Discovery and SAR of novel series of imidazopyrimidinones and dihydroimidazopyrimidinones as positive allosteric modulators of the metabotropic glutamate receptor 5 (mGlu₅)

Abstract

We report the discovery and SAR of two novel series of imidazopyrimidinones and dihydroimidazopyrimidinones as metabotropic glutamate receptor 5 (mGlu₅) positive allosteric modulators (PAMs). Exploration of several structural features in the western and eastern part of the imidazopyrimidinone core and combinations thereof, revealed compound 4a as a mGlu₅ PAM with good in vitro potency and efficacy, acceptable drug metabolism and pharmacokinetic (DMPK) properties and in vivo efficacy in an amphetamine-based model of psychosis. However, the presence of CNS-mediated adverse effects in preclinical species precluded any further in vivo evaluation.

Schizophrenia is a devastating psychiatric illness that afflicts approximately 1% of the world population. Current therapies, based on the direct modulation of the monoaminergic system (dopamine and serotonin), are associated with severe adverse effects and do not provide sufficient relief across all symptom domains of the disease.^1,2 Therefore, new treatments, possibly acting via different neurotransmitter systems are under investigation. Among the different existing alternatives, the modulation of the glutamatergic system has emerged as an obvious candidate.^3,4 Glutamate (l-glutamic acid) is the major excitatory neurotransmitter in the mammalian central nervous system (CNS), where it mediates its effects through three ionotropic receptors (d,l-α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA), N-methyl-d-aspartate (NMDA) and kainate) and eight metabotropic glutamate receptors (mGlu₁ to mGlu₈).⁵ Among the different mGlus, the mGlu₅ receptor is structurally and pharmacologically associated with the NMDA receptor playing prominent roles in synaptic plasticity, cognition, learning and memory process^1,6 and it has been reported that the activation of mGlu₅ potentiates NMDA receptor function.⁷ Provided the advantages associated to PAMs over orthosteric agonists, that is, increased subtype selectivity, improved chemical tractability, better modulatory control of receptor function and low propensity for receptor desensitization; the potentiation of mGlu₅ via activation at allosteric sites could be a preferred pharmacological mechanism to indirectly correct the hypothesized NMDA receptor hypofunction in schizophrenia.^2,8 Consequently, the development of PAMs of mGlu₅ has emerged as a potentially viable way to treat positive symptoms^9,10 and also improve cognitive deficits^11,12 in schizophrenia.

After nearly a decade of mGlu₅ PAMs research, a variety of structurally different chemotypes have been reported.¹³ Among them, we have recently described the discovery of phenoxy-containing mGlu₅ PAMs derived from dihydrothiazolopyridone¹⁴ and naphthyridinone¹⁵ bicyclic cores. As a progression of this work, and in order to further optimize the in vitro and in vivo potencies of dihydrothiazolopyridone 1, we envisioned the replacement of the thiazolyl ring system by other aromatic five-membered rings (Fig. 1).

Figure 1.

Evolution strategy of dihydrothiazolopyridone 1 to other 5,6-bicyclic systems.

Our initial approach involved the direct replacement by an oxazolyl ring affording dihydrooxazolopyridones 2. Encouraged by a comparable in vitro mGlu₅ PAM activity to dihydrothiazolopyridone 1 series, the most potent analogues were tested in the reversal of amphetamine-induced hyperlocomotion (AIH) in rats, a predictive model of antipsychotic activity.^10a,16 Unfortunately, none of the analogues tested showed robust in vivo activity after oral dosing. This lack of relevant in vivo activity could be explained by the limited absorption observed for these derivatives and thus, our subsequent exploration focused on the identification of other chemotypes with different properties (i.e., electronics, basicity, lipophilicity, etc.) that could eventually lead to an improvement in the in vivo efficacy. In this Letter we report our efforts towards the synthesis and optimization of imidazopyrimidinones 3 and dihydroimidazopyrimidinones 4 as novel series of mGlu₅ PAMs.

Synthesis of these 5,6-bicyclic systems was achieved following two parallel approaches that allowed the introduction of diverse substituents in the eastern and western part of the targeted molecules, as shown in Schemes 1 and 2. Our first strategy was focused on the preparation of the key compound 3a, with the aim to survey the eastern diversity in the last step (Scheme 1). Condensation of chloropyrimidine 5 with 1,3-dichloroacetone under acidic conditions, following a procedure already reported in the literature afforded intermediate 6.¹⁷ Introduction of phenoxy group in the western part of the molecule using potassium carbonate as base was followed by the selective hydrolysis of the thiomethyl group yielding intermediate 7. De-chlorination was performed under hydrogenation conditions affording analogue 3a. This key compound was then derivatized to the desired N-aryl imidazopyrimidinones 3e and 3h–3n, by reaction either with p-fluoroiodobenzene (3e) or various heteroaryl halides (3h–3n) using a copper iodide–diamine ligand protocol to perform a modified Ullman-type lactam cross-coupling.

Scheme 1.

Reagents and conditions: (a) 1,3-dichloroacetone, AcOH, 110 °C, 16 h, 44%; (b) PhOH, K₂CO₃, ACN, 50 °C, 18 h, 53%; (c) LiOH, H₂O/THF, 50 °C, 4 h, 98%; (d) Pd(OH)₂/C, H₂ atm, Et₃N, AcOEt/MeOH, 50 °C, 45 psi, 16 h, in a Parr reactor, 92%; (e) R¹X, CuI, N,N′-dimethylethylenediamine, K₂CO₃, toluene, 120 °C, 16 h, 22–73%.

Scheme 2.

Reagents and conditions: (a) for R¹ = Alkyl: R¹X, BuN₄OH, DMF, rt, 2 h, 57–77% and for R¹ = Aryl or heteroaryl: R¹B(OH)₂, Cu(OAc)₂, TMEDA, MeOH/H₂O, rt, 16 h, 48–65%; (b) 1,3-dichloroacetone, DMF, 150 °C, 30 min, microwave irradiation, 15–75%; (c) R²OH, K₂CO₃, ACN, 90 °C, 18 h, 58–95%; (d) I₂, NaIO₄, H₂SO₄, AcOH/water, 80 °C, 4 h, 84%; (e) MeB(OH)₂, Pd(PPh₃)₄, K₂CO₃, 1,4-dioxane/DMF, 150 °C, 45 min, microwave irradiation, 72%.

Unfortunately, functionalization in the eastern amide moiety with alkyl substituents or other aryl halides following the same Ullman-type conditions was unsuccessful and for this purpose, the approach outlined in Scheme 2 was used. This second strategy allowed the introduction of several alkyl and aryl moieties and also to survey the substitution in the western part of the molecule. Furthermore, this approach could be additionally applied to the introduction of small groups in the 7 and 8-positions of the imidazopyrimidinone core (Scheme 2). Thus, for the functionalization with alkyl and aryl groups in the eastern part of the molecule we started from commercially available cytosine derivatives 9. The reaction of 9 with a variety of alkyl halides using tetrabutylammonium hydroxide as base allowed the selective alkylation on the lactam. On the other hand, for the functionalization with aryl moieties a Chan-Lam-type copper-catalyzed coupling reaction between cytosine derivatives 9 and the corresponding aryl boronic acids as depicted in Scheme 2 was followed.¹⁸ Then, condensation with 1,3-dichloroacetone followed by introduction of the corresponding phenoxy or pyridyloxy group in the western part of the molecule, via reaction with the corresponding phenols or hydroxypyridines using potassium carbonate as base, allowed the preparation of analogues 3b–3g, 3o–3s and 3u. The 8-methyl analogue 3v was prepared by Suzuki-type coupling reaction between methylboronic acid and intermediate 13, synthesized following a similar synthetic route as described before but starting from the corresponding substituted iodocytosine 12 as shown in Scheme 2.

Hydrogenation of the double bond in analogues 3 was performed using palladium hydroxide on carbon or Raney nickel as catalysts at high temperatures (Scheme 3). All attempts to reduce the double bond in compounds with substitution in the 7 or 8-positions different than hydrogen were unsuccessful. Finally, functionalization in the imidazole ring was achieved by halogenation of the corresponding derivatives 3 and 4 with N-chlorosuccinimide (NCS) or N-bromosuccinimide (NBS). Subsequent Suzuki-type coupling reaction of 14 with methylboronic acid yielded analogue 4l as depicted in Scheme 3.

Scheme 3.

Reagents and conditions: (a) Pd(OH)₂/C, H₂ atm, MeOH, 50 °C, 50 psi, 16 h, in a Parr reactor or Raney-Ni, 80 °C, full hydrogen mode, DMF/MeOH in an H-Cube reactor, 5–78%; (b) NBS, benzoyl peroxide, DCE, rt, 16 h, 35%; (c) MeB(OH)₂, Pd(PPh₃)₄, K₂CO₃, 1,4-dioxane/DMF, 150 °C, 45 min, microwave irradiation, 56%; (d) NCS, DMF, 150 °C, 15–20 min, microwave irradiation, 27–55%.

The compounds synthesized were profiled in a human mGlu₅ low receptor expression cell line using a ‘triple add’ calcium mobilization assay, allowing the detection of agonism as well as positive and negative allosteric modulation simultaneously.¹⁹ The most relevant SAR trends for the imidazopyrimidinones 3 are shown in Table 1. Initial exploration was focused on the introduction of substituents in the amide moiety with the aim to further improve the marginal activity observed for the unsubstituted analogue 3a (EC₅₀ >30 µM, 19% Glu Max). Following a similar strategy to the dihydrothiazolopyridone series (1),¹⁴ we first targeted the introduction of alkyl substituents. Thus, N-methylation in compound 3b, despite having a positive effect in efficacy (67% Glu Max) did not improve the in vitro potency (EC₅₀ >10 µM), which was still significantly decreased compared with analogue 1. According to previous work,¹⁴ the functionalization with an alkyl chain of increased size such as cyclopropylmethyl (3c) interestingly induced a significant improvement in the in vitro mGlu₅ PAM activity, resulting in a potency in the same range as that of the dihydrothiazolopyridone prototype 1, although with reduced efficacy. Comparable potency again but significantly improved efficacy (93% Glu Max) was achieved by the introduction of a bulkier aromatic benzyl group in 3d.

Table 1.

Structures and activities of substituted imidazopyrimidinones 3

Compd	R1	R2	R3	R	mGlu5 pEC50a (±SEM)	mGlu5 EC50 (nM)	% Glu maxb (±SEM)	HLMc (%)	RLMc (%)
3a	H		H	H	<4.52d	>30,000	19	n.t.e	n.t.e
3b			H	H	<5d	<10,000	67 ± 4.2	n.t.e	n.t.e
3c			H	H	5.95 ± 0.03	1120	47 ± 3.7	n.t.e	n.t.e
3d			H	H	5.80 ± 0.06	1590	93 ± 5.0	n.t.e	n.t.e
3e			H	H	6.51 ± 0.15	310	83 ± 9.1	14	27
3f			H	H	6.29 ± 0.02	514	69 ± 0.9	47	71
3g			H	H	6.64 ± 0.05	230	93 ± 11.7	n.t.e	n.t.e
3h			H	H	6.23 ± 0.05	587	87 ± 10.2	n.t.e	n.t.e
3i			H	H	5.69 ± 0.14	2000	84 ± 15.4	n.t.e	n.t.e
3j			H	H	5.45 ± 0.11	3580	64 ± 10.3	n.t.e	n.t.e
3k			H	H	5.29 ± 0.07	5100	64 ± 9.0	n.t.e	n.t.e
3l			H	H	6.55 ± 0.08	284	87 ± 13.6	66	90
3m			H	H	6.43 ± 0.05	375	77 ± 8.0	8	50
3n			H	H	6.36 ± 0.07	435	77 ± 11.9	17	49
3o			H	H	6.64 ± 0.05	230	91 ± 11.1	13	28
3p			H	H	6.43 ± 0.06	375	85 ± 11.6	11	29
3q			H	H	6.46 ± 0.11	344	30 ± 2.8	n.t.e	n.t.e
3r			H	H	5.50 ± 0.09	3170	88 ± 12.4	n.t.e	n.t.e
3s			H	H	<5d	<10,000	65 ± 5.5	n.t.e	n.t.e
3t			Cl	H	6.66 ± 0.08f	218f	5 ± 0.0.5f	6	73
3u			H	7-Me	5.68 ± 0.08	2070	60 ± 4.5	n.t.e	n.t.e
3v			H	8-Me	6.00 ± 0.06f	1000f	16 ± 3.8f	36	71

^aCalcium mobilization assay using HEK293 cells stably expressing rat mGlu₅ receptors; values are the average of three or more independent determinations.

^bExpressed as amplitude of response using 30 µM test compound (percentage of maximal response vs 100 µM glutamate).

^cHLM and RLM data refer to % of compound metabolized after incubation of tested compound with human and rat microsomes, respectively, for 15 min at 1 µM concentration.

^dData obtained from a single experiment not replicated.

^eNot tested.

^fAntagonists/NAMs, data represents pIC₅₀/IC₅₀ from EC₈₀ window and % Glu represents E_min not E_max.

We next evaluated the effects of the introduction of different substituted aryl and heteroaryl rings on the lactam. Based on our previous results,¹⁴ SAR investigation started with the introduction of fluorine-containing phenyl rings (3e–3g) resulting in a remarkable improvement in the mGlu₅ PAM activity. Among this small set, the para- and the 3,4-substitutions (3e and 3g) proved to be more beneficial than the meta- (3f) in terms of potency and efficacy. Additionally, the meta-substitution was also found detrimental for the metabolic stability in rat liver microsomes (RLM). In a second step, we studied the effects of the replacement of the phenyl ring by more polar and weakly basic pyridines prioritizing the 2-substituted pyridine analogues over the 3 or 4 regioisomers (3h–3n).¹⁴ Thus, while only a slight decreased PAM activity was shown with compound 3h, the corresponding 2-pyridyl analogue of 3e; the introduction of a methyl substituent in 3-, 4- and 5-position of the 2-substituted pyridine ring (analogues 3i–3k), revealed a substantial loss of potency and efficacy. The recovery of both properties was achieved with a more basic analogue (3l) although unfortunately, this was accompanied by a high metabolic turnover in human liver microsomes (HLM) and RLM. Similarly, despite the combination of fluorine and methyl in the disubstituted pyridyl congeners 3m and 3n resulted in good in vitro PAM activity and efficacy, the improvement in metabolic stability in RLM was only moderate.

Encouraged by the promising profile of 3e, we next decided to evaluate the influence of the western aromatic ring in our molecules keeping constant for that purpose the para-fluorophenyl substitution on the amide moiety. Thus, the introduction of a fluorine atom in meta- and para-positions of the phenoxy ring resulted in similar PAM potencies and efficacies (EC₅₀ = 230 nM, 91% Glu Max for 3o and EC₅₀ = 375 nM, 85% Glu Max for 3p) to the parent analogue 3e. Similarly, the introduction of a methyl group also retained the in vitro activity although a remarkable decrease in efficacy was observed (3q, 30% Glu Max). In contrast, replacement of the western phenyl by a pyridine was found to be clearly detrimental for mGlu₅ activity (compounds 3r and 3s).

Additionally, we also investigated the influence of small groups in different positions of the imidazopyrimidinone core, keeping again the para-fluorophenyl substitution on the lactam (compounds 3t–3v). This brief SAR survey led to an unexpected ‘molecular switch’ that changed the mode of the pharmacology. Thus, the introduction of a chlorine atom in the 3-position (3t) or a methyl substituent in the 8-position (3v) provided weak to moderate mGlu₅ antagonists or negative allosteric modulators (NAMs) with IC₅₀s of 218 and 1000 nM, respectively, and with additionally reduced metabolic stability in RLM. On the contrary, the 7-methyl analogue 3u still behaved as mGlu₅ PAM although with a greatly reduced potency (EC₅₀ = 2070 nM, 60% Glu Max) compared to parent congener 3e.

In parallel to our exploration of the imidazopyrimidinone series 3, we also synthesized a number of the more flexible dihydroimidazopyrimidinones 4 by hydrogenation of a set of analogues 3 presenting identified preferred modifications. SAR of this series is summarized in Table 2. Alkyl analogues in the lactam were found unstable under basic conditions²⁰ and thus only aryl and heteroaryl substitutions were explored. Interestingly, this small library proved more productive than the previous one, providing several analogues with mGlu₅ PAM activities below 250 nM. The specific comparison among pairs of analogues presenting fluorine decorations in the eastern aromatic aryl confirmed this trend; with a noticeable ~3.5 fold increase in potency for the para-fluorophenyl derivative 4a versus 3e and less prominent 1.8–2.4 increase for the meta- and the 3,4-difluoro substituted analogues (4b and 4c vs 3f and 3g, respectively). The same beneficial effect was also observed for the pyridyl-containing pairs (4d–4f vs 3h,i,m) with mGlu₅ potency increases ranging from 1.6 to 4.1 fold, however counterparts 4g and 3n showed comparable activity. On the other hand, the increased flexibility in the central core did not result in any outstanding effect on microsomal stability.

Table 2.

Structures and activities of substituted dihydroimidazopyrimidinones 4

Compd	R1	R2	R3	mGlu5 pEC50a (±SEM)	mGlu5 EC50 (nM)	% Glu maxb (±SEM)	HLMc (%)	RLMc (%)
4a			H	7.07 ± 0.02	86	70 ± 0.7	18	19
4b			H	6.66 ± 0.03	218	82 ± 9.2	20	78
4c			H	6.88 ± 0.06	130	91 ± 12.6	10	37
4d			H	6.44 ± 0.08	359	66 ± 2.1	1	14
4e			H	6.31 ± 0.02	485	74 ± 9.6	35	98
4f			H	6.74 ± 0.02	183	91 ± 13.0	31	93
4g			H	6.43 ± 0.08	371	87 ± 16.1	n.t.d	n.t.d
4h			H	7.16 ± 0.01	69	72 ± 0.7	0	29
4i			H	6.70 ± 0.04	197	88 ± 12.2	3	11
4j			H	6.55 ± 0.07	285	47 ± 3.3	n.t.d	n.t.d
4k			Cl	6.88 ± 0.05e	133e	14 ± 4.8e	n.t.d	n.t.d
4l			Me	6.28 ± 0.09e	529e	6 ± 1.4e	21	46

^aCalcium mobilization assay using HEK293 cells stably expressing rat mGlu₅ receptors; values are the average of three or more independent determinations.

^bExpressed as amplitude of response using 30 µM test compound (percentage of maximal response vs 100 µM glutamate).

^cHLM and RLM data refer to % of compound metabolized after incubation of tested compound with human and rat microsomes respectively, for 15 min at 1 µM concentration.

^dNot tested.

^eAntagonists/NAMs, data represents pIC₅₀/IC₅₀ from EC₈₀ window and % Glu represents E_min not E_max.

Analogously, the introduction of a fluorine atom in the meta- (4h) and para-position (4i) of the western aryl, greatly increased mGlu₅ potency up to ~3.3 fold versus series 3 corresponding counterparts, maintaining also a high efficacy. Gratifyingly, this exploration yielded analogue 4h, the most potent mGlu₅ PAM across the two series described herein (EC₅₀ = 69 nM, 72% Glu Max). However, the introduction of a methyl group in the meta-position (4j) contributed to a significant loss in efficacy in comparison with the other dihydroimidazopyrimidinones 4, confirming the tendency previously observed for series 3 (47% Glu Max for 4j vs 30% Glu Max for 3q).

Similarly to chemotype 3, the introduction of substituents in the 3-position of the dihydroimidazopyrimidinone core was responsible again for a change in the mode of pharmacology and resulted in compounds that behaved as antagonists/NAMs. Thus, the introduction of a chlorine atom in 4k or a methyl group in 4l, afforded mGlu₅ antagonists/NAMs of moderate potency (133 and 529 nM, respectively). These data demonstrated again that with very subtle structural changes, a reasonably potent PAM can be transformed into an antagonist/NAM of comparable potency, confirming the findings previously reported for other chemotypes.^15,21

After this initial exploration, analogues 3e and 4a were selected as representative examples from both chemotypes for additional characterization (Table 3). Thus, besides increased flexibility, 4a shows also a slightly reduced cLogP (0.5 log units). Both analogues were found adequately stable in HLM and RLM with no significant differences in the extent of metabolism. Additionally, in a cocktail assay using HLM and known substrates, 4a did not show any relevant inhibition of the major human cytochrome P450 (CYP) enzymes (2C9, 2D6, 3A4, 1A2) (IC₅₀s >20 µM), while 3e was found to display moderate CYP inhibitory activity on 1A2 (IC₅₀ = 9.3 µM, others IC₅₀ >20 µM). Finally, 4a showed a slightly higher unbound fraction in both human and rat plasma compared with 3e.

Table 3.

Comparative profile of compounds 3e and 4a

Compd	mGlu5 EC50a (nM)	% Glu maxb (±SEM)	c log Pc	mGlu3 EC50 (nM)	HLMd (%)	RLMd (%)	hPPBe (%)	rPPBe (%)
3e	310	83 ± 9.1	3.95	630	14	27	96.4	93.1
4a	86	70 ± 0.7	3.42	290	18	19	90.3	88.1

^aFor the mGlu₅ PAM assay see Tables 1 and 2.

^bExpressed as amplitude of response using 30 µM test compound (percentage of maximal response vs 100 µM glutamate).

^cCalculated with Biobyte software.

^dHLM and RLM data refer to % of compound metabolized after incubation of tested compound with human and rat microsomes respectively, for 15 min at 1 µM concentration.

^ePlasma protein binding, human and rat.

Subsequent selectivity profiling against the other mGlu receptors (mGlu_1–4,6–8) revealed that both compounds were also active as full mGlu₃ NAMs, (IC₅₀ = 630 nM for 3e and 290 nM for 4a), resulting in moderate functional selectivities for mGlu₅ of ~2 fold for 3e and ~3 fold for 4a respectively (other mGlu EC₅₀ >10 µM). An analogous dual profile had already been observed for similarly substituted N-aryl analogues within our naphthyridinone series.¹⁵ Although the biological relevance of the mGlu₃ receptor is still not well understood, inhibition of mGlu₃ has been hypothesized to have potential therapeutic utility in the treatment of neurological and psychiatric disorders.²² Based on its mGlu₅ PAM potency and efficacy, higher selectivity versus mGlu₃, reduced cLogP, superior CYP profile and higher unbound fraction in plasma, 4a was selected for further in vivo pharmacokinetic (PK) and pharmacological evaluation.

Compound 4a was formulated at 1 mg/mL in 20% hydroxypropyl-β-cyclodextrin (HP-β-CD) and evaluated in in vivo PK studies in rats (1 mg/kg IV, 10 mg/kg PO).²³ Compound 4a showed a moderate clearance (CL_p = 23.3 mL/min/kg) and a terminal half-life (t_1/2) of 0.93 h. Evaluation of drug concentrations in portal vein, plasma and brain (0.5, 1.0, 2.0, 4.0 and 7.0 h) revealed a moderate first-pass effect (E_hep = 0.43) with relatively low volume of distribution (V_ss = 1.4 L/kg) and a good systemic exposure (plasma_AUC = 7.2 µM-h). In addition, 4a showed also a high brain exposure (brain_AUC = 12.7 µM-h) with a very good distribution to the brain (brain_AUC/plasma_AUC = 1.8), which together with a moderate unbound fraction in brain (rat f_u brain = 6.2%) contributed to reach high absolute brain levels (C_max = 6.05 µM). Encouraged by its overall profile, 4a was evaluated for its ability to revert AIH in rats, an established model of antipsychotic activity.^10a,16 As seen in Figure 2, 4a showed a robust dose-dependent reversal of amphetamine effects with a lowest active oral dose of 3 mg/kg, and a maximal effect of ~50% at the highest doses tested of 56.6 and 100 mg/kg. At these doses the average terminal unbound brain concentrations (382 and 363 nM, respectively) are well above the in vitro mGlu₅ PAM EC₅₀.²⁴

Figure 2.

Dose-dependent effect and calculated terminal unbound brain of 4a on the reversal of amphetamine-induced hyperlocomotion in rats.

In parallel to these explorations, potential mGlu₅ related CNS adverse-effect liabilities driven by excessive glutamate fold potentiation²⁵ or allosteric agonism^26,27 were reported, suggesting that PAMs with lower functional cooperativity with glutamate and devoid of allosteric agonism may be preferred for an adequate therapeutic index.²⁸ Concurrent with these findings and despite the fact that no significant adverse effects had previously been observed in PK and pharmacological experiments in rats up to 100 mg/kg p.o., investigation of 4a in the modified Irwin neurological test battery in rats at a high dose (120 mg/kg) and in in vivo PK studies in dogs (0.5 mg/mL i.v.) revealed clear signs of CNS-mediated side effects (pro-convulsive behavior and dizziness, respectively) which excluded any further in vivo evaluation. Although further in vitro characterization of 4a confirmed the lack of mGlu₅ agonistic activity up to the highest concentration tested of 30 µM, fold shift experiments highlighted a high degree of cooperativity with glutamate (12.5-fold leftward shift in the glutamate concentration response curve at 10 µM) that could eventually explain the observed adverse effects, although a potential contribution of mGlu₅/mGlu₃ synergistic effects cannot be completely ruled out.

In summary, further optimization of our initial dihydrothiazolopyridone lead 1 resulted in two novel series of imidazopyrimidinones and dihydroimidazopyrimidinones as mGlu₅ PAMs. Parallel SAR investigation of both series and further DMPK profiling of representative prototypes revealed dihydroimidazopyrimidinone 4a as a highly potent and efficacious mGlu₅ PAM (EC₅₀ = 86 nM, 70% Glu Max) and moderately potent mGlu₃ NAM (IC₅₀ = 290 nM, 3% Glu Max). 4a possessed an adequate PK profile in rats and showed robust dose-dependent effects in a preclinical model predictive of antipsychotic efficacy. However, the presence of CNS-mediated adverse effects in preclinical species precluded any further in vivo evaluation. Keeping in mind that more studies are necessary to address ongoing questions concerning the therapeutic index of mGlu₅, further efforts within these and related series are in progress and will be reported in due course.