Design of 4-Oxo-1-aryl-1,4-dihydroquinoline-3-carboxamides as Selective Negative Allosteric Modulators of Metabotropic Glutamate Receptor Subtype 2

Abstract

Both orthosteric and allosteric antagonists of the group II metabotropic glutamate receptors (mGlus) have been used to establish a link between mGlu_2/3 inhibition and a variety of CNS diseases and disorders. Though these tools typically have good selectivity for mGlu_2/3 versus the remaining six members of the mGlu family, compounds that are selective for only one of the individual group II mGlus have proved elusive. Herein we report on the discovery of a potent and highly selective mGlu₂ negative allosteric modulator 58 (VU6001192) from a series of 4-oxo-1-aryl-1,4-dihydroquinoline-3-carboxamides. The concept for the design of this series centered on morphing a quinoline series recently disclosed in the patent literature into a chemotype previously used for the preparation of muscarinic acetylcholine receptor subtype 1 positive allosteric modulators. Compound 58 exhibits a favorable profile and will be a useful tool for understanding the biological implications of selective inhibition of mGlu₂ in the CNS.

Graphical abstract

INTRODUCTION

Glutamate (L-glutamic acid) is the major excitatory neurotransmitter in the mammalian central nervous system (CNS) and exerts its effects through both ionotropic and metabotropic glutamate receptors (mGlus). The mGlus belong to family C of the G-protein-coupled receptors (GPCRs) and are characterized by a seven-transmembrane (7TM) α-helical domain connected via a cysteine-rich region to a large bi-lobed extracellular amino-terminal domain. The orthosteric binding site is found within this amino-terminal domain for each of the eight members of the mGlu family. The mGlus are further categorized into three groups according to their homology, preferred signal transduction mechanisms, and pharmacology. The group I mGlus (mGlu₁ and mGlu₅) are primarily located postsynaptically in neurons and coupled via G_q to the activation of phospholipase C, which leads to the elevation of intracellular calcium and activation of protein kinase C (PKC). On the other hand, group II mGlus (mGlu₂ and mGlu₃) and group III mGlus (mGlu₄, mGlu₆, mGlu₇, and mGlu₈) are primarily located presynaptically and are coupled via G_i/o to the inhibition of adenylyl cyclase activity.^1–3The expression of the group II mGlus is wide throughout the CNS; moreover, both are found in brain regions associated with emotional states such as the amygdala, hippocampus, and prefrontal cortex.^4,5

With multiple compounds having advanced into clinical trials in schizophrenic patients, the design of selective and druglike positive allosteric modulators (PAMs) of mGlu₂ is significantly more advanced than complementary research directed toward selective negative allosteric modulators (NAMs) of the same receptor.⁶ Still, the literature contains multiple examples of highly optimized orthosteric antagonists and NAMs of the group II mGlus. Though these compounds typically possess good levels of selectivity against the other members of the mGlu family, they lack appreciable selectivity between mGlu₂ and mGlu₃.⁷ Consequently, much has been learned regarding the potential utility of group II mGlu inhibition through the use of these tools in animal models of various CNS disorders. The bulk of such studies have employed the two orthosteric antagonists 1 (LY341495)⁸ and 2 (MGS0039)⁹ (Figure 1). Specifically, potential therapeutic applications of group II mGlu antagonists have been established in obsessive-compulsive disorder (OCD),^10,11 anxiety,¹² cognition,¹³ and Alzheimer’s disease.^14–16 Additionally, substantial work with these compounds has been directed toward establishing a role for mGlu_2/3 antagonists as novel antidepressants.^{9,10,12,17–22} Perhaps most intriguing are studies demonstrating efficacy in animal models of treatment-resistant depression (TRD)²³ and anhedonia.²⁴

Figure 1.

mGlu_2/3 orthosteric antagonist tools 1 and 2, mGlu_2/3 NAM tools 3 and 4, Roche mGlu_2/3 NAM clinical compound 5, first-generation selective mGlu₃ NAMs 6 and 7, and mGlu₃ NAM in vivo tool 8.

Reports of in vivo studies with mGlu_2/3 NAMs are less prevalent; yet two related compounds from a series of 4-aryl-1,3-dihydro-2H-benzo[b][1,4]diazepin-2-ones, 3 (RO4491533)²⁵ and 4 (RO4432717)^26,27 (Figure 1) are worth noting. Studies in rodent models of depression^25,28 and cognition^27,29,30 with these tools have been disclosed and show similar results as those observed with mGlu_2/3 orthosteric antagonists. Additionally, another structurally distinct mGlu_2/3 NAM, 5 (decoglurant, RO4995819)³¹ from Roche (Figure 1), advanced into a phase II trial in patients with major depressive disorder (MDD) (NCT01457677).³² Thus, the evidence for a therapeutic application with mGlu_2/3 antagonists is compelling; however, further elucidation is required regarding the individual importance of mGlu₂ and mGlu₃ in these various disorders. As such, we have been pursuing the design of selective antagonists of each receptor for use as in vivo tools. Our initial success came in the design of selective mGlu₃ NAMs from a series of 1,2-diphenylethyne compounds represented by tool compounds 6 (VU0463597, ML289)³³ and 7 (VU0469942, ML337)^34,35 (Figure 1). More recently we reported on another mGlu₃ NAM, 8 (VU0650786),³⁶ that is a superior in vivo tool and has demonstrated efficacy in rodent models of anxiety/OCD and depression.³⁶ Having selective mGlu₃ NAMs from multiple chemotypes in hand, we sought strategies for the design of selective mGlu₂ NAMs for the purpose of thoroughly evaluating the therapeutic potential of each individual target. Our first successful execution of such a strategy is described in this manuscript.

RESULTS AND DISCUSSION

Scaffold Design

In our search for new scaffolds suitable for the design of selective mGlu₂ NAMs, we were intrigued by a set of quinoline-2-carboxamide compounds 9 developed at Merck and disclosed in the patent literature (Figure 2).^7,37 A survey of the functional mGlu₂ NAM activity presented in this application showed substantial tolerance for functional diversity at the 7-position with a variety of linkers connected to a number of unsaturated and saturated ring systems (A). The 4-position demonstrated a preference for aryl and heteroaryl rings (B), and a primary amide was preferred over a nitrile at the 2-position. We prepared an exemplar compound 10 and tested it in our own cell-based functional assays for mGlu₂ and mGlu₃.³⁵ These fluorescence-based assays measure calcium mobilization induced by receptor activation in a cell line stably expressing either rat mGlu₂ or rat mGlu₃ along with the promiscuous G-protein G_α15 and are capable of detecting agonists, PAMs, and NAMs. Compound 10 exhibited potent NAM activity at mGlu₂ and no evidence of mGlu₃ activity up to the highest concentration tested (30 μM). The quinoline-2-carboxamide mGlu₂ NAMs were reminiscent of another series of allosteric modulators developed at Merck, the 4-oxo-1,4-dihydroquinoline-3-carboxylic acid muscarinic acetylcholine receptor subtype 1 (M₁) PAMs 11.^38–40 Our hypothesis was that a new mGlu₂ NAM scaffold 12 might be obtained within a 4-oxo-1,4-dihydroquinoline series by appending appropriately linked groups (A) at the 6-position and installing N-aryl rings (B) at the 1-position in the context of a primary amide at the 3-position. In addition, the extensive M₁ PAM SAR already developed in this chemotype indicated that such changes would not be favorable for that target.

Figure 2.

Merck quinoline-2-carboxamide mGlu₂ NAM scaffold 9 and representative compound 10; Merck 4-oxo-1,4-dihydroquinoline-3-carboxylic acid M₁ PAM scaffold 11; proposed 4-oxo-1-aryl-1,4-dihydroquinoline-3-carboxamide mGlu₂ NAM scaffold 12.

Synthesis of Compounds

It was envisioned that a number of interesting analogs could be prepared from versatile 6-bromo intermediate 15 (Scheme 1). The synthesis began with commercially available acid 13. Treatment of 13 with 2 equiv of butyllithium followed by addition of 5-bromo-2-fluorobenzoyl chloride provided β-ketoester 14. Reaction of 14 with N,N-dimethylformamide dimethyl acetal followed by a suitable arylamine under microwave heating afforded the desired intermediate 15. Compound 15 could be converted to primary amide 16 through heating in ammonia in methanol under microwave irradiation to give 16. Where possible, 16 was used as a common intermediate; however, certain transformations proved incompatible with the primary amide functional group and necessitated the use of ester 15 with subsequent conversion to the primary amide at a later stage. Reaction of 16 with commercially available aryl alcohols (R¹OH) in the presence of copper(I) iodide and dimethylglycine provided aryl ether analogs 17–19 (Table 1). For synthesis of ethers 23–29 (Table 1), conversion of bromide 15 to alcohol 20 was accomplished with a palladium catalyzed hydroxylation.⁴¹ Acid 20 was converted to methyl ester 21 via a Fischer esterification. A Mitsunobu coupling⁴² with commercial alcohols (R²OH) was employed for installation of the various 6-substituted ethers to afford 22. Conversion of the ester moieties to the corresponding primary amides to yield 23–29 was carried out as described previously. Finally, the synthesis of amine analogs 31–42 (Table 2) was accomplished by first reacting bromide 15 with commercially available amines in a Buchwald–Hartwig amination reaction⁴³ to yield 30. Conversion of 30 to analogs 31–42 was carried out as described previously.

Scheme 1. Synthesis of 6-Heteroatom Linked Analogs^a.

^aReagents and conditions: (a) n-BuLi, 2,2′-bipyridyl, −30 °C to −5 °C, then 5-bromo-2-fluorobenzoyl chloride, −78 °C to −30 °C, 67%; (b) N,N-dimethylformamide dimethyl acetal, DMF, microwave, 120 °C, 15 min, then ArNH₂, microwave, 150 °C, 20 min, 60–98%; (c) 7 N NH₃ in MeOH, microwave, 150 °C, 60 min, 29–99%; (d) R¹OH, CuI, Cs₂CO₃, Me₂NCH₂CO₂H, microwave, 150 °C, 15 min, 26–56% (e) KOH, Pd₂(dba)₃, t-BuXphos, dioxane, H₂O, microwave, 150 °C, 15 min, 99%; (f) MeOH, conc H₂SO₄, reflux, 74%; (g) R²OH, PPh₃, D^tBAD, THF, 40–98%; (h) HNR³R⁴, Pd₂(dba)₃, Xantphos, Cs₂CO₃, PhMe, 110 °C, 8–54%.

Table 1.

mGlu₂ NAM and in Vitro DMPK Results with 6-Substituted Ethers

compd	A	R1	R2	mGlu2 pIC50 ± SEMa	mGlu2 IC50 (nM)a	% Glu max ± SEMa,b	cLogPc	LLEd	rat plasma fue	rat CLhep (mL min−1 kg−1)f
17	I	H	H	6.07 ± 0.09	850	1.84 ± 0.47	2.80	3.27	0.083	40.6
18	I	F	H	5.89 ± 0.07	1280	0.87 ± 0.15	2.90	2.99
19	I	H	F	6.05 ± 0.12	887	1.20 ± 0.29	3.31	2.74
23	II	H	H	6.05 ± 0.07	895	1.63 ± 0.33	2.92	3.13
24	II	Me	H	6.29 ± 0.10	515	1.55 ± 0.39	3.11	3.18	0.056	43.0
25	II	H	Me	6.41 ± 0.05	386	1.31 ± 0.45	3.42	2.99	0.072	64.2
26	III	H	H	6.05 ± 0.07	895	1.63 ± 0.33	2.92	3.13
27	III	Me	H	6.39 ± 0.09	403	1.09 ± 0.16	3.11	3.28	0.091	58.0
28	III	H	Me	6.35 ± 0.06	450	1.07 ± 0.13	3.42	2.93	0.061	63.6
29	IV			6.14 ± 0.06	723	1.30 ± 0.58	2.73	3.41

^aCalcium mobilization mGlu₂ assay; values are average of n ≥ 3.

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

^cCalculated using Dotmatics Elemental (www.dotmatics.com/products/elemental/)

^dLLE (ligand-lipophilicity efficiency) = pIC₅₀ – cLogP.

^ef_u = fraction unbound.

^fPredicted hepatic clearance based on intrinsic clearance (CL_int) in rat liver microsomes.

Table 2.

mGlu₂ NAM and in Vitro DMPK Results with 6-Substituted Amines

compd	A	X	R	mGlu2 pIC50 ± SEMa	mGlu2 IC50 (nM)a	% Glu max ± SEMa,b	cLogPc	LLEd	rat plasma fue	rat CLhep (mL min−1 kg−1)f
31	I		H	6.48 ± 0.12	328	2.11 ± 0.42	2.57	3.91	0.084	48.2
32	I		Me	6.50 ± 0.37	318	−0.65 ± 1.98	2.95	3.55	0.061	60.4
33	II		Me	6.06 ± 0.06	874	1.16 ± 0.55	2.95	3.11
34	I	CH2	H	6.47 ± 0.10	341	1.38 ± 0.58	2.61	3.86	0.137	51.3
35	I	CH2	Me	6.70 ± 0.03	201	2.39 ± 0.13	2.92	3.78	0.082	67.6
36	I	CH2	Et	6.80 ± 0.01	159	2.36 ± 0.14	3.33	3.47	0.067	68.2
37	II	CH2	Me	6.70 ± 0.09	201	1.76 ± 0.74	2.92	3.78	0.089	67.0
38	III	CH2	Me	6.83 ± 0.09	147	1.86 ± 0.26	2.74	4.09	0.340	47.1
39	IV	CH2	H	6.27 ± 0.05	535	1.15 ± 0.47	2.87	3.40
40	I	CH2CH2	H	6.53 ± 0.12	296	1.98 ± 0.67	2.71	3.82	0.037	63.8
41	II	CH2CH2	H	6.42 ± 0.11	376	2.01 ± 0.40	2.71	3.71	0.172	60.5
42	V	CH2CH2	H	6.09 ± 0.07	816	1.13 ± 0.18	1.56	4.53

^aCalcium mobilization mGlu₂ assay; values are average of n ≥ 3.

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

^cCalculated using Dotmatics Elemental (www.dotmatics.com/products/elemental/).

^dLLE (ligand-lipophilicity efficiency) = pIC₅₀ – cLogP.

^ef_u = fraction unbound.

^fPredicted hepatic clearance based on intrinsic clearance (CL_int) in rat liver microsomes.

In addition to the 6-heteroatom linked analogs, 6-carbon linked compounds were prepared from intermediates 15 and 16 (Scheme 2). Methylene-linked tertiary amine analogs 45–66 (Tables 3 and 4) were accessed through bromide 16, which was first converted to vinyl intermediate 43 via a Suzuki coupling with potassium vinyltrifluoroborate.⁴⁴ Dihydroxylation of the olefin and subsequent in situ periodate cleavage of the resultant diol gave aldehyde 44. Analogs 45–66 were then prepared through reductive aminations with 44 and commercially available secondary amines (HNR²R³). For preparation of methyleneoxy linked analogs 70–78 (Table 5), bromide 15 was converted to aldehyde 67 via an analogous vinylation, dihydroxylation, and periodate cleavage as described above. Sodium borohydride reduction of 67 gave primary alcohol 68, which was reacted in a Mitsunobu coupling⁴² with commercial alcohols (R⁴OH) to give ether intermediate 69. Conversion of the ester moieties to the corresponding primary amides to yield 70–78 was carried out as described previously. Ethylene linked analogs 81–91 (Table 6 and Table 7) were also prepared from bromide 15 through initial preparation of alkynes 79. Two methods were employed for preparation of these alkyne intermediates 79, each relying on Sonogashira couplings⁴⁵ with bromide 15. A coupling with 15 and a terminal alkyne (R⁶CCH) gave 79 directly. Alternatively, a coupling with trimethylsilylacetylene followed by fluoride mediated silyl cleavage gave a 6-alkyne intermediate that was coupled to an aryl bromide (R⁶Br) to afford 79. A palladium catalyzed hydrogenation of the alkyne moiety provided 80, which was reacted with ammonia as described previously to yield the target compounds 81–91.

Scheme 2. Synthesis of 6-Carbon Linked Analogs^a.

^aReagents and conditions: (a) H₂CCHBF₃K, Pd(dppf)·CH₂Cl₂, NEt₃, n-propanol, 100 °C, 75–100%; (b) OsO₄, NMO, acetone, H₂O, then NaIO₄, 91–99%; (c) HNR²R³, NaBH(OAc)₃, AcOH, CH₂Cl₂, 7–81%; (d) NaBH₄, EtOH, 0 °C, 36–57%; (e) R⁴OH, PPh₃, D^tBAD, THF, 14–98%; (f) 7 N NH₃ in MeOH, microwave, 150 °C, 15 min, 10–94%; (g) R⁶CCH, PdCl₂(PPh₃)₂, CuI, NEt₃, DMF, microwave, 150 °C, 15 min, 23–54%; (h) Me₃SiCCH, PdCl₂(PPh₃)₂, CuI, NEt₃, DMF, microwave, 150 °C, 15 min, 83%; (i) TBAF, THF, 70%; (j) R⁶Br, PdCl₂(PPh₃)₂, CuI, NEt₃, DMF, microwave, 150 °C, 15 min, 21–47%; (k) 10% Pd/C, MeOH, H₂ (1 atm), 63–99%.

Table 3.

mGlu₂ NAM and in Vitro DMPK Results with 6-Substituted Methylene Amines

compd	A	R1	R2	mGlu2 pIC50 ± SEMa	mGlu2 IC50 (nM)a	% Glu max ± SEMa,b	cLogPc	LLEd	rat plasma fue	rat CLhep (mL min−1 kg−1)f
45	I			6.40 ± 0.01	396	1.16 ± 0.18	3.86	2.54	0.138	60.9
46	II			5.66 ± 0.09	2170	0.43 ± 0.75	4.09	1.57
47	III			5.75 ± 0.06	1770	1.44 ± 0.30	3.39	2.36
48	IV			6.67 ± 0.09	214	1.10 ± 0.12	3.32	3.35	0.156	56.5
49	V	H	H	<5.0g	>10000	37.2 ± 9.7	3.55	<1.45
50	V	H	CF3	6.21 ± 0.09	618	1.11 ± 0.28	4.21	2.00
51	V	H	CN	6.23 ± 0.07	587	0.90 ± 0.33	2.94	3.29
52	V	H	OMe	5.77 ± 0.11	1720	1.03 ± 0.26	3.08	2.69
53	V	H	SO2Me	6.05 ± 0.09	886	1.36 ± 0.26	2.34	3.71
54	V	F	F	6.79 ± 0.19	161	1.59 ± 0.27	3.77	3.02	0.109	48.6

^aCalcium mobilization mGlu₂ assay; values are average of n ≥ 3.

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

^cCalculated using Dotmatics Elemental (www.dotmatics.com/products/elemental/).

^dLLE (ligand-lipophilicity efficiency) = pIC₅₀ – cLogP.

^ef_u = fraction unbound.

^fPredicted hepatic clearance based on intrinsic clearance (CL_int) in rat liver microsomes.

^gWeak activity; concentration–response curve (CRC) does not plateau.

Table 4.

mGlu₂ NAM and in Vitro DMPK Results with 6-Substituted Methylene Amines (Continued)

compd	A	R1	R2	X	mGlu2 pIC50 ± SEMa	mGlu2 IC50 (nM)a	% Glu max ± SEMa,b	cLogPc	LLEd	rat plasma fue	rat CLhep (mL min−1 kg−1)f
55	I	Me	H		6.47 ± 0.16	341	0.98 ± 0.24	2.59	3.88
56	I	H	Me		6.26 ± 0.16	551	1.33 ± 0.15	2.59	3.67
57	I	Me	Me		6.93 ± 0.09	119	2.14 ± 0.17	2.99	3.94	0.16	51.2
58	II				6.69 ± 0.04	207	1.48 ± 0.29	3.10	3.59	0.306	24.5
59	III			S	6.38 ± 0.09	420	1.52 ± 0.63	2.76	3.62	0.205	59.6
60	III			SO2	6.06 ± 0.08	870	1.33 ± 0.38	1.35	4.71
61	III			NMe	5.95 ± 0.06	1110	1.15 ± 0.32	2.09	3.86
62	III			NCH2CF3	6.21 ± 0.07	612	1.10 ± 0.07	2.90	3.31
63	IV			CH2	5.68 ± 0.10	2080	0.55 ± 0.21	4.00	1.68
64	IV			O	<5.0g	>10000	24.2 ± 3.6	2.54	<2.46
65	IV			S	6.86 ± 0.10	138	2.19 ± 0.24	3.21	3.65	0.074	63.8
66	IV			SO2	6.38 ± 0.06	415	1.24 ± 0.04	1.80	4.58	0.253	23.7

^aCalcium mobilization mGlu₂ assay; values are average of n ≥ 3.

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

^cCalculated using Dotmatics Elemental (www.dotmatics.com/products/elemental/).

^dLLE (ligand-lipophilicity efficiency) = pIC₅₀ – cLogP.

^ef_u = fraction unbound.

^fPredicted hepatic clearance based on intrinsic clearance (CL_int) in rat liver microsomes.

^gWeak activity; CRC does not plateau.

Table 5.

mGlu₂ NAM and in Vitro DMPK Results with 6-Aryloxymethyl Ethers

compd	A	R1	R2	mGlu2 pIC50 ± SEMa	mGlu2 IC50 (nM)a	% Glu max ± SEMa,b	cLogPc	LLEd	rat plasma fue	rat CLhep (mL min−1 kg−1)f
70	I	F	H	6.13 ± 0.06	746	0.31 ± 0.30	3.42	2.71
71	I	H	F	6.35 ± 0.10	443	1.13 ± 0.38	3.42	2.93
72	I	H	Me	6.61 ± 0.11	247	1.62 ± 0.23	3.11	3.50	0.087	43.9
73	I	H	Cl	6.71 ± 0.10	193	1.70 ± 0.12	3.94	2.77	0.033	52.1
74	I	H	CF3	5.88 ± 0.04	1330	1.23 ± 0.02	3.83	2.05
75	II	F	H	6.64 ± 0.10	228	1.83 ± 0.20	3.02	3.62	0.063	59.8
76	II	H	Me	6.45 ± 0.09	351	1.90 ± 0.32	3.11	3.34	0.050	37.2
77	II	H	CF3	6.31 ± 0.06	486	1.70 ± 0.14	3.83	2.48
78	III			6.56 ± 0.08	277	1.71 ± 0.20	2.73	3.83	0.109	22.4

^aCalcium mobilization mGlu₂ assay; values are average of n ≥ 3.

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

^cCalculated using Dotmatics Elemental (www.dotmatics.com/products/elemental/).

^dLLE (ligand-lipophilicity efficiency) = pIC₅₀ – cLogP.

^ef_u = fraction unbound.

^fPredicted hepatic clearance based on intrinsic clearance (CL_int) in rat liver microsomes.

Table 6.

mGlu₂ NAM and in Vitro DMPK Results with 6-Ethylene Linked Analogs

compd	A	R1	R2	mGlu2 pIC50 ± SEMa	mGlu2 IC50 (nM)a	% Glu max ± SEMa,b	cLogPc	LLEd	rat plasma fue	rat CLhep (mL min−1 kg−1)f
81	I			5.16 ± 0.09	6890	2.07 ± 1.38	4.65	0.21
82	II	H		6.33 ± 0.09	471	1.88 ± 0.38	3.34	2.99	0.057	66.7
83	II	CF3		6.52 ± 0.09	304	0.88 ± 0.06	4.25	2.27	0.039	68.4
84	III	H	H	6.33 ± 0.08	466	1.90 ± 0.11	3.34	2.99	0.062	64.1
85	III	CF3	H	5.76 ± 0.03	1720	1.22 ± 0.14	4.25	1.51
86	III	H	F	5.93 ± 0.04	1170	0.80 ± 0.34	3.44	2.49
87	IV			6.67 ± 0.10	215	1.52 ± 0.26	3.16	3.51	0.157	46.9

^aCalcium mobilization mGlu₂ assay; values are average of n ≥ 3.

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

^cCalculated using Dotmatics Elemental (www.dotmatics.com/products/elemental/).

^dLLE (ligand-lipophilicity efficiency) = pIC₅₀ – cLogP.

^ef_u = fraction unbound.

^fPredicted hepatic clearance based on intrinsic clearance (CL_int) in rat liver microsomes.

Table 7.

mGlu₂ NAM and in vitro DMPK Results with Modified 1-Position Analogs

compd	B	mGlu2 pIC50 ± SEMa	mGlu2 IC50 (nM)a	% Glu max ± SEMa,b	cLogPc	LLEd	rat plasma fue	rat CLhep (mL min−1 kg−1)f
87	I	6.67 ± 0.10	215	1.52 ± 0.26	3.16	3.51	0.157	46.9
88	II	6.87g	136g	1.61g	3.02	3.85	0.098	53.4
89	III	6.58 ± 0.09	266	1.30 ± 0.21	3.12	3.46	0.085	53.6
90	IV	6.70 ± 0.13	198	1.52 ± 0.38	3.12	3.58	0.159	60.5
91	V	6.38 ± 0.05	414	1.20 ± 0.50	2.66	3.72	0.220	41.9

^aCalcium mobilization mGlu₂ assay; values are average of n ≥ 3.

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

^cCalculated using Dotmatics Elemental (www.dotmatics.com/products/elemental/).

^dLLE (ligand-lipophilicity efficiency) = pIC₅₀ – cLogP.

^ef_u = fraction unbound.

^fPredicted hepatic clearance based on intrinsic clearance (CL_int) in rat liver microsomes.

^gValue is average of n = 2.

In the case of the aforementioned 4-oxo-1,4-dihydroquinoline-3-carboxylic acid M₁ PAM scaffold, Merck has also shown that a 4H-quinolizin-4-one functioned as an effective bioisostere for the core of the chemotype.^46–48 As such, we decided to prepare analogs with this core to evaluate its potential as an mGlu₂ NAM scaffold as well (Scheme 3). The synthesis began with lithiation of 5-hydroxy-2-methylpyridine and subsequent in situ reaction with commercially available diethyl 2-(ethoxymethylene)malonate 92 to give 7-hydroxy-4-oxo-4H-quinolizine 93. The alcohol was protected as its methoxymethyl ether 94, which was then selectively brominated at the 1-position to afford 95. A Suzuki coupling with 4-fluorophenylboronic acid provided intermediate 96. Acidic cleavage of the protecting group and simple filtration of the precipitated product gave 97, which served as the key intermediate for the synthesis of analogs. Ether compounds 99–102 (Table 8) were prepared via Mitsunobu coupling⁴² with commercial alcohols (R¹OH) to yield 98, which was converted to primary amides 99–102 as described above. Intermediate 97 was also converted to the corresponding triflate 103, which was subjected to an analogous vinylation, dihydroxylation, and periodate cleavage as described previously to afford aldehyde 104. Finally, conversion of 104 to amines 105 and ultimately final compounds 106–108 followed methods outlined herein above.

Scheme 3. Synthesis of 4H-Quinolizin-4-one Analogs^a.

^aReagents and conditions: (a) 5-hydroxy-2-methylpyridine, n-BuLi, THF, −78 °C to −30 °C, 57%; (b) CH₃OCH₂Cl, DIEA, CH₂Cl₂, 0 °C to rt, 96%; (c) NBS, CHCl₃, 0 °C to rt, 96%; (d) 4-fluorophenylboronic acid, Pd(dppf)·CH₂Cl₂, 1 M aq Na₂CO₃, DME, 90 °C, 94%; (e) pTSA·H₂O, EtOH, DCE, 80 °C, 66%; (f) R¹OH, PPh₃, D^tBAD, THF, 0 to 45 °C, 38–82%; (g) 7 N NH₃ in MeOH, microwave, 150 °C, 2.0–3.0 h, 26–90%; (h) PhN(SO₂CF₃)₂, NEt₃, CH₂Cl₂, 0 °C, 96%; (i) H₂CCHBF₃K, Pd(dppf)·CH₂Cl₂, NEt₃, n-propanol, 90 °C, 96%; (j) OsO₄, NMO, THF, H₂O, then NaIO₄, 70%; (k) HNR⁴R⁵, NaBH(OAc)₃, AcOH, CH₂Cl₂, 27–67%.

Table 8.

mGlu₂ NAM Results with 4H-Quinolizin-4-one Analogs

compd	A	X	Y	mGlu2 pIC50 ± SEMa	mGlu2 IC50 (nM)a	% Glu max ± SEMa,b	cLogPc	LLEd	comparatore	fold decrease in potencyf
99	I	O	C–H	5.40 ± 0.02	3980	2.67 ± 0.26	2.83	2.57	24	5.8
100	I	O	N	<5.0g	>10000	19.4 ± 9.3	2.45	<2.55	29	>13
101	II	O	H	5.27 ± 0.07	5360	0.71 ± 1.86	2.63	2.64	26	6.0
102	II	O	CF3	5.34 ± 0.01	4610	2.19 ± 0.69	3.54	1.80
106	III	CH2		<5.0g	>10000	3.95 ± 1.84	2.65	<2.35	58	>48
107	IV	CH2		5.18 ± 0.13	6620	1.04 ± 2.27	2.77	2.41	65	48
108	V	CH2		5.36 ± 0.21	4400	0.93 ± 2.27	2.87	2.49	48	21

^aCalcium mobilization mGlu₂ assay; values are average of n ≥ 3.

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

^cCalculated using Dotmatics Elemental (www.dotmatics.com/products/elemental/).

^dLLE (ligand-lipophilicity efficiency) = pIC₅₀ – cLogP.

^eDirect comparator from 4-oxo-1,4-dihydroquinoline series.

^fFold decrease in potency relative to direct comparator from 4-oxo-1,4-dihydroquinoline series.

^gWeak activity; CRC does not plateau.

mGlu₂ NAM Activity and Preliminary DMPK SAR

As new analogs were evaluated for potency in our functional mGlu₂ assay, interesting compounds were further assessed in our frontline in vitro drug metabolism and pharmacokinetics (DMPK) assays. Specifically, metabolic stability was determined by measuring the intrinsic clearance of the compound when incubated with rat liver microsomes (RLMs).⁴⁹ The intrinsic clearance obtained was used to calculate a predicted hepatic clearance, and compounds were binned accordingly into low (<⅓ hepatic blood flow), moderate (⅓ to ⅔ hepatic blood flow) and high (>⅔ hepatic blood flow) groups. The extent to which the compounds were bound to rat plasma was also measured.⁵⁰ We also calculated the lipophilicity of new analogs and attempted to assess the efficiency of the structural modifications being tested.⁵¹ Much of the SAR work was conducted in the context of a 4-fluorophenyl ring at the 1-position of the scaffold, as this was a group with good potency in the quinoline series, and it was likely to be somewhat metabolically stable. As expected, the mGlu₂ NAM SAR with the new 4-oxo-1,4-dihydroquinoline ether analogs showed a good deal of tolerance at the 6-position (Table 1). The ethers with directly linked heteroaryl rings (17–19) were among the least active in this set; however, in vitro DMPK was benchmarked. The fraction unbound in rat plasma with 17 was 0.083, and the predicted hepatic clearance was moderate. The remaining analogs 23–29 possessed a single sp³ hybridized carbon between the 6-position ether oxygen and the heteroaryl ring (A). This feature generally improved mGlu₂ NAM activity. The methyl groups of analogs 24 and 25 both provided small boosts of potency relative to unsubstituted 3-pyridyl ring 23. Analogous 4-pyridyl methyl analogs 27 and 28 were approximately 2-fold more potent than unsubstituted comparator 26. Pyrimidine analog 29 was less potent than 3-pyridyl analog 24; however, installation of this nitrogen atom reduced lipophilicity, and the ligand-lipophilicity efficiency (LLE) of 29 indicated that this 2-methylpyrimidin-5-yl functional group was worthy of continued evaluation in other analogs. The compounds examined (24, 25, 27, 28) in our in vitro DMPK assays again had fraction unbound in rat plasma similar to 17, but only 24 had moderate predicted hepatic clearance with the remaining analogs having CL_hep values near liver blood flow.

4-Oxo-1,4-dihydroquinoline amine analogs further illustrated the tolerance for variation at the 6-position (Table 2). Additionally, many of these analogs exhibited superior mGlu₂ NAM potency compared to the ether analogs discussed above. Simple alkylation of the nitrogen linker generally had minimal impact on mGlu₂ NAM activity as evidenced by comparing secondary amine analogs 31 and 34 to their tertiary amine comparators 32, 35, and 36; however, in each case, the tertiary amine analogs exhibited a higher predicted hepatic clearance, possibly due to N-dealkylation. In the case of analogs with the ring (A) directly attached to the nitrogen linker, the 3-pyridyl ring (32) exhibited superior mGlu₂ NAM activity compared to the 4-pyridyl ring (33). On the other hand, the difference in potency was minimal when a methylene (35 and 37) or ethylene spacer (40 and 41) was inserted between the nitrogen atom and the ring (A). Combining the methylene spacer with the 2-methylpyrimidin-5-yl ring (38) provided the most potent compound in this set. The fraction unbound was considerably higher with 38 relative to other similar analogs (35 and 37), and the predicted hepatic clearance, though still high, was less than the majority of the other analogs in this set. Though analogs 39 and 42 were not among the most potent amine analogs, these derivatives demonstrated that saturated heteroaryl rings were also tolerated at the 6-position of the chemotype.

Having observed that aromatic rings were not required at the 6-position, we were interested to evaluate the numerous methylene amine analogs prepared at that position. Several of these analogs were simple tertiary amines without additional heteroatoms in the ring system (Table 3). The mGlu₂ NAM activity observed with these compounds was generally more dependent on minor structural changes than had been observed with previous compounds. For example, difluorocyclobutylamine 45 was approximately 5-fold more potent than cyclopentylamine 46, and difluoropyrrolidine 48 was approximately 8-fold more potent than difluoroazaspiroheptane 47. Likewise, though unsubstituted piperidine 49 was only a weak mGlu₂ NAM, inhibiting the glutamate response only at the highest concentration (30 μM), further substitution of the ring with a variety of moieties enhanced potency (50–54). Three analogs (45, 48, and 54) were evaluated in our in vitro DMPK assays, and while the protein binding results were encouraging with more than 10% unbound in each case, predicted hepatic clearance remained high (>48 mL min⁻¹ kg⁻¹).

In addition to the simple tertiary amines highlighted above, we also prepared a number of analogs with heterocyclic amines (Table 4). Substituted morpholine analogs 55–58 were potent mGlu₂ NAMs with the dimethyl substituted analogs 57 and 58 offering potency superior to monomethyl analogs 55 and 56. Particularly encouraging was analog 58, which was predicted to be a low–moderate clearance compound in rats and was approximately 30% unbound in rat plasma. On the other hand, thiomorpholine 59 exhibited high clearance in vitro, and thiomorpholine 1,1-dioxide 60 showed reduced potency. We also prepared several analogs with seven-membered rings (63–66). Though most of these medium ring-containing analogs were moderate to weak mGlu₂ NAMs, 1,4-thiazepane 65 was quite potent. Unfortunately, 65 was highly cleared in vitro; however, oxidation of the sulfur atom was a likely metabolic soft-spot, as clearance with 1,4-thiazepane 1,1-dioxide 66 was substantially reduced.

Turning our attention to the 6-aryloxymethyl ether analogs 70–78 uncovered several additional compounds with good mGlu₂ NAM potency (Table 5). Several 3-pyridyl derivatives (70–74) were prepared, and 6-methyl derivative 72 and 6-chloro derivative 73 exhibited good potency. Interestingly, a trifluoromethyl group (74) did not function as an adequate alternative at this position. The pyridyl derivatives (75–77) demonstrated more modest differences in mGlu₂ NAM activity, and in this case the trifluoromethyl (77) was only slightly less potent than its corresponding methyl comparator (76). Fraction unbound with these pyridyl analogs was in line with other similar analogs (see Table 1), and predicted clearance ranged from moderate (72 and 76) to high (73 and 75). Once again, we installed a 2-methylpyrimidin-5-yl ring (78) and observed positive results. Specifically, not only was 78 a potent mGlu₂ NAM, it exhibited more than 10% fraction unbound in rat plasma and a low predicted clearance in rat liver microsomes.

Finally, examination of 6-ethylene linked analogs 81–87 yielded a range of results (Table 6). Unsubstituted phenyl analog 81 demonstrated weak mGlu₂ NAM activity; however, modification of the aromatic ring (A) to pyridine (82 and 84) improved potency approximately 15-fold. Unfortunately, both 82 and 84 were highly cleared in vitro. Substitution of 4-pyridyl analog 82 with a trifluoromethyl group (83) modestly enhanced potency but without reducing clearance. Substitution of 3-pyridyl analog 84 with trifluoromethyl (85) and fluorine (86) was unfavorable for mGlu₂ NAM activity. Again the 2-methylpyrimidin-5-yl ring (87) proved an attractive moiety, having demonstrated the most potent activity and highest fraction unbound in this set of analogs. Also, although predicted hepatic clearance for 87 remained on the high end, it was improved relative to other analogs in this class (82–84).

Having developed substantial SAR at the 6-position of the chemotype, we wanted to conduct limited exploration of another area as well. We chose ethylene linked analog 87 as a useful comparator given its overall profile. As such, additional analogs of 87 with alternative aromatic rings (B) to the 4-fluorophenyl ring were prepared and tested (Table 7). Replacement of the 4-fluoro group (87) with a 4-methoxy group (88) improved mGlu₂ NAM potency slightly, but the predicted hepatic clearance of 88 remained high. Since methoxy groups increase electron density on the ring, fluorinated analogs 89 and 90 were prepared; however, these modifications failed to improve metabolic stability. It was encouraging to see that the 4-fluorophenyl ring could be replaced altogether with a 3-methylisothiazol-5-yl ring (91). Analog 91 demonstrated a 2-fold drop in potency relative to 87; yet, this modification was considered efficient by the LLE quotient, as it was a less lipophilic compound. Of note, 91 had an increased fraction unbound and a marginally and perhaps insignificantly lower predicted hepatic clearance than 87. Continued exploration of this region (B) of the scaffold is clearly worthwhile; however, at this point, we decided to more thoroughly profile some of the promising analogs discovered thus far to evaluate the full potential of the 4-oxo-1-aryl-1,4-dihydroquinoline-3-carboxamides as a lead series for the discovery of druglike mGlu₂ NAMs.

Prior to discussing the results of extended profiling of select 4-oxo-1-aryl-1,4-dihydroquinolines, it is worth briefly discussing the results obtained with the 4H-quinolizin-4-one analogs (Table 8). Whether in the case of the ether analogs (99–102) or methylene amine analogs (106–108), mGlu₂ NAM potency was consistently weak. In fact, when compared to their analogous compounds in the 4-oxo-1-aryl-1,4-dihydroquinoline series, these analogs were notably less potent in each case. These results are important because it illustrates that these two cores are not uniformly interchangeable. Moreover, it provides another example of the subtleties of SAR often seen in the design of allosteric modulators of class C GPCRs.^52–54

Extended Characterization of Selected Analogs

In choosing the initial compounds for more in depth evaluation, we sought molecules with both promising mGlu₂ NAM potency and in vitro DMPK profiles while also desiring some structural diversity at the 6-position substituent. As such, we selected ether 17, amine 34, methylene amines 54 and 58, and ethylene linked analog 87 for further profiling (Table 9). Since the potential therapeutic applications for an mGlu₂ NAM are in the area of CNS disorders, blood–brain barrier (BBB) penetration seemed a logical next step for evaluation in this new series. It should be noted that the fraction unbound in rat plasma ranged from 0.083 to 0.306 with these compounds; thus, consideration of unbound fraction alongside potency and CNS exposure is required to fully evaluate these compounds. Toward this end, we employed rat cassette pharmacokinetics (PK) tissue distribution studies using intravenous (iv) dosing and single time point analysis.⁵⁵ Such an approach has repeatedly proven a rapid and cost-effective mechanism for preliminary assessment of BBB penetration. In addition to the already measured protein binding in rat plasma, the protein binding of these compounds in rat brain homogenates was also assessed. Unfortunately, the observed brain to plasma ratio (K_p) for each compound was low, ranging from 0.04 (34) to 0.36 (54). Calculation of the unbound brain to unbound plasma ratio (K_p,uu) gave values that were all below 0.25, indicating possible transporter effects.⁵⁶ Thus, analogs 58 and 87 were selected for permeability studies in Madin–Darby canine kidney (MDCK) cells transfected with the human MDR1 gene to assess potential P-glycoprotein (P-gp) mediated efflux.⁵⁷ Both compounds demonstrated substantial efflux with ratios of 52 and 44, respectively. While it was disappointing to learn that this scaffold appeared to suffer from P-gp-mediated efflux, the absolute CNS concentrations observed with 58 and its other properties raised the possibility that it might still be a valuable tool.

Table 9.

Rat Intravenous Cassette and MDR1-MDCK Permeability Results

compd	mGlu2 IC50 (nM)	rat plasma fua	rat brain fua	rat iv tissue distribution resultsb				permeability in MDR1-MDCK cells
				plasma concn (nM)c	brain concn (nM)c	Kpd	Kp,uue	A–B Papp (10−6 cm/s)	B–A Papp (10−6 cm/s)	efflux ratio
17	850	0.083	0.087	55.8	9.2	0.16	0.17
34	341	0.137	0.095	83.0	3.0	0.04	0.02
54	161	0.109	0.068	60.5	21.6	0.36	0.22
58	207	0.306	0.191	56.9	18.0	0.32	0.23	1.54	79.7	52
87	215	0.157	0.160	122	5.6	0.05	0.05	1.58	69.4	44

^af_u = fraction unbound.

^bn = 2; dose = 0.2 mg/kg per compound; solution in 8% EtOH, 30% PEG 400, 62% DMSO (2 mg/mL total).

^c15 min after dose.

^dK_p = total brain to total plasma ratio.

^eK_p,uu = unbound brain (brain f_u × total brain) to unbound plasma (plasma f_u × total plasma) ratio.

Further profiling of compound 58 (VU6001192) began with determination of its full selectivity versus other members of the mGlu family. Selectivity versus fellow group II receptor subtype, mGlu₃, was particularly critical to assess. Gratifyingly, we evaluated the selectivity of 58 versus rat mGlu₃ using 10-point concentration–response curve (CRC) analysis in the presence of an EC₈₀ concentration of glutamate, and 58 was inactive up to the highest concentration tested (30 μM). For evaluation of selectivity versus other members of the mGlu family, the effect of 10 μM 58 on the orthosteric agonist CRC was measured in fold-shift experiments.^58,59 Fortunately, no activity at the other mGlus was noted in these assays. Because the genesis for the 4-oxo-1-aryl-1,4-dihydroquinoline chemotype as a mGlu₂ NAM scaffold was inspired in part by a known M₁ PAM scaffold, we also evaluated 58 in our human M₁ functional assay⁶⁰ and observed no activity up to the highest concentration tested (30 μM). Ancillary pharmacology was evaluated through screening 58 at 10 μM in a commercially available radioligand binding assay panel of 68 clinically relevant GPCRs, ion channels, kinases, and transporters,⁶¹ and no significant responses were noted.⁶²

Having established the excellent selectivity profile of 58, we progressed the compound to additional and more definitive PK studies in both rats and mice (Table 10).⁶³ In spite of its predicted low–moderate clearance, a time course study using iv dosing showed 58 to be a high clearance compound; however, the volume of distribution at steady state (V_SS) was high and the half-life was approximately 2 h. Thus, intraperitoneal (ip) dosing was chosen as a route that was both convenient for future use in behavioral models and had the likelihood of providing superior exposure to oral dosing. An ip tissue distribution study in rats at 30 mg/kg gave K_p and K_p,uu values similar to those observed previously; yet, we observed a brain concentration of 760 nM, which translates to an unbound brain concentration of 145 nM and is very near the functional mGlu₂ IC₅₀ (207 nM). An analogous study in mice at the same dose gave similar results with a brain concentration of 716 nM, which translates to an unbound brain concentration of 184 nM. Finally, to verify the role of P-gp in vivo, we repeated the study in mice with the modification of pretreating the animals with the known P-gp inhibitor 109 (elacridar).⁶⁴ The impact of this modification was profound, as the exposure of 58 in the brain was increased more than 6-fold without impacting the systemic exposure in plasma.

Table 10.

Intravenous PK and Tissue Distribution Studies with Compound 58

Protein Binding (fu)a		Rat IV PKb				Rat IP Tissue Distributionc,d
rat plasma	0.306	dose		0.2 mg/kg		dose		30 mg/kg
rat brain homogenates	0.191	t1/2		141 minutes		plasma concentration		3900 nM
mouse plasma	0.315	CLplasma		75.8 mL/min/kg		brain concentration		760 nM
mouse brain homogenates	0.257	Vss		13.3 L/kg		Kpe	Kp,uuf	0.20	0.12
		Mouse IP Tissue Distributionc,d				Mouse IP Tissue Distribution + 109c,d,g
		dose		30 mg/kg		dose		30 mg/kg
		plasma concentration		1820 nm		plasma concentration		2160 nm
		brain concentration		716 nm		brain concentration		4580 nm
		Kpe	Kp,uuf	0.39	0.32	Kpe	Kp,uuf	2.1	1.7

^af_u = fraction unbound.

^bn = 2; solution in solution in 9% EtOH, 38% PEG 400, 53% DMSO (1 mg/mL).

^cn = 2; fine homogeneous suspension in 10% Tween-80 in H₂O.

^d15 min after dose of 58.

^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.

^gCompound 109 dosed 1 h prior to compound 58 at 20 mg/kg.

CONCLUSION

A potent and highly selective mGlu₂ NAM tool compound 58 was discovered through scaffold hopping a series in the patent literature and recognizing the possibility that an established M₁ PAM series might function as a viable chemotype for new analog design. Diverse functional groups were tolerated at the 6-position of the new chemotype, and limited work established the potential for further modifications at the 1-position. While the utility of the compounds tested thus far is hampered by P-gp mediated efflux that limits CNS exposure, the overall profile of 58 remains interesting. The compound exhibits an unbound fraction of 25–30% in rodent brain homogenates, and a dose of 30 mg/kg using ip dosing produces unbound brain concentrations near the functional mGlu₂ IC₅₀. Conceivably, higher doses could be employed in order to reach pharmacologically relevant concentrations in the CNS. Perhaps more attractive is the fact that pretreatment with a commercially available P-gp inhibitor boosts the unbound brain exposure of 58 to more than 5-fold the mGlu₂ IC₅₀ at the same dose. The study of 58 in behavioral models relevant to mGlu₂ inhibition is planned and will be the subject of future communications.

EXPERIMENTAL SECTION

The synthesis of compound 58 and its associated intermediates is described below for convenience. Synthetic details for other compounds can be found in the Supporting Information.

Ethyl 3-(5-Bromo-2-fluorophenyl)-3-oxopropanoate (14)

3-Ethoxy-3-oxopropanoic acid 13 (2.16 mL, 18.3 mmol, 2.00 equiv) was dissolved in THF (91 mL) in an oven-dried round-bottom flask, and 2,2′-bipyridyl (8.00 mg, 0.0512 mmol, 0.0056 equiv) was added as an indicator. The reaction was cooled to −30 °C, and n-butyllithium (1.6 M in hexanes) (29.0 mL, 45.6 mmol, 4.00 equiv) was added dropwise over 20 min. Upon final addition the reaction turned red at which point it was allowed to warm to −5 °C. The reaction was allowed to stir at −5 °C for 15 min, during which time the red color began to dissipate. Enough n-butyllithium was added to allow the red color to persist. The reaction was then cooled to −78 °C, and 5-bromo-2-fluorobenzoyl chloride (2.17 g, 9.14 mmol, 1.00 equiv) was added dropwise as a solution in THF (6.9 mL). The reaction was allowed to stir at −78 °C for 30 min and then allowed to warm to −30 °C and stirred for an additional 30 min. The reaction was poured onto ice-cold 1 N HCl (92 mL), and the mixture was extracted with ethyl acetate (1×) and DCM (2×). The combined organics were dried (MgSO₄), filtered, and concentrated in vacuo. Purification by flash chromatography on silica gel afforded 1.78 g (67%) of the title compound as an off-white solid. ¹H NMR (400 MHz, DMSO-d₆): δ = 7.97 (dd, J = 6.5, 2.6 Hz, 1H), 7.91–7.86 (m, 1H), 7.40–7.34 (m, 1H), 4.13–4.07 (m, 4H), 1.15 ppm (t, J = 7.1 Hz, 3H). ¹³C NMR (100 MHz, DMSO-d₆): δ = 189.68 (d, J(C,F) = 3.2 Hz), 167.09, 160.26 (d, J(C,F) = 255 Hz), 138.09 (d, J(C,F) = 9.5 Hz), 132.49 (d, J(C,F) = 2.2 Hz), 126.16 (d, J(C,F) = 13.3 Hz), 119.51 (d, J(C,F) = 25.1 Hz), 116.64, 60.71, 48.81, 13.92 ppm. HRMS (ESI): calculated for C₁₁H₁₀BrFO₃ [M], 287.9797; found, 287.9794. LCMS t_R = 0.989 min, ES-MS m/z = 289.0 [M + H]⁺.

Ethyl 6-Bromo-1-(4-fluorophenyl)-4-oxo-1,4-dihydroquinoline-3-carboxylate (15, Where Ar = 4-Fluorophenyl)

Compound 14 (2.87 g, 9.93 mmol, 1.00 equiv) and N,N-dimethylformamide dimethyl acetal (1.87 mL, 14.9 mmol, 1.50 equiv) were dissolved in DMF (33 mL) in a microwave vial and heated in a microwave reactor at 120 °C for 15 min. To this mixture was then added 4-fluoroaniline (1.41 mL, 14.9 mmol, 1.50 equiv), and the reaction was heated in a microwave reactor at 150 °C for 20 min. The reaction mixture was diluted with ethyl acetate and washed with water (2×). The aqueous layers were back-extracted with ethyl acetate, and the combined organics were dried (MgSO₄), filtered, and concentrated in vacuo. Purification by flash chromatography on silica gel afforded 3.79 g (98%) of the title compound as yellow solid. ¹H NMR (400 MHz, DMSO-d₆) δ = 8.47 (s, 1H), 8.31 (d, J = 2.4 Hz, 1H), 7.81 (dd, J = 9.1, 2.4 Hz, 1H), 7.77–7.73 (m, 2H), 7.54–7.50 (m, 2H), 6.92 (d, J = 9.0 Hz, 1H), 4.20 (q, J = 7.1 Hz, 2H), 1.25 ppm (t, J = 7.1 Hz, 3H). ¹³C NMR (100 MHz, DMSO-d₆): δ = 171.80, 163.90, 162.43 (d, J(C,F) = 247.5 Hz), 149.02, 139.70, 136.38 (d, J(C,F) = 2.8 Hz), 135.37, 130.17 (d, J(C,F) = 9.0 Hz), 128.83, 128.15, 120.73, 118.06, 117.26 (d, J(C,F) = 23.2 Hz), 110.88, 60.04, 14.21 ppm. HRMS (ESI): calculated for C₁₈H₁₃BrFNO₃ [M], 389.0063; found, 389.0062. LCMS t_R = 0.934 min, ES-MS m/z = 390.2 [M + H]⁺.

6-Bromo-1-(4-fluorophenyl)-4-oxo-1,4-dihydroquinoline-3-carboxamide (16, Where Ar = 4-Fluorophenyl)

Ethyl 6-bromo-1-(4-fluorophenyl)-4-oxo-1,4-dihydroquinoline-3-carboxylate (1.00 g, 2.56 mmol, 1.00 equiv) was suspended in 7 N ammonia in methanol (30 mL) in a microwave vial, and the reaction was heated in a microwave reactor at 150 °C for 60 min. The reaction was concentrated to afford 881 mg (95%) of the title compound as a brown solid that was used without further purification. ¹H NMR (400 MHz, DMSO-d₆) δ = 9.08 (d, J = 4.0 Hz, 1H), 8.57 (s, 1H), 8.43 (d, J = 2.4 Hz, 1H), 7.85 (dd, J = 9.1, 2.4 Hz, 1H), 7.77–7.73 (m, 2H), 7.66 (d, J = 4.1 Hz, 1H), 7.56–7.50 (m, 2H), 7.02 ppm (d, J = 9.1 Hz, 1H). ¹³C NMR (100 MHz, DMSO-d₆): δ = 174.64, 164.83, 162.46 (d, J(C,F) = 247.6 Hz), 148.42, 139.85, 136.50 (d, J(C,F) = 2.8 Hz), 135.66, 129.97 (d, J(C,F) = 9.3 Hz), 128.09, 128.07, 120.86, 118.19, 117.31 (d, J(C,F) = 23.3 Hz), 112.07 ppm. HRMS (ESI): calculated for C₁₆H₁₀BrFN₂O₂ [M], 359.9910; found, 359.9909. LCMS t_R = 0.929 min, ES-MS m/z = 361.2 [M + H]⁺.

1-(4-Fluorophenyl)-4-oxo-6-vinyl-1,4-dihydroquinoline-3-carboxamide (43, Where Ar = 4-Fluorophenyl)

To a solution of 6-bromo-1-(4-fluorophenyl)-4-oxo-1,4-dihydroquinoline-3-carboxamide (450 mg, 1.25 mmol, 1.0 equiv), triethylamine (174 μL, 1.25 mmol, 1.0 equiv), and Pd(dppf)Cl₂·CH₂Cl₂ (18.2 mg, 0.025 mmol, 0.2 equiv) in 1-propanol (8.3 mL) was added potassium vinyltrifluoroborate (200 mg, 1.5 mmol, 1.2 equiv). The mixture was purged with argon and stirred at 100 °C for 16 h. The reaction was filtered through Celite and washed very well with a 5% MeOH in DCM solution. The filtrate was concentrated in vacuo to give 385 mg (100%) of the title compound, which was used without further purification. ¹H NMR (400 MHz, DMSO-d₆): δ = 9.23 (d, J = 4.2 Hz, 1H), 8.55 (s, 1H), 8.36 (d, J = 1.6 Hz, 1H), 7.89 (dd, J = 1.8, 8.9 Hz, 1H), 7.79–7.75 (m, 2H), 7.64 (d, J = 4.2 Hz, 1H), 7.54 (t, J = 8.7 Hz, 2H), 7.03 (d, J = 8.8 Hz, 1H), 6.93 (q, J = 11, 6.6 Hz, 1H), 5.95 (d, J = 17.6 Hz, 1H), 5.39 ppm (d, J = 11 Hz, 1H). ¹³C NMR (100 MHz, DMSO-d₆): δ = 175.79, 165.17, 162.36 (d, J(C,F) = 247.0 Hz), 147.77, 140.24, 136.74 (d, J(C,F) = 3.2 Hz), 135.38, 134.17, 130.1, 129.93 (d, J(C,F) = 9.0 Hz), 126.74, 123.72, 118.65, 117.35 (d, J(C,F) = 23.4 Hz), 115.9, 111.72 ppm. HRMS (ESI): calculated for C₁₈H₁₃FN₂O₂ [M], 308.0961; found, 308.0964. LCMS t_R = 0.922 min, ES-MS m/z = 309.2 [M + H]⁺.

1-(4-Fluorophenyl)-6-formyl-4-oxo-1,4-dihydroquinoline-3-carboxamide (44, Where Ar = 4-Fluorophenyl)

To a solution of 1-(4-fluorophenyl)-4-oxo-6-vinyl-1,4-dihydroquinoline-3-carboxamide (385 mg, 1.25 mmol, 1.0 equiv) in 3:1 acetone/water (8 mL) was added N-oxide-4-methylmorpholine (220 mg, 1.87 mmol, 1.5 equiv) and osmium tetroxide (6.3 mg, 0.025 mmol, 0.02 equiv). After the reaction was stirred for 1 h, sodium periodate (294 mg, 1.37 mmol, 1.1 equiv) was added. After another 2 h, the reaction was diluted with EtOAc and washed well with a 10% NaS₂O₃ solution. The organic layer was dried (MgSO₄), filtered, and concentrated in vacuo to give 365 mg (94%) of the title compound that was used without further purification. ¹H NMR (400 MHz, DMSO-d₆): δ = 10.17 (s, 1H), 9.10 (d, J = 3.8 Hz, 1H), 8.93 (d, J = 1.8 Hz, 1H), 8.62 (s, 1H), 8.12 (dd, J = 1.8, 8.8 Hz, 1H), 7.82–7.78 (m, 2H), 7.74 (d, J = 3.8 Hz, 1H), 7.56 (t, J = 8.8 Hz, 2H), 7.22 ppm (d, J = 8.8 Hz, 1H). ¹³C NMR (100 MHz, DMSO-d₆): δ = 192.23, 175.88, 164.68, 162.49 (d, J(C,F) = 247 Hz), 149.06, 144.23, 136.61 (d, J(C,F) = 3.2 Hz), 132.48, 130.98, 130.41, 130.0 (d, J(C,F) = 9.3 Hz), 126.54, 119.46, 117.37 (d, J(C,F) = 23.3 Hz), 112.78 ppm. HRMS (ESI) calculated for C₁₇H₁₁FN₂O₃ [M], 310.0754; found, 310.0757. LCMS t_R = 0.732 min, ES-MS m/z = 311.2 [M + H]⁺.

6-((cis-2,6-Dimethylmorpholino)methyl)-1-(4-fluorophenyl)-4-oxo-1,4-dihydroquinoline-3-carboxamide (58)

A solution of 1-(4-fluorophenyl)-6-formyl-4-oxo-1,4-dihydroquinoline-3-carboxamide (580 mg, 1.87 mmol, 1.0 equiv) in dichloromethane (1 mL), cis-2,6-dimethylmorpholine (461 μL, 3.74 mmol, 2.0 equiv), and acetic acid (268 μL, 4.67 mmol, 2.5 equiv) was stirred for 1 h. Sodium triacetoxyborohydride (594 mg, 2.80 mmol, 1.5 equiv) was added. After 16 h, the reaction was concentrated to dryness. Purification by reverse phase HPLC afforded 620 mg (81%) of the title compound as a white solid. ¹H NMR (400 MHz, DMSO-d₆) δ 9.23 (d, J = 4.4 Hz, 1H), 8.55 (s, 1H), 8.26 (d, J = 1.4 Hz, 1H), 7.77–7.73 (m, 2H), 7.65 (dd, J = 1.8, 8.7 Hz, 1H), 7.59 (d, J = 4.3 Hz, 1H), 7.52 (t, J = 8.7, 2H), 7.03 (d, J = 8.7, 1H), 3.57–3.52 (m, 4H), 2.65 (d, J = 10.7 Hz, 2H), 1.68 (t, J = 10.7 Hz, 2H), 1.01 ppm (d, J = 6.2 Hz, 6H). ¹³C NMR (100 MHz, DMSO-d₆): δ = 175.8, 165.26, 162.36 (d, J(C,F) = 247 Hz), 147.78, 139.9, 136.82 (d, J(C,F) = 3.1 Hz), 135.26, 133.91, 129.98 (d, J(C,F) = 8.9 Hz), 126.41, 125.85, 118.26, 117.22 (d, J(C,F) = 23.0 Hz), 111.59, 70.97, 61.23, 58.81, 18.96 ppm. HRMS (ESI) calculated for C₂₃H₂₄FN₃O₃ [M], 409.1802; found, 409.1804. LCMS t_R = 0.644 min, ES-MS m/z = 410.3 [M + H]⁺.

ABBREVIATIONS USED

7TM

seven transmembrane

Ac

acetate

BBB

blood−brain barrier

Bu

butyl

CL

clearance

CNS

central nervous system

CRC

concentration−response curve

dba

dibenzylideneacetone

DCE

1,2-dichloroethane

DIEA

N,N-diisopropylethylamine

DME

1,2-dimethoxyethane

DMF

N,N-dimethylformamide

DMPK

drug metabolism and pharmacokinetics

DMSO

dimethylsulfoxide

dppf

1,1′-bis(diphenylphosphino)-ferrocene

D^tBAD

di-tert-butyl azodicarboxylate

Et

ethyl

F_u

fraction unbound

GPCR

G-protein-coupled receptor

ip

intraperitoneal

iv

intravenous

K_p

brain to plasma ratio

K_p,uu

unbound brain to unbound plasma ratio

LLE

ligand-lipophilicity efficiency

M₁

muscarinic acetylcholine receptor subtype 1

max

maximum

MDCK

Madin-Darby canine kidney

MDD

major depressive disorder

Me

methyl

mGlu

metabotropic glutamate receptor

NAM

negative allosteric modulator

NBS

N-bromosuccinimide

NMO

N-methylmorpholine N-oxide

OCD

obsessive-compulsive disorder

PAM

positive allosteric modulator

PEG

polyethylene glycol

Ph

phenyl

P-gp

P-glycoprotein

PK

pharmacokinetics

pTSA

p-toluenesulfonic acid

RLM

rat liver microsome

SAR

structure-activity relationship

t-BuXphos

2-di-tert-butylphosphino-2′,4′,6′-triisopropylbiphenyl

TBAF

tetrabutylammonium fluoride

THF

tetrahydrofuran

TRD

treatment-resistant depression

T

time

t_1/2

half-life

V_SS

volume of distribution at steady-state

Xantphos

4,5-bis(diphenylphosphino)-9,9-dimethylxanthene