Abstract
This Letter describes the synthesis and SAR, developed through an iterative analog library approach, of a novel series of selective M1 mAChR antagonist for the potential treatment of Parkinson’s disease, dystonia and other movement disorders. Compounds in this series possess M1 antagonist IC50s in the 441 nM to 19 µM range with 8- to >340-fold functional selectivity versus rM2–rM5.
The muscarinic acetylcholine receptors (mACHRs) are members of the G Protein-Coupled Receptor (GPCR) family A that mediate the metabotropic actions of the neurotransmitter acetylcholine.1,2 To date, five distinct subtypes of mAChRs (M1–M5) have been cloned and sequenced. M1, M3 and M5 activate phospholipase C and calcium through Gq whereas M2 and M4 block the action of adenylyl cyclase through Gi/o.1,2 The cholinergic system, mediated by mAChRs, 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.1–4 As a result, agents that can selectively modulate the activity of mAChRs have the potential for therapeutic use in multiple pathological states. However, due to high sequence conservation within the orthosteric binding site of the five mAChR subtypes, it has been historically difficult to develop mAChR subtype selective ligands.1–5
To date, the majority of reported muscarinic antagonists are unselective, such as a scopolamine, 1.6 Recently, pirenzapine, 2 has emerged as a relatively selective M1 receptor antagonist (20- to 50-fold versus M2–M5) and there are numerous reports of moderately selective M3 antagonists (20- to 50-fold versus M2) such as 3.7 Interestingly, the most selective M1 antagonist, MT7, 4, the 65 amino acid peptide, (>1,000-fold versus M2–M5) was derived from venom extracts of the green mamba snake (Figure 1).8 Based on brain expression and cellular localization, data from mAChR knock-out mice and clinical trials with muscarinic agents, the M1 mAChR subtype is an attractive molecular target for the treatment of Alzheimer’s disease (AD), Parkinson’s disease (PD) and dystonia due to its role in cognition and motor control.9 Indeed, pan-muscarinic agonists, such as the M1/M4 preferring xanomeline, showed efficacy in Phase III clinical trials in AD patients; however, activation of peripheral M2 and M3 receptors led to intolerable adverse side effects.10 Moreover, anti-cholinergic agents have also demonstrated efficacy in both PD and dystonia patients, and this benefit is believed to be derived from antagonism of the M1 mAChR subtype; however, the relative contributions from M4 are unclear.1–10 In order to probe the role of M1 antagonism as a potential therapeutic approach for Parkinson’s disease, dystonia and other movement disorders, potent small molecule mAChR antagonists are required with a high degree of M1 versus M4 selectivity for study in preclinical models.
Figure 1.
Structures of representative mAChR antagonists.
The Vanderbilt Screening Center for GPCRs, Ion Channels and Transporters, and the companion Chemistry Center, were established as members of the Molecular Libraries Screening Center Network (MLSCN) initiated and supported by the NIH Molecular Libraries Roadmap.11,12 The MLSCN is a nationwide consortium of facilities that provide high-throughput small molecule screening and medicinal chemistry expertise for the development of chemical probes for use as tools to explore biological targets/pathways for which small molecule tools are unavailable.12 One such target which lacks the appropriate small molecule tools are the muscarinic acetylcholine receptors (mAChRs).1–10
Based on this unmet need in the scientific community, our MLSCN Center initiated an effort to identify potent small molecule mAChR antagonists with high specificity for M1 for use as a chemical probe and lead for further optimization towards a novel therapeutic. Towards this goal, we optimized a real-time cell-based calcium-mobilization assay employing a rat M1/CHO cell line (Z’ averaged 0.7), screened a 63,656 member MLSCN compound library and identified 2,179 primary M1 antagonist hits.13 Of these primary hits, 1,665 were available from Biofocus-DPI for re-test, and duplicate testing afforded 723 confirmed hits (43%). These compounds were then counter-screened against an mGluR4/CHO cell line which eliminated 9 hits. The remaining compounds were tested in triplicate in 10-point concentration response curves against both rat M1/CHO and rat M4/CHO cells to identify compounds with ~ 10-fold selectivity for M1 versus M4, our initial cutoff for a lead. While the vast majority of compounds displayed no selectivity for M1 versus M4, we identified two related structures based on a N-(4-(4-ethylpierazin-1-yl)phenyl amide scaffold, 5 (rM1 IC50 = 0.49 µM, rM4 IC50 = 7.9 µM) and 6 (rM1 IC50 = 0.58 µM, rM4 IC50 = 5.1 µM), which displayed ~16- to ~9-fold selectivity, respectively, for rM1 versus rM4 and displayed comparable inhibition of human M1 (Fig. 2).
Figure 2.
HTS leads 5 and 6, rM1 antagonists with selectivity versus rM4 of ~10-fold in the primary assays.
Analogues of 5 and 6 were synthesized in a library format according to Scheme 1. Both requisite anilines 7 and 8, 3-chloro-(4-(4-ethylpierazin-1yl)aniline and (4-(4-ethylpierazin-1yl)aniline, were commercially available and acylated under standard conditions employing polymer-supported reagents and scavengers to afford 24-member libraries of analogues 9 and 10, respectively.14 In the initial lead optimization phase, we prepare a 24-member library employing a diversity set of acid chlorides containing aromatic, alphatic, polar, basic and acidic moieties in order to rapidly probe the breadth and scope of the SAR; subsequent libraries will be more focused. As the chemistry was straightforward, we elected to re-synthesize the parent compounds 5 and 6 within the library. All analogues were purified by mass-guided HPLC to analytical purity.15 Surprisingly, all analogues 10, as well as the re-synthesized parent 6, were found to be inactive on rM1. Moreover, upon resynthesis in the library, 5 lost considerable efficacy as an M1 antagonist (rM1 IC50 = 13 µM), but still displayed ~10-fold selectivity versus rM4 (IC50 >150 µM).16 Not surprisingly, analysis of the original screening samples 5 and 6 indicated that there were several impurities in the wells, and we elected not to pursue a complex deconvolution exercise. Despite these findings, the strategy of employing library synthesis and exploding SAR around a primary HTS hit proved advantageous for 5, as analogues 9 proved to possess intriguing mAChR selectivity profiles.
Scheme 1.
Library synthesis of analogues 9 and 10. Reagents and conditions: (a) (i) PS-DCC, HOBt, RCOOH, (ii) MP-CO32−, 62–98% or (b) (i) RCOCl, PS-DIEA, (ii) PS-trisamine, 79–98%. All library compounds were purified by mass-guided HPLC to >98% purity.15
Table 1 highlights SAR and mAChR selectivity for analogues 9 of HTS hit 5. In general, SAR was rather flat for this series. Truncation of the pentyl side chain of 5 to simpler aliphatic chains, such as n-propyl 9a, led to a total loss of rM1 antagonist activity. Cyclization to form a cyclohexyl ring, as in 9b, afforded a selective rM1 antagonist (rM1 IC50 = 4.6 µM, >32-fold selective versus rM2–rM5), and a 3-fold increase in potency relative to HTS lead 5. The phenyl analog 9c maintained M1 activity relative to 9b, but mAChR selectivity at rM4 began to erode. However, conversion to a benzyl moiety 9d once again maintained rM1 activity (rM1 IC50 = 5.6 µM) and also displayed >26-fold selectivity for rM2–rM5 (IC50s >150 µM). Further chain homologation to the phenethyl congener 9f afforded a low micromolar potency rM1 antagonist (rM1 EC50 = 1.1 µM) with high mAChR subtype selectivity (47-fold versus rM2, 63-fold versus rM3, 16-fold versus rM4 and 6.9-fold versus rM5). Introduction of a cyclic constraint in the form of a cyclopropyl moiety in the phenethyl chain as in 9e provided a compound with an in vitro profile roughly equivalent to 9f. Incorporation of an oxygen atom in the phenylether as in 9g provided an M1 antagonist of modest potency (rM1 IC50 = 3.3 µM), but with >45-fold selectivity versus rM2–rM5 (Figure 3). Replacement of the phenyl moiety with a cyclopentyl group afforded compound 9i, with an rM1 IC50 of 441 nM and with >340-fold selectivity versus M4, but modest selectivity versus rM2, rM3 and rM5 (7.9-fold, 7-fold, and 2.4-fold, respectively). Compound 9i possessed the potentcy requirements for an MLSCN M1 antagonist probe molecule (affinity/activity >500 nM) as well as the required selectivity (>10-fold selectivity) versus rM4 (>340-fold selectivity).11,12 When evaluated against other receptors and enzymes, 9i displayed no significant ancillary pharmacology.
Table 1.
Structures and mAChR activities of analogues 9.
 | ||||||
|---|---|---|---|---|---|---|
| Cmpd | R | M1 IC50 (µM)a | M2 IC50 (µM)a | M3 IC50 (µM)a | M4 IC50 (µM)a | M5 IC50 (µM)a |
| 5 |  |
13.2 | >150 | >150 | >150 | >150 |
| 9a |  |
>150 | >150 | >150 | >150 | >150 |
| 9b |  |
4.6 | >150 | >150 | >150 | >150 |
| 9c |  |
5.0 | >150 | >150 | 66 | >150 |
| 9d |  |
5.6 | >150 | >150 | >150 | >150 |
| 9e |  |
1.15 | 29 | 24 | 20 | 13 |
| 9f |  |
1.1 | 52 | 70 | 18 | 7.6 |
| 9g |  |
3.3 | >150 | >150 | >150 | >150 |
| 9h |  |
18.8 | >150 | >150 | >150 | >150 |
| 9i |  |
0.44 | 3.5 | 3.1 | >150 | 1.1 |
| 9j |  |
>150 | >150 | >150 | >150 | >150 |
IC50s are an average of three independent experiments using rat mAChR (CHO) cell lines.
Figure 3.
Concentration-response curves for 9g on rat M1–M5. Compound 9g displays >45-fold selectivity versus M2–M5. Curves represent the average of three separate experiments.
Our attention now focused on examining mAChR subtype selectivity in binding assays to determine if the functional selectivity was mirrored in competition radioligand binding experiments and to determine whether 9i was binding at the orthosteric versus an allosteric binding site. For these experiments, we evaluated the ability of 9i to displace [3H]-N-methylscopolamine ([3H]-NMS), an orthosteric radioligand, versus all five mAChR subtypes with atropine as a positive control (Figure 5).17 In the event, 9i was shown to possess an rM1 Ki of 12.7 nM with selectivity versus rM2–rM5 (6- to 35-fold) and atropine controls demonstrated pan-mAChR antagonism as anticipated (Table 2). Gratifyingly, the functional rM1 versus rM4 selectivity was mirrored in the radioligand competition binding experiment, but the fold-selectivity had diminished ~10-fold. We often observe shifts in potency and selectivity between binding and functional assays, and we view the functional activity/selectivity as a more important measure of mAChR selectivity as a binding event does not dictate a functional response, i.e., mAChR inhibition.
Figure 5.
[3H]-NMS competition binding experiments for 9i on rat M1–M5. Compound 9i displays 27-fold selectivity versus M2, 6-fold selectivity versus M3, 35-fold selectivity versus M4 and 7-fold selectivity versus M5. Curves represent the average of three separate experiments.
Table 2.
Ki determinations and binding fold-selectivity for 9i.
 | |||
|---|---|---|---|
| mAChR | 9i Ki (nM)a | Fold Selectivity (vs M1) | atropine Ki (nM)a |
| M1 | 12.7±1.7 | 0.88±0.04 | |
| M2 | 338.0±13.5 | 27 | 2.69±0.20 |
| M3 | 74.8±4.3 | 6 | 0.96±0.03 |
| M4 | 445.1±23.8 | 35 | 0.56±0.01 |
| M5 | 85.7±15.9 | 7 | 1.80±0.11 |
Kis are an average of three independent experiments using rat mAChR (CHO) cell lines.
Phosphoinostitide (PI) hydrolysis studies and Schild analysis were performed on 9i to confirm its actvitiy in an alternate signaling pathway modulated by M1 and to further elucidate its binding mode. As shown in Figure 6, 9i causes a dose-dependent rightward shift of the ACh concentration-resposnse curve in a PI hydrolysis experiment which translates in a Schild analysis to a Kd of 10 nM and a slope of 0.98±0.10. These data strongly support the [3H]-NMS binding data and indicate that 9i is an orthosteric M1 antagonist; however, they do not rule out a binding mode wherein 9i partially overlaps with the orthosteric binding site which could account for the observed competitive binding with [3H]-NMS and high rM1 versus rM4 subtype selectivity.17 Nor do these data rule out the possibility that 9i is in fact binding to a non-overlapping allosteric site which causes a conformational exclusion of the orthosteric ligand binding site. Mutagenesis and off-rate experiments are planned to address these possibilities.
Figure 6.
PI hydrolysis studies and Schild analysis for 9i on rat M1. These data strongly support an orthosteric mode of binding for 9i. Data represent the average of three separate experiments.
In summary, an MLSCN M1 antagonist chemical probe development project afforded 9i, a selective rM1 versus rM4 orthosteric antagonist which meets the criteria for a small molecule MLSCN chemical probe. Our hit-to-lead strategy of iterative library synthesis to explode SAR and to re-synthesize HTS hits within the first generation libraries proved highly beneficial, as the intial HTS ‘hits’ lost considerable activity upon resynthesis and evaluation. Had we employed a more traditional approach wherein HTS ‘hits’ were first resynthesized and evaluated prior to generating analogues, these series would not have been pursued further, and 9i would not have been identified. Clearly, serendipity played a major role in the success of this lead optimization strategy, but this is a high risk approach that must be judiciously employed based on the chemistries involved, the assay capacity and the overall cost. Further refinements and in vitro/in vivo pharmacology will be reported for this class of M1 antagonists in due course.
Figure 4.
Concentration response curves for 9i on rat M1–M5. Compound 9i displays 7.9-fold selectivity versus M2, 7-fold selectivity versus M3, >340-fold selectivity versus M4 and 2.4-fold selectivity versus M5. Curves represent the average of three separate experiments.
Acknowledgement
The authors thank the NIH, NIMH and the MLSCN for funding of the Vanderbilt Screening Center for GPCRs, Ion Channels & Transporters (3U54MH074427), the Chemistry Supplement (3U54MH074427-02S1) and the XO1 (1XO1MH077606-01) M1 antagonist screening application which enabled this work.
Footnotes
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.
References and Notes
- 1.(a) Bonner TI, Buckley NJ, Young AC, Brann MR. Science. 1987;237:527–532. doi: 10.1126/science.3037705. [DOI] [PubMed] [Google Scholar]; (b) Bonner TI, Young AC, Buckley NJ, Brann MR. Neuron. 1988;1:403–410. doi: 10.1016/0896-6273(88)90190-0. [DOI] [PubMed] [Google Scholar]
- 2.Felder CC, Bymaster FP, Ward J, DeLapp N. J. Med. Chem. 2000;43:4333–4353. doi: 10.1021/jm990607u. [DOI] [PubMed] [Google Scholar]
- 3.Bymaster FP, McKinzie DL, Felder CC, Wess J. Neurochem. Res. 2003;28:437–442. doi: 10.1023/a:1022844517200. [DOI] [PubMed] [Google Scholar]
- 4.Birdsall NJM, Farries T, Gharagozloo P, Kobayashi S, Lazareno S, Sugimoto M. Mol. Pharmacol. 1999;55:778–786. [PubMed] [Google Scholar]
- 5.Eglen RM, Choppin A, Dillon MP, Hedge S. Curr. Opin. Chem. Biol. 1999;3:426–432. doi: 10.1016/S1367-5931(99)80063-5. [DOI] [PubMed] [Google Scholar]
- 6.(a) Burke RE. Mov. Disord. 1986;1:135–144. doi: 10.1002/mds.870010208. [DOI] [PubMed] [Google Scholar]; (b) Buckley NJ, Bonner TI, Buckley CM, Brann MR. Mol. Pharmacol. 1989;35:469–476. [PubMed] [Google Scholar]; (c) Walebroeck M, Tastenoy M, Camus J, Christophe J. Mol. Pharmacol. 1990;36:267–273. [PubMed] [Google Scholar]; (d) Dei S, Bellucci C, Buccioni M, Ferraroni M, Guandalini L, Manetti L, Martini E, Marucci G, Matucci R, Nesi M, Romanelli MN, Scapecchi S, Teodori E. J. Med. Chem. 2007;50:1409–1413. doi: 10.1021/jm061374r. [DOI] [PubMed] [Google Scholar]
- 7.Hammer R, Berrie CP, Birdsall NJM, Burgen ASV, Hulme EC. Nature. 1980;283:90–92. doi: 10.1038/283090a0. [DOI] [PubMed] [Google Scholar]
- 8.Bradley KN. Pharmacol. & Therap. 2000;85:87–109. doi: 10.1016/s0163-7258(99)00064-9. [DOI] [PubMed] [Google Scholar]
- 9.Birdsall NJM, Nathanson NM, Schwarz RD. TRENDS Pharm. Sci. 2001;22:215–219. [Google Scholar]
- 10.(a) Bodick NC, Offen WW, Levey AI, Cutler NR, Gauthier SG, Satlin A, Shannon HE, Tollefson GD, Rasmussen K, Bymaster FP, Hurley DJ, Potter WZ, Paul SM. Arch Neurol. 1997;54:465–473. doi: 10.1001/archneur.1997.00550160091022. [DOI] [PubMed] [Google Scholar]; (b) Bodick NC, Offen WW, Shannon HE, Satterwhite J, Lucas R, van Lier R, Paul SM. Alzheimer Dis Assoc Disord. 1997;11 Suppl 4:S16–S22. [PubMed] [Google Scholar]
- 11. For information on the Molecular Library Screening Center Network see: http://nihroadmap.nih.gov/molecularlibraries.
- 12. For information on the Vanderbilt Molecular Library Screening Center Network Center please refer to: http://www.vanderbilt.edu/mlscn/Templates/index.htm.
- 13.Functional assay. Chinese Hamster Ovary (CHO) cells containing rat M1 (rM1) receptor were plated at 10,000 cells/well and rM2/Gqi5, rM3, rM4/Gqi5, and rM5 were all plated in 384 well plates at 25,000 cells/well in assay media 1 (F12 (Ham), 10% FBS, 2 mM GlutaMAX (Invitrogen), 20mM HEPES), except rM4/Gqi5 which used assay media 2 (DMEM, 20 mM HEPES, 10% FBS, 2 mM L-glutamine, 1X Nonessential Amino Acids (NEAA) and 1 mM Na pyruvate). The plates were incubated overnight at 37°C in 5% CO2. Media was removed and assay buffer (Hanks Balanced Salt Solution, 20 mM HEPES, 2.5 mM Probenecid, pH 7.4) containing 4.0µM Fluo4-AM dye (Invitrogen) was added. Cells were incubated for 45 minutes (37°C, 5% CO2) for dye loading. Cell plates were loaded into the Hamamatsu FDSS equipped with 480 nanometer excitation and 540 nanometer emission filters. Test compound in assay buffer + 0.1% DMSO was added at 5 seconds and simultaneously the plate was kinetically imaged. Subsequently, 8 nM acetylcholine (EC80) in assay buffer was added at 197 seconds and imaging continued for a total of 4 minutes acquisition time. 0.1% DMSO, compound vehicle, and 80 nM acetylcholine (ECMax) were added to each plate as controls. Compounds were tested in concentrations ranging from 150 µM to 2.5nM in triplicate on three different days. Kinetic data were transformed and fit with GraphPad Prism version 4.0 to a 4 parameter logistic equation to determine IC50 values.
- 14.(a) Lindsley CW, Weaver D, Conn PJ, Marnett L. ACS Chem. Biol. 2007;2:17–21. doi: 10.1021/cb6004867. [DOI] [PubMed] [Google Scholar]; (b) Lindsley CW, Weaver D, Bridges TM, Kennedy JP. Wiley Encyclopedia of Chemical Biology. in press. [Google Scholar]
- 15.Leister WH, Strauss KA, Wisnoski DD, Zhao Z, Lindsley CW. J. Comb. Chem. 2003;5:322–329. doi: 10.1021/cc0201041. [DOI] [PubMed] [Google Scholar]
- 16.Typical experimental for the synthesis of 9i: To a solution of 3-chloro-4-(4-ethylpiperazin-1-yl)aniline 7 (100 mg, 0.417 mmol) in 9:1 DMF:DIEA (2 mL) was added 3-cyclopentyl propanoyl chloride (63.9 µL, 0.417 mmol) all at once. The reaction was stirred at room temperature for 12 h at which time it was determined complete by LC/MS. The reaction was comcentrated in situ, redissolved in 1 mL DMSO, and purified on the Agilent 1200 preparative LCMS. Concentration of purified fractions afforded the TFA salt of N-(3-chloro-4-(4-ethylpiperazin-1-yl)phenyl)-3-cyclopentylpropanamide 9i as a white solid (80 mg, 42%). 1H NMR (400 MHz, DMSO-d6) δ 10.03 (br s, 1H), 7.84 (d, J = 2.0 Hz, 1H), 7.44 (dd, J = 2.0, 8.4 Hz, 1H), 7.16 (d, J = 8.4 Hz, 1H), 3.57 (m, 2H), 3.35 (m, 2H), 3.21 (m, 2H), 3.13 (m, 2H), 2.99 (m, 2H), 2.29 (t, J = 7.6 Hz, 2H), 1.73 (m, 3H), 1.58 (m, 4H), 1.47 (m, 2H), 1.25 (t, J = 7.2 Hz, 2H), 1.08 (m, 3H). 13C NMR (100 MHz, DMSO-d6) δ 171.4, 142.3, 136.2, 127.5, 121.1, 120.6, 118.5, 50.8, 50.7, 48.1, 39.3, 35.7, 32.0, 31.3, 24.7, 8.9; LCMS, single peak, 2.79 min, m/e, 364.2 (M+1).
- 17.Phosphoinositide (PI) Hydrolysis. Hamster Ovary (CHO) cells containing rat M1 (rM1) were plated at 120,000 cells per well in standard growth media (F12 (HAM), supplemented with 10% fetal bovine serum, and 20 mM HEPES) in 24 well plates twenty-four hours prior to assay. Cell media was replaced late in the day with standard growth media containing 1 µCi/mL [3H]inositol (Perkin-Elmer LAS) and cells were incubated overnight at 37°C in 5% CO2. [3H]inositol-containing media was removed and the rM1 cells were treated with either vehicle or fixed concentrations of antagonist (2X, 500 µL) in a modified Krebs’-bicarbonate buffer (108 mM NaCl, 4.7 mM KCl, 2.5 mM CaCl2, 1.2 mM MgSO4, 1.2 mM KH2PO4, 25 mM NaHCO3, 10 mM Glucose, pH 7.4) equilibrated to 37°C and 5% CO2 and supplemented with 30 mM LiCl. Following vehicle or antagoist addition, agonist was added (2X, 500 µL) in modified Krebs’-bicarbonate buffer and cells were incubated for 1 hour at 37°C and 5% CO2. The accumulation of phosphoinositides was terminated by aspiration, followed by the addition of 1 mL of 10 mM formic acid. Cells were incubated in the 10 mM formic acid for at least 30 minutes at room temperature to insure extraction of phosphoinositides. Columns packed with a 1 mL bed of AG® 1-X8 Resin 100–200 mesh anion-exchange resin (formate form) (Bio-Rad; Hercules, CA) were washed with 10 mL of water twice. Following the water washes, the entire 1 mL sample volumes were added to columns, avoiding the transfer of any cells. Columns were washed with 2 mL of water, then with 10 mL of water, followed by 10 mL of 5 mM myo-inositol made up in water. Total phosphoinositides (PIs) were eluted with 10 mL of 0.1 M formic acid / 0.2 M ammonium formate into vials containing 3a70B liquid scintillation cocktail (Research Products International; Elk Grove Village, IL) and radioactivity was measured by liquid scintillation counting. Following use, columns were regenerated with 10 mL of 0.1 M formic acid / 1M ammonium formate and washed with 30 mL of water. Data were fit with GraphPad Prism version 4.0 to a 4 parameter logistic equation to determine IC50 values and Schild Dose-Ratios.