Discovery of ONO-TR-772 (VU6018042): A Highly Selective and CNS Penetrant TREK Inhibitor in Vivo Tool Compound

Abstract

Herein we describe our continuing work on the K₂P family of potassium ion channels with the chemical optimization of a selective and CNS penetrant series of TREK inhibitors, culminating in the discovery of ONO-TR-772 (VU6018042). From an HTS hit harboring a benzyl ether linker, SAR proved intractable until an acetylene linker was identified as an isosteric replacement. Robust SAR was then observed, and a key fluorination to enhance PK and CNS penetration provided ONO-TR-772 (VU6018042), a potent (TREK-1 IC₅₀ = 15 nM), selective (>10 μM versus other K₂P channels except TREK-2), and CNS penetrant (rat K_p = 0.98) TREK inhibitor. ONO-TR-772 (VU6018042) demonstrated robust efficacy in an MK-801 challenge NOR paradigm, with an MED of 10 mg/kg.

Potassium (K⁺) channels represent a major thrust of drug development activities targeting both shaker type voltage-gated (K_v) and inward rectifier (K_ir) channels.¹⁻⁶ Far less explored is the third family, the two pore domain (K₂P) family of potassium channels, consisting of 15 K₂P subtypes, within six distinct subfamilies: tandem of P domains in a weakly inward rectifying potassium channel (TWIK), TWIK related K⁺ channels (TREK), TASK (TWIK related acid-sensitive K⁺ channels), TWIK related ALkaline pH-activated K⁺ channels (TALK), tandem pore domain halothane inhibited K⁺ channels (THIK), and TWIK related spinal cord K⁺ channel (TRESK).¹⁻⁶ Our first priority was to develop potent and selective in vivo tool compounds to modulate TREK-1. TREK-1 is highly expressed in the central and peripheral nervous systems, and modulation of these channels are proposed to play key roles in multiple CNS and peripheral disorders; unfortunately, a deeper understanding of the therapeutic potential of TREK-1 has been hampered by a lack of small molecule tools.¹⁻¹²

Recently (Figure 1), we disclosed ONO-2920632 (1, VU6011887),¹³ a highly selective and CNS penetrant TREK-2 preferring activator (structurally distinct from BL-1249 (2)^14,15), with potential for nonopiate pain management. Attention now turned to the development of TREK inhibitors for the potential treatment of cognitive deficits, as numerous studies implicate TREK-1.⁷⁻¹² For example, inhibition of TREK-1 protects mice from cognitive impairment induced by anesthesia,⁷ and TREK-1 gene expression is increased in the hippocampus of schizophrenic patients relative to normal controls.¹⁶ Likewise, inhibition of TREK-2 is required for the neurotensin-mediated facilitation of spatial learning, and TREK-2 expression is increased in the cortex and hippocampus of several CNS pathologies.^17,18 Taken together, these data argue for the need to pharmacologically validate the role of TREK-1 and TREK-2 inhibition in preclinical rodent cognition models. Yet, the tools to do so are limited to peptides such as spadin,¹⁹ weak (IC₅₀s 2 to >10 μM) off-target small molecule inhibitors such as fluoxetine,²⁰ atypical antipsychotics,²¹ antihypertensives,²² and antiarrhythmic drugs.²³ A state-dependent TREK-1 inhibitor, TKIM (3), has been reported with an IC₅₀ of 2.96 μM²⁴ as well as more recently the CNS-penetrant isobenzofuran-1(3H)-one derivative, Cpd8 (4) (TREK-1 IC₅₀ of 810 nM), which displayed selectivity versus TREK-2 and neuroprotective effects in vitro and in vivo.²⁵ In this letter, we describe our efforts toward the development of a potent, selective, and CNS penetrant TREK inhibitor derived from a weak HTS hit.

Figure 1.

Structures of reported TREK-1 and TREK-2 activators 1 and 2, and the more recently developed TREK-1 inhibitors 3 and 4.

A high throughput screen employing a functional TREK-1 thallium (Tl⁺) flux assay identified (Figure 2) ONO-1530283 (5) as a weak TREK-1 inhibitor (IC₅₀ = 7.7 μM). A hit expansion exercise increased TREK-1 potency ∼17-fold (IC₅₀ = 0.46 μM) to afford ONO-0606822 (6), the lead for the program despite superhepatic rodent clearance and high clogP (5.2). Moreover, we were concerned about the latent quinone moiety in 6; thus, the initial aim of the optimization was to remove the latent quinone by the addition of heteroatoms to the phenyl core while also reducing lipophilicity.

Figure 2.

Structures of the HTS hit 5, and a rapid hit expansion exercise improved potency ∼17-fold to afford submicromolar TREK-1 lead 6.

Scheme 1 highlights the routes to access key heteroaryl congeners 10 of TREK inhibitor lead 6. For both pyridine isomers and the pyrazine analogue, commercial starting materials 7–9 underwent either alkylation reactions or a Mitsunobu reaction to install the benzyl ether moiety, followed by nitro reduction, in the case of 8, and a polyphosphonic anhydride (T₃P) coupling with 5-(tert-butoxycarbonylamino)-2-chlorobenzoic acid to provide amide analogues 10. TREK-1 inhibitory data is shown in Table 1; while the SAR was without texture, clogPs did decrease by almost a log (∼4.4), but both in vitro and in vivo clearance in rat remained suprahepatic (CL_p > 100 mL/min/kg). The N-Boc moiety proved to be, surprisingly, not responsible for the poor PK. However, these data prompted the team to modify the core for subsequent optimization to the pyridine congener 10a.

Scheme 1. Synthesis of Heteroaryl Analogues 10.

Reagents and conditions: (a) BnBr, Cs₂CO₃, DMF, rt, 12%; (b) 5-(tert-butoxycarbonylamino)-2-chlorobenzoic acid, DIEA, DMF, HBTU, rt to 60 °C, 18%; (c) BnOH, DIAD, PPh₃, THF, rt, 24%; (d) Fe, NH₄Cl, THF, H₂O, 75 °C, 87%; (e) 5-(tert-butoxycarbonylamino)-2-chlorobenzoic acid, DIEA, DCM, T₃P, rt, 42%; (f) BnOH, NaH, 0–120 °C, 34%; (g) 5-(tert-butoxycarbonylamino)-2-chlorobenzoic acid, DIEA, DCM, T₃P, rt, 26%.

Table 1. TREK-1 Inhibitory Activity and Lipophilicity of Analogues 10.

From previous work in our laboratories,²⁶ we found that an acetylene was a productive bioisosteric replacement for a benzyl ether moiety, and we applied this tactic to both 10a and 10c (Figure 3) modifying Scheme 1 to replace the alkylation step with a Sonogashira coupling (see Supporting Information for full details) to provide 11 and 12. In the case of 10a, the acetylenic variant 11 displayed an ∼ 4-fold increase in TREK-1 potency (IC₅₀ = 0.14 μM) and lowered human microsomal intrinsic clearance 10-fold (h CL_INT = 13 mL/min/kg). Interestingly, TREK-1 potency diminished in the pyrazine matched pairs 10c (IC₅₀ = 0.56 μM) and 12 (IC₅₀ = 1.4 μM), but in vivo rat clearance improved dramatically with installation of the acetylene (10c: suprahepatic CL_p= 135 mL/min/kg; 12: CL_p = 25 mL/min/kg). Therefore, 11 became the new lead, and we elected to explore 2-position substituents on the pyridine core in analogues 13 (Table 2) as well as substituents on the distal phenyl ring pendant to the acetylene, as in derivatives 14 (Table 3).

Figure 3.

Impact of replacement of the benzyl ether with an acetylene moiety on both TREK-1 potency and in vitro and in vivo clearance.

Table 2. TREK-1 Inhibitory Activity and in Vitro DMPK Profiles of Analogues 13.

compd	R	h TREK-1 Tl+ IC50 (μM)	LMS CLINT human/mouse (mL/min/kg)	PPB human/mouse (%)
11	Me	0.14	13/34	99.5/98.7
13a	F	0.11	10/6	99.8/98.9
13b	Cl	0.43	NDa	ND/>99.9
13c	OMe	0.91	ND	ND

^aND = not determined.

Table 3. TREK-1 Inhibitory Activity and in Vitro DMPK Profiles of Analogues 13^a.

^aND = not determined.

In the 3-positon on the pyridine core (Table 2), a fluoro congener 13a was equipotent to the parent methyl 6 and displayed similar or better microsomal stability and a little higher protein binding profiles. TREK-1 potency eroded as either a chlorine atom (13b, IC₅₀ = 0.43 μM) was introduced or an electron-donating methoxy group (13c, IC₅₀ = 0.91 μM). While the introduction of substituents to the distal phenyl ring of the acetylene moiety had no impact on TREK-1 potency (Table 3), a 4-F phenyl congener 14c possessed improved microsomal stability (human CL_INT = 7.5 mL/min/kg, mouse CL_INT = 13 mL/min/kg). Thus, all efforts shifted toward a more detailed evaluation of 14c (ONO-TR-772).

Up to this point in the program, SAR was guided by a thallium flux TREK-1 assay, it was now time to evaluate broader K₂P and ion channel activity for 14c. Gratifyingly, 14c proved to be >67-fold selective over other K₂P channels (TASK-3, TRESK, TWIK-2, TRAAK, TASK-1, TASK-2). Only TREK-2 was shown to be equipotent with TREK-1 in the profiling electrophysiological assays. In human manual patch clamp (MPC), 14c was a potent TREK-1 inhibitor (IC₅₀ = 15 nM), and therefore, more accurately described as a dual TREK-1/2 inhibitor. As mouse would be the species used in our cognition PD assay, we evaluated 14c in mouse MPC, and it was a potent TREK-1 inhibitor (IC₅₀ = 67 nM). Prior to performing cognition studies, we needed to assess broader ancillary pharmacology to ensure we would be evaluating selective TREK-1/2 inhibition. When gauged against a cardiovascular (CV) ion channel panel, 14c was similarly highly selective (<50% at 10 μM versus: Ca_v1.2, HNC4, K_ir2.1, K_v1.5, K_v4.3, KCNQ1, Na_v1.5 and hERG). The only activity >50% at 10 μM was Kir2.2 (52%). The clean ion channel pharmacology profile prompted the team to collect a broader ancillary pharmacology profile across GPCRs, ion channels and transporters in the Eurofins lead profiling screen. Here, 14c showed >50% inhibition at 10 μM for only four of 68 targets: A₃ (84%, K_i = 0.48 μM), Calcium L-type/benzothiazepine (56%, K_i = 2.6 μM), Calcium L-type/dihydropyridine (76%, K_i = 0.54 μM) and sodium channel, site 2 (84%, K_i = 0.65 μM). Weak activity at these targets would not obscure our proof-of-concept studies in rodent cognition models, so 14c advanced into DMPK profiling.

In liver microsomes, CL_INT for 14c was low across all species (h, r, m: 7.5, 29, and 13 mL/min/kg); however, the lipophilic nature of 14c (clogP = 5.53) led to high plasma protein binding (h, r, m: 99.4%, 99.9% and 98.4%), brain homogenate binding (r, m: >99.9% and 99.4%) and low aqueous solubility (<5 μM). In a rat IV:PBL PK cassette study, 14c demonstrated a good IVIVC with low clearance (CL_p = 21.2 mL/min/kg), high volume of distribution (6.7 L/kg) long half-life (t_1/2 = 6.78 h) and a K_p of 0.98. In a 10 mg/kg PO mouse PBL study, 14c displayed a K_p of 0.44, highlighting excellent CNS exposure in both mouse and rat; however, due to the high brain homogenate binding (driven by the lipophilic nature of 14c), a K_p,uu could not be accurately ascertained. From the 10 mg/kg PO CD-1 male mouse study with 14c, C_max total brain concentrations achieved 252 nM (T_max = 4 h, AUC 2269 h*nM), which we felt was too low to proceed with a novel object recognition (NOR) cognition assay. Switching to an IP route of administration (10 mg/kg) increased total brain exposure (C_max of 888 nM, 5.3 nM unbound based on mouse brain homogenate binding, T_max = 4 h, AUC 12,261 h*nM). Based on the lipophilicity of 14c, we assumed the unbound brain level was a low estimate, so we elected to proceed with the NOR study at doses of 3, 10, and 30 mg/kg IP. Since the brain T_max was at 4 h, 14c or the clozapine positive control was dosed IP to male CD-1 mice 3.5 h prior to the MK-801 (0.2 mg/kg IP) challenge, and then the discrimination index was assessed (Figure 4). As expected, MK-801 induced a significant cognitive deficit,²⁷ which was dose dependently reversed by 14c, with a minimum effective dose (MED) at 10 mg/kg and comparative efficacy at 30 mg/kg to the internal control clozapine (1 mg/kg IP). Thus, selective inhibition of TREK-1/2 was shown, for the first time, to be efficacious in the NOR paradigm and suggests therapeutic relevance for TREK-1 and or TREK-1/2 inhibition to treat cognitive disorders.

Figure 4.

Effect of 14c (ONO-TR-772/VU6018042) in the MK-801 challenged novel object recognition task. 14c dose-dependently enhanced recognition memory in male CD-1 mice after challenged with MK-801. Pretreatment with 3, 10, and 30 mg/kg 14c IP 3.5 h prior to MK-801 treatment and exposure to identical objects significantly enhanced recognition memory assessed 90 min later. Minimum effective dose (MED) is 10 mg/kg. N = 15/group of male CD-1 mice. # p < 0.05, ###, p < 0.001 vs normal (t test), **p < 0.01; ***p < 0.001 vs vehicle (Dunnett posthoc test). †††, p < 0.001 vehicle group (t test).

In parallel with the in vivo POC work, efforts were being made to find alternatives for the privileged N-Boc moiety, which contributed to the high lipophilicity and structural concern. Only one viable replacement emerged, a cyclopropyl amide (Figure 5), as shown in 15a (TREK-1 IC₅₀ = 0.18 μM) and 15b (TREK-1 IC₅₀ = 0.29 μM). In these matched pairs, the impact of the 4-F phenyl moiety was profound on both in vitro and in vivo clearance (10-fold reduction in in vivo rat CL_p); however, both analogues still possessed high plasma protein binding (>99.7%) and no improvement on aqueous solubility. Further optimization on this series is required to improve fraction unbound, solubility and reduce lipophilicity. Therefore, 14c ONO-TR-772 (VU60108042) stands as a valuable rodent in vivo tool compound to explore selective inhibition of TREK-1 and TREK-2.

Figure 5.

Cyclopropyl amides are viable replacements for the N-Boc moiety. Introduction of the 4-F phenyl moiety reduces in vivo rat clearance almost 10-fold.

In summary, we have developed a potent, selective, and CNS penetrant TREK inhibitor 14c (ONO-TR-772, VU6018042) that provided proof of concept for pro-cognitive efficacy in an MK-801 challenge NOR paradigm with an MED of 10 mg/kg IP. While a valuable in vivo probe for the community, further efforts are ongoing and will be reported in due course.

Glossary

Abbreviations

TREK

TWIK RElated K⁺ channels

MED

minimum effective dose

PK

pharmacokinetics

PBL

plasma:brain level study

NOR

novel object recognition

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.5c00215.

• Additional experimental details, methods for the synthesis and characterization of all compounds (¹H NMR, ¹³C NMR, 2-D NMR, HRMS), in vitro and in vivo DMPK protocols and supplemental figures (PDF)

Author Contributions

C.W.L., H.K., J.S.D., D.W.E., J.W. and T.M.B. oversaw the medicinal chemistry, target selection and interpreted biological/DMPK data. C.W.L. wrote the manuscript. M.T., E.S.C., S.R.B., T.C.C., D.S., Z.H. and M.W. performed chemical synthesis. T.M. and J.SD performed and analyzed in vitro pharmacology assays. G.H., K.M. and A.K. performed in vivo behavior pharmacology assays. T.M.B. and W.N. performed in vitro and in vivo DMPK studies. All authors have given approval to the final version of the manuscript.

The authors declare the following competing financial interest(s): While we abandoned the patent application for the compounds described in this manuscript, Vanderbilt and Ono hold patents on other TREK inhibitors.