A Soluble Epoxide Hydrolase Inhibitor, 1-trifluoromethoxyphenyl-3-(1-propionylpiperidin-4-yl) Urea, Ameliorates Experimental Autoimmune Encephalomyelitis

Polyunsaturated fatty acids (PUFAs) are essential FAs for human health. Cytochrome P450 oxygenates PUFAs to produce anti-inflammatory and pain-resolving epoxy fatty acids (EpFAs) and other oxylipins whose epoxide ring is opened by the soluble epoxide hydrolase (sEH/Ephx2), resulting in the formation of toxic and pro-inflammatory vicinal diols (dihydroxy-FAs). Pharmacological inhibition of sEH is a promising strategy for the treatment of pain, inflammation, cardiovascular diseases, and other conditions. We tested the efficacy of a potent, selective sEH inhibitor, 1-trifluoromethoxyphenyl-3-(1-propionylpiperidin-4-yl) urea (TPPU), in an animal model of multiple sclerosis (MS), experimental autoimmune encephalomyelitis (EAE). Prophylactic TPPU treatment significantly ameliorated EAE without affecting circulating white blood cell counts. TPPU accumulated in the spinal cords (SCs), which was correlated with plasma TPPU concentration. Targeted lipidomics in EAE SCs and plasma identified that TPPU blocked production of dihydroxy-FAs efficiently and increased some EpFA species including 12(13)-epoxy-octadecenoic acid (12(13)-EpOME) and 17(18)-epoxy-eicosatrienoic acid (17(18)-EpETE). TPPU did not alter levels of cyclooxygenase (COX-1/2) metabolites, while it increased 12-hydroxyeicosatetraenoic acid (12-HETE) and other 12/15-lipoxygenase metabolites. These analytical results are consistent with sEH inhibitors that reduce neuroinflammation and accelerate anti-inflammatory responses, providing the possibility that sEH inhibitors could be used as a disease modifying therapy, as well as for MS-associated pain relief.

Multiple sclerosis (MS) is a neurological disease that is pathologically characterized by the central nervous system (CNS)-specific demyelination and inflammation [13]. Although the cause of MS still remains unclear, both genetic and environmental factors are highly involved in MS pathogenesis [14]. An animal model of MS, experimental autoimmune encephalomyelitis (EAE) [15], is a useful tool to study MS pathology. This model revealed the abnormality of neuro-immune interactions among autoreactive helper T cells secreting IL-17 (T H 17) or IFN-γ (T H 1) [16], B cells [17], and the CNS resident immune competent cells (microglia and astrocytes) [18,19]. We previously applied targeted lipidomics and transcriptomics approaches to EAE spinal cords (SCs), which identified the neuroinflammatory functions of PGE 2 in EAE pathogenesis [20,21]. EAE studies using knockout (KO) mice demonstrated the involvement of cytosolic phospholipase A 2 α (cPLA 2 α) in disease development [22], PGE 2 receptors (EP 2 and EP 4 ) [23] and the LTB 4 receptor (BLT 1 ) in T H 17 differentiation [24], and platelet-activating factor (PAF) receptors in macrophage/microglial phagocytic activity [25,26]. One of the significant contributions of the lipid biology field to the MS research is the approval of several sphingosine 1-phosphate (S1P) receptor modulators (fingolimod, siponimod, ozanimod, and ponesimod) by the U.S. Food and Drug Administration (FDA) [7,27]. Based on biochemical and pharmacological studies, S1P receptor modulators down-regulate S1P receptor 1 (S1P 1 ) expression on the cell surface as functional antagonists [28], resulting in sequestration of pathogenic lymphocytes from the circulation to secondary lymphoid organs [29]. Moreover, these drugs inhibit astrocyte activation during EAE development [30]. These studies clearly demonstrated that lipid signaling pathways are druggable for MS patients.

TPPU Protects Mice from EAE
We first tested the efficacy of TPPU, a sEH-selective inhibitor, on EAE development. C57BL/6 female mice were subcutaneously immunized with complete Freund's adjuvant containing myelin oligodendrocyte glycoprotein peptide (MOG 35-55 ) on day 0 and prophylactically treated with TPPU (10 mg/kg, s.i.d.) by oral administration starting from day 0. TPPU treatment significantly ameliorated EAE disease course as compared to controls (treatment, p < 0.0001; time, p < 0.0001; interaction, p = 0.89; by two-way ANOVA) ( Figure 1A). Cumulative scores, that were defined as the sum of the clinical scores from days 0 to 23, in the TPPU-treated group were significantly reduced over the vehicle group ( Figure 1B). TPPU treatment reduced the incidence and the mean maximal scores (average of the maximal score of the mice in the group), but were not significant ( Figure 1B). We detected a considerable concentration of TPPU in both spinal cords (SCs) and plasma, which showed a significant positive correlation ( Figure 1C). The white blood cell (WBC) counts and the proportions of WBCs in TPPU-treated EAE mice were equivalent to those in the vehicle-treated EAE mice ( Figure 1D). These results suggest that TPPU is effective for treating EAE, and its mechanism of action is different from fingolimod (Gilenya ® , Novartis), siponimod (Mayzent ® , Novaritis), ozanimod (Zeposia ® , Bristol Myers Squibb), and ponesimod (Ponvory TM , Johnson & Johnson), which reduce the circulating pathogenic lymphocytes via S1P 1 down-regulation [7].
containing myelin oligodendrocyte glycoprotein peptide (MOG35-55) on day 0 a lactically treated with TPPU (10 mg/kg, s.i.d.) by oral administration starting f TPPU treatment significantly ameliorated EAE disease course as compared (treatment, p < 0.0001; time, p < 0.0001; interaction, p = 0.89; by two-way ANOV 1A). Cumulative scores, that were defined as the sum of the clinical scores from 23, in the TPPU-treated group were significantly reduced over the vehicle gro 1B). TPPU treatment reduced the incidence and the mean maximal scores (ave maximal score of the mice in the group), but were not significant ( Figure 1B). W a considerable concentration of TPPU in both spinal cords (SCs) and plas showed a significant positive correlation ( Figure 1C). The white blood cell (W and the proportions of WBCs in TPPU-treated EAE mice were equivalent to t vehicle-treated EAE mice ( Figure 1D). These results suggest that TPPU is e treating EAE, and its mechanism of action is different from fingolimod (Gileny tis), siponimod (Mayzent ® , Novaritis), ozanimod (Zeposia ® , Bristol Myers Sq ponesimod (Ponvory TM , Johnson & Johnson), which reduce the circulating lymphocytes via S1P1 down-regulation [7].  Next, the EAE SCs were stained with hematoxylin and eosin (H&E) and luxol fast blue (LFB)-cresyl violet to assess the degree of inflammation and demyelination ( Figure 2). The vehicle-treated group displayed inflammatory cell infiltration into the perivascular regions and parenchyma (Figure 2A), which was associated with myelin pallor (demyelination) and tissue vacuolation ( Figure 2B). Importantly, tissue vacuolation was one of the features of EAE and was not observed in the naïve SCs [25]. Immunohistochemstry (IHC) for Iba-1, a microglia/macrophages marker, showed that Iba-1-positive cells accmulated in the EAE lesions and localized in the blood vessels-like structures of the grey matter ( Figure 2C). Moreover, astrogliosis was also determined by IHC for glial fibrillary acidic protein (GFAP) ( Figure 2C). TPPU treatment showed lesser degree of inflammation, demyelination and astrogliosis ( Figure 2D-F) as compared to vehicle controls, while a similar degree of tissue vacuolation was observed between TPPU-treated vs. control groups. The brain RNA-seq database showed specific expression of sEH/Ephx2 in astrocytes [35], suggesting that TPPU may inhibit astrogliosis as well as inflammation and demyelination. Next, the EAE SCs were stained with hematoxylin and eosin (H&E) and luxol fast blue (LFB)-cresyl violet to assess the degree of inflammation and demyelination ( Figure  2). The vehicle-treated group displayed inflammatory cell infiltration into the perivascular regions and parenchyma (Figure 2A), which was associated with myelin pallor (demyelination) and tissue vacuolation ( Figure 2B). Importantly, tissue vacuolation was one of the features of EAE and was not observed in the naïve SCs [25]. Immunohistochemstry (IHC) for Iba-1, a microglia/macrophages marker, showed that Iba-1-positive cells accmulated in the EAE lesions and localized in the blood vessels-like structures of the grey matter ( Figure 2C). Moreover, astrogliosis was also determined by IHC for glial fibrillary acidic protein (GFAP) ( Figure 2C). TPPU treatment showed lesser degree of inflammation, demyelination and astrogliosis ( Figure 2D-F) as compared to vehicle controls, while a similar degree of tissue vacuolation was observed between TPPU-treated vs. control groups. The brain RNA-seq database showed specific expression of sEH/Ephx2 in astrocytes [35], suggesting that TPPU may inhibit astrogliosis as well as inflammation and demyelination.

TPPU Blocked Dihydroxy-FA Production in EAE Plasma and Spinal Cords
We applied targeted lipidomics approaches to analyze lipid profiles in both plasma and SCs of EAE mice that were collected in the chronic phase of EAE. To investigate TPPU effects on lipid metabolism, we first analyzed total lipid levels in the COX, 5-LO, 12/15-LO, and CYP-sEH pathways by calculating the sum of metabolite levels in each pathway.

TPPU Blocked Dihydroxy-FA Production in EAE Plasma and Spinal Cords
We applied targeted lipidomics approaches to analyze lipid profiles in both plasma and SCs of EAE mice that were collected in the chronic phase of EAE. To investigate TPPU effects on lipid metabolism, we first analyzed total lipid levels in the COX, 5-LO, 12/15-LO, and CYP-sEH pathways by calculating the sum of metabolite levels in each pathway.
AA metabolites produced by 12/15-LO were rich in EAE plasma (~1 µmol/L) and were up-regulated by TPPU (~2 µmol/L; Figure 3A). TPPU did not alter COX and 5-LO-mediated AA fluxes, but did significantly reduce COX-mediated EPA metabolites and significantly elevated the 12/15-LO metabolites ( Figure 3A). EpFAs were abundantly present (200-300 nmol/L), except for EpETE (~10 nmol/L), in the TPPU-treated and control groups ( Figure 3B). As expected from the TPPU inhibitory actions to the sEH, TPPU effectively and significantly blocked the sEH metabolites including dihydroxy-octadecenoic acid (DiHOME), dihydroxy-icosatrienoic acid (DiHETrE), dihydroxy-octadecadienoic acid (DiHODE), and dihydroxy-eicosatetraenoic acid (DiHETE) ( Figure 3B). We also found that epoxy-octadecenoic acid (EpOME), a precursor of DiHOME, was significantly elevated in the TPPU-treated group as compared to controls ( Figure 3B). Correlation analyses revealed positive relationships within C18-PUFA metabolites and within C20-and C22-PUFA metabolites ( Figure 3C). This suggested the association of carbon chain lengths with the substrate preferences in CYPs and sEH activities. All the dihydroxy-FAs showed strong negative correlation with the regioisomeric epoxides of linoleate EpOME (Figure 3C), suggesting a potential anti-inflammatory role for EpOME in EAE or possibly a toxic or inflammatory role for the corresponding diols or DiHOMEs (sometimes termed leukotoxin diols). mediated AA fluxes, but did significantly reduce COX-mediated EPA metabolites and significantly elevated the 12/15-LO metabolites ( Figure 3A). EpFAs were abundantly present (200-300 nmol/L), except for EpETE (~10 nmol/L), in the TPPU-treated and control groups ( Figure 3B). As expected from the TPPU inhibitory actions to the sEH, TPPU effectively and significantly blocked the sEH metabolites including dihydroxy-octadecenoic acid (DiHOME), dihydroxy-icosatrienoic acid (DiHETrE), dihydroxy-octadecadienoic acid (DiHODE), and dihydroxy-eicosatetraenoic acid (DiHETE) ( Figure 3B). We also found that epoxy-octadecenoic acid (EpOME), a precursor of DiHOME, was significantly elevated in the TPPU-treated group as compared to controls ( Figure 3B). Correlation analyses revealed positive relationships within C18-PUFA metabolites and within C20-and C22-PUFA metabolites ( Figure 3C). This suggested the association of carbon chain lengths with the substrate preferences in CYPs and sEH activities. All the dihydroxy-FAs showed strong negative correlation with the regioisomeric epoxides of linoleate EpOME ( Figure  3C), suggesting a potential anti-inflammatory role for EpOME in EAE or possibly a toxic or inflammatory role for the corresponding diols or DiHOMEs (sometimes termed leukotoxin diols). In EAE SCs, AA metabolites via the COX-1/2 pathway were abundant (~500 pmol/g). TPPU treatment did not affect fluxes in the COX and 5-LO pathways but showed a similar In EAE SCs, AA metabolites via the COX-1/2 pathway were abundant (~500 pmol/g). TPPU treatment did not affect fluxes in the COX and 5-LO pathways but showed a similar trend in the 12/15-LO pathway with that of plasma ( Figure 4A). Levels of EpETrE and EpDPE (100-200 pmol/g) in SCs were equivalent to those in plasma (100-200 nmol/L), while C18-PUFA metabolites (EpOME and EpODE) were~50-fold lower than those in plasma ( Figure 4B). Similar trends of EpFA and dihydroxy-FA profiles were observed between SCs and plasma, including the inhibition of DiHETrE and DiHODE, as well as an increase of EpOME ( Figure 4B) possibly due to TPPU penetrating efficiently into the SCs ( Figure 1B). Positive correlations within C20 or C22-PUFA metabolites were found, such as EpETrE vs. EpDPE and DiHDPE vs. DiHETE ( Figure 4C). trend in the 12/15-LO pathway with that of plasma ( Figure 4A). Levels of EpETrE and EpDPE (100-200 pmol/g) in SCs were equivalent to those in plasma (100-200 nmol/L), while C18-PUFA metabolites (EpOME and EpODE) were ~50-fold lower than those in plasma ( Figure 4B). Similar trends of EpFA and dihydroxy-FA profiles were observed between SCs and plasma, including the inhibition of DiHETrE and DiHODE, as well as an increase of EpOME ( Figure 4B) possibly due to TPPU penetrating efficiently into the SCs ( Figure 1B). Positive correlations within C20 or C22-PUFA metabolites were found, such as EpETrE vs. EpDPE and DiHDPE vs. DiHETE ( Figure 4C).

TPPU Reduced Dihydroxy-FA Production with an Accompanying Increase of EpFAs in EAE Mice
Differential lipid levels were computed for TPPU vs. control groups that were represented as a scatter plot ( Figure 5). This plot clearly displayed the aggregation of almost all the dihydroxy-FAs (such as 12,13-DiHOME and 15,16-DiHODE) and trihydroxy-FAs (such as 9,10,13-TriHOME and 9,12,13-TriHOME) into quadrant III (log 10 (TPPU/vehicle) < 0 in both plasma and SC) with a few exceptions such as 4,5-DiHDPE ( Figure 5). This was accompanied by an up-regulation of some EpFA species including 12(13)-EpOME and 17(18)-EpETE ( Figure 5), indicating that TPPU inhibited sEH activity effectively and increased Ep-FAs consequentially. Furthermore, AA and EPA metabolites in the 12/15-LO pathway (such as 12-HETE) were located in quadrant I of the scatter plot (log 10 (TPPU/vehicle) > 0 in both plasma and SC; Figure 5), suggesting that sEH inhibition caused a potential re-diversion of PUFA to the 12/15-LO pathway, which was also reported in the Ephx2 deficient mice. pathway (such as 12-HETE) were located in quadrant I of the scatter plot (log10(TPPU/vehicle) > 0 in both plasma and SC; Figure 5), suggesting that sEH inhibition caused a potential re-diversion of PUFA to the 12/15-LO pathway, which was also reported in the Ephx2 deficient mice. Figure 5. Differential lipid profiles of TPPU-treated vs. vehicle-treated EAE mice. The scatter plot shows the effect of TPPU on lipid levels in EAE SCs (x-axis) and plasma (y-axis). Each symbol represents lipid species coded by color and shape. Representative lipids are displayed as bar graphs. P values were determined by t-test or Mann-Whitney U test. N.S., nonsignificant. n.d., not detected.

Discussion and Conclusions
In the present study, we demonstrated the beneficial effect of TPPU in the EAE mice without changing the number of circulating lymphocytes, and also showed that it effectively reduced pro-inflammatory dihydro-FAs in SCs and blood. Currently available disease modifying therapies (DMTs) for MS treatment are mostly immunomodulatory drugs that decrease circulating T and B lymphocytes, and thus prevent pathogenic lymphocytes from penetrating the CNS [36]. DMTs that induce lymphopenia increase serious infection risk, including John Cunningham virus (JCV) infection that causes progressive multifocal leukoencephalopathy (PML) [37]. Although the incidence rates of these infections are reported to be low and similar between DMTs [37], immunomodulatory DMTs appear to be unsafe for use in immunocompromised patients. This study, along with a recent report testing TPPU in EAE [38], provides a novel therapeutic strategy for using TPPU and re- Figure 5. Differential lipid profiles of TPPU-treated vs. vehicle-treated EAE mice. The scatter plot shows the effect of TPPU on lipid levels in EAE SCs (x-axis) and plasma (y-axis). Each symbol represents lipid species coded by color and shape. Representative lipids are displayed as bar graphs. P values were determined by t-test or Mann-Whitney U test. N.S., non-significant. n.d., not detected.

Discussion and Conclusions
In the present study, we demonstrated the beneficial effect of TPPU in the EAE mice without changing the number of circulating lymphocytes, and also showed that it effectively reduced pro-inflammatory dihydro-FAs in SCs and blood. Currently available disease modifying therapies (DMTs) for MS treatment are mostly immunomodulatory drugs that decrease circulating T and B lymphocytes, and thus prevent pathogenic lymphocytes from penetrating the CNS [36]. DMTs that induce lymphopenia increase serious infection risk, including John Cunningham virus (JCV) infection that causes progressive multifocal leukoencephalopathy (PML) [37]. Although the incidence rates of these infections are reported to be low and similar between DMTs [37], immunomodulatory DMTs appear to be unsafe for use in immunocompromised patients. This study, along with a recent report testing TPPU in EAE [38], provides a novel therapeutic strategy for using TPPU and related sEH inhibitors, which might be effective for all types of MS patients. sEH inhibitors stabilize most EpFAs studied to date to varying degrees [9]. In general, this is considered beneficial because these epoxides appear to be inflammation resolving agents that reduce ER stress [39]. A recent study showed that TPPU induced neuroinflammatory resolution in a mouse model of Alzheimer's disease (AD) and increased EpFAs (EpETE and EpDPE) in the brain [31]. However, the findings of the present study were not entirely consistent with these results possibly because of the differences in the MS and AD mechanisms underlying neuroinflammation. We showed that TPPU effectively blocked production of most dihydroxy-FA species, resulting in a compensatory increase of a few EpFA species including 12(13)-EpOME and 17(18)-EpETE ( Figure 5). The 12(13)-EpOME (leukotoxin) was believed to be involved in multiple organ failure and adult respiratory distress syndrome until the discovery of the ultimate toxic metabolite, 12,13-DiHOME (leukotoxin diol) [40]. Although sEH induction and subsequent DiHOME production are involved in thermogenesis in brown fat adipose [41], diols of linoleate at high concentrations induce deleterious consequences in vascular and pulmonary permeability [40]. Importantly, the plasma levels of 12,13-DiHOME were associated with severe cases of COVID-19 (coronavirus disease 2019) [42], further supporting the detrimental effects of 12,13-DiHOME in respiratory failure. Since TPPU significantly and robustly reduced the toxic 12,13-DiHOME in EAE plasma and SCs, inhibition of this pathway might be a key mechanism for TPPU's preventative effects. Moreover, the DiHOMEs should be evaluated as possible biomarkers in EAE, and potentially in MS and related neuroinflammatory diseases.
Anti-inflammatory effects of 17(18)-EpETE have been proposed in several diseases including contact hypersensitivity [43] and non-alcoholic fatty liver disease [44], which may be mediated through one of the three FA GPCRs, GPR40 [43,45], and/or peroxisome proliferator-activated receptor gamma (PPARg) [46]. Although the relative amount of EPA-derived 17(18)-EpETE was small, its increase appeared to be important for neuroinflammatory resolution in EAE. On the other hand, DHA metabolites were mostly down-modulated by TPPU. A unique exception was an increase of 4,5-DiHDPE, whose functions remain elusive, while it most likely shares a similar pro-inflammatory function with other dihdroxy-FAs.
Pharmacological and genetic sEH inhibition appears to alter FA fluxes towards the 12/15-LO pathway. This increased flux may be used for SPM production. Since specialized pro-resolving mediators (SPMs; including lipoxins, hepoxillins, resolvins, protectins) require 12/15-LO activity for their biosynthesis, sEH inhibition may enhance SPM production when substrate PUFAs are sufficiently provided. 12/15-LO deficiency aggravated EAE [47], supporting the pro-resolving and anti-inflammatory effects of 12/15-LO metabolites in EAE and MS. Indeed, resolvin D 1 , which is produced from DHA by the actions of 15-LO and 5-LO, ameliorated EAE [48]. On the other hand, selective sEH inhibition did not affect metabolic pathways mediated through COX-1/2 and 5-LO in either SCs or plasma. Although COX-2 might not be profoundly involved in EAE/MS pathogenesis [49], many eicosanoid species produced downstream of the COX-1/2 and 5-LO pathways show pro-inflammatory action in EAE [20]. Therefore, dual inhibitors for sEH/COX-2 that are currently under development might be beneficial for MS patients and probably more effective in MS-associated pain [10]. TPPU is highly present in the CNS, and its concentration was significantly correlated between SCs and plasma ( Figure 1C), supporting a direct action of TPPU in the CNS to suppress neuroinflammation. Taken together, TPPU and other sEH-selective inhibitors appear to be beneficial for the treatment of MS in this murine model and possibly other neurological diseases.
Spinal cords were homogenized in ice-cold methanol containing 0.1% butylated hydroxytoluene and 0.1% acetic acid. Internal standard mixtures were spiked into the homogenates and stored at −80 • C for 20 h. Samples were processed by solid-phase extraction (OasisHLB Cartridge, Waters, Milford, MA, USA), reconstituted in 200 nM CUDA in methanol, and analyzed by modified LC-MS/MS [34].

Statistics
As appropriate, data were analyzed statistically using Prism software (GraphPad, San Diego, CA, USA) including Student's t test, Mann-Whitney U test, Fisher's exact test, Tukey's multiple comparison test, or two-way repeated measures ANOVA. A p-value of <0.05 was considered to be statistically significant.

Informed Consent Statement: Not applicable.
Data Availability Statement: The data presented in this study are available on request from the corresponding author.