Discovery of VU6025733 (AG06827): A Highly Selective, Orally Bioavailable, and Structurally Distinct M₄ Muscarinic Acetylcholine Receptor Positive Allosteric Modulator (PAM) with Robust In Vivo Efficacy

Abstract

This work describes progress toward an M₄ PAM preclinical candidate. The SAR to address potency, clearance, subtype selectivity, CNS exposure, and P-gp efflux are detailed within. A novel 1-(7,8-dimethyl-[1,2,4]­triazolo­[4,3-b]­pyridazin-6-yl)­piperidin-4-ol scaffold was identified, and optimization provided a highly potent analog VU6025733 (hM₄ EC₅₀ = 23 nM; rM₄ EC₅₀ = 55 nM). Further characterization revealed a highly selective compound across muscarinic acetylcholine receptor subtypes with exceptional DMPK properties (in vivo rat CL_p = 5.9 mL/min/kg; t _1/2 = 4.8 h; CYP1A2 & CYP2C9 IC₅₀s > 30 μM, CYP2D6 IC₅₀ > 9 μM; CYP3A4 IC₅₀ > 25 μM). Moreover, VU6025733 demonstrated robust in vivo efficacy in a rat amphetamine-induced hyperlocomotion model in a dose-dependent manner. However, hepatotoxicity risk precluded further development.

Introduction

Cholinergic neurotransmission involves the binding of an orthosteric endogenous agonist, acetylcholine (ACh), to activate receptors such as the nicotinic acetylcholine (nAChRs) or muscarinic acetylcholine receptors (mAChRs). Often conditions involving cognitive impairment, e.g., Alzheimer’s disease (AD), have been associated with lower levels of acetylcholine in the brain. Such reductions in the levels of acetylcholine are perceived to be a consequence of the deterioration of cholinergic neurons of the basal forebrain, which widely innervates multiple areas of the brain crucial for higher processes. , Moreover, cholinergic hypofunction has been clinically linked to cognitive deficits in patients suffering from schizophrenia.

Previous endeavors to increase acetylcholine levels have taken one of two approaches: (1) increasing the levels of the acetylcholine precursor, choline; or (2) inhibition of the enzyme responsible for the metabolism of acetylcholine, acetylcholinesterase (AChE). Attempts to augment central cholinergic function through administration of choline or phosphatidylcholine have proved futile. Nonetheless, AChE inhibitors (AChEI’s) have been approved for the use in palliative, but not disease-modifying, treatments of cognitive deficits in Alzheimer’s patients. While AChEI’s have demonstrated therapeutic efficacy, they induce cholinergic side effects due to peripheral acetylcholine stimulation. These side effects have been observed in nearly one-third of the patients treated and include abdominal cramps, nausea, vomiting, and diarrhea. Additionally, some AChEI’s, such as tacrine, cause significant hepatotoxicity with elevated liver transaminases in nearly 30% of patients treated. Such adverse effects greatly hinder the clinical utility of AChEI’s and highlight the glaring need for an alternative approach to pharmacologically target cholinergic hypofunction. One such approach, as showcased in this paper, is to target the activation of mAChRs, which are widely expressed throughout the body, including the brain.

Muscarinic acetylcholine receptors are members of the Class A family of G-protein coupled receptors (GPCRs) and include five subtypes of receptors, designated M₁ - M₅. The subtypes can be grouped into two main categories: (1) those that are mainly G_q-coupled and activate Phospholipase C (M₁, M₃, and M₅) and (2) those that mainly couple to G_i/o and effector systems (M₂ and M₄). − Not only are these five distinct mAChR subtypes prevalent and differentially expressed in the mammalian central nervous system, but they also play varying roles in cognitive, sensory, motor, and autonomic functions. Thus, it has been hypothesized that selective agonists of mAChR subtypes involved in regulating processes associated with cognitive function could provide a superior avenue in the treatment of psychosis, schizophrenia, and related disorders.

The activation of peripheral M₂ and M₃ mAChRs has been linked to the most prominent side effects of AChE inhibitors and other cholinergic agents (i.e., bradycardia, GI stress, and excessive salivation and sweating). , Alternatively, the muscarinic M₄ receptor has been demonstrated as playing a major role in cognitive processing and is viewed as the most plausible subtype for mediating the effects of mAChR dysfunction in psychotic disorders, including cognition disorders, neuropathic pain, and schizophrenia. − As a result, considerable effort has been put forth to develop selective M₄ agonists for the treatment of such disorders; however, attempts have fallen short due to the inability to design and develop highly selective compounds for mAChR M₄. Past shortcomings can be attributed to targeting the highly conserved orthostatic ACh binding site.

Further target validation by Eli Lilly and Co., in collaboration with Novo Nordisk, came with the development of xanomeline (an M₁/M₄ preferring agonist) which further solidified the mACh system as a mechanism for treating psychosis and behavioral disturbances observed in both schizophrenia and AD patients. , However, due to peripherally mediated cholinergic side effects which were attributed to the lack of mAChR selectivity, xanomeline’s clinical development was discontinued. To overcome these adverse events, Karuna Therapeutics (acquired by Bristol Myers Squibb) developed KarXT (Cobenfy) which was recently approved by the FDA as the first antipsychotic drug for the treatment of schizophrenia which targets the cholinergic receptors. Cobenfy is a treatment that coadministers xanomeline with trospium chloride (a peripherally restricted, pan-selective mAChR antagonist) which aids in minimizing the cholinergic adverse events observed when xanomeline is administered alone.

To circumvent issues arising from targeting the highly conserved orthostatic binding site (e.g., lack of subtype selectivity), our approach is to develop compounds that act at allosteric sites of mAChRs that are less likely to be highly conserved. Allosteric activators can include (1) allosteric agonists, which directly activate the receptor in the absence of ACh at a site removed from the orthosteric site and (2) positive allosteric modulators (PAMs), which do not activate the receptor directly but potentiate activation of the receptor by the endogenous orthosteric agonist, ACh. , It should be noted that it is possible for a single molecule to have both allosteric potentiator and allosteric agonist activity. It has been reported that a selective M₄ PAM not only demonstrated robust efficacy in preclinical models of antipsychotic-like activity and enhancement of cognition but also, and perhaps more importantly, lacked the adverse cholinergic-related side effects previously observed with xanomeline. Therefore, one strategy to improve tolerability and safety profiles is the development of receptor-subtype-selective M₄ PAMs. Cerevel Therapeutics (acquired by AbbVie) developed selective M₄ PAM CVL-231 (Emraclidine), which is undergoing clinical investigation as an adjunct treatment for schizophrenia and neurodegenerative psychosis.

Despite advances in mAChR research, there is still a scarcity of potent, efficacious, and selective activators of M₄ mAChR that are also effective in the treatment of neurological and psychiatric disorders associated with cholinergic activity and diseases in which the muscarinic M₄ receptor is involved. This paper details a recent effort to develop one such compound, VU6025733, a muscarinic M₄ PAM.

Results and Discussion

Synthesis and SAR

We began our current effort with a high-throughput screen (HTS) identifying VU0641491 and VU0641483 as weak M₄ PAMs with potencies in the micromolar range (Figure ). Following an investigation into recent research in the field, we came across a structurally similar series which incorporates a 2-methoxy-5-(piperidin-4-yloxy)­pyridine moiety. Our first step was to replace the (4-fluorophenyl)­(piperazin-1-yl)­methanone tail of VU0641491 and VU0641483 with the 2-methoxy-5-(piperidin-4-yloxy)­pyridine moiety (Figure , shown in red) while retaining our [1,2,4]­triazolo­[4,3-b]­pyridazine headgroup (Figure , shown in blue). This resulted in lead compounds VU6015863 and VU6020378, which displayed much improved human M₄ (hM₄) potency profiles when compared to predecessors VU0641491 and VU0641483 (18–52-fold). Unfortunately, this first-generation iteration suffered from many shortcomings, such as human-rat potency discrepancies typically 4–5 times less potent when screened against rat M₄ (rM₄), moderate to high predicted human hepatic clearance (CL_hep), and/or inhibition of cytochrome P₄₅₀s (CYPs).

1.

Scaffold hybridization provided a novel chemotype for the first-generation M₄ PAM analogs.

To overcome these hurdles, we devised a multidimensional optimization approach summarized in Figure . We began our investigation with the substitution of the 2-methoxypyridine group of the ether linkage to generate our second generation of analogs (Table ). These analogs were synthesized according to Scheme . In general, commercial alcohol 1 was converted into mesylate 2 and followed by nucleophilic substitution with various commercial alcohols and subsequent Boc-deprotection yielded intermediates 3. The 6-chloro-[1,2,4]­triazolo­[4,3-b]­pyridazines 4 could then undergo nucleophilic aromatic substitution (S_NAr) with free amines 3 to give final compounds 5 or 6. Select analogs were screened against hM₄ to determine PAM activity with results highlighted in Table . The 4-fluoro benzene derivatives (5a: hM₄ EC₅₀ = 152 nM; 6a: hM₄ EC₅₀ = 131 nM), while highly potent in relation to hM₄, still suffered from a ∼3–4-fold human-rat M₄ discrepancy (5a: rM₄ EC₅₀ = 605 nM; 6a: rM₄ EC₅₀ = 324 nM). It also became apparent that the location of the fluoro-substituent was greatly important, as analogs with a 3-fluorophenyl were well tolerated (5b, 6b, 5d, and 6d; hM₄ EC₅₀ ∼240–330 nM) as opposed to 2-fluorophenyl analogs which displayed a great reduction in potency (5c, 6c, 5e, and 6e; hM₄ EC₅₀ > 1.4 μM). Both the m-methylphenyl (5h: hM₄ EC₅₀ = 160 nM, rM₄ EC₅₀ = 575 nM; and 6h: hM₄ EC₅₀ = 159 nM, rM₄ EC₅₀ = 482 nM) and p-methylphenyl (5g: hM₄ EC₅₀ = 324 nM, rM₄ EC₅₀ = 893 nM; and 6g: hM₄ EC₅₀ = 280 nM, rM₄ EC₅₀ = 985 nM) were tolerated; however, the m-methylphenyl derivatives were ∼2-fold more potent in both rM₄ and hM₄. Interestingly, introduction of a nitrogen meta to the ether linkage to generate pyridines 5i, 6i, 5j, and 6j (similar to lead compounds VU6015863 and VU6020378) led to diminished activity; however, this phenomena was less pronounced when a 7,8-dimethyl-[1,2,4]­triazolo­[4,3-b]­pyridazine headgroup (R¹ = Me) was employed (6g vs 6j; ∼1.3-fold loss of activity) versus when a 7-methyl-[1,2,4]­triazolo­[4,3-b]­pyridazine headgroup (R¹ = H) was present (5g vs 5j; ∼2.7-fold loss of activity). Even more detrimental was the pyridine analog in which the nitrogen was ortho to the ether linkage (5k), resulting in significant loss of activity. Additional SAR revealed that exchanging the p-methylphenyl (5g and 6g) moiety with 4-trifluormethylphenyl resulted in a loss of potency (5f and 6f), although, once again, this phenomenon was less severe when a 7,8-dimethyl -[1,2,4]­triazolo­[4,3-b]­pyridazine headgroup was utilized (6f). Other R³ groups that provided low nanomolar compounds included the benzo­[d]­thiazole analogs (5l: hM₄ EC₅₀ = 80 nM; and 6l: hM₄ EC₅₀ = 68 nM) and the 1-methyl-1H-indazole analogs (5m: hM₄ EC₅₀ = 216 nM; and 6m: hM₄ EC₅₀ = 180 nM); however, these compounds were not pursued as they were shown to inhibit a multitude of CYP enzymes including CYP2C9, CYP2D6, and CYP3A4 (Table ). Moreover, the benzo­[d]­thiazole analogs also displayed moderate to high predicted human hepatic clearance. We also examined naphthalene as a substitute to the 2-methoxypyridine group which resulted in a loss of hM₄ PAM potency (5o and 6o: hM₄ EC₅₀s > 1.1 μM). Intriguingly, introducing flexibility into the bicycle to give the 1,2,3,4-tetrahydronaphthalene analogs 5p and 6p enhanced potency by 2–2.5-fold. Further exploration led to the discovery of the 2,3-dihydrobenzo­[b]­[1,4]­dioxine analogs 5q (hM₄ EC₅₀ = 134 nM, rM₄ EC₅₀ = 48 nM) and 6q (hM₄ EC₅₀ = 38 nM, rM₄ EC₅₀ = 95 nM), both of which displayed low nanomolar potencies in hM₄ and rM₄. The addition of an extra methyl group on the triazolopyridazine headgroup (R¹ = Me) of 6q provided us with an analog with an EC₅₀ < 100 nM in both human and rat M₄ as well as improved rat and human predicted CL_hep (Table ). Unfortunately, 6q suffered from a short in vivo elimination half-life (t _1/2= 0.6 h), suggesting potential extrahepatic clearance mechanisms.

2.

Library optimization strategy to improve M₄ PAM potency as well as DMPK properties.

1. SAR of Second Generation M₄ PAM Analogs 5 or 6 .

^aCalcium mobilization assays with hM_4/Gqi5-CHO cells performed in the presence of an EC₂₀ fixed concentration of acetylcholine; n ≥ 1 independent experiment in triplicate.

1. Synthesis of M₄ PAM Aryl-Ethers 5 and 6 .

^a Reagents and conditions: (a) MsCl, TEA, DCM, 0 °C, 2 h, 99%; (b) i. ArOH, Bu₄NCl, K₂CO₃, H₂O, 100 °C; ii. TFA, DCM, 1–18 h, 10–86% (2 steps); (c) PPh₃, DBAD, THF, 18 h, 8–77%; (d) 4, DIEA, NMP, 175 °C, 18 h, 4–73%; (e) DIEA, MeCN, 85 °C, 4 d, 39–54%.

2. In Vitro Rat and Human Hepatic Clearance, Human M₂ Selectivity, and CYP Inhibition Data for Select Analogs.

	CL hep (mL/min/kg)				CYP 450 IC 50 (μM)
Cmpd.	Human	Rat	rM 4 EC 50 (nM) (%Ach)	hM 2 EC 50 (μM) (%Ach)	1A2	2C9	2D6	3A4
5l	14	29	n.d.	Inactive	>30	3.6	4.4	9.3
6l	13	42	n.d.	Inactive	27.9	16.4	>30	5.1
5m	5.3	25	740 (96)	Inactive	>30	4.0	9.2	14.2
6m	10	29	271 (84)	Inactive	>30	>30	>30	12.8
5q	11	58	48 (93)	1.38 (52)	>30	6.1	>30	20.3
6q	7.3	46	95 (81)	Inactive	>30	>30	9.2	25.1
29b	2.4	55	231 (70)	0.67 (52)	>30	4.5	>30	9.3
29c	8.8	32	82 (73)	Inactive	>30	20.8	>30	16.7
29d	2.1	46	81 (80)	>10 (43)	>30	7.8	6.1	>30
29e	5.4	37	96 (85)	Inactive	>30	12.6	3.6	10.8
29j	5.8	41	100 (44)	4.00 (27)	>30	15.1	1.3	16.9
33h	8.5	40	70 (80)	2.99 (54)	>30	>30	2.3	11.4
33i	2.2	29	80 (68)	>10 (41)	>30	>30	2.9	20.4
33j	10	37	130 (85)	>10 (47)	27	>30	5.2	12.3
33k	7.5	11	92 (87)	>10 (33)	>30	>30	5.1	29
33n	9.1	42	188 (66)	Inactive	>30	>30	>30	29.5
33o	7.9	33	154 (72)	Inactive	>30	28.6	>30	20.7
33p	9.9	34	49 (73)	Inactive	>30	26.1	8.1	19.1
33q	2.1	34	62 (77)	Inactive	>30	>30	8.4	19.3
33r	0.36	43	55 (53)	1.32 (55)	>30	7.7	10.2	1.2
39b	9.9	44	233 (72)	5.59 (40)	>30	29.5	>30	>30
39d	12	51	214 (101)	Inactive	>30	5.9	>30	18.7
39i	1.7	29	289 (96)	Inactive	>30	3.1	4.0	4.7
39n	20	60	240 (59)	Inactive	>30	>30	>30	>30
39r	3.4	45	348 (96)	Inactive	>30	29.7	13.8	19.6

^aCalcium mobilization assays with hM_4/Gqi5-CHO cells performed in the presence of an EC₂₀ fixed concentration of acetylcholine, n ≥ 1 independent experiment in triplicate. (n.d. = not determined).

To improve half-life, we devised a strategy for our third generation of analogs in which we modified the 2,3-dihydrobenzo­[b]­[1,4]­dioxine ring. These analogs were synthesized according to Schemes and . We first began with the synthesis of the various 2,3-dihydrobenzo­[b]­[1,4]­dioxin-6-ols as highlighted in Scheme . Aldehyde 7 was first treated with 1,2-dibromoethane-1,1,2,2-d ₄, then subsequently oxidized with mCPBA to yield intermediate 8. Intermediate 8 was further modified by first protecting the alcohol as the THP-ether followed by selective bromination using 1,2- dibromotetrafluoroethane and n-BuLi. Deprotection of the THP-ether yielded intermediate 9. Starting diols 13 and 16 were likewise treated with 1,2-dibromoethane-1,1,2,2-d ₄ to yield intermediates 14 and 17, respectively. Bromides 14 and 17 were then converted into their respective pinacol boranes which were further converted into alcohols 15 and 18 via oxidation-hydrolysis. Starting alcohol 10 was first alkylated with 1-bromopropan-2-ol then converted into intermediate 11 via an intramolecular Pd-catalyzed carbon–oxygen bond formation. Methyl ester 11 was then reduced with LAH to the benzyl alcohol followed by Dess-Martin oxidation to yield the corresponding aldehyde which was then converted into alcohol 12 in a similar manner as intermediate 8. Final analogs could then be synthesized according to Scheme . With alcohols 8, 12, 15, and 18 in hand as well as commercially available 2,2,3,3-tetrafluoro-6-hydroxybenzodioxene, we could easily generate piperidines 20–24 via substitution followed by Boc-deprotection. Intermediate 23 (R⁵ = Me) was then purified by supercritical fluid chromatography (SFC) to yield enantiomerically pure material which was then carried forward. Alcohol 9 reacted with mesylate 19 to afford bromide 25. Intermediate 25 then underwent either a Pd-catalyzed cyanation or Suzuki coupling reaction followed by Boc-deprotection to afford piperidines 26–28. All piperidines reacted with chloride 4 to give the S_NAr products 29. Analogs were screened against hM₄ with results highlighted in Table .

2. Synthesis of Modified 2,3-Dihydrobenzo­[b]­[1,4]­Dioxin-6-ol Intermediates .

^a Reagents and conditions: (a) K₂CO₃, acetone, 60 °C, 18 h, 60%; (b) mCPBA, K₂CO₃, DCE, 50 °C, 18 h; then K₂CO₃, MeOH, rt, 3 h, 80%; (c) DHP, PPTS, DCM, rt, 80%; (d) 1,2-dibromotetrafluoroethane, n-BuLi, THF, −50 °C, 1 h, 82%; (e) HCl, THF, rt, 1 h, 93%; (f) KI, K₂CO₃, DMF, 80 °C, 72 h, 98%; (g) t-BuXPhos, Pd­(OAc)₂, Cs₂CO₃, toluene, 110 °C, 18 h, 37%; (h) LAH, THF, 0 °C to rt, 1 h, 88%; (i) DMP, DCM, 4 h, rt, 61%; (j) Br₂, NaOAc, chloroform, 0 °C to rt, 67%; (k) B₂pin₂, KOAc, Pd₂(dppf)­Cl₂·CH₂Cl₂, DMF, 90 °C, 18 h, 63%; (l) NaOH, H₂O₂, THF, 0 °C to rt, 1 h, 38%.

3. Synthesis of M₄ PAM Analogs 29 .

^a Reagents and conditions: (a) 18, Bu₄NCl, K₂CO₃, DMF/H₂O, 100 °C, 18 h, 40–60%; (b) TFA, DCM, rt, 1 h; 67 −85%; (c) trimethylboroxine or cyclopropyl boronic acid, Pd₂(dppf)­Cl₂·Cs₂CO₃, 1,4-dioxanes, 80 °C, 2 h, 36–47%; (d) Zn­(CN)₂, Pd­(PPh₃)₄, DMF, microwave irradiated at 140 °C, 0.5 h, 77%; (e) DIEA, NMP, 175 °C, 18 h, 20–49%.

3. SAR of Third Generation M₄ PAM Analogs 29 .

^aCalcium mobilization assays with hM4/Gqi5-CHO cells performed in the presence of an EC₂₀ fixed concentration of acetylcholine; n ≥ 1 independent experiment in triplicate. *Single enantiomer of unknown absolute configuration.

Substituting the 2,3-dihydrobenzo­[b]­[1,4]­dioxine ring to give the tetrafluoro derivative 29a resulted in a loss of potency (hM₄ EC₅₀ = 3.4 μM). The two 2-methyl-2,3-dihydrobenzo­[b]­[1,4]­dioxine enantiomers (29b and 29c) provided analogs that were potent on both hM₄ and rM₄. Upon comparison, one enantiomer (29b: R¹ = H) displayed low human CL_hep (2.4 mL/min/kg) but unfortunately exhibited hM₂ activity (EC₅₀ = 671 nM) (Table ). Conversely, the other enantiomer (29c: R¹ = Me) was inactive on hM₂, but was determined to have a less desirable human CL_hep (8.8 mL/min/kg). It can be hypothesized that the loss of hM₂ activity is due the presence of the additional methyl group on the [1,2,4]­triazolo­[4,3-b]­pyridazine ring (R¹ = Me), as loss of hM₂ activity is also observed with analog 29d versus 29e, in which the only point of difference is the additional 7-methyl group. We believe moderate human CL_hep can be attributed to the 2-methyl-2,3-dihydrobenzo­[b]­[1,4]­dioxine ring as opposed to the additional methyl on the [1,2,4]­triazolo­[4,3-b]­pyridazine ring. This is supported by the fact that the additional methyl did not prove unfavorable in regard to human CL_hep of compound 29e versus 29d. In fact, the additional methyl group on the [1,2,4]­triazolo­[4,3-b]­pyridazine ring of 6q (hCL_hep = 7.3 mL/min/kg; rCL_hep = 30 mL/min/kg) versus 5q (hCL_hep = 11 mL/min/kg; rCL_hep = 58 mL/min/kg) improved both rat and human predicted hepatic clearance values.

Replacement of the hydrogens on the 2,3-dihydrobenzo­[b]­[1,4]­dioxine ring with deuterium (29d and 29e) did not greatly affect hM₄ potencies when compared to 5q and 6q; however, this modification yielded compounds with low human predicted hepatic clearance (29d: CL_hep = 2.14 mL/min/kg; 29e: CL_hep = 5.4 mL/min/kg), which was an improvement in comparison to 5q and 6q. More importantly, this modification improved the poor elimination half-life observed for 6q (t _1/2 = 0.6 h) to the more desirable half-life of 29e (t _1/2 = 8.8 h) but, unfortunately, we also observed increased CYP2D6 inhibition (IC₅₀ = 3.6 μM). Further analysis revealed that substituting the aryl ring of the 2,3-dihydrobenzo­[b]­[1,4]­dioxine at the 5-position was detrimental to hM₄ activity (29f, 29g, 29h, and 29i). While fluorine was at least tolerated, larger groups at the 5-position greatly reduced potency and even afforded inactive compounds. Alternatively, substituting the aryl ring of the 2,3-dihydrobenzo­[b]­[1,4]­dioxine at the 8-position (29j) resulted in human M₄ EC₅₀’s < 100 nM as well as low human CL_hep (5.8 mL/min/kg) and moderate rat CL_hep (41 mL/min/kg). This modification, however, resulted in even greater CYP2D6 inhibition (IC₅₀ = 1.3 μM).

With the improved half-life of 29e, we turned our attention to rectifying the CYP inhibition profile. The fourth generation of analogs focused on alteration to the core piperidine ring (Table ). Piperidine replacements 33a–33g were synthesized in a straightforward manner from the commercially available Boc-protected amine alcohols in a similar manner as depicted in Scheme . Ring expansion of the piperidine (33a and 33b) as well as ring contraction (33c and 33g) afforded analogs that are either less potent than the parent piperidine compound (5q or 6q) or inactive when screened against hM₄. Piperidine bioisosteres 33e and 33f also proved to be significantly less potent than the original piperidine analog.

4. SAR of Fourth Generation M₄ PAM Analogs 33 .

^aCalcium mobilization assays with hM_4/Gqi5-CHO cells performed in the presence of an EC₂₀ fixed concentration of acetylcholine; n ≥ 1 independent experiment in triplicate.

As deviation from the piperidine core was detrimental to the potency of our compounds, we focused our attention on generating substituted piperidine cores (33h–33r). Scheme shows a representative synthetic route by which analogs 33h–33o were all synthesized. In general, commercially available Boc-protected amines underwent a Mitsunobu reaction with phenol 8 or 8’ and, following Boc-deprotection, afforded amines 31 or 31’. Following our standard S_NAr conditions with aryl chloride 32, final compounds 33h–33o were generated. Scheme details a representative process to synthesized analogs 33p–33r. In short, commercially available 1-boc-4-piperidone (34) was reduced with sodium borodeuteride. The resulting alcohol was then tosylated with TsCl to afford intermediate 35. Following a substitution reaction with alcohol 8 or 8’ then subsequent Boc-deprotection yielded amines 36 or 36’. This newly synthesized piperidine-4-d intermediate was then subjected to our standard S_NAr conditions with aryl chloride 32 to afford final compound 33p and 33q.

4. Synthesis of M₄ PAM Analogs 33h and 33i .

^a Reagents and conditions: (a) ADDP, P­(n-Bu)₃, toluene, 80 °C, 91%; (b) TFA, DCM, rt, 1h; 44%; (c) 32, DIEA, NMP, 175 °C, 18 h, 56%.

5. Synthesis of M₄ PAM Analog 33p and 33q .

^a Reagents and conditions: (a) NaBD₄, MeOH, 0 °C to rt, 4 h, 99%; (b) TsCl, DMAP, pyridine, 18 h, 82%; (c) 20, Bu₄NCl, K₂CO₃, DMF/H₂O, 100 °C, 18 h; (d) TFA, DCM, rt, 1 h; 46% (2 steps); (e) 32, DIEA, NMP, 175 °C, 18 h, 60%.

All fluoropiperidine analogs tested were highly potent against hM₄ with EC₅₀ s < 200 nM; however, the fluorinated (3S, 4S)-trans isomer (33l and 33m) was the least favorable isomer in relation to hM₄ potencies (EC₅₀ > 140 nM) (Table ). Conversely, the fluorinated (3R,4R)-trans isomers 33n and 33o were ∼2–4-fold more potent when screened for hM₄ activity with moderate predicted human hepatic clearance (33o, CL_hep = 7.9 mL/min/kg) in comparison to its nonfluorinated counterpart which displayed low predicted human clearance (29e, CL_hep = 5.4 mL/min/kg) as shown in Table . By comparison, both fluorinated cis-isomers displayed ≤100 nM potencies against hM₄. While the fluorinated (3S, 4R)-cis isomers 33h and 33i improved CYP2C9 inhibition (IC₅₀’s = 30 μM) compared to 29e (IC₅₀ = 12.6 μM), both analogs suffered from an increase in CYP2D6 inhibition (33h: IC₅₀ = 2.3 μM and 33i IC₅₀ = 2.9 μM) when compared to analog 29e (IC₅₀ = 3.6 μM). Additionally, the fluorinated (3S, 4R)-cis isomer resulted in a drastic decrease in the elimination half-life (33i, t _1/2 = 0.83 h) when compared to 29e (t _1/2 = 8.8 h). Once again, the tetradeutero substitution on the 2,3-dihydrobenzo­[b]­[1,4]­dioxine ring displayed a trend of improved predicted hepatic clearance of 33i (rCL_hep = 28.7 mL/min/kg; hCL_hep = 2.2 mL/min/kg) when compared to the nondeutero analog, 33h (rCL_hep = 40.4 mL/min/kg; hCL_hep = 8.5 mL/min/kg). This trend was also observed with the (3R, 4S)-cis isomers 33k rCL_hep = 11.1 mL/min/kg; hCL_hep = 7.5 mL/min/kg and 33j (rCL_hep = 37 mL/min/kg; hCL_hep = 10 mL/min/kg). Moreover, the fluorinated (3R, 4S)-cis isomers 33j and 33k modestly improved CYP2D6 inhibition (33j: IC₅₀ = 5.2 μM and 33k: IC₅₀ = 5.1 μM) when compared to analog 29e (IC₅₀ = 3.6 μM). Fluorination of the piperidine ring also improved predicted rat hepatic clearance (33k; rCL_hep = 11.1 mL/min/kg; hCL_hep = 7.5 mL/min/kg) when compared to the nonfluorinated analog 29e (rCL_hep = 36.8 mL/min/kg; hCL_hep = 5.4 mL/min/kg) while having minimal effect on predicted human hepatic clearance. Moreover, the fluorinated (3R, 4S)-cis isomer also reduced the elimination half-life of our molecule (33k, t _1/2 = 2.66 h), although to a lesser extent than the (3S, 4R)-cis isomer.

The most profound effect was noticed when a deuterium was incorporated into the piperidine ring (33p and 33q; Tables and ). This modification not only improved CYP2D6 inhibition (33p: IC₅₀ = 8.2 μM and 33q: IC₅₀ = 8.4 μM) but also CYP2C9 and CYP3A4 inhibition (IC₅₀s > 19 μM) in comparison to 29e (CYP2D6 IC₅₀ = 3.6 μM; CYP2C9 IC₅₀ = 12.6 μM; CYP3A4 IC₅₀ = 10.8 μM) while maintaining hM₄ (33p: EC₅₀ = 25 nM 33q: EC₅₀ = 33 nM) and rat potency (33p: EC₅₀ = 49 nM 33q: EC₅₀ = 62 nM). Once again, we noticed the tetradeutero substitution on the 2,3-dihydrobenzo­[b]­[1,4]­dioxine ring benefited human predicted hepatic clearance (33q, hCL_hep = 2.14 mL/min/kg) when compared to the nondeutero analog 33p (hCL_hep = 9.9 mL/min/kg). Interestingly, we observed a less than desirable CYP profile upon deletion of the 7-methyl (R¹) of the 7,8-dimethyl-[1,2,4]­triazolo­[4,3-b]­pyridazine headgroup (33r). The most profound effect was in relation to CYP3A4 (IC₅₀ = 1.2 μM) as compared to the corresponding dimethyl analog 33q (IC₅₀ = 19.3 μM).

Finally, we shifted our focus toward modifications of the 7,8-dimethyl-[1,2,4]­triazolo­[4,3-b]­pyridazine headgroup to generate our fifth generation of analogs, 38 and 39. To begin, we evaluated analogs containing historical head groups we have employed in the past when designing M₄ PAMs. Heteroaryl bromides (37) underwent Buchwald-Hartwig aminations to afford analogs 38a-I (Scheme ). Disappointingly, as highlighted in Table , this approach proved unfruitful as these analogs were either inactive or showed micromolar activity when screened for activity against hM₄. Undeterred by these results, we turned our attention to the synthesis of novel head groups that more closely resemble the original 7,8-dimethyl-[1,2,4]­triazolo­[4,3-b]­pyridazine headgroup or were direct modifications thereof. Briefly, heteroaryl chlorides 37 were subjected to standard S_NAr conditions to generate analogs 39 with results of this endeavor showcased in Table . The importance of the nitrogen at the 2-position of the [1,2,4]­triazolo­[4,3-b]­pyridazine ring was apparent as deletion of this nitrogen (39a) led to a 5.5-fold decrease in hM₄ potency. Replacement of the original headgroup with a 2,7-dimethyl-[1,2,4]­triazolo­[1,5-a]­pyridine ring (39b) led to a slight decrease in potency in hM₄ (2-fold). Additionally, this motif was detrimental to hCL_hep (9.9 mL/min/kg) and revealed hM₂ activity (5.6 μM; Table ). In general, substitution of the 7,8-dimethyl-[1,2,4]­triazolo­[4,3-b]­pyridazine ring at the 3-position (39c – 39e) resulted in a loss of hM₄ potency. Interestingly, the difluoromethyl analog 39d (hM₄ EC₅₀ = 78 nM) was ∼9-fold more potent than the corresponding trifluoromethyl analog 39c (hM₄ EC₅₀ = 699 nM); however, 39d was still over 2-fold less potent than the parent compound 33q (hM₄ EC₅₀ = 35 nM). Moreover, the difluoromethyl substitution had an undesirable effect on the hCL_hep (12 mL/min/kg).

6. Synthesis of M₄ PAM Analogs 38 and 39 .

^a Reagents and conditions: (a) Pd₂(dba)₃, Xantphos, Cs₂CO₃, 1,4-dioxanes, 100 °C, 13–90%; (b) ^t BuXPhos Pd G1, ^t BuXPhos, ^t BuONa, ^t BuOH, 50 °C (16–41%); (c) DIEA, NMP, 175 °C, 10–95%; (d) Pd₂(dba)₃, rac-BINAP, ^t BuONa, toluene, 120 °C, 12%; (e) Pd₂(dba)₃, Xantphos, RuPhos Pd G3, Cs₂CO₃, 1,4-dioxanes, 100 °C, 33%.

5. SAR of Analogs 38 Containing Historical M₄ PAM Head Groups .

^aCalcium mobilization assays with hM_4/Gqi5-CHO cells performed in the presence of an EC₂₀ fixed concentration of acetylcholine, n = 1 in triplicate. ^b Synthesized according to Scheme , condition a; ^c Synthesized according to Scheme , condition b.

6. SAR of Fifth Generation M₄ PAM Analogs 39 .

^aCalcium mobilization assays with hM_4/Gqi5-CHO cells performed in the presence of an EC₂₀ fixed concentration of acetylcholine; n ≥ 1 independent experiment in triplicate. ^bSynthesized according to Scheme , condition c; ^csynthesized according to Scheme , condition d; ^dsynthesized according to Scheme , condition e.

Varying the substitutions at the 7 or 8-positions of the [1,2,4]­triazolo­[4,3-b]­pyridazine ring consistently produced less active compounds (39f–39m and 39o–39s). The importance of the 8-methyl (R²) of the 7,8-dimethyl-[1,2,4]­triazolo­[4,3-b]­pyridazine ring was demonstrated by 7-methyl analog 39o (hM₄ EC₅₀ = 869 nM) which was 6.5-fold less potent when compared to the 8-methyl analog 5q (hM₄ EC₅₀ = 134 nM) and 23-fold less potent when compared to the 7,8-dimethyl analog 6g (hM₄ EC₅₀ = 38 nM). More surprising was the ∼23-fold loss in activity when 39o was compared to the corresponding 7,8-dimethyl analog 6q (hM₄ EC₅₀ = 38 nM). Intriguingly, introducing larger groups at the 7-position, such as an ethyl (39h, hM₄ EC₅₀ = 717 nM) or diethylamine (39j, hM₄ EC₅₀ = 362 nM), resulted in a 1.2–2.4-fold increase in hM₄ activity, respectively, when compared to 39o. We postulate that bulkier groups at the 7-position are extending into a hydrophobic pocket originally occupied by the 8-methyl group of the 7,8-dimethyl-[1,2,4]­triazolo­[4,3-b]­pyridazine headgroup. This trend is also observed when comparing analogs 39p and 39q; the presence of the 8-methyl group yields a 17.6-fold more active analog. Replacing the 8-methyl group of analog 33r (hM₄ EC₅₀ = 23 nM) with a larger ethyl group yielded analog 39i (hM₄ EC₅₀ = 109 nM). Although 39i is ∼4.7-fold less potent than 33r, it still displayed low nanomolar activity; however, further evaluation revealed that this modification was a detriment to the CYP profile (CYP2C9 IC₅₀ = 3.1 μM; CYP2D6 IC₅₀ = 4 μM; CYP3A4 IC₅₀ = 4.7 μM) as shown in Table .

A more in-depth investigation into varying the substitutions at the 7 and 8-positions of the [1,2,4]­triazolo­[4,3-b]­pyridazine ring revealed a loss in potency of all analogs tested (39f, 39g, 39p, 39r, and 39s). Simply substituting the 7-methyl group of 6q with a trifluoromethyl-group to yield 39p led to a 5.3-fold decrease in hM₄ activity. When compared to parent analog 33q (hM₄ EC₅₀ = 33 nM), the cyclopropyl derivatives 39f and 39g also exhibited 5.7-fold and 6.2-fold losses in hM₄ activity, respectively. Likewise, analogs 39r and 39s displayed a 3.8-fold and 31-fold loss in hM₄ activity, respectively, when compared to parent analog 6q (hM₄ EC₅₀ = 38 nM). While the 7-cyclopropyl analog (39f) and the 8-cyclopropyl analog (39g) were nearly equipotent to one another, the 8-methoxymethyl derivative (39r) was nearly 8.2-fold more potent than the 7-methoxymethyl compound (39s). Attempts to “tie-back” the 7-methyl and 8-methyl into a tricyclic ring system (39l and 39m) also resulted in a loss of hM₄ potency. Additionally, deviation from the [1,2,4]­triazolo­[4,3-b]­pyridazine core, in general, afforded analogs displaying a loss of hM₄ activity (39t, 39u, and 39v) with the exception of 39n (hM₄ EC₅₀ = 43 nM). Although the most active analog of our fifth generation series, compound 39n suffered from a human-rat M₄ potency discrepancy (4.4-fold, rM₄ EC₅₀ = 188 nM).

Molecular Pharmacology and DMPK Profiling

Using the data summarized in Table , we rapidly deprioritized compounds with moderate to high predicted hepatic clearance (rat or human), potency discrepancies between species, rat potency (rM₄ EC₅₀ ≥ 200 nM), lack of hM₂ selectivity, and/or less desirable CYP450 inhibition. As a result, only one compound stood out as a compound of interest for further profiling; thus, compound 33q (VU6025733) was carried forward and evaluated for muscarinic selectivity as well as further in vitro and in vivo DMPK profiling (Table ). Regarding physicochemical properties, VU6025733 possesses an attractive molecular weight of <400 Da as well as a desirable CNS xLogP (2.99) and tPSA (74 Ų). When screened against other subtypes of muscarinic acetylcholine receptors (M₁, M₂, M₃, and M₅), VU6025733 displayed high receptor subtype selectivity as it was inactive on both the human and rat isoforms of all other subtypes. Moreover, VU6025733 exhibited no appreciable species differences in M₄ activity between human and rat (∼2-fold) (Tables and ). VU6025733 demonstrated acceptable CYP₄₅₀ profiles against CYP1A2, CYP2C9, and CYP3A4 (IC₅₀s ≥ 19.3 μM, > 500-fold selectivity) as well as CYP2D (IC₅₀s = 8.4 μM, 240-fold selectivity). VU6025733 displayed low human (CL_hep = 2.14 mL/min/kg) and moderate rat (CL_hep = 34 mL/min/kg) hepatic clearance based on microsomal intrinsic clearance (CL_int). VU6025733 had moderate fraction unbound in rat and human plasma (f _u,plasma = 0.010 and f _u,plasma = 0.051, respectively) and moderate rat brain homogenate binding (f _u,brain = 0.016). Next, we assessed in vitro brain

7. Muscarinic Selectivity Data and DMPK Analysis for VU6025733 (33q)­ .

Property	33q VU6025733
MW (g/mol)	386.5
xLogP	2.99
TPSA (Å2)	74
Muscarinic selectivity
Human M1, M2, M3, M5	Inactive
Rat M1, M2, M3, M5	Inactive
In vitro PK parameters
CLint (mL/min/kg), rat	66
CLhep (mL/min/kg), rat	34
CLint (mL/min/kg), human	2.3
CLhep (mL/min/kg), human	2.1
Rat f u,plasma	0.010
Human f u,plasma	0.051
Rat f u,brain	0.016
Brain distribution (0.25 h) (SD Rat; 0.2 mg/kg IV)
K p, brain:plasma	0.39
K p,uu, brain:plasma	0.78
Rat IV PK
t 1/2 (hr)	5.67
MRT (hr)	3.83
CLp (mL/min/kg)	5.26
Vss (L/kg)	1.21
Rat PO PK
T max (hr)	0.75
C max (ng/mL)	1,783
AUC0‑∞ (hr·ng/mL)	6,753
%F	74.1

^aCalcium mobilization assay; values are an n ≥ 1 independent experiments in triplicate.

^b f _u = Fraction unbound; equilibrium dialysis assay; brain = rat brain homogenates.

^c K _p = total brain to total plasma ratio.

^d K _p,uu = unbound brain (brain f _u × total brain) to unbound plasma (plasma f _u × total plasma) ratio.

^eMale Sprague–Dawley rats (n = 2); vehicle = 10% EtOH, 40% PEG 400, 50% saline (1 mL/kg); dose = 1 mg/kg.

^fMale Sprague–Dawley rats (n = 2); vehicle = 0.1% Tween-80, 0.5% methyl cellulose, 99.4% water (10 mL/kg); dose = 3 mg/kg.

penetration potential utilizing MDCKII-MDR1 transfected cells. VU6025733 exhibited an efflux ratio (ER) of 0.80 and a P _app (A-B) of 12.7 × 10^–6 indicating high brain penetration and lack of P-glycoprotein 1 (P-gp) efflux transport. Furthermore, VU6025733 was administered in a rat PBL IV cassette study to determine the plasma/brain partition ratio. This analysis revealed our candidate showed a K _p = 0.39 and a K _p,uu = 0.78. Additionally, our candidate showed low plasma clearance in rat (CL_p = 5.26 mL/min/kg) with an acceptable volume of distribution (V _ss = 1.21 L/kg), and a desirable half-life of 4.8 h. When VU6025733 was administered to rats at a PO dose of 3 mg/kg, our candidate displayed moderate to high oral bioavailability (%F = 74.1) with rapid absorption and low interanimal variability.

Behavioral Pharmacology

With VU6025733 in hand, we evaluated this compound in a preclinical model of antipsychotic-like activity utilizing VU0467154 as a positive comparator. , VU6025733 demonstrated a robust dose-dependent blockade of amphetamine-induced hyperlocomotion (AHL) after oral administration following a 30 min pretreatment interval in rats (MED = 10 mg/kg, Figure ). At the end of the study, brain:plasma K_ps and K_p,uus were determined at all dose groups (K_ps = 0.25–0.35; K_p,uus = 0.39–0.44) with mean C _{brain,unbound} ranging from 17.4 ng/g (10 mg/kg) to 50.5 ng/g (30 mg/kg) (Table ). These data were in alignment with our PBL cassette data. Given the promising profile of VU6025733 thus far, the compound was progressed toward a battery of genotoxicity and multiparametric cytotoxicity assays.

3.

Systemic PO administration of VU6025733 (33q) blocked amphetamine-induced hyperlocomotion in male Sprague–Dawley rats. (A) The time course of locomotor activity and (B) Total locomotor activity during the 55 min period following amphetamine administration. Data are means ± SEM of 7–8 animals per group. *p < 0.05, **p < 0.01, ***p < 0.001 vs Vehicle + Amphetamine. Vehicle = 10% Tween 80 in H₂O. VU0467154 is a positive control.

8. Relationship between Total (Mean C _brain) and Unbound (Mean C _brain,u) Brain Concentrations of VU6025733 (33q) and Pharmacodynamic Effects on Amphetamine (0.75 mg/kg , SC)-Induced Hyperlocomotion in Rats at 1.5 h.

Dose (mg/kg)	Mean reversal of AHL (%)	Mean C plasma (ng/mL)	Mean C plasma,u (ng/mL)	Mean C brain (ng/g)	Mean C brain,u (ng/g)	Brain:plasma mean K p	Brain:plasma mean K p,uu
10	25.6	3,974	39.7	1,086	17.4	0.27	0.44
15	34.0	6,463	64.6	1,587	25.4	0.25	0.39
30	39.7	11,477	115	3,157	50.5	0.28	0.44

^aAt 1.5 h postadministration.

^bEstimated unbound plasma concentration based on the rat f _u,plasma (0.010).

^cEstimated unbound brain concentration based on the rat f _u,brain (0.016).

Cytotoxicity and Toxicology Profile

With VU6025733 displaying an attractive profile thus far, attention turned to assessing its viability as a development candidate. To assess potential cardiotoxicity, VU6025733 was evaluated in a hERG SyncroPatch assay and was determined to have an IC₅₀ of 4.6 μM, which was considered concerning, as human exposure projections were 1.06 μM, providing a narrow 4.4x margin. In genotoxicity assays, VU6025733 was negative in both AMES (5 strain with and without S9) as well as negative in the in vitro micronucleus assay. PAM VU6025733 was then advanced into a multiparametric cytotoxicity assay (Figure ) in HepaRG 3D spheroids. Here, VU6025733 had a very concerning profile, decreasing spheroid size, increasing oxidative stress, decreasing glutathione content (MEC = 9.7 μM, AC₅₀ = 39 μM) and decreasing cellular ATP (MEC = 4.2 μM, AC₅₀ = 6.0 μM). A follow-up evaluation of VU6025733 in HepG2 cells demonstrated a decrease in oxygen consumption rate (MEC = 2.0 μM) and an increase in extracellular acidification rate (MEC = 7.1 μM). Combined, these data indicate that VU6025733 is an electron transport chain inhibitor with only a ∼2-fold margin of the human exposure projection. Thus, there is predicted to be a very high risk of hepatotoxic side effects for VU6025733. Coupled with the narrow hERG margin, development of VU6025733 was terminated.

4.

Multiparametric cytotoxicity in HepG2 cells obtained by high content imaging using 4 separate fluorescent and/or potentiometric probes. PAM VU6025733 (33q) was broadly cytotoxic and proven to be an electron transport chain inhibitor (MEC ∼ 2 mM) with less than a 2-fold margin for human exposure projection. The graph shows a concentration range on its radials, from 0 (center) to 100 μM (edge). The lower the “lowest effective concentrations” (LEC) for each measured parameter, the larger the red area of the graph will be (contrary to the green center), suggesting an unfavorable safety profile for VU6025733. MMP = mitochondrial membrane potential.

Conclusions

In summary, hybridizing the chemical scaffolds of previously disclosed M₄ PAMs with unique scaffolds identified via an HTS (VU0641491 and VU0641483, Figure ) resulted in the discovery of novel PAMs VU6015863 and VU6020378. Further medicinal chemistry efforts identified several highly potent (hM₄ EC₅₀ < 100 nM) M₄ PAM analogs. Of these, analog VU6025733 (33q) provided a superior overall profile that supported further progression. VU6025733 not only displayed high selectivity over the other mAChRs evaluated (M_1–3,₅) but also demonstrated M₄ potency agreement between species (human and rat). Moreover, VU6025733 exhibited a low predicted hepatic clearance profile in human as well as low in vivo plasma clearance in rat. This was a considerable improvement over lead compounds VU6015863 (hCL_hep = 11 mL/min/kg) and VU6020378 (hCL_hep = 16 mL/min/kg). VU6025733 displayed moderate to high CNS distribution of unbound drug (K _p,uu = 0.78) as well as modest brain and plasma fraction unbound in rat. Not only was VU6025733 highly brain penetrant and not a substrate for the efflux transporter P-gp (ER = 0.80; a P _app (A-B) = 12.7 × 10^–6) but the compound also demonstrated an acceptable CYP inhibition profile with IC₅₀s ≥ 8.4 μM. Due to its attractive DMPK profile, VU6025733 was advanced into in vivo pharmacokinetic/pharmacodynamic (PK/PD) profiling. VU6025733 showed robust efficacy in a preclinical model of antipsychotic activity (AHL) after oral administration with an MED of 15 mg/kg. It is important to note that due to other setbacks, nonmuscarinic off-target activity, such as dopaminergic modulation, was not assessed during this study.

Finally, we evaluated the cytotoxicity and toxicology profile of VU6025733. Although VU6025733 was negative in genotoxicity assays (AMES and in vitro micronucleus), further profiling revealed a narrow safety margin in relation to hERG inhibition. Advancement into a multiparametric cytotoxicity assay indicated a VU6025733 is an electron transport chain inhibitor likely to have a very high risk of hepatotoxic side effects. For these reasons, further development of VU6025733 was discontinued. Subsequent efforts suggest the [1,2,4]­triazolo­[4,3-b]­pyridazine headgroup as the key toxicophore, details of which will be provided in due course.

Methods

General Information

All chemicals were purchased from commercial vendors and used without further purification. All NMR spectra were recorded on a 400 MHz AMX Bruker NMR spectrometer. ¹H and ¹³C chemical shifts are reported in δ values in ppm downfield with the deuterated solvent as the internal standard. Low resolution mass spectra were obtained on an Agilent 6120/6150 or Waters QDa (Performance) SQ MS with ESI source. High resolution mass spectra were obtained on an Agilent 6540 UHD Q-TOF with ESI source. Normal phase column chromatography was performed on a Teledyne ISCO CombiFlash Rf+ system. For compounds that were purified on a Gilson preparative reversed-phase HPLC, the system comprised of a 333 aqueous pump with solvent selection valve, 334 organic pump, GX 271 or GX-281 liquid hander, two column switching valves, and a 155 UV detector. Solvents for extraction, washing and chromatography were HPLC grade. All final compounds were found to be >95% pure by HPLC-MS analysis.

Synthesis

Synthesis of 6-Chloro-7,8-dimethyl-[1,2,4]­Triazolo­[4,3-b]­pyridazine (32)

3,6-Dichloro-4,5-dimethylpyridazine (1.0 g, 5.6 mmol) and potassium carbonate (79 mg, 0.56 mmol) were dissolved in THF (28 mL) before the addition of hydrazine (890 μL, 28.2 mmol) dropwise under N₂ atmosphere. The reaction mixture was heated to reflux. After 72 h, the reaction was concentrated in vacuo and used without further purification (975 mg). LRMS: C₆H₉ClN₄ [M + H]⁺ calc. mass 173.0, found 173.3. The crude residue of 3-chloro-6-hydrazineylidene-4,5-dimethyl-1,6-dihydropyridazine (975 mg, 5.6 mmol) and formic acid (1.06 mL) were added to a sealed vessel. After heating at 100 °C for 1 h, the mixture was concentrated in vacuo. The crude material was dissolved in DCM, washed with 10% aqueous K₂CO₃ and back extracted with DCM (2×). The combined organic layers were dried (MgSO₄), filtered, and concentrated in vacuo. The crude material was purified using flash chromatography on silica gel to afford the title compound (755 mg). ¹H NMR (400 MHz, CDCl₃) δ 8.89 (s, 1H), 2.67 (s, 3H), 2.36 (s, 3H). LRMS: C₇H₇ClN₄ [M + H]⁺ calc. mass 183.0, found 183.4.

tert-Butyl 4-(Tosyloxy)­piperidine-1-carboxylate-4-d (35)

To a 0 °C solution of 1-tert-butyl-4-piperidone (10 g, 50 mmol) in methanol (250 mL) was added sodium borodeuteride (3.2 mL, 100 mmol). The resulting mixture was stirred for 4 h at room temperature. The reaction was quenched with saturated NH₄Cl and extracted with EtOAc (3x). The combined organic layers were dried (MgSO₄), filtered, and concentrated. To a suspension the crude residue and 4-dimethylaminopyridine (0.6 g, 4.9 mmol) in pyridine (45 mL) was added tosyl chloride (11.8 g, 62 mmol). The mixture stirred at room temperature for 18 h. The reaction was quenched with saturated aqueous NaHCO₃ solution and extracted with EtOAc (2×). The combined organic layers were washed with water (2×), brine (2×), dried (MgSO₄), filtered, and concentrated. The crude oil was purified via normal-phase chromatography on silica gel (0–20% EtOAc/Hexanes) to provide the title compound (14.3 g). ¹H NMR (400 MHz, CDCl₃) δ 7.79 (d, J = 8.4 Hz, 2H), 7.34 (d, J = 8.0 Hz, 2H), 3.61–3.55 (m, 2H), 3.28–3.22 (m, 2H), 2.45 (s, 3H), 1.79–1.73 (m, 2H), 1.70–1.64 (m, 2H), 1.43 (s, 9H). LRMS: C₁₇H₂₄DNO₅S [M + Na]⁺ calc. mass 379.1, found 379.4.

4-((2,3-Dihydrobenzo­[b]­[1,4]­dioxin-6-yl-2,2,3,3-d ₄)­oxy)­piperidine-4-d (36)

To a round-bottom flask were added 2,2,3,3-tetradeuterio-1,4-benzodioxin-6-ol (1.0 g, 6.7 mmol), tert-butyl 4-deuterio-4-(p-tolylsulfonyloxy)­piperidine-1-carboxylate (2.0 g, 5.6 mmol), potassium carbonate (2.4 g, 16.8 mmol), and tetrabutylammonium chloride (0.31 g, 1.1 mmol) in water (25 mL) and DMF (1.3 mL). The reaction was heated at reflux for 18 h. The reaction was diluted with 3:1 CHCl₃/IPA and the layers were separated. The aqueous layer was extracted with 3:1 CHCl₃/IPA (2x) and the combined organics were washed with water, brine, then dried (MgSO₄), filtered, and concentrated. The crude oil was purified by using normal phase chromatography on silica gel (0–20% EtOAc/Hexanes) to provide the Boc-protected intermediate which was dissolved in DCM (9 mL) followed by addition of trifluoroacetic acid (2.1 mL, 28 mmol). After 1 h, the solvents were removed in vacuo. The oil was dissolved in MeOH and loaded onto SCX cartridge. The cartridge was rinsed with MeOH and 7N NH₃/MeOH solution. The solvents were removed to afford the title compound (725 mg). ¹H NMR (400 MHz, CDCl₃) δ 6.75 (d, J = 8.7, 1H), 6.45 (d, J = 2.8, 1H), 6.41 (dd, J = 8.8, 2.9 Hz, 1H), 3.21–3.15 (m, 2H), 2.87–2.81 (m, 2H), 2.07–2.00 (m, 2H), 1.78–1.72 (m, 2H). LRMS: C₁₃H₁₂D₄NO₃ [M + H]⁺ calc. mass 241.2, found 241.2.

6-(4-((2,3-Dihydrobenzo­[b]­[1,4]­dioxin-6-yl-2,2,3,3-d₄)­oxy)­piperidin-1-yl-4-d)-7,8-dimethyl-[1,2,4]­triazolo­[4,3-b]­pyridazine (33q, VU6025733)

6-Chloro-7,8-dimethyl-[1,2,4]­triazolo­[4,3-b]­pyridazine (250 mg, 1.4 mmol), 4-deuterio-4-[(2,2,3,3-tetradeuterio-1,4-benzodioxin-6-yl)­oxy]­piperidine (345 mg, 1.4 mmol), and N,N-diisopropylethylamine (0.9 mL, 5.5 mmol) were combined in NMP (7 mL) and the vial heated at 175 °C for 18 h. The reaction was passed through a syringe filter and purified by reverse phase HLPC (20–60% MeCN/0.1% aqueous TFA) to afford the title compound (361 mg). ¹H NMR (400 MHz, CDCl₃) δ 8.82 (s, 1H), 6.77 (d, J = 8.8 Hz, 1H), 6.49 (d, J = 2.7 Hz, 1H), 6.45 (dd, J = 8.7, 2.8 Hz, 1H), 3.44–3.38 (m, 2H), 3.08–3.02 (m, 2H), 2.65 (s, 3H), 2.30 (s, 3H), 2.12–2.06 (m, 2H), 1.98–1.91 (m, 2H). ¹³C NMR (101 MHz, CDCl₃) δ 160.1, 151.6, 145.0, 144.0, 138.5, 138.3, 133.8, 125.5, 117.6, 110.1, 105.8, 72.9–71.7 (m), 64.6–63.2 (m, 2C), 47.5 (2C), 30.6 (2C), 14.6, 13.8. HR-MS (Q-TOF, ES+) calc’d for C₂₀H₁₈D₅N₅O₃, 387.2187; found, 387.2190.

Molecular Pharmacology

Calcium Mobilization Assay

Compound-evoked increases to an EC₂₀ concentration of acetylcholine (ACh) in intracellular calcium were measured using Chinese hamster ovary (CHO) cells stably expressing human, rat, dog, cyno, or minipig muscarinic receptors (M₁–M₅; M₂ and M₄ cells were cotransfected with G_qi5). The stable cells were cultured in F12 medium containing 10% fetal bovine serum, 20 mM HEPES, 100 units/mL antibiotics/antimycotic, 0.5 mg/mL G418, and 0.2 mg/mL hygromycin (M₂ and M₄ G_qi5 coexpressing cells only). All reagents used were from Life Technologies (Carlsbad, CA) unless otherwise noted.

Briefly, the day before the assay, cells (15,000 cells/20 μL/well) were plated in black-walled, clear-bottomed, 384 well plates (Greiner Bio-One, Monroe, NC) in the culture medium without G418 and hygromycin, and then incubated overnight at 37 °C in the presence of 5% CO₂. The next day, calcium assay buffer (Hank’s balanced salt solution (HBSS), 20 mM HEPES, 2.5 mM probenecid, 4.16 mM sodium bicarbonate Sigma-Aldrich, St. Louis, MO) was prepared to dilute compounds, agonists, and Fluo-4-acetomethoxyester (Fluo-4-AM), fluorescent calcium indicator dye. Compounds were serially diluted 1:3 into 10-point concentration response curves in DMSO using the Bravo Liquid Handler (Agilent, Santa Clara, CA), transferred to a 384 well daughter plates using an Echo acoustic liquid handler (Beckman Coulter, Indianapolis, Indiana), and diluted in assay buffer to a 2X final concentration. The agonist plates were prepared using acetylcholine (ACh, Sigma-Aldrich, St. Louis, MO) concentrations for the EC₂₀, EC₈₀, and EC_MAX responses by diluting in assay buffer to a 5X final concentration. The 2X dye solution (2.3 μM) was prepared by mixing a 2.3 mM Fluo-4-AM stock in DMSO with 10% (w/v) pluronic acid F-127 in a 1:1 ratio in assay buffer. Using a microplate washer (BioTek, Winooski, VT), cells were washed with assay buffer 3 times to remove medium. After the final wash, 20 μL of assay buffer remained in the cell plates. Immediately, 20 μL of the 2X dye solution (final 1.15 μM) was added to each well of the cell plate using a Multidrop Combi dispenser (Thermo Fisher, Waltham, MA). After cells were incubated with the dye solutions for 50 min at 37 °C in the presence of 5% CO₂, the dye solutions were removed and replaced with assay buffer using a microplate washer, leaving 20 μL of assay buffer in the cell plate, and the cell plate allowed to incubate for 10 min at 37 °C. The compound, agonist, and cell plates were placed inside the Functional Drug Screening System 7000 (FDSS7000, Hamamatsu, Japan) to measure the calcium flux. After establishment of a fluorescence baseline for 2–3 s (2–3 images at 1 Hz; excitation, 480 ± 20 nm; emission, 540 ± 30 nm), 20 μL (2X) of test compound or vehicle was added to the cells, and the response was measured. 140 s later, 10 μL (5X) of an EC₂₀ concentration of ACh (Sigma-Aldrich, St. Louis, MO) or vehicle was added to the cells, and the response of the cells was measured. Approximately 125 s later, an EC_80 or EC_MAX concentration of ACh was added. Calcium fluorescence was recorded as fold over basal fluorescence and raw data were normalized to the maximal response to ACh. Compound-evoked increase in calcium response in the absence of ACh agonist was determined as ago activity of positive allosteric modulators. Compound-evoked increase in calcium response in the presence of ACh EC₂₀ agonist was determined as potentiator activity of positive allosteric modulator. Potency (EC₅₀) and maximum response (% ACh Max) for compounds was determined using a four-parameter logistical equation using GraphPad Prism (La Jolla, CA) or the Dotmatics software platform (Woburn, MA):

$$y=\mathrm{bottom}+\frac{\mathrm{top}-\mathrm{bottom}}{1+{10}^{(\mathrm{Log}\mathrm{EC}50-A)\mathrm{Hill}\mathrm{slope}}}$$

where A is the molar concentration of the compound; bottom and top denote the lower and upper plateaus of the concentration–response curve; HillSlope is the Hill coefficient that describes the steepness of the curve; and EC₅₀ is the molar concentration of compound required to generate a response halfway between the top and bottom.

Glossary

Abbreviations

ACh

acetylcholine

AD

Alzheimer’s disease

AHL

amphetamine-induced hyperlocomotion

BHB

brain homogenate binding

CL_hep

hepatic clearance

CL_int

intrinsic clearance

CNS

central nervous system

CYP

cytochrome P450

DMPK

drug metabolism and pharmacokinetics

FDA

food and drug administration

hM₄

muscarinic acetylcholine receptor subtype 4, human

IND

investigational new drug

mAChR

muscarinic acetylcholine receptor

MED

minimum effective dose

NHP

nonhuman primate

PAM

positive allosteric modulator

P-gp

P-glycoprotein 1

PBL

plasma-to-brain levels

PD

pharmacodynamic

PK

pharmacokinetic

PPB

plasma protein binding

rM4

muscarinic acetylcholine receptor subtype 4, rat

SAR

structure-activity relationship

μW

microwave irradiated

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acschemneuro.5c00963.

• General methods for the synthesis and characterization of key compounds and experimental details for all assays performed (DMPK, behavioral pharmacology, ancillary pharmacology) (PDF)

A.R.G., C.P., M.F.L., L.A.B., K.A.B., A.E.R., and K.J.T. performed synthetic chemistry and scaled up key compounds. V.B.L., A.L.R., H.P.C., and C.M.N. performed and analyzed molecular pharmacology. M.B. and C.K.J. performed and analyzed behavioral pharmacology experiments. T.M.B. performed and analyzed DMPK experiments. P.J.C., C.W.L., K.J.T., D.W.E., C.M.N., and C.K.J. oversaw experimental design, and K.J.T. wrote the manuscript with input from all authors. A.R.G., C.P., M.F.L., L.A.B., K.A.B., A.E.R., T.M.B., C.W.L., P.J.C., D.W.E., and K.J.T. are inventors on applications for composition-of matter patents that protect several series of M₄ positive allosteric modulators.

The authors declare the following competing financial interest(s): The authors are currently developing next generation M4 PAMs in collaboration with Neumora Therapeutics and hold patents on M4 PAMs.