Discovery, synthesis and characterization of a highly mAChR selective M₅ orthosteric antagonist, VU0488130 (ML381): a novel molecular probe

Abstract

Of the five G-protein-coupled muscarinic acetylcholine receptors (mAChRs or M₁-M₅), M₅ is the least explored and understood due to a lack of mAChR subtype selective ligands. We recently performed a high-throughput functional screen and identified a number of weak antagonist hits that were selective for M₅. An iterative parallel synthesis and detailed molecular pharmacologic profiling effort, led to the discovery of the first highly selective, CNS penetrant M₅ orthosteric antagonist tool compound, with submicromolar potency (hM₅ IC₅₀ = 450 nM, hM₅ K_i = 340 nM, M₁-M₄ IC₅₀s >30 μM), enantiospecific inhibition and an acceptable DMPK profile for in vitro and electrophysiology studies.

Numerous studies have suggested that subtype-selective muscarinic acetylcholine receptor (mAChR) antagonists may provide novel approaches to treat a number of central nervous system (CNS) disorders;^[1–7] however, the majority of existing antagonists are not selective for specific mAChR subtypes, forcing the field to rely on biochemical and or genetic approaches to discern therapeutic potential.^[1–7] Of the central mAChRs, M₅ expression is the lowest (<2%), and M₅ is predominantly localized to dopaminergic neurons originating in the substantia nigra pars compacta (SNc) and the ventral tegmental area (VTA).^[4–10] Studies with M₅^−/− mice have confirmed the hypothesis that M₅ modulates dopaminergic neurotransmission and that M₅ functions in addiction/reward mechanisms, with M₅^−/− mice displaying reduced cocaine place preference and self-administration without affecting food intake.^[4–10] Further biochemical studies with M₅-targeted antisense oligonucleotides, or scopolamine, infused into the VTA have recapitulated the genetic findings,^[11,12] strongly suggesting that selective M₅ inhibition may provide a novel therapeutic approach for the treatment of addiction.^[1–12]

Only recently have M₅-preferring and M₅-selective ligands begun to emerge. By targeting allosteric sites^[13,14] on mAChRs, we recently described the first M₅-selective ligands, a series of isatin-based M₅ positive allosteric modulators (PAMs) (1-3, Figure 1).^[15–17] Shortly thereafter, Dwoskin and co-workers described an M₅-preferring orthosteric antagonist (4, K_i= 2.2 μM) with 11-fold selectivity for M₅ over M₁, but no functional (IC₅₀) or DMPK data was reported.^[18] Last year, we reported the first results from a functional M₅ high-throughput screen (HTS), where we identified and optimized the first M₅ selective negative allosteric modulator (NAM), 5, with submicromolar potency (hM₅ IC₅₀ = 300 nM, hM₁-M₄ IC₅₀s >30 μM).^[19] Based on the Dwoskin work,^[18] and our earlier studies that had yielded a highly (>75-fold versus M₂-M₅) subtype-selective M₁ orthosteric antagonist 6,^[20] we revisited our HTS hits in an attempt to identify and optimize a highly M₅ selective orthosteric antagonist to complete our tool kit of M₅ probes to dissect the role of M₅ in the CNS.

Figure 1.

Structures of recently reported M₅ PAMs (1-3), an M₅-preferring orthosteric antagonist 4, the first highly selective M₅ NAM 5 and a highly M₁ selective orthosteric antagonist 6.

Under the auspice of the Molecular Libraries Probe Center Network (MLPCN),^[21] a triple-add, functional HTS was performed evaluating 360,000 compounds for M₅ modulatory activity. In parallel, we screened the same collection in triple-add, functional mode against M₁, M₄ and the parental Chinese hamster ovary (CHO) cell line, resulting in nine confirmed and selective inhibitors of M₅, from which the M₅ NAM 5 resulted.^[19] Evaluation of the eight remaining hits led us to quickly focus on a weak, yet selective, isoxazole-based amide 7 (Figure 2), which confirmed in the HTS center with an M₅ IC₅₀ of 9.3 μM and selectivity versus M₁ and M₄ (IC₅₀s > 30 μM). Resynthesis of 7 (VU0480131), and screening fresh powder significantly increased our enthusiasm for the hit, as 7 displayed improved potency for both human and rat M₅ (hM₅ IC₅₀ = 1.12 μM (pIC₅₀ = 5.94±0.03), rM₅ IC₅₀ = 3.47 μM (pIC₅₀ = 5.46±0.07)) and high selectivity versus human M₁-M₄ (IC₅₀s > 30 μM). We then assessed the mechanism of action of 6 and performed competition binding studies with [³H]-N-methylscopolamine (NMS) and an atropine control, and observed that 7 displaced [³H]-NMS in a concentration dependent manner with a K_i of 1.3 μM, pK_i = 5.88±0.05, (comparable to its IC₅₀ of 1.1 μM), indicating competitive, and thus, orthosteric antagonism, reminiscent of the highly selective M₁ orthosteric antagonist 6.^[19,20]

Figure 2.

Structure and pharmacological profile of HTS hit 7 (VU0480131). A) Structure and pharmacology of 7 as an HTS stock solution (hM₅ IC₅₀ = 9.3 μM) and as resynthesized, fresh powder (hM₅ IC₅₀ = 1.12 μM). B) Concentration-response curves (CRCs) of 7 for human and rat M5, as well as human M₁-M₄ (IC₅₀s >30 μM, n=3). C) [³H]NMS competition binding [n = 3] in membranes prepped from human M₅ cells, showing competitive displacement (K_i = 1.3 μM).

Figure 3 highlights the chemical optimization strategy for 7, wherein we evaluated multiple dimensions of the ligand simultaneously through iterative parallel synthesis, as SAR for the analogous 6 was very steep. The first round of library synthesis surveyed alternate amides; thus, commercial acid 8 was acylated under HATU conditions to afford analogs 9 in yields ranging from 56–90% (Scheme 1). SAR was extremely steep, with all analogs tested, even regioisomeric pyridines, possessing hM₅ IC₅₀s >10 μM. This effort only afforded one weakly active compound 9a (hM₅ IC₅₀ = 5.6 μM, rM₅ IC₅₀ >10 μM, an N-ethyl congener of 7 (Figure 4).

Figure 3.

Chemical optimization plan for HTS hit 7 (VU0480131) via multidimensional, iterative parallel synthesis.

Scheme 1.

Preparation of amide analogs 9. Reagents and conditions: a) R₁R₂NH, HATU, CH₂Cl₂, rt, 56–90%.

Figure 4.

Structures and activities of amide analogs 9. All were uniformly inactive, except for the N-ethyl analog 9a.

With the amide region of 7 intolerant to modification, we next prepared a library of diverse aryl ethers to assess SAR and potentially identify a replacement for the ketone (Scheme 2). Alcohol 10 smoothly reacted with a variety of phenols under Mitsunobu conditions to provide a library of ether analogs 11, in yields ranging from 35–70%.^[22] Nothing in this collection showed activity (hM₅ IC₅₀s >10 μM), despite surveying nitro, halogen, alkoxy and alkyl moieties around the pendant phenyl ring (See experimental section for full listing of analogs). Replacement of the ether oxygen with an NH, or acylated congener, also led to a complete loss of M₅ inhibitory activity, as did amide analogs in place of the aryl ether moiety. Replacement of the central isoxazole with a phenyl ring also led to inactive compounds. Again, akin to the steep SAR noted for the highly selective M₁ orthosteric antagonist 6, very little modification to 7 was tolerated. Thus, we elected to next synthesize the single enantiomers of 7 and assess if enantioselective inhibition was observed.

Scheme 2.

Preparation of aryl ether analogs 11. Reagents and conditions: a) ArOH, PPh₃, DIAD, CH₂Cl₂, rt, 2h, 35–70%.

Starting with commercial (S)- or (R)-1-(pyridin-2-yl)ethanamine 12, treatment with Nosyl chloride provides sulphonamides 13 in good yield (Scheme 3). The N-methyl moiety was installed via a Mitsunobu reaction with MeOH to afford tertiary sulphonamides 14. Deprotection was accomplished by treatment with 3-chlorothiophenol in a modest 20% yield to provide chiral secondary amines 15. Finally, a HATU mediated coupling with acid 8 afforded the single enantiomers of 7, (S)-16 and (R)-16, in overall yields of ~15%.^[22]

Scheme 3.

Preparation of (S)-16 and (R)-16, single enantiomers of racemic 7. Reagents and conditions: a) NosCl, DIEPA, CH₂Cl₂, rt, 78–90% b) MeOH, PPh3, DIAD, CH₂Cl₂, rt, 74–90% c) 3-ClPhSH, K₂CO₃, MeCN, rt, 20% d) 8, HATU, DIEPA, CH₂Cl₂, rt, 89–96%. Enantiopurity >99% ee as analysed by chiral SFC.

When screened in the human M₅ cell-based assay, enantiospecific inhibition was observed, with (R)-16 possessing no activity (hM₅ IC₅₀ >10 μM), yet (S)-16 displayed (Figure 5) submicromolar potency (hM₅ IC₅₀ = 450 nM, pIC₅₀ = 6.34±0.05) and exceptional mAChR selectivity (hM₁-M₄, IC₅₀s >30 μM). These findings led us to then profile (S)-16 against the rat mAChRs, and once again, (S)-16 was selective (rM₁-rM₄ >30 μM), but a 4-fold diminution in potency was noted for rat M₅ (rM₅ IC₅₀ = 1.65 μM, pIC₅₀ = 5.78+0.04). Based on the potency and selectivity of (S)-16, it was approved as an MLPCN probe and given the identifier ML381 (VU0488130).^[21] As expected, competition binding with [³H]NMS indicated that ML381 was an orthosteric antagonist, with a K_i of 341 nM, pK_i = 6.47±0.03, (comparable to its IC₅₀ of 450 nM, Figure 5B).

Figure 5.

Structure and pharmacological profile of (S)-16, ML381 (VU0480131). A) Structure and pharmacology of ML381. B) [³H]NMS competition binding [n = 3] in membranes prepared from human M₅ cells, showing competitive displacement (K_i = 340 nM). C) CRCs of ML381 for human M₅ (hM₅ IC₅₀ = 450 nM) and human M₁-M₄ (hM₁-M₄ IC₅₀s >30 μM). D) CRCs of ML381 for rat M5 (rM₅ IC₅₀ = 1.65 μM) and rat M₁-M₄ (rM₁-M₄ IC₅₀s >30 μM).

Next, we began to assess the physiochemical and DMPK profile of ML381. ML381 is a low molecular weight probe (MW = 379) with favourable lipophilicity (cLogP = 2.74) and an acceptable TPSA of 86. Moreover, ML381 was found be both soluble in phosphate buffered saline, PBS, (96 μM@ pH = 7.4 at 23 °C) and stable (99% parent remaining at 48 hours in PBS at 23 °C). In rat and human hepatic microsomes, ML381 exhibited high intrinsic clearance (rat CL_int: 770 mL/min/kg, human CL_int: 93 mL/min/kg) with predicted hepatic clearance values (rat CL_hep: 64 mL/min/kg, human CL_hep: 17 mL/min/kg) near the respective rates of hepatic blood flow in each species (Table 1). In plasma protein binding experiments, the compound exhibited exceptionally high fraction unbound (fu_plasma) in rat (58.4%) and moderately high fu_plasma in human (11%), which prompted investigation into the compound’s stability over time (4 hours; 37 °C) in plasma of both species, ultimately revealing a pronounced instability in rat plasma (approximately 1 % of parent remaining) and a moderate instability in human plasma (approximately 76 % of parent remaining), which ultimately prevented accurate determination of fu_plasma. A high fraction unbound in brain (fu_brain: 0.14) was observed for ML381 in rat whole brain homogenate. In order to determine the relevant biotransformation pathway(s) contributing to the poor metabolic stability observed in vitro, metabolite identification (Met ID) experiments were performed using hepatic microsomes and plasma from both species (Figure 6). These analyses revealed hydrolytic cleavage of the amide as the principle pathway of biotransformation in microsomes of both species as well as in rat plasma. Additionally, a number of NADPH-dependent mono-oxidation pathways were also identified in rat and human microsomes, including O-dearylation (M2), phenyl hydroxylation (M3), keto-reduction (M4), and N-dealkylations at the amide (M5, M6; Figure 6).

Table 1.

DMPK profile of ML381.

Species	Property	Value1
Sprague Dawley Rat	Hepatic Microsomal CLint	770 mL/min/kg
	Predicted CLhep	64 mL/min/kg
	fuplasma; fubrain	0.582; 0.14
	Cbrain:Cplasma (Kp)	0.583
Human	Hepatic Microsomal CLint	93 mL/min/kg
	Predicted CLhep	17 mL/min/kg
	fuplasma	0.112
	MDCK-MDR1 ER	1.6
	P450 1A2 IC50	> 30 μM
	P450 2C9 IC50	6.3 μM
	P450 2D6 IC50	> 30 μM
	P450 3A4 IC50	27 μM

¹values represent means of at least two replicates with similar results

²fu_plasma values not accurate due to instability in rat plasma

³K_p determined at 0.25 hr following a 0.2 mg/kg IV dose (n = 2)

Figure 6.

Metabolite identification for 7, ML381 (VU0480131), in rat and human, indicating six major routes of metabolism (M1-M6).

In order to gauge distribution to the central nervous system (CNS), concentrations of ML381 in whole brain and plasma at a single time point (0.25 hr) were measured following a single intravenous (IV) administration (0.2 mg/kg) to male, Sprague Dawley rats (n = 2). This study revealed a brain:plasma partition coefficient (K_p) of 0.58, suggesting moderate distribution to the CNS (Table 1); determination of an unbound brain:plasma partition coefficient (K_p,uu) was precluded by an absence of accurate fu_plasma data due to the aforementioned plasma instability. In a bidirectional MDCK-MDR1 transwell assay, ML381 (5 μM) exhibited an efflux ratio (ER) of 1.6, suggesting an absence of P-glycoprotein (P-gp)-mediated active efflux liabilities at the blood-brain barrier (BBB; Table 1). Furthermore, favorably low inhibition of four major human cytochrome P450 enzymes by ML381 was observed in human hepatic microsomes, with inhibition of 2C9 (6.3 μM) representing the predominant liability (Table 1). Moreover, ML381 was screened in a Eurofins radioligand binding panel of 68 GPCRs, ion channels and transporters at a concentration of 10 μM,^[23] and no significant activity was noted (no inhibition >50%@10 μM). Thus, in addition to unprecedented selectivity versus M1-M4, ML381 possessed clean ancillary pharmacology against a diverse array of discrete molecular targets. Together, these findings suggest that ML381 possesses an overall acceptable drug metabolism and pharmacokinetic (DMPK) profile for pharmacodynamic studies in rat,^[24] with the exception of poor metabolic stability and a potential for amide hydrolysis in plasma; however, ML381 is best suited as an in vitro/electrophysiological probe.

In summary, we have developed (S)-16 (also known as ML381 or VU0480131), the most potent and selective M₅ orthosteric antagonist (hM₅ IC₅₀ = 450 nM, M₅ K_i = 340 nM, hM₁-hM₄ IC₅₀s >30 μM) reported to date. Moreover, ML381 possesses a favourable DMPK profile and CNS penetration; however, instability in rat plasma indicates the utility of ML381 will be primarily as a molecular probe for in vitro and electrophysiology studies. SAR was extremely steep for this series, with subtle modifications leading to complete loss of M₅ inhibitory activity, reminiscent of the M₁ ligand 6; however, enantiospecific inhibition was observed, with all of the activity residing in the (S)-enantiomer. Current efforts are focused on further chemical optimization to address the plasma instability identified in Met ID studies and to improve PK and CNS penetration. Studies are underway and will be reported in due course. ML381 is an MLPCN probe and is freely available upon request.

Experimental Section

Experimental procedures for medicinal chemistry, pharmacology and drug metabolism as well as characterization of compounds are provided in the Supporting information, available at http://dx.doi.org/10.1002/cmdc.20xxxxxxx.