Li et al. (2016) identified a PDE1 inhibitor clinical candidate, Compound 1 (ITI-214), via systematic pyrazolo-pyrimidinone scaffold optimizations. Compound 1 showed picomolar potency (Ki: 58 pM), excellent selectivity (2759-fold selective over PDE4), and good in vivo efficacy in the rat novel object recognition (NOR) test and has been nominated for clinical development. The optimization that led to ITI-214 involved changing substitution patterns at four distinct positions within the pyrazolo-pyrimidinone scaffold: N-2, C-3, C-6, and N-7. Fusing six-membered rings at positions C-6 and N-7 substantially improved PDE1 potency. However, reducing the size of the 6-membered rings to 5-membered rings significantly enhanced inhibitory activity and facilitated additional interactions with PDE1. The best improvement of potency and selectivity at this position was achieved with a cyclopentyl-fused 5-membered ring. Notably, incorporating a 2-(4-methyl-phenyl) pyridine sidechain at the N-2 position resulted in good PDE1 inhibitory potency and enhanced selectivity against PDE4. Amidst the SAR analysis, multiple substitutions were incorporated at the C-3 position. They discovered that larger alkyl groups, such as 2-methylbutyl-NH and cyclohexyl-NH, could enhance PDE1 inhibitory potency. These findings led to the identification of 2-(4-ethylphenyl)-6-fluoropyridine at the C-3 position as the optimal substitution. The combination of optimal substitutions at N-2, C-3, C-6, and N-7 as discussed above is epitomized by Compound 1 with a PDE1 Ki of 58 picomolar, which is over 100-fold more potent than other derivatives. ITI-214 also possesses 2759-fold selectivity for PDE1 over PDE4 (Figures 4A,B). Crystallographic X-ray examination of Compound 1 with PDE1B disclosed divergent interactions with the enzyme. The orientation of Compound 1 inside the catalytic cavity is inverted relative to that of cGMP. The pyrazolo-pyrimidinone core forms hydrophobic interactions and is stabilized by a “hydrophobic clamp.” The C-3 phenylamino group of the A-ring is deeply buried in a hydrophobic sub-pocket, establishing interactions with F392, L409, L388, and V417. The exocyclic C=O moiety of the B-ring and the C-3 amino group of the A-ring formed bidentate H-bonds with conserved Q421 (Figures 4A,B). The molecular weight (MW) of Compound 1 is 507.56 g/mol, exceeding the 500 g/mol threshold of Lipinski’s rules, suggesting potential synthetic challenges (Daina et al., 2017). Therefore, to efficiently synthesize Compound 1, Compound 2 was secured with a para-methoxybenzyl (PMB) group, followed by the replacement of chlorine to produce Compound 3. In the presence of DMF, Compound 3 reacted with POCl₃ to yield Compound 4. Subsequent benzylation of intermediate Compound 4, followed by its deprotection, resulted in the formation of N₂-substituted Compound 5. Direct amination of Compound 5 mediated by BOP afforded Compound 6. Further cyclization and chlorination of Compound 6 yielded Compound 7, which, through a palladium-catalyzed cross-coupling reaction, produced the target molecule, Compound 1 (Scheme 1) (Li et al., 2016).

The design and synthesis of thienotriazolopyrimidinone derivatives as PDE1 inhibitors were described later, focusing on various substitutions at positions C-6, C-8, and C-9 (Dyck et al., 2017). This led to Compound 13 (DNS-0056), a nanomolar (IC₅₀: 26 nM), orally bioavailable (90%), and brain-penetrating (maximum brain plasma ratio: 1.7) PDE1B inhibitor with good memory-enhancing effects in the rat NOR test. Initially, lead 8 was identified as a potential PDE1B inhibitor by screening a compound library. Further modifications, particularly the fusion of a cyclopropyl-pyridine-methanone ring at the C-8/C-9 positions and chlorobenzyl substitution at the C-6 position, led to Compound 9, exhibiting enhanced PDE1B inhibitory activity with moderate hERG activity. However, the incorporation of difluorocyclopropyl-pyridin-methanone and methoxybenzyl substitutions at positions C-6/C-8 and C-9, respectively, led to Compound 10, which not only inhibited PDE1B effectively but also exhibited potent inhibition of PDE10 (IC₅₀: 59 nM). Therefore, pyran-methyl-pyridine and methoxy-methylbenzene were substituted at the same positions to enhance selectivity toward PDE1B, leading to Compound 11. Improved potency and selectivity of Compound 11 toward PDE1B suggest that further substitution, such as difluoro-pyran-methyl-pyridine and methoxybenzyl at the C-6/C-8 and C-9 positions, can be employed to develop more potent PDE1B inhibitors. Such substitutions led to Compound 12 with excellent nanomolar inhibitory activity. However, Compound 12 was unfortunately identified as a substrate of multidrug resistance protein 1 (MDR1, P-gp) in Madin–Darby canine kidney (MDCK) cells with an efflux ratio greater than 5, which precluded its development as a candidate for CNS drugs. Ultimately, Compound 13 (DNS-0056) with pyran-methyl-pyridine and fluoro-methoxybenzyl substitutions at positions C-6/C-8 and C-9, respectively, emerged as the most potent derivative and was not a substrate of MDR1 (efflux ratio: 1.0) (Figure 5). Evaluation of Compound 8 co-crystallized with unveiled several significant interactions. The carbonyl group of Compound 8 formed a polar interaction with Q421 and H373. This orientation of Compound 8 within the catalytic domain positioned the 4-chlorobenzyl moiety within a newly created lipophilic pocket, arising from the rotation of M389. Additionally, the cyclohexyl ring of Compound 8 protrudes into a solvent-exposed region, contributing to its specificity toward other PDEs (Figure 6A). Furthermore, Compound 11 bound to the enzymatic region of PDE1B exhibited binding mode similarity with Compound 8, albeit with some differences, such as the pyran-methyl group of Compound 11 sticking out into a pocket made up of P408, F392, and T271. Additionally, the oxygen group of pyran formed a hydrogen interaction with Q421 (Figure 6B). The synthetic accessibility score and MW of DNS-0056 were 3.09 and 483.56 g/mol, respectively, lower than those of ITI-214; this indicates a comparatively simpler synthesis approach and fewer synthetic steps for DNS-0056 compared to ITI-214. The synthesis of DNS-0056 starts with treating Boc-protected piperidinones 14 with 15. Subsequently, utilizing pyridine under standard conditions, the triazolo-pyrimidinone ring was incorporated, yielding Compound 16. Compound 17 was obtained through N-alkylation. Substitution on the piperidine nitrogen atom was achieved through reductive alkylation under standard conditions, resulting in the target Compound 13 (Scheme 2) (Dyck et al., 2017).

The fusion of theophylline, ITI-214, and SCH-51866 backbones, followed by SAR exploration at N-7, C-2, and C-3 positions of the pyrazolo-pyrimidone core, resulted in the development of a series of pyrazolo-pyrimidone derivatives. Compound 18 (2j) emerged as a potent PDE1 inhibitor (PDE1B IC₅₀: 21 nM) with good metabolic stability in rat liver microsomes (T_1/2: 28.5 min). Detailed binding interaction analysis of ITI-214 and SCH-51866 with PDE1 revealed that the N₇-bound nitrogen in the pyrazolo-pyrimidone core acts as an H-bond acceptor, forming an H-bond with Y222 in the PDE1 catalytic site. As the substitution at this position was bulky, it was hypothesized that replacing it with smaller hydrogen bond acceptor groups could enhance inhibitory activity. Following this hypothesis, introducing a cyclopropylmethyl group at the N₇-position resulted in exceptional PDE1 inhibitory activity. In contrast, larger substitutions, such as the 2-oxo-2-phenylethyl group at the same position, yielded significantly reduced PDE1 inhibitory activity, indicating a preference for smaller groups at this position. However, the compound with a cyclopropylmethyl group at the N₇-position demonstrated poor microsomal stability, prompting structural optimization at the N₂-position to improve metabolic stability. It was anticipated that substituting saturated heterocycles such as oxetane at the N₂-position could modify metabolic pathways and enhance metabolic stability. Furthermore, substituting 4-(methylsulfonyl) benzyl at the N₂-position increased PDE1 inhibitory activity and improved microsomal stability. Moreover, the SAR investigation indicated that aerobic metabolism of the phenylamino ring at the C₃-position contributes to structural instability. Replacing this group with (4-fluorophenyl) amino moiety, along with optimal substitutions at the N-7, C-2, and C-3 positions discussed above, resulted in Compound 18, which exhibited the most potent (IC₅₀: 21 nM) and selective (over 480-fold versus PDE8) PDE1 inhibitor among the synthesized derivatives, along with good metabolic stability in rat liver microsomes (T_1/2: 28.5 min). Structural optimization revealed that incorporating small hydrophobic groups, such as cyclopropylmethyl, at the N₇-position enhanced potency, while a sulfonyl group at the N₂-position and (4-fluorophenyl) amino at the C₃-position improved metabolic stability (Figures 7A,B). Binding interaction analysis of Compound 18 with PDE1 revealed that the pyrazolo-pyrimidone core of Compound 18 positioned itself between F424 and L388, establishing H-bonds with Q421. The 4-substituted benzyl chain formed a hydrophobic bond with F424 and F427, while the C3-(4-fluorophenyl) amino side chain occupied a lipophilic pocket made up of F392, L409, and V417 (Figures 7A,B). The synthetic accessibility score and MW of 18 were 3.73 and 497.54 g/mol, respectively, which lie between those of ITI-214 and DNS-0056, indicating a moderate synthesis approach for Compound 18 compared to ITI-214 and DNS-0056 (Daina et al., 2017). The synthesis of 18 followed the same method used for ITI-214 with some modifications, starting with the protection of Compound 19 with a PMB group and subsequent chlorine displacement with hydrazine hydrate to obtain Compound 20. Phenyl isothiocyanate was then treated with Compound 20 and afforded the pyrazolo-pyrimidinone ring-containing Compound 21, which was treated with a substituted benzyl halide to afford Compound 22. Finally, the PMB group was eliminated from Compound 22, yielding an intermediate which, after reaction with halides in the presence of K₂CO₃, produced Compound 18 (Scheme 3) (Zhang et al., 2021).

Recent computational drug design efforts have discovered novel 2,3-dihydrobenzofuran scaffold-containing molecules, 27 (MDZ7) and 28 (MDZ12), as potential PDE1B inhibitors, exhibiting notable binding affinities of −9.6 and −9.5, respectively, with the protein. A pharmacophore model from the 5UOY–16j complex was used to screen the ZINC database, yielding 11,126 hits, which were narrowed down to 4 candidates, namely, compounds 23, 24, 25, and 26 (ZINC00552773, ZINC75914738, ZINC73091800, and ZINC15933417) using Lipinski’s rule, the PFS filter, and the SwissADME server (Figure 8). These hits were docked with five PDE1B crystal structures (PDB IDs: 5UOY, 5UP0, 4NPW, 4NPV, and 5B25) to analyze protein–ligand interactions. In the context of these co-crystal structures, Compound 24 consistently engaged key conserved hydrophobic residues (M336, L388, F392, and F424) and formed hydrogen bonds with Q421, while additional interactions with H223, Y222, I371, H373, I428, P408, L409, M389, and V417 contributed to specificity (Figures 9A–E).

Compound 24 exhibited various hydrophobic interactions with them, and its planar ring occupation within the active site suggested its potential as a lead molecule for further structural modification. The modifications focused on the 1,4-dioxin moiety and the isopropyl acetate bond as the 2,3-dihydrobenzofuran core and benzyl ring of the 2,3-dihydrobenzodioxin groups were crucial for preserving hydrophobic contacts with the P-clamp residues. Initially, based on the understanding that amine or amide groups could enhance hydrophobic and hydrophilic interactions, the isopropyl acetate linkage was replaced with these groups. Subsequently, to increase hydrogen bond interactions with Q421 and π-stacking interactions with F424 in PDE1B, the 1,4-dioxin group was replaced with phenol or phenoxyamine groups. Additionally, to enhance halogen interaction opportunities, a fluorine group was incorporated at the ortho site of the benzyl group in 2,3-dihydrobenzofuran. Interestingly, through these structural optimizations, compounds 27 and 28 were obtained as the highest-affinity compounds. In compounds 27 and 28, the 1,4-dioxin group was replaced with phenol, the isopropyl acetate linkage was replaced with amine and amide, respectively, and the benzyl ring of the 2,3-dihydrobenzodioxin groups was replaced with 6-fluoro and 5,6-difluoro substitutions, respectively. This suggests that the phenol ring is most favorable due to its significant hydrophilic and hydrophobic interactions, while the amine and amide groups, being polar, offer hydrogen bonding interactions. Additionally, the fluorine groups facilitate halogen interactions with the active site residues. Therefore, compounds 27 and 28 exhibited the most significant affinity for PDE1B among all the compounds, designating them as potential candidates for synthesis and preclinical assessment to evaluate their PDE1B inhibition and antipsychotic properties (Al-Nema et al., 2023). However, studies related to this have yet to be published.

Another important class of PDE1B inhibitors, containing the quinazoline nucleus, was discovered by Nadur et al. Initial high-throughput screening (HTS) led to the discovery of compounds 31 (PF-04471141) and 32 (PF-04822163) as potent, brain-penetrating, and selective PDE1B inhibitors. Based on their previous work on PDE10A, it was hypothesized that the 6, 7-dimethoxyquinazoline group of hits 1 and 2 can form a bidentate hydrogen bonding interaction between a PDE-invariant glutamine residue of PDE1B. Therefore, two different approaches focused on modifying the quinazoline structure. In the first approach, a series of 4-chloro-6,7,8-trimethoxyquinazoline (29) derivatives were synthesized and coupled with a variety of amines, which indicated that 7,8-dimethoxy functionalization on the quinazoline ring enhanced PDE1B inhibition, while the presence of a third 6-methoxy group was neither beneficial nor detrimental. The most potent compound from this series, Compound 31, exhibited an IC₅₀ value of 35 nM against PDE1B. In contrast, the second approach explored replacing the quinazoline nucleus with alternative templates such as phthalazine, cinnoline, and quinoline. Although these templates showed activity against PDE10A, they were ineffective against PDE1B, highlighting the essential role of the quinazoline core in PDE1B inhibition. Furthermore, the synthesis and SAR exploration of another series, 4-benzyl-7,8-dimethoxyquinazoline (30) derivatives, indicated that removing the chlorine and benzylic methyl group from the quinazoline structure reduced binding affinity toward PDE1B, whereas structural rigidification enhanced potency. Notably, transposing the methoxy group from the 6- to the 8-position on the quinazoline ring significantly increased potency, while replacing the chlorine atom with bromine or fluorine in the indane group decreased potency, likely due to size constraints within the binding pocket. This structural optimization led to the identification of Compound 32, a highly potent indane derivative with improved PDE1B inhibitory activity (PDE1B IC₅₀: 2.4 nM) (Figure 10). The quinazoline scaffold plays a critical role in PDE1B inhibition. In Compound 31, the 7- and 8-methoxy groups form a bifurcated hydrogen bond with the side chain nitrogen of Q421, while the N₁ atom engages His373. Hydrophobic interactions, including π–σ and π–π stacking with L388 and F424, further enhance potency (Figure 11A). Compound 32 binds similarly, with the chlorine atom occupying the hydrophobic pocket defined by L388 and F392, which may account for its slightly lower potency compared to Compound 31 (Figure 11B).

The synthetic accessibility score and MW of Compound 31 were 2.61 and 291.35 g/mol, respectively, marking the lowest values among the molecules discussed above, thereby indicating the most facile synthetic route. In addition, Compound 31 exhibits favorable pharmacokinetic properties, including high GI absorption and significant brain penetration, with no violations of Lipinski’s rules, which also supports its commercial availability from Sigma-Aldrich (Daina et al., 2017). The synthesis of Compound 31 began with the formation of intermediate 34 through the reaction of Compound 33 with ammonium acetate, followed by subsequent treatment with phosphoryl chloride. Intermediate 34 was then refluxed with triethylamine to afford the final product, Compound 31 (Scheme 4). Conversely, Compound 32, with a synthetic accessibility score of 3.23 and MW of 340.80 g/mol, higher than that of Compound 31, was synthesized via a multi-step reaction (Daina et al., 2017). The synthesis initiated with the direct chlorination of Compound 35, yielding chloroindane methyl ester 36, which, after chiral resolution using HPLC, afforded Compound 32 (Scheme 5) (Humphrey, 2014).

An HTS campaign and subsequent structural optimization of Compound 37, an initially identified PDE10A inhibitor, led to the discovery of imidazo[1,2-a] pyrazine-containing Compound 40, a potent and selective PDE10A inhibitor exhibiting promising in vivo efficacy across various rodent behavioral models of schizophrenia, along with favorable pharmacokinetic properties in rats. Compound 37, though effective, displayed limited selectivity toward PDE1, and this may be attributed to the presence of the trimethoxyphenyl moiety in Compound 37. As a result, the optimization process focused on substituting the trimethoxyphenyl group with various functional groups. Compound 38, containing a (methoxyethyl)pyrazole group, emerged as a successful substitution, enhancing selectivity (pIC₅₀ for other PDEs <5) while reducing potency toward PDE10A (pIC₅₀: 6.0), related to Compound 37. In order to enhance PDE10A potency, modifications were introduced to the pyrazole moiety, leading to the identification of Compound 39, which exhibited improved PDE10A inhibitory activity. To gain insight into the critical interaction characteristics of the imidazo[1,2-a] pyrazine series, crystallographic studies of the Compound 39 complex with PDE10A were analyzed. The binding interactions revealed a characteristic mode, where the morpholine ring expanded into the Q1 pocket, while the R-group on the pyrazole moiety projected into the solvent-exposed region, establishing hydrophobic interactions with F689, I682, and F719 (Figure 13A). To further enhance PDE10A potency, diverse substitutions were explored at the 2-position of the bicyclic imidazo[1,2-a] pyrazine scaffold. This effort resulted in the identification of Compound 40, which exhibited improved PDE10A inhibitory activity compared with other compounds (Figure 13B). Subsequent evaluation of these compounds for in vivo PDE10A inhibition using an apomorphine-induced stereotypy model in rats revealed that Compound 40 exhibited exceptional in vitro and in vivo potency, along with favorable microsomal stability. Given the encouraging results in the apomorphine-induced stereotypy model, Compound 40 was further tested for antipsychotic activity in a phencyclidine (PCP)-induced hyperlocomotion model in rats. Compound 40 effectively reversed PCP-induced hyperlocomotion. Additionally, protein binding in the rat brain was validated by the displacement of the selective PDE10A ligand [3H]MP-10 in the striatum. Finally, pharmacokinetic analysis revealed that Compound 40 possessed a well-balanced profile among all the derivatives. The synthetic accessibility score and MW of Compound 40 were 3.46 and 342.40 g/mol, respectively, indicating a relatively more complex synthetic route compared to other compounds (Daina et al., 2017). The synthesis began with commercially available starting material Compound 41, which was treated with ammonium hydroxide, followed by condensation with chloroacetone to yield intermediate 42. Selective bromination at position 3 was achieved using N-bromosuccinimide (NBS), resulting in intermediate 43. Subsequent reaction of intermediate 43 with morpholine led to the formation of intermediate 44. Finally, Compound 40 was synthesized via a Suzuki−Miyaura palladium-catalyzed cross-coupling reaction between intermediate 44 and 1-(2-methoxyethyl)-4-(4,4,5,5-tetramethyl[1,3,2]dioxaborolan-2-yl)-1H-pyrazole, completing the desired structure (Scheme 6) (Bartolomé-Nebreda et al., 2014).

The development of dihydro-imidazo benzimidazole-based PDE10A inhibitors was reported by Chino et al. (2019), focusing on overcoming the blood–brain barrier permeability issue identified in the previous molecule. Initially, Compound 45 demonstrated strong inhibitory activity against PDE10A, but poor brain penetration (Pint: 3.2 × 10⁻⁶ cm/s) limited its therapeutic potential. To address this, Compound 47 was developed through structural modifications, significantly reducing P-glycoprotein (P-gp) efflux liability, enhancing PDE10A inhibition, and improving brain penetration, thereby reversing MK-801-induced working memory deficits in preclinical models. SAR studies focused on modifying the pyrimidoindazole core. Various replacements, including aminoquinoline, quinoxaline, and quinazoline derivatives, were explored, in which the benzimidazole-based 46 demonstrated superior potency and permeability. Further refinement of the scaffold resulted in the identification of (S)-2-methyl-2,3-dihydro-1H-benzo[d]imidazo[1,2-a] imidazole containing Compound 47 (Figures 14A,B). The X-ray co-crystal structure of Compound 47 in complex with PDE10A enzyme demonstrated that the nitrogen atom at the 9-position of the imidazo[1,2-a]benzimidazole ring interacts with Y693. Additionally, the nitrogen group of the imidazo[1,2-a]benzimidazole ring engages in hydrogen bonding with Q726, which likely contributes to its enhanced potency. While Compound 47 showed improved the PDE10A inhibitory activity, it continued to exhibit significant P-gp liability (Figures 14A,B). In order to overcome this, efforts focused on modifying the nitrogen arrangement on the triazolopyrimidine ring. By varying the number and position of nitrogen atoms, particularly at key positions, researchers aimed to reduce P-gp interaction. Substitution with imidazopyrimidine and triazolopyridine moieties successfully lowered P-gp efflux but resulted in a marked decrease in PDE10A inhibitory potency, highlighting the importance of the 1- and 4-nitrogen atoms in maintaining both high inhibitory activity and minimal P-gp susceptibility. Finally, the triazolopyrazine-substituted Compound 48 (6d) exhibited balanced PDE10A inhibition and low P-gp liability, along with favorable pharmacokinetics (PK). It also showed minimal CYP enzyme inhibition, highlighting its therapeutic potential. The MW of Compound 48 was 361.40 g/mol, with a synthetic accessibility score of 3.74, slightly higher but still within a drug-likeness range, with a bioavailability score of 0.55. Synthesis began with the cyanation of chloromethyl-triazolopyrazine 49, followed by hydrolysis with NaOH to form intermediate 50. Subsequent treatment with WSC·HCl and (2S)-2-methyl-2,3-dihydro-1H-imidazo[1,2-a]benzimidazole yielded Compound 48 (Scheme 7) (Chino et al., 2019).

Chino et al. (2018) discovered a novel series of pyrimidoindazole derivatives as potent PDE10A inhibitors through fragment-based drug discovery (FBDD). The optimization began with Compound 51, a fragment hit with promising ligand efficiency (PEI: 0.39) for PDE10A. Structural modifications led to the development of Compound 52, a pyrimidoindazole derivative that exhibited potent PDE10A inhibition and favorable physicochemical properties. The X-ray co-crystallographic analysis revealed that the pyrazolopyrimidine ring of Compound 51 formed H-bonds with Q726 and hydrophobic interaction with the P-clamp in the PDE10A active site, although no interaction with Y693 in the selectivity pocket was observed (Figures 15A,B). To enhance binding to Y693 and improve inhibitory activity, the 4-chlorophenyl group at the 7-position was replaced with a hydroxyl group, aiming to reduce lipophilicity and promote stronger interactions. Subsequent modifications explored heteroaromatic substitutions with nitrogen atoms to target Y693. Among these, Compound 53, a triazolopyrimidine derivative, demonstrated a balance of strong PDE10A inhibition (97.3% inhibition at 4 µM) and improved hydrophilicity (log D7.4: 0.3), making it a candidate for further optimization. Further refinement of the triazolopyrimidine and pyrazolopyrimidine cores yielded Compound 52, which emerged as the most potent PDE10A inhibitor in the series, with an IC₅₀ value of 2.0 nM (Figure 16). Crystallographic analysis confirmed that the 4-nitrogen atom of Compound 53 formed a key hydrogen bond with Y693, enhancing selectivity and potency (Figures 15A,B). Synthesis of Compound 52 began with the cyclocondensation of Compound 54 and succinic anhydride 55, followed by condensation with acetylacetone to yield intermediate 56. The final product was obtained by reacting 56 with monoethyl potassium malonate and 3-aminoindazole in 1,4-dioxane (Scheme 8) (Chino et al., 2018).

In another study, Koizumi et al. (2019) identified 57 (MT-3014), a potent pyrazolo[1,5-a]pyrimidine derivative, as a promising PDE10A inhibitor with an IC₅₀ value of 0.062 nM. This compound emerged from a core structure transformation of a previously identified stilbene compound, 58, which exhibited high selectivity for PDE10A but was deprioritized due to concerns over E/Z isomerization and glutathione-adduct formation. Compound 57 was selected for further profiling in clinical trials based on its favorable pharmacokinetic properties and efficacy in the rat-conditioned avoidance response test. Structural optimization focused on designing various 6,5-fused heterocyclic derivatives, which identified Compound 59 with the best PDE10A inhibitory activity among other derivatives. Further modifications targeted the 5- to 8-positions of the quinoxaline ring, yielding Compound 60, which showed potent inhibition but exhibited strong hERG inhibition, necessitating additional optimization. Crystallographic analysis revealed that the quinoxaline core of Compound 60 formed key π–π interactions with F719 and F686 at the substrate-binding site of PDE10A. However, the steric clash between the 3-CH₃ of quinoxaline and 3-H of the pyrazolopyrimidine ring controlled the torsion angle and binding properties. Furthermore, the nitrogen atom of the pyrazolopyrimidine ring makes a polar hydrogen bond with the OH of Y683 (Figure 17A). Subsequent modifications at position-7 of the pyrazolo[1,5-a]pyrimidine core led to the discovery of derivative 61, which showed high PDE10A inhibition (IC₅₀ of 0.036 nM) and improved solubilities, although Compound 61 also demonstrated reduced hERG inhibition (29%). Introducing a fluorine atom at position 3 of the quinoxaline ring further enhanced the solubility and reduced the lipophilicity, leading to the identification of Compound 57, which balanced PDE10A inhibitory activity (IC₅₀ of 0.090 nM), low hERG inhibition (23% at 1 µM), and favorable solubility in artificial intestinal fluid (118 μg/mL) (Figure 17B). MT-3014 (57) was synthesized via a linear route, starting with the reaction of quinoxaline 62 with an acetonitrile anion to produce Compound 63 with a 77% yield. Compound 63 was then refluxed with hydrazine hydrate to yield intermediate 64, which reacted with N,N′-diisopropylcarbodiimide to afford Compound 65. Further reaction of Compound 65 with DMAP resulted in the formation of Compound 66. Chlorination of Compound 66 was performed using phosphorus(V) oxychloride, followed by successive treatment with methylated-ethanolamine and substituted pyrrolidine, which produced Compound 67. Finally, Compound 67 was reacted with an HCl solution and recrystallized to obtain Compound 57 as an HCl salt (Scheme 9) (Koizumi et al., 2019).

A novel chromone scaffold-containing molecule, Compound 68, has been introduced as a PDE10 inhibitor, demonstrating significant PDE10 inhibition with an IC₅₀ value of 6.5 nM, high selectivity (>95-fold over other PDEs), and excellent metabolic stability (RLM t_1/2: 105 min). The development of Compound 68 resulted from three rounds of structure optimization and critical analysis of binding patterns with PDE10. The chromone scaffold of Compound 69 interacts with conserved Q726 via hydrogen bonding and with a hydrophobic clamp formed by F729 and I692/F696 through π–π stacking interactions. Additionally, the ethyl linker extends into the Q2 pocket, allowing the thiazol-5-yl methyl benzimidazole motif of 1 to form a hydrogen bond with Y693. As is known, the Q2 pocket provides unique selectivity to PDE10 inhibitors; therefore, structural optimization focused on the benzimidazole moiety was employed to enhance selectivity. The first-round optimization led to Compound 70, which contains a 5-methyl-1-phenyl benzimidazole moiety and exhibited an excellent increase in potency (IC₅₀: 52 nM). Compound 70 showed a 10-fold increase in potency compared to Compound 69, indicating that the phenyl group at the N₅-position of the benzimidazole moiety was preferred as it fit well into the Q2 pocket. The binding interaction analysis of Compound 70 with PDE10 showed that the 5-methyl-1-phenyl benzimidazole group of Compound 70 fit properly into the Q2 pocket. The oxygen atom of pyranone and the nitrogen atom of the benzimidazole group formed an H-bond with F726 and Y693 (Figure 18A). Further optimization at the C₆-position of the chromone ring proposed that the halogenation at this position could enhance potency due to proximity (3.8 Å) to the phenolic hydroxyl group of Y524. The second-round optimization introduced chlorine at the C₆-position, resulting in Compound 71, which showed 2–3 times increased inhibitory activity compared to Compound 70. The co-crystal structure of Compound 71 bound to PDE10A revealed that, in addition to existing interactions, an extra halogen bond formed between the chlorine atom of Compound 71 and Y524, contributing to enhanced inhibitory activity (Figure 18B). The third-round optimization aimed to improve the metabolic stability and pharmacokinetic properties of Compound 71. Molecular docking suggested that a vinyl group, rather than an ethyl linker, would better extend into the Q2 pocket, resulting in molecules with improved metabolic stability and inhibitory potency. The extra halogen bond with Y524 indicated the C₆-position of the chromone scaffold as a probable metabolic site. Introducing a fluorine atom at this position yielded Compound 68, which demonstrated slightly improved inhibitory potency (IC₅₀: 6.5 nM) and significantly enhanced metabolic stability (RLM t_1/2: 105 min) (Figure 19). The binding mode of Compound 68 with PDE10A showed that the vinyl linker and benzimidazole group fit well into the narrow Q2 pocket, while the chromone scaffold engaged the hydrophobic clamp (F729 and F696/I692). Notably, H-bonds with Q726 and Y693, along with a halogen bond with Y524, contributed to the strong inhibitory potency of Compound 68. Compound 68, with a synthetic accessibility score of 3.27, indicates a moderately challenging synthesis. The synthesis of Compound 68 began with a Knoevenagel−Doebner condensation between the previously synthesized Compound 72 and malonic acid, yielding Compound 73. Subsequent condensation of intermediate 73 with 4-chloro-N1-phenylbenzene-1,2-diamine, followed by cyclization, produced Compound 68 (Scheme 10) (Yu et al., 2020).

In a recent study, Al-Nema et al. (2022b) identified 74 (Zinc42657360), containing a cyclopenta[4,5]thieno[2,3-d]pyrimidin-4-one scaffold, as a promising new PDE10A inhibitor. A systematic approach included a structure-based drug design program, combining pharmacophore modeling, molecular docking, and molecular dynamics simulations, followed by biological evaluation through the PDE-Glo phosphodiesterase assay, leading to the discovery of potent Compound 74, with a PDE10A IC₅₀ value of 1.60 µM. The development of Compound 74 involved a comprehensive virtual screening process. The screening identified six key pharmacophoric features in the co-crystallized structure of PDE10A (PDB ID: 5UWF) and the ligand 16d, which included one hydrogen bond acceptors (HBA) interacting with Q726, an aromatic ring (AR) represented by the thiophene moiety, and four hydrophobic areas (HA) represented by phenol, sulfide, and fluorine groups. A series of multi-step virtual screenings, such as pharmacophore-based screening, drug-likeness analysis, and pharmacophore fit score, resulted in 7541, 516, and 14 hits, respectively, which were subsequently subjected to the PAINS filter. These compounds passed the PAINS filter and were subjected to molecular docking studies. The docking studies revealed that compounds 74 and 75 (Zinc47464611) and the standard TAK063 exhibited the lowest binding energies, indicating high affinity for PDE10A. Structural analysis of Compound 74 with PDE10A showed that its cyclopenta-thiophene moiety occupies the P-clamp region, stabilized by aromatic interactions with the pyrimidine substituents. Notably, Compound 74 forms unique hydrogen bonds with Y524, Q726, and H525 and a coordination interaction with Mg, which collectively contribute to its high selectivity and affinity for PDE10A (Figure 20A). Compound 75 also engages the P-clamp region and forms aromatic interactions with F726, F696, and I692, along with interactions with H525 (Figure 20B). While both compounds 74 and 75 share P-clamp occupancy and aromatic contacts, Compound 74 distinguished itself by establishing unique hydrogen bonds.

Furthermore, the inhibitory potency of Compound 74 (IC₅₀: 1.6 µM) was observed to be relatively weak compared with standard PDE10A inhibitors such as TAK-063 (IC₅₀: 0.34 nM). Although the study provides a strong foundation for initial hit identification, it lacks critical follow-up characterization, including selectivity profiling against other PDE isoforms and essential ADME (Absorption, Distribution, Metabolism, and Excretion) parameters (e.g., aqueous solubility, microsomal stability, and plasma protein binding) required to assess the scaffold’s overall drug-likeness and development. Given the observed hydrogen-bonding interactions with Q726 and Y524 and considering the contribution of the P-clamp region to PDE10A selectivity, future optimization should prioritize (i) enhancing potency by incorporating structural constraints that properly orient key substituents to reinforce the Q726/Y524 hydrogen-bond network and (ii) modulating lipophilicity and polar surface area to achieve an optimal balance between CNS penetration and effective P-clamp engagement. The study demonstrated that a multistep in silico workflow (Figure 21), followed by in vitro analysis, enabled the identification of Compound 74; however, the scaffold still requires comprehensive PK and preclinical assessment to establish its suitability as a suitable drug candidate (Al-Nema et al., 2022b).

Layton et al. (2023) described a potent 2-methyl-4-amino-6-methoxypyrimidine inhibitor selective for PDE10A. Their earlier structural optimization efforts led to the identification of Compound 76, a potent PDE10A inhibitor, with poor physicochemical properties and off-target activities. Furthermore, the rational design, combined with synthetic techniques and guided by inhibitor-bound X-ray crystal structures, led to the discovery of Compound 77 (MK-8189) with a PDE10A Ki value of 1.6 nM and PDE selectivity >500,000. A novel scaffold, Compound 76, was identified through a fragment screening as a PDE10A inhibitor with high ligand binding efficiency (LBE). It was optimized using structure-based design and parallel library synthesis, leading to the identification of Compound 78. Compound 78 showed excellent potency, engaging key residues such as Y683 in the PDE10A selectivity pocket, but it still exhibited a poor pharmacokinetic profile, low aqueous solubility, and off-target ion channel activity against hERG. To address these challenges, further optimization targeted the ether linkages, resulting in Compound 79 with a methyl-pyrazole moiety. This step improved metabolic stability but still needed enhancement in CYP inhibition and clearance. Hypothesizing that the chloro-methyl-pyrimidine core was responsible for the remaining issues, they explored replacing it with a bicyclic core, leading to Compound 80. This bicyclic pyrazolopyrimidine core improved permeability and minimized P-gp transporter liability. Subsequent optimization focused on improving potency by replacing the N-methyl-pyrazole with a 2-methylpyrimidine, yielding Compound 81. This modification led to a notable improvement in selectivity, solubility, and pharmacokinetic properties, although with a slight compromise in potency. Efforts to improve potency while retaining the optimal profile led to the most balanced Compound 77, which combines a methyl-pyridine and methyl-1,3,4-thiadiazole motif, achieving high potency, selectivity, and favorable PK, making it a lead candidate for further evaluation. Compound 77 demonstrated functional inhibition of PDE10A with a Ki value of 0.029 nM and over 500,000-fold selectivity over other PDE enzymes, making it a promising PDE10A inhibitor (Figures 22A,B). The co-crystal structure of Compound 77 with PDE10A reveals its binding interaction. The pendant 5-methyl-pyridine forms an H-bond with Y683 in the “selectivity pocket” of PDE10A, while the pyrimidine core of Compound 77 establishes π–π stacking with the side chain of F719. Additionally, a hydrogen bond is observed between N₁ nitrogen and the side chain of Q716. The N₃ nitrogen of the pyrimidine core is ideally positioned to form additional interactions with the binding site of PDE10A, enhancing the potency of Compound 77 (Figures 22A,B). The synthetic accessibility score and MW of Compound 77 were 4.13 and 382.48 g/mol, respectively, suggesting a more complicated synthetic strategy compared with other compounds (Daina et al., 2017). The synthesis started with pyrimidine 82, which was exposed to a microwave-assisted reaction with (5-methyl-1,3,4-thiadiazol-2-yl)methanamine hydrochloride, yielding intermediate 83. This intermediate was then treated with a mixture of ((1S,2S)-2-(5-methylpyridin-2-yl)cyclopropyl)methanol and tetrahydrofuran (THF), which yielded Compound 77 (Scheme 11). The SAR and mechanistic insights emerging from the MK-8189 highlight several key design principles. First, the pyrimidine-derived hinge-binding core provides a rigid π-stacking framework, most notably interacting with F719, while orienting heteroatoms to form essential H-bond interaction within the PDE10A catalytic pocket. Second, the pendant 5-methylpyridine, a motif preserved throughout late-stage optimization, projects deeply into the PDE10A selectivity pocket, forming a critical H-bond with Y683, which is a key contributor to the exceptional isoform selectivity of MK-8189. Third, systematic replacement of flexible ether linkages and metabolically vulnerable substituents with bicyclic and conformationally constrained motifs produced substantial improvements in physicochemical and ADME parameters, including increased solubility, enhanced microsomal stability, reduced CYP inhibition, and lower P-gp liability, while maintaining sub-nanomolar to picomolar potency (Layton et al., 2023).