Discovery of MK-8189, a Highly Potent and Selective PDE10A Inhibitor for the Treatment of Schizophrenia

Phosphodiesterases regulate cAMP/cGMP signaling in neurons. PDE10A is highly expressed in the striatum, and abnormal striatal output is implicated in schizophrenia. By increasing striatal cAMP/cGMP, PDE10A inhibitors could ameliorate positive symptoms, cognitive impairment, and possibly negative symptoms. Current atypical antipsychotics have limited efficacy for negative/cognitive symptoms and carry tolerability liabilities. Preclinical data suggested PDE10A inhibition could address these limitations. The study aimed to discover and optimize potent, selective, and brain-penetrant PDE10A inhibitors with favorable physicochemical, safety, and pharmacokinetic properties suitable for oral dosing and CNS activity, culminating in a clinical candidate for schizophrenia.

The paper situates PDE10A among 11 PDE families, emphasizing its striatal localization and role in schizophrenia pathophysiology. Prior work identified fragment-derived pyrimidine PDE10A inhibitors with high ligand binding efficiency but with poor PK, solubility, hERG liability, and CYP interactions. The clinical landscape notes limitations of atypical antipsychotics and mixed outcomes of earlier PDE10A inhibitors in clinical trials. Preclinical studies support antipsychotic-like and pro-cognitive effects with a differentiated side-effect profile versus D2 antagonists.

- Lead identification and optimization: Began from a fragment (2-chloro-3-methyl pyrimidine) with high ligand binding efficiency. Structure-based design using inhibitor-bound X-ray crystal structures and parallel library synthesis drove potency to picomolar range. Addressed liabilities by exploring western ether and eastern heteroaryl substituents and by swapping the central core (bicyclic pyrazolopyrimidines and multiple naphthyridine isomers; ultimately an isomeric 2-methyl-pyrimidine core). - Synthetic chemistry: Detailed multi-step syntheses for cyclopropyl-methanol intermediates and target compounds via Mitsunobu ether formation, SNAr displacement, palladium-catalyzed C–O and C–N cross-couplings, Corey–Chaykovsky modification, and chiral SFC resolution to obtain (S,S) alcohols. Libraries of monocyclic/bicyclic cores and eastern/western substituents were prepared; final MK-8189 (compound 18) synthesized by sequential SNAr (amine then alcohol) on 4,6-dichloro-2-methylpyrimidine. - In vitro enzyme assays: PDE activity measured via IMAP fluorescence polarization assays. Apparent Ki values determined for PDE1–PDE11 using defined Km values per enzyme/substrate. - Cellular assay: IC50 measured in cells recombinantly expressing full-length human PDE10A. - SPR binding kinetics: Biacore T-200 with His-tagged PDE10A immobilized on CM5-NTA; association/dissociation profiles obtained over 10 concentrations; kinetic parameters extracted. - Crystallography: PDE10A catalytic domain (residues 439–779) co-crystallized/soaked with compound 18; structure solved at 2.1 Å (PDB 8DI4) using autoPROC and BUSTER. - Physicochemical profiling: Molecular weight, LogD, ligand binding efficiency (LBE), ligand-lipophilicity efficiency (LLE), aqueous solubility (HTS at pH 7; FaSSIF and SGF solubility). - Off-target/safety profiling: Ion channel panel (Iks, Cav1.2, Nav1.5; functional hERG Ikr), broad receptor binding panel (Panlabs), genotoxicity (Ames; chromosome aberration in CHO cells), PXR activation, reversible CYP inhibition (CYP2C9, CYP3A4). - Transport and permeability: Passive permeability and P-gp substrate assessment (bidirectional transport, BA/AB ratios) in human/monkey/rat. - Pharmacokinetics: IV and PO studies in Wistar–Hannover rats and rhesus monkeys with serial plasma sampling; noncompartmental analysis for CLp, Vdss, t1/2, and bioavailability. - Target engagement (occupancy): Ex vivo PDE10A occupancy in rat striatum using [3H]MK-8193 displacement after oral dosing of 18; occupancy calculated from specific binding reduction vs vehicle and related to plasma concentrations. - In vivo efficacy models: MK-801-induced hyperlocomotion assay (antipsychotic-like effect) and novel object recognition (episodic memory) in rats; dose- and exposure-response related to PDE10A occupancy. - Tolerability observations: Prolactin levels, weight gain, and catalepsy window (data referenced).

- Compound 18 (MK-8189) emerged from an isomeric 2-methyl-pyrimidine series with optimal balance of potency, selectivity, PK, solubility, and safety. - Enzymology and cellular potency: Human PDE10A functional Ki = 0.029 nM; cellular IC50 = 1.6 nM; >500,000-fold selectivity over PDE1–PDE11. - Binding kinetics (SPR): kon ≈ 8.0 × 10^7 M−1·s−1; in vitro enzyme residence half-life ≈ 6.7 min at 25°C; reversible binding. - Structural biology: X-ray co-complex (2.1 Å; PDB 8DI4) shows 5-methyl-pyridine engages Tyr683 (selectivity pocket) via H-bond; pyrimidine core π-stacks with Phe719 and H-bonds (N1 to Gln716); N3 and 4-amine linker form water-mediated H-bonds with Ser667 and Tyr514. - Physicochemical efficiency: MW = 382; LogD = 2.1; LBE = 0.54; LLE = 7.8; high-throughput solubility at pH 7 = 167 μM; free-base solubility = 0.17 mg/mL in FaSSIF and 5 mg/mL in SGF. - Pharmacokinetics: Rat CLp 21.3 mL/min/kg, Vdss 0.93 L/kg, t1/2 4.8 h, F 46% (2 mg/kg IV; 10 mg/kg PO). Rhesus CLp 9.3 mL/min/kg, Vdss 1.4 L/kg, t1/2 4.2 h, F 41% (0.5 mg/kg IV; 5 mg/kg PO). Plasma protein binding high; fu: rat 8.2%, monkey 8.7%, human 4.0%. - Drug–drug interaction and nuclear receptor: Not a potent reversible inhibitor of CYP2C9 or CYP3A4 (IC50 ≥ 50 μM); inactive in PXR assay (EC50 > 30 μM). - Safety/off-target: Ion channels Iks, Cav1.2, Nav1.5 > 30 μM; functional hERG Ikr IC50 = 33 μM. One notable off-target: SSTR2 binding IC50 = 2.8 μM. Negative in Ames (5-strain) and CHO chromosomal aberration assays. - Transport/permeability: High passive permeability (35.4–42.6 × 10−6 cm/s). Not a substrate of human/monkey P-gp (BA/AB < 2). Weak rat P-gp substrate (BA/AB ≈ 2.6) but achieved full striatal occupancy in rats. - Target engagement: In rats, 50% PDE10A occupancy at plasma ≈ 52 nM; full occupancy at higher concentrations. - In vivo efficacy: MK-801-induced hyperlocomotion attenuated dose-dependently at 0.25–0.75 mg/kg PO with end-of-study plasma 17–50 nM; efficacy seen at ~25% occupancy and above. Novel object recognition improved at 0.16–0.25 mg/kg with plasma 18–29 nM; significant effects at ~29% occupancy and above. Efficacy windows similar across psychosis and cognition models. - Tolerability (preclinical): No increase in circulating prolactin; reduced weight gain in rats; larger therapeutic window to catalepsy relative to atypical antipsychotics (data referenced). - Clinical status: MK-8189 generally well tolerated in Phase 1 at doses achieving high trough PDE10A occupancy; initial proof-of-concept in acute schizophrenia provided evidence for antipsychotic efficacy; ongoing Phase 2b development.

The study demonstrates that rational, structure-guided optimization from a fragment hit can yield an exceptionally potent and selective PDE10A inhibitor with a superior developability profile. By shifting to an isomeric 2-methyl-pyrimidine core and optimizing eastern/western substituents, the team overcame earlier liabilities (hERG, CYP inhibition, PXR activation, poor solubility and PK). Structural insights (Tyr683 selectivity pocket interactions and additional H-bonding via the N3 of the pyrimidine) explain the improved potency/selectivity and ancillary profile of MK-8189. The compound shows strong CNS drug-like properties (high permeability, minimal P-gp efflux in human/monkey), robust target engagement in vivo, and translation of occupancy to efficacy in rodent models predictive of antipsychotic and pro-cognitive effects. Safety and off-target data suggest a differentiated profile versus dopamine D2 antagonists, aligning with the therapeutic hypothesis that PDE10A modulation may treat schizophrenia symptoms with improved tolerability. Early clinical data showing high sustained PDE10A occupancy and antipsychotic signal further supports the mechanism and the compound’s potential.

MK-8189 (compound 18) is a potent, highly selective, and brain-penetrant PDE10A inhibitor with favorable physicochemical, PK, safety, and efficacy profiles. The integrated medicinal chemistry, structural biology, and pharmacology program transitioned an early fragment into a clinical candidate that demonstrates robust target engagement and efficacy in relevant rodent models, with improved tolerability markers. MK-8189 has progressed through Phase 1 with good tolerability and shows initial antipsychotic efficacy in acute schizophrenia; it is currently in Phase 2b clinical development. Future work should further characterize efficacy across symptom domains, evaluate long-term safety/tolerability, and explore biomarker-driven dose optimization tied to PDE10A occupancy.

- Most efficacy data presented are preclinical; detailed clinical efficacy and safety outcomes are not reported in this paper. - A measurable off-target interaction with somatostatin receptor 2 (SSTR2; IC50 ~2.8 μM) was identified, the clinical relevance of which remains to be determined. - MK-8189 is a weak substrate for rat P-gp (BA/AB ~2.6), though not for human/monkey; interspecies differences could affect translation, albeit full occupancy was achieved in rats. - Prior PDE10A inhibitors have had mixed clinical results; while initial proof-of-concept is encouraging, broader efficacy and long-term outcomes require confirmation in larger clinical trials.