Discovery of VU6016235: A Highly Selective, Orally Bioavailable, and Structurally Distinct Tricyclic M₄ Muscarinic Acetylcholine Receptor Positive Allosteric Modulator (PAM)

Abstract

Herein, we report structure–activity relationship (SAR) studies to develop novel tricyclic M₄ PAM scaffolds with improved pharmacological properties. This endeavor involved a “tie-back” strategy to replace a 5-amino-2,4-dimethylthieno[2,3-d]pyrimidine-6-carboxamide core, which led to the discovery of two novel tricyclic cores. While both tricyclic cores displayed low nanomolar potency against both human and rat M₄ and were highly brain-penetrant, the 2,4-dimethylpyrido[4′,3′:4,5]thieno[2,3-d]pyrimidine tricycle core provided lead compound, VU6016235, with an overall superior pharmacological and drug metabolism and pharmacokinetics (DMPK) profile, as well as efficacy in a preclinical antipsychotic animal model.

Introduction

Muscarinic acetylcholine receptors (mAChRs) are G protein-coupled receptors that have been linked to a variety of central nervous system functions, including cognition, motor control, and sleep–wake architecture.¹ Muscarinic acetylcholine receptor subtype 4 (M₄) positive allosteric modulators (PAMs) have gained interest as a novel therapeutic target for the treatment of behavioral and cognitive disturbances associated with both schizophrenia and Alzheimer’s disease (AD), as well as other neurological disorders, such as Parkinson’s disease (PD) and Huntington’s disease (HD).²⁻⁸ Xanomeline, an M₁/M₄-preferring agonist, was evaluated in clinical trials for the treatment of AD and schizophrenia. Unfortunately, adverse cholinergic side effects, associated with a lack of receptor subtype selectivity, halted xanomeline’s clinical development. Nevertheless, data obtained from these trials further validated the muscarinic cholinergic system as a treatment for the psychosis and behavioral disturbances observed in both Alzheimer’s and schizophrenia patients.^9,10 To overcome these side effects, Karuna Therapeutics (acquired by Bristol Myers Squibb) developed KarXT, which coadministers xanomeline with a pan-selective peripheral mAChR antagonist (trospium chloride) to combat the adverse events associated with xanomeline administration alone (Figure 1).¹¹ The New Drug Application (NDA) for KarXT was accepted for review by the FDA in late 2023.¹² Recently, a selective M₄ PAM was shown to not only be efficacious in preclinical assays but also exhibited fewer and less severe adverse cholinergic-related side effects when compared to rats treated with xanomeline.¹³ Thus, data suggest that development of a receptor subtype-selective M₄ PAM could result in improved safety profiles. Emraclidine (CVL-231), a selective M₄ PAM developed by Cerevel Therapeutics, is currently undergoing clinical trials (Figure 1).¹⁴⁻²⁰

Figure 1.

Structures of clinically advanced M₄-targeting therapeutics.

Our lab has heavily focused on developing novel, selective M₄ PAM chemotypes bereft of the traditional β-amino carboxamide moiety historically believed to be essential for M₄ PAM activity (Figure 2, dashed circle).²¹⁻²⁷ This key pharmacophore gave rise to M₄ PAMs with poor solubility, species potency discrepancies, and poor brain exposure.²⁸⁻³⁷ Most recently, our laboratory disclosed a novel 5,6,6-tricyclic scaffold and a novel 6,5,6-tricyclic scaffold that still afforded potent and CNS-penetrant M₄ PAMs (Figure 2, VU6007215 and VU6017649).^24,25 To identify additional novel M₄ PAM chemotypes, we elected to further explore these tricyclic scaffolds. An historic β-amino carboxamide-containing M₄ PAM (4), developed by the Capuano laboratory, formed the base of our novel tricycles.³⁸ This exercise resulted in the discovery of a novel M₄ PAM chemotype containing a 7,9-dimethylthieno[2,3-d:4,5-d′]dipyrimidine core, 5. Further exploration revealed a second, novel, tricyclic M₄ PAM chemotype containing a 2,4-dimethylpyrido[4′,3′:4,5]thieno[2,3-d]pyrimidine core, 6. This body of work details the development of these two novel M₄ PAM chemotypes.

Figure 2.

Exploration of novel tricyclic cores as M₄ PAMs revealed two unique M₄ PAM tricyclic chemotypes: 7,9-dimethylthieno[2,3-d:4,5-d′]dipyrimidine core (5) and 2,4-dimethylpyrido[4′,3′:4,5]thieno[2,3-d]pyrimidine core (6).

Results and Discussion

The synthesis of tricyclic core 5 began with reacting commercially available acyl chlorides 7, 3-aminocrotonitrile (8), and ammonium thiocyanate to afford the tetra-substituted pyrimidine 9 (Scheme 1). Similar to published protocols, treatment with 2-chloroacetamide (10) and potassium carbonate followed by irradiation in a microwave reactor generated carboxamide 11.³⁹ Treatment of 11 with trimethylorthoformate under heat afforded pyrimidone intermediate 12. Pyrimidone 12 was then converted into chloride 13 with POCl₃, which then readily underwent nucleophilic aromatic substitution with a variety of amines to yield desired analogues 14. For this exercise, we decided to forego exploring amino azetidines of past M₄ PAMs because of their potential metabolic instability.⁴⁰ Instead, we focused on examining small aliphatic amines, which typically provided potent M₄ PAMs, as well as substituted benzylamines, which could aid in increasing solubility.

Scheme 1. Synthesis of M₄ PAM Analogues 14.

Reagents and conditions: (a) 8, NH₄SCN, 1,4-dioxane, 110 °C, 2 h, 46%; (b) 10, K₂CO₃, N-methyl-2-pyrrolidone (NMP), microwave-irradiated at 60 °C for 30 min, then microwave-irradiated at 100 °C for 1 h, 43%; (c) CH(OEt)₃, 150 °C, 3 h, 99%; (d) POCl₃, TEA, 1,2-dichloroethane (DCE), 110 °C, 3 h, 81%; (e) amine, N,N-diisopropylethylamine (DIEA), NMP, 60 °C for 2 h or microwave-irradiated at 100 °C for 10 min, 30–77%.

Select analogues 14 were screened against human M₄ (hM₄) to determine EC₅₀ values with results highlighted in Table 1. These results emphasize the importance of the amine tail on potency. In the context of the 7,9-dimethylthieno[2,3-d:4,5-d’]dipyrimidine core (5), the benzyl amine-containing analogues 14j and 14k both have hM₄ EC₅₀ values > 1 μM. However, several of the small aliphatic amine tails provided compounds with hM₄ EC₅₀ values < 500 nM (14g, hM₄ EC₅₀ = 290 nM; 14h, hM₄ EC₅₀ = 100 nM; and 14i, hM₄ EC₅₀ = 160 nM). It was noticed that minor modifications, such as fluorine substitutions on the pyrrolidine (14e vs 14g), led to a 2.2-fold loss in potency. Moreover, removal of the oxygen from the morpholine analogue 14b (hM₄ EC₅₀ = 5.6 μM) to give the piperidine analogue 14a (hM₄ EC₅₀ > 30 μM) resulted in a loss of activity. Additionally, substituting the 2-oxa-6-azaspiro[3.3]heptane tail of 14c (hM₄ EC₅₀ = 1.6 μM) with the 6,6-difluoro-2-azaspiro[3.3]heptane tail of 14h (hM₄ EC₅₀ = 100 nM) resulted in a 16-fold increase in functional human M₄ potency. Interestingly, we also observed an 11-fold loss of potency when comparing the pyridine predecessor compound VU6007215 (Figure 1; hM₄ EC₅₀ = 144 nM) to the pyrimidine compound 14c. One could speculate that this result is due to the difference in the tricyclic cores; however, other amine tails were nearly equipotent [e.g., pyrrolidine (14g)].²⁵ These results reiterate the importance the amine tail can have on potency, particularly in combination with different tricycles.

Table 1. Structures and Activities for Analogues 14.

^aCalcium mobilization assays with hM_4/Gqi5-CHO cells were performed in the presence of an EC₂₀ fixed concentration of acetylcholine. EC₅₀ values for hM₄ represent at least one experiment performed in triplicate.

We next turned our attention to evaluating the relevance of the pyrimidine nitrogen at the 5-position by synthesizing analogues 22 and 23. Previously, conversion of this pyrimidine ring into a pyridine ring was not detrimental to hM₄ activity; rather, several analogues displayed increased potency.⁴¹ To determine if this trend proved true for our current scaffold, we synthesized analogues with the tricyclic core 6. The synthesis was accomplished by first treating intermediate 9b with ethyl 2-chloroacetate (15) and sodium carbonate to generate carboxylate 16 (Scheme 2). A copper-catalyzed Sandmeyer reaction gave bromide 17, which could then undergo Suzuki–Miyaura coupling with 1,3,2-dioxaborolane 18 to afford carboxylate 19. In a two-step process, intermediate 19 was first treated with TFA and heat to generate a pyranone intermediate that was further transformed in the presence of NH₄OH and heat to yield pyridinone 20. Treatment with POCl₃ yielded chloride 21, which could readily undergo nucleophile aromatic substitution to generate desired analogues 22. Alternatively, chloride 21 could be subjected to an iron-catalyzed Kumada coupling with various Grignard reagents to afford analogues 23.⁴²

Scheme 2. Synthesis of M₄ PAM Analogues 22 and 23.

Reagents and conditions: (a) 8, NH₄SCN, 1,4-dioxane, 110 °C, 2 h, 21–42%; (b) 15, Na₂CO₃, DMF, 45 °C, 3 h, 52–91%; (c) CuBr₂, ^tBuONO, acetonitrile (ACN), 1 h, 40–75%; (d) 18, Cs₂CO₃, Pd(dppf)Cl₂, 1,4-dioxanes/H₂O (10:1), 80 °C, 18 h, 44–76%; (e) (i) TFA, 120 °C, 2 h; (ii) NH₄OH, 100 °C, 4.5 h, 30–88% over two steps; (f) POCl₃, microwave-irradiated at 120 °C, 30 min, 13–92%; (g) aliphatic amine, DIEA or K₂CO₃, NMP, microwave-irradiated at 120–180 °C, 0.5–2 h, 8–20%; (h) benzyl amine, Pd₂(dba)₃, Xantphos, Cs₂CO₃, 1,4-dioxane, 110 °C, 6–18 h, 14–51%; (i) R₂MgCl, Fe(acac)₃, THF/NMP (5:1), 1–18 h, 31–41%.

Select analogues 22 and 23 were screened against hM₄ to determine potency with results highlighted in Table 2. It was evident that 2,4-dimethylpyrido[4′,3′:4,5]thieno[2,3-d]pyrimidine core (6) provided a boost in functional human hM₄ potency. For instance, comparing the benzyl amine tail analogues 22g (hM₄ EC₅₀ = 140 nM) and 22j (hM₄ EC₅₀ = 170 nM) to 14j (hM₄ EC₅₀ = 1.3 μM) and 14k (hM₄ EC₅₀ = 1.2 μM) resulted in a 9-fold and 7-fold increase in EC₅₀, respectively. This trend was also observed with small aliphatic groups. Comparing analogue 22k (hM₄ EC₅₀ = 73 nM) to 14g (hM₄ EC₅₀ = 290 nM) and 22f (hM₄ EC₅₀ = 33 nM) to 14e (hM₄ EC₅₀ = 640 nM) resulted in a 4-fold and 19-fold increase in EC₅₀, respectively. Most notable was the potency difference in the pyrrolidine analogues 22e (hM₄ EC₅₀ = 73 nM) and 14a (hM₄ EC₅₀ > 30 μM). Taken together, these data suggest that the basicity of the nitrogen at the 7-position plays a key role in hM₄ PAM potency and may function as a critical H-bond acceptor, while the nitrogen at the 5-position is not essential for activity.

Table 2. Structures and Activities for Analogues 22 and 23.

^aCalcium mobilization assays with hM_4/Gqi5-CHO cells were performed in the presence of an EC₂₀ fixed concentration of acetylcholine. EC₅₀ values for hM₄ represent at least one experiment performed in triplicate.

We next wanted to explore various substitutions on the pyridine ring of the tricyclic core 6 to expand upon our SAR studies, as well as elucidate the spatial constraints of the binding pocket. Being para to the amine linkage and, thus, a potential site of metabolism, we chose to explore substitutions at the 5-position of the tricyclic ring (R₂). The preparation of analogues 27 and 28 began with chlorinating pyridinone 20b in the presence of N-chlorosuccinimide to give intermediate 24 (Scheme 3). Treatment with POCl₃ afforded dichloride 25, which then underwent either a nucleophilic aromatic substitution or a Buchwald–Hartwig amination to generate analogues 26. Analogue 26a was further derivatized utilizing Suzuki–Miyaura coupling conditions to yield analogue 27. Alternatively, analogue 26a was subjected to palladium-catalyzed cyanation to afford analogue 28. We decided to focus mainly on analogues containing the 2-[4-(aminomethyl)phenyl]propan-2-ol tail as this would aid in solubility, as well as the small aliphatic pyrrolidine tail for comparison.

Scheme 3. Synthesis of M₄ PAM Analogues 26, 27, and 28.

Reagents and conditions: (a) N-chlorosuccinimide (NCS), DMF, 18 h, 55%; (b) POCl₃, microwave-irradiated at 150 °C, 1 h, 45%; (c) pyrrolidine, DIEA, NMP, microwave-irradiated at 150 °C, 30 min, 15%; (d) 2-[4-(aminomethyl)phenyl]propan-2-ol, Pd₂(dba)₃, Xantphos, Cs₂CO₃, 1,4-dioxane, microwave-irradiated at 130 °C, 1 h, 29%; (e) cyclopropylboronic acid, Cs₂CO₃, Pd(dppf)Cl₂, 1,4-dioxanes/H₂O (10:1), microwave-irradiated at 110 °C, 30 min, 11%; (f) Zn⁰, Zn(CN)₂, NiCl₂, dppf, 4-(dimethylamino)pyridine (DMAP), ACN, 80 °C, 2 h, 43%.

Select analogues 22, 26, 27, and 28 were screened against hM₄ to determine potency with results highlighted in Table 3. It was evident that substitutions larger than methyl at the R₁ position were tolerated; however, a decrease in the EC₅₀ was observed. In fact, an increase in steric bulk [22p (ethyl) < 22o (cyclopropyl) < 22n (isobutyl)] coincided with more dramatic reductions in activity (4.0-fold, 9.3-fold, >66-fold decrease, respectively). This was also observed with the pyrrolidine analogue 22q (>13-fold loss of potency). Similarly, substitutions at the 5-position of the ring (R₂) also led to a loss of potency (analogues 26a, 26b, 27, and 28).

Table 3. Structures and Activities for Analogues 22, 26, 27, and 28.

^aCalcium mobilization assays with hM_4/Gqi5-CHO cells were performed in the presence of an EC₂₀ fixed concentration of acetylcholine. EC₅₀ values for hM₄ represents at least one experiment performed in triplicate.

Of these compounds, 14h, 14i, 22e, 22f, 22g, 22j, and 22k were advanced into a battery of in vitro drug metabolism and pharmacokinetics (DMPK) assays and our standard rat plasma-to-brain level (PBL) cassette paradigm (Table 4). These compounds were chosen to move forward based on their hM₄ potency (EC₅₀ < 190 nM), as well as their chemical diversity across subseries. In regards to physicochemical properties, these analogues all possessed molecular weights less than 400 Da and desirable topological polar surface area (TPSA) (<80 Å) with 14h, 14i, 22e, 22f, and 22k having the most attractive CNS xLogP values (2.52–3.68).^43,44 Of the analogues tested, several (14i, 22e, 22f, 22j, and 22k) displayed high predicted hepatic clearance in both human and rat on the basis of microsomal CL_int data (rat CL_hep ≥ 46 mL/min/kg; human CL_hep ≥ 15 mL/min/kg). These analogues were not progressed forward because of their high predicted hepatic clearance in rat and/or human. By contrast, analogues 14h and 22g display moderate human and rat predicted hepatic clearance on the basis of microsomal CL_int data (rat CL_hep of 37–44 mL/min/kg; human CL_hep of 12–14 mL/min/kg).

Table 4. In Vitro DMPK and Rat PBL Data for Select Analogues 14 and 22.

property	14h	14i	22e	22f	22g	22j	22k
	VU6015976	VU6015368	VU6016233	VU6017369	VU6016235	VU6016234	VU6016225
MW	347.39	297.38	298.41	320.36	378.49	368.43	284.38
xLogP	3.06	2.52	3.68	3.45	4.43	4.23	3.23
TPSA	54.8	54.8	41.9	41.9	70.9	59.9	41.9
in vitro pharmacokinetic (PK) parameters
CLint (mL/min/kg), rat	121	1048	3444	1518	78	386	3055
CLhep (mL/min/kg), rat	44	66	69	67	37	59	68
CLint (mL/min/kg), human	45	177	134	148	26	50	124
CLhep (mL/min/kg), human	14	19	18	18	12	15	18
rat fu,plasmaa	0.132	0.048	0.019	0.043	0.024	0.014	0.038
human fu,plasmaa	0.035	0.013	0.007	0.010	0.014	0.009	0.023
rat fbraina	0.033	0.012	0.004	0.015	0.016	0.009	0.007
brain distribution (0.25 h) (SD rat; 0.2 mg/kg IV)
Kp, brain:plasmab	5.07	7.27	7.77	8.33	1.14	1.69	18.6
Kp,uu, brain:plasmac	1.27	1.82	1.64	2.91	0.76	1.01	3.43

^af_u = fraction unbound; equilibrium dialysis assay; brain = rat brain homogenates.

^bK_p = total brain to total plasma ratio.

^cK_p,uu = unbound brain (brain f_u × total brain) to unbound plasma (plasma f_u × total plasma) ratio.

All compounds tested had low-to-moderate rat plasma protein binding (rat f_u,plasma = 1.5–13.2%). Analogue 22e was highly bound to human plasma protein (human f_u,plasma < 1%), while all other analogues analyzed displayed moderate human plasma protein binding profiles (human f_u,plasma = 1.0–3.5%). Analogues 22e, 22j, and 22k were highly bound to rat brain homogenate (f_u,brain < 1%). Conversely, analogues 14h, 14i, 22f, and 22g were moderately bound to rat brain homogenate (f_u,brain = 1.2–3.3%). All compounds tested achieved acceptable CNS penetration (K_p ≥ 1) and unbound brain to unbound plasma ratios (K_p,uu ≥ 0.76). Because of their superior DMPK profiles, 14h and 22g were screened against rat M₄ (rM₄) to ensure there was no appreciable species differences in M₄ activity.⁴⁶ Both analogues were potent against rM₄ (14h, rM₄ EC₅₀ = 380 nM; 22g, rM₄ EC₅₀ = 42 nM) and minor species differences were within the acceptable range (2–3.5-fold); however, 22g was 9-fold more potent against rM₄ than 14h. Thus, 22g (VU6016235) was selected for further evaluation in an array of in vitro and in vivo experiments.

When evaluated, VU6016235 displayed high subtype selectivity across the mAChRs (M₁, M₃, and M₅ = inactive; M₂ = 148-fold selectivity) and was highly brain-penetrant with a P-glycoprotein (P-gp) efflux ratio of 1.5 (Table 5). Additionally, VU6016235 displayed an excellent cytochrome P450 (CYP450) inhibition profile with IC₅₀ values ≥ 17 μM across all isoforms tested (1A2, 2D6, 2C9, and 3A4). Highlighted in Table 6 are the in vivo rat PK parameters. VU6016235 displayed high oral bioavailability (≥100% at a 3 mg/kg dose and 84% at a 30 mg/kg dose) and low plasma clearance (4.9 mL/min/kg) in rat. Volume of distribution was moderate (1.3 L/kg), and elimination t_1/2 was 3.4 h. In a rat dose escalation pharmacokinetic (PK) study, VU6016235 demonstrated an increase in the mean AUC_0-∞ (3 to 30 mg/kg, po). In a preclinical pharmacodynamic (PD) model of antipsychotic-like activity, VU6016235 demonstrated a robust dose-dependent reversal of amphetamine-induced hyperlocomotion (AHL) significant at doses of 1.0, 3.0, and 10.0 mg/kg after oral administration (MED = 1 mg/kg, Figure 3). At 90 min postdose, brain/plasma K_p and K_p,uu were determined at 1 mg/kg (K_p = 0.62; K_p,uu = 0.41), 3 mg/kg (K_p = 0.79; K_p,uu = 0.53), and 10 mg/kg (K_p = 0.82; K_p,uu = 0.55) with mean C_brain,u values of 13.9, 60.3, and 204 nM, respectively (Table 7).

Table 5. Further In Vitro Characterization of VU6016235 (22g).

muscarinic selectivity45
	EC50 (nM)	[%Ach Max]	pEC50
rat M4a	42	[89]	7.38
human M1a	inactive
human M2a	6200	[46]	5.21
human M3a	inactive
human M5a	inactive

permeability assay
cell line	Papp (A–B) (cm/s)	Papp (B–A) (cm/s)	efflux ratio
MDCK-MDR1	47 × 10–6	68 × 10–6	1.5

P450 inhibition IC50 (μM)b
1A2	2D6	2C9	3A4
23	>30	17	17

^aCalcium mobilization assay; EC₅₀ values for rM₄ and hM₂ represent three experiments performed in triplicate; EC₅₀ values for hM_1,3,5 represent one to two experiments performed in triplicate.

^bAssayed in pooled human liver microsomes (HLM) in the presence of NADPH with CYP-specific probe substrates.

Table 6. In Vivo Rat and Dog iv/po PK Profile of VU6016235.

iv PK
species	dose (mg/kg)	t1/2 (hr)b	MRT (hr)b	CLp (mL/min/kg)b	Vss (L/kg)b
rat (SD)a	1.0	3.4	4.3	4.9	1.3
dog (beagle)c	0.5	2.4	1.6	23.8	2.3

po PK
species	dose (mg/kg)	Tmax (hr)	Cmax (ng/mL)	AUC0-∞ (hr·ng/mL)	%F
rat (SD)d	3.0	3.0	657	8408	>100
rat (SD)d	30.0	5.0	4455	59164	84
dog (beagle)e	1.0	0.83	62.4	200	27.5

^aMale Sprague–Dawley (SD) rats (n = 2); vehicle = 10% EtOH, 40% PEG 400, 50% Saline (1.0 mL/kg).

^bt_1/2 = terminal phase plasma half-life; MRT = mean residence time; Cl_p = plasma clearance; V_ss = volume of distribution at steady state.

^cMale beagle dogs (n = 3); vehicle = 10% 2-hydroxypropyl-β-cyclodextrin (HP-β-CD).

^dMale Sprague–Dawley rats (n = 2); vehicle = 10% Tween-80 in water (10 mL/kg; suspension).

^eMale beagle dogs (n = 3); fasted; vehicle = 0.5% hydroxypropyl methylcellulose (HPMC) in water.

Figure 3.

VU6016235, administered orally, reverses amphetamine-induced hyperlocomotion in male Sprague–Dawley rats. (Left) Time course of locomotor activity. (Right) Total locomotor activity during the 60 min period following amphetamine administration. Data are means ± SEM of 6–8 animals per group. *p < 0.05, **p < 0.01,***p < 0.001 vs Vehicle + Amphetamine. VU0467154 = positive control. Vehicle = 10% Tween80 in H₂O.

Table 7. Relationship between Total (Mean C_brain) and Unbound (Mean C_brain,u) Brain Concentrations of VU6016235 and Pharmacodynamic Effects on Amphetamine (0.75 mg/kg, sc)-Induced Hyperlocomation in Rats.^a.

dose (mg/kg)	mean reversal of AHL (%)	mean Cplasmab (μM)	mean Cplasma,uc (nM)	mean Cbrainb (μM)	mean Cbrain,ud (nM)	brain/plasma mean Kp	brain/plasma mean Kp,uu
1	30.5	1.37	33.2	0.85	13.9	0.62	0.41
3	41.4	4.66	113	3.70	60.3	0.79	0.53
10	54.2	15.3	372	12.5	204	0.82	0.55

^aBioanalysis only performed for efficacious doses.

^bAt 1.5 h proadministration.

^cEstimated unbound plasma concentration based on the rat f_u,p (0.024).

^dEstimated unbound brain concentration based on the rat f_u,b (0.016).

On the basis of the encouraging in vivo rat PK and PD profile of VU6016235, the off-target and safety/toxicity profiles for this compound were further investigated. An ancillary pharmacology screen (Eurofins Panlabs) revealed no off-target activity (≤26% inhibition at 10 μM across all off-targets screened) (see the Supporting Information for the full ancillary pharmacology profile).⁴⁷ Additionally, VU6016235 was negative in a Mini Ames test (4 strains ± S9). Potential cardiac risks were also assessed (Charles River ChanTest Cardiac Channel Panel), and VU6016235 was not found to strongly inhibit any cardiac ion channels tested (all ion channel inhibitions ≤14% at 10 μM (patch clamp, see the Supporting Information for additional details).⁴⁸ With such promising data, VU6016235 was progressed into higher species in vivo PK (Table 6). VU6016235 displayed moderate oral bioavailability (27.5% at a 1 mg/kg dose) in the dog; however, high plasma clearance (∼24 mL/min/kg, nearly 78% hepatic blood flow) in the dog halted further progress of VU6016235.

Conclusion

In summary, a scaffold hopping exercise utilizing a “tie-back” strategy to design M₄ PAMs 1 and 2 proved to be a successful strategy in converting an early micromolar M₄ PAM, 4, into potent tricyclic M₄ PAM analogues devoid of the classic β-amino carboxamide moiety. Two analogues within the tricycle series containing the 7,9-dimethylthieno[2,3-d:4,5-d’]dipyrimidine core core (5) were potent M₄ PAMs (14h and 14i; hM₄ EC₅₀ < 200 nM). Both analogues displayed favorable fraction unbound in regard to human plasma protein and rat brain homogenate binding (BHB) (1% < f_u < 5%) and low rat plasma protein binding (PPB) (f_u > 5%). Additionally, both analogues were highly CNS-penetrant (rat brain/plasma K_p > 5, K_p,uu > 1). Points of difference were observed in the predicted hepatic clearance. While 14i displayed high human (CL_hep = 19 mL/min/kg) and rat (CL_hep = 66 mL/min/kg) predicted hepatic clearance, 14h displayed moderate human (CL_hep = 14 mL/min/kg) and rat (CL_hep = 44 mL/min/kg) predicted hepatic clearance.

Moreover, several analogues (22e-g and 22j,k) containing the 2,4-dimethylpyrido[4′,3′:4,5]thieno[2,3-d]pyrimidine core (6) were potent M₄ PAMs (hM₄ EC₅₀ < 200 nM). Four of the five analogues profiled (22e, 22f, 22j, and 22k) displayed high human (CL_hep ≥ 15 mL/min/kg) and rat (CL_hep ≥ 50 mL/min/kg) predicted hepatic clearance. Conversely, 22g displayed moderate human (CL_hep = 12 mL/min/kg) and rat (CL_hep = 37 mL/min/kg) predicted hepatic clearance. While analogues 22e, 22i, and 22k were all highly bound to rat brain homogenates (f_u < 1%), analogues 22f and 22g had modest fraction unbound (1% < f_u < 5%). Most analogues in this series had moderate rat and human plasma protein binding except for 22e, which was highly bound to human plasma proteins. Furthermore, all five analogues evaluated were highly CNS-penetrant (rat brain/plasma K_p > 5, K_p,uu > 1).

Our efforts to mask the β-amino carboxamide moiety of an early M₄ PAM compound 4 resulted in the discovery of two new tricyclic cores (5 and 6), which provided potent M₄ PAM analogues that were highly brain-penetrant. This endeavor also provided analogues 14h (VU6015976) and 22g (VU6016235), which display 24-fold and 30-fold increase in human M₄ activity, respectively, when compared to parent compound 4 (hM₄ EC₅₀ = 4.5 μM). Moreover, tricycles VU6015976 and VU6016235 exhibited moderate predicted hepatic clearance profiles in rat and human similar to parent compound 4; this was a great improvement over previously reported tricycle VU6007215 (Figure 2).²⁵ Additionally, both VU6015976 and VU6016235 displayed a higher CNS distribution of unbound drug (K_p,uu = 0.76 and 1.27, respectively) compared to VU6007215 (K_p,uu = 0.40). Overall, VU6015976 and VU6016235 have similarly attractive DMPK profiles; however, because of VU6016235 exhibiting a nearly 9-fold increase in rat M₄ activity compared to VU6015976, we selected VU6016235 to carry forward for further in vivo PK/PD profiling. Although VU6016235 possessed a favorable rat PK profile, displayed efficacy in a rat AHL-induced hyperlocomotion PD assay, and was determined to have minimal safety and toxicity risks, its inferior dog in vivo PK necessitated evaluation in alternative nonrodent species. Efforts aimed at ameliorating higher species PK are ongoing and will be reported in due course.

Glossary

Abbreviations

ACh

acetylcholine

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

mAChR

muscarinic acetylcholine receptor

MED

minimum effective dose

PAM

positive allosteric modulator

PBL

plasma-to-brain level

PPB

plasma protein binding

rM₄

muscarinic acetylcholine receptor subtype 4, rat

SAR

structure activity relationship

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acschemneuro.4c00465.

• Detailed methods for the synthesis and characterization of key compounds, experimental details for all assays performed [in vitro pharmacology; DMPK (in vitro and in vivo); rat AHL], and additional ancillary pharmacology data (Eurofins Panlabs and Cardiac Ion Channel Panel) (PDF)

Author Contributions

J.L.E., L.A.B., 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. W.P., J.M.R., A.T.G., and C.K.J. performed in vivo pharmacokinetic experiments. S.C., T.M.B., and O.B. performed and analyzed DMPK experiments. J.L.E., K.J.T., P.J.C, C.W.L., 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.

The authors declare the following competing financial interest(s): J.L.E., L.A.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.