Discovery and optimization of 3-(4-aryl/heteroarylsulfonyl)piperazin-1-yl)-6-(piperidin-1-yl)pyridazines as novel, CNS penetrant pan-muscarinic antagonists

Abstract

This Letter describes the synthesis and structure activity relationship (SAR) studies of structurally novel M₄ antagonists, based on a 3-(4-aryl/heteroarylsulfonyl)piperazin-1-yl)-6-(piperidin-1-yl)pyridazine core, identified from a high-throughput screening campaign. A multidimensional optimization effort enhanced potency at human M₄ (hM₄ IC₅₀s < 200 nM), with only moderate species differences noted, and with enantioselective inhibition. Moreover, CNS penetration proved attractive for this series (rat brain:plasma K_p = 2.1, K_p,uu = 1.1). Despite the absence of the prototypical mAChR antagonist basic or quaternary amine moiety, this series displayed pan-muscarinic antagonist activity across M_1-5 (with 9- to 16-fold functional selectivity at best). This series further expands the chemical diversity of mAChR antagonists.

Graphical abstract

Selective inhibition of the M₄ receptor, one of five muscarinic acetylcholine receptors (mAChRs or M_1-5), a class A family of G protein-coupled receptors (GPCRs), has emerged as an exciting new approach for the symptomatic treatment of Parkinson’s disease.^1–4 However, the development of small molecule ligands that are selective for M₄, or any of the individual mAChRs, has proven to be challenging due to the high sequence homology amongst the receptor subtypes.^5–8 Historically, mAChR antagonists possessed somewhat conserved chemotypes, exemplified by a basic tertiary or quaternary amine, and compounds with these functional groups represent the bulk of high-throughput screening (HTS) hits (the ‘usual suspects’) and marketed drugs 1-4 (Figure 1).^9–11 Thus, we were delighted to identify fundamentally new chemotypes in an M₄ functional HTS campaign, which was subsequently optimized to deliver potent and CNS penetrant antagonists such as 5; however, despite the absence of the classical pharmacophore, 5 proved to be a pan-muscarinic antagonist that bound to the orthosteric (ACh) site.¹² Another departure from the classical mAChR chemotype was found in HTS hit 6, an ~ 1 μM hM₄ antagonist. This Letter details the synthesis, SAR, pharmacology and DMPK profiles of analogs of 6, and the finding, once again, of pan-mAChR inhibition.

Figure 1.

Chemical structures of known muscarinic antagonists 1-4, the newly optimized pan-mAChR antagonist 5, and a novel hit 6 from an M₄ antagonist high-throughput screen.

Compound 6 was resynthesized as shown in Scheme 1 as both the racemate, as well as the single (R)- and (S)-enantiomers. Briefly, 3,6-dichloropyridazine was subjected to an S_NAr reaction with 2-methylpiperidine in NMP at 200 °C, followed by a second S_NAr with piperazine to afford racemic 8 or chiral 8a/8b in 32–39% isolated yields in a one-pot procedure. Standard sulfonamide formation with 2-chlorobenzenesulfonyl chloride delivered racemic 6 and chiral 6a/6b in moderate yields.¹³

Scheme 1.

Synthesis of racemic compound 6 as well as discrete enantiomers.^a

^aReagents and conditions: (a) (R,S), (R) or (S)-2-methylpiperidine, DIPEA, NMP, microwave, 200 °C; (b) piperazine, 200 °C, 36% (R,S), 32% (R), 39% (S); (c) 2-chlorobenzenesulfonyl chloride, DIPEA, DCM, r.t., 38% (R,S), 35% (R), 44% (S).

After resynthesis, racemic 6 was found to have submicromolar potency at human M₄ (hM₄ IC₅₀ = 530 nM, pIC₅₀=6.28±0.07, 9.2±5.4 % ACh Min). Excitingly, 6a, the (S)-enantiomer showed enhanced potency (hM₄ IC₅₀ = 440 nM, pIC₅₀ = 6.37±0.08, 4±0 % ACh Min), while 6b, the (R)-enantiomer was significantly less active (hM₄ IC₅₀ = 6.08 μM, pIC₅₀ = 5.24±0.09, 12±2 % ACh Min). Based on the unique and non-basic chemotype, coupled with the noted enantioselectivity (~14-fold) of hM₄ inhibition, we anticipated that 6a would be selective for M₄, akin to our related efforts on M₁ and M₅.^14–20 Interestingly, 6a inhibited the other four mAChRs (hM₁ IC₅₀ = 1.3 μM, 2.8% ACh Min; hM₂ IC₅₀ = 1.2 μM, 2.9% ACh Min; hM₃ IC₅₀ = 7.0 μM, 4.3% ACh Min; hM₅ IC₅₀ = 6.5 μM, 3.1% ACh Min), but was moderately M₄-preferring (2.7- to 16-fold). We then determined activity at rat mAChRs, and here, 6a was a weaker antagonist (rM₁ IC₅₀ = 1.4 μM, 2.6% ACh Min; rM₂ IC₅₀=5.9 uM, 7.1% ACh Min; rM₃ IC₅₀ >10 μM, 40 % ACh Min; rM₄ IC₅₀ = 1.0 μM, 4.8% ACh Min; rM₅ IC₅₀ >10uM, 37% ACh Min), only partially diminishing an EC₈₀ of ACh at M₃ and M₅. While not the result we were hoping for in terms of mAChR selectivity, the series was still deemed worthy of further optimization. Figure 2 highlights the chemical optimization plan for 6a.

Figure 2.

Chemical optimization plan for 6a, surveying multiple dimensions of the novel mAChR antagonist chemotype.

Our initial survey held the eastern portion of 6a constant while evaluating alternative sulfonamides, according to the route depicted in Scheme 1, which afforded analogs 9 (Table 1). Clear SAR was noted with analogs 9. Substitutions in the 2-position of the phenylsulfonamide were preferred, as potency decreased from 6a (hM₄ IC₅₀ = 440 nM), to the 3-Cl congener 9a (hM₄ IC₅₀ = 760 nM), and to the 4-Cl analog 9b (hM₄ IC₅₀ = 2.34 μM). Unsubstituted phenyl, 9c, was weak, as were electron-donating moieties in the 2-positon (9d). A 2-CF₃ derivative (9e) was essentially equipotent to 6a. A 2,5-dimethylisoxazole (9g) afforded the best activity in this series (hM₄ IC₅₀ = 90 nM), and a piperonyl congener 9i (hM₄ IC₅₀ = 200 nM) was 10-fold more potent than a 3,4-dimethoxy analog 9j (hM₄ IC₅₀ = 2.80 μM). However, all analogs 9 (and including 6a) displayed high predicted hepatic clearance (CL_hep) based on microsomal intrinsic clearance (CL_int) data in both rat and human near hepatic blood flow rates (>60 mL/min/kg and >20 mL/min/kg, respectively), as well as high plasma protein binding. Based on these results, we performed metabolite identification studies in rat and human hepatic microsomes to assess soft spots and attempt to understand the high clearance, which appeared to be independent of the nature of the sulfonamide moiety. For this work, we evaluated 6a (VU6008887) in the presence and absence of NADPH, and found no NADPH-independent metabolism. 6a was more stable in human, forming 8 NADPH-dependent metabolites, but the parent remained the major species. In rat microsomes, the major peak (by UV and extracted ion chromatograms) was metabolite F, resulting from extensive oxidative metabolism of the piperidine moiety (Figure 3), and very little parent remained.

Table 1.

Structures and mAChR activities of analogs 6a, 9a-k.

Compound	R	hM4 IC50 (μM)a [% ACh Min ±SEM]	hM4 pIC50 (±SEM)a
6a		0.44 [4.0±0]	6.37±0.08
9a		0.76 [22±2]	6.15±0.11
9b		2.34 [30±2]	5.64±0.05
9c		2.57 [8.0±1]	5.59±0.02
9d		2.38 [11±0]	5.62±0.02
9e		0.52 [4.0±0]	6.28±0.02
9f		1.60 [8.0±0]	5.80±0.06
9g		0.09 [6.0±0]	7.04±0.03
9h		1.17 [7.0±1]	5.95±0.09
9i		0.20 [9.0±1]	6.71±0.06
9j		2.80 [46±3]	5.58±0.10
9k		4.02 [51±0]	5.40±0.08

^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 3.

Metabolite identification studies in rat and human liver microsomes in the presence or absence of NADPH. No NADPH-independent metabolites were noted in either species, but rat specifically produced metabolite F.

Based on these data, we first explored alternative heteroaryl replacements for the pyridazine ring, to assess if electronics could modulate metabolism of the piperidine ring (or the pendant methyl group) while maintaining M₄ inhibitory activity. Employing variations of 7 in Scheme 1, the two regioisomeric pyridine cores 10 and 11 were prepared, as well as a pyrazine core 12 and a phenyl congener 13 (Figure 4). Clearly, the pyridazine of 6a and 9a-k is essential for M₄ activity, and these alternate heterocyclic analogs did not provide superior compounds.

Figure 4.

Structures and activities of pyridazine replacements 10-13. Note, the parent pyridazine analog, 9i, has an hM₄ IC₅₀ of 200 nM.

Therefore, all efforts now focused on a multidimensional array surveying the most active sulfonamide moieties (6a, 9g and 9i) in combination with replacements for the 2-methlypiperidine moiety (Table 2). SAR was steep, with chiral 2-methyl morpholine surrogates (14a-c) devoid of M₄ activity (hM₄ IC₅₀s > 10 μM), as was a piperidinone (14o). Ring contraction to a chiral 2-methyl pyrrolidine as in 14d-f lost 5- to 6-fold in hM₄ inhibitory activity relative to 6a, as did deletion of the chiral methyl group (14g, h). Interestingly, ring expansion to an unsubstituted homopiperidine, as in 14i-k, afforded potent hM₄ antagonists (hM₄ IC₅₀s as low as 290 nM). This result led us to re-explore the piperidine core, but with a gem-dimethyl moiety in the 3-position (14l-n) to mimic the steric bulk of the homopiperidine. From this effort, 14n (hM₄ IC₅₀ = 120 nM, 12% ACh min) emerged as an attractive antagonist for further profiling, as it was also active on rat M₄ (rM₄ IC₅₀ = 365 nM, pIC₅₀ = 6.43±0.01, 32±1% ACh Min).

Table 2.

Structures and mAChR activities of analogs 14.

Cmpd	R	R1	hM4 IC50 (μM)a [% ACh Min ±SEM]	hM4 pIC50 (±SEM)a
14a			>10 [56]	>5
14b			>10 [63]	>5
14c			>10 [66]	>5
14d			2.90 [9.0±0]	5.54±0.02
14e			2.73 [14±5]	5.59±0.12
14f			3.31 [10±1]	5.48±0.03
14g			5.58 [12±1]	5.27±0.08
14h			0.73 [55±8]	6.18±0.14
14i			1.27 [6.0±0]	5.90±0.06
14j			0.29 [20±3]	6.54±0.02
14k			0.43 [8.0±1]	6.39±0.09
14l			0.94 [5.0±0]	6.03±0.03
14m			0.47 [14±2]	6.34±0.06
14n			0.12 [12±1]	6.96±0.14
14o			>10 [51]	>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.

Based on the unexpected results with 5 and 6a, we next evaluated the molecular pharmacology profile of 14n (VU6009833, Figure 5). Like 5, 14n proved to bind at the orthosteric ACh site (or possibly overlapping with the orthosteric site), affording a K_i of 7.78±0.86 nM in a radioligand binding assay displacing [³H]-NMS in human M₄ membranes. Compound 14n inhibited all five human mAChRs, but displayed weak, partial antagonism^21,22 of M₃ and M₅, and was a potent full antagonist of M₁ and M₂ (rat data similar but not shown). Thus, like atypical pyrimidine 5, 14n represents another novel chemotype with pan-mAChR inhibitory activity.

Figure 5.

In vitro molecular pharmacology profile of 14n (VU6009833). A) Radioligand binding assay displacing [³H]-NMS in human M₄ membranes. 14n afforded a K_i of 7.78±0.86 nM, while atropine, in agreement with historical data afforded a K_i of 0.64±0.09 nM. B) Concentration response curves (CRCs) for 14n in calcium mobilization assays in stably expressing hM_1-5 Chinese hamster ovary cells (for Gi-coupled hM₂ and hM₄, co-expressed with Gq_i5) in the presence of an approximate EC₈₀ of ACh. Data are presented as the percentage of the EC₈₀ ACh response of each cell line. All data points represent the mean of triplicate in single assay. (M₁ IC₅₀ = 288 nM (11% ACh min), M₂ IC₅₀ = 352 nM (17% ACh min), M₃ IC₅₀ = 782 nM (55% ACh min), M₅ IC₅₀ = 1,030 nM (43% ACh min).

14n possessed attractive physiochemical properties (clogP of 2.9, TPSA of 96 and a molecular weight of 434), which translated into acceptable plasma protein binding (f_u (r,h) = 0.07 and 0.08) and rat brain homogenate binding (f_u = 0.04). In our standard rat plasma:brain level (PBL) cassette study,²³ 14n displayed a K_p of 2.1 and a K_p,uu of 1.1, indicating excellent total and free CNS distribution. However, predicted hepatic clearance remained near hepatic blood flow rates in both rat and human (CL_hep = 66.6 mL/min/kg and 20 mL/min/kg, respectively), suggesting that co-administration with an inhibitor of metabolism (putatively CYP₄₅₀) and/or non-oral route(s) of administration may be required for in vivo studies with 14n. Currently, efforts are focused on improving metabolic stability of 14n and related analogs, as well as homology modeling and mutagenesis to better understand the mode of interaction with the ACh binding site.

In summary, an M₄ HTS campaign identified 6 as a novel M₄ antagonist chemotype, devoid of the ‘usual suspect’ pharmacophores. Synthesis of the discrete enantiomers indicated enantioselectivity with regards to M₄ inhibition, yet this series proved to lack mAChR subtype selectivity. Further optimization afforded 14n, a potent and highly CNS penetrant pan-mAChR antagonist with attractive physiochemical properties. Additionally, 14n and related analogs do not feature either a strong basic amine or the prototypical tropane structure of classical muscarinic antagonists. Thus, these analogs represent a next generation of pan-mAChR antagonists that could serve as leads for the development of potentially safer or differentiating anti-cholinergic agents. Studies are underway to better understand the nature of the binding mode, as well as efforts to reduce metabolic clearance and ultimately improve in vivo pharmacokinetics.