Polynitrogen-containing compounds as multi-target sEH/FAAH inhibitors: structure-activity relationship and pharmacological studies

Abstract

Both soluble epoxide hydrolase (sEH) and fatty acid amide hydrolase (FAAH) are involved in degradation of anti-inflammatory and antinociceptive lipids, thus inhibition of these enzymatic pathways represents a novel strategy in the discovery of non-opioid drugs for treating inflammatory pain. We previously discovered several multi-targeted designed ligands and described a pharmacophore necessary for inhibition of both sEH and FAAH. The potential for optimization on the left side of the pharmacophore led us to exploration of different heterocyclic moieties with the hope to keep strong inhibition potencies, but to increase the metabolic stability and solubility of new analogs. Eighteen analogs containing various substituted and unsubstituted pyrimidinyl-, quinoxalinyl- and tetrazolyl- rings are synthesized and tested for inhibition potency in human FAAH, and human, rat and mouse sEH. The structure-activity relationship study revealed quinoxalinyl- analog 4m, the most potent dual inhibitor reported to date, with IC₅₀ values of 2.9 nM in human FAAH and 0.7 nM, 39.1 nM and 0.3 nM in human, mouse and rat sEH, respectively. 4m showed no binding to opioid and most serotonin receptors and was tested in the human, mouse and rat liver microsomes stability assays where it exhibited good and/or moderate clearance rates. Lastly, we evaluated 4m in vivo in a wheel running assay to determine its effects on voluntary locomotor behavior. Both 4m and the traditional opioid morphine exhibited significant depression of wheel running after intraperitoneal administration indicating that 4m may produce undesirable behavioral effects, which will be the basis for future studies.

Introduction

Fatty acid amide hydrolase (FAAH) hydrolyzes the endocannabinoid lipid anandamide (AEA) into ethanolamine and arachidonic acid (AA), which do not activate cannabinoid receptors (Fig. 1).¹ FAAH plays a significant role in anti-inflammatory processes by modulation cannabinoid CB1 and CB2 receptors, making it a promising target for pain treatment without risks related to opioid agonists.² FAAH inhibitors have shown potential in treating various conditions like pain, anxiety, and inflammation, with several preclinical studies demonstrating their efficacy in modulating inflammatory responses and antitussive activity.^{3, 4} Previous research studies have focused on creating irreversible FAAH inhibitors, such as URB-937 and URB-597 (Fig. 2), which exhibit high selectivity and potency, however challenges observed in solubility and selectivity limited their clinical development and application.⁵ Soluble epoxide hydrolase (sEH) plays a crucial role in inflammation by metabolizing epoxyeicosatrienoic acids (EpFAs), which are derived from arachidonic acid, into dihydroxy acids (DHETs) that might possess pro-inflammatory properties (Fig. 1).⁶ The sEH increases during inflammation, reducing the anti-inflammatory effects of EETs and making sEH a valid target for drug discovery in inflammatory-related diseases.⁷ Inhibitors of sEH, particularly urea-based compounds (Fig. 2 – TPPU and t-TUCB), have been developed to block this metabolic pathway, thereby preserving the beneficial effects of EETs and reducing inflammation. These inhibitors have shown promise in several preclinical models for treating chronic inflammatory diseases and pain.⁸ Continued research into sEH inhibitors aims to enhance their stability, solubility and efficacy, and explore their potential in managing various inflammation-related conditions and effective use in vivo.

Figure 1.

Main metabolic pathways and targets of this study. Simultaneous inhibition of sEH and FAAH (shown in red) will lead to increased levels of AEA and EpFAs, both possessing anti-inflammatory and analgesic effects.

Figure 2.

Previously reported FAAH inhibitors (URB937 and URB597), sEH inhibitors (TPPU and t-TUCB) and dual sEH/FAAH inhibitors (SP 4–5 and JA 112)

Simultaneous administration of FAAH inhibitor URB937 and sEH inhibitor TPPU in rat and mouse models significantly elevated levels of AEA and EETs, reducing inflammation and pain.⁹ This suggests potential for developing hybrid inhibitors targeting both pathways, i.e. multi-target designed ligands (MTDLs), as co-administration required lower doses for effective analgesic effects compared to separate administration. Moreover, MTDLs are proven to have more predictive pharmacokinetics, better patient compliance and reduced risk of drug interactions.¹⁰ MTDLs have been an emerging research topic for the treatment of multiple diseases such as Alzheimer’s, breast cancer, prostate cancer and chronic pain.^11–14

We have previously discovered a unique pharmacophore required to simultaneously inhibit both FAAH and sEH activities composed of a phenyl ring linked to the piperidine ring via an amide bond (Fig. 3). The right side of the sEH/FAAH MTDL has an aromatic ring linked through a sulfonamide bond to the piperidine ring. The initial SAR study showed that both enzymes can yield potent dual inhibitors if the aromatic ring on the right side of the pharmacophore possesses different halogens or small alkyl groups in the ortho postion.¹⁵ In addtion, the previous SAR showed that the left side of the pharmacophore has potential for the optimization, since binding pockets of both enzymes can fit various heterocyclic aromatic moities.¹⁶ We reported SP 4–5, a benzothiazole dual inhibitor with IC₅₀ value of 7 nM in FAAH and 9.6 nM in human sEH (Table 1).¹⁵ This dual inhibitor does not bind to opioid receptors and does not affect the wheel running behavior in vivo.¹⁷ SP 4–5 was also evaluated in a rat formalin model of acute inflammatory pain resulting in antinociception at a much lower dose than the control, a non-steroidal inflammatory drug (NSAID) ketoprofen.¹⁶ Hower, SP 4–5 showed low metabolic stability in liver microsomal assay, low activity in mouse sEH and poor water solublity (Table 1). Next, the benzothiazole moiety was succesfully replaced with 4-phenylthiazole¹⁸, quinoline¹⁹ and isoquinoline²⁰ moieties. As examplifed in JA-112, a quinolynyl-derivative (Table 1), the strong inhibitory potency was preserved and it showed high stability in human and rat plasma stabililty assays. This MTDL was also not interacting with opioid receptors in vitro, and was active in mouse sEH inhibition assay. We evaluated in Formalin Test and it showed significant nociceptive effects in rats at the dose of 1 mg/kg, compared to 30 mg/kg of ketoprofen.¹⁹ Hovewer, this compound also showed low metabolic stability in liver microsomal assay and poor water solubility, i.e. in general solubility below 31 μM is considered a poor water solubility.

Figure 3.

Active pharmacophore for dual sEH/FAAH inhibition. A heteroaromatic ring on the left side of the molecule represents the potential for optimization.

Table 1.

Fatty acid amide hydrolase (FAAH) and soluble epoxide hydrolase (sEH) inhibitory activities and solubilities of analogs 4a–4r.

Compound	R	Human FAAH IC50 (nM)a	Human sEH IC50 (nM)a	Mouse sEH IC50 (nM)a	Rat sEH IC50 (nM)a	Solubility (μM)b
URB597	-	33	-	-	-	> 63
TPPU	-	-	2.2	5.8	0.9	60
SP 4–5		7.0	9.6	812.8	3.9	12.85
JA112		19.6	1.7	17.2	2.0	30.85
4a		174.9	19.7	342.5	18.1	> 63
4b		196.5	5.6	904.2	4.9	> 63
4c		289.7	23.4	82.5	47.7	31 < x < 63
4d		459.7	89.4	89	82.3	31 < x < 63
4e		307.9	73.6	79.8	73.9	31 < x < 63
4f		181.5	26.3	32.4	55.9	31 < x < 63
4g		175.4	27.2	160.8	43.7	31 < x < 63
4h		258.4	32.6	510.2	71.5	31 < x < 63
4i		1239.1	5.6	213.9	12.1	31 < x < 63
4j		461.9	7.7	378.6	17.9	31 < x < 63
4k		327.1	94.6	1129	133.7	31 < x < 63
4l		1996.4	78.8	40.4	15.4	31 < x < 63
4m		2.9	0.7	39.1	0.3	31 < x < 63
4n		>10000	232.1	201.6	83.4	31 < x < 63
4o		>10000	22.0	12.5	29.6	31 < x < 63
4p		>10000	9.8	105.8	25.5	31 < x < 63
4q		>10000	213.4	263.9	304.6	31 < x < 63
4r		>10000	19.4	9.0	21.8	31 < x < 63

^aReported IC₅₀ values are the average of three replicates. The assay as performed here has a standard error between 10 and 20% suggesting that differences of two-fold or greater are significant.

^bSolubilities were measured in 0.1 M sodium phosphate buffer (pH 7.4) containing 1% of DMSO. The data present a number or a range where the solubility is greater than the lower value. Results are the means of three separate experiments.

In this follow up SAR study, a library of dual inhibitors, containing various polynitrogen-containing heterocycles, including substituted and unsubstituted pyrimidine, quinoxaline and tetrazole rings, was designed with the aim to preserve strong potencies at both FAAH and sEH, improve solubility and potentially stability in the liver microsomes. It also further explores the medicinal chemistry space of the left side dual sEH/FAAH inhibitors.

Results and Discussion

Design, Synthesis and structure-activity relationship studies

Our previous SAR studies showed the importance of bulky heterocyclic aromatic moieties placed on the left side of the pharmacophore for FAAH inhibition, while sEH showed to be more tolerant to various alkyl or aryl functionalities placed on the same side (Fig. 3). Here, we hypothesized that the introduction of a more compact aromatic rings such as a tetrazole (electron rich heterocycle) or pyrimidine (electron deficient) might lead to potency loss for FAAH, but the addition of more nitrogen atoms will improve the solubility and overall metabolic stability.^{21, 22} Since our lead compounds SP 4–5 and JA-112 showed low nanomolar potencies in both sEH and FAAH (Table 1), we expected that this chemical change will not lead to the significant decrease in the potency, i.e. we will still be able to obtain potent dual inhibitors (IC₅₀ values in both sEH and FAAH <50 nM), but will yield dual inhibitors with better physicochemical properties. The addition of nitrogen atoms into the aromatic ring to improve solubility and metabolic stability is a common and previously described method in medicinal chemistry.²³ First, a nitrogen atom with its lone electron pair can act as a hydrogen bond acceptor, thus incorporating more nitrogens might provide new possible hydrogen bonding with the surrounding amino acid residues in the binding pocket of both enzymes, which will in turn increase water solubility. Solubility represents one of the most important factors that contributes to drug bioavailability. Next, if the nitrogen atom’s lone electron pairs are delocalized, it makes the heteroaromatic ring less susceptible to P450 enzymes. Finally, nitrogen incorporation lowers c_log P values and decreases lipophilicity²⁴ which will impair the inhibitor’s ability to cross BBB, the property that we seek in our drug design approach. All compounds reported here were synthesized using a 4-step synthetic route, previously described by us (Scheme 1).¹⁹ First, the sulfonamide 1 was obtained by coupling of 2-chlorobenzenesulfonyl chloride and methyl isonipecotate, followed by ester hydrolysis with lithium hydroxide, yielding carboxylic acid 2 in ~70% over two steps. Next, the commercially available 4-aminophenylboronic acid pinacolyl ester was coupled to 2 under microwave irradiation conditions using EDC as a coupling reagent, furnishing a key intermediate 3 in moderate yields. Lastly, we performed the direct biaryl cross coupling using Suzuki-Miyaura microwave assisted reaction to obtain eighteen target compounds 4a-4r in low to moderate yields. All analogs were evaluated in human FAAH and human, mouse and rat sEH enzymatic inhibition assays (Table 1). Some sEH inhibitors previously identified in our lab (e.g. benzothiazolyl- and 4-phenylthiazolyl- analogs) have shown significant species differences, i.e. analogs were very potent in human and rat sEH but possessed diminished activity or were completely inactive in mouse sEH. However, quinolinyl- and isoquinolinyl- analogs were active in all three tested species. We hypothesized that certain heterocycles present at the left side of the pharmacophore might have unfavorable interactions with the binding pocket of mouse sEH. In this study we tested three additional heterocycles to better understand this species selectivity observed in sEH.

Scheme 1.

Reagents and conditions: (a) Et₃N, CH₂Cl₂, 20 min, 80 °C, microwave irradiation, 78%; (b) 4.6 M aq LiOH; THF/H₂O, 16 h, rt, 78%; (c) EDC, DMAP, CH₂Cl₂, 20 min, 80 °C, microwave irradiation, 65%; (d) R-Br (see Table 1 for R), Pd(PPh₃)₄; K₂CO₃, THF/H₂O, 30 min, 90 °C, microwave irradiation, 7–64%.

Our SAR investigation of the structural requirements for dual sEH/FAAH activity started with two unsubstituted 2- and 5-pyrimidinyl- analogs, 4a and 4b, respectively (Table 1). Both compounds showed similar moderate inhibition for FAAH with IC₅₀ values of 174.9 nM and 196.5 nM for 4a and 4b respectively, and strong potencies for human sEH (19.7 nM for 4a and 5.6 nM for 4b) and rat sEH (18.1 nM for 4a and 4.9 nM for 4b), while only moderate to weak potencies for mouse sEH (342.5 nM and 904.2 nM for 4a and 4b respectively). Again, we noticed species differences characteristic for sEH inhibitors, however, biological results suggest that 2-pyrimidinyl ring is significantly more favorable group than the 5-pyrimidinyl ring in the mouse sEH binding pocket. We were excited to see if this trend will remain throughout the rest of the substituted pyrimidinyl- analogs library. Although the introduction of the pyrimidine moiety led to decreased inhibition potency in FAAH compared to benzothiazole (SP 4–5) or quinoline (JA-112) rings, the pyrimidine ring offers access to many diverse chemical substituents that could improve FAAH inhibition. Additionally, both 4a and 4b performed much better in our solubility tests compared to previously identified dual inhibitors (Table 1). We started with the placement of fluoro-, chloro-, methyl- and methoxy- substituents in the position 5 of the 2-pyrimidinyl ring, as shown in analogs 4c, 4d, 4e and 4f, respectively. These modifications still led to only moderate inhibitory potency in FAAH relative to unsubstituted 2-pyrimidinyl analog 4a and slightly decreased potencies in human and rat sEH. However, all these compounds were active in mouse sEH, with methoxy analog 4f being the most active with IC₅₀ value of 32.4 nM. Additionally, these substitutions did not improve general water solubility (Table 1). Next, we introduced fluoro-, chloro-, methyl-, methoxy- and trifluoromethyl- groups in the position 2 of the 5-pyrimidinyl ring, as exemplified with 4g, 4h, 4i, 4j and 4k, and two chloro- groups in the positions 2 and 4 of the 5-pyrimidinyl ring as shown in analog 4l. These modifications did not improve FAAH activity and even led to the significant decrease in FAAH potency, with IC₅₀ = 1239.1 nM in methyl analog 4i and IC₅₀ = 1996.4 nM in disubstituted chloro- analog 4l. The potencies in human and rat sEH were still in the low nanomolar range, except for the trifluoromethyl analog 4k which showed a significant decrease compared to 4b. The activities in mouse sEH were similar to unsubstituted pyrimidinyl- analogs. Solubilities of 5-pyrimidinyl analogs 4g-4l were in the same range as the 2-pyrimidinyl analogs 4c-4l. We completed the testing of heterocycles containing two nitrogen atoms with 2-quinoxaline moiety, 4m. This compound showed excellent inhibition potencies in both FAAH (IC₅₀ = 2.9 nM) and sEH (IC₅₀ values of 0.7 nM, 39.1 nM and 0.3 nM in human, mouse and rat sEH, respectively), and represents the most potent dual sEH/FAAH inhibitor reported to date. This is not surprising, since our previous studies revealed that aromatic two-ring fused heterocycle systems (benzothiazoles, quinolines and isoquinolines) are all well-tolerated in both sEH and FAAH binding pockets. However, the introduction of two nitrogen atoms in the aromatic heterocyclic rings seems to be very beneficial for the activity but did not improve the overall solubility. At this stage, we hypothesized that introduction of more than two nitrogen atoms into aromatic heterocycles on the left side of the molecule should not (significantly) alter the activity but could help increasing the solubility. We turned our attention to tetrazole ring and made a small library of tetrazolyl-analogs, 4n-4r to better understand the impact of these modifications on the activity. Surprisingly, the introduction of tetrazole ring led to complete loss of activity in FAAH. However, most of these compounds were very potent in human, mouse and rat sEH, with 4r being the most active with IC₅₀ values of 19.4 nM, 9.0 nM and 21.8 nM in human, mouse and rat sEH, respectively. To our knowledge this is the first reported amide-piperidine-containing sEH inhibitor with no FAAH activity. Our further testing (see Supplemental Data) revealed that 4r only inhibits sEH, and it does not inhibit microsomal EH nor EH3. Most of the previously reported sEH inhibitors inhibit EH3 as well. Analog 4q with eight nitrogen atoms in total, is a compound with the most nitrogen atoms that we synthesized and tested in our lab, however, it possesses only moderate sEH activity and no FAAH activity at all. Since sEH is an important target in many current research settings, including Alzheimer’s disease, anxiety and pulmonary diseases, tetrazole analogs 4n-4r may be of interest for others in the field.

We decided to continue our studies on 2-quinoxalinyl-analog 4m; this compound was scaled up and further evaluated in vitro in human and rat microsomal liver stability assays, binding to opioid receptors and in vivo to determine its effects on voluntary locomotor activity as measured by home cage wheel running.

Microsomal liver assays and NIMH PDSP evaluation of 4m

Microsomal liver assays (MLAs) were used to preliminary assess the metabolic stability of dual inhibitors before we test them in vivo. The drug’s half-life (t_1/2), which is helpful to predict the duration of drug action in the body were determined. In addition, this analysis can predict the rate at which the drug will be cleared from the body, i.e. apparent intrinsic clearance (CL_int). In general, intrinsic clearance in microsomal liver assay below 13.2 μL/min/mg are considered low clearance, while those above 71.9 are classified as high clearance compounds.²⁵ Drugs that are rapidly metabolized may have shorter durations of effect and may require more frequent dosing.

The results for human, mouse and rat microsomal liver assays for 4m and our previously identified dual inhibitor SP 4–5 are summarized in the Table 2. 4m has a good half-life, t_1/2 > 60 minutes, and a low to moderate intrinsic clearance of 22.46 μL/min/mg protein in human liver microsomes (see Supplemental document for the detailed calculations). Compared to SP 4–5 we observed significant improvement in stability in human MLA, suggesting that the 2-quinoxaline moiety is less prone to degradation than the benzothiazole moiety. In addition, SP 4–5 with the apparent intrinsic clearance value of 76.5 μL/min/mg is considered a high clearance drug. Interestingly, there is a significant ~2-fold decrease in half-lives of 4m in mouse and rat microsomes, i.e. t_1/2 values of 32 minutes and 26 minutes, respectively. The intrinsic clearance of 4m showed moderate clearance rates in both mouse and rat liver microsomes. We observed similar trends in the half life and apparent intrinsic clearance values of SP 4–5. The half-life values for 4m are in the ranges to predict the acceptable stability in liver metabolism, and this is especially important for the oral applications since drugs administrated per os undergo “First-pass metabolism”. More importantly, we were able to improve the metabolic stability (and keep the activity) of our lead compound SP 4–5 by replacing the benzothiazole ring with 2-quinaxolinyl moiety, supporting our hypothesis that this part of the molecule is probably involved in the initial liver microsomes degradative processes.

Table 2.

Stability in Human, Mouse and Rat Liver Microsomes (t_1/2) for 4m and SP 4–5

Compound	Species	LM t1/2 (min)a	CLint,app (μL/min/mg)b
4m	Human	> 60	22.46
	Mouse	32	42.8
	Rat	26	52.6
SP 4–5 c	Human	18.1	76.5
	Mouse	11.3	122.6
	Rat	8.82	157.1

^aData represents averages of duplicate or triplicate determination. LMt_1/2 is the half-life in human, mouse or rat liver microsomes.

^bApparent intrinsic clearance (CL_int,app) was calculated based on CL_int = k/P, where k is the elimination rate constant and P is the protein concentration in the incubation.

^cPreviously published.

Since our long-term goal is to develop a non-opioid drug for the treatment of pain, the compounds are designed to not to cross blood-brain barrier (BBB), but to rather inhibit sEH and FAAH on the periphery. Currently, we only use in silico predictions by Gupta et al.²⁶, to predict whether the tested drug will be able to cross BBB (see Table 6). To complement this prediction, 4m was sent to National Institute of Mental Health Psychoactive Drug Screening Program (NIMH PDSP) to evaluate if it binds to serotonin (5-HT) and opioid receptor subtypes: delta (DOR), kappa (KOR) and mu (MOR), and to nociceptin (NOP). As shown in Table 3, 4m did not significantly bind to any of the four opioid receptors tested, where significant inhibition is considered if the values are > 50%. In cases where negative inhibition (−) is seen, this represents a stimulation of binding. Occasionally, compounds at high concentrations will non-specifically increase binding. However, 4m appears to bind to serotonin receptor subtypes 1B and 2C (Table 3), but no other subtypes. These results confirmed that if any potential analgesic effects of 4m would not be from stimulation of any of these opioid receptors. Additionally, not binding to opioid receptors will also avoid other opioid-related adverse effects, such as sedation, respiratory depression and tolerance and addiction.

Table 6.

Selected ADMET predicted properties for 4m and other standard compounds

Compound	BBB Score	Human Plasma t1/2 (h)	log Caco2 (cm/s)	log LD50
TPPU	3.17	9.79	−4.73	1.68
URB-597	2.43	2.13	−4.98	1.70
SP 4–5	1.84	1.44	−5.27	1.85
JA-112	1.96	2.60	−5.26	1.96
4m	1.18	2.37	−5.25	1.27

Table 3.

NIMH PDSP screening results for binding of 4m to opioid and serotonin receptor subtypes.

Receptor	Inhibition (%) a	P(value = 0)b
Delta opioid (DOR)	−0.23 ± 11.89	0.9858
Kappa opioid (KOR)	5.59 ± 6.57	0.4574
Mu opioid (MOR)	−1.00 ± 1.63	0.5829
Nociceptin (NOP)	−5.22 ± 3.98	0.2810
5-HT1A	−18.02 ± 14.59	0.3047
5-HT1B	22.18 ± 6.67	0.0069*
5-HT1D	−8.07 ± 3.46	0.1022
5-HT2A	1.96 ± 1.88	0.1284
5-HT2B	−4.41 ± 7.06	0.3002
5-HT2C	−18.22 ± 3.45	0.0018*
5-HT4	29.56 ± 23.22	0.0842

^aData represent average represent average ± SEM of inhibition (n = 4) for 4m tested at 10 μM.

^bProbability that results are equal to zero (no inhibition nor activation (negative inhibition) of the receptor).

^*indicates p < 0.05

Docking experiments

To better understand the binding of 4m and its interactions with both FAAH and sEH, docking experiments were performed. A homology model for human FAAH was previously prepared since there is no crystal structure currently available.²⁷ Since 4m is a direct follow up analog of the previously published quinoline library, we compared binding of JA-112 (the most potent analog identified in quinoline series) in both sEH and FAAH with 4m. Briefly, using X-ray crystal structure of rat FAAH (PDB: 3LJ6), we constructed a human FAAH model and after energy minimization and validation using several prediction tools, we proceeded with the identification of potential binding pockets. 4m was subsequentially docked into the model and our visual analysis revealed that it binds between the Catalytic Triad (S241-S217-K142) and Oxyanion hole, which is composed of S241, G240, G239, and I238 amino acid residues, (Figs. 4A and 4B, and Table 4). We observed one hydrogen bond with S241, a residue that is a part of the FAAH Catalytic Triad, where the hydrogen from the amide bond of 4m acts as a hydrogen bond donor and the oxygen on the serine residue is a hydrogen bond acceptor. The terminal 2-chlorophenyl moiety of 4m is surrounded with several hydrophobic residues V491, L380, M495, F432, F381, T488 and G239 (a residue from the Oxyanion hole). Piperidine ring, a part of the central pharmacophore, is engaged in the non-covalent interactions with S193, G240 (from the Oxyanion hole) and S217 (from the Catalytic Triad). Next, the phenyl ring is interacting with F192, I238 (from the Oxyanion hole), S190 and M191 amino acid residues. Finally, the 2-quinoxaline moiety has numerous hydrophobic non-covalent interactions including C269, V270, Y271, L266, K267, G268, L278 and F192. The broken black line around 4m in Fig. 4b represents the open space towards solvent. This is important for the future drug design, since we believe that various groups still can be placed at the positions 5 and 6 of the 2-quinoxalinyl moiety and further improve the binding interactions with FAAH. We previously published a binding pose of JA-112¹⁹ and here we analyzed the binding of this quinoline analog compared to 4m (Table 4) in order to explain why the quinoxaline moiety provides better inhibition than quinoline at the same position. We first observed less non-covalent interactions of JA-112 compared to 4m, i.e. one hydrogen bond and eleven hydrophobic interactions compared to one hydrogen bond and 16 hydrophobic interactions. Another noticeable feature, 4m establishes a hydrogen bond with S241, a residue that is a part of Catalytic Triad, while JA-112 has potential hydrogen bond with L278, a residue that is located more towards the entrance to the binding site. We hypothesize that superior FAAH potency of 4m (IC₅₀ value of 2.9 nM) compared to IC₅₀ = 19.6 nM for JA-112 is due to these additional non-covalent interactions and direct hydrogen bonding to the Catalytic Triad residue S241.

Figure 4.

A) 4m in the binding pocket of human FAAH – 2D representation: green shading represents hydrophobic region; gray parabolas represent accessible surface for large areas; broken thick line around ligand shape indicates accessible surface; size of residue ellipse represents the distance from the surrounding amino acid residues. B) 3D-docking pose of 4m in the binding site of the human FAAH with the surrounding amino acid residues within 5 Å distance.

Table 4.

The list of hydrogen bonds and hydrophobic interactions of 4m and JA-112 docked in human FAAH with the distances from amino acid residues.

4m			JA-112
Residue	Type	Distance (Å)	Residue	Type	Distance (Å)
S241	H-bond	2.77	L278	H-bond	2.91
M191	Hydrophobic	3.63	M191	Hydrophobic	3.79
F192	Hydrophobic	3.36	F192	Hydrophobic	3.34
S193	Hydrophobic	3.56	I238	Hydrophobic	3.14
Y194	Hydrophobic	4.43	S262	Hydrophobic	4.44
I238	Hydrophobic	3.26	K263	Hydrophobic	3.95
G239	Hydrophobic	3.62	C269	Hydrophobic	4.33
L266	Hydrophobic	3.76	V270	Hydrophobic	3.53
K267	Hydrophobic	4.49	Q273	Hydrophobic	3.95
G268	Hydrophobic	3.27	R277	Hydrophobic	4.41
Y271	Hydrophobic	3.24	T308	Hydrophobic	3.58
L278	Hydrophobic	3.10	V309	Hydrophobic	4.38
L380	Hydrophobic	4.41
F381	Hydrophobic	3.65
L404	Hydrophobic	3.82
F432	Hydrophobic	4.04
T488	Hydrophobic	4.47
V491	Hydrophobic	4.31
M495	Hydrophobic	3.86

For docking purposes in sEH, a crystallographic structure of human sEH (PDB: 4HAI) with piperidine-amide inhibitor in it was used.²⁸ The active site of sEH, where EpFAs are hydrolyzed into DHETs, is composed of D335-Y383-Y466 catalytic amino acid residues, and two binding pockets on each side of the catalytic triad that can fit various aliphatic and aromatic bulky hydrophobic groups. After docking experiments (Figs. 5A, 5B and Table 5), we first observed two hydrogen bonds: the oxygen of the amide bond acts as a hydrogen bond acceptor with the Y466, a residue that is a part of the catalytic triad, and the oxygen from the sulfonamide group act as a hydrogen bond acceptor with Q384 amino acid residue. Also, there are several non-covalent interactions with 2-chlorophenyl ring including I363, M503, F381, M469, P361, T360 and M339 amino acid residues. The piperidine ring is engaged in the interactions with L499 and Y383 (a part of the Catalytic Triad), followed by the phenyl ring that interacts with W336, W525, H524, V498 and D335 (a part of the Catalytic Triad) residues. The 2-quinoxalinyl forms several hydrophobic interactions with the L408, L417, R410, S407 and A411. Again, the 2-quinoxalinyl moiety is orientated towards the solvent, similar to FAAH docking experiments, which suggests that more potent dual inhibitors might be obtained by introducing hydrophobic/lipophilic groups on the 2-quinoxaline ring. Using our published docking experiments of JA-112 in sEH¹⁹, we compared the binding poses of 4m and JA-112 and their interactions with the binding pocket of sEH to better understand why the quinoxaline analog is better tolerated than the quinoline moiety (Table 5). We observed the same number of hydrogen bonds in both 4m and JA-112, i.e. two hydrogen bonds, but 4m has 15 hydrophobic interactions compared to 11 in JA-112. Both compounds were very potent in human sEH, with the IC₅₀ values of 0.7 and 1.7 nM for 4m and JA-112, respectively, and we believe this is due to hydrogen bonds with the residues from the sEH Catalytic Triad.

Figure 5.

A) 4m in the binding pocket of human sEH – 2D representation: green shading represents hydrophobic region; gray parabolas represent accessible surface for large areas; broken thick line around ligand shape indicates accessible surface; size of residue ellipse represents the distance from the surrounding amino acid residues. B) 3D-docking pose of 4m in the binding site of the human sEH with the surrounding amino acid residues within 5 Å distance.

Table 5.

The list of hydrogen bonds and hydrophobic interactions of 4m and JA-112 docked in human sEH with the distances from amino acid residues.

4m			JA-112
Residue	Type	Distance (Å)	Residue	Type	Distance (Å)
Q384	H-bond	3.30	Y383	H-bond	1.41
Y466	H-bond	2.68	D335	H-bond	2.62
W336	Hydrophobic	3.59	F267	Hydrophobic	4.38
M339	Hydrophobic	3.21	W336	Hydrophobic	3.52
T360	Hydrophobic	3.80	M339	Hydrophobic	3.31
P361	Hydrophobic	4.00	P361	Hydrophobic	4.15
I363	Hydrophobic	3.74	L417	Hydrophobic	4.07
F381	Hydrophobic	3.47	M419	Hydrophobic	3.60
Y383	Hydrophobic	4.13	Y466	Hydrophobic	3.90
L408	Hydrophobic	3.92	V498	Hydrophobic	3.17
R410	Hydrophobic	4.00	L499	Hydrophobic	3.63
L417	Hydrophobic	4.11	H524	Hydrophobic	2.93
V498	Hydrophobic	3.23	W525	Hydrophobic	3.77
L499	Hydrophobic	4.22
M503	Hydrophobic	3.75
H524	Hydrophobic	3.01
W525	Hydrophobic	3.66

ADMET Predictions

Using the ICM Pro – Chemist tool we predicted several absorption, distribution, metabolism, excretion and toxicity (ADMET) properties. Table 6 has summarized predicted BBB and log Caco-2 scores, predicted human plasma half-lives and log LD₅₀ values. We compared these scores for the standards, sEH inhibitor TPPU and FAAH inhibitor URB-597, previously identified dual sEH/FAAH inhibitors SP 4–5 and JA-112, with the most potent dual inhibitor identified in this study – 4m. An important part of our non-opioid drug design process is to predict if the analogs will cross the BBB, i.e. we prefer our compounds not to penetrate BBB and thus not interact with opioid receptors or any other receptor within the central nervous system (CNS) to produce CNS-dependent side effects. The ICM Pro predicting tool uses an algorithm based on Gupta et al.²⁹, and all predicted scores > 4 indicate that the drug can penetrate the BBB. As outlined in Table 6, none of compounds tested in our lab has predicted score above 4, suggesting that these analogs will not cross the BBB. One of the important factors we want to predict is the plasma half-life of the designed dual inhibitors. Our assessment showed that inhibitors 4m and JA-112 have better predicted half-lives compared to our first lead compound SP 4–5. Previously we were able to experimentally determine the half-life value in human plasma for JA-112, and we obtained a value of 7.26 hours.¹⁹ This is a much higher number than the predicted value obtained by ICM Pro, but we are looking at these numbers more as a guide and to determine the ratio than the real values. In summary, we expect that 4m has a higher plasma stability than SP 4–5 and should be stable as JA-112. The next pharmacokinetic parameter we were interested in is permeability, and we assessed it using the Caco-2³⁰ prediction tool. The ICM Pro Chemist prediction scores are expressed as logarithm and values higher than −5 suggest a highly permeable drug candidate, while scores of below −6 represent a poorly permeable compound. As shown in Table 6, all three assessed dual inhibitors have predicted scores very similar and values are between −5 and −6, suggesting moderately permeable drug candidates. As a part of toxicity assessment, LD₅₀ values were obtained. ICM Pro Chemist tool provides the logarithm of the LD₅₀, where log LD₅₀ values equal to or below 0 indicate high toxicity, i.e. these values correlate to LD₅₀ values below 1 mg/kg. On the other hand, log LD₅₀ values that are equal or above 2 correlate to LD₅₀ values above 100 mg/kg, meaning very low toxicity. As shown in Table 6, all tested compounds have moderate toxicity, i.e. log LD50 values between 1 and 2. However, our tool predicts that 4m has the lowest log LD₅₀ value of 1.27, i.e. the closest to 1, therefore we wanted to examine whether 4m affects behavior in vivo in rats and to validate these in silico results.

In vivo evaluation of 4m

Wheel running, a voluntary behavior in rats, was measured in the rat’s home cage following drug administration to evaluate effects of 4m on locomotion. The amount of wheel revolutions in the 15 min after drug injection significantly differed between rats receiving an intraperitoneal administration of 3 mg/kg 4m, 3 mg/kg morphine, or vehicle [F(2,21) = 6.93, p = 0.005]. A Tukey post-hoc test revealed wheel running was significantly lower in rats receiving 3 mg/kg morphine (Vehicle vs. 3 mg/kg morphine; p = 0.005) and 3 mg/kg 4m (Vehicle vs. 3 mg/kg 4m, p = 0.03), but not rats receiving 3 mg/kg 4m and 3 mg/kg morphine (3 mg/kg 4m vs. morphine, p > 0.05). These results suggest that administration of 4m likely produces an undesirable effect that interferes with normal behavior.

The present data indicate that administration of 4m significantly reduces voluntary wheel running in rats shortly after injection, similar to the depression of wheel running observed following morphine injection (Fig. 6). Many studies have demonstrated the depression of wheel running³¹ and other locomotor behaviors³² following administration, especially at therapeutic doses³³. These effects are likely due to the sedative properties of morphine due to its action at the mu-opioid receptor within the central nervous system. However, given 4m does not bind to opioid receptors (Table 3), the data suggest that 4m, at 3 mg/kg, may engage other pathways that produce undesirable side effects. In contrast, our previous dual FAAH/sEH inhibitor SP 4–5, at the same dose of 3 mg/kg, did not inhibit wheel running in male rats¹⁷. Thus, this depression of activity may not be a property of dual FAAH/sEH inhibition itself, but of various structural properties or off-target interactions of 4m. There are several other potential reasons for this discrepancy. While 4m lacked measurable affinity for opioid receptors, it may interact with other neurotransmitter systems that could interfere with wheel running. Table 3 revealed that 4m binds to certain serotonin receptor subtypes, which may impact mood or locomotor activity. For example, 4m may enhance GABA receptor modulation, interact with histamine receptors to produce drowsiness, or interact with dopamine receptors to blunt motivation/reward especially considering wheel running is a rewarding activity. Many of these effects may only be possible if 4m penetrates the BBB; however, a more detailed screen of relevant CNS-mediated receptors and BBB penetrability may reveal these compound-specific properties that are not universal to all dual FAAH/sEH inhibitors.

Figure 6.

Wheel running following injections of vehicle, morphine, or 4m. Wheel running was measured for 15 min following intraperitoneal administration of the drug. Wheel running significantly differed between animals receiving vehicle, morphine, and 4m. * indicates p < 0.05; ** indicates p < 0.01). n = 8/group.

The utility of wheel running as a tool in preclinical drug discovery is especially apparent in this context. Given the voluntary and rewarding nature of wheel running to rats³⁴, drug-induced depression of wheel running can be interpreted in many ways. For example, this depression may not necessarily be opioid-like sedation, but may actually be anxiety, anhedonia, or other forms of toxicity. Thus, depression of wheel running in an early-stage drug discovery context represents an early warning sign of potential tolerability issues or may confound findings from assays of pain-evoked behaviors^{35, 36}. Given that 4m reduced wheel running at a dose comparable to that used to test for antinociceptive effects highlights the importance of incorporating these measures early in the preclinical evaluation of dual FAAH/sEH inhibitors.

Future experiments will be designed to determine whether the locomotor-depressing effects of 4m reflects sedation or is secondary to other effects such as anxiety. While these experiments only tested one dose, testing across a broader dose range may reveal whether depression of activity is only specific to certain doses or whether it is simply an unavoidable feature of 4m. Dual FAAH/sEH inhibition remains a promising therapeutic strategy; however, the current work indicates that each candidate compound must be evaluated independently. This depression of wheel running emphasizes that even within the same mechanistic drug class, structural features will influence tolerability, safety, and efficacy.

Conclusions

Following-up on our previous findings where compounds have low solubility and significant species differences in dual sEH/FAAH inhibitors, we decided to test several heteroaromatic moieties, including pyrimidine, quinoxaline, and tetrazole rings. Using a four-step synthetic route eighteen compounds were successfully synthesized and evaluated their inhibitory effects on FAAH and sEH enzymes across different species, including humans, mice, and rats. Modifications to the pyrimidine ring led to slight decreased FAAH inhibition but improved solubility compared to previous dual inhibitors (Fig. 7). The introduction of tetrazoles led to complete loss of potencies in FAAH but showed that this moiety can still yield a potent sEH inhibitors, particularly 4r, which has selective activity for sEH only and no other epoxy hydrolases. The introduction of heterocycles containing two nitrogen atoms demonstrated excellent inhibition potency, particularly 4m, an inhibitor with the 2-quinoxaline moiety, which represents the most active dual sEH/FAAH inhibitor reported to date, with IC₅₀ values of 2.9 nM in human FAAH and 0.7 nM, 39.1 nM and 0.3 nM in human, mouse and rat sEH, respectively. 4m was scaled up for further in vitro and in vivo evaluations. Microsomal liver assays revealed moderate clearance rates for compound 4m, indicating acceptable stability for potential oral administration. In addition, this dual inhibitor does not interact with opioid or many serotonin receptors. Docking experiments highlighted the numerous interactions of 4m with FAAH and sEH, explaining its high potencies for both enzymes and suggesting that introducing hydrophobic groups on the 2-quinoxaline ring may further enhance inhibitor potency. However, similar to the opioid morphine, intraperitoneal administration of 4m significantly depressed voluntary wheel running in rats indicating that 4m may produce undesirable effects, which will be further investigated in our next work. Based on our previous work (Fig. 7) the future optimization of the right side should still yield potent dual inhibitors which will provide more insights if the unexpected in vivo results of 4m are due to 2-quinoxaline moiety (and its metabolite(s)) or it is something only observed in this particular compound.

Figure 7.

A summary of previous SAR studies and main SAR of this study (shown in red). A pharmacophore (in green) and future directions (potential for optimizations) are also presented.

Materials and methods

All starting materials, reagents and solvents were commercially available and used without further purification. Precursors for tetrazolyl analogs 4n, 4p, 4q and 4r were prepared at Purdue University and are previously reported.^37–39 Synthesis and analytical data for 5-bromo-1-cyclopropyl-1H-tetrazole, a precursor for 4o, are provided in Supporting material. Reactions were monitored by thin-layer chromatography (TLC) performed on silica gel 60 F₂₅₄ precoated aluminum plates obtained from Sigma–Aldrich. All intermediates and final compounds were purified via flash chromatography using Teledyne CombiFlash Rf+ Separation System and prepacked silica gel columns from Teledyne. The purity of all final compounds was determined by TLC, LC-MS, and proton and carbon NMR, and it was equal or higher than 95%. Proton (¹H) and carbon (¹³C) NMR spectra were recorded on a Bruker Avance-II spectrometer at 400 MHz and 101 MHz, respectively, in deuterated chloroform (CDCl₃) with tetramethylsilane as an internal reference, or deuterated dimethyl sulfoxide (DMSO-d₆). Chemical shifts δ are reported in parts per million (ppm) and coupling constants (J) are expressed in Hz. Abbreviations for multiplicities are reported as follows: s = singlet, bs = broad singlet, d = doublet, t = triplet, q = quartet, m = multiplet. LCMS analysis was performed on a Shimadzu 2050 System with LC grade solvents. The following conditions were used: eluent A, water +0.05% TFA; eluent B, acetonitrile +0.05% TFA with isocratic gradient conditions (5 min, A = 20% and B = 80%) with a flow rate of 0.4 mL/min; analytical column: Phenomenex Kinetex 1.7 μm C18 100 Å, 50 mm × 2.1 mm. The measurements were run in positive ion mode. HRMS analysis was performed on a Thermo Scientific Q Exactive Focus Orbitrap LC-MS/MS System with LC grade solvents and following conditions were used: eluent A, water +0.05% TFA; eluent B, acetonitrile +0.05% TFA with linear gradient conditions (0–4 min, linear increase from A = 60% and B = 40% to A = 0% and B = 100%; 4–5 min, B = 100%; 5 min, decrease to B = 10%; 5–7 min B = 10%) with a flow rate of 0.4 mL/min; analytical column: Phenomenex Kinetex 5 μm XB-C18 100 Å, 250 mm × 4.6 mm The measurements were run in positive ion mode. Microwave reactions were carried out in a CEM 2.0 Discover microwave synthesizer. Melting points were measured with a DigiMelt MPA-160 melting point apparatus and are reported uncorrected. Fatty Acid Amide Hydrolase (human, recombinant) was purchased from Cayman Chemical and Soluble Epoxide Hydrolase (human, mouse, and rat, recombinant) were obtained from UC Davis. Figs. 1 and 3 were created with BioRender.com. Molecular modeling studies and docking experiments were performed using ICM Pro Molsoft software.

Chemistry

General procedure for the preparation of analogs 4a-4r

Preparation of intermediates 1, 2 and 3 has been previously reported by us.¹⁹ The final analogs 4a-4r were obtained as follows: 3 (0.2604 mmol), corresponding bromo-polynitrogen containing compound (0.2604 mmol), and potassium carbonate (0.6512 mmol) were suspended in a mixture of solvents tetrahydrofuran and water (4 to 1). Palladium-tetrakis(triphenylphosphine) (0.0130 mmol) was added and the reaction mixture was subjected to microwave irradiation at 90 °C for 30 min. The solvent mixture was removed under vacuo, and the residue was dissolved in dichloromethane (20 mL) and water (20 mL). The organic layer was transferred to the separatory funnel, separated, dried over anhydrous sodium sulfate, filtered, and concentrated. The crude product was purified using a CombiFlash system using ethyl acetate and hexane, or dichloromethane and methanol as solvents, over 20 min and final products 4a-4r were obtained.

1-((2-chlorophenyl)sulfonyl)-N-(4-(pyrimidin-2-yl)phenyl)piperidine-4-carboxamide (4a) was obtained as a white solid in the amount of 0.037 g (41% yield); mp 190–191 °C. ¹H NMR (400 MHz; DMSO‑d₆): δ 10.17 (s, 1H), 8.85 (d, J = 4.8 Hz, 2H), 8.33 (d, J = 8.7 Hz, 2H), 8.02–8.00 (m, 1H), 7.75–7.67 (m, 4H), 7.60–7.56 (m, 1H), 7.38 (t, J = 4.8 Hz, 1H), 3.79–3.76 (m, 2H), 2.87–2.81 (m, 2H), 2.55–2.53 (m, 1H), 1.91–1.87 (m, 2H), 1.67–1.57 (m, 2H). ¹³C NMR (101 MHz; DMSO‑d₆): δ 173.4, 163.4, 158.0, 142.0, 136.3, 134.9, 132.8, 132.3, 132.0, 131.4, 128.8, 128.3, 119.8, 119.4, 100.0, 45.2, 42.2, 28.5 ppm. ESI-MS (M⁺+H): 457.

1-((2-chlorophenyl)sulfonyl)-N-(4-(pyrimidin-5-yl)phenyl)piperidine-4-carboxamide (4b) was obtained as a white solid in the amount of 0.059 g (65% yield); mp 256– 257 °C. ¹H NMR (400 MHz; DMSO-d₆): δ 10.13 (s, 1H), 9.15–9.10 (m, 3H), 8.02–8.00 (m, 1H), 7.78–7.67 (m, 6H), 7.60–7.58 (m, 1H), 3.79–3.76 (m, 2H), 2.87–2.79 (m, 2H), 2.55–2.51 (m, 1H), 1.90–1.86 (m, 2H), 1.63–1.59 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆): δ 173.4, 157.3, 154.6, 140.4, 136.3, 134.9, 133.2, 132.8, 132.0, 131.4, 128.6, 128.3, 127.7, 120.2, 45.2, 42.2, 28.5 ppm. ESI-MS (M⁺+H): 457.

1-((2-chlorophenyl)sulfonyl)-N-(4-(5-fluoropyrimidin-2-yl)phenyl)piperidine-4-carboxamide (4c) was obtained as a white solid in the amount of 0.0137 g (12% yield); mp 230– 231 °C. ¹H NMR (400 MHz; DMSO-d₆): δ 10.17 (s, 1H), 8.92 (s, 2H), 8.26 (d, J = 8.7 Hz, 2H), 8.01 (dd, J = 7.8, 0.8 Hz, 1H), 7.75–7.67 (m, 4H), 7.60–7.56 (m, 1H), 3.78 (d, J = 12.6 Hz, 2H), 2.87–2.81 (m, 2H), 2.55–2.52 (m, 1H), 1.90–1.87 (m, 2H), 1.67–1.57 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆): δ 173.4, 160.10, 160.05, 155.6, 146.0, 145.8, 141.9, 136.3, 134.9, 132.8, 132.0, 131.43, 131.35, 128.8, 128.3, 119.5, 45.2, 42.2, 28.5 ppm. ESI-MS (M⁺+H): 475.

1-((2-chlorophenyl)sulfonyl)-N-(4-(5-chloropyrimidin-2-yl)phenyl)piperidine-4-carboxamide (4d) was obtained as a white solid in the amount of 0.024 g (24% yield); mp > 260 °C. ¹H NMR (400 MHz; DMSO-d₆): δ 10.13 (s, 1H), 9.09–9.01 (m, 2H), 8.02–8.00 (m, 1H), 7.79–7.67 (m, 6H), 7.60–7.56 (m, 1H), 3.79–3.76 (m, 2H), 2.87–2.81 (m, 2H), 2.55–2.51 (m, 1H), 1.90–1.87 (m, 2H), 1.66–1.57 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆): δ 173.4, 158.8, 157.9, 140.7, 136.3, 134.99, 134.95, 132.8, 132.27, 132.08, 132.01, 131.4, 128.3, 127.8, 127.3, 120.4, 120.1, 45.2, 42.2, 28.5 ppm. ESI-MS (M⁺+H): 491.

1-((2-chlorophenyl)sulfonyl)-N-(4-(5-methylpyrimidin-2-yl)phenyl)piperidine-4-carboxamide (4e) was obtained as a white solid in the amount of 0.0320 g (34% yield); mp 251–252 °C. ¹H NMR (400 MHz; DMSO-d₆): δ 10.13 (s, 1H), 8.69 (s, 2H), 8.30–8.28 (m, 2H), 8.02–8.00 (m, 1H), 7.71–7.69 (m, 4H), 7.60–7.58 (m, 1H), 3.78 (dt, J = 11.7, 0.3 Hz, 2H), 2.87–2.81 (m, 2H), 2.55–2.52 (m, 1H), 2.28 (s, 3H), 1.90–1.87 (m, 2H), 1.67–1.58 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆): δ 173.3, 161.2, 157.9, 141.7, 136.4, 134.9, 132.8, 132.4, 132.0, 131.4, 128.8, 128.48, 128.29, 119.4, 45.2, 42.2, 28.5, 15.4 ppm. ESI-MS (M⁺+H): 471.

1-((2-chlorophenyl)sulfonyl)-N-(4-(5-methoxypyrimidin-2-yl)phenyl)piperidine-4-carboxamide (4f) was obtained as a white solid in the amount of 0.0511 g (53% yield); mp 239–242 °C. ¹H NMR (400 MHz, DMSO-d₆) δ 10.09 (d, J = 0.7 Hz, 1H), 8.60 (s, 2H), 8.23 (dd, J = 6.8, 0.3 Hz, 2H), 8.03–8.01 (m, 1H), 7.72–7.70 (m, 4H), 7.60–7.58 (m, 1H), 3.94 (s, 3H), 3.79–3.76 (m, 2H), 2.88–2.81 (m, 2H), 2.54–2.51 (m, 1H), 1.90–1.87 (m, 2H), 1.64–1.58 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆): δ 173.3, 156.6, 152.2, 144.2, 141.1, 136.4, 134.9, 132.8, 132.3, 132.0, 131.4, 128.3, 128.1, 119.4, 56.6, 45.2, 42.1, 28.5 ppm. ESI-MS (M⁺+H): 487.

1-((2-chlorophenyl)sulfonyl)-N-(4-(2-fluoropyrimidin-5-yl)phenyl)piperidine-4-carboxamide (4g) was obtained as a pale pink solid in the amount of 0.0744 g (64% yield); mp > 260 °C. ¹H NMR (400 MHz, DMSO-d₆) δ 10.11 (s, 1H), 9.09 (d, J = 1.5 Hz, 2H), 8.02 (dd, J = 7.9, 1.5 Hz, 1H), 7.80–7.67 (m, 6H), 7.58 (ddd, J = 7.8, 7.1, 1.6 Hz, 1H), 3.80–3.76 (m, 2H), 2.85 (td, J = 12.3, 2.2 Hz, 2H), 2.55–2.51 (m, 1H), 1.91–1.87 (m, 2H), 1.67–1.57 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆): δ 173.3, 163.0, 160.9, 159.18, 159.06, 140.4, 136.4, 134.9, 132.8, 132.22, 132.17, 132.00, 131.4, 128.3, 127.7, 127.5, 120.1, 45.2, 42.2, 28.5 ppm. ESI-MS (M⁺+H): 475.

1-((2-chlorophenyl)sulfonyl)-N-(4-(2-chloropyrimidin-5-yl)phenyl)piperidine-4-carboxamide (4h) was obtained as a white solid in the amount of 0.0051g (7% yield); mp > 260 °C. ¹H NMR (400 MHz, DMSO-d₆) δ 10.14 (s, 1H), 9.08 (s, 2H), 8.02–8.00 (m, 1H), 7.79–7.67 (m, 5H), 7.65–7.56 (m, 2H), 3.79–3.76 (m, 2H), 2.87–2.81 (m, 2H), 2.54–2.51 (m, 1H), 1.90–1.86 (m, 2H), 1.66–1.57 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆): δ 173.4, 158.8, 157.9, 140.7, 136.3, 135.0, 132.8, 132.3, 132.00, 131.89, 131.4, 129.28, 129.17, 128.3, 127.8, 127.3, 120.1, 45.2, 42.2, 28.5 ppm. ESI-MS (M⁺+H): 491.

1-((2-chlorophenyl)sulfonyl)-N-(4-(2-methylpyrimidin-5-yl)phenyl)piperidine-4-carboxamide (4i) was obtained as a white solid in the amount of 0.0564 g (60 % yield); mp > 260 °C. ¹H NMR (400 MHz, DMSO-d₆) δ 10.08 (s, 1H), 8.98 (s, 2H), 8.02 (dd, J = 7.9, 1.5 Hz, 1H), 7.74–7.67 (m, 6H), 7.58 (ddd, J = 7.8, 7.1, 1.6 Hz, 1H), 3.78 (d, J = 12.7 Hz, 2H), 2.84 (td, J = 12.3, 2.1 Hz, 2H), 2.64 (s, 3H), 2.54–2.51 (m, 1H), 1.91–1.87 (m, 2H), 1.67–1.58 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆): δ 173.3, 166.1, 154.7, 140.1, 136.4, 134.9, 132.8, 132.0, 131.4, 130.1, 128.9, 128.3, 127.4, 120.1, 45.2, 42.2, 28.5, 25.7 ppm. ESI-MS (M⁺+H): 471.

1-((2-chlorophenyl)sulfonyl)-N-(4-(2-methoxypyrimidin-5-yl)phenyl)piperidine-4-carboxamide (4j) was obtained as a white solid in the amount of 0.0581 g (58% yield); mp 239 – 242 °C. ¹H NMR (400 MHz, DMSO-d₆) δ 10.09 (d, J = 0.7 Hz, 1H), 8.60 (s, 2H), 8.23 (dd, J = 6.8, 0.3 Hz, 2H), 8.03–8.01 (m, 1H), 7.72–7.70 (m, 4H), 7.60–7.58 (m, 1H), 3.94 (s, 3H), 3.79–3.76 (m, 2H), 2.88–2.81 (m, 2H), 1.90–1.87 (m, 2H), 1.64–1.58 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆): δ 173.3, 156.6, 152.2, 144.2, 141.1, 136.4, 134.9, 132.8, 132.3, 132.0, 131.4, 128.3, 128.1, 119.4, 56.6, 45.2, 42.1, 28.5 ppm. ESI-MS (M⁺+H): 487.

1-((2-chlorophenyl)sulfonyl)-N-(4-(2-(trifluoromethyl)pyrimidin-5-yl)phenyl)piperidine-4-carboxamide (4k) was obtained as a white solid in the amount of 0.031 g (26% yield); mp > 260 °C. ¹H NMR (400 MHz; DMSO-d₆): δ 10.11 (s, 1H), 9.31 (s, 2H), 7.96 (dd, J = 7.9, 1.0 Hz, 1H), 7.82 (d, J = 8.7 Hz, 2H), 7.73 (d, J = 8.7 Hz, 2H), 7.68–7.61 (m, 2H), 7.55–7.51 (m, 1H), 3.74–3.71 (m, 2H), 2.82–2.76 (m, 2H), 2.49–2.47 (m, 1H), 1.85–1.82 (m, 2H), 1.62–1.51 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆): δ 173.4, 155.8, 153.8, 153.4, 141.2, 136.3, 135.3, 134.96, 134.90, 132.7, 132.0, 131.3, 128.36, 128.29, 128.25, 127.1, 121.6, 120.1, 118.9, 45.1, 42.2, 28.4 ppm. ESI-MS (M⁺+H): 525.

1-((2-chlorophenyl)sulfonyl)-N-(4-(2,4-dichloropyrimidin-5-yl)phenyl)piperidine-4-carboxamide (4l) was obtained as a white solid in the amount of 0.038g (36% yield); mp 157–159 °C. ¹H NMR (400 MHz, DMSO-d₆) δ 10.25 (s, 1H), 9.04 (s, 1H), 8.03–8.01 (m, 1H), 7.82–7.75 (m, 5H), 7.73–7.68 (m, 2H), 7.61–7.57 (m, 1H), 3.80–3.76 (m, 2H), 2.88–2.82 (m, 2H), 2.54–2.50 (m, 1H), 1.91–1.88 (m, 2H), 1.66–1.57 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆): δ 173.6, 166.0, 163.5, 158.9, 142.1, 136.4, 134.9, 132.8, 132.0, 131.4, 130.7, 130.2, 128.3, 118.9, 117.6, 45.2, 42.2, 28.5 ppm. ESI-MS (M⁺+H): 525.

1-((2-chlorophenyl)sulfonyl)-N-(4-(quinoxalin-2-yl)phenyl)piperidine-4-carboxamide (4m) was obtained as a white solid in the amount of 0.034g (34% yield); mp: 212 – 213 °C. ¹H NMR (400 MHz; DMSO-d₆): δ 10.19 (s, 1H), 9.54–9.52 (m, 1H), 8.31–8.28 (m, 2H), 8.07 (ddd, J = 8.0, 3.9, 1.7 Hz, 2H), 8.00 (dd, J = 7.9, 1.6 Hz, 1H), 7.86–7.76 (m, 4H), 7.72–7.65 (m, 2H), 7.56 (ddd, J = 7.8, 7.1, 1.6 Hz, 1H), 3.78–3.75 (m, 2H), 2.86–2.80 (m, 2H), 2.52 (s, 1H), 1.88 (dd, J = 13.3, 2.8 Hz, 2H), 1.66–1.56 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆): δ 173.4, 150.9, 143.9, 141.88, 141.78, 141.3, 136.3, 134.9, 132.7, 132.0, 131.3, 131.0, 129.9, 129.45, 129.26, 128.5, 128.3, 119.8, 45.2, 42.2, 28.5 ppm. ESI-MS (M⁺+H): 507. HRMS-ESI+: calculated for C₂₆H₂₃ClN₄O₃S + H: 507.1258; Found: 507.1236.

N-(4-(1H-tetrazol-5-yl)phenyl)-1-((2-chlorophenyl)sulfonyl)piperidine-4-carboxamide (4n) was obtained as a white solid in the amount of 0.0054 g (7% yield); mp N/A. ¹H NMR (400 MHz, DMSO-d₆) δ 9.98 (s, 1H), 8.02 (dd, J = 7.9, 1.5 Hz, 1H), 7.74 (dd, J = 8.0, 1.5 Hz, 1H), 7.70 (td, J = 7.5, 1.6 Hz, 1H), 7.64 (d, J = 8.8 Hz, 2H), 7.61–7.56 (m, 3H), 3.79–3.75 (m, 2H), 2.87–2.80 (m, 2H), 2.47–2.46 (m, 1H), 1.90–1.85 (m, 2H), 1.67–1.56 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆): δ 173.0, 138.7, 136.4, 134.94, 134.82, 132.8, 132.0, 131.4, 128.3, 126.8, 120.0, 45.2, 42.1, 28.5 ppm. ESI-MS (M⁺+H): 447.

1-((2-chlorophenyl)sulfonyl)-N-(4-(1-cyclopropyl-1H-tetrazol-5-yl)phenyl)piperidine-4-carboxamide (4o) was obtained as a pink yellow thick oil in the amount of 0.0238g (25% yield). ¹H NMR (400 MHz, DMSO-d₆) δ 10.27 (d, J = 0.2 Hz, 1H), 9.97 (t, J = 0.6 Hz, ), 8.03–8.01 (m, 1H), 7.96 (d, J = 8.8 Hz, 2H), 7.83 (d, J = 8.8 Hz, 2H), 7.75–7.68 (m, 3H), 7.65–7.53 (m, 12H), 4.05–4.00 (m, 1H), 3.78 (dd, J = 12.5, 0.4 Hz, 2H), 2.89–2.83 (m, 2H), 2.57–2.53 (m, 1H), 1.92–1.89 (m, 2H), 1.68–1.58 (m, 3H), 1.25–1.20 (m, 2H), 1.11–1.08 (m, 6H). ¹³C NMR (101 MHz, DMSO-d₆): δ 173.6, 154.9, 142.2, 136.4, 135.0, 133.7, 132.77, 132.71, 132.50, 132.47, 131.99, 131.89, 131.4, 129.9, 129.27, 129.16, 128.3, 126.8, 120.0, 119.5, 118.6, 65.4, 45.2, 42.2, 29.9, 28.5 ppm. ESI-MS (M⁺+H): 489.

N-(4-(1-benzyl-1H-tetrazol-5-yl)phenyl)-1-((2-chlorophenyl)sulfonyl)piperidine-4-carboxamide (4p) was obtained as a white solid in the amount of 0.0238 g (22% yield); mp 155–156 °C. ¹H NMR (400 MHz; DMSO-d₆): δ 10.17 (s, 1H), 8.85 (d, J = 4.8 Hz, 2H), 8.33 (d, J = 8.7 Hz, 2H), 8.02–8.00 (m, 1H), 7.75–7.67 (m, 4H), 7.60–7.56 (m, 1H), 7.38 (t, J = 4.8 Hz, 1H), 3.79–3.76 (m, 2H), 2.87–2.81 (m, 2H), 2.55–2.53 (m, 1H), 1.91–1.87 (m, 2H), 1.67–1.57 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆): δ 173.4, 163.4, 158.0, 142.0, 136.3, 134.9, 132.8, 132.3, 132.0, 131.4, 128.8, 128.3, 119.8, 119.4, 100.0, 45.2, 42.2, 28.5 ppm. ESI-MS (M⁺+H): 537.

1-((2-chlorophenyl)sulfonyl)-N-(4-(1-(3,5-dimethoxypyrazin-2-yl)-1H-tetrazol-5-yl)phenyl)piperidine-4-carboxamide (4q) was obtained in the amount of 0.007 g (8% yield). ¹H NMR (400 MHz; DMSO-d₆): δ 10.23 (s, 1H), 8.02–8.00 (m, 2H), 7.73 (dd, J = 8.0, 1.6 Hz, 1H), 7.71–7.67 (m, 3H), 7.58 (ddd, J = 7.9, 7.0, 1.6 Hz, 1H), 7.51–7.47 (m, 2H), 4.06 (s, 3H), 3.86 (s, 3H), 3.78–3.73 (m, 2H), 2.87–2.80 (m, 2H), 2.54–2.50 (m, 1H)1.88–1.84 (m, 2H), 1.64–1.53 (m, 2H). ¹³C NMR (101 MHz, DMSO-d₆): δ 175.9, 163.5, 156.9, 156.0, 144.8, 138.6, 137.2, 135.0, 134.2, 133.6, 131.5, 130.5, 127.0, 125.3, 122.0, 119.6, 57.6, 57.3, 47.3, 44.3, 30.6 ppm. ESI-MS (M⁺+H): 585.

1-((2-chlorophenyl)sulfonyl)-N-(4-(6,8-dimethoxy-[1,2,4]triazolo[4,3-a]pyrazin-3-yl)phenyl)piperidine-4-carboxamide (4r) was obtained as a white solid in the amount of 0.0483 g (50% yield); mp 236–237 °C. ¹H NMR (400 MHz; DMSO-d₆): δ 10.24 (s, 1H), 9.92 (d, J = 0.3 Hz, ), 8.03–8.00 (m, 1H), 7.91–7.81 (m, 4H), 7.75–7.68 (m, 3H), 7.61–7.56 (m, 1H), 7.54–7.51 (m, 1H), 4.13 (s, 3H), 3.87 (s, 3H), 3.80–3.76 (m, 2H), 2.90–2.83 (m, 2H), 2.59–2.53 (m, 1H), 1.93–1.89 (m, 2H), 1.69–1.59 (m, 2H). ¹³C NMR (101 MHz; DMSO-d₆): δ 173.5, 152.8, 152.1, 148.4, 141.4, 138.9, 136.4, 135.2, 134.9, 132.8, 132.0, 131.4, 129.0, 128.3, 121.0, 119.9, 91.8, 56.0, 55.0, 49.1, 45.2, 42.2, 28.5 ppm. ESI-MS (M⁺+H): 557.

Measurement of Inhibitor Potencies on sEH and FAAH

The reported IC₅₀ values were obtained using previously published FAAH and sEH fluorescent-based assays.^{40, 41} The substrate for FAAH inhibition assay N-(6-methoxypyridin-3-yl) octanamide (OMP), was synthesized following a previously reported synthetic procedure and reaction conditions.²⁷ The substrate cyano(6-methoxynaphthalen-2-yl)methyl((3-phenyloxiran-2-yl)methyl)carbonate (CMNPC) for sEH inhibition assay was synthesized at UC Davis.⁴¹

FAAH Inhibition Assay:

Measurement of human FAAH inhibition potency was done using the substrate OMP ([S]_final = 50 μM) in sodium phosphate buffer [0.1 M, pH = 8.0, and 0.1 mg/mL bovine serum albumin (BSA)]. Progress of the reaction was measured by fluorescence detection of 6-methoxypyridin-3-amine at an excitation wavelength of 303 nm and an emission wavelength of 394 nm at 37 °C using a microplate reader M2 (Molecular Devices., CA, USA).

sEH Inhibition Assay:

The substrate CMNPC ([S]_final = 5 μM) was added to wells containing sEH (~2nM mouse sEH, ~1nM rat sEH, or ~1 nM human sEH) in sodium phosphate buffer (0.1 M, pH = 7.4 and 0.1 mg/mL BSA, and formation of the fluorescent 6-methoxynaphthaldehyde (λ_excitation = 330 nm, λ_emission = 465 nm, 37 °C) was measured using of a microplate reader M2 (Molecular Devices., CA, USA).

The fluorescence generated by hydrolysis was quantified every 30 s for 10 min, and the linear portion of the curve was used to generate the reaction velocity (v_inhibitor). Values were subtracted from wells containing no enzyme. Next, the IC₅₀ values were quantified by simple linear regression of the log [I] vs. % remaining activity (v_inhibitor/v_DMSO) and determining x when y = 0.50 with at least 2 points on either side of 50% activity mark. All measurements were the average of triplicates. For all assays, the final DMSO concentration was 2%.

Microsomal stability assay in liver microsomes

The test compound was dissolved in DMSO (1 mM) and 1 μL was preincubated at 37 °C for 5 minutes in 924 μL of phosphate buffer (0.1 M, pH = 7.4) and 25 μL of microsome mix from the liver. After 5 min, the reaction was started by the addition of 50 μL of NADPH regenerating system (NADPH regenerating system: dissolve 121 mg of glucose-6-phosphate and 59 mg of NADPH in 2.2 mL of 100 mM PBS buffer, followed by 800 μL (100 Un/1 mL) glucose 6-phosphate dehydrogenase from baker’s yeast and 1 mL of 400 mM MgCl₂). The reaction was stopped by adding 450 μL of methanol at 0, 10, 20, 30 and 60 min. The samples were centrifuged at 13,500 × g for 5 min at 4 °C, filtered and analyzed in the SIM mode on LC-MS using the method described in Material and methods section. The peak area response ratio to internal standard was collected and the in vitro intrinsic clearance was calculated (see Supplemental info), wherein k represents the -gradient of the ln peak area ratio plotted against time.⁴²

Solubility assay

This assay is based on turbidity (light scattering) principle: a 10 mM solution of each drug is prepared in DMSO, and 50 μL is mixed with sodium phosphate buffer (pH 7.4) through serial dilutions. Aliquots are then transferred to a 96-well plate, pre-filled with the same buffer, and absorbance is measured at 800 nm using a spectrophotometer. Blank values are subtracted, and solubility is assessed through linear extrapolation of the lowest concentrations with measurable absorbance.⁴³

Molecular modeling and ADMET predicitions

ICM Pro software (Molsoft LLC) was used for docking experiments and ADMET predictions. Human FAAH homology model was prepared following previously published procedure.²⁷ All inhibitors 4a-4r, and URB597 were docked into a homology model following the program guidelines and using 50 thoroughness. Human sEH model was prepared from PDB: 4HAI²⁸ as follows: the PDB file was first converted to an ICM object and optimized according to the default program settings. The inhibitor, N-cycloheptyl-1-(mesitylsulfonyl) piperidine-4-carboxamide, was manually removed, and binding pockets were identified, followed by docking experiments.

In vivo evaluation of 4m

Subjects

Data were collected from male Sprague-Dawley rats purchased from Charles River (Hollister, CA, USA). All rats were at least 50 days old at the start of the study and randomly assigned to treatment groups. Rats were housed in standard Plexiglas cages containing a 33 cm metal running wheel (Starr Life Sciences) on a 12/12-hour light/dark cycle (lights off at 1100 h). Food and water were available ad libitum. All procedures were approved by the California State University, East Bay Institutional Animal Care and Use Committee.

Drugs

4m was dissolved in vehicle (20% DMSO, 20% cremophor, 60% saline). Morphine was dissolved in saline. DMSO, cremophor, and morphine was purchased from Sigma-Aldrich (St. Louis, Missouri). Drugs were administered via intraperitoneal injection in the 10 min prior to the beginning of the dark phase (1100 h) to capture the effects of the drug on wheel running behavior. Rats were assigned to drug groups in a randomized order.

Wheel running

Rats were allowed to habituate to the running wheel cages for 3 days before data collection. All wheel revolutions were logged by a computer containing VitalView^® Activity software to count the number of wheel revolutions. The number of wheel revolutions were collected in 5 min bins for 15 min following drug injection. No experimenters were in the room as wheel running data were collected.

Data analysis

All data are expressed as mean ± SEM except where stated. The number of revolutions in the 15 min period after drug injected were averaged and analyzed using a one-way analysis of variance (ANOVA) on GraphPad Prism. Statistical significance was defined as a probability of < 0.05.