Discovery of Inhibitors of Soluble Epoxide Hydrolase: A Target with Multiple Potential Therapeutic Indications

1. Introduction

Epoxide hydrolases have been detected in prokaryotes and eukaryotes ranging from plants to mammals.^{1,2, 3} In mammals these include the soluble epoxide hydrolase (sEH), microsomal epoxide hydrolase (mEH), cholesterol epoxide hydrolase, and leukotriene A₄ (LTA₄) hydrolase. These enzymes mediate the addition of water to both exogenous and endogenous epoxides, leading to the corresponding vicinal diols except for LTA₄ hydrolase, and they display different substrate selectivity. For example, the mammalian sEH is selective for aliphatic epoxides and particularly fatty acid epoxides whereas mEH is more selective for cyclic and arene epoxides. Studies on the mEH have focused on its role in xenobiotic metabolism, but its distribution, particularly in the brain and adrenal gland suggests a possible endogenous role.⁴ Although its catalytic activity on fatty acid epoxides is low, the high level of the mEH in some brain regions may contributeto their hydrolysis. The catalytic activity of the sEH on arene oxides and other cyclic epoxides is so low that its contribution appears insignificant compared to the mEH as well as for chemical and glutathione S-transferase-catalyzed conjugation of reactive epoxides. Although the sEH can metabolize some aliphatic natural products, the sEH is thought to be involved largely in the metabolism of regulatory epoxylipids, particularly those of the arachidonic acid cascade (Figure 1). Titers of free arachidonic acid are very low, but when it is released it is converted to a wide variety of biologically active metabolites. Most research has focused on the cyclooxygenase and lipoxygenase pathways, but increasing attention is being paid to the cytochrome P450 branch of the cascade. One set of P450 enzymes carry out allylic and ω and ω-1 oxidation. Another set of P450 enzymes form regioisomeric epoxides of arachidonic acid and other unsaturated lipids. In the arachidonate series these epoxides are called epoxyeicosatrienoic acids (EETs). The EETs are metabolized by incorporation into phospholipids, chain shortening, chain elongation, hydroxylation and other pathways.⁵ However, the dominant pathway is hydration of the epoxides to the corresponding 1,2-diols by sEH. It should be noted that multiple drugs have already been discovered to act on the cyclooxygenase and lipoxygenase branches of the arachidonic acid metabolic cascade. For example, numerous non-steroidal anti-inflammatory drugs (NSAIDs) are inhibitors ofcyclooxygenase-1 (COX-1) and cyclooxygenase-2 (COX-2).⁶ In addition, montelukast is a leukotriene (LT) receptor antagonist blocking the action of LTD₄ and secondary ligands LTC₄ and LTE₄.⁷ Zileuton inhibits 5-lipoxygenase, an enzyme of the eicosanoid synthesis pathway for the production of LTs.⁸ Both montelukast and zileuton are effective therapies for treatment of asthma. Lastly, laropiprant, an antagonist of the DP1 receptor of prostaglandin D₂, is used in combination with niacin to suppress the niacin-induced vasodilation.⁹

Figure 1.

Major pathways of arachidonic acid metabolism.

As shown in Figure2, sEH in human (hsEH, EPHX2, EC 3.3.2.10) is a bifunctional homodimeric enzyme located in both cytosol and peroxisomes with both epoxide hydrolase and phosphatase activity.¹ Specifically, the C-terminus epoxide hydrolase motif of sEH transforms four regioisomers of EETs, namely 5,6-, 8,9-, 11,12- and 14,15-EETs which are the endogenous chemical mediators derived from arachidonic acid by cytochrome P450 epoxygenases, to the corresponding dihydroxyeicosatrienoic acids (DHETs), whereby the biological effects of EETs are diminished, eliminated or altered.¹⁰ Recent work has shown that some fatty acid diols have unique biological activities, but the diols are far more polar than the epoxides and thus quickly move out of cells and often are conjugated.^{11, 12} The sEH hydrates all fatty acids so far tested. For example, sEH converts linoleic acid epoxides to proinflammatory linoleic acid diols which are proposed endogenous chemical mediators as well. Regarding the enzyme activity in tissues, it was found that human liver possesses the highest sEH specific activity followed by the kidney.^13,14 Of note, specific cell types in the heart, vasculature, brain, lung, and kidney have quite high levels of enzyme. However much less is known about the N-terminal phosphatase regarding its endogenous substrates and physiological roles.

Figure 2.

Left: X-ray cocrystal structure of murine sEH (PDB code: 1CR6). One of the antiparallel monomers is shown in blue and the other in yellow. Right: X-ray cocrystal structure of a single subunit of the human sEH (PDB code: 1ZD3).

The catalytic mechanism of epoxide hydrolases was worked out by a series of biochemical studies in several laboratories based on the homology of the mEH and sEH to haloalkane dehalogenase. Similarly the hypothetical arrangement of binding sites in the enzyme was predicted by three-dimensionalquantitative structure-activity relationship (3D QSAR).¹⁵ The X-ray crystal structure of human sEH complexed with an sEH inhibitor (PDB code: 1ZD3) and later structures revealed the catalytic pocket and the key structural features required to inhibit the epoxide hydrolase activity of this enzyme in great detail (Figure 3).¹⁶ The epoxide hydrolase catalytic pocket consists of two tyrosine residues (Tyr381 and Tyr465) which activate the epoxide ring opening by Asp333. The resulting ester is then rapidly hydrolyzed into DHETs. It has been recognized that amide, carbamate and urea groups fit well in the hydrolase catalytic pocket. Specifically, the carbonyl oxygen of amide or urea is involved in a hydrogen bonding interaction with Tyr381 and Tyr465, and the N-H of ureas or amides acts as a hydrogen bond donor to Asp333. Therefore, various ureas and amides (1 and 2) have been developed as competitive, reversible and often tight binding sEH inhibitors. Several of these inhibitors bind to the sEH in the picomolar range. The X-ray structure also showed that the two domains of the sEH were joined by a proline rich bridge, the enzyme in mammals is an anti parallel homodimer, and the putative catalytic site of the N-terminal domain suggested an active phosphatase of unknown role. The W334 niche and F265 binding pocket, as depicted in Figure 3, can each accommodate a variety of functional groups of sEH inhibitors. These structures not only have assisted with optimization of sEH inhibitors but have proven valuable to evaluate the biology associated with single nucleotide polymorphisms (SNPs) in the enzyme.

Figure 3.

X-ray cocrystal structure of human sEH C domain showing the binding site which exists as an “L” shape hydrophobic tunnel. (PDB code: 1ZD3)

In mammals the mEH (EPHX1, EC 3.3.2.9) is membrane bound and largely in the endoplasmic reticulum.^17,18 In some pathological states the mEH dissociates from the membrane and appears in the blood where it is known as the preneoplatic antigen and is a marker for tissue damage including cancer.¹⁹ Activity and polymorphisms of the enzyme have been associated with a variety of diseases, and this is a quite active area of research.²⁰ With regard to this review it is important to note that although the mEH and sEH have similar catalytic mechanisms and are members of the α/β-hydrolase fold family of proteins, their evolutionary paths diverged at the level of prokaryotes. Thus it is possible to differentially inhibit the mEH and sEH with high selectivity.²¹

Evidence is excellent that the enzyme referred to as hepoxilin epoxide hydrolase is in fact the sEH. The cholesterol 5,6-epoxide hydrolase and leukotriene A₄ hydrolase are in different enzyme families working by different catalytic mechanisms and will not be discussed further. In studying the evolution of the sEH, related genes were further found. These hypothetical products of these genes are known as EH3 and EH4. Their expression in vertebrates, particularly in man is under investigation and their biological activity is so far not clear.¹⁸ In light of the recent progress in this field, this perspective will provide biological rationales, medicinal chemistry approaches, and possible paths for sEH inhibitors to enter clinical trials.

2. Biological rationales and potential indications for sEH inhibitors

It was envisaged that sEH inhibition may lead to elevated levels of EETs, which in turn could elicit various beneficial biological effects. ²² These effects may be translated into therapeutic treatment for hypertension, atherosclerosis, pulmonary diseases, diabetes, pain, inflammation, immunological disorders and other indications.

The first major therapeutic area pursued for sEH inhibitors was hypertension. This target selection was based on extensive previous biology on the endothelium derived hyperpolarizing factor and the important roles of EETs in the regulation of renal tubular and vascular function.^{23, 24} It was an early target probably also because the role of EETs and other epoxylipids in inflammation and pain was not widely appreciated. The hypertension target is supported by many preclinical studies and multiple models of hypertension, but a major limitation is that the vast majority of studies have been done in rodent species.²⁴ For example, the intra-arterial infusion of 14,15-EET transiently reduced arterial blood pressure in rats.²⁵ Furthermore, certain preclinical studies supported that sEH inhibitors could be used for hypertension treatment and disease-modifying end organ protection. It has been reported that sEH inhibitors could lower blood pressure in angiotensin II (AngII)-treated rats ²⁶ and spontaneously hypertensive rats (SHRs).¹⁰ An sEH inhibitor in AngII-infused hypertensive rats successfully attenuated the afferent kidney arteriolar diameter and reduced urinary albumin secretion, a marker of compromised renal function. On the other hand, there are also conflicting reports which showed that highly potent sEH inhibitors failed to induce blood pressure lowering effects in some SHR strains which are often used as models of human hypertension by industry.^{27, 28} Most importantly, a clinical evaluation of an sEH inhibitor,1-(1-acetyl-piperidin-4-yl)-3-adamantan-1-yl-urea (28,AR9281), most likely showed no efficacy in phase II studies.²⁹ Although compound 28 is not a highly potent inhibitor of the human sEH and it has a short blood half life, ³⁰ it showed a clear reduction of blood pressure in an AngII model of hypertension.³¹

There are multiple cautions about targeting human hypertension as an early clinical trial for sEH inhibitors in addition to the failure of compound 28 in human trials. There is of course the ongoing question in the field whether any rodent models predict efficacy in man. There are a variety of SHR strains. It was reported by that beneficial effects could be obtained in from one SHR strain without lowering hypertension. ³² Several studies from the Doris laboratory caution that there are haplotypes of SHR making them insensitive to the action of sEH inhibitors.^{33, 34} A second caution is that, with hypertension and some other indications, EETs and other fatty acid epoxides appear to modulate blood pressure back toward the normotensive state, rather than being true blood pressure lowering agents. Finally the action of sEH inhibitors are likely to be influenced by the dietary status of the patient with major effects from salt consumption and relative amounts of ω-3 and ω-6 dietary lipids as well as lipid chain length.

The most studied targets for intervention by sEH inhibitors are related to reduction in inflammation. sEH inhibitors are effective as an anti-atherosclerosis therapy in murine models and for improving blood flow and reducing inflammation in models of stroke and myocardial infarction. This is supported by the roles of EETs in dilating coronary arterioles via the activation of BK_Ca ion channels.³⁵ In addition, EETs inhibited leukocyte adhesion to the vascular wall, presumably by inhibiting NF^-κB and IKB kinases.³⁶ As an endothelium-derived hyperpolarizing factor (EDHF) to relax vascular smooth muscle, ³⁷ 11,12-EET also displayed anti-inflammatory activity, apparently by decreasing the expression of cytokine induced endothelial cell adhesion molecules, such as vascular cell adhesion molecule 1 (VCAM-1), intercellular adhesion molecule-1 (ICAM-1), and E-selectin. ³⁸ Preclinically, a short-term intracoronary infusion of EETs reduced infarct size in a canine ischemia-reperfusion model.³⁹ Thus, sEH inhibitors may elevate levels of EETs thereby potentially ameliorating vascular inflammation, endothelial dysfunction, and ultimately atherosclerosis.⁴⁰ The utility of sEH inhibitors in cardiovascular animal models has been reported. For example, an sEH inhibitor 12-(3-adamantan-1-yl-ureido)-dodecanoic acid butyl ester (AUDA-BE) reduced the infarct size after a middle cerebral artery occlusion (MCAO) in mice.⁴¹ An sEH inhibitor was shown to prevent the development of cardiac hypertrophy and failure, and to suppress cardiac arrhythmias in thoracic aortic constriction (TAC) mice.⁴² Therefore, sEH inhibitors may be useful for treating cardiomyopathy and inhibiting cardiac arrhythmia. Furthermore, Arête disclosed that an sEH inhibitor effectively decreased total cholesterol, low density lipoprotein cholesterol (LDL-C), triglyceride (TG), and glucose in AngII-infused ApoE deficient mice. ⁴³ Blood eicosanoids and particularly fatty acid epoxide to diol absolute levels and ratios are good indicators of in vivo target engagement in these and other studies. Despite these observations, the major challenges to pursue anti-atherosclerosis and many other indications driven by inflammation with sEH inhibitors are the lack of disease-modifying efficacy biomarkers, absence of data in non-rodent models, and the high cost of outcome trials.

Arête disclosed several studies of sEH inhibitors in metabolic disease models.⁴⁴ One sEH inhibitor was shown to slow down the weight gain of diet-induced obese (DIO) mice relative to vehicle. This compound reduced plasma cholesterol and serum glucose in an interperitoneal glucose tolerance test, and decreased blood pressure at 60 mg/kg (mpk) (b.i.d.) in C57B1/6 mice fed on high-fat and high-fructose diet without significantly changing the heart rate. In another related patent application,⁴³ Arête also indicated that an sEH inhibitor could effectively reduce total cholesterol, LDL, TG, and glucose in AngII-infused ApoE deficient mice. Therefore, it appears that sEH inhibition may be useful in treating metabolic syndromes including obesity, hypertension, diabetes, and hypercholesterolemia. However, a mechanism for these effects remains elusive⁴⁵ and sEH inhibitors have not proven universally successful in reducing metabolic disease in rodent models.

Another therapeutic area of interest for sEH inhibitors is neuropathic or inflammatory pain. ⁴⁶ An sEH inhibitor provided similar efficacy with respect to morphine (1 mpk) in a pain alleviation model and far greater potency in another model.⁴⁷ Interestingly, sEH inhibitors were also found to synergize activity of COX and 5-lipoxygenase (5-LOX) inhibitors.^{48, 49} In a pain model, efficacy of pain tolerance after lipopolysaccharide (LPS) exposure appeared to be similar for Vioxx^™ (10 mpk) and AUDA-BE (20 mpk). In addition, 12-[3-adamantane-1-yl ureido]-dodecanoic acid (AUDA, 3) analogs blocked LPS-elicited thermal hyperalgesia in rats. ⁵⁰ Topical application of either an sEH inhibitor or EETs reduced inflammatory pain in rats, and the combination was far more effective.⁵¹ Of particular interest sEH inhibitors reduced neuropathic pain in a number of rodent models including nerve damage and diabetic neuropathic pain. This is a largely unmet medical need, and sEH inhibitors appeared superior to the gabapentin family of drugs while not causing changes in behavior or coordination associated with opiates.⁵² Interestingly sEH inhibitors seemed to reduce the perception of pain in models where pain perception was enhanced (allodynia and hyperalgesia) but to not influence pain perception in normal animals. This may be due to cyclic nucleotides being needed for sEH inhibitors to act.⁵³ Interestingly sEH inhibitors synergized in reducing neuropathic pain with COX inhibitors such as diclofenac.⁵⁴

sEH inhibitors also displayed reasonable rheumatoid arthritis assessment score improvement in a mouse model.⁵⁵ One patent application claimed that the intraocular high pressure caused by inflammation could be attenuated by using EETs or sEH inhibitors.^{56, 57}

Boehringer Ingelheim discovered that pyrazole aniline-derived amides were sEH inhibitors which may be effective in treating T-lymphocyte mediated immunological disorders in their preliminary in vitro and in vivo studies.⁵⁸

Inhibitors of sEH reduced pulmonary infiltration by neutrophils, and reduced leukotoxin diols which are toxic to pulmonary and vascular epithelium cells associated with adult respiratory distress syndrome.^{59, 60} The dosing of an EET and sEH inhibitor appeared to be synergistic in reducing the number of neutrophils in lung, which implies their potential utility to treat obstructive pulmonary diseases, restrictive airway diseases, and asthma.⁵⁹

sEH inhibitors may also treat smooth muscle disorders such as erectile dysfunction, overactive bladder, uterine contractions, and irritable bowel syndrome.⁶¹ A patent application from Roche claimed a method of treating genitourinary disorders and particularly overactive bladder by using sEH inhibitors.⁶² They reported that an sEH inhibitor reduced the bladder pressure and decreased the bladder contraction frequency as well as amplitude in anesthetized SHRs. These data indicate that fatty acid epoxides and particularly EETs may be the hyperpolarizing factor of the urinary epithelium.

A common theme among sEH inhibitors in different models is that the compounds seem to act more to return a physiological system toward a normal state, rather than being overtly hypotensive, hypoalgesic, or anti-inflammatory. For example, there is little change in the plasma oxylipin profile following administration of sEH inhibitors to normal animals. However, in inflamed animals there is a dramatic shift toward profiles indicating resolution of inflammation rather than its propagation.^{63, 64} The eicosanoid profile seen post administration of sEH inhibitors suggests that they should synergize with NSAIDs, COX-2 blockers (COXIBs), and inhibitors of the 5-LOX pathway. This was verified experimentally.^{65, 66, 67} Co-treatment with sEH inhibitors also reduced the thrombotic events associated with the massive increase in 20-HETE by some COXIBs.⁶⁸ The additive to synergistic effect and the reduction of side effects of COXIBs with sEH inhibitors indicate that they might be attractive in drug combinations, and joint inhibitors have also been prepared.⁶⁹

3. Medicinal Chemistry

The potential therapeutic utility of sEH inhibition, based on the aforementioned biological rationales, prompted the discovery of sEH inhibitors with enhanced potency, improved solubility and pharmacokinetics (PK) properties, and high target selectivity. This is a case where a detailed understanding of the catalytic mechanism of the enzyme and later multiple crystal structures allowed the development of theoretical transition state mimic inhibitors of the enzyme from first principles based on the pioneering idea of Linus Pauling. ⁷⁰ Several pharmaceutical companies and academic institutions, including Arête, Astellas, Boehringer Ingelheim, Chinese Academy of Sciences, Dainippon Sumitomo, GlaxoSmithKline, Merck, Roche, Taisho, Takeda, University of California Davis, Columbia University, and Shanghai Institute of Biological Sciences, etc, have entered the arena of developing sEH inhibitors to target various therapeutic indications. As a result, more than 100 patent applications⁷¹ and over 30 medicinal chemistry articles on sEH inhibitors have been published to date. Overviews of medicinal chemistry are presented in at least two comprehensive reviews.^{72, 73} The medicinal chemistry section of this perspective is primarily devoted to the reported approaches to identify tool compounds with desirable properties for pharmacological studies. The biological data of benchmark sEH inhibitors will also be highlighted and used to indicate strengths and limitations of the approach of using sEH inhibitors in medicine.

The availability of homogenous, recombinant sEH from human and model species was critical for the medicinal chemistry efforts described below. Modern sEH inhibitors are competitive, slowly reversible inhibitors acting at a low concentration and competing with a substrate likely present at a very low concentration. Thus, assay conditions may violate Michaelis-Menton assumptions, and it is difficult to compare inhibitor potencies among laboratories. The ability of assays to distinguish among the most potent inhibitors has improved with time. Early sEH assays were unable to distinguish among the most potent inhibitors.⁷⁴ As new assays were developed still more active inhibitors followed which challenged the ability of enzymologists to distinguish among them. For example the early colorimetric high throughput assay was unable to distinguish effectively among inhibitors in the nanomolar range.^{74, 75} Also as the properties of these assays approach their limit to distinguish the most potent inhibitors, their reliability decreases. A high throughput fluorescent assay is commercially available.⁷⁶ With a highly purified substrate it can reach inhibitors acting in the sub-nanomolar range. Glutathione S-transferases and esterase interfere with this assay, and it is often used incorrectly. The development of assays to distinguish among the most potent inhibitors was driven by the synthesis of such potent inhibitors. The use of suspected endogenous substrates as substrates for assay is intellectually compelling, and certainly should be used to verify the activity of lead inhibitors. However, these substrates are lipophilic, relatively water insoluble, and hard to detect –LC/MS/MS shows nanomolar limit of quantification (LOQ).^{77, 78} Even with these hypothetical endogenous substrates, it is a challenge to distinguish among inhibitors with picomolar IC₅₀ values.⁷⁷

The structural types of sEH inhibitors are extremely broad, which is consistent with the wide binding pocket of the enzyme (Figure 3). The major inhibitor chemotypes include urea, amide, carbamate, thioester, carbonate, ester, thiourea, thioamide, amidine, guanadine, heterocycles, aminoheterocycles, aminoheteroaryls, chalcone oxides, acyl hydrazones, chalcone oxides and trans-3-phenylglycidols, of which the most advanced series in terms of potency and pharmacological studies are ureas and amides.⁷¹ Interestingly even the early kinetically irreversible inhibitors including chalcone oxides and glycidols provided astructure-activity relationship (SAR) that gave a general indication of the shape of the catalytic site.⁷⁹

3.1. University of California Davis

The University of California Davis (UCD) conducted pioneering research in the field of sEH inhibitors. Their early alternate substrate inhibitors were used to show in vivo effects of sEH inhibition, but these compounds were far too quickly metabolized to be useful drug leads. They subsequently reported a series of 1,3-disubstituted ureas and related amides and carbamates (1, Figure 4).⁸⁰ These pharmacophores were thought to be mimics of EETs and possible mimics of a reaction transition state or transient intermediate to the diol. For example, their first lead was a common reaction side product in organic chemistry, dicyclohexyl urea (DCU, 2). It elicited dose-dependent inhibition of sEH but no inhibition of the mEH. Since the mEH is involved in the metabolism of some dangerous xenobiotics, counter screens on the mEH have been employed since the first papers in the field. Using the tail cuff method to measure blood pressure, an average of 22 mmHg reduction of systolic blood pressure in SHRs was observed 6 h after the i.p. dose of 3 mpk of solubilized compound 2, but no effect was observed in WKY rats. Similarly, compound 2 decreased the systolic blood pressure an average of 12 mmHg in SHRs 6 hours post dosing.¹⁰ With typical SAR approaches, the group optimized steric and lipophilic parameters for the substituents on the N and N′ positions of the urea as well as the amide pharmacophores. Aromatic groups with ortho substituents on the N and N′ positions of the urea dramatically reduced inhibitory potency. Aromatic groups with substituents in other than the ortho positions led to high activity with steric parameters generally dominating over electronic in the SAR. By introducing a terminal solubilizing group to mimic the carboxylic acid in a putative endogenous substrate, the benchmark compound 3 was discovered. It was reported that this compound improved renal hemodynamics, causing reduced sodium retention and vascular dilation in rats fed on high-fat diet.⁸¹ This commercially available compound has been used in a wide variety of studies to knock out the sEH chemically and test for biological activity. The observation that compound 3 could dilate mesenteric arteries lacking the sEH provided evidence that these pharmacophores could mimic the epoxide at putative fatty acid oxide receptors.⁸² This led to a new generation of EET mimics.⁸³

Figure 4.

Evoluation of early sEH inhibitors and pharmacophores.

Although compounds such as 3were potent and competitive inhibitors with low nanomolar K_i values and subnanomolar binding, the lack of solubility in water and common formulation solvents and high melting points limited the use of compound 3 for pharmacodynamics (PD) studies. Compound 3 is also rapidly metabolized by cytochrome P450 action on the adamantyl substituent and particularly by beta oxidation. Thus the in vivo efficacy of compound 3 is brief in most systems. To improve physical properties, particularly solubility and PK of these disubstituted ureas, extensive SAR studies were carried out to introduce new polar pharmacophores and conformational constraints. For these studies drug-like properties, high oral bioavailability as well as good PK properties were seen as major drivers for improved structures. Since the adamantyl substituent gave a 10–50 fold increase in sensitivity of detection of the compounds on positive ion LC-MS, the adamantyl group was included in many derivatives during the optimization of structures on the right side of the central pharmacophore to fit into the F265 pocket. This allowed cassette dosing of animals at low drug levels and the determination of PK using only microliters of blood. This work designated the urea or amide moiety of sEH inhibitors as the primary pharmacophore that interacts with the catalytic triad of sEH (Figure 4).⁸⁴ Polar functional groups such as carbonyl, ester, ether, sulfonamide, or amide that are ~7Å away from the urea carbonyl were defined as a secondary pharmacophore which typically improved aqueous solubility and PK properties while maintaining potency. In addition, a polar tertiary pharmacophore such as an ester, ether, acid, or amine that is 13 atoms or ~17 Å away from the urea carbonyl has also been identified. The linker that connects the primary and the secondary pharmacophores, or the secondary and the tertiary pharmacophores, can be either a conformationally flexible alkyl, or restricted cyclic structures such as aryl, cycloalkyl or cycloamino groups.

The UCD group adopted multiple approaches to improve the physical properties as well as metabolic stability of the disubstituted urea sEH inhibitors (Figure 5). The first approach involved the derivatization of compound 3 to provide improved inhibition potenciesand physical properties such as a 23–66 °C lower melting point and up to 5-fold better solubility in oil (triglyceride).⁸⁵ For example, the ester derivatives of compound 3 (4), gave appreciable increase of oral exposure compared to 3when dosed in triglyceride formulations. The corresponding terminal amides, however, offered no advantage in potency or physical properties. It became clear that for the sEH inhibitors to be generally useful to biologists one needed material which had improved pharmacokinetic properties while preserving the high potency of compound 3. The UCD group’snext approach employed an amide to replace the urea moiety in compound 3.⁸⁶ In earlier studies, the small advantage of the corresponding amides as central pharmacophore, in terms of physical properties were offset by decreased potency. However, these studies showed that with the proper substituents amides could be remarkably powerful inhibitors. In general the SAR predicted from ureas applied as well to the corresponding amides. In this series the resulting amide analog 5a gave a ~2-fold less potency but 10-fold higher solubility in water than the corresponding urea analog 5b.

Figure 5.

Various secondary urea-based sEH inhibitors.

Approximately 300 hits were found following the screening of the 300,000 NIH ‘drug like’ compound library. Most of these hits were amide or urea compounds that fit with published SAR work. This work also showed that the basic pharmacophore demonstrated high selectivity for the sEH since the compounds were screened against numerous other targets. These data are in the public domain (http://pubchem.ncbi.nlm.nih.gov/). Starting with a quite active hit (6) from this screen, the optimization of the amine partner on the right and the sulfonamide on the left led to modest success in terms of potency improvement.⁸⁷ The most active analog resulting from such efforts contained a cycloalkyl amine moiety on the left (7)

Large peptidyl moieties were also incorporated into sEH inhibitors. It was found that this modification was acceptable if the peptidyl group was located at a proper distance from the primary pharmacophore.⁸⁸ Despite the lack of advantage over compound 3 regarding PK, analog 8 suggested that large terminal groups were tolerated in terms of sEH inhibitory activity. These groups could potentially be used for innovative delivery systems and large reporter groups can fit into the right side cavity of the enzyme and even reach into the aqueous environment outside of the catalytic tunnel.

The introduction of certain functional groups significantly altered the physical properties of sEH inhibitors. For example, a 5-substituted piperazine (9) was weaved by the Long group into a series of sEH inhibitors as a novel secondary pharmacophore which greatly enhanced the solubility of the urea-based analogs despite some loss of potency.⁸⁹ Urea ether analogs containing a diethylene glycol (10) or morpholine (11) motif exhibited good potency, markedly increased water solubility, and eliminated the beta oxidation that led to rapid degradation of compound 3.⁹⁰ Such derivatives could have utility in organ baths, intravenous applications or oral hygiene.

Another strategy to improve physical properties of disubstituted ureas is to impose conformational constraints. Namely, the linker between the primary and secondary pharmacophores is no longer an alkyl chain but a saturated ring, in the case of piperidine or cyclohexane, or an unsaturated ring, in the case of a phenyl group (Figure 6). For instance, a piperidine moiety was identified to rigidify the linker region between the primary and secondary pharmacophores resulting in highly potent analogs such as 12, 13, 27 and 28.^{78, 91} Although large substituents on the right side of the sulfonamide or amide secondary pharmacophore could result in increased potency as in 12, they also resulted in much lower blood levels following oral administration to dogs.⁷⁸ This work also showed that the adamantyl moiety gave marginally acceptable pharmacokinetic properties with a piperidine acetamide group on the right as in compounds 12 and 28. However, when the lipophilicity of the whole molecule was increased even slightly, the adamantane was so rapidly hydroxylated that these derivatives were not very useful in vivo

Figure 6.

Conformationally constrained sEH inhibitors.

Further modifications of the left-side of the scaffold using aryl to replace admantyl resulted in several analogs with subnanomolar IC₅₀’s against sEH.⁴⁷ The improved murine PK of analog 14 allowed the evaluation of this compound in a mouse inflammatory pain model, in which this compound reduced local inflammatory pain caused by carrageenan at a far lower dose than morphine. These compounds gave good PK results in rats, dogs, and monkeys.^{54, 78, 92} Besides piperidinyl, cyclohexyl could also be placed as an alternative conformational constraint for the linker region connecting the primary and secondary pharmacophores. ⁹³ The solid-phase combinatorial approach successfully yielded highly potent inhibitors such as compound 15. It should also be noted that free carboxylic acid or ortho-substitution of the phenyl group on the left-hand side of the primary pharmacophore led to dramatic decrease of potency. In contrast, the addition of a carboxylic acid to the terminal phenyl group on the right led to an orally bioavailable and potent sEH inhibitor 16.⁹⁴ Interestingly, the 1,4-trans cyclohexane isomers were more metabolically stable in human hepatic microsomes than the cis isomers although their in vivo AUCs were similar. The AUC of 16 in dogs was 40-fold higher than that of compound 3, and the half time and oral bioavailability were both excellent.⁷⁸ Replacing the adamantane with other aliphatic or aromatic systems on the basic structure of 16 did not have a major effect on canine AUC but dramatically increased the blood levels in other species. These compounds represent another series with high potency on the enzyme and good exposure in multiple species following oral administration. Unlike the piperidines which appeared quite active on rodents and primates but poorly active on the sEHs from other species, the cyclohexyl ethers were active in multiple species. These compounds are remarkably similar to the commercial drug Sorafenib which is both a potent kinase and potent sEH inhibitor.⁹⁵ This observation allowed the synthesis of compounds which had no activity on known kinases and others which were selective inhibitors of both sEH and a subset of kinases. These studies suggested only limited structural space where compounds had both kinase and sEH inhibitory action. Sorafenib has a terminal carboxylic acid function converted to an amide. Although this increased the potency of compounds such as 16 into the low nanomolar range for sEH, it also decreased their solubility. Aromatic systems such as a benzoic acid moiety may also be used as a conformationally restrained linker to increase the solubility of compounds.^{78, 85, 86, 96}

Due to its excellent activity in both human and other species, compound 16 was subsequently evaluated in the LPS-induced sepsis model in mice. At 1 mpk p.o. dose, this compound ameliorated hypotension caused by LPS in mouse, which was equivalent to the effect of 10 mpk of compound 3 under the same conditions. Furthermore, the increase in blood EETs/DHETs ratio correlated with the reversal of LPS-induced hypotension suggesting that the observed efficacy may be mediated via sEH inhibition. When the linker is a phenyl group in the case of a salicylate urea-based sEH inhibitor, compound 17 exhibited good potency and excellent stability in human hepatic S9 fraction, which was further reflected by its superior oral drug exposure compared with compound 3.⁹⁶

Lastly, tertiary pharmacophores of sEH inhibitors, as colored in purple in Figure 7, were successfully explored to improve PK properties. For example, a second ester can be added to the long terminal alkane moiety (18) without loss of activity, but the PK was not acceptable.⁹⁷ The incorporation of a piperazine functionality into the 1,3-disubstituted urea series led to potent analogs 19, 20 and 21.⁸⁴ These compounds also demonstrated modest to good water solubility, excellent oral exposure and good half-lives. Specifically, compound 20 provided a remarkable half-life and oral exposure, and compound 21 gave good solubility and potency. The extensive SAR developed with these compounds has allowed exploitation of biological observations that sEH inhibitors reduced the undesirable side effects and synergized with NSAIDs and COX2 inhibitors,^{65, 66} aspirin and FLAP inhibitors,⁶⁷ and some phosphodiesterase (PDE) inhibitors.⁵³ For example, Hwang et al. have prepared pyrazole derivatives that were both more potent and more efficacious in some in vivo assays than a combination of a good sEH inhibitor and celecoxib.⁶⁹

Figure 7.

The modifications of the tertiary pharmacophore to improve PK.

3.2. Arête Therapeutics

Although the majority of reported sEH inhibitors are amides and ureas, Arête Therapeutics investigated alternative primary pharmacophores such as sulfonamide, thiourea, sulfonylurea, aminomethylene, hydroxyamide, and ketoamide in order to identify chemotypes to cover potential intellectual property (IP) space. They found that almost all replacements led to substantial loss of sEH potency with the exception of α-hydroxyamides, which gave acceptable sEH inhibitory activity as shown by analog 22 (Figure 8).⁹⁸ The hydroxy group in this position could be anticipated from earlier irreversible glycidol inhibitors.⁹⁹ To avoid the metabolic oxidation associated with the admantane ring of compound 3, the Arête team reported unsymmetrical non-adamantyl N, N′-diaryl urea and amide inhibitors such as disubstituted ureas 23 and 24, which had excellent enzyme activity and significantly improved oral exposure.¹⁰⁰ Further SAR optimization of R, L¹, P², L², and P³moieties resulted in several potent sEH inhibitors including analogs 25 and 26 with low nanomolar IC₅₀s.

Figure 8.

The modifications of primary, secondary and tertiary pharmacophores by Arête Therapeutics.

Arête disclosed a series of studies of sEH inhibitors in metabolic disease models.⁴⁴ For example, compound 27 caused diet-induced obese (DIO) mice to gain less weight relative to vehicle. This analog also reduced plasma cholesterol, decreased serum glucose relative to the control group in an interperitoneal glucose tolerance test, and reduced blood pressure at 60 mpk (b.i.d.) in C57B1/6 mice fed on high-fat and high-fructose diet without significantly changing the heart rate. Therefore, compound 27 may be useful in treating metabolic syndromes including obesity, hypertension, diabetes, and hypercholesterolemia.⁴⁵ However, the performance of sEH inhibitors in animal models of metabolic syndromes has been erratic. In another related patent application,⁴³ Arête disclosed for the first time an sEH inhibitor that effectively reduced total cholesterol, low-density lipoproteins (LDL), triglycerides (TG), and glucose in AngII-infused ApoE deficient mice.

The structure of Arête’s phase IIa clinical candidate 28 has recently been revealed.³⁰ This previously published compound⁹¹ exhibited good potency against sEH, good Caco-2 permeability, surprisingly high water solubility, and reasonable plasma protein binding across species. Regarding off-target profiles, compound 28 displayed excellent selectivity with extremely low inhibition of the microsomal epoxide hydrolase and little inhibition or binding with a panel of about 150 other enzyme and receptor targets, and minimal hERG and cytochrome P450 (CYP) liability. Furthermore, compound 28 had good oral bioavailability ranging from 25% in cynomolgus monkey to 100% in rat. A robust correlation between the plasma concentration of compound 28 and inhibition of sEH, which was measured by the total EET/DHET and EpOME/DiHOME ratios, was observed in a murine model. When orally dosed at a relatively high 50 mpk (b.i.d.) in AngII-induced hypertensive rats, compound 28 provided 14–16 mmHg reduction of systolic blood pressure relative to vehicle control treatment.³¹ It should be noted that this compound was found to be only moderately efficacious at 300 mpk dose in SHRs, showing a lack of correlation between the target engagement and antihypertensive effects in this model. Compound 28 was also tested in DIO mouse for antidiabetic effects. At 100 mpk b.i.d. oral dose for 4 weeks, compound 28 reduced glucose AUC and maximal glucose excursion with respect to the vehicle-treated control group. In this study, blood sEH activity was inhibited over ~90% up to 7 h post dosing, and then ~70% at the 12 h time point. Compound 28 is a better inhibitor of the rodent sEHs than the human and far less potent on canine and feline enzymes.⁷⁸ Based on its efficacy in rodent models, and the fact that compound 28 was well tolerated in preclinical toxicology studies, ^{101, 102} this compound was advanced to human clinical trial involving obese patients with stage 1 hypertension and impaired glucose tolerance. In a phase I trial involving healthy volunteers, compound 28 was well-tolerated in single and multiple oral doses with a mean terminal half-life of 3–5 hours. 90% or greater of sEH inhibition was achieved over 8 h at the 250 mg dose and over 12 h at the 500 mg dose.¹⁰³ Multiple doses at 100–400 mg every 8 h resulted in more ≥ 90% of sEH inhibition during the trough. Although a compound with a short half life and used at high doses, compound 28 could be the first proof-of-concept molecule for this mechanism, and could become a useful tool compound to test for other therapeutic indications. A caution is that the high throughput fluorescent screening assay for epoxide hydrolase inhibitors tends to overestimate the potency of some piperidines such as 28 and this compound is rapidly metabolized in a variety of animal models including mice⁴⁷ and dogs⁷⁸.

3.3. Boehringer Ingelheim

By performing a highthroughput screening (HTS) researchers at Boehringer Ingelheim identified N-(3,3-diphenylpropyl)-nicotinamide as a potent hit (29, Figure 10).¹⁰⁴ Interestingly, their lead compound was found in a cell based HTS for inflammation with a biological readout. In brilliant work the late Thomas Warren developed a photoaffinity label of the hit and identified the sEH as the biochemical target. It is likely that the left side pyridine in the structure binds to a unique site in the W334 niche and the diphenyl group takes advantage of the large F265 pocket. This branched chain approach could likely be used in the generation of more selective sEH inhibitors. The compound also benefits from the increased solubility of the amide central pharmacophore. Further profiling of this compound revealed short metabolic half-lives in both human and rat liver microsomes as a result of extensive oxidation of the two phenyl groups of the benzhydryl moiety. The inclusion of fluorine atoms at the 4-position of the two aryl groups shifted the major metabolic route to the oxidation of the pyridyl group on the left-hand side. A trifluoroethoxy group was then introduced to the pyridyl (compound 30) leading to significantly improved half-lives in liver microsomes. However, the increased microsomal stability did not translate into reasonable drug level at 4 h post dosing of 5 mpk of 30 (p.o.) presumably due to poor absorption. Inspired by the cocrystal structure of 30 with sEH, one of the two fluorine atoms in 30 was substituted with a polar group such as methyl sulfonyl, and the trifluoroethoxy substituent of the pyridyl group was replaced by nitrile allowing for the discovery of 31 as a potent and metabolically stable analog with good oral exposure. In addition, this analog displayed good selectivity for sEH over CYP enzymes as well as excellent cell permeability.

Figure 10.

SAR development of an amide series byBoehringer Ingelheim.

Another HTS hit series, exemplified by urea 32, provided extended half-life in microsomes, but lacked desirable sEH potency. The hybridization of the nicotinamide series and the urea series then led to another attractive class of sEH inhibitors with lower molecular weight. Specifically, the benzhydryl moiety of hit 29 can be replaced with an aryl group which is one carbon linker away from the amide.¹⁰⁵ Despite the poor rat liver microsomal stability, compound 33 gave acceptable oral exposure and long half-life. The replacement of one of the two chlorine atoms in compound 33with a methyl sulfone then yielded analog 34, which had a 5-fold increase in plasma concentration compared with compound 34 at 1h post dosing. As such, the combination of a metabolically more stable benzyl amine and nicotinamide provided a new class of sEH inhibitors with balanced potency and metabolic profiles. The benzyl of 33 and 34 moved the ortho substituent far enough away from the amide NH to allow its hydrogen bonding with the catalytic Asp333 of the active site of sEH. In addition, the large ortho chlorine probably reduced benzylic hydroxylation by cytochrome P450 enzymes.

The Boehringer Ingelheim team also explored piperidyl-ureas derived from a potent HTS hit (35) in their lead optimization efforts (Figure 11). The primary efforts were focused on the replacement of two potential toxicophores, the aniline and the pyrrole moieties. First, it was discovered that the aniline component could be replaced by 2,4-dichlorobenzyl (36), which is reminiscent of the aforementioned nicotinamide series, without loss of potency. Keeping the 2,4-dichlorobenzyl group constant, the team examined various substitutions of the piperidyl ring, and ultimately found that aryl-oxy groups provided good potency. Although pyrimidyl-oxy analog 37 was less active than the phenoxy analog 36 in terms of enzyme and cell sEH inhibitory activity, analog 37 was more selective against the CYP enzymes responsible for EETs production. Furthermore, compound 37 led to a prolonged half-life in liver microsomes and a 13-fold higher drug level at 4 h post oral dosing (5 mpk). The co-crystal structure of 37 and sEH revealed that the left-hand side aromatic moietyoccupied the W334 niche pocket. This observation prompted the SAR study of the aryl substitution in order to improve the PK while maintaining good sEH inhibitory activity. Despite similar enzyme inhibitory activity (IC₅₀=4–7 nM), compounds with polar aryl substituents including cyano, sulfonylmethyl, methyl sulfonylamido, or methyl amido were typically more potent in the cell-based assay (IC₅₀<1 nM) than those bearing halogen or trifluoromethyl groups. These analogs, however, did not have acceptable in vitro metabolic stability with the exception of analog 38 containing a carboxylic acid group at the 4-position of the phenoxy moiety. This compound had high and sustained drug exposure over 6 h, and was significantly more selective against sEH over several CYP enzymes.

Figure 11.

Piperidyl urea sEH inhibitors discovered by Boehringer Ingelheim.

The modular scaffold of the trisubstituted urea prompted the Boehringer Ingelheim team to utilize parallel solid-phase synthesis to enable rapid access of a 270-member library analogs.¹⁰⁶ This approach started with a pharmacophore-based virtual screening based on a set of 287 known sEH ligands. The combined pharmacophore and shape model was then used in the context of a virtual screening software called PharmShape to screen a virtual library of 5510 compounds (29 R¹R² cyclic amines and 190 R³ aniline-replacements), of which 291 was removed. The similarity of the remaining virtual library compounds were compared with 93 known sEH inhibitors using 3D pharmacophore fingerprints. Each building block was assigned a consensus score by averaging the similarity scores of all enumerated compounds containing this building block. These building blocks were then clustered, and representatives of each cluster with the highest consensus scores were selected for parallel synthesis. The synthesis commenced with the conversion of solid-support amine (R³NH-CH₂-solid support) reagents with 4-nitrophenyl chloroformate to generate activated carbamates, which then reacted with R¹R² cyclic amines followed by solid-support removal to afford the desired trisubstituted ureas without the need for chromatographic purification. As a result, several potent analogs such as compound 39 resulting from a modest variation of the lead structure were obtained.

Boehringer Ingelheim also described pyrazole aniline-derived amide (40) as a lead to potent sEH inhibitors (Figure 12).¹⁰⁷ The structure of the co-crystallized sEH with inhibitor 40 suggested that both pyridyl groups were facing solvent and the ethyl group resides in a deep hydrophobic pocket of sEH. Three changes were adopted to provide more potent analogs such as 41. First, the ethyl group was replaced with a trifluoromethyl group. Second, the left-hand side 3-pyridyl was replaced by a phenyl group. Third, a pyridyl group was selected after an extensive exploration of heteroaryl replacement of the central benzene group due to the mutagenic concern related to the central aniline moiety of 40. Eventually, analog 41 offered good sEH inhibitory activity, a modest half-life in human liver microsomes, and no CYP3A4 inhibition, thus presenting an attractive lead for further optimization.

Figure 12.

Pyrazole amides as sEH inhibitors developed by Boehringer Ingelheim.

3.4. Dainippon Sumitomo

Ligand efficiency (LE) is an important parameter to ascertain the hit quality in lead identification. A Dainippon Sumitomo team used LE indices (LEI) in a virtual screening-initiated lead generation study of sEH inhibitors.¹⁰⁸ High lipophilicity and high molecular weight were considered as risk factors of promiscuous association with off-target activity, causing undesired side effects. As such, the team selected compound 42 of low molecular weight (MW = 262) and less lipophilicity (AlogP = 2.76), but a high LEI value (0.43) during the hit triage process (Figure 13). The right-hand cyclopropyl phenyl group was not optimized since the preparation of substituted phenyl group was considered challenging. The conjugated cyclopropyl group offers a number of advantages in electronically stabilizing the urea, avoiding aniline as a possible metabolite, and avoiding a labile benzylic carbon. Thus the left-hand moiety went through a hit-to-lead process overcoming a series of issues including poor aqueous solubility, microsomal stability, and CYP inhibition while maintaining good LEI values. Ultimately, analog 43 was identified as a desirable lead for further optimization. This fragment-based drug design approach was described as fragment-inspired medicinal chemistry by the authors.

Figure 13.

SAR evolution of sEH inhibitors with high ligand efficiencies by Dainippon Sumitomo.

3.5. Merck

The target validation efforts at Merck were aimed to establish whether sEH inhibitors can provide mechanism-based blood pressure lowering efficacy in SHR, a common preclinical hypertension model that responds to various antihypertensive drugs. This objective was rigorously pursued by using several series of highly selective and potent sEH inhibitors. The off-targets routinely monitored included CYP enzymes, particularly CYP2C8 and 2C9 which are involved in the production of EETs, ion channels with cardiac implications, and mEH, a significant player in xenobiotic detoxification¹⁰⁹ and steroid metabolism.¹¹⁰

The first series reported by Merck covered a range of 3,3-disubstituted piperidine-derived trisubstituted ureas (Figure 14). A highly potent analog in both enzyme and cell-based assays, the lead compound 44 had some CYP and ion channel off-target activity, and relatively high clearance and low oral bioavailability. Lead optimization studies were directed at the left-hand amine/aniline, and two of the piperidinyl substituents colored by purple, blue and red, respectively. 3,3-Disubstituted piperidine-derived trisubstituted urea ent-45 was discovered ultimately as a highly potent and selective sEH inhibitor. ¹¹¹ It is worth noting that the presence of a carboxylic acid with a proper linker length was key to the dramatically improve PK and selectivity against all of the off-targets. Despite the good compound oral exposure, excellent sEH inhibition in whole blood, and remarkable selectivity, compound ent-45 failed to lower blood pressure acutely in SHRs at 50 mpk p.o. dose.

Figure 14.

3,3-Disubstituted piperidine-derived trisubstituted ureas as sEH inhibitors discovered by Merck.

To add additional conformational constraint to urea structures, spirocyclic secondary amine-derived trisubstituted ureas were identified as highly potent, bioavailable and selective sEH inhibitors (Figure 15). ¹¹² Two major subseries, chromanones and chromans, represented by analog ent-46 and ent-47, are both potent and reasonably clean in off-target screening. The presence of the ketone moiety in analog ent-46 might account for its higher clearance than ent-47, which had remarkably high oral exposure and low clearance. However, the in vitro DHET production was suppressed more by ent-46 (93%) vs ent-47 (68%) when tested at 1 μM of compound concentration in whole blood. In addition, it has been shown that compound rac-46suppressed DHET production rate by >95% from 0.5 h to 6 h post dosing (50 mpk, po). As such, this compound was subjected to a PD efficacy study in SHR. Despite good oral exposure and excellent ex vivo target engagement in blood, rac-46 failed to lower blood pressure acutely in SHRs at 300 mpk dose (p.o.) over 8 h.

Figure 15.

Spirocyclic secondary amine-derived trisubstituted urea inhibitors of sEH discovered by Merck.

4-Substituted piperidine-derived trisubstituted ureas were also reported as highly potent and selective inhibitors for sEH (Figure 16).¹¹³ By judiciously varying the heteroaryl substitution of the piperidinyl group, compound ent-48 was discovered as a good inhibitor against sEH. In addition, this analog has minimal mEH, ion channel, CYP as well as other off-target liability against a panel of 166 counterscreening targets (IC₅₀>10 μM) including a subset involved in blood pressure regulation. The acute action of ent-48 on vascular tone was further evaluated by SHR mesenteric artery assay, in which ent-48 was able to reverse vascular contraction induced by methoxamine (IC₅₀=6.5 μM). No vasoconstriction was observed on basal tension with this compound up to 100 μM. These results were in line with the vasodilatory effect of EETs previously reported in literature.¹¹⁴

Figure 16.

Optimization of the 4-heteroaryl piperidine-derived ureas by Merck.

Due in part to its excellent bioavailability (77%), moderate clearance (29 mL/min/kg), and good normalized oral exposure (1.2 μM. h. kg/mg), compound ent-48 also inhibited sEH activity effectively ex vivo and in vivo. After the treatment with ent-48 for eight days (300 mg/pk, p.o., q.d.), the endogenous epoxide/diol ratios in SHR kidney were increased 2 to 9-fold (Figure 3). In telemetrized SHRs, ent-48 failed to lower systolic blood pressure acutely or chronically despite the excellent compound exposure and >95% sEH inhibition as reflected by the ex vivo measurement of DHET production over a time course of ~24 h. It has then been proposed that the lack of blood pressure lowering activity by ent-48 might be due to the observed 2.5-fold increase in 20-hydroxyeicosatetraenoic acid (20-HETE), a potent vasoconstrictor in kidney,¹¹⁵ which could potentially negate the vasodilatory effect of elevated EET levels. The lack of robust blood pressure reduction effects of several potent, structurally distinct, and orally bioavailable sEH inhibitors 28, ent-45, rac-46 and ent-48 in SHRs, suggests that sEH inhibitors may not be a valid target for hypertension if the strains of SHRs used in these studies are a robust model for human hypertension. These data bring up the common problem in pharmacology of what animal models are predictive of human disease in drug development. This is particularly complex for hypertension which appears to be a symptom of a number of biochemical imbalances.

Distinct from urea and amide inhibitors, novel aminoheteroaryl analogs were also prepared as potent sEH inhibitors (Figure 17).¹¹⁶ It was envisioned that the N-H group of this class of inhibitors interacted with Asp333, and the heteroatom of the heteroaryl group formed hydrogen bonds with Tyr381 and Tyr465. The SAR study revealed that aminobenzisoxazoles emerged was optimal in terms of its inhibitory activity against sEH. Analog 49 was identified as a nanomolar inhibitor of sEH with a good PK profile. The compound is particularly attractive in being of low molecular weight, excellent physical properties, and of a highly constrained structure. The strategy of employing such aminoheterocycles as amide replacements represents a useful approach to develop mimics of known hydrolase or protease inhibitors containing an amide moiety.

Figure 17.

Amino heteroaryls as novel sEH inhibitors discovered by Merck.

The aforementioned series developed by Merck demonstrated their judicious consideration of both intellectual property and physical properties of compounds. When they started the program, only disubstituted ureas were reported. Realizing that trisubstituted ureas often offer better physical properties and PK, the Merck team engineered some novel tertiary ureas bearing novel spirocycles (46, 47) or heteroaryls (48), and demonstrated that these are useful tool compounds with good potency, selectivity, and PK profiles. Their aminoheteroaryls represented a nice extension distinct from amide or urea type of common motifs for sEH inhibitors.

4. Possible approaches to exploiting the P450 branch of the arachidonate cascade

The cytochrome P450 branch of the arachidonate cascade has so far not been exploited successfully by the pharmaceutical industry. Now that multiple natural eicosanoids in this cascade are commercial along with analytical methods for them, and multiple pharmacological probes are available, knowledge about the biology of this branch of the cascade is expanding dramatically. There are many possible ways to apply the understanding of this cascade and these important regulatory lipids to improve human health. In addition, there are multiple pharmacological approaches including inhibition of synthesis of the predominantly inflammatory 20-HETE, but exploitation of the knowledge that the epoxides of fatty acids and particularly long chain ω-6 and ω-3 polyunsaturated lipids are powerful chemical mediators are the most advanced. Although diols of linoleate appear to be proinflammatory mediators of vascular permeability and diols of arachidonate appear to mediate stem cell mobility and other biologies, so far EETs and their ω-3 homologs appear to be the major anti-inflammatory, anti-hypertensive, and analgesic fatty acid epoxides with implications for treating diseases. In the indications discussed below the diols are simplistically regarded as highly polar inactivation products.

There are good pharmacological probes available to mimic and to inhibit the action of EETs and ω-3 epoxides, but so far mimics have not been exploited as drug-like molecules. On analogy with angiotensin converting enzyme (ACE) inhibitors and AngII receptor blockers (ARBs), both EET mimics and sEH inhibitors are attractive as potential drug candidates. The concept of mimicking the bioactive eicosanoid is appealing in that the mimics will be active even in the absence of the natural eicosanoid. Thus the maximal effects of these materials should be greater than the sEH inhibitors emphasized in this review. However, sEH inhibitors should stabilize epoxy lipids generally in the ratio of homologs, optical and regioisomers that are produced naturally. The increase in these chemical mediators also should be greatest near the site of production. This local effect could be pharmacologically important because lipid chemical mediators commonly have different ratios and activities in different tissues. sEH inhibitors have been repeatedly shown to increase the concentrations of epoxylipids since the sEH is the major route of metabolism of these compounds. However, since the sEH is only one of multiple pathways of degradation for epoxylipids, the increase in epoxylipids and decrease in corresponding diols will be limited. Thus on the negative side, the maximal efficacy of sEH inhibition is thus restricted to stabilizing endogenous fatty acids. In contrast, theoretical mimics of these mediators would only be limited by receptor saturation and signal transduction. On the positive side, the possible side effects of dramatically increased EETs will also be limited with sEH inhibitors. This may explain the exceptionally positive therapeutic index so far observed with sEH inhibitors. As we learn more about the biological activities of the resulting diols, the reduction in diols might cause additional benefit or off-target effects. At this point reduction in diols seems to be generally positive in reducing the production of another inflammatory mediator. Of course sEH inhibitors will synergize the effects of exogenous epoxy fatty acids or other pharmaceutical agents such as NSAIDs and PDE inhibitors which increase them.

The stabilization of epoxylipids with sEH inhibitors is the approach closest to the clinic which addresses this branch of the arachidonate cascade. As discussed above the sEH inhibitors stabilize endogenous lipid epoxides and increasing these metabolites has generally proven beneficial in the reduction of hypertension, a variety of inflammatory disorders and pain. Analogously, NSAIDs and COXIBs inhibit the biosynthesis of prostaglandins and thromboxanes. In general these eicosanoids have proven to increase pain and inflammation so COXIBs are considered beneficial. However, a caution is that the effects of thromboxanes and prostaglandins are not in all cases detrimental. On analogy neither should we expect an increase in epoxylipids caused by sEH inhibitors to be beneficial in all cases. As sEH inhibitors are investigated, it is critical that experiments into their biology are also designed in an attempt to elucidate possible detrimental as well as beneficial effects.

5. Possible routes for sEH inhibitors to reach the clinic

As discussed in this review a number of powerful inhibitors of the sEH have been prepared in several chemical series. Many of these compounds have drug-like properties in terms of excellent potency, selectivity, and oral exposure. This class of drugs appears to have a massive therapeutic index with few side effects in animals. These properties yield good target engagement as shown by an increase in epoxides in tissue fluids and often a decrease in the corresponding diols. In a variety of cases this has been coupled with demonstration of efficacy in numerous animal models of a variety of human diseases. These diverse series of inhibitors have effectively mapped the catalytic site of the enzyme. This knowledge coupled with several crystal structures of the murine and human enzyme will facilitate the synthesis of still other potent inhibitors.

sEH inhibitors have never been shown to reduce the blood pressure in normotensive animals. However, they are moderately to dramatically anti-hypertensive in a variety of animal models examined in many laboratories. They are particularly active in models where angiotensin is a major driver of hypertension and where renal damage is involved. Some SHR strains have proven particularly valuable for extrapolating from treating animal to treating human hypertension. It is worth noting that hepatic sEH levels in all rat strains examined so far are far lower than those in any other mammalian species reported.¹⁴ In various SHR models the response from sEH inhibitors has been erratic. Some but not all of these observations can be attributed to documented genetic variations among the SHR strains used, as well as within strain variations in both the sEH titers and the sensitivity to treatment with sEH inhibitors. Some SHRs are even resistant to treatment with commercial anti-hypertensive compounds. Several extensive studies have shown good blood pressure reduction in SHR strains with sEH inhibitors. However, as discussed above the failure of potent sEH inhibitors with good exposure using their SHR models apparently led both Boehringer Ingelheim and Merck to put on hold their sEH inhibitors for blood pressure treatment. This appears to have been a good decision given the failure of Arête Therapeutics to control blood pressure in phase II clinical trials, the diversity of causes of hypertension in man, and the high cost and high risk of hypertension trials. One approach might be to segregate the human population into those with SNPs showing high sEH activity or those with angiotensin-driven hypertension, but such an approach is not now commercially attractive. The tendency of sEH inhibitors to reduce blood pressure could make them attractive add-on drugs for severe inflammation where extensive use of some NSAIDs exacerbates hypertension.

The major limitation in the field is that, in spite of these powerful chemical probes and a variety of attractive biological targets, there is not an obvious and inexpensive route to the clinic for a major market with these molecules. For example, diabetes is an attractive and timely target, and sEH inhibitors have been shown to reduce symptoms of diabetes in several models and to reduce symptoms dramatically in several. A caution is that in still other models there have been either marginal effects or no beneficial effects in overcoming insulin resistance. Furthermore, it is well-known that there is a lack of predictability of rodent diabetic model for humans. In addition, human genetic studies have not yet disclosed a strong correlation between sEH and diabetes. In the absence of a theoretical mechanism of action in treating diabetes, a compelling case for human clinical trials for diabetes has not been made. This caution was realized in a recent failure by Arête Therapeutics in a phase II trial to improve blood markers of metabolic syndromes. On the other hand, animal models show strong efficacy of sEH inhibitors for a number of co-morbidities associated with diabetes, including in particular renal failure, vascular inflammation, atherosclerosis, fibrosis, heart failure (HF), neuropathic pain, stroke and other indications. Compounds that offer these attributes and the ability to improve insulin sensitivity and glucose tolerance even marginally in some models are attractive. However, it remains challenging to pursue these co-morbidity indications without showing that sEH inhibitors can control blood sugar as a stand-alone therapy. This opinion could be reversed with demonstrated efficacy in other and particularly non rodent models, or a mechanistic understanding that allows extrapolation of rodent results to man.

On the other hand, even in SHRs resistant to blood pressure reduction with sEH inhibitors, other positive biological effects have been seen including a dramatic reduction in renal, cardiac ischemia reperfusion and stroke damage. Separate from hypertension, sEH inhibitors are effective with a variety of cardiovascular indications ranging from atherosclerosis through atrial fibrillation and fibrosis. For example, compound 3 was recently used in a small clinical trial to ascertain the effects of an sEH inhibitor on vascular tone.¹¹⁷ It was found that compound 3 alone, or in the presence of urotensin II, increased flux in healthy controls and HF patients. However, most of the cardiovascular indications involve long and expensive outcome trials. This is in part due to a lack of biomarkers that indicate efficacy for anti-inflammatory drugs to treat atherosclerosis in both preclinical and clinical setting. As such the pursuit of such indications for sEH inhibitors would be highly unlikely unless they could at least demonstrate utility in non-human primate disease models.

Diabetes and hypertension represent large potential markets. Multiple animal models suggest a role for sEH inhibitors as add-on or combination therapies. However, with the phase II failure of Arête Therapeutics with metabolic syndromes, the great expense of the trials for these indications, and the uncertainties with some of the animal models, to repeat trials in the near term is unlikely. There are a variety of potential short trials where animal models suggest a good chance of success. One example would be to test sEH inhibitors to reduce the renal toxicity of contrast agents. Other examples with similar limitations include treatment of restenosis, ateriovenous graft failure, inflammation from hemodialysis, or Raynaud’s syndrome. A somewhat larger market suggested from rodent data is renal protection from the toxicity of chemotherapeutic agents. However, these potential markets seem of little interest to pharmaceutical companies due to their small size. Such potential utility of sEH inhibitors for minor indications may be better addressed with physician-initiated trials. A major limitation in the field is the absence of IND enabling work on an sEH inhibitor other than compound 28 which would permit clinical trials by small companies or in the public sector. Should such a public sector investigational new drug (IND) compound become available there are a variety of indications ranging from psoriasis to eclampsia which could be approached in the public or private sector.

Numerous other indications appear attractive and represent large markets. However, they also involve high risk, long, and/or expensive trials. Examples of this dilemma include renal and vascular inflammation where evidence from animal models is very supportive. The sEH inhibitors dramatically reduce inflammation and damage associated with ischemia reperfusion injury in a number of systems including stroke models, myocardial infarction (MI), and transplant. MI is one of the few indications where the rodent studies have been reproduced in canine systems. A caution is that some sEH inhibitors that are quite active on rodent and primate sEH are almost inactive with the feline and canine sEH. A variety of other inflammatory conditions such as chronic obstructive pulmonary disease (COPD), asthma, irritable bladder syndrome, pancreatitis, and inflammatory bowel disease (IBD) are attractive, but they also have long trials. Possibly the best predictive evidence from both animal models and human genetics for a clinical trial is heart failure, but this trial is intimidating as well.

The positive effects of sEH inhibitors alone and in combination with NSAIDs and COXIBs to reduce both inflammatory and neuropathic pain could represent a relatively short trial to address a poorly met medical need. Since sEH inhibitors are known to reduce some of the toxicities associated with high dose use of NSAIDs and COXIBs, and to synergize with them to reduce both inflammation and pain, a combination of a sEH inhibitor with a NSAID or COXIB appears attractive in increasing efficacy, reducing side effects and possibly extending patent life. An example could be arthritis where long term use of NSAIDs and COXIBs can be associated with hypertension, blood clotting, and other side effects. The ability to drop the NSAID – COXIB dose while reducing some deleterious side effects is reasonable. Along the same line PDE inhibitors have long been of therapeutic interest for several diseases. Since particularly PDE 4 and 5 inhibitors are synergized by sEH, possibly these drug combinations could be used to avoid the severe dose-limiting side effects of PDE inhibitors.

The pain indication appears particularly attractive in representing a large potential market, an unmet medical need while offering phase II trials which are not prohibitively long and expensive. sEH inhibitors not only reduced inflammatory pain in a variety of animal models, but surprisingly they reduced neuropathic pain as well. When acting alone they are far more active than the common standard for treatment of neuropathic pain, gabapentin, while not causing severe sedation as morphine does using either rotorod or open field experiments. Although NSAIDs and COXIBs are widely used for treating neuropathic pain, the literature shows them to be of variable and questionable benefit. For example, the mixed COX inhibitor diclofenac varies between slight analgesia and statistically significant enhancement of pain perception in rodent models depending on time post treatment and dose. However, when diclofenac is used with a low dose of sEH inhibitor, there is dramatic synergism in reducing pain resulting from induced type I diabetes in rodents. A problem with human pain trials is of course that there is often a major placebo effect. Therefore, careful controls are critical. On the other hand, patients often suffer from both inflammatory and neuropathic pain, and the sEH inhibitors may be able to reduce both. In addition, the trials can be relatively short, and neuropathic pain in particular represents an unmet medical need.

Numerous indications appear promising for treatment by sEH inhibitors varying from large markets such as co-morbidities associated with diabetes to smaller markets such as periodontal disease or enhanced wound healing. One limitation with sEH inhibitors is that most published models have been based on rodents. It is not clear if these rodent data will apply to human disease. It would thus require a compelling biological rational, cogent data in human genetics, and/or non-human primate data to support their entry to clinics. As far as tool compounds are concerned, these compounds should not have off-target liabilities that could contribute to observed PD effects in animals. As such, a thorough investigation of potential off-target activity that could confound the data would be imperative. Another aspect when validating indications of sEH inhibitors is a robust PK/PD correlation, which will enable the translation of target engagement in a proper compartment to observed PD effects.

Certainly for the field to advance one needs to see the compounds evaluated by either a pharmaceutical company, or the availability of a compound with IND status from the FDA to allow physician-initiated clinical trials in the public sector. Excellent drug-like structures exist both in the private and public sector which have properties indicating that they could be excellent candidates for INDs.

6. Conclusion

Early sEH inhibitors were conceived based on a knowledge of the enzyme’s catalytic mechanism in particular the transition state. They were optimized via classical SAR approaches coupled with moderate throughput enzyme assays and rodent pharmacokinetics. Three dimensional SAR helped to design a second generation of sEH inhibitors which were more sophisticated and with better physical properties and pharmacokinetics. The more recent structure-based drug design was possible due to the availability of several solidco-crystal structures of inhibitors bound to sEH. These co-crystal structures also elucidated the precise information on the key binding interactions of the inhibitor with the enzyme, and suggested further structure modifications which could fit well in the binding pocket. In all cases, sEH inhibitors possess two key structural features in the hydrogen bond network. The first is a hydrogen bond acceptor, such as the carbonyl group of either an amide or a urea, which interacts with Tyr383 and Tyr466 mimicking the oxygen atom of an EET epoxide. The second structural feature is a hydrogen bond donor, such as the N-H group of an amide or a urea, which engages a hydrogen bond with Asp335. The triple interaction is indispensable for inhibitors to achieve high activity, as the absence of either carbonyl or amide N-H group typically resulted in significant loss of potency.

During the lead identification and optimization processes, appropriate balance of potency and physical properties, ligand efficiency, polarity, metabolic stability, PK, and off-target profiles are essential to provide tool compounds suitable for pharmacological proof of concept. Several useful approaches have been applied in the medicinal chemistry work on sEH inhibitors. One such is to introduce secondary or tertiary polar pharmacophores to increase solubility and polarity. Conformational restraint to boost potency while improving metabolic stability is particularly powerful.

The lack of efficacy of compound 28 to treat patients with mild to moderate hypertension and impaired glucose tolerance in a phase II clinical trial cautions whether sEH inhibition is a robust mechanism to treat hypertension and/or diabetes. However, it is conceivable that other promising therapeutic applications of sEH inhibitors, including inflammation, pain, cardiovascular diseases, and co-morbidities of diabetes, warrant further investigations. Possible routes do exist to explore the P450 branch of the arachidonate pathway, and in particular, to enable sEH inhibitors to be tested in clinic.

Figure 9.

Key sEH inhibitors developed by Arête Therapeutics.

Nonstandard abbreviations

sEH

soluble epoxide hydrolase

mEH

microsomal epoxide hydrolase

LTA₄

leukotriene A₄

EETs

epoxyeicosatrienoic acids

NSAIDs

non-steroidal anti-inflammatory drugs

COX-1

cyclooxygenase-1

COX-2

cyclooxygenase-2

LT

leukotriene

DHETs

dihydroepoxyeicosatrienoic acids

3D QSAR

three-dimensional quantitative structure-activity relationship

SNPs

single nucleotide polymorphisms

AngII

angiotensin II

SHRs

spontaneously hypertensive rats

EDHF

endothelium-derived hyperpolarizing factor

AUDA-BE

12-(3-adamantan-1-yl-ureido)-dodecanoic acid butyl ester

MCAO

middle cerebral artery occlusion

thoracic aortic constriction

TAC

LDL-C

low density lipoprotein cholesterol

TG

triglyceride

mpk

mg/kg

5-LOX

5-lipoxygenase

LPS

lipopolysaccharide

AUDA

12-[3-adamantane-1-yl ureido]-dodecanoic acid

COXIBs

COX-2 blockers

PK

pharmacokinetics

LOQ

limit of quantification

SAR

structure-activity relationship

UCD

University of California Davis

DCU

dicyclohexyl urea

PD

pharmacodynamics

PDE

phosphodiesterase

IP

intellectual property

CYP

cytochrome P450

HTS

high throughput screening

LE

ligand efficiency

LEI

LE indices

20-HETE

20-hydroxyeicosatetraenoic acid

ACE

angiotensin converting enzyme

ARBs

AngII receptor blockers

IND

investigational new drug

MI

myocardial infarction

COPD

chronic obstructive pulmonary disease

IBD

inflammatory bowel disease

Biographies

Dr. Hong C. Shen received his B.S. degree in chemistry from Peking University under the direction of Professor Yunhua Ye. He subsequently obtained his M.S. degree from University of Minnesota working with Professor Richard Husng. Hong Shen then conducted his Ph.D. work on transition metal catalysis and total synthesis with Professor Barry Trost at Stanford University. In 2003, he joined the Merck Research Laboratories in New Jersey, working on drug discovery for cardiovascular, thrombosis, and metabolic diseases. Recently Hong Shen assumed a position of director and section head at the Roche R&D center in China working in the therapeutic areas including oncology, virology and metabolic disorders. He has more than 55 publications and 17 patent applications.

Dr. Bruce D. Hammock is a Distinguished Professor and a toxicologist in the Department of Entomology at the University of California, Davis. He holds a joint appointment in the Cancer Center of the School of Medicine at UCD. He has over 800 peer reviewed publications and 100 patents and is a member of the US National Academy of Sciences. His laboratory has pursued the biochemistry and physiology of the sEH following its discovery in 1970 with Sarjeet Gill and John Casida, and has been engaged in development of sEH inhibitors for human therapy, and metabolomics approaches using mass spectrometry for evaluating the role of the cytochrome P450 pathway and the sEH in the arachidonic acid cascade.