Soluble Epoxide Hydrolase as an Anti-inflammatory Target of the Thrombolytic Stroke Drug SMTP-7*

Background: The small molecule thrombolytic SMTP-7 excels in the treatment of ischemic stroke through an unidentified anti-inflammatory activity. Results: Affinity purification and inhibition studies reveal that SMTP-7 and its congeners inhibit soluble epoxide hydrolase (sEH), an enzyme responsible for inflammation. Conclusion: sEH is the plausible in vivo anti-inflammatory target of SMTP-7. Significance: Dual-targeting of thrombolysis and sEH is a promising strategy for novel stroke therapy. Although ischemic stroke is a major cause of death and disability worldwide, only a small fraction of patients benefit from the current thrombolytic therapy due to a risk of cerebral hemorrhage caused by inflammation. Thus, the development of a new strategy to combat inflammation during thrombolysis is an urgent demand. The small molecule thrombolytic SMTP-7 effectively treats ischemic stroke in several animal models with reducing cerebral hemorrhage. Here we revealed that SMTP-7 targeted soluble epoxide hydrolase (sEH) to suppress inflammation. SMTP-7 inhibited both of the two sEH enzyme activities: epoxide hydrolase (which inactivates anti-inflammatory epoxy-fatty acids) and lipid phosphate phosphatase. SMTP-7 suppressed epoxy-fatty acid hydrolysis in HepG2 cells in culture, implicating the sEH inhibition in the anti-inflammatory mechanism. The sEH inhibition by SMTP-7 was independent of its thrombolytic activity. The simultaneous targeting of thrombolysis and sEH by a single molecule is a promising strategy to revolutionize the current stroke therapy.

Although ischemic stroke is a major cause of death and disability worldwide, only a small fraction of patients benefit from the current thrombolytic therapy due to a risk of cerebral hemorrhage caused by inflammation. Thus, the development of a new strategy to combat inflammation during thrombolysis is an urgent demand. The small molecule thrombolytic SMTP-7 effectively treats ischemic stroke in several animal models with reducing cerebral hemorrhage. Here we revealed that SMTP-7 targeted soluble epoxide hydrolase (sEH) to suppress inflammation. SMTP-7 inhibited both of the two sEH enzyme activities: epoxide hydrolase (which inactivates anti-inflammatory epoxyfatty acids) and lipid phosphate phosphatase. SMTP-7 suppressed epoxy-fatty acid hydrolysis in HepG2 cells in culture, implicating the sEH inhibition in the anti-inflammatory mechanism. The sEH inhibition by SMTP-7 was independent of its thrombolytic activity. The simultaneous targeting of thrombolysis and sEH by a single molecule is a promising strategy to revolutionize the current stroke therapy.
Stroke is a major cause of death and disability worldwide (1). Improving the case fatality rates and long term disability after stroke continues to be a challenge (2). During the past decade thrombolytic enzyme tissue plasminogen activator-based treatment has been the standard therapy for acute ischemic stroke (2). However, due to its risk of parenchymal hemorrhage and a limited therapeutic time window (3)(4)(5), only a small fraction of patients (1.8 -8.9%) benefits from tissue plasminogen activator-based therapy (6,7). The development of an alternative therapeutic agent is urgently needed. In this context, suppressing inflammation within the infarction area to rescue pen-umbral tissue and reduce hemorrhagic transformation is particularly important (8,9). SMTPs 3 (named after Stachybotrys microspora triprenyl phenols) are a family of novel small molecules produced by the fungus Stachybotrys microspora (10 -12). SMTP-7 (Fig. 1A), one of the SMTP family compounds with profound biological activities, enhances plasminogen activation by modulating plasminogen conformation (10,(13)(14)(15). SMTP-7 thus promotes plasmin formation and clot clearance in vivo (14,16), and it has been used to treat thrombotic and embolic strokes in experimental models in rodents and nonhuman primates (17)(18)(19)(20). Unexpectedly, SMTP-7 was demonstrated to reduce hemorrhagic transformation and to have a wide therapeutic time window (17,18,20). These excellent activities are partly explained by the finding that, unlike tissue plasminogen activator, SMTP-7 suppresses inflammatory/oxidative responses after thrombolytic reperfusion (16,18,19,21). These beneficial properties prompted the development of SMTP-7 as an alternative stroke drug that can treat patients who do not benefit from tissue plasminogen activator-based therapy (20).
The aim of the present study was to explore the anti-inflammatory mechanism of SMTP-7. We first compared the efficacy of two SMTP congeners differing in the potential of thrombolytic activity using three inflammatory disease animal models. The results demonstrate that the anti-inflammatory action of SMTP is independent of its thrombolytic activity. We next searched for an anti-inflammatory target using SMTP affinity beads resulting in the identification of soluble epoxide hydrolase (sEH). sEH is a bifunctional enzyme with an epoxide hydrolase activity at the C-terminal domain (Cterm-EH) and a lipid phosphate phosphatase activity at the N-terminal domain (Nterm-phos) (22). The Cterm-EH catalyzes the hydrolysis of epoxy fatty acids such as epoxyeicosatrienoic acids (EETs), which are potent endogenous signaling molecules implicated in anti-inflammation, vascular dilation, endothelial cell hyperpolarization, angiogenesis, neuroprotection, and analgesia (antihyperalgesia) (23)(24)(25)(26)(27). The Nterm-phos hydrolyzes lipid phosphates such as lysophosphatidic acid and intermediates of the cholesterol biosynthesis (28 -30). SMTP-7 inhibited both of the two activities of sEH: competitively for Cterm-EH and pseudononcompetitively for Nterm-phos. The treatment of HepG2 cells with of SMTP-7 or its congeners suppressed the hydrolysis of EET to dihydroxyeicosatrienoic acid (DHET). Here, we describe the anti-inflammatory action of SMTP-7 and its mechanism.

EXPERIMENTAL PROCEDURES
Animal Experiments-All of the animal protocols were approved by the institutional animal experiment committees at the Tokyo Noko University and Nihon Pharmaceutical. Male C57BL/6J mice (7 weeks old; Japan SLC, Hamamatsu) and male Lewis rats (6 weeks old; Charles River Laboratories Japan, Yokohama) were used after 1 week of preliminary rearing for the inflammatory disease models. Retired male ICR mice (Japan SLC, Hamamatsu) were used to obtain the livers for the sEH purification.
Crohn Disease Model-Male C57BL/6J mice (8 weeks old) were intrarectally injected with 2,4,6-trinitrobenzene sulfonic acid (TNBS) (40) solution (20 mg ml Ϫ1 in 50% ethanol) at a dose of 100 mg kg Ϫ1 . SMTP-7, SMTP-44D, 5-aminosalicylic acid, or prednisolone was administered as described for the dextran sulfate model. These treatments were made 30 min before the TNBS injection as well as 24, 48, and 72 h after the TNBS injection. Control animals received no drug treatment. Normal animals received neither TNBS nor any drug. The disease activity index score was measured on days 1, 2, 3, and 4 after the TNBS injection. There were five animal in each group.
SMTP-47-SMTP-47 was synthesized by the microbial amine feeding method (31). For the preparation of feeding amine, N ␣ -tert-butoxycarbonyl-N ␦ -Fmoc-L-ornithine (60 mg ml Ϫ1 in tetrahydrofuran) was treated with equal volume of trifluoroacetic acid, affording N ␦ -Fmoc-L-ornithine. The culture of S. microspora IFO 30018 (100 ml) was fed with 100 mg of N ␦ -Fmoc-L-ornithine, and the resulting SMTP-47 (89 mg) was purified by reverse-phase HPLC developed with MeOH, 0.1% formic acid (85:15). 1  SMTP-50-coupled Affinity Matrix-One column volume (1 ml for target identification or 5 ml for target purification) of 1.2 mM SMTP-50, dissolved in 10 mM sodium phosphate, pH 8.3, was applied to a HiTrap NHS-activated HP column (GE Healthcare) at room temperature for 30 min. The column was then treated with monoethanolamine to block residual N-hydroxysuccinimidyl group of the matrix.
Identification of SMTP-binding Protein-The livers from male ICR mice were perfused with ice-cold saline and homogenized in 4 volumes of 25 mM Tris-HCl, pH 7.4, containing 150 mM NaCl and 0.2% (wt vol Ϫ1 ) sodium deoxycholate (buffer A). A supernatant fraction was obtained after centrifugation at 1000 ϫ g for 15 min followed by 20,000 ϫ g for 15 min at 4°C. After filtration, 15 ml of the supernatant was applied to a 1-ml SMTP-affinity column pre-equilibrated with buffer A at 20°C. The column was washed with 10 ml of buffer A and developed with 10 ml of buffer A containing additional NaCl (500 mM, finally). Aliquots of the eluate were resolved by reduced SDSpolyacrylamide gel electrophoresis. Coomassie Brilliant Blue R250-stained protein bands were excised from the gel and digested with trypsin. The resulting peptides were subjected to chemically assisted fragmentation post-source decay analysis by matrix-assisted laser desorption/ionization (MALDI) timeof-flight (TOF) mass spectrometry on a Ettan MALDI-TOF Pro (Amersham Biosciences) or LC-MALDI-TOF/TOF analysis on a 4700 Proteomics Analyzer (Applied Biosystems, Framingham, MA). Detected masses were subjected to amino acid sequence analysis and to comparison with theoretical peptide masses on MASCOT search engine (Matrixscience, Boston, MA) to identify the protein species.
Purification of sEH by SMTP Affinity Chromatography-All of the following operations were carried out at 0 -4°C. The livers from male ICR mice were homogenized in 11.5 volumes of 76 mM sodium phosphate, pH 7.4. A supernatant fraction was obtained after centrifugation at 1000 ϫ g for 10 min followed by 10,000 ϫ g for 25 min and then 100,000 ϫ g for 60 min. The resulting cytosol fraction (250 ml) was applied to a 5-ml SMTPcolumn pre-equilibrated with 76 mM sodium phosphate, pH 7.4, containing 0.1 mM EDTA (buffer B). After washing with buffer B (200 ml), the column was developed with 60 ml of buffer B containing 10 M 12-(3-adamantan-1-yl-ureido)dodecanoic acid (AUDA). The eluate was dialyzed against buffer B to remove AUDA and ultrafiltered to concentrate and exchange buffer to 100 mM sodium phosphate, pH 7.4, containing 3 mM dithiothreitol. From 3 batches of affinity chromatography, 2.2 mg of homogeneous sEH was purified. The purified sEH had specific activities of 511 nmol min Ϫ1 mg Ϫ1 for the Cterm-EH and 2850 nmol min Ϫ1 mg Ϫ1 for the Nterm-phos when we used trans-stilbene oxide and p-nitrophenyl phosphate as respective substrates.
LC-MS Analysis of EET and DHET-Samples to be analyzed were extracted with ethyl acetate. After centrifugation, super-natant was concentrated to dryness. The resulting materials were dissolved in methanol and subjected to LC-MS analysis for 14,15-EET and 14,15-DHET on a Micromass Quattro Ultima triple quadrupole tandem mass spectrometer equipped with an electrospray ionization interface (Waters, Tokyo, Japan). Samples (10 l) were resolved on a silica-ODS column (100 ϫ 2 mm; Pegasil ODS SP100 -3, Senshu Scientific, Tokyo, Japan) developed at 0.2 ml min Ϫ1 with a linear gradient of acetonitrile in 0.1% formic acid (60 -100%) for 15 min). The electrospray ionization was performed in the negative ion mode with a capillary voltage at 3.0 kV. The cone voltages were set at 35 V for 14,15-DHET/14,15-DHET-d11, and 30 V for 14,15-EET/14,15-EET-d11. Data were acquired in the multichannel analysis mode and analyzed using the MassLynx software (Version 3.5; Waters).
Assay for sEH-The Cterm-EH activity was assayed with 14,15-EET or (3-phenyl-oxiranyl)-acetic acid cyano-(6-methoxy-naphthalen-2-yl)-methyl ester (PHOME) as a substrate. When using PHOME, we preincubated mouse sEH (60 ng) for 10 min in 80 l of 25 mM Bis-Tris-HCl, pH 7.0, containing 0.1 mM MgCl 2 and 0.1 mg ml Ϫ1 bovine serum albumin (buffer C) with or without a compound to be tested. The composition of buffer C was based on the method by Tran et al. (28), unless MgCl 2 was added to unify the buffer composition with that for Nterm-phos determination, in which MgCl 2 is essential. After adding 20 l of the substrate, fluorescence (excitation, 355 nm; emission, 460 nm) of the reaction product was measured kinetically at 30°C. The final concentrations of sEH and PHOME in the standard assay conditions were 4.7 nM and 12.5 M, respectively. When using EET, we incubated mouse sEH (0.18 ng) with 14,15-EET in 60 l of buffer C at 30°C for 20 min. After the addition of 14,15-EET-d 11 and 14,15-DHET-d 11 (25 pmol and 5 pmol, respectively) as internal standards, 14,15-DHET formed was determined by LC-MS as described above. The Ntermphos activity was assayed with AttoPhos as a substrate. Mouse sEH (30 ng) was preincubated for 10 min in 80 l of buffer C with or without a compound to be tested. After adding 20 l of AttoPhos, fluorescence (excitation, 450 nm; emission, 545 nm) of the reaction product was measured kinetically at 30°C. The final concentrations of sEH and AttoPhos in the standard assay conditions were 2. Global Analysis of Arachidonate Metabolites of the Cytochrome P450, Cyclooxygenase, and Lipoxygenase Pathways-Plasma (400 l) obtained from Guillain-Barré syndrome model rats 2 h after SMTP-7 or saline treatment (n ϭ 8) on day 13 was randomly paired within each group, and the mixture (n ϭ 4 for each group) was centrifuged at 5000 ϫ g for 10 min. The resulting supernatant was mixed with 80 l of formic acid, and the mixture was applied to Sep-Pak C18 Plus Short Cartridges (Waters, Tokyo, Japan). The column was washed with EtOHwater-formic acid (10:100:1, vol Ϫ1 vol Ϫ1 ), and metabolites of interest were eluted with 5 ml of EtOH. The eluate was evaporated and dissolved with 40 l of 50% aqueous MeOH. Aliquots (10 l) were subjected to LC-MS/MS analysis on an L-column2 ODS (2 m, 1 ϫ 150 mm, CERI, Tokyo, Japan) developed at a rate of 0.1 ml Ϫ1 with a liner gradient (10-85%) of acetonitrile in 5 mM ammonium formate-formic acid (1000:1, vol Ϫ1 vol Ϫ1 ) for 26 min. Eluates were ionized with electrospray ionization, and negative ions of oxylipins were monitored on API 3200 QTRAP (AB SCIEX, Tokyo, Japan). Metabolites to be analyzed were

RESULTS
Anti-inflammatory Action Independent of Plasminogen Modulation-Although SMTP-7 reduces inflammatory responses in thromboembolic stroke models, it was unclear whether or not this outcome is a consequence of the recanalization by thrombolytic enhancement. We thus employed inflammatory disease models (Guillain-Barré syndrome, ulcerative colitis, and Crohn disease models) that were irrelevant to thromboembolic complications to directly assess the anti-inflammatory action of SMTP-7. We compared the efficacy of SMTP-7 with that of its congener, SMTP-44D (11) (Fig. 1A), which is essentially inactive in the plasminogen modulation activity (Fig. 1B). In the Guillain-Barré syndrome model in rats, SMTP-7 (10 mg kg Ϫ1 ) and SMTP-44D (10 mg kg Ϫ1 ) both ameliorated neuritis symptoms as did the clinically used immunoglobulin formula-  DECEMBER 26, 2014 • VOLUME 289 • NUMBER 52 tion (400 mg kg Ϫ1 ) (Fig. 1C). Both SMTP-7 (10 mg kg Ϫ1 ) and SMTP-44D (10 mg kg Ϫ1 ) alleviated the disease-associated body weight loss, stool inconsistency, and intestinal bleeding in the models of ulcerative colitis and Crohn disease in mice (Fig. 1,  D-G and H-K). These effects were comparable with more prominent than the effects of the standard drug 5-aminosalicylic acid (100 mg kg Ϫ1 ) and those of prednisolone (20 mg kg Ϫ1 ). Thus, the anti-inflammatory action of SMTP is independent of the plasminogen modulation activity.

Anti-inflammatory Target of a Novel Thrombolytic Stroke Drug
Identification of sEH as a Target-To identify the molecule that is involved in the anti-inflammatory mechanism of SMTP, we designed an affinity matrix that contained an essential part of SMTP, tricyclic ␥-lactam with a geranylmethyl side chain ( Fig. 2A). To prepare the SMTP congener with a primary amine FIGURE 2. Identification of sEH as an intracellular target of SMTP. A, the synthesis of SMTP-50-coupled affinity matrix. We first synthesized SMTP-47 by the precursor amine-fed culture of S. microspora IFO 30018 using N ␦ -Fmoc-L-ornithine as a feeding amine. The Fmoc group was removed to afford SMTP-50, which was then coupled with N-hydroxysuccinimidyl-activated cross-linked agarose, affording SMTP-50-coupled affinity matrix. B, SDS-polyacrylamide gel electrophoresis of the eluates from the affinity chromatography of mouse liver homogenates. Arrowheads denote the protein bands specifically bound to the affinity matrix, and asterisks denote nonspecifically bound protein bands as judged by comparison with the results obtained with control matrix. C, summary of the peptide mass fingerprinting analysis of protein bands 1-4. D, results of the peptide mass fingerprinting analysis of protein band 1. The bars are for the found peptides cover the peptide sequences in Table 1. E, SDS-polyacrylamide gel electrophoresis of the preparation specifically eluted with AUDA by SMTP-affinity chromatography.

Results of the peptide mass fingerprinting analysis of protein band 1 in Fig. 2B
The propionamide modification of cysteine residue may be due to the reaction with acrylamide monomer remaining in polyacrylamide. DECEMBER 26, 2014 • VOLUME 289 • NUMBER 52 at the N-linked side chain (SMTP-50), we first synthesized a precursor with the primary amine of SMTP-50 protected by a 9-fluorenylmethyloxycarbonyl group (SMTP-47) ( Fig. 2A) by the precursor amine-fed culture method (31). After eliminating the protective group, the resulting SMTP-50 was coupled to cross-linked agarose beads.

Anti-inflammatory Target of a Novel Thrombolytic Stroke Drug
Detergent-solubilized homogenates of the mouse liver were subjected to affinity chromatography on the SMTP-coupled beads. The resulting eluate gave several protein bands on SDSpolyacrylamide gel electrophoresis. We subjected four specifically eluted protein species to peptide-mass fingerprinting analysis (Fig. 2B), and every protein species was identified as a full-length form or a fragment of sEH (Fig. 2, C and D, and Table  1). The purification of sEH to homogeneity was achieved when SMTP-affinity chromatography was performed using a cytosol fraction of detergent-free homogenates and an elution buffer containing AUDA, a competitive inhibitor of the Cterm-EH of sEH (41) (Fig. 2E).
Inhibition of sEH-SMTP-7 inhibited both the Cterm-EH and the Nterm-phos of sEH (their respective IC 50 values were 23 Ϯ 1 and 6 Ϯ 1 M) when we used the synthetic substrates PHOME and AttoPhos, respectively (Fig. 3, A and B). Similarly, SMTP-44D inhibited Cterm-EH and Nterm-phos (IC 50 27 Ϯ 2 and 24 Ϯ 3 M, respectively) (Fig. 3, A and B). In addition, the structurally simplest congener SMTP-0 (which lacks the N-linked side chain) was inhibitory to both activities (IC 50 6 Ϯ 1 and 14 Ϯ 1 M, respectively) (Fig. 3, A and B). Thus, the structural requirement for sEH inhibition is clearly distinguishable from that for plasminogen modulation activity in which the N-linked side chain plays a crucial role (11,12,32).
We performed detailed kinetic analyses of sEH inhibition using SMTP-0 and the natural substrate 14,15-EET. The use of SMTP-0 was to avoid complexity of data analysis: SMTP-0 consists of a single core unit of SMTP (Fig. 3), whereas SMTP-7 has two core units that are asymmetrically configured (Fig. 1A), and each of these may differently interact with the enzyme. The kinetic results revealed a positive cooperativity for the hydrol-ysis of 14,15-EET (Fig. 4A) (Hill coefficient of 1.9 for the substrate binding; Fig. 4B), suggesting an allosteric interaction between the two monomers of sEH. The data were, therefore, analyzed based on a nonlinear mathematical model that allowed allostericity between the two equivalent catalytic sites of Cterm-EH (Fig. 4C). The pattern of the Cterm-EH inhibition by SMTP-0 fitted well to a competitive model (Fig. 4, A and C). Moreover, Cterm-EH inhibition by SMTP-0 was competed for by AUDA (this class of inhibitor binds to the catalytic site in Cterm-EH (42)) (Fig. 4, D and E). These results are consistent with the observation that sEH is specifically eluted with AUDA in SMTP affinity chromatography. Regarding the Nterm-phos, the kinetic data fitted to a linear mathematical model (Fig. 5, A-C), suggesting no cooperativity between the two Ntermphos domains (Hill coefficient of 0.91; Fig. 5B). The inhibition of Nterm-phos by SMTP-0 was pseudo-noncompetitive with respect to the substrate AttoPhos (Fig. 5, A and B). The pseudononcompetitive mechanism suggests that the Nterm-phos inhibition is mediated by the SMTP-0 binding to an allosteric site other than the substrate binding site in the Nterm-phos. Because there remained a possibility that the binding of SMTP-0 to the substrate binding site in the Cterm-EH might affect the activity of the Nterm-phos, we assessed the inhibitory activity of SMTP-0 toward the Nterm-phos in the presence of AUDA, which competed with SMTP-0 for binding to the catalytic site in the Cterm-EH (Fig. 4D). As a result, the presence of AUDA did not affect the SMTP-0 inhibition of the Nterm-phos (Fig. 5, D and E). Thus, SMTP-0 should bind to two distinct sites in sEH; one is the catalytic site in the Cterm-EH, and the other is an allosteric site that affects the Nterm-phos.
The structure-activity relationships of SMTP congeners differing in the N-linked side chain are summarized in Fig. 6. Although the minimum structural requirement for the sEH inhibition was the tricyclic ␥-lactam with a geranylmethyl side chain (represented by SMTP-0), the difference in the N-linked side chain variably affected the potency of the inhibition of the Cterm-EH and Nterm-phos. The congener with a naphthalene (SMTP-16) (32) or a glucose moiety as the N-linked side chain (SMTP-33) (32) was essentially inactive in inhibiting the Cterm-EH and Nterm-phos. SMTP-54 (with a glutamine moiety) (33) and SMTP-55 (with a glutamic acid moiety) (33) were relatively specific for the inhibition of the Cterm-EH. On the other hand, SMTP-5D (with a D-leucine moiety) (34) was relatively specific for the inhibition of the Nterm-phos. The inhibitory activity of the congeners with a phenylamine derivative as the N-linked side chain (such as SMTP-26, SMTP-27, and SMTP-28) (12) were potent with respect to both the Cterm-EH and Nterm-phos inhibitions. The variability of the inhibition selectivity (IC 50 for the Cterm-EH versus that for the Ntermphos) supports the idea that SMTPs bind to two distinct sites in sEH.  (Fig. 7A). Along with the inhibition of the 14,15-EET hydrolysis, the level of 14,15-EET was elevated in the presence of SMTP-7 (Fig. 7A). SMTP-0 and SMTP-44D were also active in inhibiting 14,15-DHET formation from 14,15-EET in HepG2 cells (IC 50 1.2 and 9.2 M, respectively) (Fig. 7B).

Inhibition of EET
Effects on the Level of Arachidonate-derived Lipid Mediators in Plasma from Guillain-Barré Syndrome Model Rats-To confirm the impact of sEH inhibition by SMTP-7 and its selectivity on sEH, we measured the levels of metabolites in the cyclooxygenase, lipoxygenase, and cytochrome P450 pathways using  ). B, Hill plots of the control data in panel A. C, a diagram of the equilibrium in which allostericity for the substrate binding between the two equivalent catalytic sites of the Cterm-EH is allowed (as sigmoidal velocity curves with a Hill coefficient of 1.9 that suggested an allosteric kinetic model for the substrate binding). As sEH consists of two identical monomers (E 1 and E 2 ), we hypothesized that the equilibrium substrate dissociation constants (K S1 , the primary substrate dissociation constant, and K S2 , the secondary substrate dissociation constant) for each subunit were identical (this means that E 1 and E 2 are indistinguishable from each other in this model). Allostericity for the inhibitor binding to each monomer is not considered in this model. In addition, no change in the dissociation constant for the substrate binding to the remaining vacant site is taken into account. S, substrate (14,15-EET); I, inhibitor (SMTP-0); P, product (14,15-DHET); K I , inhibitor dissociation constant. The equation is derived from the model. v, reaction velocity; V max , maximum reaction velocity. To determine each constant, we tried various combinations of kinetic parameters to fit the velocity results to the equation. Among the parameters tested, the most probable values obtained were: V max 11.5 mol min Ϫ1 mg Ϫ1 ; K S1 59 M; K S2 0.21 M; K I 21.2 M; r 2 0.99. D, competition of SMTP-0 with AUDA in the Cterm-EH inhibition. The inhibition of Cterm-EH by SMTP-0 was determined with 4.7 nM sEH and 12.5 M PHOME in the presence of the indicated concentrations of AUDA. Inset shows the secondary plots. Dashed lines represent the theoretical lines derived from the mathematical model in panel E fitted to the experimental data. E, a diagram of the equilibrium in which two inhibitors (SMTP-0, I 1 , and AUDA, I 2 ) competes with each other for binding to the catalytic site in an enzyme (Cterm-EH, E). S, substrate (PHOME); P, product; K I1 , inhibitor dissociation constant for I 1 ; K I2 , inhibitor dissociation constant for I 2 . The equation is derived from the model. v, reaction velocity; V max , maximum reaction velocity. To determine each constant, we fitted the velocity results obtained in the presence of 5 M of S (PHOME) with various concentrations of I 1 and I 2 to the equation using the value of V max and K S obtained in the absence of the inhibitor (V max 1.8 ϫ 10 4 ⌬fluorescence min Ϫ1 , K S 59 M). The values obtained were: K I1 , 6.8 M; K I2 , 6.3 nM; r 2 , 0.97. Each value represents the mean Ϯ S.D. from triplicate determinations. plasma obtained from Guillain-Barré syndrome model rats. A global analysis revealed that SMTP-7 did not significantly affect the levels of 47 out of 48 metabolites examined (see "Experimental Procedures" for metabolites analyzed), resulting in a small change in the distribution of the three classes of arachidonate metabolites (Fig. 8A). The only one metabolite that was significantly affected by the treatment with SMTP-7 was 11,12-DHET. The level (% distribution among the 48 metabolites) of 11,12-DHET in SMTP-7-treated rats was significantly lower than that in saline-treated rats (13.5 Ϯ 0.17% compared with 20.2 Ϯ 0.04% in saline group, p Ͻ 0.05; Fig. 8B). The levels of 5,6-, 8,9-, and 14,15-DHETs, however, were not significantly changed by the treatment with SMTP-7 (Fig. 8B). The levels of all regioisomers of EETs were too low to be detected by the method employed. The result that SMTP-7 selectively decreased 11,12-DHET can partly be explained by inhibition of sEH. Details of this interpretation are described under "Discussion." DISCUSSION SMTP-7 has a plasminogen modulation activity that leads to a thrombolytic enhancement as observed in several animal models (14,16,20). SMTP-7 effectively treats thrombotic and embolic stroke models (17,18,20), and the involvement of an additional mechanism that leads to anti-inflammation has been suggested (16,18,19,21). In the present study we observed the SMTP anti-inflammatory action that is independent of the plasminogen modulation activity. Although the in vivo models used in this study (Guillain-Barré syndrome, ulcerative colitis, and Crohn disease) are apparently irrelevant to thromboembolic complication, there remained a possibility that a local generation of plasmin might affect disease progress. Therefore, we compared the effect of SMTP-7 with that of the congener SMTP-44D, which is essentially inactive in plasminogen modulation activity (11). The results clearly demonstrate that both compounds are effective in treating these inflammatory disease models (Fig. 1). Thus we conclude that SMTPs have an antiinflammatory activity independent of their plasminogen modulation activity. A target molecule for the anti-inflammatory action should, therefore, exist. In addition, the structural requirement for the anti-inflammatory action could be different from that for the plasminogen modulation activity.
To identify the anti-inflammatory target, we designed an affinity matrix that can bind a target protein. In our preliminary experiments, we observed anti-inflammatory effects with SMTP congeners with varying N-linked side-chain structures. We, therefore, employed the strategy of coupling the core SMTP structural moiety (tricyclic ␥-lactam with a geranylmethyl side-chain) to a bead via an N-linked side chain. The major protein that bound to the resulting affinity matrix was sEH (Fig. 2). We purified a homogeneous sEH preparation by performing a single-step chromatography on this matrix using AUDA, which binds to the catalytic site of the Cterm-EH, as an eluent. This result is consistent with the fact that SMTP-0 competitively inhibits the Cterm-EH (Fig. 4). In addition to its inhibition of Cterm-EH, SMTP-0 inhibits the Nterm-phos pseudononcompetitively (Fig. 5). The result that the inhibition of the Cterm-EH, but not the Nterm-phos, by SMTP-0 is competed FIGURE 6. Structure-activity relationships of SMTP congeners in the inhibition of Cterm-EH and Nterm-phos. The IC 50 values of the Cterm-EH and Nterm-phos inhibitions, determined with PHOME and AttoPhos, respectively, are plotted for each congener, shown as the SMTP number. The following concentrations of enzyme and substrate were used: 4.7 nM sEH and 12.5 M PHOME for Cterm-EH and 2.3 nM sEH and 5 M AttoPhos for Nterm-phos. The inhibition selectivity (the ratio of the IC 50 of the Cterm-EH over that of the Nterm-phos) of each SMTP congener is also shown as a bar graph. The N-linked side-chain structure for each congener is shown. IC 50 values (M) of each SMTP congener for Cterm-EH and Nterm-phos, respectively, were the following: SMTP-0, 6 Ϯ 1 and 14 Ϯ 1; SMTP-5D, 70 Ϯ 7 and 8 Ϯ 0; SMTP-7, 23 Ϯ 1 and 6 Ϯ 1; SMTP-16, Ͼ 100 and Ͼ 100; SMTP-26, 12 Ϯ 1 and 12 Ϯ 2; SMTP-27, 8 Ϯ 1 and 14 Ϯ 1; SMTP-28, 5 Ϯ 2 and 9 Ϯ 1; SMTP-33, Ͼ100 and Ͼ100; SMTP-44D, 27 Ϯ 2 and 24 Ϯ 3; SMTP-54, 17 Ϯ 1 and 67 Ϯ 5; SMTP-55, 19 Ϯ 1 and 60 Ϯ 6. Each value represents the mean Ϯ 95% confidential interval from triplicate determinations. for by AUDA suggests a mechanism in which SMTP-0 binds to two distinct sites in sEH; one is the substrate binding site in the Cterm-EH and the other is an allosteric site that affects the Nterm-phos. We observed a positive cooperativity in Cterm-EH, whereas no cooperativity (43) or a negative cooperativity in Cterm-EH (44) has been reported in previous studies. These variable findings may be due to the use of enzyme from different sources (native or recombinant) and/or buffer compositions. As expected from the crystal structure, sEH has higher order intradomain and interdomain interactions (45), and environmental conditions would affect the conformational status of sEH to exhibit cooperativity. Nevertheless, kinetic parameters obtained with these three investigations are relatively close (V max 11.5 mol min Ϫ1 mg Ϫ1 and apparent K m ((K S1 ϫ K S2 ) 1/2 ] 3.52 M in this study (Fig. 4); V max 9.0 mol min Ϫ1 mg Ϫ1 and K m 4 M for 14(R),15(R)-EET, and V max 1.36 mol min Ϫ1 mg Ϫ1 and K m 5 M for 14(S),15(S)-EET by Zeldin et al. (43); V max 20 mol min Ϫ1 mg Ϫ1 and apparent K m ((K S1 ϫ K S2 ) 1/2 ) 6.3 M by Marowsky et al. (44)).
The structure-activity relationship results, which reveal variable inhibition selectivity (the ratio of the IC 50 of the Cterm-EH over that of the Nterm-phos) among congeners with different N-linked side chains (Fig. 6), support this idea. Although the physiological function of the Cterm-EH has been extensively characterized by the use of the combination of sEH-deficient animals and specific inhibitors of the Cterm-EH (46 -56), the role of the Nterm-phos remains elusive because of the lack of information about the relevant substrate and a potent specific inhibitor. The existence of an allosteric site that affects the Nterm-phos has not been reported. Our structure-activity relationship data suggest the possibility that a selective Nterm-phos inhibitor can be developed based on the allosteric mechanism.
EETs are signaling molecules implicated in anti-inflammation (57,58). SMTP-7 and its congeners inhibited the hydrolysis of 14,15-EET to the inactive 14,15-DHET in HepG2 cells (Fig.  7). In addition, the plasma level of 11,12-DHET in Guillain-Barré syndrome model rats was significantly decreased by SMTP-7 treatment (Fig. 8), whereas the levels of 5,6-, 8,9-, and 14,15-DHETs were unaffected. The levels of 5,6-and 8,9-DHETs were low, which is consistent with the finding by Li et al. (59). The lack of effect of SMTP-7 on the levels of these metabolites can be explained by the low catalytic activity of sEH toward these regioisomers (43,60), letting sEH to contribute lesser to these levels. We speculate the lack of effect on the 14,15-DHET level is due to complex mechanisms of the formation/catabolism (degradation and excretion) of EETs and DHETs under physiological conditions. According to previous literatures (43,60,61) the specific activity of human sEH for the hydrolysis of 14,15-EET is approximately Ͼ2 times higher than that for 11,12-EET. In our animal model, the circulating level of 14,15-DHET was only 1.1 times higher than that of 11,12-DHET in control animals. Thus, it is likely that the rates of degradation and/or excretion of 14,15-DHET in this model is higher than those of 11,12-DHET. This may make it difficult to reflect the impact of sEH inhibition by SMTP-7. The selective change in the 11,12-DHET level by SMTP-7 suggest a specificity of SMTP-7 in the arachidonate metabolisms. Along with the in vitro data, this in vivo result supports the idea that sEH is an anti-inflammatory target of SMTP-7.
The deficiency of sEH and the inhibition of the Cterm-EH have been reported to be protective against disease progression in animal models of inflammatory bowel diseases (62). In thrombotic and embolic stroke models in rodents and primates, SMTP-7 exhibited excellent activities, with a wide therapeutic time window and a reduced cerebral hemorrhage (17,18,20), that were not achieved by the conventional thrombolytic therapy. The finding that SMTP-7 inhibits sEH suggests that this activity, aside from the thrombolytic enhancement by plasminogen modulation, may account for these additional pharmacological potentials of SMTP-7. The observations that the sEH deficiency or the Cterm-EH inhibition is protective against experimental stroke (63-65) support this hypothesis. The pharmacological significance of the SMTP-7 ability to inhibit the Nterm-phos remains to be investigated. SMTP congeners that selectivity inhibit the Nterm-phos can be useful tools to investigate the physiological role of the Nterm-phos.
In conclusion, our results demonstrated SMTP-7's anti-inflammatory action that is independent of its plasminogen modulation activity. The finding that SMTP-7 inhibits sEH and the resulting hydrolysis of the anti-inflammatory signaling molecule EET suggests that sEH inhibition is involved in the antiinflammatory action of SMTP-7. The combination of throm- bolysis and sEH inhibition explains the excellent activity of SMTP-7 in treating thrombotic and embolic strokes. SMTP-7 is under development as an alternative drug that could be effective in stroke patients who do not benefit from the standard thrombolytic therapy.