Febuxostat, an Inhibitor of Xanthine Oxidase, Suppresses Lipopolysaccharide-Induced MCP-1 Production via MAPK Phosphatase-1-Mediated Inactivation of JNK

Excess reactive oxygen species (ROS) formation can trigger various pathological conditions such as inflammation, in which xanthine oxidase (XO) is one major enzymatic source of ROS. Although XO has been reported to play essential roles in inflammatory conditions, the molecular mechanisms underlying the involvement of XO in inflammatory pathways remain unclear. Febuxostat, a selective and potent inhibitor of XO, effectively inhibits not only the generation of uric acid but also the formation of ROS. In this study, therefore, we examined the effects of febuxostat on lipopolysaccharide (LPS)-mediated inflammatory responses. Here we show that febuxostat suppresses LPS-induced MCP-1 production and mRNA expression via activating MAPK phosphatase-1 (MKP-1) which, in turn, leads to dephosphorylation and inactivation of JNK in macrophages. Moreover, these effects of febuxostat are mediated by inhibiting XO-mediated intracellular ROS production. Taken together, our data suggest that XO mediates LPS-induced phosphorylation of JNK through ROS production and MKP-1 inactivation, leading to MCP-1 production in macrophages. These studies may bring new insights into the novel role of XO in regulating inflammatory process through MAPK phosphatase, and demonstrate the potential use of XO inhibitor in modulating the inflammatory processes.


Introduction
Inflammation plays a fundamental role in a variety of chronic diseases such as atherosclerosis, rheumatoid arthritis, and chronic obstructive lung disease [1][2][3]. During inflammation, activated macrophages release pro-inflammatory cytokines that amplify cellular responses to injury as well as generating reactive oxygen species (ROS), which play an important role in the defense against invading organisms [4]. However, excess ROS production by activated cells can also provoke inflammation and tissue damage.
Xanthine oxidase (XO) is one of the major enzymatic sources of ROS. It is derived from xanthine dehydrogenase (XDH) either by proteolytic modification or reversible sulfhydryl oxidation [5]. XO catalyzes the oxidation of purine substrates, such as xanthine and hypoxanthine, producing uric acid and ROS [5]. XO has been reported to be up-regulated by various inflammatory stimuli such as LPS, hypoxia, and cytokines [6][7][8][9]. Augmented XO eventually causes excess ROS formation, leading to tissue damage. Pharmacological inhibitors of XO, such as febuxostat, allopurinol and oxypurinol have been reported to have an anti-inflammatory effect in various diseases such as atherosclerosis, chronic heart failure, acute lung injury, renal interstitial fibrosis and ischemic-reperfusion injury [10][11][12][13][14][15]. Thus, these findings demonstrate the essential role of XO in inflammatory conditions. However the mechanisms that link XO production to inflammation are not well understood.
We have studied the effects of modulating XO activity by LPS, in particular its effects on MCP-1, a potent chemotactic factor for monocytes and dendritic cells, and investigated if inhibition of XO leads to an altered inflammatory response. LPS is a main component of the outer membrane of Gramnegative bacteria, and previous studies have demonstrated that LPS up-regulates XO expression and activity. Pharmacological inhibition of XO protects against LPS-induced tissue injury [6,7]. Febuxostat, is a potent and selective XO inhibitor that effectively inhibits the formation of uric acid [16]. In this study, we found that febuxostat significantly suppresses LPS-induced MCP-1 production in macrophages and mice. These inhibitory effects were mediated by decreasing ROS formation and activating MKP-1, which leads to dephosphorylation and inactivation of JNK. These results may bring new insights into the novel role of XO in regulating inflammatory process through MAPK phosphatase.

In vivo experiments
Female C57BL/6 mice (from Charles River) between 8-12 weeks of age were used for experiments. Animal experiments were performed in strict accordance to the Swiss Federal Regulations. The protocol was approved by the "Service de la consommation et des affaires vétérinaires du Canton de Vaud", Switzerland. All efforts were made to minimize suffering and minimize the number of mice needed to assess statistical significance and experimental reproducibility. Mice were intraperitoneally treated with 0.5 ml of febuxostat at 1 mM or an equal volume of vehicle (0.6% DMSO) 30 min before an intraperitoneal dose of 300 µg LPS (Escherichia coli O111:B4) in 0.5 ml sterile phosphate buffered saline (PBS). In addition control mice injected with vehicle or febuxostat alone were performed. After 6 hours, blood was collected and mice were euthanized by CO 2 administration and peritoneal exudate cells were subsequently harvested by performing lavage with 3 ml of PBS. Lavage fluids were centrifuged at 450g for 10 min.
Supernatants of lavage fluids and sera were used for MCP-1 ELISA measurement.

Measurement of intracellular ROS level
Differentiated THP-1 cells were loaded for 30 min with 5 µM H 2 DCFDA (Life Technologies), and washed twice with HEPESbuffered saline (Life Technologies). Then, cells were pretreated for 10 min with the indicated concentration of febuxostat, allopurinol, oxypurinol or NAC, and stimulated with 100 ng/mL

Measurement of ROS level produced by XO/xanthine
Buttermilk XO (Merck Millipore, Billerica, MA) was incubated for 60 min with 100 µM of xanthine, and ROS levels in the solutions were measured using the DCF indicator of Oxiselect ™ in vitro ROS/RNS assay kit (Cell Biolabs, Inc., San Diego, CA) according to manufacturer's instructions.

Measurement of XO activity
Cells were homogenized with 50 mM phosphate buffer containing protease inhibitor cocktail, and sonicated for 10 min. The homogenate was incubated with 5 µM of pterine (Sigma-Aldrich) for 60 min, and fluorescent intensity on 360 nm (excitation) and 400 nm (emission) was measured with Paradigm fluorescent plate reader.

Statistical analysis
All data are expressed as mean±SEM. For two-group comparisons, student's t test was used. For multiple comparisons, one-way ANOVA followed by a Dunnett's test was used to compare each group versus a vehicle-treated group. In time-dependent experiments, two-way ANOVA was used for two-group comparisons and Bonferroni test was used for comparisons at each time point. All data were statistically analyzed using GraphPad PRISM software version 4.01 (GraphPad, La Jolla, CA). Differences with a probability value of <0.05 were considered significant.

Febuxostat inhibits LPS-induced XO activity and intracellular ROS formation in PMA-treated THP-1 cells
In this study, we first examined whether LPS enhanced XO activity and whether febuxostat inhibits LPS-induced XO activity and intracellular ROS formation. PMA-treated THP-1 cells were stimulated with 100 ng/mL LPS in the presence or absence of febuxostat. As shown in Figure 1A, LPS treatment resulted in the rapid and sustained increase in XO activity 0.25-4 h after treatment. Febuxostat effectively inhibited basal and LPS-induced XO activity ( Figure 1A). In addition, the analysis using H 2 DCFDA, an indicator of ROS, showed that intracellular ROS formation was significantly increased by LPS treatment in a time-dependent manner ( Figure 1B). Febuxostat inhibited LPS-induced intracellular ROS formation whereas Nacetyl-L-cysteine (NAC), an antioxidant, deleted intracellular ROS completely ( Figure 1B and 1C). These data demonstrated that febuxostat inhibits LPS-induced XO activity and XOderived intracellular ROS formation.

Febuxostat suppresses LPS-induced MCP-1 production by inhibiting XO-derived ROS formation in macrophages and mice
To determine the effects of febuxostat on LPS-induced inflammatory responses, we examined whether febuxostat suppresses LPS-induced MCP-1 production in macrophages. PMA-treated THP-1 cells were pretreated with 30 µM of febuxostat and then stimulated with 100 ng/mL LPS for 20 h. Although LPS treatment resulted in a large amount of MCP-1 in the supernatant, febuxostat significantly and dose-dependently inhibited LPS-induced production of MCP-1 without any effect on cytotoxicity (Figure 2A and 2B). NAC also inhibited LPSinduced production of MCP-1. Additionally, in human primary monocyte-derived M1 and M2 macrophages, febuxostat also suppressed MCP-1 production induced by LPS ( Figure S1 in File S1). To confirm the inhibitory effect of febuxostat was dependent on the inhibition of XO, we tested whether XO siRNA, as well as febuxostat, suppressed LPS-induced MCP-1 production. XO knockdown mediated by XO siRNA but not control siRNA resulted in the inhibition of LPS-induced MCP-1 production ( Figure 2C and S2 in File S1). Taken together, these results demonstrated that febuxostat suppresses LPSinduced MCP-1 production through inhibiting XO-derived ROS formation in macrophages.
Furthermore, we examined whether febuxostat also inhibits LPS-induced MCP-1 secretion in a in vivo model. Mice were intraperitoneally treated with febuxostat, then after 30 min mice were injected with 300 µg of LPS. In vehicle-treated mice, robust increases in MCP-1 levels were observed in peritoneal lavage whereas MCP-1 levels in febuxostat-treated mice were significantly reduced compared to vehicle-treated mice ( Figure  2D). Similar data were obtained in serum from vehicle-and febuxostat-treated mice (vehicle: 98001±8624, febuxostat: 72545±7279, p<0.05). These data demonstrated that febuxostat inhibits LPS-induced MCP-1 production not only in a in vitro cell system but also in a in vivo mice model.

Febuxostat suppresses LPS-induced MCP-1 mRNA expression without affecting mRNA stability
Next, we examined whether febuxostat inhibits LPS-induced MCP-1 mRNA expression in PMA-treated THP-1 cells. LPS treatment results in the increase in MCP-1 mRNA expression in a time-dependent manner whereas febuxostat suppressed LPS-induced MCP-1 mRNA expression ( Figure 3A). Consistent with the effects on protein level in Figure 2A, febuxostat significantly inhibited the induction of MCP-1 mRNA by LPS in a dose-dependent manner ( Figure 3B). In addition, LPSinduced TNF-α expression was also inhibited by febuxostat ( Figure S3 in File S1). Inhibitory effects of febuxostat were observed 2 h after LPS treatment, suggesting that febuxostat affected the transcription of MCP-1 mRNA. We next tested whether febuxostat affected the stability of MCP-1 mRNA. As shown in Figure 3C, febuxostat failed to affect the stability of MCP-1 mRNA. These results suggest that febuxostat inhibits

Febuxostat suppresses LPS-induced ROS formation and MCP-1 production more effectively than allopurinol and oxypurinol
It has been reported that other XOR inhibitors such as allopurinol and oxypurinol suppress intracellular ROS formation by XO, but their inhibitory effects are less potent than that of febuxostat [17]. Febuxostat (100 nM) potently inhibited ROS production by cell free XO/xanthine system, whereas the inhibitory effects of allopurinol or oxypurinol were partial, even at 1000-fold concentrations of febuxostat ( Figure 4A). Similarly, in PMA-treated THP-1 cells allopurinol or oxypurinol had very weak or no inhibitory effects on LPS-induced intracellular ROS ( Figure 4B). Consistent with these results, allopurinol or oxypurinol did not inhibit LPS-induced MCP-1 production and mRNA as efficiently as febuxostat ( Figure 4C and 4D). These results indicated that febuxostat inhibits LPS-induced ROS formation and MCP-1 production more effectively than allopurinol and oxypurinol.

Febuxostat suppresses LPS-induced activation of JNK
In LPS signaling, activation of transcription factor NF-κB and MAPKs play essential roles in transcriptional induction of genes involved in inflammation, such as iNOS, COX-2 and TNF-α [18]. We therefore sought to investigate whether febuxostat inhibits signaling pathways, which play important roles in LPSmediated MCP-1 induction in macrophages. As shown in Figure 5A, LPS transiently decreased total IκB-α level, showing the activation of NF-κB. However, pretreatment with febuxostat failed to change total IκB-α level. In addition, LPS treatment resulted in a time-dependent phosphorylation of JNK and p38 ( Figure 5A), but not of ERK and MEK1 (data not shown).
Febuxostat suppressed LPS-induced phosphorylation of JNK but not of p38. To determine the involvement of JNK in LPSinduced MCP-1, we examined the effects of LPS on MCP-1 production and mRNA expression in the absence or presence of SP600125, which is the specific inhibitor of JNK. As shown in Figure 5B and 5C, SP600125 suppressed LPS-induced MCP-1 production as well as mRNA expression, indicating that JNK plays essential role in LPS-induced MCP-1 production. These results suggest that febuxostat suppresses LPSmediated activation of JNK.

Febuxostat suppresses LPS-induced MCP-1 production via MKP-1-dependent inhibition of JNK through increasing MKP-1 activity
Because protein phosphatases have been shown to regulate the activation of JNK [19], we therefore examined whether protein phosphatases are involved in febuxostat-mediated inhibition of JNK phosphorylation. We first evaluated the effects of febuxostat on LPS-induced phosphorylation of JNK, and LPS-mediated MCP-1 expression in the presence or absence of protein phosphatase inhibitors. Here, we determined the effects of orthovanadate and Ro-31-8220, which are protein phosphatase inhibitor and MKP-1 inhibitor, respectively. As shown in Figure 6A, febuxostat-mediated inhibitions of phosphorylation of JNK were abolished by both orthovanadate and Ro-31-8220. Febuxostat no longer significantly inhibited LPS-induced MCP-1 expression in the presence of these inhibitors ( Figure 6B). Moreover, the inhibitory effects of NAC were also abolished by these inhibitors. These data demonstrate that febuxostat inhibits LPS-induced MCP-1 expression via MKP-1-dependent inhibition of JNK activation.
Furthermore, we determined whether febuxostat affects the activity of MKP-1. After LPS treatment, MKP-1 activity was rapidly decreased, and followed by gradual increase ( Figure  6C), suggesting that the change in phosphatase activity after LPS treatment was likely to affect the phosphorylation levels of JNK. Consistent with time points at which febuxostat-mediated inhibition of JNK activation was observed, febuxostat significantly increased phosphatase activity 1-2 hours after LPS treatment ( Figure 6C). Taken together, these results suggested that febuxostat suppresses LPS-induced MCP-1 production via MKP-1-dependent inhibition of JNK through increasing MKP-1 activity.

Discussion
Although several studies have reported that XO plays the fundamental roles in inflammation, the molecular mechanisms underlying the involvement of XO in inflammatory responses remain unclear. We here examined the effects of febuxostat, a selective and potent inhibitor of XO, on LPS-induced inflammatory signals. In this study, we found that febuxostat significantly suppresses LPS-induced MCP-1 production in human macrophages and in vivo in mice ( Figure 2 and Figure  S1 in File S1). The inhibitory effects of febuxostat on LPSinduced MCP-1 were mimicked by XO knockdown (Figure 2C), as well as the other XO inhibitors such as allopurinol and oxypurinol ( Figure 4C). In addition, exogenous XO induced febuxostat-inhibitable MCP-1 production in PMA-treated THP-1 cells ( Figure S4 in File S1). Taken together, these data suggested that XO plays essential roles in LPS-induced MCP-1 production, and that our in vitro system using PMA-treated THP-1 cells is well suitable for investigating the mechanisms underlying XO-mediated inflammatory pathways.
MKP-1, a protein phosphatase with dual specificity (Ser/Thr or Thr/Tyr), is responsible for the inactivation of JNK and p38, and therefore controls MAPK-dependent inflammation during the innate immune response [20,21]. Because MKP-1 activity has reported to be regulated by oxidative stress [22], we hypothesized that XO-derived ROS might contribute to decreased MKP-1 activity following LPS stimulation. Supporting this hypothesis, NAC-mediated inactivation of JNK was abolished by orthovanadate or Ro-31-8220 ( Figure 6A), and decrease in MKP-1 activity by LPS was recovered by NAC (data not shown). Thus, LPS decreases MKP-1 activity via XOderived ROS, leading to enhancement of JNK phosphorylation and ultimately to the induction of MCP-1. On the other hand, febuxostat inhibits XO-derived ROS and increases MKP-1 activity, which leads to dephosphorylation and inactivation of JNK. This is supported by the findings that febuxostat failed to inhibit LPS-induced MCP-1 production and phosphorylation of JNK in the presence of orthovanadate or Ro-31-8220 ( Figure  6A and 6B). Additionally, febuxostat enhanced MKP-1 mRNA levels, whereas it did not alter the mRNA levels of the other MKPs, such as MKP-3 and -5 (data not shown). To our knowledge, this is the first report showing a link between XO and MKP-1, representing an important and new regulatory mechanism in inflammation.
XO is well known to produce not only ROS but also uric acid. Several studies have reported that uric acid induced MCP-1 production in tubular epithelial cells, adipocytes and vascular smooth muscle cells [23][24][25]. However, our data revealed that XO-derived ROS is involved in LPS-induced MCP-1 production in PMA-treated THP-1 cells. In fact, LPS enhanced XO activity and increased intracellular ROS formation ( Figure 1). In addition, febuxostat as well as NAC attenuated increases in ROS formation and MCP-1 production (Figures 1 and 2). Furthermore, uric acid at concentrations equivalent to that found in hyperuricemic blood failed to affect basal and LPSinduced MCP-1 levels ( Figure S5 in File S1). Taken together, these findings demonstrate the crucial role of ROS in LPSinduced MCP-1 production in macrophages.
Inhibitory effects of febuxostat on the expression of cytokines/chemokines were selective: febuxostat also inhibited LPS-induced TNF-α, VCAM-1 and MMP9 expression, as well as MCP-1, but did not affect IL-8 ( Figure S3 in File S1 and data not shown). Additionally, inhibitory profiles of febuxostat did not completely correspond to that of SP600125 (data not shown), suggesting the possible involvement of signal molecules other than JNK in XO-mediated inflammatory responses. Further studies are required to clarify this possibility.
The inhibitory effect of febuxostat against LPS-induced MCP-1 production obtained in the human THP-1 cell line was reproduced in human primary macrophages. In both cell types, the IC50 dose was close to 10 µM. It is interesting to note that in clinical studies when febuxostat was given at 120 mg once daily for 13 days, the C max of febuxostat in plasma is around 17 µM [26,27]. Thus, the inhibitory doses of febuxostat in vitro are in the same range of the doses found in plasma from febuxostat-treated patients, suggesting that the antiinflammatory effects of febuxostat found in vitro may be relevant in vivo. Further clinical studies are needed to confirm the potential anti-inflammatory role of XO inhibiotrs.
In conclusion, febuxostat suppresses LPS-induced MCP-1 production and mRNA expression via activation of MKP-1 which, in turn, leads to dephosphorylation and inactivation of JNK in macrophages. These findings suggest that XO contributes to inflammatory responses via MKP-1-dependent JNK activation through intracellular ROS formation. These studies may bring new insights into a novel role of XO in regulating inflammatory process through MAPK phosphatase, and demonstrate the potential use of XO inhibitor in modulating the inflammatory processes.

Supporting Information
File S1. Supporting figures (Figures S1-S5). Figure S1. Human primary monocyte-derived macrophages were differentiated from CD14-positive monocytes with 50 ng/mL GM-CSF or M-CSF for 7 days, then stimulated with 15.6 ng/mL LPS in the absence or presence of febuxostat. The levels of MCP-1 in the supernatants were measured by ELISA. Data are shown as mean of two independent experiments in which similar results were obtained. Figure S2. PMA-treated THP-1 cells were transfected with 50 or 100 nM of XO siRNA or control siRNA, and then incubated for 20 h with 100 ng/mL LPS. Total RNA was extracted from 3 wells, and qRT-PCR was performed as described in materials and methods. Data are shown as mean of two independent experiments in which similar results were obtained. Figure S3. PMA-treated THP-1 cells were stimulated with 100 ng/mL of LPS for 2 h in the absence or presence of febuxostat (3, 10 and 30 µM) or NAC (10 mM). Total RNA was extracted, and qRT-PCR was performed as described in materials and methods. Data are shown as mean±SEM (n=3) of three independent experiments. # P<0.001 versus control-group, * P<0.05, ** P<0.01 versus vehicle/LPS-treated group. Figure S4. PMA-treated THP-1 cells were stimulated with 10 mU/mL of XO for 20 h in the absence or presence of febuxostat (10 µM). The levels of MCP-1 in the supernatants were measured by ELISA. Data are shown as mean of two independent experiments in which similar results were obtained. Figure S5. PMA-treated THP-1 cells were stimulated with 100 ng/mL of LPS for 20 h in the absence or presence of uric acid (0, 500 and 1000 µM). The levels of MCP-1 in the supernatants were measured by ELISA. Data are shown as mean±SEM (n=3).
(TIF) The effect of age and gender on pharmacokinetics, pharmacodynamics, and safety of febuxostat, a novel nonpurine