Synthesis and SAR of a novel metabotropic glutamate receptor 4 (mGlu₄) antagonist: Unexpected ‘molecular switch’ from a closely related mGlu₄ positive allosteric modulator

Abstract

Herein we report the discovery and SAR of a novel antagonist of metabotropic glutamate receptor 4 (mGlu₄). The antagonist was discovered via a molecular switch from a closely related mGlu₄ positive allosteric modulator (PAM). This antagonist (VU0448383) displays an IC₅₀ value of 8.2 ± 0.4 μM and inhibits an EC₈₀ glutamate response by 63.1 ± 6.6%.

G-protein coupled receptor (GPCR) orthosteric ligands currently represent the largest class of drug molecules. However, over the past several years, modulation of these targets via a novel mechanism, allosteric modulation, has generated much attention and research focus.^1–3 One phenomenon that has been reported recently is the notion of ‘molecular switching’, referring to allosteric ligands that switch modes of pharmacology upon subtle modifications of the chemical scaffold.⁴ While there have been a number of reports of molecular switches within the Groups I^5,6 and II⁷ metabotropic glutamate receptors, to the best of our knowledge, this is the first report of a molecular switch within allosteric ligands of the Group III subfamily.

As part of the Molecular Libraries Production Center Network (MLPCN),⁸ we had previously performed a high throughput screen of the NIH Molecular Libraries Small Molecule Repository (MLSMR) collection of compounds to identify muscarinic acetylcholine receptor 4 (M₄)-specific antagonists using a cell line expressing M₄ with a chimeric G protein, G_qi5, to induce calcium mobilization. During primary and confirmation screening, VU0219493 was flagged as an M₄ antagonist when screened at a single dose of 10 μM. The compound was then evaluated for concentration-dependent activity at M₄, a stage which involved counter screening using four cell lines, one of which was a cell line co-expressing metabotropic glutamate receptor 4 (mGlu₄) and Gqi₅.⁹ The purpose of this counter screen was to control for nonspecific GPCR or Gqi₅-mediated effects of HTS hits. While VU0219493 failed to confirm as active in these follow-up screens for muscarinic receptors (IC₅₀ >30 μM for both M₄ and M₁), surprisingly, the compound robustly potentiated the mGlu₄/Gq_i5 calcium response with a potency of 1.4 ± 0.4 μM (Fig. 1). The activity of this hit was also verified in cells co-expressing rat mGlu₄ with GIRK potassium channels¹⁰ (EC₅₀ = 1.1 μM) and a 30 μM concentration of the compound shifted a concentration-response curve for glutamate ninefold to the left (data not shown).

Figure 1.

VU0219493 (inset structure) potentiates glutamate-induced calcium stimulation in human mGlu4/Gqi₅ cells in a concentration-dependent fashion. Data represent the mean ± SEM of three independent experiments performed in triplicate.

With a verified positive allosteric modulator (PAM) hit in hand,¹¹ we began a chemical optimization program around this scaffold. SAR exploration was begun by reacting commercially available α-bromoketones with 2-bromophenol to yield the desired compounds. Methylation of the phenolic groups or replacing the phenols with fluorine led to inactive compounds (2b and 2c). The 2,4-dihydrophenyl group could be replaced with 2-pyridyl (2d; 4.6 ± 1.4 μM); however, a significant loss of potency was observed.

In our screening program, we routinely assess compound activity in the presence of glutamate concentrations that generate a response that is either 20% (to detect potentiators) or 80% (to identify antagonists) of a maximal agonist response. Using this strategy, we found that inclusion of a 4-chloro-3-pyridyl group engendered a ‘molecular switch’ within the scaffold, resulting in a compound with mGlu₄ antagonist activity (2e, 7.8 μM, blockade of the EC₈₀ glutamate response by approximately 75%, Table 1).

Table 1.

Initial SAR for mGlu₄ PAM VU0219493

Compd	R	hmGlu4 EC50 (μM)a	Efficacy %PHCCC
2a		1.4 ± 0.4	79.0 ± 3.1
2b		Inactive	12.1 ± 3.1
2c		Inactive	12.2 ± 4.7
2d		4.6 ± 1.4	81.9 ± 2.7
2e		Antagonistb(IC50 = 7.8)	%Inhibition 74.4

^aPotentiation EC₅₀ and efficacy (% PHCCC) are the average of at least three independent determinations performed in triplicate (mean ± SEM shown in table). The maximal response generated in mGlu₄ CHO cells in the presence of mGlu₄ PAMs varies slightly in each experiment; therefore, PAM data are normalized to a control PAM, PHCCC, response obtained in each day’s run.

^bFor 2e, potency and efficacy (% inhibition) data are for blockade of an EC₈₀ glutamate response.

Based on the activity of 2e, a brief SAR around the left-hand pyridyl moiety was undertaken (Table 2). Removal of the pyridine nitrogen led to inactive compounds (3a, 3b, and 3e). In addition, moving the nitrogen around the ring also was unproductive and led to inactive compounds (3c and 3d). Interestingly, the 2-pyridyl compound was an activator (2d), but other regioisomers were inactive.

Table 2.

SAR of the pyridyl group

Compd	R	hmGlu4 IC50 (μM)a	% Inhibition
2e		8.2 ± 0.4	63.1 ± 6.6
3a		Inactive	−3.1 ± 1.3
3b		Inactive	−6.9 ± 3.0
3c		Inactive	−4.0 ± 6.1
3d		Inactive	−1.5 ± 1.5
3e		Inactive	2.3 ± 1.6

^aAntagonist activity at mGlu4 is assessed by measuring the ability of a compound to concentration-dependently inhibit an EC₈₀ glutamate response. IC₅₀ and % inhibition values are the average of at least three independent determinations (mean ± SEM shown in table). Compounds were ranked as active if they inhibited the EC₈₀ response by at least 20%.

A more extensive SAR optimization was undertaken to further evaluate compounds as antagonists of the mGlu₄ receptor (Table 3). The initial antagonist hit was re-synthesized and evaluated in antagonist mode where potency and efficacy were confirmed (8.2 ± 0.4 μM; 63.1% inhibition). Further evaluation by displacement of the 4-chloro, however, was not productive as all compounds were inactive (4a–h and 4j–m) or very weak in their activity (4i, >10 μM, 25.5% inhibition).

Table 3.

SAR of substituted pyridine compounds

Compd	R-X	hmGlu4 IC50 (μM)a	% Inhibition
2e	cl	8.2 ± 0.4	63.1 ± 6.6
4a		Inactive	−5.4 ± 6.1
4b		Inactive	−7.3 ± 3.9
4c		Inactive	−7.2 ± 4.8
4d		Inactive	−7.2 ± 2.3
4e		Inactive	−9.9 ± 4.9
4f		Inactive	−2.9 ± 5.8
4g		Inactive	−15.8 ± 6.1
4h		Inactive	2.2 ± 6.0
4i		>10	25.5 ± 9.2
4j		Inactive	−4.8 ± 3.0
4k		Inactive	0.9 ± 5.4
4l		Inactive	14.0 ± 15.5
4m		Inactive	4.3 ± 13.9

^aAntagonist activity at mGlu4 is assessed by measuring the ability of a compound to concentration-dependently inhibit an EC₈₀ glutamate response. IC₅₀ and % inhibition values are the average of at least three independent determinations performed in triplicate (mean ± SEM shown in table). Compounds were ranked as active if they inhibited the EC₈₀ response by at least 20%.

In contrast to the 4-chloro-3-pyridyl portion of the molecule, there was somewhat more tractable SAR on the right-hand aryl portion of the scaffold (Table 4). These compounds were either made in an analogous fashion as shown in Table 1, or, for the biaryl compounds, they were synthesized from compound 2e via a palladium- mediated Suzuki cross coupling.¹² The ether could not be replaced by a thio-ether (5a) or an unsubstituted phenyl derivative (5b); however, the 2-bromo could be substituted with a 3-bromo without significant loss in potency or efficacy (5d, 7.7 ± 0.8 μM; 58.5% inhibition). 2-fluoro (5c), 3-methoxy (5e), 3-dimethyl amine (5f), 4-methyl (5h), 3-methyl ester (5g) and 4-methoxy (5i) groups were not tolerated. In contrast, 4-trifluoromethyl was similar in potency to the original antagonist (5j, 4.7 ± 0.5 μM; 59.8% inhibition). The SAR of the 2-substituted phenyl derivatives was not predictive and rather shallow. For example, the 2-phenyl and 2-(2-pyridyl) derivatives were active with weak inhibition (5n, >10 μM; 31.1% inhibition and 5o, 8.0 ± 0.9 μM; 60.0% inhibition, respectively). In contrast, the other two pyridyl regioisomers (5p and 5q), as well as a cyclohexyl or other replacements (5s, 5u, 5v, and 5x), were inactive. Several other compounds evaluated exhibited weak antagonist activity (5t; >10 μM, 56.2% inhibition, 5w; >10 μM; 36.5% inhibition).

Table 4.

Compd	R1	hmGlu4 EC50 (μM)a	% Inhibition
5a		Inactive	12.0 ± 12.2
5b		Inactive	−8.7 ± 4.2
5c		Inactive	−10.7 ± 10.2
5d		7.7 ± 0.8	58.5 ± 16.3
5e		Inactive	−2.8 ± 5.0
5f		Inactive	4.4 ± 9.9
5g		Inactive	12.9 ± 13.2
5h		Inactive	16.9 ± 11.0
5i		Inactive	−18.3 ± 12.0
5j		4.7 ± 0.5	59.8 ± 14.6
5l		Inactive	−17.0 ± 3.6
5m		Inactive	−14.8 ± 8.7
5n		>10	31.1 ± 15.8
5o		8.0 ± 0.9	60.0 ± 9.4
5p		Inactive	−3.4 ± 2.4
5q		Inactive	−4.8 ± 6.4
5r		Inactive	7.8 ± 3.3
5s		Inactive	7.4 ± 5.2
5t		>10	56.2 ± 12.7
5u		Inactive	−10.4 ± 9.4
5v		Inactive	−0.1 ± 4.6
5w		>10	36.5 ± 13.7
5x		Inactive	−7.1 ± 9.2

^aIC₅₀ and % EC₈₀ values are the average of at least three independent determinations performed in triplicate (mean ± SEM shown in table). Compounds were ranked as active if they inhibited the EC₈₀ response by at least 20%.

To confirm that these compounds were specific to mGlu₄, we counter screened the entire set against the original screening cell line which contains the human M4 muscarinic receptor co-expressed with Gq_i5. No compounds significantly affected an EC₂₀ response to the M4 agonist acetylcholine; 5r and 5s, which were inactive at mGlu4, weakly inhibited EC₈₀ responses to ACh by 26% and 23%, respectively. As a further validation that these compounds were inhibitors of mGlu4, the lead antagonist for this series, VU0448383, was confirmed to be active in the rat mGlu₄ thallium flux assay (Fig. 2). Although this antagonist, and the series in general, is of low potency, it is important to note that in our previous reports exploring the structure activity relationships of other mGlu₄ PAM scaffolds (representing approximately 5,000 compounds), we have never observed a similar pharmacology ‘switch’ for mGlu₄, suggesting that this compound represents a major advance as a tool compound for the study of mGlu4 biology.

Figure 2.

VU0448383 antagonizes glutamate induced responses in human mGlu4/Gq_i5 cells (calcium assay) and rat mGlu4/GIRK cells (thallium assay) in a concentration-dependent fashion.

In conclusion, we have shown that slight structural modifications to a novel mGlu₄ PAM scaffold lead to a ‘molecular switch’, resulting in a novel mGlu₄ antagonist. Specifically, changing the 2,4-dihydroxyphenol to a 4-chloro-3-pyridyl moiety resulted in the switch from PAM to antagonist (2e). Further modifications of the 4-chloro-3-pyridyl moiety were not tolerated suggesting a shallow binding pocket for that portion of the molecule. The far right side of the molecule did allow for some modification; however, all compounds were shown to be equipotent with 2e. Although ‘molecular switches’ have been reported for other mGlu receptors, to our knowledge this is the first report for a Group III mGlu receptor, namely, mGlu₄. Further studies are planned and will be reported in due course.