Synthesis and evaluation of 4,6-disubstituted pyrimidines as CNS penetrant pan-muscarinic antagonists with a novel chemotype

Abstract

This Letter describes the synthesis and structure activity relationship (SAR) studies of structurally novel M₄ antagonists, based on a 4,6-disubstituted core, identified from a high-throughput screening campaign. A multi-dimensional optimization effort enhanced potency at both human and rat M₄ (IC₅₀s < 300 nM), with no substantial species differences noted. Moreover, CNS penetration proved attractive for this series (brain:plasma K_p,uu = 0.87), while other DMPK attributes were addressed in the course of the optimization effort, providing low in vivo clearance in rat (CL_p = 5.37 mL/min/kg). Surprisingly, this series displayed pan-muscarinic antagonist activity across M_1–5, despite the absence of the prototypical basic or quaternary amine moiety, thus offering a new chemotype from which to develop a next generation of pan-muscarinic antagonist agents.

Graphical Abstract

The muscarinic acetylcholine receptors (mAChRs) are members of the class A family of G protein-coupled receptors (GPCRs), which modulate the activity of the neurotransmitter acetylcholine. Currently, five mAChR subtypes (M₁–M₅) have been characterized. M₁, M₃ and M₅ are G_q-coupled, and activate phospholipase C and calcium mobilization. M₂ and M₄ are coupled to G_i/o, and thus inhibit the actions of adenylyl cyclase. The mAChRs are distributed throughout the central and peripheral nervous systems, where they modulate a wide variety of neuronal and autocrine functions relating to memory, nociception, gastrointestinal function, and many others. The development of therapeutics targeting the mAChRs is therefore widely pursued, and such agents have the potential to be useful for a number of different pathologies.^1,2

The majority of known muscarinic antagonists such as atropine (1) and the anti-emetic scopolamine (2) (Figure 1) are non-selective across the five receptor subtypes (pan-muscarinic antagonists). Other muscarinic antagonists include tiotropium bromide (3), a quaternary ammonium salt currently used for the treatment of chronic obstructive pulmonary disease (COPD),³ and AZD8683 (4), an M₃-preferring antagonist with a long duration of action and reduced side effect profile compared to tiotropium (Figure 1).⁴ Muscarinic antagonists are also used to treat overactive bladder and have shown promise for the treatment of irritable bowel syndrome (IBS).^5–7

Figure 1.

Chemical structures of known muscarinic antagonists 1–4, and the newly identified hit 5 from an M₄ antagonist high-throughput screen.

Since the M₄ receptor subtype is co-localized with dopaminergic D₁ receptors in the striatum, M₄ modulation has been identified as potentially playing a role in various movement disorders such as Parkinson’s disease and dystonia; hence, our interest in development of selective M₄ antagonists.^8–11 Development of agents that are selective for M₄, or any of the individual mAChRs, has traditionally proven difficult due to the high sequence homology amongst the receptor subtypes.² In an effort to identify structurally novel and selective M₄ antagonist chemotypes, a high-throughput screen was performed, and we identified the 4,6-disubstituted pyrimidine 5 (Figure 1) as a putative lead. This result was exciting, as 5 (VU0623863) represents a novel muscarinic antagonist scaffold, and we surmised that the potential for selectivity across the mAChR subtypes was high (especially given the lack of the basic nitrogen found in other non-selective muscarinic scaffolds 1–4), as we have previously demonstrated for both M₁ and M₅.^12–18

Compound 5 was resynthesized as shown in Scheme 1. Briefly, a HATU-mediated amide coupling between 1,2,3,4-tetrahydroisoquinoline (THIQ) 6 and N-Boc-isonipecotic acid 7 affords 8 in 91% yield. Subsequent Boc-deprotection yielded HCl salt 9 in quantitative yield. Nucleophilic aromatic substitution (S_NAr) between 4,6-dichloropyrimidine 10 and 4-ethylphenol 11 provided 12, which could be coupled to 9 via microwave-assisted S_NAr.¹⁹

Scheme 1.

Synthesis of compound 5 and route for analog synthesis.^a

^aReagents and conditions: (a) HATU, DIPEA, DCM, 91%; (b) HCl, 1,4-dioxane, 99%; (c) 12, DIPEA, NMP, microwave, 150 °C, 59%; (d) K₂CO₃, DMF, 100 °C, 38%.

After resynthesis, compound 5 was found to have comparable, sub-micromolar potency at both human and rat M₄ (IC₅₀s of 380 nM (pIC₅₀ = 6.46±0.09, 5.3±0.7 % ACh Min) and 590 nM (pIC₅₀ = 6.25±0.10, 13.1±2.5 % ACh Min), respectively). Due to the unique and non-basic chemotype, we anticipated that this would be selective for M₄, akin to our related efforts on M₁ and M₅.^12–18 Surprisingly, the compound was also found to have substantial antagonist activity at all of the mAChRs (M_1–3,5 IC₅₀s < 1.5 μM), but was weakly M₄-preferring (2- to 5-fold). We also assessed the DMPK profile of 5, and found it to be highly bound in plasma (f_{u,p lasma} (r, h) = 0.004, 0.001), with a high rat brain:plasma K_p (1.45), but lower K_p,uu (0.36)(rat f_u,brain = 0.001). Furthermore, the predicted hepatic clearance of 5 was near hepatic blood flow in both human and rat (CL_hep = 20.3 and 68.6 mL/min/kg, respectively) based on CL_int data from microsomes. Despite these blemishes, this scaffold was attractive from an academic standpoint to assess if either M₄ selectivity could be improved or if a next generation pan-mAChR antagonist could be generated with favorable PK and CNS penetration to compliment 1–4.

Toward these parallel goals, an SAR campaign focused around modifications to the eastern diaryl ether motif was pursued. In order to streamline the synthesis of these compounds for a more high-throughput approach, chloropyrimidine intermediate 13 was first synthesized from HCl salt 9 (Scheme 2). This intermediate could then be used as a platform for library synthesis, and a number of substituted phenols were coupled to 13 through microwave-assisted S_NAr in moderate yields (14a–o).

Scheme 2.

Synthesis of analogs 14.^a

^aReagents and conditions: (a) 4,6-dichloropyrimidine, DIPEA, DMF, 100 °C, 88%; (b) substituted phenol, Cs₂CO₃, NMP, microwave, 180 °C, 45–67%.

All analogs 14 were of comparable potency at rat M₄ (IC₅₀s within 2- to 3-fold, n = 1), and all were uniformly predicted to be highly cleared (predicted CL_hep of ~18–20 mL/min/kg and ~60–68 mL/min/kg, for human and rat respectively). Similarly, SAR was rather ‘flat’ with little texture with respect to M₄ inhibition, suggesting other areas of the molecule may be better targets for future optimization. However, analog 14g showed high brain distribution (rat K_p ~0.68, K_p,uu ~0.75), but its high clearance hindered any further advancement. Moreover, when evaluated in a [³H]-NMS binding assay with human M₄ cell membranes, 14k (VU6008913) displaced the radioligand binding with a K_i of 10.3 nM (cf. to atropine, K_i = 1.6 nM),²⁰ which translated into no mAChR selectivity (M₁ IC₅₀ = 250 nM, M₂ IC₅₀ = 330 nM, M₃ IC₅₀ = 380 nM, M₅ IC₅₀ = 510 nM, all n = 1 and <6% ACh Min). Similarly, other potent antagonists such as 14g (M₁ IC₅₀ = 322 nM, M₂ IC₅₀ = 384 nM, M₃ IC₅₀ = 404 nM, M₅ IC₅₀ = 732 nM, all n = 1 and <6% ACh Min) and 14h (M₁ IC₅₀ = 313 nM, M₂ IC₅₀ = 352 nM, M₃ IC₅₀ = 365 nM, M₅ IC₅₀ = 695 nM, all n = 1 and <6% ACh Min) were pan-mAChR antagonists. Thus, it became readily apparent that this series was not likely to afford selective M₄ antagonists, but that the goal should instead focus on development of a next generation pan-mAChR antagonist with exceptional PK and CNS penetration to compliment 1–4.

Due to a putative blockade of oxidative metabolism of the phenyl moiety via fluoro substitution, we next decided to hold the 2,4-difluoro phenylether eastern piece constant and explore an amide library to evaluate alternatives for the western THIQ moiety. Starting from commercially available methyl 1-(6-chloropyrimidin-4-yl)piperidine-4-carboxylate 15, a one-pot S_NAr/methyl ester hydrolysis was performed under microwave irradiation to give carboxylic acid 16 in moderate yield, which could then be used for HATU-mediated amide couplings to give final analogs 17a–h (Scheme 3).

Scheme 3.

Synthesis of compounds 17a–h.^a

^aReagents and conditions: (a) 2,4-difluorophenol, Cs₂CO₃, NMP, microwave, 180 °C, 49%; (b) amine, HATU, DIPEA, DMF, 32–67%.

As shown in Table 2, SAR was steep with respect to replacements for the THIQ moiety. Benzyl amines such as 17g and 17h were inactive, as was a CF₃-substituted THIQ, 17a. Electron-donating moieties, such as the regioisomeric OMe-substituted THIQ analogs 17b and 17c, retained M₄ inhibitory activity, as did a fully saturated congener 17d. However, the isoindoline derivative 17e emerged as the most exciting analog. 17e was a potent antagonist at both human and rat M₄ (hM₄ IC₅₀ = 340 nM, pIC₅₀ = 6.49±0.06, 18.4±2.3% ACh Min and rM₄ IC₅₀ = 560 nM, pIC₅₀ = 6.32±0.18, 42.7±3.2% ACh Min), but displayed partial antagonist activity.^21,22 Moreover, 17e proved to also be a pan-mAChR antagonist (Figure 2). However, the disposition of 17e was superior to all other analogs 14 and 17 evaluated. 17e (VU6009229) displayed good CNS penetration (rat brain:plasma K_p = 0.67, K_p,uu = 0.87) and low to moderate predicted hepatic clearance (rat CL_hep = 36.5 mL/min/kg and human CL_hep = 5.56 mL/min/kg). A robust in vitro:in vivo correlation (IVIVC) was noted, with 17e displaying a low rat in vivo clearance (CL_p = 5.37 mL/min/kg; predicted CL_hep with inclusion of binding terms in the well-stirred model = 5.93 mL/min/kg) with a 3.7 hour elimination half-life and modest volume (V_ss = 1.35 L/kg) in an IV cassette (0.2 mg/kg; n = 1) study. Thus, 17e emerged as a next generation pan-mAChR partial antagonist tool compound, with attractive in vivo rat PK and excellent CNS penetration, suitable for in vitro and in vivo studies.

Table 2.

Structures and mAChR activities of analogs 17a–h.

Compound	R	hM4 IC50 (μM)a [% ACh Min ±SEM]	hM4 pIC50 (±SEM)a
17a		>10	>5
17b		0.58 [5.5±2.9]	6.25±0.09
17c		2.47 [11.8±1.5]	5.62±0.08
17d		0.41 [5.1±0.6]	6.39±0.05
17e		0.34 [18.4±2.3]	6.49±0.06
17f		1.04 [6.0±0.5]	6.00±0.05
17g		>10	>5
17h		>10	>5

^aMean of three independent determinations in a calcium mobilization assay using recombinant hM₄-expressing Chinese hamster ovary cells co-transfected with chimeric G_qi5 in the presence of an ACh EC₈₀.

Figure 2.

Concentration response curves (CRCs) for 17e (VU6009229) in calcium mobilization assays with recombinant hM_1–5 Chinese hamster ovary cells (co-transfected with G_qi5) in the presence of an approximate EC₈₀ of ACh. (M₁ IC₅₀ = 540 nM (18.2% ACh min), M₂ IC₅₀ = 520 nM (8.7% ACh min), M₃ IC₅₀ = 660 nM (2.3% ACh min), M₅ IC₅₀ = 1,130 nM (42.6% ACh min), all n = 1.

In summary, an M₄ HTS campaign identified 5 as a novel M₄ antagonist chemotype, that proved to lack mAChR subtype selectivity. Further optimization afforded 17e, a potent and highly CNS penetrant pan-mAChR antagonist with an attractive rat in vivo PK profile. Additionally, 17e and related analogs do not feature the prototypical tropane structure of classical muscarinic antagonists, nor a strong basic amine. Thus, these analogs represent a next generation of pan-mAChR antagonists that could serve as leads for the development of potential safer or differentiating anti-cholinergic agents. The expedient and straightforward synthesis of these analogs will allow us to further explore the requirements for muscarinic selectivity, as well as fine-tune the DMPK properties of this series.

Table 1.

Structures and mAChR activities of analogs 5, 14a–o.

Compound	R	hM4 IC50 (μM)a [% ACh Min ±SEM]	hM4 pIC50 (±SEM)a
5	4-Et	0.38 [5.3±0.7]	6.46±0.09
14a	H	0.65 [2.7±0.3]	6.21±0.09
14b	4-Me	0.31 [3.3±0.2]	6.52±0.08
14c	2-naphthyl	0.71 [6.8±2.2]	6.16±0.06
14d	3,4-methylenedioxy	0.51 [4.1±0.3]	6.30±0.09
14e	4-F	0.24 [3.5±0.2]	6.64±0.02
14f	3-F	0.31 [3.2±0.3]	6.52±0.06
14g	2-F	0.13 [3.2±0.5]	6.91±0.08
14h	4-Cl	0.14 [3.0±0.2]	6.88±0.07
14i	4-OMe	0.46 [5.3±1.0]	6.30±0.06
14j	2,3-diF	0.19 [3.1±0.1]	6.77±0.15
14k	2,4-diF	0.10	7.00±0.08
14l	2,5-diF	0.08 [3.1±0.2]	7.11±0.11
14m	2,6-diF	2.72 [7.0±0.9]	5.57±0.05
14n	2-F, 4-Cl	0.15 [3.1±0.3]	6.93±0.06
14o	2,4-diCl	0.27 [3.0±0.1]	6.58±0.09

^aMean of three independent determinations in a calcium mobilization assay using recombinant hM₄-expressing Chinese hamster ovary cells co-transfected with chimeric G_qi5 in the presence of an ACh EC₈₀.