Naturally occurring monoepoxides of eicosapentaenoic acid and docosahexaenoic acid are bioactive antihyperalgesic lipids.

Beneficial physiological effects of long-chain n-3 polyunsaturated fatty acids are widely accepted but the mechanism(s) by which these fatty acids act remains unclear. Herein, we report the presence, distribution, and regulation of the levels of n-3 epoxy-fatty acids by soluble epoxide hydrolase (sEH) and a direct antinociceptive role of n-3 epoxy-fatty acids, specifically those originating from docosahexaenoic acid (DHA). The monoepoxides of the C18:1 to C22:6 fatty acids in both the n-6 and n-3 series were prepared and the individual regioisomers purified. The kinetic constants of the hydrolysis of the pure regioisomers by sEH were measured. Surprisingly, the best substrates are the mid-chain DHA epoxides. We also demonstrate that the DHA epoxides are present in considerable amounts in the rat central nervous system. Furthermore, using an animal model of pain associated with inflammation, we show that DHA epoxides, but neither the parent fatty acid nor the corresponding diols, selectively modulate nociceptive pathophysiology. Our findings support an important function of epoxy-fatty acids in the n-3 series in modulating nociceptive signaling. Consequently, the DHA and eicosapentaenoic acid epoxides may be responsible for some of the beneficial effects associated with dietary n-3 fatty acid intake.

ton, WI) with a differential refractometer (model R401) detector (Waters, Milford, MA). The oxylipids were stored under nitrogen at Ϫ 80°C until use. Details of this method are given in the supplemental data. An LC/MS/MS method, similar to the one used for EET ( 28 ), was developed and optimized to analyze and quantify these novel oxylipins.

LC/MS/MS instrumentation
All HPLC/MS analyses were performed with a Waters ULPC separation module equipped with a 2.0 × 150mm, 5µm Luna C18 column (Phenomenex, Torrence, CA) held at 40°C. The sample chamber was held at 10°C ( 28 ). The HPLC was interfaced to the electrospray ionization probe of a Quattro Ultima tandemquadrupole mass spectrometer (Waters). Solvent fl ow rates were fi xed at 350 l/min with a cone gas fl ow of 125 l/h, desolvation gas fl ow of 650 l/h, a source temperature of 125°C, and a desolvation temperature of 400°C. Electrospray ionization was accomplished in the negative mode with a capillary voltage fi xed at 3.0 kV. For MS/MS experiments, argon was used as the collision gas at a pressure of 2.3 × 10 3 Torr while quadrupole mass resolution settings were fi xed at 12.0 (i.e., 1.5 Da resolution). The photo multiplier voltage was 650 V. Optimal cone voltage and collision voltages were established experimentally.

Optimization of MS/MS parameters
To maximize analytical sensitivity using multi-reaction monitoring, we optimized the declustering potential (DP) needed to produce the molecular ion and the collision energy (CE) to maximize characteristic fragmentation ion abundance ( 28 ). First, the intensity of the deprotonated molecular ion was evaluated at DP between Ϫ 50 and Ϫ 100V. Epoxides showed maximal ionization with a cone voltage of Ϫ 65 V (supplementary Table I). The ionization of most of the diols was also maximal for a DP of Ϫ 65 V, except for ␣ 12,13-, ␣ 15,16 and ␥ 12,13-diHODe that required a DP of Ϫ 80 V for optimal ionization. Then at optimal DP values, collision-induced dissociation was optimized to maximize the transition ion abundance. For each analyte, pseudo molecular ion was isolated in the fi rst quadrupole and accelerated into the collision cell gas, leading to collision-induced dissociation of the ion. The composition and intensity of the ions produced during this process are a function of the inherent molecular stability and the energies applied. CEs were increased from Ϫ 10 V to Ϫ 30 V. As observed before for linoleate and arachidonate epoxides and diols ( 28 ), the most characteristics daughter ion for each analyte was the one resulting from breaking the epoxide ring or the bond between the two alcohols. The optimized CE and characteristic transition ions for each target analyte is displayed in supplementary Table I.
In parallel to their biosynthesis by CYP enzymes, the degradation of n-3 epoxy-fatty acids could potentially be mediated by the soluble epoxide hydrolase (sEH; epxh2). The sEH is the primary mammalian enzyme that degrades n-6 fatty acid epoxides to their corresponding vicinal diols ( 25 ). In fact, the bioactivity of the EETs is strongly regulated by their catabolism by sEH ( 25 ). In vivo, inhibition of sEH has been demonstrated to stabilize EETs and presumably other lipid epoxides to lead to diverse biological activities including antihyperalgesia, antihypertension, and anti-infl ammation in animal models. Inhibition of sEH is being considered as a potential therapeutic approach for hypertension, vascular infl ammation, and pain ( 10,13,26,27 ). Here, we tested the hypothesis that EpDPE and EpETE are substrates for sEH, and that they possess parallel biological activity with EET in reducing pain associated with infl ammation. A sensitive analytical method described herein for simultaneous detection of the epoxide and diol metabolites from tissue extracts enabled us to support these hypotheses.

Chemicals
Unsaturated lipids were purchased from NuChek Prep (Elysian, MN). The 11(12)-EET*d8 was purchased from Cayman Chemical (Ann Arbor, MI). HPLC grade hexane, ethyl acetate, and glacial acetic acid were purchased from Fisher Scientifi c (Pittsburgh, PA). Acetonitrile and methanol were purchased from EM Science (Gibbstown, NJ) and were used for all reversephase HPLC analyses. All other chemical reagents were purchased from either Sigma (St. Louis, MO) or Aldrich Chemical Co. (Milwaukee, WI) unless indicated.

Oxylipin preparation and analysis
EpETE and EpDPE regioisomers were generated by reacting the methyl ester of EPA and DHA receptively with meta -chloroperbenzoic acid ( m CPBA) as described ( 15 ). Epoxy and diol lipids were purifi ed on a Gilson (model 302) HPLC (Gilson, Middle-used for each substrate) was added to 100 l of the enzyme solution at 0.2 g/ml ([E] fi nal ≈ 3 nM) in Bis-Tris HCl buffer (25 mM, pH 7.0) containing 0.1 mg/ml of lipid free BSA. The reaction mixtures were incubated at 30°C for 5 to 15 min. The reactions were then quenched by adding 400 l of methanol and 1 l of internal standard (9,10-dihyoxy-heptadecanoic acid). The quantity of diol formed was then determined by LC/MS/MS analysis as described above. The kinetic constants (K M and V m ) were calculated by nonlinear fi tting of the Michaelis equation using the enzyme kinetic module of Sigma-Plot version 9.01 (Systat Software Inc., Chicago IL). The k cat constant was then calculated from V m using a concentration of enzyme of 3 nM.

Behavioral nociceptive testing
This study was approved by University of California Davis Animal Care and Use Committee. Male Sprague-Dawley rats weighing 250-300 g were obtained from Charles River Inc., and maintained in UC Davis animal housing facilities with ad libitum water and food on a 12 h:12 h light-dark cycle. Data were collected during the same time of day for all groups. Behavioral nociceptive testing was conducted as described previously by assessing thermal hind paw withdrawal latencies using a commercial Hargreaves apparatus ( 32 ) and by assessing mechanical withdrawal threshold using an electronic Von Frey anesthesiometer apparatus (IITC, Woodland Hills, CA). Three measurements were taken at 1 to 2 min interstimulus intervals. Mean area under the normalized time-response curve (0-120 min after intraplantar compound administration) was calculated using the trapezoidal method and areas under the time-response curve in arbitrary units are presented in the fi gures. Data were normalized to percentage values using the formula: (thermal withdrawal latency at time point × 100)/ thermal withdrawal latency before treatment. For the mechanical withdrawal threshold measurements, the controller was set to 'maximum holding' mode so that the highest applied force (in grams) upon withdrawal of the paw was displayed. EET, EpDPE, and EpETE were tested using the intraplantar carrageenan elicited pain associated with infl ammation model ( 26 ) by injecting them directly in the paw or in the spinal cord.
Injection in the paw. Following baseline measurements, carrageenan (50 l, 1% solution of carrageenan, 0.5 mg) was administered into the plantar area of one hind paw (n = 6 per group). Four hours following this, postcarrageenan responses were determined. Immediately after (t = 0), the vehicle (10% EtOH in saline), the free fatty acids or regioisomeric mixtures of EET, EpDPE, and EpETE methyl esters, or corresponding diols were administered into the same paw by intraplantar injection in a volume of 10 l vehicle solution in quantities indicated in fi gure legends, as described before ( 26 ). Responses following compound administration were monitored over the course of 2 h. Injection in the spinal cord. Intrathecal catheters were implanted according to Yaksh and Rudy ( 33 ). Following baseline measurements, animals were injected with carrageenan in saline in the plantar surface of one hind paw (0.5 mg; n = 6 per group). Four hours after carrageenan, baseline thermal withdrawal latency and mechanical withdrawal thresholds were measured. Following this, 10 l of artifi cial cerebrospinal fl uid containing vehicle (1% DMSO) or EpDPE regioisomeric mixture was administered through the catheters (t = 0). After 45 min, thermal withdrawal latency and mechanical withdrawal thresholds were measured again. contained all EpOME, EET, EpETE, EpDPE, DiHOME, DHET, DiHETE, and DiHDPE isomers.

Determining accuracy and precision
Excellent linearity (r 2 > 0.999) was observed for all analytes over the extended concentration range tested (from 5 to 1,000 fmol/ l). This results in a limit of quantifi cation of about 1 nM for each analyte in the enzyme incubations. A series of matrix spikes and standard addition experiments were performed in order to evaluate precision and accuracy (or bias due to sample matrix) of the present method. In the matrix spike experiments, both phosphate buffer and rat serum were spiked to yield a fi nal concentration of ‫ف‬ 3, 20, and 50 nM of each analyte, extracted and analyzed. The resulting concentrations were compared with standard addition experiments in which the samples were fortifi ed with 0, 3, 20, and 50 nM analyte post extraction. All samples were extracted and analyzed in triplicate. The standard additions (post-extraction fortifi cations) allowed evaluation of matrix interference on the ability to quantify a known amount of compound. For all the concentrations, analytes were detected accurately with variations ranging from 20 to 50%.

Tissue preparation
Brain and spinal cord tissues ( ‫ف‬ 100 mg) were collected postmortem from male Sprague-Dawley rats (300-350 g). Animals were anesthetized with Nembutal and perfused with cold saline to eliminate blood from the tissues. All samples were fl ash-frozen with liquid nitrogen and stored at Ϫ 80°C until used. After thawing and weighing, 10 µL of anti-oxidant solution (0.2 mg/ml of BHT and EDTA), 10 µL of surrogate standards solution, and 400 l of ice-cold methanol with 0.1% of acetic acid and 0.1% of BHT were added onto the tissue samples and incubated at Ϫ 80°C for 30 min. Samples were then homogenized by using ultrasonichomogenizer at 30 Hz for 10 min and centrifuged at 10,000 rpm for 10 min at 4°C. The supernatants were collected and pellets were reextracted with 100 µL of ice-cold methanol with 0.1% of acetic acid and 0.1% of BHT and centrifuged. The supernatant of each extract combined, and an equal volume of 0.2 N NaHCO 3 in water was added to hydrolyze lipid esters. The mixtures were allowed to react at 4°C overnight. The solutions were then acidifi ed with 1.5 ml of 0.2 N acetic acid in water and loaded onto solid phase extraction (SPE) cartridges for SPE.

SPE
Oasis cartridges (Waters) were washed with ethyl acetate and methanol and equilibrated with [20% methanol (MeOH), 80% water] to the initial condition. Diluted supernatants of the samples were loaded on the cartridges and washed twice with SPE solution (water with 20% MeOH). The cartridges were dried under vacuum followed by elution using 0.5 ml of methanol and 3 ml of ethyl acetate into tubes with 10 µL of trapping solution (30% glycerol in MeOH). The eluates were dried by vacuum and reconstituted with an internal standard solution. Recovery for the epoxide surrogate (D 11 11,12-EET) was 77%, and recovery for the diol surrogate (D 11 14,15-DHET) was 82%. No D 11 11,12-DHET (the diol resulting from the epoxide surrogate) was detected, indicating that no signifi cant hydrolysis of the epoxides took place during the sample preparation and analysis. This is consistent with previous observations ( 28 ).

Kinetic assay conditions
Kinetic parameters for a series of monoepoxy-fatty acids were determined under steady-state conditions using the purifi ed recombinant human sEH ( 30,31 ). One l of substrate solution in ethanol ([S] fi nal from 1.0 to 50 M; 7 to 8 concentrations compared with the other epoxy-fatty acids, the EpETEs in the CNS were in similar quantities to the EET regioisomers ( Table 1 ). Similarly, for the C 18 fatty acids, the n-3 ␣ -EpODEs were present in similar quantities to the n-6 EpOME.

Kinetic constants of epoxy-fatty acids for the human sEH
Next, based on the ability of sEH to hydrolyze epoxy C:18 (EpO and EPOME) and C:20 fatty acids (EET), we hypothesized that EpODE, EpDPE, and EpETE might also be hydrolyzed by sEH ( 28 ). We tested this hypothesis by using individual regioisomers of these fatty acids and recombinant human sEH, quantitatively measuring the products, corresponding diols, with the LC-MS/MS method described above. As shown in Table 2 , all epoxyfatty acids tested were substrates for the human sEH though signifi cant differences were observed. The results for all the substrates tested fi tted well with the Michaelis-Menten equation (r 2 > 0.94; Fig. 1 ). Although apparently better fi tting (higher r 2 ) was obtained with a cooperative model ( 37 ), the velocity values calculated from the cooperative model were not statistically different ( P > 0.05) than the values obtained with the Michaelis model. This suggests that the apparent better fi tting with the cooperative model is an artifact. Because the sEH has a two-step mechanism involving the formation and hydrolysis of a covalent intermediate ( 38 ), K M in this case was not a measure of the affi nity of the substrate for the enzyme. Rather, K M refl ects the concentration of substrate for which the velocity is half maximal. As expected, using various epoxy-fatty acids, we observed increases of the observed K M values with a parallel increase in k cat values. However, 8,9-EET, 8,9-EpETE, and 19,20-EpDPE are exceptions to this rule

Statistics
Results are presented as average ± standard deviation unless noted otherwise. Kinetic parameters were determined by nonlinear regression of the Michaelis-Menten equation using SigmaPlot (v9.01, Systat Software Inc., Chicago, IL). F-test and t -test ( P < 0.05) were performed to ensure the quality of the fi tting. Data from nociceptive testing were analyzed using ANOVA followed by Dunnett's 2-sided t -test for group comparisons. Statistical signifi cance was set to P р 0.05.

Occurrence and distribution of epoxy-fatty acids in rat tissues
Several studies established that DHA and EPA can be oxygenated by CYP enzymes in a manner parallel to oxygenation of arachidonic acid ( 5,(14)(15)(16)(17). This prompted us to hypothesize that EpDPEs and EpETEs occur in mammalian tissues. Because n-3 fatty acids are known to accumulate in the central nervous system, we concentrated on the brain and spinal cord. Eicosanoids are known to be incorporated in various lipid esters, affecting their bioactivity ( 34 ). Because fatty acids and their metabolites could still be released from membranes postmortem ( 35 ), we measured total oxylipin concentrations after base hydrolysis of the lipid esters. As presented in Table 1 , these organs contained signifi cantly more EpDPE than EpETE, consistent with the observation that there is more DHA than EPA in the central nervous system (CNS) ( 36 ). Although only the 17,18-EpETE regioisomer of parent EPA was detected in the brain and spinal cord, all of the EpDPE regioisomers of parent DHA were present. Except for the 7,8-EpDPE, which was present in relatively high quantity Animals were sacrifi ced and perfused with cold saline. Tissues were excised and extracted two times with acidifi ed methanol, followed by base hydrolysis of the lipid esters. Surrogates were added before extraction. Extracts were further purifi ed using C18 SPE cartridges. Internal standards were added and samples were subjected to LC/MS/MS analysis. Surrogates recoveries of ‫ف‬ 80% were obtained. Values are tissue average ± standard deviation (n = 3). The abbreviations used for oxidized fatty acids follow the recommendations of Smith ( 47 ), adopted by the International Union of Pure and Applied Chemistry (IUPAC). BLD, below the limit of detection of 0.1 pmol/mg of tissue.
it is clear that in vivo inhibition of sEH will affect the metabolism of all of these epoxy-fatty acids, and the extent of this effect will mostly depend on the relative abundance of each regioisomer.

EETs, EpDPEs, and EpETEs are peripherally acting antihyperalgesic molecules
Next, we tested the hypothesis that EpDPEs are analogous to EETs in function. One of the landmark bioactivities of EETs and sEH inhibitor is their strong antihyperalgesic effects ( 26,32 ). Because EpDPE and EpETE are apparently better substrates for sEH than EET, we hypothesized that they may also have better antihyperalgesic effects.
We tested this hypothesis using the carrageenan elicited local pain associated with infl ammation model ( 26 ). Carrageenan administration, as expected, signifi cantly (>4-fold) decreased both thermal withdrawal latencies and mechanical withdrawal thresholds ( Fig. 2 ). We previously showed that the carrageenan painful effect lasts at least 8 h ( 26 ). The bioactivity of regioisomeric mixtures of EETs, EpETEs, and EpDPEs were then determined. Intraplantar administration of a low dose of 300 ng/paw of the epoxyfatty acids resulted in signifi cant antihyperalgesic activity that lasted for at least 2 h. Among the epoxy-fatty acids, the EpDPEs were surprisingly the most effective in reducing pain associated with infl ammation, followed by the EETs and the EpETEs ( Fig. 2A, B ). By contrast, the parent free fatty acids were inactive even when administered at 10-fold higher quantity (3,000 ng/paw; Fig. 2C ), and the corresponding diols were also inactive when administered at an identical dose (300 ng/paw; data not shown). Overall, these data support the hypothesis that epoxygenated metabolites are involved in pain signaling. because they have relatively poor K M (>25 M) with intermediate k cat s. The second step of the sEH reaction mechanism is at least an order of magnitude slower than the fi rst step ( 38 ). Thus, k cat values represent largely the rate of hydrolysis of the covalent intermediate. However, the k cat of the EpOs, those that are from monounsaturated fatty acids (MUFAs), were much smaller than those from the PUFAs, suggesting that intermediate formation could be rate limiting for the MUFA. For the PUFA, k cat values varied an order of magnitude depending on the position of the epoxide on the fatty acid structure. In general, k cat s were the largest when the epoxide was on the 14th or 15th carbon of the fatty acid chain.
The ratio k cat /K M is a specifi c measurement of the overall rate of the reaction for a particular substrate. Using this ratio and its value for cis 9,10-EpO as a reference, a relative preference index was calculated. More than a 60-fold difference between the best (13,14-EpDPE) and the worst substrate (19,20-EpDPE) was evident. Overall, it was clear that the sEH preferred epoxides situated in the middle of the fatty acid chain, and that its activity decreased significantly if the epoxide was closer to the acid function or the -carbon. This preference probably refl ects the structure of the human sEH active site, which is proposed to have large hydrophobic pockets on either side of the catalytic residues ( 38 ). For the EpOME and EET, the rodent and human sEH showed similar preferences ( 28 ). Thus, one could assume that these two enzymes will have overall similar preferences for all the epoxy-fatty acids studied herein. When comparing n-6 to n-3 epoxy-fatty acids in either the ␣ -and ␥ -epODE or the EET to the EpETE series, there was no clear preference for any particular kind of fatty acid. Although overall sEH seemed to prefer EpDPE over EET, mals. The EpDPE (3 g/animal) did not change thermal withdrawal latency or mechanical withdrawal threshold of noninfl amed animals ( P > 0.1 for both tests, not shown). However, they were highly effi cacious when administered into the spinal cord of infl amed animals ( Fig. 3 ). This high concentration of EpDPE completely reversed the decrease in thermal withdrawal latency and partially reversed mechanical withdrawal threshold. These fi ndings indicate a selective role of the epoxides of DHA in modulating nociceptive physiology.

DISCUSSION
Many of the past efforts on bioactive lipids focused on ARA metabolites and their biological relevance ( 8,13,25 ). Here, we report on the presence, distribution, and regulation of the levels by sEH of a set of 18, 20, and 22 carbon epoxy-fatty acids in both n-3 and n-6 series. We also report a direct antinociceptive role for n-3 epoxy-fatty acid metabolites and more specifi cally those of DHA.
Lipid molecules are ubiquitous messengers that are known to participate in intracellular signaling, cell-to-cell communication, serve as neurotransmitters, and regulate specifi c physiologic functions, one of which is the transmission of noxious sensory information both in the periphery and the CNS ( 1,9,25,39 ). Analysis of alterations in the type, amount, and organization of lipids can thus provide critical information leading to the understanding of mechanism of action of each molecule, diagnostic tools, therapeutic strategies, and identifi cation of the mechanisms underlying the disease processes. Analytical methods based on LC/MS technology are highly sensitive methods that simultaneously provide information on quantity and identity of multiple analytes in a complex sample ( 28 ). Accordingly, the methods developed herein show promise in the quantitative analysis of epoxy-fatty acids both from in vitro experiments and from tissue extracts. Using this technology, we studied the role of n-3 epoxy-fatty acids in nociceptive pathophysiology.

sEH effi ciently degrades n-3 and n-6 series epoxy-fatty acids
In mammals, the sEH was historically found in liver, kidneys, and at lower levels in other organs ( 25 ). Recently, the sEH was found in the brain, specifi cally in astrocytes and in the body of the neuronal cells ( 40,41 ). The sEH is the primary mammalian enzyme that degrades EET in vivo ( 25 ). In murine brain, sEH is responsible for at least twothirds of the EET hydrolysis ( 41 ). Pharmacological inhibition of sEH has been demonstrated to stabilize EETs and is being considered as a potential therapeutic approach for hypertension, vascular infl ammation, and pain ( 10,13,26,27 ). Inhibition of sEH is benefi cial also in ischemic brain injury and stroke ( 42 ).
Although it was not surprising that sEH would degrade epoxy-fatty acids other than those in the C 18 and C 20 n-6 series such as the EET and EpOME, an unexpected outcome of this study was that EpDPE and EpETE are in fact degraded effi ciently by sEH (Table 2) ( 25 ). We showed We then investigated whether individual regioisomers of EpDPE were active because there were different quantities of each regioisomer in the spinal cord and the brain. Three regioisomers were selected based on information presented in Table 2 . Even though each regioisomer was active, there were signifi cant differences among the regioisomers ( Fig. 2D ). The position of the epoxide functionality was important in the antihyperalgesic activity. The rank order of effi cacy in reducing pain associated with infl ammation (13,14-EpDPE > 16,17-EpDPE > 19,20-EpDPE) was in parallel to the rank order of preference by the sEH ( Table 2 ).

EPDPEs are also antihyperalgesic in the central nervous system
Because the EpDPEs had signifi cant peripheral antihyperalgesic activity, we investigated whether the EpDPEs were also active in the CNS. We administered the regioisomeric mixture of EpDPE into the spinal cord through chronically implanted intrathecal catheters, fi rst to noninfl amed animals and then to carrageenan infl amed ani- affect EpDPE levels in the brain similarly to the way it affects EET levels ( 41,42 ). Like the EET ( 34 ), it is likely that under normal physiologic conditions sizable amounts of EpDPE are esterifi ed in lipids. These esters could have direct biological activities or act as stores from which the oxylipins could be released.

n-3 Series epoxy-fatty acid regioisomers are selectively antihyperalgesic
A signifi cant aspect of the role of lipids in neuronal function is their ability to modify the functional responses of ion channels, synaptic transmission, and cellular signaling cascades through which neuronal cell function is altered to meet physiologic demand ( 39 ). We previously showed that the EET are antihyperalgesic in infl ammatory and neuropathic pain models ( 26,32,43 ). Furthermore, the pharmacological inhibition of sEH, which resulted in the stabilization of the EET levels, is also strongly antihyperalgesic ( 32,43 ). In this study, we demonstrate that n-3 that the epoxides of EPA and DHA are indeed preferred substrates for sEH over the other epoxy-fatty acids tested including the EET and EpOME. Thus, in vivo inhibition of sEH or an alteration in its titer will not only affect the metabolism of EETs and EpOME but also the metabolism and bioactivity of EpDPE and EpETE. Obviously, the extent of such an effect will depend on the relative abundance of each epoxy-fatty acid in addition to the extent of their respective interactions with molecular targets. Although EPA concentrations are relatively low in the body, DHA concentrations, especially in neuronal and retinal tissues, are similar to or higher than those of ARA ( 36 ). Accordingly, in rat brain and spinal cord we did not fi nd detectable levels of EpETE, but we found similar and signifi cant amounts of EET and EpDPE ( Table 1 ). Although representing a small amount of the parent fatty acid, the EET are very bioactive ( 8,10 ). Our data suggest a similar scenario for the DHA epoxides. Taken together, these data suggest that pharmacological inhibition of sEH will Fig. 2. Epoxy-fatty acids but not parent fatty acids reduce pain associated with infl ammation induced by a single intraplantar injection of 0.5 mg of carrageenan in rat paws (n = 6 per group). A: Regioisomeric mixtures of EET, EpETE, and EpDPE (300 ng /paw) all signifi cantly reversed carrageenan induced pain ( P < 0.01), measured as a percent of mechanical withdrawal thresholds before infl ammation. The EET and the EpETE gave a signifi cant action for 90 min, while EpDPE were still signifi cantly active after 120 min. B: The area under the curve (integrated from 0-120 minutes), which gives an overall measurement of the effi cacy of a treatment, is reported on the y axis for each compound from panel A. Compared with the vehicle control (10% ethanol in saline), the mixture of epoxy-fatty acids signifi cantly reduced the pain (* P < 0.001; ‡ P < 0.01). EpETE had lower effi cacy than EpDPE ( P = 0.009) but not lower than EET ( P = 0.12). C: The parent free fatty acids (ARA, EPA, and DHA) at 3000 ng/paw, failed to signifi cantly reduce pain ( P = 0.1). D: All three individual EpDPE regioisomers (300 ng /paw) tested signifi cantly reduced pain (* P < 0.001; ‡ P = 0.001; # P = 0.03, compared with the vehicle). 16, 17-EpDPE and 19,20-EpDPE were similar in effi cacy ( P = 0.29), whereas 13,14-EpDPE was more effi cacious than them ( P < 0.01).
mers to decrease pain associated with infl ammation ( Fig.  2D ). Because the 11,12-and 14,15-EET are the most effective mediators of the above biology ( 45 ), it was not surprising that the antihyperalgesic activity is related to the position of the epoxide functionality among the EpDPE regioisomers. The rank order of potency of the tested EpDPE regioisomers is 13,14-EpDPE > 16,17-EpDPE > 19,20-EpDPE, with a clear decline in biological activity as the epoxide function is moved further from the acid function ( Fig. 2D ). At the same time, this rank order of antihyperalgesic effects remarkably follows the preference of the sEH for these substrate molecules ( Table 2 ). This trend was previously observed for the biological activity of the EET regioisomers ( 25,45 ). The sEH is expressed in the CNS, particularly in neurons and astrocytes of the brain ( 41 ). Even though spinal expression of sEH has not been demonstrated by immunohistochemistry ( 40 ), we detected a considerable quantity of sEH mRNA in whole spinal cord. Consistent with the presence of sEH in the spinal cord, we report the presence of epoxy-fatty acids and their corresponding diols in the spinal cord ( Table 1 ). Our earlier work suggested that the antihyperalgesic effect of sEH inhibitors are at least partially mediated by the CNS. Here, those earlier observations are supported because EpDPE effectively eliminated thermal hyperalgesia when administered into the spinal cord ( Fig. 3 ). The differential effects of peripheral and spinal administration of EpDPE in thermal versus mechanical withdrawal tests imply a selective mechanism of action of the EpDPE because these responses were shown to be mediated by different types of neurons ( 44,46 ).
Although the detail mechanism(s) by which epoxy-fatty acids lead to antihyperalgesia is still under investigation ( 43 ), it is hypothetically possible that, because of their structural similarity, EET, EpETE, and EpDPE will affect a limited number of molecular targets. Our fi ndings support this concept given that the antihyperalgesic activity reported herein is regioisomer selective ( Fig. 2D ). Furthermore, the parent fatty acids as well as the corresponding diols had no effect even when administered at higher doses, supporting the selectivity of the biological activity observed. We previously showed that the EET lead to antihyperalgesia in both infl ammatory and neuropathic pain states by at least two spinal mechanisms; fi rst, by repressing the induction of the COX2 gene and second, by rapidly upregulating an acute neurosteroid-producing gene, StARD1 ( 43 ). We hypothesize that EpETE and EpDHE are antihyperalgesic through mechanisms similar to that of the EET.
At least theoretically, inhibition or downregulation of sEH in vivo is expected to result in increased levels of multiple different epoxy-fatty acids ( 25 ). For the linoleic and arachidonic epoxy-fatty acids this is the most frequently observed outcome. Although in other cases when sEH is inhibited, a decrease in sEH products, the corresponding diols of EET and EpOME, is observed ( 25 ). Data presented here clearly suggest EpDPE and EpETE should be taken into consideration when sEH inhibitors are used in various disease models. If inhibition of sEH results in the epoxy-fatty acids are also strongly antihyperalgesic, especially EpDPE ( Fig. 2 ). Because no sEH inhibitors were used, the epoxy-fatty acids injected in the paw probably hydrolyzed by sEH to the diols, which are not active. This could explain the slow reduction of the effi cacy of the compounds over the length of the experiment (2 h). Although EpDPE and EET displayed a similar effi cacy in reducing pain ( ‫ف‬ 50%; Fig. 2A, B ), the DHA epoxides acted longer. The EpETEs were less effective than the EETs. The EETs and EpDPEs have strong peripheral anti-hyperalgesic properties with potencies similar to the endocannabinoids ( 43,44 ). For example, a 300 ng dose EpDPE regioisomer mixture gave a similar reduction in pain as 75 ng of anandamide in a similar model ( 44 ). Given the intensity of the painful state induced by carrageenan, the observed effects are highly signifi cant.
Because the bioactivity of the regioisomers of EET vary dramatically in assays for angiogenesis, infl ammation, and vasodilatation ( 45 ) and the EpDPE regioisomers aare not uniformly distributed in rat tissues ( Table 1 ), we evaluated the ability of EpDPE regioisomers in reducing pain. There were signifi cant differences in the ability of the regioiso- stabilization of a wide range of epoxy-fatty acids, then investigating the individual epoxy-fatty acids and their regioand stereoisomers should increasingly provide insight into the pleitropic effects of inhibiting sEH such as antihypertensive and anti-infl ammatory effects ( 27 ). Here, we demonstrated the presence and bioactivity of epoxy-fatty acid regioisomers and the preference of sEH toward these substrates, indicating that both sEH and fatty acid epoxides are involved in the same system of chemical mediation. In summary, the parent fatty acids DHA and EPA are receiving increasing attention in particular in regard to their benefi cial effects on cardiovascular and neurological systems. However, given that CYPs effi ciently convert DHA and EPA to their epoxygenated metabolites ( 5,17 ), the EpDPEs and EpETEs are likely to be responsible for at least some of the benefi cial effects associated with dietary DHA and EPA.