Discovery and SAR of a novel series of potent, CNS penetrant M₄ PAMs based on a non-enolizable ketone core: Challenges in disposition

Abstract

This letter describes the chemical optimization of a novel series of M₄ PAMs based on a non-enolizable ketone core, identified from an MLPCN functional high-throughput screen. The HTS hit was potent, selective and CNS penetrant; however, the compound was highly cleared in vitro and in vivo. SAR provided analogs for which M₄ PAM potency and CNS exposure were maintained; yet, clearance remained high. Metabolite identification studies demonstrated that this series was subject to rapid, and near quantitative, reductive metabolism to the corresponding secondary alcohol metabolite that was devoid of M₄ PAM activity.

Graphical abstract

Selective activation of the M₄ muscarinic acetylcholine receptor has emerged as a very promising, and mechanistically distinct, therapeutic approach for the treatment of numerous neuropsychiatric and rare genetic CNS disorders.^1-8 Until recently,⁹ the M₄ PAMs reported to date have been primarily within a single amide-bearing chemotype (Figure 1),^10-19 wherein non-obvious, subtle structural modifications engender steep SAR, species differences in PAM potency as well as affinity/cooperativity and subtype selectivity, solubility, and/or P-gp efflux.^10-21 As such, all of the known chemotypes of M₄ PAMs (1-6) pose one or more major challenges en route to a clinical devleopment candidate. In this Letter, we detail the identification, SAR and DMPK profile of a new, non-enolizable ketone chemotype, related to 1, 2 and 4, identified in a high-throughput screening (HTS) campaign.²²

Figure 1.

Structures of reported M₄ PAMs 1-6.

Under the auspice of the MLPCN, we performed a functional M₄ HTS in 1536-well assay format against the ∼360,000 compound MLSMR collection, and four structurally novel M₄ PAMs were identified.²² Of the hits, VU0009153 (7, CID:682264) represented an interesting lead (Figure 2), as it was similarly potent on both human and rat M₄ (EC₅₀s of 120 nM (76% ACh Max) and 76 nM (71% ACh Max), respectively (using DMSO stock from the source plates). In addition to no apparent species differences, 7 was was also inactive on both human M₁ and M₅ (built-in HTS counterscreen). Upon re-synthesis of 7, M₄ PAM potency remained intact (hM₄ EC₅₀ = 138 nM, pEC₅₀ = 6.86±0.07, ACh Max = 76.8±4.1 and rM₄, EC₅₀ = 145 nM, pEC₅₀ = 6.84±0.10, ACh Max = 80.5±6.3). Moreover, it was a des-NH variant of the prototypical M₄ PAM chemotypes 1, 2, and 4, replacing the amide linker with a non-enolizable ketone. PAM 7 displayed very high in vitro intrinsic clearance (CL_int) in both rat and human hepatic microsomes with predicted hepatic clearance (CL_hep) near hepatic blood flow rates (Q_hep), but was this due to the trimethyl moiety, or an inherent feature of the ketone-linker? Figure 2 also highlights the three regions of 7 targeted for SAR exploration with focus on alternate aryl/heteroaryl moieties, evaluating our preferred aza systems of 2 and 4, while also exploring ketone bioisosteres.

Figure 2.

Structure and key pharmacology and in vitro DMPK data for the M₄ PAM HTS hit VU0009153 (7), and three areas to address in the lead optimization campaign. Potency, efficacy, and selectivity was determined via calcium mobilization assays with hM_4/Gqi5-CHO or rM_4/Gqi5-CHO cells performed in the presence of an EC₂₀ fixed concentration of acetylcholine; values represent means from three (n=3) independent experiments performed in triplicate. Plasma protein binding and hepatic microsomal intrinsic clearance values represent data from a single (n=1) experiment performed in triplicate.

The chemistry to produce analogs of 7 was straightforward (Scheme 1), and employed advanced mercaptonitrile intermediates 8-10 previously described^3,15-19 in a one-step protocol. Here, intermediates 8-10 are treated with a diverse array of commercial bromomethylketones 11 under basic conditions to afford, via a one-pot alklyation/cyclization sequence analogs 12-14, respectively in yields ranging from 55-90%.

Scheme 1.

Reagents and conditions: a) KOH, EtOH, rt, 6 hr, 55-90%.

Table 1 highlights select SAR for the first generation of analogs 12 surveying diverse aryl/heteroaryl moeities. The direct analog of 7, 12a, lost ∼7- to 10-fold potency at both human and rat M₄, whereas the direct ketone congener of 2 (12c) was essentially equipotent at both human and rat M₄, but with diminished efficacy (∼50% ACh Max). The 2,4-difluorophenyl analog 12e and the 4-pyridyl congener 12f displayed good activity at human M₄, but exhibited a significant (4- to 7-fold) rightward shift in potency at rat M₄. In general, ketone-linked subseries 12 possessed weaker efficacy than the corresponding amide-linked derivatives. Relative to 7, these findings suggest substitution in the 5-position may be required for M₄ PAM potency.

Table 1.

Structures and activities for M₄ PAM analogs 12.

Cpd	Ar (Het)	hM4 EC50 (nM)a [% ACh Max ±SEM]	hM4 pEC50 (±SEM)	rM4 EC50 (nM)a [% ACh Max ±SEM]	rM4 pEC50 (±SEM)
12a	Ph	891 [47±6%]	6.05±0.12	1,500 [46±4%]	5.82±0.03
12b	4-OHPh	355 [70±6%]	6.45±0.04	380 [73±4%]	6.42±0.02
12c	4-OMePh	479 [48±6%]	6.32±0.11	490 [54±4%]	6.31±0.05
12d	3-Cl,4-FPh	240 [45±9%]	6.62±0.11	195 [41±2%]	6.71±0.03
12e	2,4-diFPh	562 [39±3%]	6.25±0.28	3,980 [35±7%]	5.40±0.31
12f	4-pyridyl	589 [71±5%]	6.23±0.12	2,239 [64±6%]	5.65±0.22

^aCalcium mobilization assays with hM_4/Gqi5-CHO or rM_4/Gqi5-CHO cells performed in the presence of an EC₂₀ fixed concentration of acetylcholine; values represent means from three (n=3) independent experiments performed in triplicate.

While in a very limited number of cases (Table 2), substitution at the 5-position with a chlorine atom afforded active M₄ PAMs with good efficacy as in 13a-b, solubility with other analogs proved poor (driven by cLogPs >5) and hindered evaluation. Thus, our attention focused on the 5,6-dimethyl pyridazine core of 4 and ketone analogs 14.

Table 2.

Structures and activities for M₄ PAM analogs 13.

Cpd	Ar (Het)	hM4 EC50 (nM)a [% ACh Max ±SEM]	hM4 pEC50 (±SEM)	rM4 EC50 (nM)a [% ACh Max ±SEM]	rM4 pEC50 (±SEM)
13a	Ph	372 [45±5%]	6.43±0.13	692 [43±4%]	6.16±0.12
13b	3-pyridyl	105 [74±5%]	6.98±0.16	417 [95±7%]	6.38±0.06

^aCalcium mobilization assays with hM_4/Gqi5-CHO or rM_4/Gqi5-CHO cells performed in the presence of an EC₂₀ fixed concentration of acetylcholine; values represent means from three (n=3) independent experiments performed in triplicate.

As shown in Table 3, analogs 14 restored M₄ PAM potency (hM₄, EC₅₀s 102 nM to 891 nM) and efficacy (ACh Max 55-76%), akin to 7; however, unpredictable species differences emerged (cf, 14a versus 14d-e) with a trend towards diminished potency at rat M₄. Encouragingly, analogs within these three subseries retained selectivity for M₄ versus M_1-.3,5 (data not shown).

Table 3.

Structures and activities for M₄ PAM analogs 14.

Cpd	Ar (Het)	hM4 EC50 (nM)a [% ACh Max ±SEM]	hM4 pEC50 (±SEM)	rM4 EC50 (nM)a [% ACh Max ±SEM]	rM4 pEC50 (±SEM)
14a	Ph	102 [76±3%]	6.99±0.04	166 [67±8%]	6.78±0.04
14b	4-OMePh	646 [63±6%]	6.19±0.21	1,450 [70±7%]	5.84±0.20
14c	2,4-diFPh	331 [55±5%]	6.48±0.28	380 [66±4%]	6.42±0.05
14d	3-pyridyl	105 [74±5%]	6.98±0.16	417 [95±7%]	6.38±0.06
14e	4-pyridyl	891 [76±4%]	6.05±0.05	3,550 [68±3%]	5.45±0.07

^aCalcium mobilization assays with hM_4/Gqi5-CHO or rM_4/Gqi5-CHO cells performed in the presence of an EC₂₀ fixed concentration of acetylcholine; values represent means from three (n=3) independent experiments performed in triplicate.

From our ongoing M₄ PAM program, we had additional mercaptonitrile intermediates beyond 8-10, so we elected to convert them into ketone analogs 15-16 and assess their profiles (Figure 3). Gratifyingly, significant diversity was tolerated, providing, in the case of 15, our most potent M₄ PAM in the ketone-linked series (hM₄ EC₅₀ =37 nM, pEC₅₀ = 7.12±0.09, ACh Max = 63.8±5.9 and rM₄ EC₅₀ = 182 nM, pEC₅₀ = 6.74±0.05, ACh Max = 52.7±5.4).

Figure 3.

More highly functionalized M₄ PAMs 15-17 in the ketone series. Potencies were determined via calcium mobilization assays with hM_4/Gqi5-CHO or rM_4/Gqi5-CHO cells performed in the presence of an EC₂₀ fixed concentration of acetylcholine; values represent means from three (n=3) independent experiments performed in triplicate.

Initial in vitro and in vivo DMPK evaluation of analogs 12-17 showed a range of low to moderate fraction unbound in rat and human plasma (f_us from 0.005 to 0.180) coupled with high CNS distribution (K_ps of 0.83 to 6.95 and K_p,uus of 0.48 to 2.12, but low absolute concentrations in plasma and brain).² In addition, analogs 12-17 showed generally clean CYP₄₅₀ inhibition profiles with the exception of often potent CYP1A2 inhibition (data not shown). However, 12-17 all shared one common property, predicted in vitro clearance in hepatic microsomes near hepatic blood flow rates (hCL_hep ≥ 20 mL/min/kg and rCL_hep ≥ 68 mL/min/kg) despite their structural diversity. The only commonality amongst them was the non-enolizable ketone linker. In vitro metabolite identification (Met ID) studies performed with 12f (as a representative of the series) in rat and human cryopreserved hepatocytes demonstrated a rapid, and near quantitative, reductive metabolism route affording the generic secondary alcohol 18 as the single major metabolite (Figure 4). We synthesized these known and putative metabolites by simple reduction of the ketones, and found that unfortunately, analogs 18 were devoid of M₄ PAM activity (human and rat EC₅₀s > 10 μM), i.e, a primary pharmacology inactivation metabolism pathway. Subsequent rat IV PK studies using a subset of representative compounds demonstrated that these PAMs were also highly cleared in vivo (CL_ps >70 mL/min/kg) indicating a good in vitro:in vivo correlation (IVIVC). Interestingly, even the secondary alcohols 18 had high predicted hepatic clearance (CL_hep >15 and >55 mL/min/kg for human and rat, respectively).

Figure 4.

Metabolite identification studies in rat and human cryopreserved hepatocytes with 12f suggest 12-17 undergo rapid and near quantitative reductive metabolism to the corresponding inactive secondary alcohols, 18. Predicted hepatic clearance values (from microsomal CL_int assays) represent data from a single (n=1) experiment performed in triplicate.

Based on these data, we decided to pursue two divergent courses of action: 1) survey enolizable ketone congeners, and 2) survey alternatives for the ketone functional group. Figure 5 highlights two representative examples, 19 and 20, of the first approach employing a phenethyketone derivative. Despite the significant structural change, 19 (hM₄ EC₅₀ = 479 nM, pEC₅₀ =6.32±0.12, ACh Max = 59.1±4.8 and rM₄ EC₅₀ = 540 nM, pEC₅₀ = 6.27±0.02, ACh Max = 67.9±5.9) and 20 (hM₄, EC₅₀ = 380 nM, pEC₅₀ = 6.42±0.11, ACh Max = 69.1±4.3 and rM₄ EC₅₀ = 254 nM, pEC₅₀ = 6.60±0.10, ACh Max = 77.2±4.4) were potent M₄ PAMs. Once again high brain distribution was observed (K_ps ∼11, with very low absolute concentrations)²³, but both compounds were highly cleared in vitro (predicted CL_hep =18 and 62 mL/min/kg for human and rat, respectively). Thus, this approach was terminated in favor of exploring alternatives for the ketone moiety.

Figure 5.

Phenethylketone-based M₄ PAMs 19 and 20. Potencies were determined via calcium mobilization assays with hM4/Gqi5-CHO or rM4/Gqi5-CHO cells performed in the presence of an EC20 fixed concentration of acetylcholine; values represent means from three (n=3) independent experiments performed in triplicate.

A number of substitutions for the ketone moiety were prepared and assessed as M₄ PAMs. In short, we synthesized and assessed mono-fluoro, gem-difluro, ether and amino (1° and 2°) derivatives; however, all were devoid of M₄ PAM activity (EC₅₀s > 10 μM). All attempts to make the 2- and 3-spirooxetane ketone bioisosteres proved unsuccessful, as the 3-amino moiety of the azathiophene bicycle precluded successful chemistry. By this time, it was clear that the ketone series was not tractable for the development of a clinical candidate, despite similar non-enolizable ketone moieties found in commercial drugs wherein ketone reduction is not a major route of metabolism.^24,25 In the present system, the presence of the 3-amino moiety, essential for M₄ PAM activity, likely serves as internal H-bond donor to activate the carbonyl moiety for metabolic reduction.

In summary, an HTS campaign identified a new M₄ PAM chemotype, based on a non-enolizable ketone moiety, as a substitute for the classical amide linker in 1-4. Despite attractive M₄ PAM potency and CNS penetration, the compounds uniformly were unstable in hepatic microsomes as well as in vivo, and underwent a rapid and near quantitative reductive metabolism to the corresponding secondary alcohol – an inactive metabolite, confirmed via synthesis and biological evaluation. All attempts to replace the ketone were unsuccessful, rendering this new core not viable to deliver a clinical candidate. Further study and optimization with other HTS series are ongoing, and will be reported in due course.