|Home | About | Journals | Submit | Contact Us | Français|
Herein we describe the discovery and development of a novel class of M4 positive allosteric modulators, culminating in the discovery of ML293. ML293 exhibited modest potency at the human M4 receptor (EC50 = 1.3 µM) and excellent efficacy as noted by the 14.6-fold leftward shift of the agonist concentration-response curve. ML293 was also selective versus the other muscarinic subtypes and displayed excellent in vivo PK properties in rat with low IV clearance (11.6 mL/min/kg) and excellent brain exposure (PO PBL, 10 mg/kg at 1 h, [Brain] = 10.3 µM, B:P = 0.85).
The neurotransmitter acetylcholine (ACh) regulates a diverse set of physiological actions through the activation of two classes of cell-surface receptors. These receptors, the nicotinic ACh receptors (nAChRs)1 are ACh-gated cation channels and the muscarinic ACh receptors (mAChRs) are G protein-coupled receptors (GPCRs).2,3 Both sets of receptors are involved in numerous physiological processes. Molecular cloning has identified five separate subtypes of the mAChRs (M1 – M5) which are subclassified due to their coupling patterns to different G proteins.2–4 One class couples to Gq, activating phospholipase C (M1,3,5), and the second class couples to Gi/o (M2,4), regulating cAMP levels and various ion channels.2–4 Two mAChRs are widely distributed in the CNS (M1 and M4) and have been shown in pre-clinical experiments to be potentially viable therapeutic targets for Alzheimer’s disease and schizophrenia.5 In addition to pre-clinical validation, these mechanisms have been shown to be clinically effective utilizing the non-selective M1/M4-preferring agonist xanomeline.6 Finally, significant work utilizing knockout mice has shown the importance of the role of M1 and M4 in cognition and psychosis, respectively.2, 4
Unfortunately, due to high sequence homology and conservation of the ACh binding site among the mAChRs, development of selective orthosteric modulators has been difficult. Due to this, we and others have been investigating the use of allosteric modulators in order to overcome the homology of the orthosteric site and develop selective tools for these receptors. As part of the NIH-sponsored Molecular Libraries Probe Production Center Network, we have recently reported three selective M4 positive allosteric modulator probe molecules (Figure 1). These molecules have been important tool compounds and have shown efficacy in pre-clinical in vivo animal models of schizophrenia; however, these compounds suffer from less than ideal pharmacokinetic properties.7
The first SAR library kept the left-hand benzothiazole portion constant and looked at a variety of amide analogs (Table 1). These compounds were synthesized by utilizing the commercially available 4-methoxy-7-methylbenzo[d]thiazol-2-amine and coupling with an appropriate acid chloride under basic conditions. The original molecule, 1, reconfirmed and exhibited micromolar potency (EC50 = 1.8 µM). Addition of a methyl group, 18, led to an inactive compound; in addition, replacement of the dioxine moiety with a saturated pyran was deleterious to activity (16 and 17). Replacement with other heteroaryl groups proved more fruitful with the 2-furyl group (4, 1.3 µM) showing a slight improvement in activity. In addition, 3-pyridyl (20, 2.33 µM), 4-pyridyl (21, 1.3 µM) and 2-pyrazine (23, 1.1 µM) modifications were all well tolerated. However, these analogs showed significant differences in efficacy (%ACh Max) with the 4-pyridyl analog, 20, proving to be the most efficacious (65% vs. <40% for 20 and 23). Replacement of the heterocyclic moiety with alkyl, cycloalkyl or aromatic groups led to inactive compounds.
Due to these shortcomings, we were still interested in the discovery of novel M4 tool compounds that offer improved in vivo PK properties. To this end, we performed a screen at Vanderbilt which revealed a novel structural class of compounds as potential M4 positive allosteric modulators (Figure 2, 1). A small library was generated around this molecule and from this effort there were many compounds that potentiated a submaximal (EC20) concentration of ACh (Figure 2). Based on these promising initial results, a lead optimization campaign was initiated in our laboratories.
Next, we turned our attention to amide replacements – an exercise which proved to be less informative and highlights the shallow nature of allosteric modulator SAR (Table 2).9 Moving from the 4-pyridyl amide to the 4-pyridyl urea (29, inactive) led to a complete loss of activity. In fact, all ureas that were analyzed were inactive, with the lone exception being 35, which was weak in terms of both potency (7.1 µM) and efficacy (36%). Other amide replacements (sulfonamide/reverse amides) were not tolerated, resulting in either inactive or weakly active compounds.
Lastly, we investigated replacements of the benzothiazole moiety (Table 3). Removal of both the 4-methoxy and 7-methyl groups were not tolerated (50). In addition, replacement of the benzothiazole with thiazolo[5,4-b]pyridine groups also resulted in inactive compounds (51 – 55). However, deletion of the 7-methyl group (while maintaining the 4-methoxy) was tolerated, resulting in compound 56 (2.8 µM).
We determined that compound 21 possessed the appropriate balance between in vitro potency and efficacy, thus, this compound was further evaluated in our muscarinic selectivity panel as well as in a fold-shift assay (an evaluation of a compound’s ability to produce a left-ward shift of the ACh curve). Compound 21 was examined at the other four muscarinic subtypes (M1,2,3,5) and was inactive against each of these receptors at a concentration up to 30 µM. In our fold-shift assay (at a constant concentration, 30 µM), 21 displayed a robust left-ward shift of the ACh curve (14-fold) showing a potentiation of the ACh potency (Figure 3B). Based on the hM4 potency, fold shift and selectivity profile, 21 was declared an MLPCN probe molecule and redesignated ML293.10
ML293 was next profiled in a number of Tier 1 in vitro pharmacokinetic assays (Table 4). ML293 was assessed using an intrinsic clearance assay (CLINT) in hepatic microsomes to evaluate the oxidative metabolism potential. This assay also provides an in vitro prediction for eventual assessment of clearance values in in vivo PK studies (CLHEP). ML293 was predicted to display moderate clearance in both human and rat microsomes (14.9 and 48.5 mL/min/kg). In addition to intrinsic clearance, we determined the plasma protein binding (human and rat equilibrium dialysis studies) of ML293. ML293 showed high plasma protein binding in human plasma (1.0%, unbound); however, in rat plasma, ML293 showed a more favorable free fraction (3.2%, unbound). Lastly, ML293 was evaluated in a rat brain homogenate binding study to predict the fraction of unbound compound in brain. ML293 showed higher brain homogenate binding profile versus the plasma binding (BHB, 0.9% unbound vs. 3.2% unbound, respectively).
As a follow-up to the in vitro studies, a standard rat IV PK experiment was performed to determine the in vivo clearance of ML293. In this experiment (1 mg/kg, IV), ML293 exhibited low clearance (CLp = 11.5 mL/min/kg) with a modest half-life (t1/2 = 57 mins) and significant plasma levels (AUC = 1445 hr•ng/mL, 4.8 µM). These values are in contrast to the previously reported M4 probe molecules (Figure 1), as each of those molecules exhibit high clearance in rat when assessed using IV PK experiments (CLp > 100 mL/min/kg).7b,11 Lastly, we evaluated the CNS exposure of ML293 in a PBL (Plasma:Brain level) experiment after a single dose and at a single time point. ML293 was dosed orally at 10 mg/kg and after 1 hour the Brain:Plasma (B/P) ratio was 0.85 with absolute brain levels of 3093 ng/g (~10 µM). In addition, in this experiment we assessed the concentrations in the hepatic portal vein as well as plasma. The levels found in the HPV and plasma can be used as another assessment of the extent of oxidative metabolism, as a ratio at or close to 1:1 indicates a low level of metabolism; ML293 exhibits a roughly 0.8 ratio in these studies. Based on these experiments, ML293 is a potent and selective M4 PAM with favorable pharmokinetics that demonstrates significant CNS penetration following oral administration.
In conclusion, we have discovered an additional M4 tool compound (ML293) with superior in vivo PK properties compared to previously published molecules. ML293 represents a novel chemical scaffold as other reported M4 PAMs have been derived from a common 3-amino-N-(aryl)-4,6-disubstitutedthieno[2,3-b]pyridine-2-carboxamide moiety, and have all suffered from similar poor in vivo PK properties. ML293 is a potent and selective small molecule M4 positive allosteric modulator that significantly leftward shifts the ACh curve (14.9 Fold Shift). In addition, ML293, in contrast to the previously reported compounds, displays low clearance when evaluated in an in vivo rat IV clearance assay and is highly brain penetrant (B:P = 0.85, [brain] = ~10 µM).
The authors would like to thank Tammy Santomango, Frank Byers, and Ryan Morrison for technical assistance with the PK experiments and Nathan Kett and Sichen Chang for the purification of compounds. Vanderbilt is a member of the MLPCN and houses the Vanderbilt Specialized Chemistry Center for Accelerated Probe Development (U54MH084659).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.