug resistance protein, ABCG2 has a wide range of substrates, such as mitoxantrone, topotecan, rhodamine 123, methotrexate, estrone3sulfate, and porphy rins [40]. ABCG2 is expressed in the plasma membranes of a variety of tissues, including placenta, pharynx, bladder, brain, and intestine, and mediates the efflux of xenobiotics. In the kidney, ABCG2 is expressed at the apical membrane of the proximal tubule [41] (Figure 1). Although ABCG2 was identified pathophysiologically as a gene partially responsible for porphyria, ABCG2's role in vivo remains unclear [42].
In genomewide association studies, SNPs in ABCG2 have been found to be related to serum urate concentrations [23,27]. The ability of ABCG2 to transport urate was recently confirmed by measuring urate efflux from ABCG2 expressing Xenopus oocytes [43]. It would follow, there fore, that ABCG2 would excrete urate at the renal proximal tubule apical membrane. The frequency of the mutation Q141K, encoded by the common SNP rs2231142, is highly variable in the human population; the frequency of the A allele ranges from 1 to 5% in Africans, to approximately 30% in Asians. Urate transport by the ABCG2 mutant Q141K is about half that of the wild type [43]. Q126X shows stronger effects on gout development than Q141K, confer ring an odds ratio of 5.97 [44]. Furthermore, 10% of patients with gout had genotype combinations resulting in more than 75% reduction of ABCG2 function (odds ratio 25.8). These findings indicate that nonfunctional variants of ABCG2 essentially block gut and renal urate excretion and cause gout.

SLC22A11 (OAT4)
OAT4 encoded by SLC22A11 is expressed in the kidney and placenta at moderate levels [45]. OAT4 is localized to the apical membrane of proximal tubular cells in the kidney ( Figure 1). OAT4 exhibits 53% amino acid homology with URAT1. OAT4 functions as an organic anion/dicarboxylate exchanger and is responsible for the reabsorption of organic anions driven by an outwardly directed dicarboxy late gradient [46]. Substrates for OAT4 include sulfate conjugates such as estrone sulfate and dehydroepiandro sterone sulfate, prostaglandins E 2 and F 2α , and urate [45,47,48]. Since OAT4 is thought to be an asymmetric carrier, it may transport organic anions such as glutarate and paminohippurate outward into the lumen and act as an entry route for urate and estrone sulfate into the proximal tubule cell. Probenecid inhibits OAT4 with K i values of approximately 50 μM, which would be sufficient to decrease urate reabsorption by OAT4 [49,50].
Hagos et al. [48] reported OAT4 to be a lowaffinity urate transporter, using cells stably expressing OAT4 and OAT4 expressing oocytes in plasmaequivalent concentrations (up to 400 μM). The contribution of OAT4 to urate transport at the apical membrane of proximal tubular cells under physiological conditions is unclear; however, genomewide association studies have reported an association between OAT4 and serum urate concentrations [27]. Thus, OAT4 may share with URAT1 the physiologic function of urate reabsorption at the apical membrane.

SLC17A1 (NPT1), SLC17A 3 and SLC17A4
Metaanalysis of genomewide association studies showed that a region mapped to chromosome 6p23-p21.3, including the SLC17A1, SLC17A3, and SLC17A4 genes, is associated with serum urate concentrations [27]. Renal sodiumdependent phosphate transport protein 1 (NPT1) is encoded by SLC17A1 [51]. NPT1, which was first cloned as a phosphate transporter, is located in the proximal convoluted renal tubule (Figure 1). NPT1 mediates voltage sensitive transport of organic anions, including urate, and is suggested to function as a urate secretor [52]. In SNP analysis of NPT1 in patients with gout, the T allele frequency of rs1165196 (T806C) was significantly higher in patients than in control individuals and T806C showed significant association with reduced serum urate concen trations in obese individuals in spite of a negative asso cia tion in all controls [53]. However, the precise mechanisms at a molecular level remain to be clarified.
NPT4, encoded by SLC17A3, is expressed in the kidney, brain, and liver, while a sodium/phosphate cotransporter encoded by SLC17A4 is expressed in the intestine, bladder, and liver, and weakly in the kidney and testis. NPT4's biological function has not been clarified in detail, and although a heterozygous transition of NPT4 in a patient with glycogen storage disease type Ic has been reported, a causal relationship has not been established [54].

PDZK1
PDZK1, coding for PDZ domain containing 1, acts as a scaffolding protein for a large variety of transporter and regulatory proteins and has been identified in the kidney, liver, small intestine, and adrenal cortex [55]. Within the kidney, PDZK1 is localized in the apical membrane of the proximal tubule. PDZK1 contains four PDZbinding domains, each of which binds independently a sequence specific PDZ motif at the carboxyterminal end of transporters ( Figure 2). PDZ domains are thought to play important roles in targeting of proteins to specific cell membranes, assembling proteins into signaling complexes for efficient transduction, and regulating the function of transporters. The urate transporters URAT1, OAT4, and NPT1 interact with PDZK1 via a class I PDZ motif (S/TX Φ, where X is any amino acid and Φ is a hydrophobic amino acid) [5558]. Coexpression of URAT1 or OAT4 and PDZK1 in HEK293 cells increases transport activity through increasing cellsurface expression of the trans porters. This effect is suggested to result from stabilization and/or anchoring of URAT1 and OAT4 at the cell membrane by PDZK1. PDZK1 may also provide a structural basis for functional coupling of transporters. For example, binding of both URAT1 and sodiumcoupled monocarboxy late transporter (SMCT) to PDZK1 may induce efficient substrate transport because monocarboxylates such as lactate, pyruvate, β-hydroxybutyrate and acetoacetate are substrates for SMCT; thus, an outwardly directed gradient Schematic representation of the interaction between PDZ proteins and urate transporters at the apical membrane of the proximal tubule. PDZK1 binds another PDZ protein, Na + /H + exchanger regulatory factor 1 (NHERF1). NHERF1 contains two tandem PDZ domains and a carboxyterminal domain that binds members of the ERM (ezrinradixinmoesin) family of membranecytoskeletal adaptors. The carboxyl terminus of URAT1, OAT4, and NPT1 binds with PDZK1 [81]. URAT1 interacts with PDZK1 via PDZ domains 1, 2, and 4, while OAT4 interacts with PDZ domains 1 and 4 [56,57]. OAT4, NPT1, and MRP4 also bind NHERF1 [82,83]. Abbreviations: MRP4, multidrug resistance protein 4; NHERF, Na + /H + exchanger regulatory factor; NPT1, sodiumdependent phosphate transport protein 1; OAT, organic anion transporter; URAT1, urate transporter 1. of these monocarboxylates created by the sodiumcoupled uptake by SMCT drives URAT1mediated urate reabsorp tion [56,5962]. Consequently, SNPs and some mutations of PDZK1 would influence serum urate concentrations.
Fibrates, the peroxisome proliferator-activated receptor-α (PPAR-α) agonists, decrease hepatic levels of PDZK1 in a PPAR-α-dependent fashion in mouse liver, while PPAR-α upregulates the expression of PDZK1 in humans [63,64]. PPAR-α, which belongs to the nuclear receptor superfamily and regulates the expression of genes responsible for fatty acid β-oxidation and energy homeostasis, is one of the key molecules involved in metabolic disorders. Regulation of PDZK1 expression by PPAR-α has not been fully clarified, and it is difficult to explain the induction of renal urate underexcretion in obesity on the basis of human PDZK1 upregulation by PPAR-α. However, further elucidation of the relationship between PDZK1 and PPAR-α would help determine the role of PDZK1 in metabolic disorders.

SLC16A9 (MCT9)
SLC16A9, encoding monocarboxylate transporter 9 (MCT9), was identified purely from analysis of human genomic expressed sequence tag (EST) databases [65]. MCT9 is expressed in the parathyroid, kidney, trachea, spleen and adrenal gland. Metaanalysis of genomewide association studies showed a relationship between SLC16A9 and serum urate concentrations, but the function of MCT9 remains unknown.

LRRC16A (CARMIL) and SCGN
Metaanalysis of genomewide association studies showed a region containing the genes leucinerich repeat containing protein 16A (LRRC16A) and SCGN, chromo somal locus 6p22.2, as a region associated with serum urate concentrations. However, the mechanism of their involvement remains elusive. CARMIL, encoded by LRRC16A, is important for actinbased motility and can bind to actin capping protein, an essential element of the actin cytoskeleton. Actin capping protein regulates polymerization by binding to the barbed ends of actin filaments. CARMIL inhibits the binding ability of actin capping protein and regulates its interaction with actin filaments. SCGN, coding for Secretagogin, a calcium binding protein, is expressed in neuroendocrine tissue and pancreatic betacells. The function of Secretagogin is unknown, but it has been suggested to influence calcium influx, insulin secretion and proliferation in β-cells [66].

GKRP
Glucokinase, expressed exclusively in liver and pancreatic β-cells, plays an essential role in glucose metabolism by catalyzing the phosphorylation of glucose (Figure 3). Glucokinase regulatory protein (GKRP) acts as a compe titive inhibitor as well as a nuclearbinding protein for glucokinase. Glucokinase is located in the nucleus, bound to GKRP as an inactive complex under basal glucose conditions. In highglucose conditions, most studies have suggested that GKRP releases glucokinase, which is Relationships between purine, fructose, and glucose metabolism. Abbreviations: GAPDH, glyceraldehyde3phosphate dehydrogenase; PRPP, phosphoribosyl pyrophosphate. Hyperuricemia has been reported to be associated with the metabolic syndrome based on insulin resistance and hyperinsulinemia, since insulin decreases renal urate clear ance [4,72]. Another hypothesized mechanism for hyperuricemia due to insulin resistance is that adenine nucleotide translocator inhibition by increased intra cellular longchain fatty acylCoA ester in insulin resistant states leads to high cytosolic AMP concentrations; this results in hyperuricemia by a high rate of breakdown to urate [73]. Insulin resistance may mediate the association of hyperuricemia with GKRP, as identified by a genomewide association analysis [27].

Clinical implications of genetic and genomic data
Among the disorders associated with hyperuricemia or hypouricemia, gout is the most common disease, and its incidence is increasing. Hyperuricemia is classified into urate overproduction type, underexcretion type, and mixed type. Antihyperuricemic agents include xanthine dihydro ge nase inhibitor and uricosuric agent. Japanese guidelines for management of gout recommend the use of xanthine dihydrogenase inhibitor for overproductiontype hyper uricemia, and uricosuric agent for underexcretiontype hyperuricemia [74]. Uricosuric agents, however, need to maintain a high urinary output or alkalinization of urine for prevention of urolithiasis. Furthermore, the indication of benzbromarone, a main uricosuric agent, has been controversial because of rare but serious hepatotoxicity. Thus, uricosuric agents have been regarded as a second line agent of the xanthine dihydrogenase inhibitor [75,76]. However, urate excretion is the major factor in the regulation of serum urate concentrations, supported by the results of genomewide association studies. Recently, identified molecules such as GLUT9 and ABCG2 should be candidates for targeting the development of new anti hyperuricemic agents.

Conclusions
Recent progress in the study of renal urate transport has identified transporters for urate reabsorption and secretion at apical and basolateral membranes. The extent of the contribution of some transporters such as URAT1 and GLUT9 to urate transport has also been clarified, and a synoptic renal urate transport model has been developed.
Genomewide association studies have led to verification of the model and the identification of novel candidate genes related to urate metabolism and transport. Most candidates have been categorized as transporters. These results are consistent with the fact that about 90% of hyperuricemic patients suffer from underexcretiontype hyperuricemia. However, functional and physiological roles of several candidates are as yet uncertain, and the influence of acquired factors, including obesity, diet or alcoholic beverages, also requires further investigation.
Future research should elucidate the urate transport systems in the liver and intestine, since hepatocytes are major contributors to purine metabolism and about one third of daily urate disposal is excreted into the intestine.
The mechanism of overproductive hyperuricemia should also be established.

Competing interests
The author declares that he has no competing interests.