Hyperuricemia as a Mediator of the Proinflammatory Endocrine Imbalance in the Adipose Tissue in a Murine Model of the Metabolic Syndrome

Mouse preadipocyte cells 3T3-L1 obtained from ATCC (Manassas, VA) were maintained in high glucose Dulbecco’s modified Eagle’s medium (Invitrogen, Carlsbad, CA) supplemented with 10% FBS and antibiotics. For adipocyte differentiation, confluent cells were treated with 10 μg/mL insulin, 0.25 μmol/L dexamethasone, and 0.5 mmol/L isobutyl-1-methylxanhine for 2 days followed by 3- to 5-day treatment with insulin alone. Differentiation of adipocytes was confirmed by lipid staining with Red Oil and mRNA expression of PPAR-γ and adiponectin as described earlier (15). Human primary subcutaneous adipocytes were obtained from ZenBio (Research Triangle Park, NC) and cultured following the instruction of the manufacturer in subcutaneous adipocyte medium AM-1 (ZenBio). Uric acid solution for cell treatments (Ultrapure, 1–15 mg/dL; Sigma) was prepared in prewarmed cell culture medium as previously described (15). If treatments with uric acid were continued for more than 3 days, medium was replaced with freshly prepared identical medium. By the end of the 3-day treatment, there was no difference in pH between cell culture medium containing uric acid and control medium. In some experiments, conditioned medium was collected and stored at −80°C for measuring the levels of adipokines.

The Animal Care and Use Committee of the University of Florida approved protocols for all animal experiments. To test the effect of lowering uric acid, we used the Pound mouse, a recently established prediabetes/metabolic syndrome model (Charles River, Wilmington, MA). These mice have a mutation in the leptin receptor (a deletion of the exon 2), and they develop obesity, insulin resistance, dyslipidemia, and fatty liver disease but do not progress to type 2 diabetes (39). In preliminary experiments we showed that these mice develop hyperuricemia (13results). Obese (C57BL/6NCrl-Lepr^db-lb/Crl) and lean (C57BL/6NCrl) 6-week-old mice were maintained on a standard Purina 5001 Diet (PMI, Richmond, IN) under a 12-h light/12-h dark cycle with access to food and water ad libitum. To assess the effects of lowering uric acid levels, obese and lean animals (8–10 animals per group) were either administered for 8 weeks with allopurinol, an inhibitor of XOR (125 mg/L in the drinking water; Sigma, St. Louis, MO) or remained untreated. Every 3 weeks, samples of blood were obtained through facial vein to monitor the level of uric acid. Insulin tolerance testing was performed 1 week prior to the end of the experiment. Animals were fasted for 4 h and then injected intraperitoneally with insulin (Humulin R; Eli Lilly, Indianapolis, IN) at 1 unit/kg body wt. Glucose in tail vein blood was measured immediately before injection (time 0) and at 15, 30, 60, 90, and 120 min after injection using OneTouch Ultra glucometer (LifeScan, Milpitas, CA). At the end of the treatment period, the animals were killed by the intraperitoneal injection of 150 mg/kg sodium pentobarbital. Blood was collected via cardiac puncture, and serum was stored in aliquots at −80°C. For the mean arterial blood pressure (MAP) measurements, animals were placed under anesthesia (1–3% isoflurane), and arterial blood pressure was measured via the carotid artery by inserting a 27-gauge needle connected to a pressure transducer as described (40,41) followed by blood collection. While measurement of blood pressure under anesthesia may not recapitulate values in the conscious, unstressed state, we have previously reported that differences in MAP between experimental groups are maintained (40). Samples of visceral adipose tissue were collected on ice and rinsed in ice-cold PBS. For RNA isolation, tissue samples were preserved in RNAlater Solution (Applied Biosystems/Ambion, Austin, TX) at −80°C.

Total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA). Trace DNA was removed using a DNA-free kit (Applied Biosystems/Ambion, Austin, TX), and 0.5 μg of total RNA was converted to cDNA using iScript cDNA synthesis kit (Bio-Rad, Hercules, CA). Quantitative real-time RT-PCR was performed using SsoFast EvaGreen Supermix (Bio-Rad) with primers spanning two or more exons and optimized for real-time PCR, which were designed using Geneious Pro software (v. 4.8; Biomatters Ltd., Auckland, New Zealand) and Beacon Designer (v. 7.70; Premier Biosoft International, Palo Alto, CA). Sequences for all used primers are shown in Table 1. Real-time PCR was performed using CFX96 real-time PCR detection system as follows: 50°C for 2 min, then 95°C for 2 min, then 40 cycles of denaturation at 95°C for 15 s, annealing at 55°C for 30 s, and extension at 72°C for 30 s. Reaction specificity was confirmed by 1) electrophoresis of products in 2% agarose gel after real-time RT-PCR to check if bands of expected size were detected; 2) melting curve analysis; and 3) sequencing amplified fragments. Relative gene expression was analyzed by ΔΔC(t) method using the CFX Manager software (v. 1.6; Bio-Rad, Hercules, CA).

To assess the macrophage infiltration in the adipose tissue, samples of the visceral adipose tissue were collected in 10% buffered formalin, processed, embedded in paraffin, and sectioned. Macrophages were stained as described (42) with the F4/80 antibody from Invitrogen/Caltag Laboratories (Carlsbad, CA), secondary horseradish peroxidase–conjugated antibody, diaminobenzidine chromogen, and a hematoxylin counterstain.

Levels of MCP-1 and adiponectin in the cell culture medium were measured using the multiplex mouse adipocyte LINCOplex kit (Linco Research, St. Charles, MO) and the Luminex 100 system (Luminex, Austin, TX). Serum levels of MCP-1 and adiponectin were measured with the Milliplex MAP mouse single-plex kit and the mouse serum adipokine Milliplex MAP kit (Millipore, Billerica, MA), correspondingly using the Luminex 100 system.

Uric acid was measured by an enzymatic assay with urate oxidase (Diagnostic Chemicals Limited, Oxford, CT). Thiobarbituric acid reactive substances were measured with the assay from Cayman Chemical (Ann Arbor, MI).

At least three independent experiments were performed in the case of in vitro studies, and/or 6–10 animals per group were used for the in vivo studies. Each experiment was performed in triplicate. Data were analyzed by one-way ANOVA followed by Fisher least significant test, unpaired Student t test, or Mann-Whitney U test, with a value of P < 0.05 considered to represent a significant difference. Comparisons between two values were performed by t test or U test. ANOVA was used to test differences among several means.

To test the effects of uric acid on the adipokine production in vitro, we used differentiated mouse 3T3-L1 adipocytes and human primary subcutaneous adipocytes. When differentiated mouse adipocytes were incubated in the presence of 5 or 15 mg/dL uric acid for 3 days, we observed a significant dose-dependent increase in the mRNA abundance for MCP-1 (Fig. 1A). Even the lower dose of 5 mg/dL was significant (P < 0.05, U test). A time course experiment showed that 15 mg/dL uric acid induced an increase in MCP-1 mRNA at day 3, continued to increase to day 7, and plateaued between days 7 and 14 (Fig. 1B). The mRNA expression for adiponectin, which was induced by in vitro adipocyte differentiation, was lowered by both 5 mg/dL and 15 mg/dL of uric acid (Fig. 1C). In time course studies, uric acid also induced a gradual decrease in the adiponectin mRNA in adipocytes starting from day 3 to day 14 (Fig. 1D). Uric acid produced a similar effect in human primary subcutaneous adipocytes. Treatment with uric acid at a concentration of 15 mg/dL for 7 days induced a dramatic increase in mRNA level for MCP-1 and a decrease in the mRNA abundance for adiponectin (Supplementary Fig. 1A and B). A concentration of 7.5 mg/dL, which is a urate concentration typical for mild hyperuricemia in modern humans (43), did not produce statistically significant effects in this experiment (Supplementary Fig. 1).

We showed previously (15) that uric acid induced NOX-dependent reactive oxygen species (ROS) production in 3T3-L1 adipocytes, which triggered redox-dependent activation of the proinflammatory signaling by protein kinase p38. Therefore, we tested the hypothesis that production of proinflammatory factors in adipocytes in response to uric acid is also redox-dependent, and NOX activity is required for this effect. 3T3-L1 adipocytes were treated with uric acid for 7 days with or without Mn(II) tetrakis(1-methyl-4-pyridyl)porphyrin (MnTMPyP), a superoxide scavenger, or apocynin, an inhibitor of NADPH oxidase, and the MCP-1 and adiponectin mRNA and protein were determined. Uric acid induced a highly significant increase in MCP-1 mRNA in cells and protein concentration in the conditioned medium (Fig. 2A and B) while the presence of MnTMPyP or apocynin in the medium completely prevented these effects (Fig. 2A and B). These data suggest that urate-induced activation of the MCP-1 expression and secretion in adipocytes is mediated by the superoxide-dependent ROS, likely generated by NADPH oxidase. Moreover, urate-induced MCP-1 expression was abrogated by rosiglitazone, a PPAR-γ agonist (Fig. 2C and D), suggesting that activation of PPAR-γ prevents the effect of urate.

Treatment of adipocytes with 15 mg/dL uric acid for 7 days induced a dramatic fall in the adiponectin mRNA level, which was accompanied by moderate but significant decrease in the adiponectin concentration in the incubation medium (Fig. 3A and B). In contrast to MCP-1, the effect of uric acid on adiponectin production was not prevented by either MnTMPyP or apocynin (Fig. 3A and B), while rosiglitazone restored adiponectin mRNA in cells and adiponectin levels in conditioned media to control levels (Fig. 3C and D). These experiments suggest that uric acid–induced inhibition of adiponectin production is not mediated by redox-dependent signaling but rather by a mechanism involving PPAR-γ.

PPAR-γ activation by rosiglitazone prevented the effects of uric acid in our experiments. On the other hand, XOR, an enzyme-producing uric acid, is a crucial upstream regulator of PPAR-γ activity (38). Therefore, we hypothesized that uric acid could induce downregulation of PPAR-γ by affecting XOR via negative feedback mechanism. We tested an expression of the XOR mRNA in mouse and human adipocytes treated with uric acid for 7 days and found that uric acid induced a significant decrease in the XOR mRNA abundance in both cell types (Supplementary Fig. 2A and B).

We analyzed several animal models of obesity and metabolic syndrome or type 2 diabetes including Zucker diabetic fatty rats, ZSF1 rats, and the Pound mouse as well as corresponding lean control animals of the same age to compare the blood levels of uric acid in obese and lean animals. As shown in Fig. 4A, for each analyzed model, the level of uric acid in 8–week-old obese animals was substantially higher than in the corresponding lean animals. The greatest ratio between the level of uric acid in the lean control and obese animals was found in the Pound mice (Fig. 4A). These mice have metabolic syndrome but do not develop type 2 diabetes (39) and showed hyperuricemia compared with their littermates (Fig. 4A). Therefore, this model was selected for further experiments. To lower blood levels of uric acid, both lean and obese mice were treated with or without allopurinol for 8 weeks. Obese animals were hyperuricemic from the beginning of the study, and hyperuricemia progressed to even higher levels of uric acid during the course of the experiment (Fig. 4B). The correlation between the body weight and the level of uric acid in the Pound mouse is highly significant (R = 0.860, N = 33, P < 0.001). Serum uric acid substantially decreased in obese and lean animals after 2 weeks of treatments and remained low to the end of the study (Fig. 4B). The body weight gain was not affected by treatments with allopurinol in either obese (Fig. 4C) or lean animals (not shown).

As expected, MCP-1 mRNA was about 10-fold higher in the adipose tissue of obese animals compared with lean control animals whereas adiponectin mRNA expression was about 20-fold lower than that observed in lean animals (Fig. 5A and C). Serum concentrations of MCP-1 and adiponectin protein changed correspondingly (Fig. 5B and D). Lowering uric acid with allopurinol reduced the MCP-1 mRNA expression and serum levels in the Pound mice by >50% while inducing a threefold increase in adiponectin mRNA level with an increase in the level of the circulating adiponectin. Thus, hyperuricemia in the obese mice with metabolic syndrome might be partially responsible for the proinflammatory generation of MCP-1 by adipose. In addition, elevated levels of uric acid might contribute to the decrease in adiponectin. Because lowering urate blood levels with allopurinol improved the balance of the production of MCP-1 and adiponectin, hyperuricemia might contribute to the proinflammatory endocrine imbalance in the adipose tissue.

In order to check whether lowering uric acid reduced signs of inflammation in the adipose tissue in addition to MCP-1 production, we examined the visceral adipose tissue for the presence of macrophages, a key step in proinflammatory changes in obesity (42,44). The number of the F4/80-stained macrophages was negligent in lean mice and increased dramatically in obese mice (Fig. 6A, B, and D) while it significantly decreased in obese mice treated with allopurinol (Fig. 6C and D). Because expression of the proinflammatory cytokine TNF-α in the adipose tissue is elevated in obesity and it is macrophage-specific (44), we measured the abundance of mRNA for TNF-α in the adipose tissue. As expected, the TNF-α mRNA was induced dramatically in the Pound mice while allopurinol treatment reduced it (Fig. 6E), consistent with changes in macrophage infiltration and changes in the MCP-1 production. Obesity is associated also with oxidative stress in the adipose tissue (45). Proinflammatory effects of uric acid in our experiments were also redox-dependent. We assessed the effect of lowering uric acid on the level of oxidative stress by measuring malondialdehide (MDA), a product of lipid peroxidation. MDA levels in the visceral adipose tissue and serum of obese mice was increased while allopurinol treatment substantially reduced it (Fig. 6F and G).

The proinflammatory endocrine imbalance and low-grade inflammation in the visceral adipose tissue is an underlying mechanism in the insulin resistance and cardiovascular abnormalities in subjects with metabolic syndrome (18,19,29,30). To test whether lowering uric acid can improve signs of metabolic syndrome, we performed insulin tolerance tests and measured blood pressure in obese mice treated with allopurinol. As expected, obese mice had dramatically reduced insulin sensitivity (Fig. 7A) and elevated MAP in comparison with lean mice (Fig. 7B and C). Lowering uric acid significantly improved both parameters (Fig. 7A–C).

We found that uric acid can increase expression and release of MCP-1 and reduce production of adiponectin in cultured adipocytes. We further showed that an increase in MCP-1 with a reduction in adiponectin occurs in both adipose tissue and serum of the obese, hyperuricemic Pound mice, and that lowering uric acid can attenuate these changes. Obesity is associated with an increase in MCP-1 production (27,28) and a decrease in the adiponectin production (19,46) in the adipose tissue. These changes contribute substantially to the obesity-related low-grade inflammation and metabolic syndrome, including insulin resistance and hypertension (17,25). A dramatic increase in macrophage infiltration with expression of the proinflammatory cytokine TNF-α was observed in visceral adipose tissue in obese mice and is consistent with reported results (42,44). Importantly, lowering uric acid caused a reduction in macrophage infiltration and TNF-α expression as well as improved insulin sensitivity and blood pressure. This study may provide a mechanism linking hyperuricemia, obesity, and metabolic syndrome that has been shown previously in clinical studies (4,5,12). The observation that lowering uric acid improved but did not reverse the changes in MCP-1 and adiponectin, inflammation, and insulin resistance in the Pound mouse is consistent with uric acid being a modifying factor—but not the sole causal factor—in driving these changes.

We observed an increase in serum uric acid in the Pound mice and other animal models of obesity in comparison with the corresponding lean controls. This observation is in agreement with published data (38,47). Hyperuricemia progressed as animals gained weight, and the level of uric acid correlated with the body weight. As these mice are obese and insulin-resistant because of the mutation in the leptin receptor (39), hyperuricemia cannot be considered as a causal factor of obesity in this case. However, an increase in uric acid contributed to the low-grade inflammation and metabolic syndrome via its direct effect on the production of MCP-1 and adiponectin in the adipose tissue.

In mouse 3T3-L1 adipocytes and human primary adipocytes, uric acid can induce a direct dose-dependent increase in the production of MCP-1 and a decrease in the production of adiponectin at the level of the mRNA expression and protein release. For mice, cells responded to uric acid beginning at concentrations of 5 mg/dL, which is similar to the levels of uric acid in the obese Pound mice. In the case of human adipocytes, cells responded to 7.5–15 mg/dL uric acid, which encompass the range observed in subjects with asymptomatic hyperuricemia to severe gout (3).

The ability of soluble uric acid to induce MCP-1 expression was first demonstrated in rat vascular smooth muscle cells (31). The effect of uric acid was mediated by MAP kinases ERK1/2 and p38 and nuclear factor-κB in a redox-dependent fashion (31). In our previous work with 3T3-L1 adipocytes (15), we showed that uric acid induced ROS production via activation of NADPH oxidase followed by phosphorylation of p38. The urate-induced increase in the MCP-1 expression in adipocytes was downregulated by a superoxide scavenger and a NOX inhibitor. Collectively, these data suggest that uric acid induced MCP-1 production in adipocytes via the redox-dependent signaling initiated by NOX activation. Furukawa et al. (45) reported that obesity was associated with oxidative stress in adipose tissue, which, in turn, caused an overexpression of a variety of proinflammatory cytokines including MCP-1. We observed an increase in oxidative stress in our mouse model of obesity by measuring MDA in the adipose tissue and serum. This was reduced by allopurinol, suggesting that hyperuricemia induced by obesity is another mechanism triggering oxidative stress followed by induction of the MCP-1 expression and inflammation in the adipose tissue.

The uric acid–induced increase in the MCP-1 production observed in adipocytes was prevented not only by antioxidants but also by activation of PPAR-γ with rosiglitazone. An activation of PPAR-γ is known to block the proinflammatory effects of TNF-α in adipocytes by affecting the proinflammatory branch of the nuclear factor-κB–dependent signaling (48). XOR in adipocytes is thought to be a crucial upstream regulator of PPAR-γ activity (38). Treating adipocytes with exogenous uric acid reduced expression of XOR, which could be a cause of downregulated anti-inflammatory activity of PPAR-γ and facilitated redox-dependent MCP-1 production.

The mechanism of downregulation of the adiponectin production in obesity is not completely understood but appears to involve proinflammatory pathways (48,49), oxidative stress (45), and a deficiency in the PPAR-γ activity (48,50). The uric acid–induced decrease in adiponectin expression and secretion in our experiments was preventable by the PPAR-γ agonist, rosiglitazone, suggesting an involvement of the same kind of proinflammatory mechanism that was responsible for the induction of MCP-1. On the other hand, antioxidants could not improve adiponectin production affected by uric acid. We can suggest therefore that the deficiency of PPAR-γ but not oxidative stress was a primary trigger of this downregulation.

Because rosiglitazone prevented the uric acid–induced induction of MCP-1 and downregulation of adiponectin, these effects of uric acid might depend on PPAR-γ. XOR-dependent regulation of the adipocyte differentiation via control of PPAR-γ activity is known to be a finely orchestrated mechanism, which could be affected by manipulating the expression of XOR (38). Our data suggest that elevated concentrations of uric acid may reduce expression of XOR and attenuate PPAR-γ–dependent endocrine regulation in adipocytes.

In summary, uric acid can affect adipocytes directly by inducing effects resembling those observed in obesity: upregulation of proinflammatory factors and downregulation of the production of the insulin sensitizer and anti-inflammatory factor adiponectin via redox-dependent mechanisms, which can be prevented by an agonist of PPAR-γ (Fig. 8). In the mouse model of the metabolic syndrome, we observed hyperuricemia, which progressed with an increase in body weight. Lowering uric acid by inhibiting xanthine oxidoreductase in obese mice with the metabolic syndrome could improve the proinflammatory endocrine imbalance in the adipose tissue by lowering production of MCP-1 and increasing production of adiponectin.

This work was supported by American Heart Association Grant 0755595B (to Y.Y.S.) and by U.S. Public Health Service Grant HL-68607 (to R.J.J. and Y.Y.S.).

R.J.J. has several patent applications related to lowering uric acid as a means to treat hypertension and metabolic syndrome. He also has a lay book, The Sugar Fix (Rodale, 2008), that discusses the potential effects of fructose on obesity, diabetes, and cardiovascular disease.

No potential conflicts of interest relevant to this article were reported.

W.B. and S.M. researched data and reviewed and edited the manuscript. G.M., D.W., and V.P. researched data. C.B. contributed blood pressure measurements, contributed to discussion, and reviewed and edited the manuscript. R.J.J. contributed to discussion and reviewed and edited the manuscript. Y.Y.S. researched data and wrote the manuscript.

The University of Florida Hypertension Center and Bruce Cunningham, University of Florida, Gainesville, Florida, are acknowledged for help with measuring blood pressure in mice. Molecular Pathology and Immunology Core, University of Florida, Gainesville, Florida, is also acknowledged for help with histochemistry. The authors thank Charles River, Wilmington, Massachusetts, for generously providing serum samples for several models of obesity.