N-Alkylpyrido[1′,2′:1,5]pyrazolo-[4,3-d]pyrimidin-4-amines: A new series of negative allosteric modulators of mGlu_1/5 with CNS exposure in rodents

Abstract

Selective negative allosteric modulators (NAMs) of each of the group I metabotropic glutamate receptors (mGlu₁ and mGlu₅) have been well characterized in the literature and offer potential as therapeutics in several disorders of the central nervous system (CNS). Still, compounds that are potent mGlu_1/5 NAMs with selectivity versus the other six members of the mGlu family as well as the balance of properties required for use in vivo are lacking. A medicinal chemistry effort centered on the identification of a lead series with the potential of delivering such compounds is described in this Letter. Specifically, a new class of pyrido[1′,2′:1,5]pyrazolo[4,3-d]pyrimidin-4-amines was designed as a novel isosteric replacement for 4-aminoquinazolines, and compounds from within this chemotype exhibited dual NAM activity at both group I mGlus. One compound, VU0467558 (29), demonstrated near equipotent activity at both receptors, selectivity versus other mGlus, a favorable ancillary pharmacology profile, and CNS exposure in rodents.

Graphical Abstract

Glutamate (L-glutamic acid), the major excitatory transmitter in the mammalian central nervous system (CNS), produces its effects through binding to both ionotropic and metabotropic glutamate receptors (mGlus). The mGlus comprise a family of eight G-protein-coupled receptors (GPCRs) that are further divided according to their structure, preferred signal transduction mechanisms, and pharmacology. The group I mGlus (mGlu₁ and mGlu₅) are located post-synaptically and are coupled via G_q to the activation of phospholipase C, an event that leads to the elevation of intracellular calcium (Ca²⁺) and activation of protein kinase C (PKC). Conversely, both the group II (mGlu₂ and mGlu₃) and group III (mGlu₄, mGlu₆, mGlu₇, and mGlu₈) mGlus are found predominantly pre-synaptically and are coupled via G_i/o to the inhibition of adenylyl cylase activity. Orthosteric binding sites are located in the extracellular N-terminal domain within the mGlu family. In contrast, the majority of allosteric binding sites that have been discovered to date are contained in the transmembrane domain.^1-3

Given the high homology across orthosteric binding sites within the mGlu family, the design of highly selective orthosteric ligands has been quite challenging. One approach to circumvent this challenge that has proven successful in many instances has been the development of selective allosteric modulators of the individual mGlus.^4-8 Among the most advanced areas within this field are small molecule negative allosteric modulators (NAMs) of mGlu₁^9,10 and mGlu₅.^11-15 In fact, at least one mGlu₁ NAM, a 1,4-diaryl-5-methyl-1,2,3-triazole developed by researchers at Merck-Banyu, reached clinical candidate status.^16,17 The mGlu₅ NAM field is still more advanced, with multiple compounds advancing to the clinic, including phase II studies with basimglurant (Roche),^18,19 dipraglurant (Addex),²⁰ and mavoglurant (Novartis).^21,22

Highly selective mGlu₁ and mGlu₅ NAMs are both clearly interesting, and many high quality small molecule probes from multiple chemotypes exist for each. Still, we were surprised to find no reports of compounds suitable for rodent studies that are potent NAMs of both group I receptors while maintaining selectivity against the other mGlus. Numerous preclinical studies with selective mGlu₁ and mGlu₅ NAM tools demonstrate overlap in potential therapeutic applications. For example, accounts of efficacy in preclinical models of anxiety,^23-28 addiction,^29-45 and pain^46-50 point toward a potential role for antagonism of each group I receptor in these disorders. Still, the literature has documented concerns regarding the potential for motor and cognitive side effects with mGlu₁ NAMs,⁹ and concerns regarding the psychotomimetic effects of certain mGlu₅ NAMs have been documented as well.^11,51 Thus, it is reasonable to question whether there will be a negative impact in combining such activities; however, one should not assume such will necessarily be the case. It may even be possible that a dual mGlu_1/5 NAM could offer an improved safety profile by reducing the occupancy at each receptor required for efficacy relative to agents selective for only one of these receptors. Such studies suggest the possibility of a potential benefit with dual mGlu_1/5 inhibition and warrant efforts to identify a lead series capable of delivering that profile.

During the course of our own efforts to identify and optimize selective mGlu₅ NAMs, we identified a series of 6-substituted-4-anilinoquinazolines represented by screening hit 1 (Fig. 1).⁵² Interestingly, compound 1 and additional analogs within this series were essentially equipotent against mGlu₁ and mGlu₅ while demonstrating selectivity versus other members of the mGlu family. The mGlu₁ NAM co-activity in 1 was not entirely unexpected as Lilly had previously identified and characterized a selective mGlu₁ NAM tool, LY456236 (2) from the same chemotype.⁵³ Furthermore, more highly optimized mGlu₁ NAMs from quinazoline-like scaffolds had also been previously reported, including compound 3 from Pfizer⁴⁸ and LY456066 (4) from Lilly.⁵⁴

Figure 1.

Quinazoline and quinazoline-like group I mGlu NAMs

Based on these observations, we rationalized that it might be possible to design a selective mGlu_1/5 NAM with properties suitable for use in rodent behavioral assays. Quinazolines are among the most studied scaffolds in medicinal chemistry and are extensively described in the primary and patent literature;⁵⁵ thus, we sought to develop a novel isostere for this moiety. Literature searching revealed that one chemotype that had yet to be prepared and studied to any degree biologically was a series of pyrido[1′,2′:1,5]pyrazolo[4,3-d]pyrimidin-4-amines (Fig. 2). Given such novelty and a high probability that such compounds would function as quinazoline isosteres, we were quite interested in pursuing these targets. Our own aforementioned work related to mGlu₅ NAMs⁵² led us to believe that secondary amines would be required for activity at that target. Evaluation of the Pfizer series of mGlu₁ NAMs⁴⁸ indicated that saturated lipophilic groups would be favorable for mGlu₁ activity. We were keen to incorporate saturated groups at this position as increased sp³ character has been linked to improved properties and drug-likeness.⁵⁶ Finally, given the substitution pattern in LY456066 (4, Fig. 1), we identified the 2-position worthy of SAR development as well.

Figure 2.

Proposal for development of novel mGlu_1/5 NAMs from a pyrido[1′,2′:1,5]pyrazolo[4,3-d]pyrimidin-4-amine scaffold

It was envisioned that analogs could be accessed through known key intermediate amine 7 (Scheme 1). The synthesis of 7 from commercially available methyl ester 5 was previously described in the literature.⁵⁷ We employed a modified version of the published synthesis of 7. Specifically, nitration of 5 with potassium nitrate in sulfuric acid afforded 3-nitro intermediate 6 in good yield, and a tin(II) chloride reduction of the nitro group gave amine 7 in moderate yield. For synthesis of intermediate 8 (R¹ = H), 7 was heated via microwave irradiation in formamide. Synthesis of intermediate 9 (R¹ = CH₃) was accomplished through microwave irradiation in triethyl orthoacetate under mildly acidic conditions; however, this transformation was much more sluggish than was observed for the preparation of 8. Conversion of 8 and 9 to heteroaryl chlorides 10 and 11, respectively, was carried out in moderate yield by microwave irradiation in phosphorous oxychloride. Installation of 4-position amines was accomplished through a nucleophilic aromatic substitution reaction with the requisite primary amine (R²NH₂).

Scheme 1.

Reagents and conditions: (a) Reagents and conditions: (a) KNO₃, H₂SO₄, 0 °C to r.t., 79%; (b) SnCl₂, con. HCl, dioxane, 65%; (c) For R¹ = H, HCONH₂, microwave, 200 °C, 60 min., 98%; (d) For R¹ = CH₃, CH₃C(OEt)₃, formic acid, microwave, 180 °C, 60 min., 24%; (e) POCl₃, microwave, 140 °C, 20 min., 59% (R¹ = H), 54% (R¹ = CH₃); (f) R²NH₂, DIEA, DMF, microwave, 150 °C, 15 min., 43-84%.

Synthesis of analogs with amines at the 2-position of the scaffold employed an alternative route that also began with key intermediate 7 (Scheme 2). Heating of 7 with ethyl chloroformate in dioxane, followed by treatment with potassium tert-butoxide under microwave irradiation, afforded intermediate 38 in near quantitative yield. Conversion of 38 to dichloro intermediate 39 was accomplished in high yield through microwave heating in phosphorous oxychloride. Room temperature nucleophilic aromatic substitution of trans-methylcyclohexylamine proceeded selectively at the 4-position of the template to provide penultimate intermediate 40. The selectivity for the 4-position mirrors results obtained in similar transformations involving 2,4-dichloroquinazolines.^58,59 Final stage diversification was carried out as described above through heating 40 with the desired amine (R³NH₂) via microwave irradiation.

Scheme 2.

Reagents and conditions: (a) (i) ClCO₂Et, dioxane 100 °C, 16 h, (ii) KO^tBu, isopropanol, microwave, 160 °C, 15 min., 99%; (b) POCl₃, microwave, 140 °C, 20 min., 84%; (c) trans-4-methylcyclohexylamine, DIEA, DMF, rt, 16h., 76%; (d) R³NH₂, DIEA, DMF, microwave, 150 °C, 10 min., 48-77%.

Evaluation of all new compounds for their activity at mGlu₁ and mGlu₅ was conducted using functional cell-based assays. These fluorescence-based assays employ cell lines stably expressing either rat mGlu₁ or rat mGlu₅ and measure the ability of the compound to block the mobilization of calcium induced by an EC₈₀ concentration of glutamate.^60,61 The format for these assays employs a “triple-add” protocol that allows for the detection of agonists, positive allosteric modulators (PAMs), and NAMs.⁶² A number of 4-position amine analogs (12–19) were prepared that were considered somewhat flexible, meaning that the sp³ carbon to which the amine was attached was not part of a cyclic ring system (Table 1). Several analogs were prepared and tested as pairs of enantiomers with a chiral methyl group (12–17). In each case, the (R)-enantiomer analogs (13, 15, and 17) were preferred relative to their (S)-enantiomer counterparts (12, 14, and 16) with regards to mGlu₁ NAM activity. The mGlu₁ NAM activity was also improved by moving from increasingly lipophilic cyclopropyl (12, 13) to tert-butyl (14, 15) to cyclohexyl (16, 17) analogs. The 3,3-dimethylbutan-1-amine analog 18 proved the most potent mGlu₁ NAM in this set. On the other hand, 2-methylbutan-2-amine analog 19 was only a weak mGlu₁ NAM, inhibiting the glutamate response only at the top concentration tested (30 μM). Finally, mGlu₅ NAM activity with these compounds was generally weak or inactive (> 30 μM); however, 18 demonstrated modest mGlu₅ NAM potency that was approximately ten-fold less than its mGlu₁ NAM activity. Such a result confirmed the possibility of accessing dual mGlu_1/5 NAMs from within this scaffold.

Table 1. 4-Position SAR for flexible alkyl amines.

Cpd	R	A	cLogPa	mGlu1 pIC50 (± SEM)b	mGlu1 IC50 (nM)b	% Glu Max (± SEM)b,c	mGlu5 pIC50 (± SEM)b	mGlu5 IC50 (nM)b	% Glu Max (± SEM)b,c
12	I	cyclopropyl	2.26	< 4.5	> 30,000	—	< 4.5	> 30,000	—
13	II	cyclopropyl	2.26	< 5.0d,e	>10,000d,e	71.6e	< 4.5	> 30,000	—
14	I	tert-butyl	3.30	5.18 ± 0.05	6,540	−0.2 ± 3.0	< 4.5	> 30,000	—
15	II	tert-butyl	3.30	6.05 ± 0.07	883	2.2 ± 0.1	< 4.5	> 30,000	—
16	I	cyclohexyl	3.52	5.65 ± 0.03	2,240	0.9 ± 0.2	< 5.0d	> 10,000d	61.8 ± 4.3
17	II	cyclohexyl	3.52	6.33 ± 0.10	462	2.4 ± 0.3	< 5.0d	> 10,000d	18.4 ± 6.9
18	III	—	3.26	6.47 ± 0.07	341	1.5 ± 0.3	5.44e	3,600e	1.3e
19	IV	—	2.58	< 5.0d	> 10,000d	12.6 ± 4.3	< 4.5	> 30,000	—

^aCalculated using ChemDraw Professional, version 15.0 (PerkinElmer Informatics, Inc.)

^bCalcium mobilization assay; values are average of n ≥ 3

^cAmplitude of response in the presence of 30 μM test compound as a percentage of maximal response (100 μM glutamate); average of n ≥ 3

^dConcentration-response curve (CRC) does not plateau

^eAverage of n = 2

Results obtained with cycloalkyl amine analogs 20–26 provided further encouragement (Table 2). Potency at both mGlu₁ and mGlu₅ was generally enhanced by increasing size and lipophilicity; however, a preference for mGlu₁ versus mGlu₅ was noted with all analogs. For example, a clear potency enhancement at both receptors was observed in moving from cyclohexyl analog 20 to cycloheptyl analog 21 to cycloctyl analog 22. Additionally, while potency was good with 2,3-dihydro-1H-inden-2-amine 23, adamantyl derivatives 24 and 25 were among the most potent compounds in this set. Notably, installation of a tertiary hydroxyl group on the adamantine ring (26), a modification that reduced lipophilicity substantially relative to direct comparator 25, also reduced potency at both receptors. Finally, while mGlu₁ NAM potency was excellent in the case of a few analogs in this set, mGlu₅ NAM potency remained modest in those same analogs.

Table 2. 4-Position SAR for cycloalkyl amines.

Cpd	R	n	cLogPa	mGlu1 pIC50 (± SEM)b	mGlu1 IC50 (nM)b	% Glu Max (± SEM)b,c	mGlu5 pIC50 (± SEM)b	mGlu5 IC50 (nM)b	% Glu Max (± SEM)b,c
20	I	1	2.77	5.57 ± 0.13	2,720	−0.5 ± 0.8	< 4.5	> 30,000	—
21	I	2	3.19	6.65 ± 0.07	225	2.0 ± 0.3	< 5.0d	> 10,000d	20.0 ± 14.1
22	I	3	3.60	7.03 ± 0.09	94	1.9 ± 0.1	5.48e	3,290e	1.3e
23	II	—	3.19	6.61 ± 0.05	247	2.1 ± 0.2	5.57 ± 0.13	2,700	17.2 ± 10.6
24	III	—	3.23	7.08 ± 0.05	84	2.9 ± 0.6	5.76 ± 0.41	1,750	3.4 ± 1.3
25	IV	—	3.05	6.83 ± 0.04	148	2.6 ± 0.5	5.63 ± 0.25	2,360	4.8 ± 5.3
26	V	—	1.74	6.32 ± 0.04	484	2.4 ± 0.3	< 4.5	> 30,000	—

^aCalculated using ChemDraw Professional, version 15.0 (PerkinElmer Informatics, Inc.)

^bCalcium mobilization assay; values are average of n ≥ 3

^cAmplitude of response in the presence of 30 μM test compound as a percentage of maximal response (100 μM glutamate); average of n ≥ 3

^dCRC does not plateau

^eAverage of n = 2

Having demonstrated that increased lipophilicity and/or steric bulk on the 4-position increased potency, we decided to prepare substituted analogs of cyclohexyl amine 20. New analogs (Table 3) contained functional groups designed to further probe this trend while continuing to search for increased mGlu₅ NAM activity. Difluorocyclohexyl amine 27 reduced mGlu₁ NAM activity relative to 20 (Table 2) and failed to enhance mGlu₅ NAM activity. Fortunately, other analogs were more promising. For instance, both trans-4-methylcyclohexyl amine 28 and 4,4-dimethylcyclohexyl amine 29 demonstrated excellent mGlu₁ NAM potency, and 29 proved a breakthrough compound with excellent activity at mGlu₅ as well. Additional compounds prepared from various commercially available chiral amines (30–33) offered good potency as mGlu₁ NAMs (30–32); however, mGlu₅ potency failed to approach that observed with 29. Such results are yet another example of the narrow SAR that is often discovered in the design of allosteric modulators of GPCRs and a reminder of the importance of preparing and testing multiple analogs even when there are only subtle structural differences across compounds.

Table 3. 4-Position SAR for substituted cyclohexyl amines.

Cpd	R	cLogPa	mGlu1 pIC50 (± SEM)b	mGlu1 IC50 (nM)b	% Glu Max (± SEM)b,c	mGlu5 pIC50 (± SEM)b	mGlu5 IC50 (nM)b	% Glu Max (± SEM)b,c
27	I	2.33	< 5.0d	> 10,000d	39.0 ± 7.7	< 4.5	> 30,000	—
28	II	3.10	7.03 ± 0.08	93	2.0 ± 0.2	5.64 ± 0.22	2,270	2.0 ± 0.5
29	III	3.57	7.16 ± 0.08	69	2.9 ± 0.2	6.89 ± 0.29	129	1.6 ± 0.5
30	IV	2.58	6.63 ± 0.05	232	2.5 ± 0.6	< 5.0d	> 10,000d	9.9 ± 9.2
31	V	3.71	6.82 ± 0.03	152	2.7 ± 0.2	< 5.0d	> 10,000d	9.0 ± 8.3
32	VI	3.71	6.63 ± 0.12	233	3.1 ± 0.4	5.68e	2,090e	2.3e
33	VII	3.92	5.79 ± 0.04	1,610	7.1 ± 2.9	< 4.5	> 30,000	—

^aCalculated using ChemDraw Professional, version 15.0 (PerkinElmer Informatics, Inc.)

^bCalcium mobilization assay; values are average of n ≥ 3

^cAmplitude of response in the presence of 30 μM test compound as a percentage of maximal response (100 μM glutamate); average of n ≥ 3

^dCRC does not plateau

^eAverage of n = 2

Prior to the discovery of 29 and in parallel to its preparation, we evaluated substitution of the 2-position of the scaffold with methyl (34–37) and amine (41–45) groups (Table 4). The patent application that covers LY456066 (4) (Fig. 1) indicates that alkyl and N-alkyl groups can function as alternatives to the S-alkyl groups of that compound,⁶³ and we viewed these as more drug-like alternatives. For these analogs, we employed four of the 4-position amines that offered good to excellent mGlu₁ NAM activity and modest mGlu₅ NAM potency. Unfortunately, modest reduction in mGlu₁ NAM potency was observed with a 5-methyl substituent relative to hydrogen, and no mGlu₅ activity was noted with these compounds. The 5-amino analogs were even more disappointing, with only primary amine 41 offering weak mGlu₁ NAM activity and other compounds inactive at both receptors. Given the intolerance of substitution at this position, we decided to move forward with further profiling of 29 as it was the only compound that was essentially equipotent at both receptors. It should be noted that it has yet to be established that a compound that is equipotent at both group I mGlus will ultimately prove optimal. If a potent mGlu₁ NAM with moderate mGlu₅ NAM co-activity was of interest, this chemotype appears highly tractable for that profile. Still, compound 29 (VU0467558) represented a novel and potent dual mGlu_1/5 NAM with reduced lipophilicity (3.57 vs. 5.09) and enhanced sp³ character relative to hit 1 (Fig.1). Thus, 29 was a logical choice for further profiling from within this new lead series.⁶⁴

Table 4. 2-Position SAR.

Cpd	Series	R	cLogPa	mGlu1 pIC50 (± SEM)b	mGlu1 IC50 (nM)b	% Glu Max (± SEM)b,c	mGlu5 pIC50 (± SEM)b	mGlu5 IC50 (nM)b	% Glu Max (± SEM)b,c
34	I	CH3	4.01	5.75 ± 0.09	1,800	2.6 ± 0.7	< 4.5	> 30,000	—
35	II	CH3	4.62	6.12 ± 0.04	759	2.4 ± 0.7	< 4.5	> 30,000	—
36	III	CH3	4.83	5.79 ± 0.04	1,610	7.1 ± 2.9	< 4.5	> 30,000	—
37	IV	CH3	4.14	5.67 ± 0.07	2,120	0.6 ± 0.8	< 4.5	> 30,000	—
41	I	NH2	2.81	5.25e	5,600e	0.58e	< 4.5	> 30,000	—
42	I	NHCH3	3.11	< 4.5	> 30,000	—	< 4.5	> 30,000	—
43	I	N(CH3)2	3.90	< 4.5	> 30,000	—	< 4.5	> 30,000	—
44	I	NHCH(CH3)2	3.76	< 4.5	> 30,000	—	< 4.5	> 30,000	—
45	I	NHCH2CH2OCH3	2.95	< 4.5	> 30,000	—	< 4.5	> 30,000	—

^aCalculated using ChemDraw Professional, version 15.0 (PerkinElmer Informatics, Inc.)

^bCalcium mobilization assay; values are average of n ≥ 3

^cAmplitude of response in the presence of 30 μM test compound as a percentage of maximal response (100 μM glutamate); average of n ≥ 3

^dCRC does not plateau

^eAverage of n = 2

First, selectivity was assessed versus the remaining six members of the mGlu family. The effect of 10 μM 29 on the orthosteric agonist concentration-response curve (CRC) was assessed in fold-shift experiments,^65,66 and no significant effects were observed. Hence, 29 exhibited excellent selectivity (>75-fold) for the group I mGlu receptors. To evaluate the ancillary pharmacology profile (Table 5), 29 was screened in a commercially available radioligand binding assay panel of 68 clinically relevant targets.⁶⁷ For 62 of the 68 targets no significant responses were noted; significant responses were defined as greater than or equal to 50% inhibition of binding. Four of the six remaining targets showed only modest inhibition of radioligand binding while substantial inhibition was found at the closely related receptors, adenosine A₁ and A_2A. Additionally, since 4-aminoquinazolines are well-known scaffolds for kinase inhibitors,⁵⁵ we profiled 29 in an extensive kinase panel of 369 kinases.⁶⁸ In this large panel, 10 μM 29 only inhibited more than 50% of the activity of eight kinases, and half of these were members of a single family, the casein kinase 1 family. Overall, the selectivity of 29 was considered encouraging, and evaluation of the compound in assays designed to assess its potential for use in vivo was deemed worthwhile.

Table 5.

Ancillary Pharmacology Profile for 10 μM 29 (VU0467558)

Radioligand Binding Panel Significant Resultsa
Target	Inhibition of Binding
adenosine A1	92%
adenosine A2A	94%
adenosine A3	63%
serotonin 5-HT2B	65%
norepinephrine transporter (NET)	65%
sodium channel, site 2	55%
Kinase Inhibition Profile Significant Resultsb
Kinase	Inhibition of Kinase Activity
casein kinase 1α1 (CK1α1)	65.3%
casein kinase 1δ (CK1δ)	85.1%
casein kinase 1ε (CK1ε)	75.9%
casein kinase 1γ2 (CK1γ2)	50.4%
DAP kinase-related apoptosis-inducing protein kinase 1 (DRAK1)	52.9%
Serine-arginine protein kinase 1 (SRPK1)	79.4%
Serine-arginine protein kinase 2 (SRPK2)	75.2%
TRAF2 And NCK Interacting Kinase (TNIK)	58.8%

^agreater than or equal to 50% inhibition of binding at 10 μM 29

^bgreater than or equal to 50% inhibition of kinase activity at 10 μM 29

Since unbound fraction (f_u) can be important to consider alongside potency when correlating in vivo efficacy to exposure, equilibrium dialysis protein binding assays were conducted in both plasma and brain homogenates (Table 6).^69,70 Compound 29 was highly protein bound, and the fraction unbound was higher in plasma than brain for both rat and mouse. Metabolic stability was gauged by measuring the intrinsic clearance of 29 in liver microsomes and using that value to calculate the predicted hepatic clearance (CL_HEP).^70,71 Unfortunately, the predicted hepatic clearance for 29 was near hepatic blood flow for both species. Finally, the potential for drug–drug interactions with 29 was assessed in a human liver microsomes (HLM) cocktail assay with probe substrates for four common P450s. Compound 29 was judged quite clean according to this assay, with little P450 inhibition observed. With an in vitro DMPK profile in hand, we moved forward with in vivo studies in rodents.⁷³

Table 6.

In vitro DMPK Profile for 29 (VU0467558)

Protein Binding (fu)a			Predicted Hepatic Clearance (CLHEP)b
	plasma	brain
rat	0.037	0.008	rat	63.6 mL/min/kg
mouse	0.025	0.010	mouse	84.5 mL/min/kg

P450 Inhibition IC50 (μM)c
3A4	2D6	2C9	1A2
>30	>30	>30	9.6

^af_u = fraction unbound; equilibrium dialysis assay

^bcalculated from measured intrinsic clearance in liver microsomes

^cAssayed in pooled HLM in the presence of NADPH

The clearance in rats and the CNS penetration for 29 was assessed simultaneously using intravenous (IV) dosing (Table 7).^70,74 The in vitro microsomal stability assay proved predictive as the compound exhibited high-clearance and a short half-life. Such a result may point toward microsomal stability as a useful assay for further optimization within this series. Fortunately, the CNS penetration for 29 was excellent with a brain to plasma ratio (K_p) of 5.3 in rats. Moreover, the unbound brain to unbound plasma ratio was approximately one, indicating distribution equilibrium between the compartments and a low probability of the compounds being a substrate for transporters.⁷⁵ Given the high clearance observed in rats, intraperitoneal (IP) dosing was chosen as a route of administration that might provide therapeutically relevant plasma drug levels. Thus, IP PK studies in rats and mice at 10 and 30 mg/kg, respectively, were carried out.⁷⁰ Tissue distribution studies in mice were conducted in parallel using the same conditions. CNS penetration in mice was quite similar to that observed in rats, with no evidence of efflux. Additionally, exposure using this route of administration and these doses was quite good in both species. For example, using the aforementioned data, it is estimated that 29 achieved an unbound brain concentration of approximately 100 nM in both rats and mice, which is near the functional IC₅₀ value at both mGlu₁ and mGlu₅.

Table 7.

Rodent PK Results for 29 (VU0467558)

Rat IV PKa,b		Tissue Distribution Studies
half-life	34 min		ratb	mousec,d
CLp	98 mL/min.kg	time point	15 min	30 min
Vss	3.1 mL/min/kg	Kpe	5.3	3.3
		Kp,uuf	1.1	1.3

Rodent IP PK Time Course Studies
	rata,g	mousec,d
dose	10 mg/kg	30 mg/kg
plasma Tmax	15 min	20 min
plasma Cmax	2.30 μM	3.00 μM
plasma AUC	3.23 μM·h	2.86 μM·h

^amale Sprague-Dawley rats (n = 2 per time point)

^b0.2 mg/kg; vehicle = 10% ethanol, 40% PEG 400, 50% DMSO

^cmale CD-1 mice (n = 3 per time point)

^d30 mg/kg; vehicle = 10% Tween 80 in water

^eK_p = total brain to total plasma ratio

^fK_p,uu = unbound brain (brain f_u · total brain) to unbound plasma (plasma f_u · total plasma) ratio

^g10 mg/kg; vehicle = 10% Tween 80 in water

In conclusion, we have designed a new lead series of pyrido[1′,2′:1,5]pyrazolo[4,3-d]pyrimidin-4-amines for the development of mGlu_1/5 NAMs with selectivity versus other members of the mGlu family. One member of this series, 29 (VU0467558), is a potent antagonist of both group I receptors with essentially equal activity and can be dosed IP in rats and mice to reach therapeutically relevant levels of unbound compound in the CNS. As such, 29 represents an attractive lead for further optimization. Future work to explore the SAR around modification of the tricyclic core, the identification and elimination of metabolic hotspots, and refinement of the ancillary pharmacology in this novel series is planned.