Soluble Uric Acid Is an Intrinsic Negative Regulator of Monocyte Activation in Monosodium Urate Crystal–Induced Tissue Inflammation

This article is featured in In This Issue, p.553

Uric acid (UA) is a metabolite breakdown product of purines and primarily excreted by the kidneys. In most mammals, including rodents, serum UA concentrations are low owing to the presence of the enzyme uricase, which degrades soluble UA (sUA) into the more–water soluble allantoin (1). However, humans lack functional uricase; thus, sUA is the final end product of the purine metabolism (2). As a consequence, serum UA levels are relatively high in humans and increase the risk for developing gouty arthritis or kidney stones. The excretion of sUA is largely regulated by urate transporters in the renal tubules, such as urate transporter 1 (SLC22A12/URAT1) (3, 4), glucose transporter 9 (SLC2A9/GLUT9) (5), ATP-binding cassette transporter subfamily G member 2 (ABCG2) (6), and organic anion transporters (7, 8). Hence, genetic variants, environmental factors, use of drugs that impair the UA transport capacity, and a reduced number of nephrons are frequent causes for hyperuricemia (HU) in humans (8).

sUA is the substrate for the formation of monosodium urate (MSU) crystals during gouty arthritis (9, 10). MSU crystals like many other crystals or microparticles elicit acute inflammation characterized by secretion of proinflammatory mediators, the recruitment of leukocytes from the bloodstream into the inflamed joint (11–14), and the release of neutrophil extracellular traps (12, 15, 16). Although no doubt exists on the inflammatory effects of MSU crystals, published data on the pathophysiological role of sUA remain controversial. Several in vitro studies reported proinflammatory effects of sUA, including activation of the Akt-PRAS40 pathway (17), induction of NLRP3 inflammasome-mediated IL-1β release (18–21) in immune cells, and reactive oxygen species production (22), or NF-κB and MAPK signaling in nonimmune cells (23–26). In contrast, sUA can exert antioxidant properties (8) and act as an inhibitor of protein nitrosation, lipid and protein peroxidation, low-density lipoprotein oxidation, and a deactivator of tetrahydrobiopterin (8), which were shown to elicit protective effects in neurodegenerative diseases (27) and cancer (28). In addition, sUA can inhibit Akt phosphorylation in endothelial cells (29) and modulate the activity of extracellular superoxide dismutase in atherosclerotic vessels (30).

Why do such discrepancies on the effects of sUA exist? One explanation could be the different preparation methods used in in vitro studies. Indeed, among several preparation methods, preparing sUA by prewarming or by sodium hydroxide (NaOH) solubilization are most frequently employed. We hypothesized that UA crystal contaminations may account for the false-positive reporting proinflammatory effects and that only preparing crystal-free sUA can ultimately reveal its impact on innate immunity. We speculated on immunomodulatory effects in human monocytes through the intracellular uptake of sUA in vitro, and, if so, asymptomatic HU would suppress MSU crystal–induced tissue inflammation in vivo.

All animal experiments were performed in accordance with the European protection law of animal welfare and upon approval by the local government authorities (Regierung von Oberbayern, reference number: ROB-55.2-2532.Vet_02-15-189) based on the European Union directive for the Protection of Animals Used for Scientific Purposes (2010/63/EU) and reported according to the Animal Research: Reporting of In Vivo Experiments guidelines (31). Mice were housed in groups of five in filter-top cages with unlimited access to food and water. Cages, nest lets, food, and water were sterilized by autoclaving before use.

Six-week-old Alb-creERT2;Glut9^lox/lox and Glut9^lox/lox control mice (kindly provided by Frédéric Preitner and Bernhard Thorens, University of Lausanne, Lausanne, Switzerland) were injected i.p. with tamoxifen every alternate day for 1 wk to eliminate hepatic Glut9 expression in Alb-creERT2;Glut9^lox/lox mice (32). Both groups of mice received a standard chow diet enriched with 25.6 g of inosine per kilogram (Ssniff, Soest, Germany) to induce asymptomatic HU in Alb-creERT2;Glut9^lox/lox mice for 22 d. Glut9^lox/lox mice served as controls (healthy).

Alb-creERT2;Glut9^lox/lox and Glut9^lox/lox mice were injected s.c. with MSU crystals (5 mg per mouse; Invivogen, San Diego, CA) or vehicle into a preexisting air pouch, a mouse model of tissue inflammation (33). After 12 h, both groups of mice (healthy and HU) were euthanized, and air pouch fluids were collected for further analysis.

Upon sacrifice, we sampled the kidneys from healthy and HU mice. We kept one kidney in RNA later solution at −80°C for RNA isolation, and the second kidney was fixed in 4% formaldehyde to be embedded in paraffin for histology analysis (34). For demonstrating kidney injury, we stained 2-μm-thick kidney sections with periodic acid–Schiff reagent. The RNA extraction kit from Qiagen (Düsseldorf, Germany) was used to isolate total RNA from mouse kidneys (n = 5 mice per group) following the manufacturer’s instructions. RNA quality was assessed using 1% agarose gels before being transcribed into cDNA using reverse transcriptase (Superscript II) (Invitrogen, Carlsbad, CA). Real-time RT-PCR was performed using SYBRGreen PCR master mix and analyzed with a Light Cycler 480 (Roche, Mannheim, Germany). All gene expression values were normalized using 18s rRNA as a housekeeping gene (35). All murine primers used for amplification were purchased from Metabion (Martinsried, Germany) and are listed in Table I.

Air pouch fluid was harvested by injecting 1 ml of cold PBS into the air pouch. Air pouch extracts were retrieved and centrifuged at 1200 rpm for 5 min. Supernatants were collected and stored at −20°C for further cytokine and chemokine analysis by ELISA. Air pouch extracts were resuspended in wash buffer and FcR blocked for 5 min before staining with the surface Abs FITC anti-mouse Ly6G, PE anti-mouse Ly6C, and PerCP anti-mouse CD45 (all Abs obtained from BioLegend, Fell, Germany) for 30 min. After incubation, cells were washed in Dulbecco PBS (D-PBS) and reconstituted in 1 ml of fresh wash buffer (0.1% BSA, 0.01% sodium azide in D-PBS). Flow cytometry was performed using the BD FACSCalibur (Becton Dickinson, Franklin Lakes, NJ) and data analyzed with the software FlowJo 8.7 (Tree Star, Ashland, OR). AccuCheck counting beads (Thermo Fisher Scientific, Langenselbold, Germany) were used, and the absolute number of cells per microliter was calculated according to manufacturers’ instruction.

Serum blood urea nitrogen and creatinine (DiaSys, Holzheim, Germany) and serum UA (BioAssay Systems, Hayward, CA) levels were measured using commercially available kits as per manufacturer’s protocol.

Human PBMCs were isolated from healthy individuals (n = 8; plasma UA: 4.17 ± 0.46 mg/dl; plasma urea: 28.26 ± 3.83 mg/dl; plasma creatinine: 1.45 ± 0.05 mg/dl) or hyperuricemic chronic kidney disease (CKD) patients stage G2–5 (n = 10; plasma UA: 9.73 ± 0.5 mg/dl; plasma urea: 85.65 ± 8.35 mg/dl; plasma creatinine: 5.5 ± 0.53 mg/dl) using an S-Monovette with lithium-heparin-gel (SARSTEDT, Nümbrecht, Deutschland), followed by standard dextran sedimentation and Ficoll–Hypaque density centrifugation procedures (36). We only included patients that did not receive immunosuppressive drugs and urate-lowering therapy. CD14⁺ monocytes were purified by magnetic activated cell sorting with human CD14 microbeads for positive selection (Miltenyi Biotec, Bergisch Gladbach, Germany), according to manufacturers’ protocol. Monocytes (1 × 10⁶ cells/ml) were suspended in complete RPMI medium (10% FBS, 1% penicillin/streptomycin, pH 7.4) and seeded onto 24-well plates in a 5% carbon dioxide atmosphere at 37°C for 30 min before stimulation. The study to obtain whole blood samples from healthy individuals and hyperuricemic CKD patients was approved by the local Ethical Review Board of the Medical Faculty at the Hospital of the Ludwig Maximilian University of Munich. Informed consent was obtained from all subjects prior to inclusion in the study.

Two methods described in the literature were used. First, UA (Sigma-Aldrich, Taufkirchen, Germany) was solubilized at different concentrations in prewarmed RPMI media, mixed, and incubated at 37°C for 1 h. After incubation, pH was checked, and all sUA solutions were filtered sterile using a 0.22-μm filter (17, 18, 22, 37, 38). Second, UA (25 mg) was solubilized by adding 50 ml of prewarmed RPMI medium, and 4 N of NaOH (50 μl) was added. The UA solution was mixed to dissolve crystals. Afterwards, pH was adjusted to 7.2 with 6 N of HCl (35 μl), the UA stock solution was mixed and filtered sterile (0.22-μm filter) (39). The concentration of the UA solutions (5–50 mg/dl) was measured, and the presence of UA microcrystals was visualized under the light microscope. For all in vitro cell culture experiments with CD14⁺ monocytes and THP-1 cells, we used freshly prepared UA solubilized with NaOH apart from data presented in Fig. 1C and 1H.

CD14⁺ monocytes were isolated from healthy individuals or hyperuricemic patients and pretreated with or without 10 mg/dl sUA or 20% human plasma from healthy individuals or HU patients (with or without 0.1 μg/ml rasburicase) and then stimulated either with 10 μg/ml LPS (Sigma-Aldrich) or 500 μg/ml MSU crystals (InvivoGen) or left untreated (medium control). Total mRNA was isolated for RT-PCR after 4 h, cells were prepared for intracellular cytokine staining after 8 h of stimulation by flow cytometry, or supernatants were collected for neutrophil extracellular trap formation assays.

For intracellular staining, CD14⁺ monocytes from healthy individuals were incubated with brefeldin A for 15 min. Cells were then harvested, washed, and resuspended in cell fixation/permeabilization buffer (BD Cytofix/Cytoperm Solution Kit; Thermo Fisher Scientific) for 15 min and washed in permwash buffer (BD Cytofix/Cytoperm Solution Kit; Thermo Fisher Scientific). Intracellular Abs for anti-human TNF-α and anti-human IL-6 (BioLegend) were added for 40 min at 4°C. After incubation, cells were washed with D-PBS and reconstituted in fresh wash buffer (0.1% BSA, 0.01% sodium azide in D-PBS). Flow cytometry was performed using the BD FACSCalibur (Becton Dickinson) and analyzed using FlowJo 8.7 (Tree Star).

Human monocytic THP-1 cell line was purchased from American Type Culture Collection and grown in RPMI culture medium supplemented with 10% FBS (Biochrom, Berlin, Germany), 2 mM glutamine (Life Technologies, Invitrogen, Grand Island, NY), 1 mM sodium pyruvate, 10 mM HEPES, and 1% penicillin/streptomycin (PAA Laboratories, Pasching, Austria). Cells were incubated at 37°C in 5% CO₂.

THP-1 cells were cultured in six-well plates (Costar; Corning Incorporated, Corning, NY) at a cell density of 1 × 10⁶ cells per well and preincubated with or without 10 mg/dl sUA for 1 h (short-term exposure) prior to stimulation with 10 μg/ml LPS, 200 μg/ml MSU crystals, or 100 ng/ml human TNF-α for another 4 h. For long-term exposure experiments, UA was prepared with two different preparation methods, and THP-1 cells were stimulated with UA (5–50 mg/dl UA stock solutions) for 24 h and washed with D-PBS, and fresh media were added prior to stimulation with LPS (1 μg/ml) or MSU crystals (500 μg/ml) for 24 h. Cells were harvested for RNA preparation and quantitative RT-PCR (qRT-PCR) (NFKB-p50, NLRP3, ASC, IL-6, TNF-α), and culture supernatants were collected and stored at −80°C until analysis. THP-1 cells, CD14⁺ monocytes, and human primary tubular epithelial cells (40) were quantified for mRNA expression levels of urate transporters via qRT-PCR (human primer sequences are listed in Table II) and for protein quantification of SLC2A9 Western immunoblotting.

SLC2A9 (GLUT9) and control small interfering RNA (siRNAs) (iBONi siRNA pool) were purchased from RIBOXX (Radebeul, Germany). THP-1 cells (1 × 10⁶ cells per well) were transiently transfected with 120 nM control siRNA and SLC2A9 siRNA using Neon transfection system (Thermo Fisher Scientific). The electroporation parameters were as follows: Pulse voltage 700 v, Pulse width 40 ms, and Pulse number 4. RPMI medium was changed with complete RPMI medium (10% FBS, 1% penicillin/streptomycin) after 4 h posttransfection. Forty-eight hours after transfection, THP-1 cells were preincubated with 10 mg/dl sUA for 1 h prior to stimulation with 10 μg/ml LPS or 100 ng/ml TNF-α for another 4 h.

Human THP-1 cells (1 × 10⁶ cells per well) were pretreated with or without 10 mg/dl sUA for 1 and 4 h in complete RPMI media and then treated with 10 μg/ml LPS or 100 ng/ml human TNF-α for another 4 h. Culture supernatants were collected, and the cells were harvested, washed with D-PBS, and digested (41). The concentrations of sUA in cell culture supernatants (extracellular) and cytosol (intracellular) were measured by a commercial UA assay kit (BioAssay Systems, Hayward, CA), and the intracellular sUA uptake rates were determined by adjusting UA levels to cell number in each sample.

We analyzed expression patterns of urate reabsorption transporters SLC2A9, SLC22A11, SLC22A12, and SLC22A13 and urate excretion transporters SLC2A6, ABCC4, ABCG2, SLC17A1, and SLC17A3 as previously described of either PBMCs (n = 4 controls) (17), human monocytes (n = 11 controls) (42), and human primary tubular epithelial cells (n = 3 controls) (43). All data were downloaded as processed expression matrix without quality control, which means that the reliability of the RNA sequencing and the Affymetrix GeneChip microarray is unknown. The formal analysis may deviate from the following results. RNA sequencing of PBMCs (GSE107821: platform Illumina HiSeq 2000–Homo sapiens) and human tubular epithelial cells (GSE115960: platform Illumina HiSeq 2500–Homo sapiens) used a reads per kilobase per million reads algorithm, and the microarray of monocytes (GSE9988) used the Affymetrix Human Genome U133 Plus 2.0 Array, in which the robust multiarray average algorithm is used, both to normalize and count gene expression. The genes PTPRC and CD14 were used as positive control for PBMCs and monocytes. For primary tubular epithelial cells, we used PTPRC and CD14 as negative controls. Heat maps were generated using the MultiExperiment Viewer (v4.2) software.

For indirect neutrophil extracellular traps that release via soluble mediators, supernatants from activated human CD14⁺ monocytes (as indicated above) were added to freshly isolated neutrophils (2 × 10⁵ cells per well/200 μl in eight-well microslides; Ibidi, Martinsried, Germany) from healthy individuals for 3 h. Afterwards, neutrophils were fixed with 4% paraformaldehyde and neutrophil extracellular traps stained, as previously described (36). The percent area of citrullinated histone H3 (red; Cell Signaling, Danvers, MA) and neutrophil elastase (green; Santa Cruz Biotechnology, Santa Cruz, CA) was quantified for neutrophil extracellular traps using ImageJ software.

The RNA extraction kit (Qiagen) was used to isolate total RNA from human CD14⁺ monocytes, primary tubular epithelial cells, and THP-1 cells following manufacturer’s instructions. The isolated RNAs were quantified using the Nano drop (PEQLAB Biotechnology GMBH, Erlangen, Germany), and RNA integrity was assessed with 1% agarose gels before being transcribed into cDNA using reverse transcriptase (Superscript II) (Invitrogen). Real-time qRT-PCR was performed using Light Cycler 480 (Roche) with SYBR Green PCR detection dye system (SYBR Green I 96 protocol LC480 Roche running program). All gene expression values were normalized to GAPDH as housekeeping gene. All human primers used for amplification were purchased from Metabion and are listed in Table I.

Human THP-1 cells were lysed with radioimmunoprecipitation assay buffer containing phosphatase and protease inhibitor (Roche), and protein concentration was measured (Bio-Rad Protein Assay). Protein samples (30 μg) were mixed with 4× sample loading buffer (Thermo Fisher Scientific), denatured at 95°C for 5 min, and loaded on Tris/HCl Gel 4–20% (Bio-Rad). Using wet transfer system (Bio-Rad), protein was transferred onto 0.2-μm nitrocellulose blotting membranes (GE Healthcare Life Sciences, Amersham). Membranes were blocked with 5% nonfat dry milk or 5% BSA in TBS-Tween (0.1% Tween 20) for 1 h at room temperature. The following primary human Abs were used: SLC2A9 (0.5 μg/ml, 1 h, room temperature; NOVUS Biologicals, Wiesbaden-Nordenstadt, Germany) and β-actin (1:1000, 1 h, room temperature; Cell Signaling Technology, Frankfurt am Main, Germany). After being washed with TBS-Tween 20, membranes were incubated with secondary Abs: goat anti-mouse (1:2000) or goat anti-rabbit (1:2000) IgG labeled with HRP (1 h, room temperature; both from Cell Signaling Technology). Immunostained bands were detected using a chemiluminescence kit (ECL Kit; Lumigen, Southfield, MI) and the relative protein ratio SLC2A9/β-actin quantified by ImageJ.

Concentrations of mouse IL-6 and IL-1β in air pouch extracts were measured using mouse ELISA kits (RayBiotech, Norcross, GA), and human IL-1β in THP-1 supernatants was measured using a human ELISA kit (BioLegend, San Diego, CA) according to manufacturers’ protocols. Absorbance was measured on a Multiskan EX reader (Thermo Fisher Scientific, Langenselbold, Germany).

Statistical analysis was performed using GraphPad Prism 7 software. Data were compared by Student t test to test for significance between two groups, one-way ANOVA with Tukey post hoc test between three or more groups, or by two-way ANOVA with Bonferroni comparison post hoc test when using two parameters with multiple groups. Data are presented as mean ± SD. Differences were considered significant if p < 0.05.

Two different sUA preparation methods are mainly reported in the literature: prewarming UA (Fig. 1A) (17, 18, 20, 22, 37, 38) and solubilizing UA with NaOH (Fig. 1B) (39, 44). Checking the prepared solutions for contaminant crystals by light microscopy, we observed microcrystals even after sterile filtration of the UA solutions (5–50 mg/dl) upon prewarming UA but not upon solubilizing UA with NaOH (Fig. 1C, 1D). Indeed, the final sUA concentration in the prewarming preparation containing microcrystals was much lower than predicted from the input of UA substrate (only 30% at the highest concentration), whereas the final sUA concentration upon NaOH-based sUA preparation exactly matched the UA substrate input (Fig. 1E). Obviously, prewarming solubilized only a part of the UA substrate, whereas the other part got or remained incorporated in UA crystals contaminating the final sUA preparation. We observed a pH drop in the prewarmed solutions, whereas the crystal-free NaOH-based sUA preparation solutions were stable at the adjusted pH of 7.2 (Fig. 1F). Next, we tested both preparations on the inflammatory function in THP-1 cells incubated with either sUA preparation for 24 h prior to restimulation with MSU crystals or LPS (Fig. 1G, long-term sUA exposure). Whereas prewarmed sUA alone triggered the release of IL-1β and augmented this response upon subsequent MSU crystal exposure (Fig. 1H), NaOH-solubilized sUA did not induce IL-1β release and showed only a minor effect upon MSU crystal challenge (Fig. 1I). These data indicate that the proinflammatory effects of sUA prepared by prewarming do not relate to sUA per se but to microcrystal contaminants creating erroneous results and misinterpretations. In contrast, NaOH solubilization generates a crystal-free sUA preparation, which allows studying the true biological effects of sUA.

Previously employed rodent models of HU have serious limitations, for example, they do not reach serum concentrations of UA relevant to human disease (17, 20, 21, 23, 26, 45–48). We used a transgenic mouse model based on the selective deactivation of GLUT9 in hepatocytes, which has been reported to develop clinically relevant levels of asymptomatic HU (32, 49). To test for an inhibitory effect of HU during acute tissue inflammation in vivo, we first injected Alb-creERT2;Glut9^lox/lox and Glut9^lox/lox control mice with tamoxifen to activate the transgene and placed both groups on a chow diet enriched with the urate precursor inosine (Fig. 2A). Alb-creERT2;Glut9^lox/lox mice fed an inosine-rich diet developed HU with a serum UA range of 8–12 mg/dl, whereas UA levels remained low in Glut9^lox/lox mice (healthy) (Fig. 2B). No effect on kidney function was observed between both groups as indicated by similar serum blood urea nitrogen and creatinine levels (Fig. 2C), periodic acid–Schiff staining, and intrarenal mRNA expression levels of KIM-1, Tnfα, and Fibronectin 1 (Table I, Supplemental Fig. 1A, 1B), that is, asymptomatic HU.

To mimic acute tissue inflammation as in gouty arthritis, we injected MSU crystals (or vehicle) into dorsal s.c. air pouches in HU and healthy mice (Fig. 2A, 2D). Twelve hours later, air pouch lavage fluids were collected, and flow cytometric analysis of the infiltrating immune cells was performed (Fig. 2E, gating strategy). The number of infiltrating CD45⁺ leukocytes and CD45⁺Ly6G–Ly6C⁺ monocytes in control mice injected with MSU crystals significantly increased compared with vehicle-treated mice (Fig. 2F, 2G). However, mice with HU displayed significantly fewer CD45⁺ cells and monocytes (Fig. 2F, 2G) as well as neutrophils (healthy: 18,756 ± 2011 and HU: 14,319 ± 1480, data not shown) in MSU crystal–injected pouch fluids. Pouch fluid levels of IL-1β and IL-6 were reduced in mice with HU accordingly (Fig. 2H, 2I). Taken together, the data show that HU attenuates MSU crystal–induced tissue inflammation by suppressing leukocyte recruitment and activation.

To test whether NaOH-solubilized sUA exerts inhibitory effects also in human monocytes after short-term sUA exposure, we isolated CD14⁺ monocytes from the blood of healthy subjects (Fig. 3A) and preincubated the cells with 10 mg/dl sUA prior to stimulation with LPS, MSU crystals, or medium alone (Fig. 3B, short-term exposure). Upon MSU crystal or LPS stimulation, the mRNA expression levels of NLRP3, TNF-α, and IL-6 significantly increased compared with medium (Fig. 3C). However, the presence of sUA suppressed mRNA expression levels of NLRP3, IL-6, and TNF-α in CD14⁺ monocytes stimulated with LPS or MSU crystals (Fig. 3C, Table II). Flow cytometric analysis for the intracellular cytokines IL-6 and TNF-α confirmed that sUA significantly reduced the number of IL-6 and TNF-α single-positive as well as of IL-6/TNF-α double-positive CD14⁺ monocytes compared with LPS or MSU crystal stimulation only (Fig. 3D).

To confirm a suppressive effect of sUA on monocytes also in patients, we treated CD14⁺ blood monocytes isolated from hyperuricemic CKD patients as well as healthy subjects with MSU crystals, LPS, or medium alone for 4 h. Upon stimulation with MSU crystals, the mRNA expression levels of NLRP3, ASC, IL-6, NF-κB, IL-6, TNF-α, and IL-12p40 increased in CD14⁺ monocytes from healthy individuals compared with untreated monocytes (medium) (Fig. 3E). However, this strong inflammatory response of MSU crystal–activated CD14⁺ monocytes from healthy individuals (healthy) was significantly inhibited in activated CD14⁺ monocytes from hyperuricemic patients (HU), except for the anti-inflammatory cytokine IL-10, which increased in mRNA expression (Fig. 3E). LPS stimulation gave similar results, wherein all of the tested genes were significantly downregulated in monocytes from HU patients (Fig. 3F). To test whether sUA in the plasma from HU patients contributes to the intrinsic dysfunction of monocytes from HU patients versus healthy controls, we isolated CD14⁺ monocytes from healthy subjects and preincubated the cells with plasma from either HU patients or healthy individuals. The plasma contained UA levels as follows: HU was 9.73 ± 0.50 mg/dl, and healthy was 4.17 ± 0.46 mg/dl. HU plasma decreased the inflammatory function of LPS-activated CD14⁺ monocytes from healthy donors compared with activated CD14⁺ monocytes that were incubated with plasma from healthy subjects, an effect entirely reversed by pretreating plasma with rasburicase (Fig. 3G).

As macrophage-derived soluble mediators, including IL-1β and TNF-α, can promote neutrophil extracellular trap release (50), we also tested the effect of supernatants from MSU crystal– or LPS-activated CD14⁺ monocytes on freshly isolated neutrophils from healthy subjects. Indeed, soluble mediators released from LPS- or MSU crystal–activated CD14⁺ monocytes induced neutrophil extracellular trap release (Supplemental Fig. 2). This effect was attenuated by supernatants from sUA + MSU crystals with a significant effect from sUA + LPS-stimulated monocytes (Supplemental Fig. 2). The data indicate that short-term exposure with NaOH-solubilized sUA inhibits the inflammatory effects of activated CD14⁺ monocytes, an effect that was confirmed in CD14⁺ monocytes from HU patients.

How could sUA inhibit the inflammatory function of human monocytes? As a receptor-mediated outside-in signaling mechanism seemed less likely, we speculated on a urate transporter-mediated uptake of sUA inside the cells. However, there is a paucity of data on the expression of urate transporters in immune cells; hence, we first sought to identify urate transporters expressed by monocytes. RNA-sequencing data of PBMCs and human primary tubular epithelial cells from the public domain (17, 43) and microarray data of human monocytes (42) suggested that these cells selectively express the urate reabsorption transporter SLC2A9, whereas primary tubular epithelial cells express all known urate transporters (Fig. 4A). RT-PCR of various other urate transporters confirmed the selective expression of SLC2A9 in human CD14⁺ monocytes and THP-1 cells, whereas primary tubular epithelial cells, used as positive control, expressed most of them (Fig. 4B, 4C, Table II).

If SLC2A9 is involved in the immunomodulatory effects of sUA on monocytes, it should mediate the intracellular uptake of sUA. Indeed, exposing LPS-activated THP-1 cells to 10 mg/dl sUA for 1 and 4 h resulted in an increase of intracellular sUA concentration (Fig. 4D, 4E).

Next, we preincubated THP-1 cells with or without 10 mg/dl sUA for 1 h (short-term) prior to stimulation with LPS for 4 h and found that the presence of LPS significantly reduced extracellular sUA levels and increased intracellular levels of sUA in UA-treated THP-1 cells (Fig. 4F–H). Benzbromarone, an inhibitor of several urate transporters, attenuated the LPS-induced intracellular uptake of sUA (Fig. 4H). LPS stimulation did not affect the mRNA and protein expression level of SLC2A9 (Fig. 4I, 4J). Despite the fact that LPS activation did not alter the expression of SLC2A9 in sUA-treated THP-1 cells, we observed a significant decrease in the mRNA expression of the inflammatory genes NFKB-p50, ASC, NLRP3, IL-6, and TNF-α when LPS- or human TNF-α–activated THP-1 cells were preincubated with 10 mg/dl sUA (Fig. 4K, Supplemental Fig. 3A).

To test for a role of SLC2A9 in this context, we silenced SLC2A9 by specific siRNA in THP-1 cells as confirmed by RT-PCR and immunoblotting (Fig. 5A–C). SLC2A9 silencing significantly attenuated the suppressive effect of LPS on extracellular and intracellular sUA levels, which implies that SLC2A9 mediates sUA uptake into activated monocytes (Fig. 5D, 5E). As a consequence, SLC2A9 silencing abolished the suppressive effect of sUA on the mRNA expression of NFKB-p50, IL-6, and TNF-α (Fig. 5F). The same effect of sUA uptake on the inflammatory function in THP-1 cells was observed in response to TNF-α upon silencing the SLC2A9 urate transporter (Supplemental Fig. 3B–E). We conclude that SLC2A9-mediated intracellular uptake of sUA is an essential step in UA-related suppression of activated monocytes in response to inflammatory stimuli.

We had hypothesized that the discrepant literature on the immune effects of sUA may relate to how sUA is actually prepared and that only a microcrystal-free sUA preparation will give reliable results. Indeed, our in vitro studies clearly identify that sUA prepared by prewarming produces erroneous proinflammatory results for a contamination with UA microcrystals. A proper crystal-free preparation reveals anti-inflammatory effects of sUA on activated monocytes, driven by SLC2A9-mediated intracellular uptake of sUA. The suppressed inflammatory immune response of sUA was confirmed in CD14⁺ monocytes from hyperuricemic patients. As our in vivo experiments in a meaningful model of asymptomatic HU and MSU crystal–induced tissue inflammation validate this finding, we conclude that sUA is an intrinsic inhibitor of tissue inflammation also if triggered by MSU crystals.

Several in vitro studies reported proinflammatory effects of sUA using another protocol for sUA preparation that (17, 18, 20, 22, 38, 51–55), as we show in this article, contains crystalline particles that are well-known activators of sterile inflammation (56). Therefore, such effects cannot be attributed to sUA. Previous in vitro studies reported that NaOH-solubilized sUA does not elicit an inflammatory response by immune cells (39, 44) and that sUA can act as a potent scavenger of singlet oxygen (57, 58) and inhibitor of UMP synthase to antagonize the cytotoxic effects of the chemotherapeutic agent 5-fluorouracil (59). The solubility of UA strongly depends on the preparation solution and the physiological pH range. UA can be solubilized in NaOH as recommended by commercial suppliers, such as Sigma-Aldrich, but not simply in D-PBS or prewarmed cell culture medium. In a previous study, Iwata et al. (60) found that the solubility of UA increases with the rise in pH and that long-term incubation at 37°C can decrease the total dissolved UA concentration due to crystallization. This would indicate that besides the pH range (7.0–7.4), the concentration and storage time of the stock solution can influence the solubility of UA, which should be considered when performing in vitro experiments. The mechanism of crystallization, nucleation, and crystal growth involves a combination of critical processes, including concentration of solutes, temperature, and pH. For example, crystallization of calcium oxalate crystals strongly depends on the pH (61), which is associated with kidney diseases (62, 63) because of a decline in urinary pH. Similar physiochemical processes apply to MSU and UA crystallization in gouty arthritis (64–66) and UA kidney stone disease (65, 67–69), respectively. Unlike the microcrystal-contaminated prewarmed sUA solution, a crystal-free preparation of sUA rather suppresses inflammatory responses. Our numerous in vitro experiments were consistent in this regard and were confirmed in CD14⁺ monocytes isolated from hyperuricemic patients.

The suppressive effect of sUA on the inflammatory function of activated human monocytes may occur following uptake of sUA via urate transporters (8). Although many of the urate transporters are highly expressed in the kidney and intestine, recent studies found SLC22A12 and SLC2A9 to be expressed in human endothelial cells, leukocytes, and chondrocytes (37, 70, 71). The urate transporters SLC2A9 (37) and BCRP/ABCG2 (29) can regulate the sUA uptake in human umbilical vascular endothelial cells, which further triggers cell injury and inflammation. Unlike endothelial (37), pancreatic β (72), and vascular smooth muscle (26) cells, we found that the intracellular sUA uptake by human monocytes occurs via SLC2A9 leading to this suppressed phenotype. Whether other transporters apply to other cell types, for example, urate transporters or selective organic anion channels (73), remains to be investigated in future studies.

The overall functional importance of HU in vivo remains controversial because of the lack of a suitable animal model (17, 20, 21, 26). All studies investigating the effect of HU in vivo are based either on i.p. injections of high concentrations of UA (47, 48), oral administration of potassium oxonate (23, 45, 46), unilateral ureter obstruction-induced nephropathy (20), or diabetic nephropathy (21) inducing only very low serum UA levels, which would not even be considered clinically as HU. We have now established an animal model of HU (serum UA levels of ∼10 mg/dl) without kidney impairment that mimics the range of asymptomatic HU in humans more accurately. Using this in vivo approach, we show for the first time, to our knowledge, that HU directly impairs leukocyte recruitment, which diminished MSU crystal–related tissue inflammation.

In conclusion, we identify sUA as an intrinsic inhibitor of innate immunity, a mechanism associated with the intracellular uptake of sUA via SLC2A9 by activated monocytes resulting in a suppressed phenotype. This immunomodulatory property of sUA may explain several clinical observations. First, HU is a relatively common metabolic condition, yet only a minority of patients with HU (10–15%) develop symptomatic gout (74). Second, a sudden reduction of HU with urate-lowering therapy can elicit acute gout attacks (75), probably because of a sudden decline in the negative regulator of sterile inflammation sUA. In addition, 15% or more individuals with asymptomatic HU have crystal deposits without experiencing classical symptoms (76). Fourth, despite persistent HU, not all patients with chronic kidney disease suffer from acute gout attacks (77). Further studies are needed to better understand the role of asymptomatic HU in gouty arthritis and beyond. It is important to recognize that soluble and crystalline UA have opposite effects in innate immunity.

We thank all healthy volunteers involved in this study.

The authors have no financial conflicts of interest.