Synthesis and SAR of Analogues of the M1 allosteric agonist TBPB. Part II. Amides, sulfonamides and ureas: the effect of capping the distal basic piperidine nitrogen

Abstract

This Letter describes the further synthesis and SAR, developed through an iterative analogue library approach, of analogues of the highly selective M1 allosteric agonist TBPB by deletion of the distal basic piperidine nitrogen by the formation of amides, sulfonamides and ureas. Despite the large change in basicity and topology, M1 selectivity was maintained.

The muscarinic acetylcholine receptors (mAChRs) are members of the GPCR family A that mediate the metabotropic actions of the neurotransmitter acetylcholine (ACh). Five distinct subtypes of mAChRs (M1–M5) have been cloned and sequenced.^1–3 mAChR-regulated cholinergic signaling plays a critical role in a wide variety of CNS and peripheral functions including memory and attention mechanisms, motor control, nociception, regulation of sleep wake cycles, cardiovascular function, renal and gastrointestinal function and many others.^4–6 As a result, agents that can selectively modulate the activity of specific mAChRs have therapeutic potential in multiple pathological states, including Alzheimer’s disease and schizophrenia.^1–6 However, due to high sequence conservation within the orthosteric binding site of the five mAChR subtypes, it has been historically difficult to develop subtype selective ligands though accounts of an ectopic agonist, such as AC-42, and brucine, a weak M1 positive allosteric modulator have been reported.^1–8

We recently reported on TBPB 1, a potent, centrally active and highly selective M1 allosteric agonist (Figure 1), which displayed robust efficacy in several preclinical antipsychotic models as well as significant effects on the processing of amyloid precursor protein (APP) towards the non-amyloidogenic pathway and decreased Aβ production.⁹ However, TBPB 1 was an un-optimized screening lead with antagonist activity at D2 (IC₅₀ = 5.1 μM), and despite an [¹⁸F]-fallypride micro-PET study that confirmed the antipsychotic activity observed with 1 was the result of selective M1 activation and not due inhibition of D2, we hoped to diminish D2 activity through a lead optimization campaign.^9,10 This effort initially focused on alternative benzyl moieties and afforded TBPB analogues, such as 2 which were equivalent to TBPB in terms of M1 efficacy and mAChR selectivity, but possessed no D2 inhibitory activity.¹⁰ Other analogues displayed improved M1 efficacy, but at the cost of enhanced inhibition of D2. In this Letter, we describe the synthesis, SAR and pharmacological profile of new TBPB analogues in which the benzyl moiety is replaced with either an amide, sulfonamide or urea linkage to determine the effects on mAChR activation of capping the distal nitrogen atom and altering the basicity and topology of TBPB.

Figure 1.

Concentration response curves for TBPB (1) and 2 at M1–M5. Data represent the mean±SEM of three independent determinations. TBPB (1) M1 EC₅₀ = 289 nM, 82% CCh Max, 2 M1 EC₅₀ = 410 nM, 82% CCh Max; M2–M5 EC₅₀ >50 μM.

In the initial lead optimization campaign, we prepared and evaluated one congener wherein the 2-methylbenzyl moiety in 1 was replaced with an ethyl carbamate. Surprisingly, this analogue lost all selectivity for M1 and degenerated into a potent, but pan-mAChR agonist (M1 EC₅₀ = 2.1 nM (82.6% CCh Max), M2 EC₅₀ = 11.9 nM (80.6% CCh Max), M3 EC₅₀ = 113 nM (24.1% CCh Max), M4 EC₅₀ = 9.1 nM (96.9% CCh Max), M5 EC₅₀ = 34.3 nM (50.8% CCh Max). Despite this finding, we decided to pursue amide congers, as SAR for allosteric ligands can often be very ‘flat’ and slight structural changes can have radically different pharmacological profiles. ^10–13

Amide analogues of TBPB 5 were prepared by a reductive amination sequence employing functionalized acyl piperdiones 4 and either 4-(2-keto-1-benzoimidazolinyl)piperidine or the 5-chloro conger and polymer-supported triacetoxyborohydride (Scheme 1). In short order, a 24-member library of amide analogues 5 was prepared and evaluated against M1–M5. Once again, SAR for this series was ‘flat’ with only 30% of the analogues affording M1 activation (Table 1). However, unlike the pan-mAChR carbamate agonist, all of the active amide congers 5a–5h in Table 1 were selective for M1 (i.e., no activation of M2–M5 at concentrations up to 30 μM). The SAR for this series was quite different from the benzyl amine series represented by 1 and 2, wherein substitution was only tolerated in the 2-position of the benzyl ring.¹⁰ For the amide series, methyl 5c or chloro groups 5g in the 3-position afforded M1 agonist with submicromolar EC₅₀s (510 nM and 930 nM, respectively). Uniformly, the chlorine in the 5-position of the benzimidazolone ring diminished M1 efficacy, in contrast to the benzyl amine series.¹⁰ Moreover, all of the analogues in Table 1 were found to be partial allosteric agonists (CCh Max <60%), and none possessed significant inhibition of D2 (<10% @ 10 μM). These data suggest the distal basic nitrogen is not required for M1 activation or selectivity versus M2–M5.

Scheme 1.

Reagents: (a) R¹COCl, PS-DIEA, 85–88%; (b) MP-B(OAc)₃H, 4-(2-keto-1-benzoimidazolinyl)piperidine or 5-chloro-1-(4-piperidinyl)-2-benzimidazolone, DCM, rt, 24 hr, 85–95%. All compounds purified by mass-directed HPLC to analytical purity (>98%).

Table 1.

Functional Activity of TBPB Amide Analogues 5.

Cmpd	R1	R2	M1 EC50 (μM)a	%CCh Maxa
5a	2-CH3	H	0.88	60
5b	2-CH3	CI	2.71	56
5c	3-CH3	H	0.51	53
5d	3-CH3	CI	2.11	50
5e	2-CF3	H	0.91	32
5f	2-CF3	CI	11.1	24
5g	3-CI	H	0.93	21
5h	3-CI	CI	7.11	21

^aEC₅₀S and % CCh maximum (measures activation of M1 relative to 100% response of CCh) are the mean of at least three independent detrminations. All analogues in this Table are selective for M1 (no activation of M2–M5 at conc. up to 30 μM).

These data prompted the investigation of other capping agents, such as sulfonamides and ureas, to introduce additional proton acceptors or proton donor/acceptors, respectively. In this instance, we took advantage of our large supply of 1 and removed the benzyl moiety by hydrogenation and then treated the resulting piperidine 6 with a diverse set of eleven sulfonyl chlorides 7 or eighteen isocyanates 8 to provide the corresponding 29-member library of sulfonamides 9 and ureas 10, respectively (Scheme 2). The library was triaged employing a 10 μM single point M1 activation screen in triplicate, relative to ACh, CCh and TBPB controls (Figure 2).^9,10 This effort identified only two compounds, one sulfonamide 9c and one urea 10i, that activated M1, highlighting again the ‘flat’ SAR for this M1 allosteric binding site. Full concentration-response curves were then generated for 9c and 10i at M1, along with M2–M5 selectivity assays (Figure 3). Gratifyingly, these two hits provided complete selectivity for M1 versus M2–M5 with micromolar EC₅₀s for M1 activation (9c, M1 EC₅₀ = 1.4±0.007 μM, 82% CCh Max and 10i, M1 EC₅₀ = 6.5±0.3 μM, 85% CCh Max) and neither analogue possessed significant inhibition of D2 (~10%@10 μM). Unlike the weaker partial agonist amide series 5, both 9c and 10i are more efficacious M1 agonists, which justified further optimization of these hits.

Scheme 2.

Reagents: (a) H₂, Pd/C, rt, 92%; (b) R₁SO₂Cl (7a–k) or R₁NCO (8a–r), DCM, PS-DIEA, 70–90%. All compounds purified by mass-directed HPLC to analytical purity (>98%).

Figure 2.

10 μM compound screen for M1 activation. Data represent the mean±SEM of three independent determinations with TBPB (1), ACh and CCh as positivie controls. Boxed panels show the sulfonyl chlorides 7a–7k and isocyanates 8a–8r employed in the library (Scheme 2).

Figure 3.

Concentration response curves for 9c and 10i at M1–M5. Data represent the mean±SEM of three independent determinations. 9c M1 EC₅₀ = 1.4 μM, 82% CCh Max, 10i M1 EC₅₀ = 6.5 μM, 85% CCh Max; M2–M5 EC₅₀ >30 μM.

As shown in Scheme 3, analogues of 9c and 10i were easily prepared in two steps by reacting piperidone 3 with either 7c or 8i to afford 11 or 12, respectively. Then, 11 and 12 underwent a reductive amination reaction with a series of 4-, 5- or 6-halogen functionalized piperidine benzimidazolones 13 to produce analogues 14 and 15, respectively. The SAR was ‘flat’, with all analogues 14 being of comparable M1 potency (Table 2), while most analogues 15 were inactive (Table 3). For both series, the degree of D2 inhibition varied widely (D2 IC₅₀ = 1.9 μM to >10 μM), but the %CCh max was >70% except 14a. As shown in Table 2, neither the location of fluorine incorporation nor which halogen (F, Cl or Br) was appended in the 5-position of the benzimazolone scaffold, the M1 EC₅₀s and selectivity versus M2–M5 were comparable. Of analogues 14, only the 6-F congener 14c was a weaker partial agonist at M1, with the other retaining a higher degree of partial agonism akin to the benzyl amine series and TBPB 1. We found that the variability of D2 inhibition was quite surprising, with halogens in the 5-position, 14b, 14d, and 14e, affording micromolar levels of D2 inhibition (IC₅₀s of 3.6 μM, 2.3 μM and 2.7 μM, respectively). In contrast, fluorines in the 4- or 6- position, as in 14a and 14c, did not inhibit of D2 at concentrations up to 10 μM, suggesting the 5-position of analogues 14 is a molecular switch to dial in D2 inhibitory activity. This was an important finding as a dual M1 allosteric agonist/D2 antagonist may prove to be an excellent profile for a novel schizophrenia treatment.

Scheme 3.

Reagents: (a) PhSO₂Cl, PS-DIEA, 88%; (b) MPB( OAc)₃H, 5-halo-1-(4-piperidinyl)-2-benzimidazolones 13, DCM, rt, 24 hr, 85–95%; (c) m-SCF₃NCO, PS-DIEA, 80%; All compounds purified by mass-directed HPLC to analytical purity (>98%).

Table 2.

Functional Activity of TBPB Amide Analogues 14.

Cmpd	X	M1 EC50 (μM)a	%CCh Maxa	D2 IC50 (μM)a
14a	4-F	2.6	49	>10
14b	5-F	2.3	82	3.6
14c	6-F	3.3	76	>10
14d	5-Cl	4.1	82	2.3
14e	5-Br	4.4	81	2.7

^aEC₅₀s, % CCh maximum (measures M1 activation relative to 100% CCh response) and IC₅₀s are the mean of at least three independent determinations. All analogues in this Table are selective for M1 (>30 M vs. M2–M5).

Table 3.

Functional Activity of TBPB Amide Analogues 15.

Cmpd	X	M1 EC50 (μM)a	%CCh Maxa	D2 IC50 (μM)a
15a	4-F	>10	46	1.1
15b	6-F	9.6	72	1.9
15c	5-CI	3.6	90	>10

^aEC₅₀s, % CCh maximum (measures M1 activation relative to 100% CCh response) and IC₅₀s are the mean of at least three independent determinations. All analogues in this Table are selective for M1 (>30 μM vs. M2–M5).

Activity for the urea congeners 15 was disappointing. Only two analogues, 15a and 15c, had any significant efficacy at M1 (15a, EC₅₀ = 9.6 μM, 72% CCh Max.; 15c, EC₅₀ = 3.6 μM, 90% CCh Max.). Interestingly, 15c, with the 5-Cl benzimidazolone moiety, was more potent than the parent 10i, equally efficacious and possessed no D2 inhibitory activity, whereas 15a was a ~1 μM D2 antagonist. Based on these data that allosteric M1 activation proceeds in the absence of the distal piperidine nitrogen, future exploration will center on diminishing basicity through the introduction of β-fluoroamine moieties to provide congeners 16 and 17 of TBPB (Figure 4).

Figure 4.

Future TBPB analogues 16 and 17: introducing β-fluoroamines to diminish basicity.

In summary, we have identified three novel series of allosteric partial agonists with high selectivity for M1 versus M2–M5 based on the TBPB scaffold, in which the distal piperidine nitrogen has been capped to form amides, sulfonamides or ureas. SAR was ‘flat’ within these series, with subtle changes resulting in loss of M1 efficacy, significant decreases in the degree of partial agonism or significant increases/decreases in D2 inhibition. As with other allosteric ligands, these data suggest that the allosteric binding site for TBPB and related analogues is shallow. Importantly, SAR did not track between these capped series, or with respect to the original TBPB benzyl amine series which further exemplifies the challenges in the development of allosteric ligands for GPCRs. However, we were able to demonstrate that the distal basic amine in TBPB is not required for binding and activation of M1 at the TBPB allosteric site, and within the sulfonamide series 14, the 5-position of the benzimidazolone is a molecular switch for engendering D2 inhibitory activity. Further refinements to the TBPB scaffold are in progress and will be reported in due course.