Article pubs.acs.org/jmc
Discovery of (R)‑1-(7-Chloro-2,2-bis(fluoromethyl)chroman-4-yl)-3-(3methylisoquinolin-5-yl)urea (A-1165442): A Temperature-Neutral Transient Receptor Potential Vanilloid‑1 (TRPV1) Antagonist with Analgesic Efficacy Eric A. Voight,* Arthur R. Gomtsyan,* Jerome F. Daanen, Richard J. Perner, Robert G. Schmidt, Erol K. Bayburt, Stanley DiDomenico, Heath A. McDonald, Pamela S. Puttfarcken, Jun Chen, Torben R. Neelands, Bruce R. Bianchi, Ping Han, Regina M. Reilly, Pamela H. Franklin, Jason A. Segreti, Richard A. Nelson, Zhi Su, Andrew J. King, James S. Polakowski, Scott J. Baker, Donna M. Gauvin, LaGeisha R. Lewis, Joseph P. Mikusa, Shailen K. Joshi, Connie R. Faltynek, Philip R. Kym, and Michael E. Kort Research & Development, AbbVie Inc., 1 North Waukegan Road, North Chicago, Illinois 60064, United States ABSTRACT: The synthesis and characterization of a series of selective, orally bioavailable 1-(chroman-4-yl)urea TRPV1 antagonists is described. Whereas first-generation antagonists that inhibit all modes of TRPV1 activation can elicit hyperthermia, the compounds disclosed herein do not elevate core body temperature in preclinical models and only partially block acid activation of TRPV1. Advancing the SAR of this series led to the eventual identification of (R)-1-(7-chloro-2,2bis(fluoromethyl)chroman-4-yl)-3-(3-methylisoquinolin-5-yl)urea (A-1165442, 52), an analogue that possesses excellent pharmacological selectivity, has a favorable pharmacokinetic profile, and demonstrates good efficacy against osteoarthritis pain in rodents.
■
experimental model of UV burn pain in humans.9a On the other hand, extensive characterization of selective TRPV1 ligands also has established an unequivocal role for the channel in thermoregulation. For example, the TRPV1 agonists, capsaicin and resiniferatoxin,12 have been shown to cause dose-dependent decreases in core body temperature in animals; conversely, TRPV1 antagonists such as ABT-102 (1, Figure 1) induce hyperthermia in preclinical models.13,14 Multiple TRPV1 antagonists have been evaluated in clinical trials.11a,15,16 Evidence accrued to date with AMG517 (Amgen),17 MK-2295 (Merck),18 AZD1386 (AstraZeneca),19 and our own compound 120 indicates that the hyperthermic effect of TRPV1 antagonists observed preclinically is recapitulated in humans. Moreover, TRPV1 knockout mice resist antagonist-induced hyperthermia, further implicating TRPV1 in thermoregulation.21 Therefore, pharmacological separation of analgesic and hyperthermic effects became the key challenge in developing TRPV1 antagonists as viable agents for pain management. It bears emphasizing, however, that most of the “first-generation” antagonists block, in a dose-dependent manner, the direct activation of TRPV1 by all modalities (capsaicin, endogenous lipids, acidic pH, heat).22
INTRODUCTION Transient receptor potential vanilloid-1 (TRPV1) is the most thoroughly characterized member of the TRP channel superfamily, a group of 27 cation channels with a wide variety of important biological roles.1,2 TRPV1 is localized on smalland medium-diameter neurons (C- and Aδ-fibers) in the peripheral nervous system with high levels of expression in sensory ganglia and also is expressed in areas of the central nervous system.3 As a polymodal receptor, TRPV1 is activated by “endovanilloid” endogenous lipids (e.g., anandamide, Narachidonoyl-dopamine),4 acidic pH (42 °C), and exogenous ligands (e.g., capsaicin from chili peppers, peptide toxins from tarantula venom).5 This activation allows calcium and sodium ions to flow into the cell, which eventually leads to the release of molecules known to be involved in pain transmission (e.g., CGRP, substance P, glutamate, bradykinin).6 Attenuated thermal hyperalgesic response upon genetic ablation of TRPV1 in mice provides further support for the role of the receptor as a nexus in pain transmission and suggests that efficacy in pain management could be achieved with TRPV1 antagonists.7 The identification of a number of potent and selective small molecule TRPV1 antagonists over the past decade has confirmed that pharmacological blockade of this receptor provides analgesic efficacy in preclinical models of inflammatory, osteoarthritis, and neuropathic pain,8−11 as well as in © XXXX American Chemical Society
Received: June 16, 2014
A
dx.doi.org/10.1021/jm500916t | J. Med. Chem. XXXX, XXX, XXX−XXX
Journal of Medicinal Chemistry
Article
Scheme 1a
Reagents and conditions: (a) R2CO, pyrrolidine, MeOH, 60 °C; (b) (R)-diphenyl(pyrrolidin-2-yl)methanol (5 mol %), borane−N,Ndiethylaniline complex, MTBE, 45 °C. For R2 = H, Me; (c) Ms2O, i-Pr2NEt, THF, −20 °C, n-Bu4NN3; (d) PPh3, aq THF, 60 °C; (e) (R)-(−)-mandelic acid, IPA, hexanes. For R2 = Et, n-Pr, −CH2F; (f) diphenylphosphoryl azide, DBU, THF, 0−20 °C; (g) H2, Pd−C, MeOH, 50 °C; (h) D-(−)-tartaric acid, IPA, hexanes. a
Figure 1. Temperature effects (in telemeterized rats) of TRPV1 antagonists.
Concern over the mechanism-based thermoregulatory effects associated with TRPV1 antagonists led us to develop a rat telemetry model. This model enabled the evaluation of effects of representative molecules on core body temperature in animals surgically instrumented (intraperitoneal cavity) with an internal temperature probe. Oral administration (100 μmol/kg) of 1, a potent antagonist (IC50 = 4 nM) of capsaicin activation of TRPV1 in vitro, elicited a clear elevation (0.8 °C) in core body temperature (1 h postdose) (Figure 1).23,24 An identical dose of 1 provided full efficacy in rodent models of inflammatory pain at the same time point. Tetrahydropyridine 2 and oxazole 3 also produced marked hyperthermia (1.0 and 1.4 °C, respectively) under the same conditions; the temperature increase produced by 3 persisted for 7−8 h.25,26 With the goal of identifying a TRPV1 antagonist which had robust analgesic efficacy but was devoid of dose-limiting temperature effects, we embarked on an in vivo screening campaign to assess the thermoregulatory profiles of our library of orally active TRPV1 antagonists. Among over 30 structurally diverse representatives evaluated, only the isoquinoline-derived chromanyl urea 4 with (R) configuration (TRPV1 IC50 = 2 nM) was found to be devoid of hyperthermic effects at high dose.27 These findings prompted a more focused investigation of SAR. We report herein the optimization of this chromanyl urea series, culminating in the discovery of (R)-1-(7-chloro-2,2bis(fluoromethyl)chroman-4-yl)-3-(3-methylisoquinolin-5-yl)urea (A-1165442, 52), a modality-differentiated secondgeneration TRPV1 antagonist with good analgesic efficacy and a temperature-neutral profile.28
CH2F), an alternative approach that utilized DPPA-mediated inversion, followed by azide hydrogenation, and tartaric acid salt isolation improved overall yields of amine salts 9. In either case, (R)-mandelic or D-tartaric acid chroman-4-amine salts could be isolated without intermediate purification or chromatography, making the routes rapid and scalable. To probe the effects of isoquinoline substitution on TRPV1 antagonism, several routes to trisubstituted isoquinoline-5amines were developed (Scheme 2).31,32 Nitration of 3methylisoquinoline (10) gave 11, which was subjected to methyllithium addition to give a mixture of dimethylisoquinolines 12 and 13.33 Alternatively, a vicarious nucleophilic substitution//decarboxylation sequence provided 12 selectively.33 For the preparation of 7-fluoro-substituted isoquinoline 20, the 8-position was blocked with a chloride to allow nitration at C-5 (18 → 19), while isoquinolines 18, 23, and 24 were all prepared via an oxidative Schlittler−Müller modification of the Pomeranz−Fritsch isoquinoline synthesis from benzylamines 17, 21, and 22.34 Finally, 7-trifluoromethylsubstituted isoquinoline 32 could be accessed via a propyne Sonagashira/ammonia cyclization sequence starting from benzaldehyde 29.35 All nitroisoquinolines were reduced to isoquinoline-5-amines via hydrogenation over Pd/C or Ra−Ni. To complete the synthesis of chromanyl urea compounds, two coupling procedures were employed (Scheme 3). For indazoles 33−34, hydroxytetralins (35−36), and unsubstituted isoquinolines (4, 37−40), a variety of amino arenes were coupled with chromanyl amine salts 9 using N,N-disuccinimidyl carbonate (DSC) to provide the desired urea products.36 For functionalized isoquinolines 41−60, urea formation via an intermediate phenyl carbamate or trichloroacetamide37 was required for efficient coupling (see Experimental Section for details).
■
CHEMISTRY To generate the significant quantities of chromanyl urea analogues required to develop an empirical understanding of the effects of TRPV1 antagonists on temperature in vivo, a general and scalable asymmetric synthesis of the requisite chiral chroman-4-amine precursors was developed (Scheme 1).29 Starting from readily available hydroxyacetophenones 5 or chroman-4-ones 6, a modified catalytic asymmetric CBS reduction30 provided (S)-chroman-4-ols 7 in excellent yield and >95% ee. For less lipophilic compounds (R2 = H, Me), a sequence involving mesylate displacement by azide to give 8, Staudinger reduction, and mandelic acid salt formation provided 9. For more lipophilic chromans (R2 = Et, n-Pr,
■
RESULTS AND DISCUSSION Having developed a practical synthetic route to the desired enantiomerically pure chroman building blocks, we initiated concurrent exploration of the SAR of the chroman moiety and of the pendant aryl group (Table 1). A conservative initial modification, the regioisomeric fluorinated analogue 37, maintained both the TRPV1 potency (IC50 = 4 nM, Ca2+ flux assay) against capsaicin (cap) as well as the hyperthermiaB
dx.doi.org/10.1021/jm500916t | J. Med. Chem. XXXX, XXX, XXX−XXX
Journal of Medicinal Chemistry
Article
Scheme 2a
a
Reagents and conditions: (a) KNO3, H2SO4 (86%); (b) potassium tert-butoxide, ClCH2CO2Et, THF, conc HCl (78% for 12); (c) MeLi, THF, Br2, Et3N (51%, 2.5:1 mix of 12:13); (d) H2, Ra−Ni, THF (49% for 14, 25% for 15); (e) pyruvic aldehyde dimethyl acetal, NaHB(OAc)3, THF, ClSO3H (40%); (f) KNO3, H2SO4 (97%); (g) H2, 5% Pd−C, Et3N, THF (38%); (h) pyruvic aldehyde dimethyl acetal, NaHB(OAc)3, THF, ClSO3H (82% for 23, 72% for 24); (i) KNO3, H2SO4 (23 → 25, 97%); (j) NO2BF4, sulfolane (24 → 26, 79%); (k) H2, Ra−Ni, THF (72% for 27, 85% for 28; (l) propyne, CuI, Cl2Pd(PPh3)2, Et3N, DMF; (m) NH3, MeOH; (n) KNO3, H2SO4 (64% over three steps); (o) H2, Ra−Ni, THF (61%).
Scheme 3
completely blocked activation of TRPV1 by capsaicin and Narachidonoyl-dopamine (NADA, data not shown)4 but only partially inhibit activation of the channel by protons (pH 5.0). The observed difference in the thermoregulatory profiles of two 7-substituted chroman analogues closely related to 37, isoquinolines 38 (R1 = 7-H; ΔT −0.4 °C) and 39 (R1 = 7OCHF3; ΔT +1.0 °C), is consistent with their divergent acid activation pharmacology. This stark contrast of the isoquinoline
sparing profile of 4. This 7-fluorochroman scaffold was then coupled to a selection of heteroaryl (Ar, Scheme 2) fragments, which we have previously demonstrated to impart TRPV1 activity, to provide ureas 33−36.37 However, although 33−36 maintained potent TRPV1 antagonist potency against capsaicin activation, they produced a robust increase (≥1.0 °C) in core body temperature. Close scrutiny of our pharmacology data revealed that while the temperature-neutral analogues 4 and 37 C
dx.doi.org/10.1021/jm500916t | J. Med. Chem. XXXX, XXX, XXX−XXX
Journal of Medicinal Chemistry
Article
Table 1. Effect of Arene Substitution on TRPV1 Antagonism by 1-(2,2-Dimethylchroman-4-yl) Ureas
compd 1 2 3 4 37 33 34 35 36 38 39 a
hTRPV1, capsaicin, IC50 (nM)a
R1
Ar
isoquinoline isoquinoline N-H indazole N-Me indazole (R)-tetralin (S)-tetralin isoquinoline isoquinoline
4 21 12 2 4 16 7 7 7 4 6
6-F 7-F 7-F 7-F 7-F 7-F H 7-OCHF3
± ± ± ± ± ± ± ± ± ± ±
1 3 5 1 2 2 4 1 2 1 2
hTRPV1, pH 5.0, % inhib @ 30 μM
ΔT (°C)
± ± ± ± ± ± ± ± ± ± ±
+0.8 +1.0 +1.4 0.0 −0.5 +1.2 +1.3 +1.1 +1.0 −0.4 +1.0
98 98 99 70 61 95 98 98 99 60 96
1 1 1 4 5 1 2 1 1 3 2
Assayed according to ref 28.
Table 2. Effect of Substitution of 1-(Chroman-4-yl)-3-(isoquinolin-5-yl)ureas on TRPV1 Antagonism, Cyp Inhibition, and Metabolism
compd
R1
R2
R3
37 39 4 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54
7-F 7-OCHF3 6-F 6-F 6-F 8-F 8-F 7-F 7-CF3 7,8-F 6,8-F 8-F-7-CF3 7-OCHF3 6-F 7-F 7-Cl 7-(c-Pr) 8-F
Me Me Me Et Et Et n-Pr Me Me Me Me Me Me CH2F CH2F CH2F CH2F CH2F
H H H H Me Me Me Me Me Me Me Me Me Me Me Me Me Me
a
hTRPV1, capsaicin, IC50 (nM)a 4 6 2 6 40 34 150 3 10 5 32 5 32 17 5 9 5 22
± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
hTRPV1, pH 5.0, %inhib @ 30 μM
Cyp3A4/2D6 inhib IC50 (μM)b
HLM Clint (μL/min/mg)
ΔT (°C)a
± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
0.66/20/>20 >20/>20 >20/17 >20/>20 >20/>20 >20/>20 18/>20 >20/6.0 15/20 >20/19 20/4.9 >20/>20 >20/11
210 140 180 390 480 430 500 140 130 100 100 120 110 70 82 98 240 61
−0.4 +1.0 0.0 −0.2 +0.2 −0.5 +0.1 0.0 0.0 0.0 0.0 +0.8 +1.7 0.0 0.0 0.0 +0.8 +0.2
2 2 1 3 14 2 14 1 2 1 3 1 5 1 1 1 1 5
61 96 70 66 70 59 13 70 66 58 69 86 95 86 72 62 91 60
5 2 4 3 2 4 2 3 4 2 1 5 2 4 4 3 1 2
Assayed according to ref 28. bSee Experimental Section for assay details.
acid activation of TRPV1 may predict effects on core body temperature.38,39 Armed with the insight that the recombinant TRPV1 in vitro assays could be used as surrogates for testing in telemeterized rats, we returned our attention to optimization of the isoquinoline and chroman fragments. As illustrated in Table 2, inhibition (e.g., IC50 10 μM), 55−60 produced robust hyperthermia when dosed orally in rats. From our medicinal chemistry campaign, 7-chloro chroman 52 thus emerged as a preferred lead, displaying potent, competitive antagonism at recombinant human TRPV1 activated by capsaicin (IC50 = 9 nM) and incomplete blockade of acid-evoked response (62% block at 30 μM). This pharmacology was recapitulated in recombinant rat TRPV1.28 The compound exhibited comparable differentiated effects on capsaicin and pH 5.0 activation of both recombinant rat TRPV1(specific data)28 and native TRPV1 using neurons isolated from rat dorsal root ganglia in whole cell patch-clamp electrophysiology studies (specific data). Chroman 52 possessed excellent pharmacological selectivity (>100-fold) versus other members of the TRP family (TRPA1, TRPM8, TRPV2, TRPV3) and other receptors expressed in peripheral sensory neurons including P2X2/3, Cav2.2, Nav channels, and KCNQ2/3 (data not shown). The compound showed minimal cross-reactivity upon evaluation (10 μM) in a broad screening panel (n = 74, CEREP) of cell-surface receptors, ion channels, and enzymes. The pharmacokinetic parameters of 52 administered in 10% DMSO/PEG-400 in Sprague−Dawley rats are summarized in Table 4. Modest systemic clearance (Clp
nitrogen with the heme iron of the Cyp. Introduction of steric bulk adjacent to nitrogen (R3) in the form of a methyl group (41−54) overcame this liability, reducing Cyp inhibition to IC50 values to greater than 20 μM while retaining potency at TRPV1. Surprisingly, substitution at the other carbon atom flanking the isoquinoline nitrogen dramatically eroded (IC50 >1 μM) TRPV1 antagonist potency. Extensive metabolism by human microsomes, which contrasted with favorable rodent microsome stability, was another liability of the series. Modifications at R1 or R2 of the chroman nucleus which increased lipophilicity had a detrimental effect on microsomal turnover. This trend was particularly striking with R2 substitution, where the progression from methyl (4) to ethyl (40−42) to n-propyl (43) drove human liver microsome intrinsic clearance (HLM Clint) to as high as 500 μL/min/mg for 43. Although halogen, haloalkyl, or haloalkoxy R1 substituents mitigated the in vitro metabolic stability problem within the gem-dimethyl (R2 = Me) set of methyl isoquinoline analogues (44−49), fluoromethyl substitution at the 2-position of the chroman (R2 = CH2F, 50−54) further lowered clearance to desirable (
