Discovery of VU0467485/AZ13713945: An M₄ PAM Evaluated as a Preclinical Candidate for the Treatment of Schizophrenia

Abstract

Herein, we report the structure–activity relationships within a series of potent, selective, and orally bioavailable muscarinic acetylcholine receptor 4 (M₄) positive allosteric modulators (PAMs). Compound 6c (VU0467485) possesses robust in vitro M₄ PAM potency across species and in vivo efficacy in preclinical models of schizophrenia. Coupled with an attractive DMPK profile and suitable predicted human PK, 6c (VU0467485) was evaluated as a preclinical development candidate.

The rapid onset of clinical efficacy of the M₁/M₄ preferring agonist xanomeline in both schizophrenic and Alzheimer’s patients led to a revolution in muscarinic receptor drug discovery efforts targeting allosteric mechanisms to afford highly subtype selective M₁ and M₄ positive allosteric modulators (PAMs).¹⁻⁶ Of these, M₄ has emerged as the favored mAChR subtype responsible for antipsychotic-like efficacy as well as modest cognitive enhancement in multiple preclinical rodent models via a fundamentally new molecular mechanism.⁷⁻¹⁵ While M₁ PAMs with diverse chemotypes (and without major species differences in pharmacology) have rapidly advanced to potentially address cognitive dysfunction and negative symptoms,^5,6,16,17 M₄ PAMs have progressed more slowly for many reasons, including species differences in M₄ PAM potency (i.e., affinity and cooperativity), challenges with respect to M₂ selectivity, and P-gp efflux as well as limited chemical diversity (Figure 1).^{9−14,18−20} Clearly, these are significant roadblocks en route to an M₄ PAM preclinical candidate. In this Letter, we detail our navigation of these issues within the VU0467154 series of M₄ PAMs, leading to the discovery of a potent, selective, and orally bioavailable M₄ PAM 6c (VU0467485) with robust efficacy in behavioral models that was evaluated as a preclinical candidate. This is the first disclosure of the structure–activity relationships (SAR) and preclinical profile of 6c.

Figure 1.

Structures of reported M₄ PAMs. LY2033298 (1) was the first reported M₄ PAM but was human-preferring in potency. VU0152100 (2) was the first centrally active M₄ PAM in rodents, and VU0467154 (3) has proven to be the best-in-class M₄ PAM rodent in vivo tool compound.

Our discovery of 6c originated from an optimization campaign centered on 2, and from which 3 was identified, but discontinued due to a human and rat M₄ PAM potency disconnect that precluded development.^10,12−14 A major thrust of the initial lead optimization effort was to survey alternatives for the pyridine ring in 2, and while other pyridine regioisomers and pyrimidines afforded modest M₄ potentiation, the pyridazine core, found in 3, proved optimal for imparting both potency, metabolic stability, and favorable physiochemical properties to the series.²¹ SAR was steep and divergent across rat and human M₄, and CNS penetration was low in this series; however, exceptional DMPK properties could be achieved.²¹

The synthesis of analogue 6 was straightforward and required only two steps from known materials (Scheme 1).¹²⁻¹⁴ Condensation of 3-chloro-5,6-dimethylpyridazin-4-carbonitrile 4 with methylthioglycolate under microwave irradiation delivered the core sodium 5-amino-3,4-dimethylthieno[2,3-c]pyridazine-6-carboxylate 5 in yields averaging 78%. Next, a HATU-mediated amide coupling reaction with various amines provides analogues 6 in yields ranging from 45 to 92% after HPLC purification.²¹ The direct congener of 2, VU0464090 (6a), demonstrated a ∼3-fold increase in M₄ PAM potency (human M₄ PAM EC₅₀ = 130 nM, pEC₅₀ = 6.89 ± 0.06, 83.7 ± 3.5 ACh Max and rat M₄ PAM EC₅₀ = 59.7 nM, pEC₅₀ = 7.22 ± 0.06, 78.1 ± 1.7 ACh Max), increased fraction unbound in plasma (fu_plasma(r,h) = 0.022, 0.035) and reduced hepatic microsomal intrinsic clearance (CL_int(r,h) = 81 and 36 mL/min/kg; predicted CL_hep(r,h) = 37 and 13 mL/min/kg) by virtue of the pyridazine ring system; however, metabolic instability of the PMB group precluded further advancement of 6a.¹⁴ As shown in Table 1, application of the fluorine walk strategy²²⁻²⁴ for allosteric modulator optimization proved fruitful, affording a number of potent M₄ PAMs (6c–h). Heteroatom incorporation into the benzyl ring was met with limited success, with only 3-pyridyl congeners (6i and 6j) retaining PAM activity (other regioisomeric pyridines, pyrimidines, pyrazines, and pyridazines were inactive, M₄ EC₅₀s > 10 μM). Moreover, we suspected CYP₄₅₀-mediated oxidative dealkylation of 6a led to 6b, an active putative metabolite (human M₄ PAM EC₅₀ = 96.7 nM, pEC₅₀ = 7.01 ± 0.01, 69.8 ± 2.2 ACh Max), but with high clearance in vivo (rat) and undesired activity at hM₂. As a battery of in vitro and in vivo DMPK assays (vide infra) quickly identified 6c as the most attractive PAM in the series, we prepared its putative dealkylated metabolite 6k, which also proved to be active (human M₄ EC₅₀ = 59 nM), but with no activity at hM₂, suggesting the ortho-fluoro moiety enhances selectivity versus hM₂. Incorporation of chiral methyl groups at the benzylic position of 6c led to (R)-6l (human M₄ EC₅₀ > 10 μM) and (S)-6m (human M₄ EC₅₀ = 508 nM), where enantiospecific activity was noted, but with unacceptable loss in human M₄ PAM potency. Finally, based on previous beneficial disposition by virtue of the kinetic isotope effect, we evaluated a deuterated congener, but in this instance, there was no benefit to stability.

Scheme 1. Synthesis of Analogues 6.

Reagents and conditions: (a) methyl thioglycolate, MeOH, 1 M aq. NaOH, 150 °C, microwave, 30 min, 78%; (b) NH₂CH₂Ar(Het), HATU, DMF, DIEPA, 2 h, 45–92%.

Table 1. Structures and Activities of Analogues 6.

^aCalcium mobilization assays with hM₄/G_qi5-CHO cells performed in the presence of an EC₂₀ fixed concentration of acetylcholine; values represent means from three (n = 3) independent experiments performed in triplicate.

Thus, efforts focused on 6c as a potential M₄ PAM preclinical candidate with minimal species differences in PAM potency.

The molecular pharmacology profile of M₄ PAM 6c is shown in Figure 2. PAM 6c is inactive in the absence of acetylcholine (ACh), but in the presence of an EC₂₀ concentration of ACh (Figure 2A), 6c potentiates M₄ in a concentration-dependent manner, affording potent activity at both human and rat M₄ (human M₄ PAM EC₅₀ = 78.8 nM, pEC₅₀ = 7.10 ± 0.01, 80.6 ± 0.7 ACh Max and rat M₄ PAM EC₅₀ = 26.6 nM, pEC₅₀ = 7.57 ± 0.05, 68.7 ± 3.4 ACh Max). For a preclinical candidate, comparable activity across species is essential for IND-enabling toxicology studies, and this was a major issue for previously reported M₄ PAMs for which the species disconnect averaged 10–50-fold.^{9−14,18−20} Thus, we also evaluated 6c against dog and cynomolgous monkey M₄ and the key antitarget, M₂. Beyond rat and human, 6c displayed no major species differences in potency (dog M₄ EC₅₀ = 87 nM, 49% ACh Max, dog M₂ > 30 μM and cyno M₄ EC₅₀ = 102 nM, 74% ACh Max, cyno M₂ > 30 μM). In a progressive fold-shift assay with human M₄ (Figure 2B), 6c afforded a maximal ∼45-fold leftward shift of the human M₄ ACh concentration–response curve (CRC) at 10 μM (∼40-fold shift at rat M₄). Moreover, PAM 6c was selective versus human and rat M_1–3,5. The operational model of allosterism was applied to the fold-shift data to better understand the allosteric effects of this promising candidate.^5,22 PAM 6c displayed robust potentiation of ACh and cooperativity (log αβ = 2.1, αβ = 134), significant intrinsic efficacy (log τ_B = 5.1), and an approximately 1 μM estimated affinity (pK_B 6.025, K_B = 944 nM) at human M₄.

Figure 2.

Molecular pharmacology profile of M₄ PAM 6c. (A) Enhanced intracellular calcium release induced by a subthreshold concentration of acetylcholine (EC₂₀), a PAM CRC on both rat and human M₄, with EC₅₀s of 26.6 and 78.8 nM, respectively. (B) Compound 6c induces a ∼45-fold maximal leftward shift of the hM₄ acetylcholine response curve at 10 μM. (C) Compound 6c is highly selective for hM₄ over hM_1–3,5. (D) Compound 6c is highly selective for rM₄ over rM_1–3,5. Data represent means from at least three independent determinations performed in triplicate using CHO cells stably transfected with the indicated mAChR.

Compound 6c was then evaluated in a panel of in vitro DMPK assays²³ where it displayed properties that supported continued progression. Not only did 6c possess an exceptionally clean CYP₄₅₀ inhibition (3A4, 2D6, 2C9, 1A2 IC₅₀s > 30 μM in human hepatic microsomes) and induction profile (3A4, 1A2, 2B6 EC₅₀s > 50 μM, E_maxs ≤ 1.0 in cryopreserved human hepatocytes), but also moderate plasma protein binding was noted across species (fu_plasma(r,h,c) = 0.031, 0.054, and 0.091) with moderate fraction unbound in rat brain homogenate (fu_br = 0.037). Based on hepatic microsomal CL_int data, moderate predicted hepatic clearance was also observed for 6c (predicted CL_hep(r,h,c) = 73, 26, and 38 mL/min/kg). Moreover, 6c was not a P-gp substrate (ER = 1.4 in MDCK-MDR1) and showed good apparent permeability (Caco-2 P_app = 31 × 10^–6 cm/s). In rat and dog brain distribution studies, 6c displayed moderate to high CNS penetration with K_ps of 0.31 to 1.0 and K_p,uus of 0.37 to 0.84. With regard to toxicity evaluation in vitro, 6c was clean in GSH trapping experiments (human hepatic microsomes) and a mini-Ames test (data not shown). In vivo PK (Table 2) was assessed in three species (rat, dog, and cynomolgous monkey), which revealed moderate clearance, low to moderate volume of distribution at steady-state, and short to moderate elimination half-life with high oral bioavailability in rat but low bioavailability in dog (3 mg/kg suspension dose of a mono-HCl salt). In vitro metabolite identification experiments found good coverage of human metabolites in dog and cynomolgous monkey and no evidence for human-unique metabolites (Figure 3). Human CYP₄₅₀ phenotyping experiments revealed that multiple CYPs contribute to 6c’s metabolism (Figure 4) with a generally low potential for drug–drug interactions (no metabolism-related DDI liabilities anticipated in Alzheimer’s disease clinical population as concomitant medications (acetylcholinesterase inhibitors possess a 3A4/2D6 phenotype) and, in schizophrenia populations, common antipsychotics (e.g., clozapineolanzapine) possess a 1A2/2D6/3A4 phenotype). Furthermore, in a functional hERG assay, 6c was inactive when tested at 11 μM, as well as against a larger cardiac ion channel panel (IC₅₀s > 33 μM). Finally, ancillary pharmacology was assessed in an internal AZ/Cerep panel against 200 targets, and no significant off-target activities (IC₅₀s or EC₅₀s > 10 μM) were noted, including both binding and functional assays for cardiac ion channels, with the exception of a 1.2 μM IC₅₀ (radioligand binding) at the rat GABA_A receptor.²¹

Table 2. Pharmacokinetic Parameters of 6c.

parameter	rata (SD)	doga (beagle or mongrel)	NHPa (cyno)
dose (mg/kg) i.v./p.o.	1/3	1/3	0.2 (iv only)
CLp (mL/min/kg)	29	21	25
Vss (L/kg)	1.5	1.5	0.88
elimination t1/2 (h)	4.2	1.3	0.48
Cmax (μM) p.o.	1.2	0.22
Tmax (h) p.o.	1.5	0.75
AUCo-inf (μM·h) p.o.	3.8	0.59
F (%) p.o.	79	9.5
total brain/total plasma (Kp)	0.31	1.0
unbound brain/unbound plasma (Kp,uu)	0.37	0.84

^aValues represent means from two to three animals.

Figure 3.

Metabolite identification studies for 6c across species (rat, dog, cyno monkey, and human).

Figure 4.

P450 phenotyping and CYP mapping for 6c, indicating that multiple CYP₄₅₀s (3A4/2C19/1A2 and to a lesser extent, 2C9) catalyze biotransformation.

Evaluation in a rat amphetamine-induced hyperlocomotion (AHL) study, a traditional preclinical model of antipsychotic activity,^10,12−14 revealed that 6c (Figure 5) showed robust activity with a minimum effective dose (MED) of 10 mg/kg p.o. (43.2% reversal) that correlates with terminal (time = 1.5 h) plasma and brain concentrations of 1.8 μM (0.06 μM unbound) and 0.56 μM (0.02 μM unbound), respectively, and showing greater efficacy than our previously reported rat tool M₄ PAM, VU0467154.^13,14 Beyond hyperdopaminergic states, we also assessed the ability of 6c to reverse hyperlocomotion induced by the N-Methyl-d-Aspartate (NMDA) receptor antagonist MK-801 (Figure 6) to pharmacologically model NMDA receptor hypofunction and the associated prefrontal cortex-mediate impairments.^13,24,25 Here as well, 6c displayed a dose-responsive reversal of MK-801-induced hyperlocomotion, with an MED of 30 mg/kg p.o. and a 41.1% reversal.

Figure 5.

Compound 6c has antipsychotic-like activity in an AHL rat model. Compound 6c dose-dependently (1–10 mg/kg, po) reverses AHL (amphetamine, 0.75 mg/kg, s.c., *p < 0.05 vs vehicle + amphetamine, **p < 0.01 vs vehicle + amphetamine, ***p < 0.001 vs vehicle + amphetamine). N = 6–8 rats/group.

Figure 6.

Compound 6c has antipsychotic-like activity in an MK-801 hyperlocomotion rat model. Compound 6c dose-dependently (10−30 mg/kg, po) reverses MK-801-induced hyperlocomtion (MK-801, 0.18 mg/kg, s.c., *p < 0.05 vs vehicle + MK-801. N = 6−8 rats /group.

Based on the fact that 6c was the first M₄ PAM discovered in our program with similar M₄ PAM activity across all preclinical species and man, an attractive DMPK and ancillary pharmacology profile, as well as robust efficacy in a rodent model of antipsychotic activity, 6c was further profiled as a putative preclinical candidate, including evaluation in a rat modified Irwin test for effects on autonomic and somatomotor functions. In this study, following a single high dose (56.6 mg/kg, p.o., N = 6), no significant effects were observed over a 6 h observation period. However, based on these animal model acute concentration-effect data (rat AHL) and 6c’s predicted human PK profile (Table 3), efficacious human oral doses projected to provide 12-h daily coverage of the target/mechanism were undesirably high and frequent (e.g., >450 mg, TID).

Table 3. Predicted Human Pharmacokinetics of 6c.

parameter	value
CL (mL/min/kg)a	3.7–8.9
Vss (L/kg)b	1.5–2.1
t1/2 (h)	1.9–6.6
F (%)c	71
ka (1/h)d	0.55

^aPredicted by multispecies IVIVE.

^bPredicted by scaling of unbound V_ss from rat and dog.

^cPredicted by assumption of an optimized form/formulation providing an f_abs of 1.0 with an f_gut of 1.0 and f_hep of 0.71 (i.e., ER_hep = 0.29 based on mean predicted human CL/Q_hep).

^dPredicted by the MAT method using rat oral PK data from a 3 mg/kg dose of the mono-HCl salt formulated as a suspension in 0.1% tween80 and 0.5% methylcellulose in water.

Additionally, 6c displayed low aqueous solubility (2.4 μM at pH 7.4) and evidence for solubility-limited absorption in dog was observed, even at low doses and when administered as an HCl salt.

Still, 6c represents a major advance in the field, as the first potent M₄ PAM to overcome major species differences in potency while maintaining high selectivity versus M₂ (rat, dog, cyno, and human EC₅₀s > 30 μM), CNS penetration, and in vivo efficacy. However, given the projected human efficacious dosing and solubility issues, together with an anticipated challenge in achieving sufficiently high oral exposure in IND-enabling safety studies to establish acceptable margins, further advancement of 6c was halted (pending pharmaceutical sciences work), and optimization efforts shifted toward improvement of aqueous solubility and longer predicted human t_1/2 while retaining all the desirable properties of 6c. Results from this ongoing work will be reported in due course.

Glossary

ABBREVIATIONS

DCM

dichloromethane

AHL

amphetamine-induced hyperlocomotion

MED

minimum effective dose

mAChR

muscarinic acetylcholine receptor

PAM

positive allosteric modulator

HATU

O-(7-azabenzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate

DIEPA

N,N-diisopropylamine

MED

minimum effective dose

AHL

amphetamine-induced hyperlocomotion

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsmedchemlett.6b00461.

• General methods for the synthesis and characterization of all compounds, and methods for the in vitro and in vivo DMPK protocols and supplemental tables(PDF)

Author Contributions

^⊥ M.R.W. and M.J.N contributed equally to this work. C.W.L., T.M.B., M.R.W., P.J.C., C.M.N., N.J.B., M.E.D., and C.J.K. drafted/corrected the manuscript. B.J.M., K.D.N., M.S.P., M.A.H., and D.W.E. performed the chemical synthesis. C.W.L., M.W.R., T.M.B., P.J.C., C.M.N., C.J.K., M.J.N., N.J.B., M.W.W., M.E.D., and A.L.R. oversaw the target selection and interpreted the biological data. M.J.N., A.L.R., A.L., V.B.L., and R.L.W. performed the in vitro molecular pharmacology studies. T.M.B., A.L.B., and S.C. performed the in vitro and in vivo DMPK studies. C.K.J. and M.B. performed the in vivo experiments. All authors have given approval to the final version of the manuscript.

The authors declare the following competing financial interest(s): The authors are developing M4 PAMs for the treatment of schizophrenia and hold patents on the same.