Xanthine Oxidoreductase Promotes the Inflammatory State of Mononuclear Phagocytes through Effects on Chemokine Expression, Peroxisome Proliferator-activated Receptor-γ Sumoylation, and HIF-1α*

The protective effects of pharmacological inhibitors of xanthine oxidoreductase (XOR) have implicated XOR in many inflammatory diseases. Nonetheless, the role played by XOR during inflammation is poorly understood. We previously observed that inhibition of XOR within the inflammatory mononuclear phagocytes (MNP) prevented neutrophil recruitment during adoptive transfer demonstrating the role of XOR in MNP-mediated neutrophil recruitment. To further explore the role of XOR in the inflammatory state of MNP, we studied MNP isolated from inflammatory lungs combined with analyses of MNP cell lines. We demonstrated that XOR activity was increased in inflammatory MNP following insufflation of Th-1 cytokines in vivo and that activity was specifically increased by MNP differentiation. Inhibition of XOR reduced levels of CINC-1 secreted by MNP. Expression of peroxisome proliferator-activated receptor γ (PPARγ) in purified rat lung MNP and MNP cell lines reflected both the presence of PPARγ isoforms and PPARγ SUMOylation, and XOR inhibitors increased levels of SUMO-PPARγ in MNP cell lines. Both ectopic overexpression of XOR cDNA and uric acid supplementation reduced SUMO-PPARγ in MNP cells. Levels of the M2 markers CD36, CD206, and arginase-1 were modulated by uric acid and oxonic acid, whereas siRNA to SUMO-1 or PIAS-1 also reduced arginase-1 in RAW264.7 cells. We also observed that HIF-1α was increased by XOR inhibitors in inflammatory MNP and in MNP cell lines. These data demonstrate that XOR promotes the inflammatory state of MNP through effects on chemokine expression, PPARγ SUMOylation, and HIF-1α and suggest that strategies for inhibiting XOR may be valuable in modulating lung inflammatory disorders.

Significantly, XOR has been found to be an essential component of innate immunity (37), to regulate leukocyte adhesion in vivo (38 -40), and as a product of the MNP to contribute to cytokine-induced acute lung injury (15). Inhibition of XOR within the newly recruited inflammatory MNP prevented neutrophil (polymorphonuclear phagocytes or PMN) recruitment during adoptive transfer demonstrating a key role for XOR in MNP-mediated PMN recruitment (15). Nonetheless, the role played by XOR in MNP during inflammation is still poorly understood, and in the present experiments we hypothesized that XOR may regulate basic MNP functions that contribute to inflammation. Data shown here demonstrate that XOR promotes the inflammatory state of MNP in part through effects on chemokine expression, PPAR␥ sumoylation, and HIF-1␣.
Animal Manipulations-Healthy male Sprague-Dawley rats (250 -350 g body weight, Sasco, Omaha, NE) were used in all studies. The Th-1 cytokines, 50 ng of recombinant rat IL-1␤ and 100 ng of IFN-␥ in 0.5 ml of normal saline, were delivered intratracheally into the airway of rats previously anesthetized with ketamine/xylazine, and lung tissues were harvested 5 or 24 h later as described previously (15). Control rats were insufflated with normal saline alone. The use of rats in this study was approved by the University of Colorado Animal Care and Use Committee under protocol number 21901907(08)1F.
Purification of MNP-Resident alveolar macrophages (RAM) were purified from the lungs of control rats by lavage as described previously (15). Circulating monocytes (CM) were purified from whole blood of control rats on Percoll gradients (15). Alveolar MNP (A-MNP), including both the resident AM population and the newly recruited MNP, were purified by lavage from lungs of rats insufflated with Th-1 cytokines 5 or 24 h before. Cells were washed in RPMI 1640 medium and were then plated into flasks with RPMI 1640 medium and allowed to adhere for 1 h. Nonadherent cells were removed, and the remaining adherent cells were washed four times with PBS and harvested from the plates by scraping into PBS/EDTA (42). Interstitial MNP present in the lung tissue (I-MNP) were purified from the lungs of rats following removal of the A-MNP and therefore do not contain the resident AM population. Lungs were lavaged to obtain the A-MNP population, perfused blood-free, and cells not released by lavage were prepared as described previously (43). Lung tissues were incubated under moderate agitation for 60 min in RPMI 1640 media containing 100 units/ml Ϫ1 collagenase type 1 and 50 units/ml Ϫ1 DNase at 37°C. After incubation, the digested tissue was passed through a sterile sieve to remove tissue fragments, and the cells obtained in the flow-through were adhered to flasks as described above. Adherent cells were suspended in saline to a final concentration of 2 ϫ 10 6 cells ml Ϫ1 . MNP comprised 95% of the cells recovered from adhesion as determined by microscopic examination of Wright's stained cells and macrophage-CSF-induced differentiation from both the A-MNP and the I-MNP preparations (15).
Cell Culture-Human U-937 cells (ATCC number CRL-1593.2), THP-1 cells (ATCC number TIB-202), and HL60 cells (University of Colorado Cancer Center, University of Colorado Denver, Aurora, CO) were grown in RPMI 1640 medium with 10% fetal bovine serum and 1ϫ antibiotic/antimycotic mixture (Invitrogen) at pH 7.4 and 37°C as indicated by the supplier. Cells were seeded in 12 well plates at 1 ϫ 10 6 cells well Ϫ1 . NR8383 normal rat lung alveolar macrophage cell line was obtained from ATCC (CRL-2192) and grown exactly as specified by the supplier. The adherent RAW264.7 mouse monocyte/macrophage cell line was obtained from ATCC (TIB-71) and grown in DMEM with 10% fetal bovine serum and 1ϫ antibiotic/antimycotic mixture (Invitrogen) at pH 7.4 and 37°C as indicated by the supplier.

SDS-PAGE and Western
Immunoblot Analysis-Whole cell lysates were either prepared as described previously (15) and the protein concentration determined by using the bicinchoninic acid assay (Sigma) or cells were disrupted in 2ϫ Laemmli Sample Buffer (catalog no. 161-0737, Bio-Rad) and sonicated on ice twice for 15 s, and protein concentration was determined using Pierce 660 nm protein assay (Thermo Scientific, Rockford, IL). Aliquots containing 50 g of protein were incubated with equal amounts of loading buffer (5% ␤-mercaptoethanol, 95% Laemmli loading buffer) for 10 min at 37°C and then boiled for 5 min. Samples were then separated by electrophoresis on 7.5% SDS-PAGE or 4 -15% gradient SDS-polyacrylamide gels for 40 min at 100 V and transferred to PVDF membranes (Whatman). Membranes were blocked overnight at 4°C in 5% nonfat dried milk in Tris-buffered saline, pH 7.6, containing 0.1% Tween 20. Membranes were then incubated with antibodies as indicated. Antigen-antibody complexes were detected by reaction with a chemiluminescent HRP Western blotting detection kit according to manufacturer's instruction (Millipore Corp.). Experiments were run in triplicate unless otherwise indicated, and representative immunoblots are shown. Band intensities were quantitated by scanning dosimetry with ImageQuant software.
Flow Cytometry-RAW264.7 cells were grown as indicated, shifted into growth medium containing charcoal-stripped and heat-inactivated FBS, and treated with 10 ng/ml IL-4 1 h after exposure to oxonic acid and/or uric acid at the indicated doses. Cells were harvested 24 h after treatment. Treated cells were collected into FACS running buffer (PBS with 2 mM EDTA and 0.5% BSA), washed twice in PBS, and were then fixed in 4% paraformaldehyde for 10 min. Cells were stained with anti-CD206 (50 ng/10 5 cells) and anti-CD36 (2 ng/10 5 cells) fluorescently conjugated primary antibodies in 200 l running buffer for 20 min on ice. Samples were washed and data acquired on a MACS Quant flow cytometer (Miltenyi Biotec).
XOR Quantitation-Total XOR activity was determined at 293 nm by quantitation of oxypurinol-inhibitable uric acid synthesis using a Beckman-Coulter DU640 spectrometer, 1-cm path length, and a molar extinction coefficient for uric acid of 12,600 M Ϫ1 as described previously (44). All XOR activity data show nanomoles of uric acid/min/mg of total protein.
Statistical Analyses-Data are expressed as the mean Ϯ S.E. and were assessed for significance using the Student's t test or analysis of variance. A p value of Ͻ0.05 was considered significant.

XOR Activity Is Increased in A-MNP and I-MNP following
Insufflation of Th-1 Cytokines in Vivo-XOR activity is rapidly increased in MNP purified from the bronchoalveolar la-vage of rats insufflated with the Th-1 cytokines IL-1␤ and IFN-␥ (15). In this study, we found extremely low levels of XOR activity in RAM prior to cytokine insufflation and undetectable levels in CM from the same rats (Fig. 1A). In contrast, both alveolar MNP (A-MNP) and I-MNP purified from rat lungs 24 h after Th-1 cytokine insufflation exhibited high levels of XOR activity (Fig. 1A).
Because MNP XOR activity was increased in both A-MNP and I-MNP populations and was detected previously as early as 4 h following cytokine insufflation, we hypothesized that FIGURE 1. Inflammatory MNP express high levels of XOR activity. A, CM and RAM were purified from control untreated rats, and A-MNP and I-MNP were purified from rat lungs 24 h after insufflation of the Th-1 cytokines, IL-1␤ and INF␥. XOR activity was determined from six preparations of each. Data show the mean Ϯ S.D. of XOR activity. Oxypurinol (150 M) was included in separate assays to confirm the specificity of uric acid formation. B, RAM were purified from control untreated rats and placed into 12-well culture plates. After 1 h, cells were washed and either harvested immediately (Control, 0 h) or exposed to vehicle, IL-1␤ (10 ng/ml), IFN-␥ (20 ng/ml), or LPS (1.0 g/ml) in heat-inactivated medium. Cells were harvested 24 h later, and XOR activity was determined. Data show the mean Ϯ S.D. for six determinations. C, NR8383 rat AM cell line was cultured in 12-well plates at 1.0 ϫ 10 6 cells/well and treated with vehicle, IL-1␤ (10 ng/ml), IFN-␥ (20 ng/ml), or LPS (1.0 g/ml) as in B. Cells were harvested 24 h later, and XOR activity was determined. Data show the mean Ϯ S.D. for six determinations.
XOR activity may be stimulated either by MNP differentiation per se or by exposure of RAM cells to Th-1 cytokines. To determine whether XOR activity was increased in RAM cells by Th-1 cytokines, RAM cells were purified from the lungs of native, untreated rats and exposed to Th-1 cytokines in vitro for 24 h. In addition, XOR activity was measured in Th-1 cytokine-stimulated normal rat alveolar macrophage cell line NR8383. The NR8383 cell line was derived from the rat RAM cell population and has been reported to exhibit many characteristics of normal RAM cells (45). We observed ϳ10-fold lower stimulation of XOR activity in RAM cells by the Th-1 cytokines IL-1␤, IFN-␥, IL-1␤/IFN-␥, or LPS ( Fig. 1B; Table  1) than was observed in A-MNP or I-MNP from cytokineinsufflated rats. Furthermore, Th-1 cytokine stimulation of NR8383 also exhibited ϳ10-fold lower stimulation of XOR activity than in A-MNP or I-MNP ( Fig. 1C; Table 1).
XOR Activity Is Increased by MNP Differentiation in Vitro-To determine whether XOR activity was increased by differentiation of MNP per se, we examined the effect on XOR of differentiation induced in U937 human myeloid cells in vitro. Unlike RAM that are fully differentiated macrophages, U937 cells exhibit monocyte-like characteristics and differentiate into macrophages by treatment with PMA or vitamin D 3 / TGF␤ (46). Exposure to all-trans-retinoic acid promotes differentiation along the granulocytic lineage (47). In addition, because exposure of circulating monocytes to the chemokine MCP-1 is an early event in recruitment to an inflammatory site that may be involved in differentiation along the MNP lineage (48,49), we treated U937 cells with MCP-1 as well. We observed that XOR activity was increased in U937 cells by treatment with PMA, D3/TGF␤, or MCP-1 but not by exposure to all-trans-retinoic acid ( Fig. 2A). Increased XOR activity in U937 cells following PMA or MCP-1 treatment was both dose-and time-dependent (Fig. 2, B and C). Thus, induction of macrophage differentiation or exposure to the MNP chemokine MCP-1 in human U937 cells increases XOR activity. These data demonstrate that XOR activity is increased in the inflammatory MNP by insufflation of Th-1 cytokines and that differentiation per se and the MNP chemokine MCP-1 may both contribute to the increased XOR activity observed in the inflammatory MNP.
XOR Inhibition Reduces Levels of CINC-1 Secreted by Inflammatory MNP-To determine whether XOR contributes to the inflammatory state of MNP, we quantitated levels of CINC-1 and MCP-1 in rats insufflated with Th-1 cytokines. We observed that levels of CINC-1 and MCP-1 were elevated

TABLE 1 XOR activity in RAM and NR8383 cells treated with Th-1 cytokines as described in the legend to Fig. 1
Control in both the serum-and cell-free lavage 5 h after insufflation of Th-1 cytokines compared with levels measured from salineinsufflated sham control rats (Fig. 3A). Inflammatory MNP harvested from the lungs of Th-1 cytokine-insufflated rats were cultivated in vitro, and levels of CINC-1, MCP-1, and IL-10 secreted into the culture medium were measured over  (19,36) is mediated in part by the peroxisome proliferator activator receptor-␥ (50), with high PPAR␥ promoting the M2 state. We purified RAM from untreated control rat lungs and both A-MNP and I-MNP from rats treated with Th-1 cytokines 24 h before. Western immunoblot of whole cell extracts revealed immunoreactive bands at 50/55, 62/67, and 78 kDa that were reactive to the anti-PPAR␥ antibody (Fig. 4A). Immunoreactive bands at 50/55 and 62/67 kDa were reduced in both A-MNP and I-MNP compared with the RAM. Immunoreactive bands at 50/55, 62/67, and 78 kDa were all reduced by inclusion of purified recombinant PPAR␥ in the Western immunoblot (data not shown). A similar pattern of immunoreactivity was obtained for several cultured myeloid lineage cells by reaction to the PPAR␥ antibody (Fig. 4B). Immunoreactive bands at 50/55 and 62/67 kDa were predominant, with little evidence of a 78-kDa band, in each of the cultured MNP cell lines examined. U937 and THP-1 cells expressed similar levels of the bands at 50 and 55 kDa. Treatment of RAW264.7 cells with IL-4 resulted in marked increase in the immunoreactive band migrating at 62/67 kDa (Fig. 4C), suggesting that it may constitute a modified form of PPAR␥ induced by IL-4.

Pattern of PPAR␥ Immunoreactive Material Reflects Both the Presence of PPAR␥ Isoforms and PPAR␥ Sumoylation-
PPAR␥ is encoded by a single gene in rats, mice, and humans that yields a primary translation product of 505 amino acids and 55 kDa in mass (51). Cleavage of 30 amino acids from the amino terminus of PPAR␥2 (55 kDa) produces the PPAR␥1 isoform (50 kDa). Our data indicate that U937 and THP-1 cells express both of these isoforms. PPAR␥ is subject to posttranslational sumoylation at two sites in the primary translation product as follows: at Lys-107 in PPAR␥2 and Lys-77 in PPAR␥1 and at Lys-395 in PPAR␥2 and Lys-365 in PPAR␥1 (51). Although multiple sumoylation events of PPAR␥ have been described that may contribute to the increased mass of PPAR␥ (51), a single sumoylation adduct can add 12 kDa to the mass (52). To determine whether the immunoreactive bands at 62/67 kDa reflect the presence of sumoylated PPAR␥, we transfected RAW264.7 cells with anti-SUMO-1 siRNA or siRNA for the PIAS SUMO-1-conjugating enzyme (52,53). Cells were then treated with either IL-4, PMA, and oxypurinol or PMA and Y-700 (as described in Fig. 6). Western immunoblots were performed on WCE 24 h after treatment using antibodies to PPAR␥ (Fig. 5). Scanning dosimetry revealed that exposure to anti-SUMO-1 siRNA reduced the immunoreactive band at 62/67 kDa relative to the scrambled siRNA control by 50%, although exposure to anti-PIAS siRNA reduced the same band by 75% (Fig. 5, A and C). Furthermore, exposure of RAW264.7 cells to anti-SUMO-1 siRNA followed by treatment with IL-4 decreased both the 62/67-kDa immunoreactive band and decreased the free SUMO-1 band at 12 kDa (Fig. 5B). These data indicate that the anti-PPAR␥ immunoreactive band at 62/67 kDa contains a SUMO-1 adduct.
XOR Inhibitors Increase Levels of SUMO-PPAR␥ in Cultured Myeloid Lineage Cell Lines-To determine the effect of XOR inhibition on PPAR␥ or SUMO-PPAR␥ (S-PPAR␥) in myeloid lineage cell lines, we exposed human U937, THP-1, and HL60 cells and the mouse RAW264.7 MNP cell line to the XOR inhibitors Y-700 or oxypurinol. After 1 h, cells were exposed to 30 nM PMA, and 24 or 48 h later cell lysates were prepared. Although Western immunoblot analysis showed increases in PPAR␥ by PMA alone (Fig. 4B), we observed very little effect of PMA alone on SUMO-PPAR␥ in U937, THP-1, HL60, or RAW264.7 cells (Fig. 6). However, 24 h after treatment, U937, THP-1, HL60, and RAW264.7 cells showed marked increase in levels of SUMO-PPAR␥ by both Y-700 and oxypurinol compared with control cells exposed to PMA alone. The increase in SUMO-PPAR␥ levels by XOR inhibitors occurred even in the absence of PMA stimulation. The increase in SUMO-PPAR␥ levels observed at 24 h in the presence of XOR inhibitors was frequently diminished by 48 h, although the basis for this effect is not presently understood.
Both Ectopic Overexpression of XOR cDNA and Uric Acid Supplementation Reduce SUMO-PPAR␥ in Stimulated Myeloid Lineage Cells-Further evidence that XOR was involved in modulating 62/67-kDa SUMO-PPAR␥ was obtained in three ways. First, to determine whether XOR overexpression would block IL-4 induced SUMO-PPAR␥, RAW264.7 cells were transfected with pCMV-Myc-XOR at various doses of input DNA. pCMV-Myc-XOR is a cDNA expression vector containing a functional XOR cDNA that is expressed under the control of the CMV constitutive promoter and produces ectopic overexpression of XOR (54). We observed that ec-topic overexpression of XOR dose-dependently reduced levels of SUMO-PPAR␥ in IL-4-treated RAW264.7 cells (Fig. 7A). Next, RAW264.7 cells were treated with IL-4 in the presence or absence of uric acid, the immediate product of XOR catalysis. Uric acid dose-dependently reduced IL-4-stimulated SUMO-PPAR␥ at 62/67 kDa (Fig. 7B). Finally, U937 cells were exposed to PMA/Y-700 in the presence or absence of uric acid. Again, uric acid dose-dependently reduced 62/67-kDa SUMO-PPAR␥ induced in U937 cells by PMA/Y-700 (Fig. 7C).
Arginase-1 Levels Are Modulated by Uric Acid, Oxonic Acid, and Anti-SUMO-1 or Anti-PIAS siRNA in RAW264.7 Cells-Induction of arginase-1 by IL-4 may be specifically mediated by sumoylation of PPAR␥ (36,55), and modulation of IL-4induced PPAR␥ sumoylation by uric acid (shown here) specifically predicts that uric acid will inhibit IL-4-induced arginase-1. This prediction was tested in IL-4-treated RAW264.7 cells. IL-4 was found to increase arginase-1 levels in a dosedependent fashion to 20 ng/ml, the highest dose tested (Fig.  8A). Transfection of RAW264.7 cells with anti-SUMO-1 or anti-PIAS siRNA reduced both the levels of 62/67-kDa SUMO-PPAR␥ and arginase-1 (Fig. 8B), demonstrating that IL-4-induced arginase-1 was dependent on sumoylation. To determine whether XOR overexpression would also block IL-4 induced arginase-1, RAW264.7 cells were transfected with pCMV-Myc-XOR at various doses of input DNA and treated with IL-4 for 24 h. Ectopic overexpression of XOR cDNA dose-dependently reduced arginase-1 levels in IL-4 cells (Fig. 8C). The effect of uric acid on IL-4-induced arginase-1 was tested in two ways. In the first, the enzyme uricase that degrades uric acid to allantoin was inhibited with oxonic acid, which then acts to elevate uric acid. Oxonic acid dosedependently inhibited IL-4-induced arginase-1 (Fig. 8D). In the second, uric acid was added exogenously to IL-4-treated RAW264.7 cells in the presence of 50 M oxonic acid to block uric acid degradation. Uric acid dose-dependently decreased levels of IL-4-induced arginase-1, an effect that was most pronounced at the higher doses of uric acid (Fig. 8E). These data demonstrate that ectopic overexpression of XOR or uric acid  Fig. 6. Western immunoblots were performed on WCE 24 h after treatment using antibodies to PPAR␥. B, U937 cells were plated at 0.5 ϫ 10 6 cells/well on 12-well plates and the next day were either left untreated (control no siRNA) or transfected with anti-SUMO-1 siRNA(siRNA SUMO-1) or nonspecific DNA (siRNA reagent control). After 24 h, cells were treated with PMA, and oxypurinol and Western immunoblots were performed on WCE 24 h after treatment using antibodies to PPAR␥ or SUMO-1. C, bands at 62/67 kDa were quantitated by scanning dosimetry and then averaged over all treatments (IL-4, Oxy/PMA, and Y-700/PMA). Data were normalized to the untreated control samples and shown as % of the control band. Data show the normalized average and standard deviation for six determinations in each treatment group. **, p Ͻ 0.05; ***, p Ͻ 0.01 by Student's t test comparing anti-SUMO-1 or anti-PIAS to scramble siRNA transfection. reduced levels of arginase-1 induced by IL-4 treatment through a sumoylation-dependent process.
M2 Markers CD36 and CD206 Are Also Modulated by Uric Acid and Oxonic Acid-CD36 and CD206 are additional markers of the alternative (M2) state of macrophage activation (36). RAW264.7 cells were exposed to oxonic acid or to uric acid in the presence of oxonic acid and were then treated with IL-4. Cells were analyzed by FACS 24 h later. As observed for levels of arginase-1, both oxonic acid and uric acid dose-dependently reduced levels of CD36 and CD206 detected by FACS (Fig. 8F).

HIF-1␣ Is Increased by XOR Inhibitors in Inflammatory MNP and in Cultured Myeloid Lineage Cells-
The hypoxia inducible factor-1␣ (HIF-1␣) is a transcription factor mediating both glycolytic and hypoxic effects in leukocytes. HIF-1␣ is subject to sumoylation that may affect both its stability and its transcriptional activity (56,57). To determine whether XOR also modulated levels of HIF-1␣ in inflammatory MNP, we examined levels of HIF-1␣ in I-MNP purified from Th-1 cytokine-insufflated lungs by Western immunoblot. HIF-1␣ was undetectable in I-MNP immediately after isolation. However, HIF-1␣ levels in I-MNP showed both time-and dose-dependent accumulation when cells were treated in vitro with the proteosome inhibitor MIG132 (Fig. 9A). MIG132-dependent accumulation of HIF-1␣ in I-MNP occurred during both normoxic (21% O 2 ) and hypoxic (1% O 2 ) culture in vitro (Fig. 9B). Concurrent treatment of I-MNP with MIG132 and three different inhibitors of XOR (allopurinol, oxypurinol, or Y-700) resulted in marked increase in HIF-1␣ levels over MIG132 treatment alone for cells cultured in either normoxia or hypoxia (Fig. 9B). These data suggest that XOR may exert a suppressive or modulating effect on levels of HIF-1␣ in I-MNP purified from Th-1 cytokine insufflated lungs. Furthermore, exposure of I-MNP to Th-1 cytokines IL-1␤, IFN-␥, or LPS showed increased levels of HIF-1␣ in the presence of Y-700 that did not arise in the absence of Y-700 (Fig. 9C).
To determine whether XOR exerted a similar modulating effect on HIF-1␣ in PMA-differentiated U937 cells, U937 cells were differentiated with PMA and after 48 h treated with MIG132 and Th-1 cytokines in the presence or absence of Y-700. We observed that HIF-1␣ was nearly undetectable in differentiated and normoxically grown U937 cells treated with Th-1 cytokines in the absence of Y-700, although concurrent treatment with Y-700 resulted in increased levels of HIF-1␣ (Fig. 9D). Growth of the same cells in hypoxia resulted in significantly higher levels of HIF-1␣ for cells grown in the presence of Y-700 (Fig. 9D). Thus, concurrent treatment of PMAdifferentiated U937 cells with Th-1 cytokines in the presence of the XOR inhibitor Y-700 resulted in increased levels of HIF-1␣ that were not observed in the absence of Y-700. These data suggest that XOR may also exert a modulating effect on HIF-1␣ levels in PMA-differentiated and Th-1 cytokinetreated U937 cells as well.
Vascular endothelial growth factor (VEGF) is an important target of HIF-1 regulation. We observed that I-MNP showed marked accumulation of VEGF when cultured in vitro in normoxia over the course of 48 h (Fig. 9E). Cells exposed to MIG132 in the presence or absence of XOR inhibitors for either 6 or 24 h showed a marked decrease in VEGF secretion. Because the profound suppression of VEGF by MIG132 may have obscured the effects of XOR inhibitors, cells were treated with Y-700 in the absence of MIG132, and secreted VEGF was measured in the cell-free culture medium. Y-700 alone markedly suppressed secretion of VEGF.

DISCUSSION
XOR plays a central role in the function of innate immunity and inflammation, and it is an important mediator of many inflammatory diseases in humans (37,58). XOR is up-regulated in many inflammatory models (15,18) where it shows particularly high levels of activity in the inflammatory MNP. Inhibition of XOR specifically reduces PMN recruitment to an inflammatory site in vivo (15,59), and inhibition of XOR specifically in the MNP, but not the PMN, modulates subsequent PMN recruitment (15,60). Thus, XOR contributes to FIGURE 7. Ectopic overexpression of XOR or uric acid supplementation both reduce levels of SUMO-PPAR␥ in cultured myeloid lineage cell lines. A, RAW264.7 cells were plated in 6-well plates at 0.5 ϫ 10 6 cells/well, and the next day were transfected with the indicated doses of pCMV-Myc-XOR. After 24 h, cells were shifted into standard medium with charcoalstripped and heat-inactivated FBS and treated with recombinant mouse IL-4 (10 ng/ml). Whole cell lysates were prepared 24 h after exposure to IL-4, and Western immunoblots were run sequentially with antibody to PPAR␥ and GAPDH. Representative blots are shown of three independent blots for each experiment. B, RAW264.7 cells were plated in 6-well plates at 1.0 ϫ 10 6 cells/well. The next day cells were shifted into standard medium with charcoal-stripped and heat-inactivated FBS and treated with uric acid at the indicated doses in the presence of 50 mM oxonic acid. After 1 h cells were treated with 10 ng/ml IL-4. Cells were harvested 24 h later, lysates prepared, and Western immunoblots performed sequentially against PPAR␥ and GAPDH. C, U9376 cells were plated at a density of 1 ϫ 10 6 cells/well in 6-well plates. Cells were then treated with uric acid at the indicated doses and after 1 h were exposed to PMA and Y-700 exactly as described in Fig. 6. Cell lysates were prepared after 24 h, and Western immunoblots were performed sequentially against PPAR␥ and GAPDH. the development of inflammation as a product of the newly recruited inflammatory MNP. This observation is important because the MNP themselves contribute to injury of the lung epithelium, and XOR inhibitors may provide critical support of the inflamed lung (15,21,22). Despite the vast amount of literature published on the role of XOR in inflammatory disease and the broadly protective role afforded by XOR inhibitors, surprisingly little is known about the mechanisms by which XOR contributes to inflammation. Data shown in this study demonstrate that XOR is both activated by MNP differentiation and contributes to the inflammatory state of the MNP. XOR was found to promote the inflammatory state of the MNP by contributing to inflammatory chemokine secretion, regulation of PPAR␥ sumoylation, and HIF-1␣ stability.
XOR activity was markedly elevated in both A-MNP and I-MNP following insufflation of Th-1 cytokines in contrast to the very low activity observed in CM or RAM from untreated FIGURE 8. Levels of M2 markers arginase-1, CD36, and CD206 are reduced by uric acid or oxonic acid. A, RAW264.7 cells were plated at 0.5 ϫ 10 6 cells/ well and the next day treated with recombinant mouse IL-4 from 0 to 20 ng/ml. WCE were prepared 24 h later, and Western blots were performed with antisera to arginase-1. B, RAW264.6 cells were plated as in A, and the next day were either left untreated or were transfected with scrambled siRNA, anti-SUMO-1 siRNA, or anti-PIAS siRNA. After 24 h, cells were treated with IL-4, and 24 h later WCE were prepared. Western immunoblots were performed using antisera to PPAR␥, arginase-1, or GAPDH. C, RAW264.6 cells were plated as in A and the next day were either left untreated or were transfected with pCMV-Myc vector alone or with pCMV-Myc-XOR in the presence of pCMV-Myc to fix the net input DNA at 3.0 g. After 24 h, cells were treated with IL-4 and 24 h later WCE prepared. Western immunoblots were performed using antisera to arginase-1 or GAPDH. Arginase-1 bands were quantitated by scanning dosimetry and normalized to the untransfected cells treated with IL-4. D, RAW264.7 cells were plated as in A and the next day treated with oxonic acid at the indicated doses. After 1 h, cells were treated with IL-4 and WCE prepared 24 h later. Western immunoblots were performed using antisera to arginase-1 and GAPDH. E, RAW264.7 cells were plated as in A and the next day treated with uric acid at the indicated doses in the presence of oxonic acid (50 M). After 1 h cells were treated with IL-4 and WCE prepared 24 h later. Western immunoblots were performed using antisera to arginase-1 and GAPDH. F, RAW264.7 cells were plated as in A, shifted into charcoal-stripped/heat-inactivated medium, and treated with the indicated doses of oxonic acid (O.A.) alone or in combination with uric acid (U.A.) at the indicated doses. One h later, cells were treated with 10 ng/ml IL-4. Cells were harvested 24 h after treatment and analyzed by flow cytometry using antibodies to CD36 (phycoerythrin (PE)-conjugated) and CD206 (FITC-conjugated).

FIGURE 9. Inhibition of XOR activity increases levels of HIF-1␣ protein in rat inflammatory MNP and in PMA-differentiated U937 cells.
A, I-MNP were purified from rat lungs 24 h following insufflation of Th-1 cytokines and were plated in 12-well plates at 1.0 ϫ 10 6 cells/well and grown under normoxic conditions. Nonadherent cells were removed by washing after 1 h, and cells were treated with the indicated doses of MIG132. Cells were harvested after 6 h; whole cell lysates were prepared and Western immunoblots run with antibody to HIF-1␣. In addition, washed, adherent cells were exposed to MIG132 (50 M) for the indicated times, and Western immunoblots were run on whole cell lysates and probed with antibody to HIF-1␣. Blots were subsequently stripped and re-probed with antibody to GAPDH to control for protein loading. B, I-MNP were purified and plated as in A. Plates were placed in either normoxic (21% O 2 , 5% CO 2 ) or hypoxic culture (1% O 2 , 5% CO 2 , 94% N). After 1 h, cells were treated with the XOR inhibitors allopurinol (Allo, 150 M), oxypurinol (Oxy, 150 M), or Y-700 (50 nM). One h later cells were treated with MIG132 (50 M) and grown for 6 h. Whole cell lysates were then prepared and Western immunoblots run with independent duplicate samples as indicated. Blots were first probed with antibody to HIF-1␣ and subsequently with antibody to GAPDH. Bands from Western immunoblots were quantitated by scanning dosimetry and normalized to the signal obtained from the GAPDH blots. Data show the mean Ϯ S.D. of duplicate samples. C, I-MNP were purified, plated, and grown in normoxia in the presence of MIG132 as in A. Cells were exposed to the XOR inhibitor Y-700 (50 nM) for 1 h and subsequently treated with IL-1␤ (10 ng/ml), IFN-␥ (20 ng/ml), or LPS (1.0 g/ml). Whole cell lysates were prepared and Western immunoblots run after 24 h of exposure to cytokines. Blots were run on triplicate samples, and representative blots are shown. Bands from Western immunoblots were quantitated by scanning dosimetry and normalized to the signal obtained from the GAPDH blots. Data show the mean Ϯ S.D. of triplicate blots. D, U937 cells were plated in 12-well plates at 1 ϫ 10 6 cells/well and treated with PMA (30 nM) for 48 h. Cells were then washed; the medium was replaced, and cells were grown under normoxic or hypoxic conditions for 1 h in the presence of 50 M MIG132. Subsequently, cells were treated with Y-700 (50 nM) for 1 h and cytokines added as above. Whole cell lysates were prepared after 24 h of exposure to cytokine/MIG132, and Western immunoblots were run sequentially with antibody to HIF-1␣ and GAPDH. Representative blots are shown of three independent blots for each experiment. Bands from Western immunoblots were quantitated by scanning dosimetry and normalized to the signal obtained from the GAPDH blots. Data show the mean Ϯ S.D. of triplicate blots. E, I-MNP were purified and plated in normoxia as in B. Levels of VEGF were quantitated in the cell-free supernatant over a period of 48 h in culture. In addition, cells were treated independently with MIG132 in the absence or presence of allopurinol, oxypurinol, or Y-700 as in B, and levels of VEGF were measured in the cell-free supernatant 6 or 24 h after treatment. Data show the mean Ϯ S.D. of three independent experiments. F, I-MNP were plated as above in the presence or absence of Y-700 (1.0 mM) and in the absence of MIG132. After 24 h, cell-free supernatants were collected and analyzed with the rat-specific proteome profiler. Experiments were performed in quadruplicate, and each cytokine was analyzed from two spots on each filter (boxed for VEGF), and thus all data reflect eight independent determinations for VEGF alone. Spots were quantitated as described by the supplier using an R&D transmission mode scanner and image analysis software from R&D. Data show the mean Ϯ S.D. of light transmission signals (arbitrary units) from eight spots for VEGF only and both control and Y-700 groups and were normalized first to the mean positive control spots (A1, A2, A19, A20, D1, and D2) and subsequently to the signal obtained from the control samples, which was thereby set at 1.00. ***, p Ͻ 0.02. control rats. Although XOR activity was increased by treatment of RAM or NR8383 normal rat macrophages with Th-1 cytokines, XOR activity failed to reach the levels observed in inflammatory MNP isolated from Th-1 cytokine-insufflated rats. Thus, exposure of RAM to Th-1 cytokines alone could not account for the level of XOR activity observed in the inflammatory MNP. Furthermore, XOR activity was increased by agents promoting macrophage, but not granulocyte, differentiation in U937 cells. Treatment with vitamin D 3 /TGF␤ or PMA increased XOR activity to levels approaching those observed in MNP from Th-1 cytokine-insufflated rats. XOR activity was also increased by treatment with MCP-1, and this is important because MCP-1 is a central chemokine mediating the recruitment of monocytes to an inflammatory site, including the lung (26,61), and may contribute to the inflammatory state of MNP (19,62,63). These data demonstrate that XOR activity is markedly increased by agents promoting MNP differentiation along an inflammatory pathway. Evidence that XOR contributes to the inflammatory state of MNP was derived in several ways. Insufflation of Th-1 cytokines in rats produced a rapid rise in both serum and lavage chemokines for PMN (CINC-1) and MNP (MCP-1). Both CINC-1 and MCP-1 were secreted by the MNP to a much greater extent than by the PMN isolated from the same rats (data not shown), and chemokine secretion was sustained after placing the MNP in culture. Significantly, CINC-1 levels were markedly elevated by treatment of the cultured MNP with LPS, and both uninduced and LPS-induced CINC-1 was significantly reduced by co-incubation with XOR inhibitors. These data are consistent with and extend both our previous observations demonstrating that adoptive transfer of XORinhibited MNP reduces PMN influx into the rat lung (15) and with data obtained in mice demonstrating the key role played by MNP in recruitment of PMN to the lung during inflammation (64 -66). Consistent with the role of XOR in promoting the inflammatory state of MNP, we observed that the antiinflammatory cytokine IL-10 was not modulated by inhibition of XOR.
Further evidence that XOR contributes to the inflammatory state of MNP was derived by analysis of PPAR␥. PPAR␥ is a dual-function, ligand-activated member of the nuclear hormone receptor superfamily of transcription factors. Transcriptional activation by PPAR␥ requires both ligand binding and interaction with the retinoid X receptor. Although transcriptional repression of inflammatory gene expression by PPAR␥ is well recognized, it is not yet fully understood (67,68). PPAR␥ expressed in the MNP plays an important role in regulating lung inflammation (69,70). RAM express high levels of PPAR␥, and this is down-regulated by Th-1 cytokines (69). Down-regulation of MNP PPAR␥ is associated with enhanced lung inflammation and injury (71)(72)(73), whereas activation or induction of MNP PPAR␥ is associated with an antiinflammatory state characterized by reduced leukocyte infiltration, reduced inflammatory cytokine expression, and enhanced expression of anti-inflammatory cytokines (50, 74 -76). We observed an unexpectedly complex pattern of immunoreactive material on PPAR␥ Western blots of freshly isolated rat MNP and from several cultured human and rodent MNP cells lines, including U937, THP-1, HL-60, and RAW264.7 cells. Freshly isolated MNP from the rat exhibited immunoreactive material at 78, 62/67, and 50/55 kDa using an antibody to PPAR␥, whereas we observed immunoreactive material predominantly at 62/67 and 50/55 kDa in the cultured cell lines. In rats, mice, and humans, PPAR␥ is encoded by a single gene that yields a primary translation product of 505 amino acids and 55 kDa in mass (51), and cleavage of 30 amino acids from the amino terminus of PPAR␥2 (55 kDa) produces the PPAR␥1 isoform (50 kDa). Treatment of RAW264.7 cells with IL-4 resulted in marked increase in immunoreactive material at 62/67 kDa suggesting that this higher molecular weight material reactive to the PPAR␥ antibody reflected a modified form of the primary translation product.
Sumoylation is one of a limited number of post-translational modifications that would increase apparent mass of PPAR␥ by 11 to 12 kDa (51). Sumoylation of PPAR␥ has been found to modulate its activity as a transcription factor (77,78), and sumoylation of PPAR␥ has been found specifically to mediate repression of pro-inflammatory genes (79 -81). We used siRNA technology to determine whether immunoreactive material at 62/67 kDa contained a SUMO adduct. Transient transfection of RAW264.7 cells with antisense RNA to both the PIAS-1-conjugating enzyme and SUMO-1 reduced immunoreactive material at 62/67 kDa that was induced in several ways, although antisense RNA to SUMO-2 did not affect this band (not shown). Levels of free SUMO-1 were likewise reduced by transfection with antisense RNA to SUMO-1. These data indicate that the immunoreactive material at 62/67 kDa obtained by reaction to antibody against PPAR␥ also contains a SUMO-1 adduct. Therefore, we infer that freshly isolated RAM from the rat express high levels of SUMO-PPAR␥ at 62/67 kDa and that the inflammatory MNP present in the lung following induction of inflammation express very low levels of SUMO-PPAR␥. These observations support the argument that SUMO-PPAR␥ expressed in the RAM modulates the inflammatory state of these cells, whereas the newly recruited inflammatory MNP express very little SUMO-PPAR␥ thereby contributing to their inflammatory state (79 -81).
Evidence that XOR was involved in PPAR␥ sumoylation was obtained in three ways. First, inhibitors of XOR, in the presence or absence of PMA, increased levels of SUMO-PPAR␥ at 62/67 kDa in U937, HL-60, THP-1, and RAW264.7 cells. The increase in 62/67-kDa immunoreactive material was obtained with both Y-700 and oxypurinol, and similar results were obtained with allopurinol as well (data not shown). The consistent effect of the different inhibitors is important because the inhibitors have either different ROS scavenging properties (allopurinol and oxypurinol) that are distinct from inhibition of XOR per se and potentially confounding, or they have no detectable ROS scavenging properties (Y-700) (41). Second, ectopic overexpression of XOR cDNA dose-dependently reduced levels of IL-4 induced 62/67-kDa SUMO-PPAR␥. Third, uric acid, the principal catalytic product of XOR dose-dependently reduced levels 62/ 67-kDa SUMO-PPAR␥ that were produced either by IL-4 stimulation of RAW264.7 cells or by treatment of U937 cells with Y-700 and PMA. In aggregate, these data demonstrate that XOR modulates levels of SUMO-PPAR␥-immunoreactive material at 62/67 kDa and identify uric acid as an important component of this effect.
Modulation of PPAR␥ sumoylation explicitly predicts an inhibitory effect on arginase-1, which is a canonical marker of alternatively activated (M2) macrophages that is specifically induced by IL-4 (36). We observed IL-4 dose-dependent increases in arginase-1 levels that were reduced by antisense RNA to both SUMO-1 and PIAS-1 using RAW264.7 cells, consistent with reports demonstrating the requirement of SUMO for PPAR␥ activation and PPAR␥ activation for induction of arginase-1 (36,55,68,77). Ectopic overexpression of XOR cDNA dose-dependently blocked IL-4-induced arginase-1. Furthermore, both uric acid and oxonic acid, a uric acid-elevating reagent, dose-dependently reduced arginase-1 levels that were induced with IL-4. Furthermore, we also ob-served decreases in the M2 markers CD36 and CD206 by both oxonic acid and uric acid. These data confirm the involvement of XOR and uric acid in modulating markers of M2polarized RAW264.7 cells.
The signaling pathways through which XOR activity modulates PPAR␥ sumoylation are unknown. XOR is an efficient source of both reactive oxygen and reactive nitrogen species, and both oxidative and nitrosative stress can regulate sumoylation (82)(83)(84). However, data shown here demonstrate that uric acid is at least one component of XOR biochemistry that modulates PPAR␥ sumoylation. XOR-derived uric acid may exert proinflammatory effects at several levels. On the one hand, uric acid can activate the inflammasome to promote inflammation (59,85), promote T cell activation (86), activate dendritic cells (87), and as shown here modulate PPAR␥ sumoylation. Importantly, as shown here, these effects can arise at physiological levels of uric acid that are well below the point of crystallization and may therefore be distinct from the response to monosodium urate crystals (88). Uric acid can also serve as a scavenger of ROS, and this too may underlie its effects on PPAR␥ sumoylation and inflammation (37).
HIF-1␣ plays an essential role in the development of inflammation as a product of the MNP (89 -91), and MNP HIF-1␣ is required for the expression of inflammatory cytokines in an LPS model of sepsis (90). Recent reports demonstrate that HIF-1␣ is also regulated by sumoylation. SUMO conjugation to HIF-1␣ both stabilizes the protein and reduces its activation function in transcription (56,57,92). Evidence that XOR is involved in regulating levels of HIF-1␣ protein was obtained using both I-MNP freshly isolated from Th-1 cytokine-insufflated rat lungs and U937 cells. In the presence of the proteosome inhibitor MIG132, XOR inhibitors increased both normoxic and hypoxic levels of HIF-1␣ and increased the levels of HIF-1␣ induced by treatment with LPS, IL-1, or IFN-␥. Again, the use of three different inhibitors of XOR suggests that the effect on HIF-1␣ protein levels is the result of XOR inhibition per se. We infer that in the uninhibited MNP XOR activity modulates levels of HIF-1␣ protein.
Experiments are currently underway to determine whether the effect of XOR on HIF-1␣ protein levels reflects the effect of XOR on sumoylation of HIF-1␣. However, consistent with this observation, we observed a marked decrease in VEGF, a principal target of HIF-1 regulation, in the presence of MIG132 at 6 and 24 h after exposure, when HIF-1␣ levels were most increased. Furthermore, inhibition of XOR in the absence of MIG132 alone reduced VEGF secretion from these cells.
Repair and rescue of the injured alveolar epithelium is a critical therapeutic objective for management of lung inflammatory disorders (93). Although the relatively poor understanding of repair and injury has limited current treatment strategies, MNP are of considerable interest as therapeutic targets because they contribute to both injury and repair. During inflammation classically activated (M1) inflammatory MNP contribute to lung injury and PMN recruitment. On the other hand, alternatively activated (M2) MNP exhibit antiinflammatory properties that may contribute to tissue repair. Experiments shown here demonstrate that XOR promotes the inflammatory state of MNP through effects on chemokine expression, PPAR␥ sumoylation, and HIF-1␣ and that inhibitors of XOR reduce the inflammatory state. Thus, strategies that inhibit XOR may be important considerations for modulating lung inflammatory disorders.