Solution-Phase Parallel Synthesis and SAR of Homopiperazinyl Analogs as Positive Allosteric Modulators of mGlu₄

Abstract

Using a functional high-throughput screening (HTS) and subsequent solution-phase parallel synthesis approach, we have discovered a novel series of positive allosteric modulators for mGlu₄, a G-protein coupled receptor. This series is comprised of a homopiperazine central core. The solution-phase parallel synthesis and SAR of analogs derived from this series will be presented. This series of positive allosteric modulators of mGlu₄ provide critical research tools to further probe the mGlu₄-mediated effects in Parkinson’s disease.

1. Introduction

The development of new strategies for the treatment of Parkinson’s disease (PD) continues to be a major focus of attention.^1–4 PD is a neurodegenerative disease caused by the degeneration of dopaminergic neurons in the substantia nigra pars compacta of the basal ganglia which leads to a debilitating movement disorder. Metabotropic glutamate receptor 4 (mGlu₄) is expressed at a key synapse in the indirect pathway of the basal ganglia circuitry.^{5, 6} Several studies have shown that the activation of mGlu₄ has anti-parkinsonian and neuroprotective effects in rodent PD models.^{6, 7} These results have established this receptor as a viable target for the symptomatic and disease modifying treatment of PD. We have taken the approach of selectively activating mGlu₄ using positive allosteric modulators (PAMs), compounds which increase the potency of the endogenous neurotransmitter glutamate at mGlu₄.^{8, 9}

Compounds in Vanderbilt’s small molecule library were screened in a triple-add assay format at 10 µM to identify agonist, antagonist, or allosteric modulator activity in CHO cells stably expressing human mGlu₄ and the chimeric G protein Gqi₅. Test compounds were added to cells 140 seconds prior to addition of an EC₂₀ concentration of glutamate; 90 seconds later an EC₈₀ concentration of glutamate was added. Response was measured in a fluorometric calcium assay, and data were analyzed by finding the maximum value for each fluorescence trace. This “triple add” approach to HTS is utilized in order to identify compounds with different modes of pharmacology (agonists prior to EC₂₀ addition, antagonists and negative allosteric modulators after EC₈₀ addition, and positive allosteric modulators after EC₂₀ addition) in a single experiment.

Using this functional high-throughput screening protocol, we have previously identified a number of novel mGlu₄ PAM scaffolds. In addition to known mGlu₄ PAM, PHCCC, 1,^{7, 10} we have reported a number of distinct structural classes of mGlu₄ PAMs, such as VU0155041, 2,^{11, 12} VU0080241, 3,¹³ VU0001171, 4,¹⁴ VU0092145, 5,¹⁴ VU0361737, 6,¹⁵ VU0364439, 7,¹⁶ and VU0366037, 8 (Figure 1); of which, two analogs (2 and 6) have been shown to display efficacy in anti-Parkinsonian animal models.^{11, 15} However, all of the reported compounds bear striking resemblance in that they all contain polyaromatic moieties, making these compounds planar, and they do not contain any basic nitrogens. We now report a novel scaffold, represented by VU0105737, 9, as possessing mGlu₄ PAM activity (Figure 2). This new chemotype is comprised of a nonplanar homopiperazine central core which contains a basic nitrogen (which can presumably lead to more soluble compounds than previously identified). Due to this structural novelty and the encouraging calculated properties, this scaffold provided an excellent opportunity to explore the SAR around this initial HTS lead as part of our hit-to-lead program.

Figure 1.

Structures of PHCCC, 1, and other recently reported mGlu₄ PAMs, 2–9.

Figure 2.

Areas of SAR Exploration of VU0105737, 9.

2. Results and Discussion

The SAR surrounding compound 9 centered around 4 structural motifs: 1, exploring the 2,4-dimethoxyphenylamide; 2, exploring the linker chain length; 3, exploring the central homopiperazine ring; and 4, exploring the 4-methoxyphenylsulfonamide (Figure 2). The homopiperazine, 9, was readily modified using commercially available acid chlorides, carboxylic acids, sulfonyl chlorides and other core starting materials. The parallel synthesis and biological evaluation of this structure class is detailed below. All parallel synthetic procedures were conducted utilizing either test-tube reaction blocks or stir-plate mounted vial racks.

Modification of the central homopiperazine core was performed in order to address the basicity of the homopiperazine nitrogens (piperazine, oxohomopiperazine, oxopiperazine) as well as to change the shape by introduction of the bridged [3.3.0] scaffold (Scheme 1. All experimental procedures are contained in the Supplemental Materials). The piperazine compound, 11, and [3.3.0]-compound, 13, were prepared from N-Boc starting materials, 10 and 12, via a three step protocol of sulfonylation, Boc removal and alkylation. The oxo-containing compounds started from piperazin-2-one (14a) or 1,4-diazepan-5-one (14b). Thus, sulfonamide formation with 4-methoxybenzenesulfonyl chloride (CH₂Cl₂, DIEA, 43%-quant.), followed by alkylation of the amide with methyl bromoacetate (Cs₂CO₃, CH₃CN, 80°C, 45–50%) and saponification of the ester (LiOH, THF/MeOH/H₂O, 35%-quant.) afforded the penultimate acid. Finally, HATU-mediated amide coupling with 2,4-dimethoxyaniline (DMF, DIEA) provided the final compounds, 15a and 15b.

Scheme 1.

Results from Table 1 show that little modification of the core portion of the molecule is tolerated. All modifications of the core structure led to compounds that were inactive and showed no potentiation of glutamate (as assessed by the %GluMax). We utilized the Chinese hamster ovary cells expressing human mGlu₄ and the chimeric G protein Gq_i5 to induce calcium mobilization as our pharmacological assay to determine mGlu₄ potency. This assay does show fluctuation in the day-to-day maximal PAM response; due to this, all data have been normalized to the control compound, 1, as a comparison of relative efficacy (as noted in % PHCCC).^{10, 16}

Table 1.

SAR Evaluation of Linker and Scaffold

Cmpd	R	hEC50 (µM)a	% PHCCCa	Yieldc (%)
11		Inactiveb	20.8 ± 1.1	45
13		Inactiveb	25.3 ± 2.3	49
15a		Inactiveb	18.1 ± 1.3	19
15b		Inactiveb 3.1 ± 0.3	16.4 ± 0.8	60

^aEC₅₀ and GluMax, are the average of at least three independent determinations performed in triplicate (Mean ± SEM shown in table). PHCCC is run as a control compound each day and it, and the maximal response generated in mGlu₄ CHO cells in the presence of mGlu₄ PAMs varies slightly in each experiment. Therefore, efficacy data were further normalized to the relative PHCCC response obtained in each day’s run.

^bInactive compounds are defined as %GluMax did not surpass 2X the EC₂₀ value for that day’s run.

^cAll yields were obtained by reverse phase preparative HPLC and were optimized for purity (>95%) not yield.

Keeping the right-side of the molecule constant, investigation of replacement of the amide aromatic ring with various heterocycles and alkyl groups, and exploration of the linker was next undertaken (Scheme 2, see Supplemental Material for full experimental procedures). To this end, commercially available N-Boc homopiperazine, 16, was reacted with 4-methoxybenzenesulfonyl chloride (CH₂Cl₂, DIEA) to give 17 in high yield (98%). Deprotection of the Boc group (4 M HCl, dioxane, quant.) furnished the key intermediate 18 which was converted to the urea, 20i, via reaction with 2,4-dimethoxyphenyl isocyanate (CH₂Cl₂, DIEA) in good yield (80%). Alternatively, 18 was N-alkylated with methyl bromoacetate (Cs₂CO₃, CH₃CN, 80 °C, 76%) followed by saponification of the ester (LiOH, THF/H₂O/MeOH, quant.) which gave the penultimate intermediate, 19, in good overall yield. The carboxylic acid was then coupled to a variety of anilines and amines utilizing HATU as the coupling reagent (HATU, DMF, DIEA, 45–85%) to provide the desired compounds (20a–j).

Scheme 2.

Synthetic procedure for the initial homopiperazine analogs, 20a–j.

The results in Table 2 show that the original hit with the 2,4-dimethoxyphenyl substituent was weakly active (9, >10 µM, 69.5% PHCCC) as was the 2-methoxyphenyl substituent (20a, >10 µM, 63.9% PHCCC); however, both compounds showed good efficacy (>100% PHCCC). Replacing the 2-methoxyphenyl with 4-methoxyphenyl led to an inactive compound (20b, 14.8% PHCCC). Other substitution patterns, such as 2,4-difluorophenyl (20c), 2-fluorophenyl (20d), 4-pyridyl (20e), cycloalkyl (20f, 20g) and cyclicdioxoaryl compound (20h) were all inactive. In addition, changing the length of the carbon linker also produced inactive compounds (20i, 20j). This rather ‘shallow’ SAR is a common occurrence with allosteric modulators.^4,5

Table 2.

SAR Evaluation of Amide Aromatic Ring

Cmpd	n	R	hEC50 (µM)a	% PHCCCa	Yieldc (%)
9	1	2,4-dimethoxyphenyl	>10	69.5 ± 7.7	61
20a	1	2-methoxyphenyl	>10	63.9 ± 3.6	64
20b	1	4-methoxyphenyl	Inactiveb	14.8 ± 0.8	20d
20c	1	2,4-difluorophenyl	Inactiveb	17.7 ± 1.6	3d
20d	1	2-fluorophenyl	Inactiveb	13.2 ± 0.4	2d
20e	1	4-pyridyl	Inactiveb	13.9 ± 1.2	2d
20f	1	Cyclohexyl	Inactiveb	15.5 ± 1.3	12d
20g	1	Isopropyl	Inactiveb	15.0 ± 0.8	19d
20h	1	2,3-dihydrobenzo[b][1,4]dioxin-6-yl	Inactiveb	17.6 ± 1.0	12d
20i	0	2,4-dimethoxyphenyl	Inactiveb	13.5 ± 0.7	80
20j	3	2,4-dimethoxyphenyl (±)-PHCCC	Inactiveb 3.1 ± 0.3	14.8 ± 1.3	70

^aEC₅₀ and GluMax, are the average of at least three independent determinations performed in triplicate (Mean ± SEM shown in table). PHCCC is run as a control compound each day, and the maximal response generated in mGlu₄ CHO cells in the presence of mGlu₄ PAMs varies slightly in each experiment. Therefore, efficacy data were further normalized to the relative PHCCC response obtained in each day’s run.

^bInactive compounds are defined as %GluMax did not surpass 2X the EC₂₀ value for that day’s run.

^cAll yields were obtained by reverse phase preparative HPLC unless otherwise stated and were optimized for purity (>95%) not yield.

^dYields obtained by mass directed HPLC¹⁷

The last portion of the molecule for modification was the right-hand sulfonamide moiety (Scheme 3, Table 3, see Supplemental Material for full experimental details). Starting with 2,4-dimethoxyaniline, 21, amide formation with bromoacetyl bromide (CH₂Cl₂, DIEA, 81%) gave the desired α-bromo-amide, 22, which was N- alkylated with Boc-homopiperazine (Cs₂CO₃, CH₃CN, 80 °C, 79%) to give 23 in 64% yield (2 steps). The Boc protecting group was removed (4 M HCl, dioxane, quant.) giving the key homopiperazine intermediate, 24, which was coupled with the appropriate sulfonyl chloride (CH₂Cl₂, DIEA, 45–85%) to furnish the desired sulfonamides 25a–al.

Scheme 3.

Table 3.

SAR Evaluation of Sulfonamide Aromatic Ring

Cmpd	R	hEC50 (µM)	% PHCCC	Yieldc (%)
25a	phenyl	3.5 ± 0.7	69.7 ± 5.1	80
25b	2-fluorophenyl	3.5 ± 0.5	36.6 ± 1.8	3d
25c	2-chlorophenyl	>10	25.8 ± 2.5	58
25d	2-methoxyphenyl	Inactiveb	21.8 ± 1.0	87
25e	2-(trifluoromethyl)phenyl	Inactiveb	20.1 ± 0.2	24d
25f	3-(trifluoromethyl)phenyl	Inactiveb	18.2 ± 1.9	6d
25g	4-(trifluoromethyl)phenyl	2.5 ± 0.5	26.6 ± 2.3	36
25h	4-(trifluoromethoxy)phenyl	>10	48.2 ± 4.6	5d
25i	4-methylphenyl	3.3 ± 0.2	48.7 ± 1.1	77
25j	4-fluorophenyl	>10	27.6 ± 1.3	33d
25k	4-chlorophenyl	4.1 ± 0.6	43.6 ± 4.4	41
25l	4-tertbutylphenyl	>10	32.6 ± 3.1	32
25m	4-acetylphenyl	>10	42.9 ± 2.7	65
25n	2,4-dimethylphenyl	1.3 ± 0.5	42.8 ± 4.1	58
25o	2,5-dimethylphenyl	2.7 ± 0.6	32.5 ± 1.2	67
25p	2,5-dichlorophenyl	Inactiveb	22.7 ± 0.7	67
25q	2-chloro-6-methylphenyl	Inactiveb	18.1 ± 0.9	75
25r	2,6-dichlorophenyl	Inactiveb	22.3 ± 1.7	23
25s	3,4-dimethylphenyl	3.2 ± 0.2	42.4 ± 5.0	67
25t	3-chloro-4-methylphenyl	3.4 ± 0.2	39.2 ± 1.2	43
25u	3,4-dichlorophenyl	>10	24.3 ± 1.5	31
25v	3,4-difluorophenyl	>10	27.0 ± 1.8	64
25w	4-chloro-3-fluorophenyl	4.5 ± 0.7	25.5 ± 1.2	53
25x	2,4-dichloro-5-methylphenyl	3.1 ± 0.9	43.6 ± 4.7	44
25y	4-chloro-2-fluoro-5-methylphenyl	2.7 ± 0.2	36.9 ± 2.7	36
25z	benzyl	>10	46.5 ± 4.8	12d
25aa	2,4-dichlorobenzyl	2.8 ± 0.01	34.4 ± 3.5	44
25ab	isopropyl	Inctiveb	17.5 ± 0.7	22d
25ac	isobutyl	Inactiveb	21.3 ± 0.9	27d
25ad	2-pyridyl	Inactiveb	20.6 ± 0.3	36
25ae	2-thiophene	1.8 ± 0.3	52.9 ± 5.0	83
25af	2-furyl	3.3 ± 0.3	36.2 ± 2.2	98
25ag	3-pyridyl	>10	25.6 ± 4.6	20
25ah	3-furyl	>10	34.1 ± 3.6	70
25ai	3-thiophene	>10	35.0 ± 3.3	79
25aj	2-acetamidothiazol-5-yl	>10	46.4 ± 5.5	60
25ak	2,4-dimethylthiazol-5-yl	>10	47.7 ± 2.1	73
25al	2-methyl-5-(trifluoromethyl)thiazol-5-yl (±)-PHCCC	2.5 ± 0.4 3.1 ± 0.3	28.1 ± 4.8	49

^aEC₅₀ and GluMax, are the average of at least three independent determinations performed in triplicate (Mean ± SEM shown in table). PHCCC is run as a control compound each day, and the maximal response generated in mGlu₄ CHO cells in the presence of mGlu₄ PAMs varies slightly in each experiment. Therefore, data were further normalized to the relative PHCCC response obtained in each day’s run.

^bInactive compounds are defined as %GluMax did not surpass 2X the EC₂₀ value for that day’s run.

^cAll yields were obtained by reverse phase preparative HPLC unless otherwise stated and were optimized for purity (>95%) not yield.

^dYields obtained by mass directed HPLC¹⁷

The compounds evaluated in Table 3 cover a range of phenyl, benzyl, alkyl and heteroaryl groups. The phenyl group showed good activity and efficacy (25a, 3.5 µM, 69.7% PHCCC). However, substitution at the 2- position only tolerated F (25b, 3.5 µM, 36.6% PHCCC) and Cl (25c, >10 µM, 25.8% PHCCC) with the latter being only a weak potentiator. The 4-position was the most tolerated with 4-CF₃-phenyl (25g, 2.1 µM, 26.6% PHCCC) and 4-Me-phenyl (25i, 3.3 µM, 48.7% PHCCC) being the most potent. Disubstituted phenyl groups in general showed good activity especially with dimethylated phenyl groups where the 2,4-dimethylphenyl was the most potent compound of this series (25n, 1.3 µM, 42.8% PHCCC) along with 2,5-dimethyl phenyl (25o, 2.7 µM, 32.5% PHCCC). 2,4,5-Trisubstituted phenyl groups also showed potencies as good or better than the original HTS hit (25x, 3.1 µM, 43.6% PHCCC and 25y, 2.7 µM, 36.9% PHCCC).

The limited SAR around the benzyl group showed contrasting activity compared to the phenyl group. For example, the benzyl group was a weak potentiator (25z, >10 µM, 46.5% PHCCC). Furthermore the 2,4-dichlorobenzyl group showed good potency (25aa, 2.8 µM, 34.4% PHCCC) which in the case of the dihalogenated phenyl groups were all inactive. Compounds 25ab and 25ac showed that alkyl groups were not tolerated. Substitution with heteroaryl groups led to 3 compounds with acceptable potency. The most potent was the 2-thienyl (25ae, 1.8 µM, 52.9% PHCCC) followed by the 2-methyl-4-trifluoromethylthiazol-5-yl (25al, 2.5 µM, 28.1% PHCCC) and the 2-furyl (25af, 3.3 µM, 36.2% PHCCC).

Compounds 9, 20a, 25a, 25i, 25ae were selected based on potency and efficacy for further in vitro pharmacokinetic (PK) studies which included CYP450 inhibition, metabolic stability (Cl_INT)¹⁸ and rat plasma protein binding (rPPB) (Table 4, see Supplemental Material for assay details). Although these compounds possessed favorable free fraction when tested for plasma protein binding in rats (>3% free), this class of compounds was very unstable in liver microsomes (Cl_INT) and were equipotent with inhibiting CYP activity (2C9 and 3A4 < 10 µM).

Table 4.

In Vitro Pharmacokinetic Data for Selected Compounds.

Cmpd	CYP				rPPB, %fua	Human CLINT (CLHEP)b
	1A2	2C9	2D6	3A4
9	>30 µM	1.8 µM	>20 µM	4.2 µM	3.9	908.58 (20.53)
20a	>30 µM	2.0 µM	>20 µM	4.6 µM	3.5	893.34 (20.52)
25a	>20 µM	2.9 µM	>20 µM	7.0 µM	11.5	559.86 (20.24)
25i	16.29 µM	2.3 µM	13.9 µM	3.0 µM	3.1	731.45 (20.41)
25ae	8.46 µM	2.6 µM	10.1 µM	7.4 µM	27.3	517.92 (20.18)

^arat protein binding, % free unbound.

^bIntrinsic clearance, human (CL_INT). Predicted hepatic clearance (CL_HEP).

Conclusions

In summary, through a functional HTS campaign, we identified a novel chemotype as an mGlu₄ PAM and SAR was explored around 4-structural motifs. These compounds represent a series of homopiperazine sulfonamides. Limited SAR around this scaffold suggests a possible alternate allosteric binding site compared to previously disclosed mGlu₄ PAMs where more robust SAR was identified. Unfortunately, these compounds possess less than ideal PK properties (poor metabolic stability, CYP 2C9 and CYP 2D6 inhibition) preventing their further study. However, as this class of compounds represent a novel chemotype in the mGlu₄ PAM area, we anticipate these compounds will inform the community for future scaffold design.