Xanthine Oxidase Mediates Axonal and Myelin Loss in a Murine Model of Multiple Sclerosis

Objectives Oxidative stress plays an important role in the pathogenesis of multiple sclerosis (MS). Though reactive oxygen species (ROS) are produced by various mechanisms, xanthine oxidase (XO) is a major enzyme generating ROS in the context of inflammation. The objectives of this study were to investigate the involvement of XO in the pathogenesis of MS and to develop a potent new therapy for MS based on the inhibition of ROS. Methods XO were assessed in a model of MS: experimental autoimmune encephalomyelitis (EAE). The contribution of XO-generated ROS to the pathogenesis of EAE was assessed by treating EAE mice with a novel XO inhibitor, febuxostat. The efficacy of febuxostat was also examined in in vitro studies. Results We showed for the first time that the expression and the activity of XO were increased dramatically within the central nervous system of EAE mice as compared to naïve mice. Furthermore, prophylactic administration of febuxostat, a XO inhibitor, markedly reduced the clinical signs of EAE. Both in vivo and in vitro studies showed infiltrating macrophages and microglia as the major sources of excess XO production, and febuxostat significantly suppressed ROS generation from these cells. Inflammatory cellular infiltration and glial activation in the spinal cord of EAE mice were inhibited by the treatment with febuxostat. Importantly, therapeutic efficacy was observed not only in mice with relapsing-remitting EAE but also in mice with secondary progressive EAE by preventing axonal loss and demyelination. Conclusion These results highlight the implication of XO in EAE pathogenesis and suggest XO as a target for MS treatment and febuxostat as a promising therapeutic option for MS neuropathology.


Introduction
Multiple sclerosis (MS) is an inflammatory demyelinating disease of the central nervous system (CNS) and is a leading cause of disability in young adults [1]. It is characterized by CNS autoimmunity that evokes a cascade of pathological processes, including demyelination and axonal injury, leading to impaired mobility in patients with MS [2,3]. Reactive oxygen species (ROS) have been suggested to play an important role in the myelin and axonal impairments in MS [4][5][6]. Indeed, various oxidative stress markers are expressed in brain tissues from MS patients, particularly in apoptotic oligodendrocytes, axonal spheroids and degenerating neurons [7][8][9]. The production of ROS by macrophages/microglia is also increased in the CNS of experimental model of MS, namely, autoimmune encephalomyelitis (EAE) [4,10]. In addition, several studies have reported that ROS scavengers, such as edaravone [11] and alpha-lipoic acid [12], produce therapeutic benefits in EAE, thereby supporting the notion that ROS play a role in the pathogenesis of MS.
In the context of inflammation, several distinct pathways can generate ROS and reactive nitrogen species (RNS); these include uncoupling of mitochondrial oxidative phosphorylation, induction of nitric oxide synthase (NOS), and the activation of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase and xanthine oxidoreductase (XOR) systems [13,14]. Among these catalyzing enzymes, the relationship between NADPH oxidase and ROS generation has been well characterized [15], with studies showing that NADPH oxidase is expressed in activated microglia and infiltrating macrophages in the active lesions of MS [16,17]. By contrast, relatively little is known concerning the relationship between another ROS generating enzyme, XOR, and CNS autoimmunity. Xanthine dehydrogenase (XDH) and xanthine oxidase (XO) are inter-convertible forms of XOR, and the best characterized function of both enzymes is the conversion of hypoxanthine to xanthine and the conversion of xanthine to uric acid (UA) [18]. In addition to its involvement in the purine metabolism, there is growing evidence that XOR plays more diverse pathophysiological roles through the generation of ROS. Indeed, XOR produces various types of ROS, including superoxide, hydrogen peroxide and nitric oxide, by distinct mechanisms, and these ROS are involved in several types of tissue injury, such as acute lung injury, ischemia and reperfusion injury, and atherosclerosis [19][20][21][22]. Febuxostat is a thiazolecarboxylic acid derivative that is a novel, selective, non-purine inhibitor of XOR. It has potent and well-tolerated effects on uric acid synthesis, and its use for the treatment of gout was recently approved in several countries, including the US and throughout the European Union [23,24]. Thus, the objectives of this study were to investigate the involvement of XOR in the pathogenesis of MS and to develop a potent new therapy for MS based on the inhibition of ROS.

Induction and Clinical Evaluation of EAE
EAE was induced in 8-week-old female SJL/J mice (Oriental Yeast Corp., Tokyo, Japan) by subcutaneous injection of 150 mg of PLP 139-151 (Biologica, Nagoya, Japan) peptide emulsified in complete Freund's adjuvant (CFA; Sigma, St. Louis, MO, USA ) supplemented with 400 mg of Mycobacterium tuberculosis (H37Ra strain; Difco Laboratories, Detroit, MI, USA) [11]. In non-obese diabetic (NOD)/ShiJcl mice, EAE was induced by immunization with MOG 35-55 peptide (Biologica) emulsified in CFA at a dose of 200 mg, followed by the administration of pertussis toxin (200 ng; List Biological Laboratories Inc., Colorado, USA) on days 0 and 2 post-immunization. For prophylactic treatment, mice were given febuxostat (0.75 or 1.5 mg/kg body weight per day; Teijin Pharma Ltd., Tokyo, Japan) from day 10 post-immunization to the end of the experiment. For therapeutic purpose and in NOD/ShiJcl mice, febuxostat was administered from day 20 post-immunization. Control animals were given water. Mice were monitored daily for clinical signs of EAE, graded on a scale from 0 to 5 as follows: 0, normal; 1, limp tail; 2, hind limb paresis; 3, hind limb paralysis; 4, paraplegia; 5, moribund [25]. All experimental procedures were approved by the Institutional Animal Care and Use Committee of Osaka University Graduate School of Medicine. All possible efforts were made to minimize animal suffering and limit the number of animals used.

Histology and Immunohistochemistry
At days 16 and 40 after immunization, mice were perfused with phosphate-buffered saline (PBS) containing 4% paraformaldehyde (PFA). Lumbar spinal cords were dissected and immersed in 4% PFA overnight and then transferred to 30% sucrose overnight and embedded in optimal cutting temperature (OCT) compound and frozen in liquid nitrogen. Frozen sections were cut at a thickness of 10 mm and were processed for immunofluorescence. Sections were incubated with one of the primary antibodies e.g., rabbit anti-iNOS (inducible nitric oxide synthase), mouse anti-Iba1, rabbit anti-XO (Abcam; Cambridge, MA, USA), mouse anti-GFAP (glial fibrillary acidic protein) combined with Alexa Fluor 433 (Cell Signaling Technology, Beverly, MA, USA), rat anti-F4/80 (eBiosciences, San Diego, CA, USA ) overnight at 4uC, followed by incubation with appropriate secondary antibodies. For histological and immunohistochemical analyses, the spinal cord was embedded in paraffin after perfusion with 4% PFA. Axonal loss and demyelination were assessed by staining with Bielschowsky's silver and Luxol fast blue, respectively. Semi-quantitative histological analysis of inflammatory cellular infiltration was performed according to the following scoring system: 0, no inflammation; 1, cellular infiltrates only in the perivascular areas and meninges; 2, mild cellular infiltration in the parenchyma; 3, moderate cellular infiltration in the parenchyma; 4, severe cellular infiltration in the parenchyma [26].

XOR Activity and Oxidative Stress Assay
Frozen tissues were homogenized with potassium phosphate buffer, pH 7.4, containing 1 mM ethylenediaminetetraacetic acid (EDTA) and protease inhibitors. The homogenates were centrifuged at 12,000 rpm for 15 min at 4uC, and supernatants were used for the assay. For XOR activity assay, supernatants were coincubated with 50 mM pterin solution or pterin with 50 mM methylene blue solution to assay XO or both XO and XDH activity, respectively. After 60 min incubation at 37uC, fluorometric assays were performed to calculate the production of isoxanthopterin in the absence or presence of febuxostat (1 mM) in order to subtract the background. XDH activity was determined by subtracting XO activity from XOR activity. Activity of purified XO derived from buttermilk (Calbiochem, Merck, Darmstadt, Germany) and protein concentration were measured and used to normalize the sample activity [29]. The measurement of thiobarbituric acid-reactive substances (TBARS) was based on the formation of malondialdehyde as assessed by the TBARS assay kit (Cayman Chemical) according to the manufacturer's instructions.

Plasma Uric Acid Level
Plasma uric acid levels were determined by an enzymatic method based on the uricase-peroxidase system (autosera S UA, Sekisui Medical Co., Ltd., Japan).

Statistical Analysis
Student's t-test was used for two groups comparisons (semiquantitative analysis of iNOS and XO staining, relapse rate, TBARS). One way ANOVA followed by the Tukey-kramer multiple comparisons test was used to compare results from Griess reaction, XO activity assay, iNOS relative expression. One way ANOVA and student t-test were used to compare data from clinical and histological scores.

Xanthine Oxidase Expression is Increased in the CNS of EAE
XOR activity is enhanced under various inflammatory conditions, leading to further tissue damage through ROS generation [18,22]. To examine whether XOR is involved in the pathogenesis of CNS autoimmunity, we investigated the expression of XOR in EAE mice. Immunohistochemical analysis revealed that the XO expression was first detected in a few cells around day 8 postimmunization in the spinal cords and had markedly increased at the clinical peak of EAE, while naïve mice showed no apparent   (Fig. 1A-1C). This trend was also confirmed by realtime PCR (Fig. S1). The majorities of XO-positive cells were round-shaped cells and were mainly found at the periphery of the spinal cord and the perivascular areas. Most of these cells were double-labeled with F4/80. Some of the Iba-1-positive processbearing cells, presumably microglia, were also double-labeled with anti-XO antibody (Fig. 1D-1I). A few of CD3-positive cells were also double-labeled with anti-XO antibody, but none of GFAPpositive astrocytes showed the immunoreactivity to XO (Fig. 1J-1O). These data indicate that the expression of XO is markedly up-regulated in the context of CNS autoimmunity and that XOpositive cells mainly consisted of infiltrating macrophages and microglia. These data also suggest the possible involvement of XOR in the pathogenesis of EAE.

Prophylactic Administration of Febuxostat Ameliorates EAE
Because the expression of XOR was increased in EAE, we examined whether the up-regulated XO contributes to the development of EAE. EAE mice were treated orally with a selective XOR inhibitor, febuxostat (0.75 mg/kg body weight), from day 10 post-immunization (before the onset of the disease) to the end of the experiment. The severity of the disease was significantly ameliorated in the febuxostat-treated group when compared with the control group ( Fig. 2A). High-dose febuxostat (1.5 mg/kg body weight) did not produce additional efficacy in terms of ameliorating the severity of the disease (Fig. S2). Histological examination of the spinal cords demonstrated that less inflammatory cellular infiltration was observed in febuxostattreated group when compared with control group. Semi-quantitative evaluation of the cellular infiltration in the spinal cords revealed a significant difference between the two groups (Fig. 2B). Moreover, the increased expression of XOR in the control EAE spinal cord was significantly inhibited by treatment with febuxostat (Fig. 2C). These results suggest that XOR plays a pivotal role in the pathogenesis of EAE.
Peripheral T cell responses were analyzed in both the control and the febuxostat-treated mice to examine whether febuxostat affects the peripheral immunity. No significant differences were observed between the two groups in the production of IFN-c and interleukin-17 (IL-17) after PLP 139-151 stimulation of T cells derived from the draining lymph nodes (Fig. S3).

Febuxostat Inhibits XOR Activity, Oxidative Stress and Glial Activation in the CNS of EAE
We next investigated whether febuxostat affects the redox environment in the CNS by inhibiting the XOR activity of the inflamed tissues. A marked increase in the activity of both XO and XDH was revealed in the CNS and in the serum of EAE mice when compared with naive mice. This increase was particularly observed in the cerebellum and the spinal cord, where the inflammatory cells had predominantly infiltrated (Fig. 3A). Febuxostat treatment resulted in significant inhibition of XOR activity in the CNS of EAE mice as well a reduction in XO activity in the serum (Fig. 3A, 3B) when compared with the control group. Serum uric acid level was shown to be elevated after the induction of EAE, and febuxostat-treatment significantly reduced the increase of uric acid (Fig. S4). Moreover, TBARS, which are well-established surrogate markers of oxidation, were decreased in the CNS of the febuxostat-treated group when compared with the control group (Fig. 3C).
In accordance with the amelioration of oxidative stress in the CNS, reduced glial activation suggested by GFAP-and Iba1immunoreactivities was observed. iNOS-positive cells, predominantly found in Iba1-expressing cells (Fig. S5), were also reduced in the febuxostat-treated group when compared with the control (Fig. 3D). The increase in levels of iNOS mRNA in the control EAE spinal cord was inhibited in the febuxostat-treated group, though this difference did not reach the level of statistical significance (Fig. 3E). These observations suggest that febuxostat reduces oxidative stress in the CNS by inhibiting XOR activity, leading to the amelioration of infiltrating macrophages and activated glial cells.

Febuxostat Inhibits the Activity of XOR in vitro
Previous studies have demonstrated that macrophages and microglia are important sources of ROS and that the oxidative stress induced by these cells is involved in the pathogenesis of EAE and MS [4,6]. When J774 macrophage cells were incubated with febuxostat under LPS and IFN-c stimulation, the inhibitory effect of febuxostat on NO production and iNOS induction was observed (Fig. 4A, 4B). Immunocytochemical analysis also showed that the number of iNOS-positive cells was reduced following incubation with febuxostat (Fig. 4B). The inhibitory effect of febuxostat on NO production and iNOS induction was also observed in the microglial GMI-M6-3 cells (Fig. 4C, 4D). The inhibition of NO production by febuxostat was dose-dependent (Fig. 4C). Notably, the suppressive effect of febuxostat on redox responses in J774 cells and GMI-M6-3 cells correlated well with the reduction of XOR expression in these cells (Fig. 4E). These results suggest that febuxostat reduces ROS production by infiltrating macrophages and activated microglia in accordance with a decrease in XOR expression.

Therapeutic Application of Febuxostat Ameliorates Both Relapsing-remitting and Secondary Progressive EAE
As we have shown that prophylactic administration of febuxostat ameliorates EAE, we further examined whether it has a therapeutic effect when administered after establishment of symptoms. Administration of febuxostat to EAE mice was started on day 20 post-immunization. The febuxostat-treated group showed significantly less severe symptoms when compared with the control group. In addition, the treated group showed less frequency of relapse (Fig. 5A, 5B). Thus, febuxostat also has a therapeutic effect on the relapsing-remitting form of EAE.
Extensive axonal degeneration and demyelination constitute major determinants of progressive neurological disability in MS, which is coined as a progressive form of the disease. Therefore, we investigated whether febuxostat could also have a therapeutic effect on disease progression in the secondary progressive form of EAE. Induction of EAE in NOD mice with MOG peptide showed a relatively short period of acute CNS inflammation followed by a progressive disease course in the control group, which recapitulates the secondary progressive form of MS [30,31]. Administration of febuxostat from day 20 post-immunization resulted in a significant reduction in the severity of the disease in febuxostattreated group compared with the control group (Fig. 5C). Histological analysis showed that febuxostat treatment resulted in apparently less demyelination and axonal loss when compared with the control group (Fig. 5D).

Discussion
In this study, XOR activity was excessively elevated in the serum and in the CNS of EAE mice, particularly within the cerebellum and the spinal cord, where inflammation predominantly occurs. Moreover, these increases in XOR activity were inhibited by the novel XOR inhibitor, febuxostat, and this inhibition was associated with the amelioration of EAE.
Various studies suggest that activated microglia and infiltrated macrophages are among the main sources of ROS in CNS inflammation in patients with MS [4,32]. In accordance with this, immunohistochemical studies showed that the strong XO-positive cells in our model were macrophages and microglia. In vitro studies suggested that febuxostat effectively inhibits XOR expression by both macrophages and microglia, leading to the reduction of ROS generation. Consistent with these observations, febuxostat inhibited the activation of Iba1-positive cells in our model and furthermore led to the reduction of iNOS expression, which is reported to be enhanced by inflammatory stimuli and further increases ROS generation [33]. These results support a role for XOR within phagocytes in the establishment of CNS inflammation. Though XO immunoreactivity was detected in a few CD3positive T cells in the CNS, the role in EAE is currently unclear. At least febuxostat seems not to affect the peripheral T cell activation, as it did not inhibit the IFN-c and IL-17 production in recall assay.
Oxidation of myelin and axons plays a pivotal role in terms of the contribution of ROS to the pathogenesis of EAE [6,7]. ROS play an important role in myelin phagocytosis, as ROS inhibitors or scavengers reduce the phagocytosis of myelin by macrophages [34]. Oligodendrocytes are highly susceptible to oxidative injury, as they contain high levels of lipid and iron as well as reduced levels of antioxidant enzymes and free radical scavengers [35,36]. Human oligodendrocyte death associated with ROS induced by XO is reported [37]. ROS also have direct effects on the lipid and protein components of myelin through peroxidation [38]. In addition, a recent study reported that macrophages/microgliaderived ROS and RNS trigger axonal degeneration, and neutralization of ROS and RNS rescues axons [39]. Consistent with these studies, axonal loss and demyelination in our EAE models were markedly ameliorated by febuxostat, correlating with the decrease of ROS and RNS. These results suggest that antioxidant therapy with febuxostat may be effective for both EAE and MS. One possible drawback is that the reduction of uric acid by febuxostat may contribute to the deterioration of the disease, as uric acid is suggested to have a role as antioxidant [40]. However, the amelioration of EAE by the treatment with febuxostat suggests that the reduction of XOR is more important than the reduction of presumed protection by uric acid.
While 85% of MS patients initially present with a relapsingremitting course, the majority of these patients evolve into a secondary progressive course where disabilities worsen without recovery [41]. The pathology of progressive form of MS is characterized by extensive demyelination, gliosis, and axonal and neuronal losses [42]. Because current disease modifying therapies used for the relapsing-remitting MS (RRMS) are not effective enough for the treatment of progressive form of MS, it is important to identify new therapies for these patients [43]. EAE induced in NOD mice immunized with MOG resembles with secondary progressive MS (SPMS) and is known as secondary progressive EAE [30,31]. Febuxostat treatment ameliorated the disease progression in this secondary progressive EAE in accordance with the reduction of demyelination and axonal loss. These data suggest that febuxostat may be effective for patients with SPMS.
In summary, the XOR inhibitor, febuxostat, was effective as a therapeutic agent in EAE, likely due to its antioxidant activity. The effect was observed when febuxostat was administered either before or after establishment of symptoms of EAE. In addition, the effect was observed also in secondary progressive EAE. Therefore, XOR represents a novel therapeutic target for the treatment of MS, and febuxostat, which is already approved for treatment of gout in several countries, may be an important agent to modulate XOR activity and improve outcomes in patients with both RRMS and SPMS.