Design and Synthesis of (2-(5-Chloro-2,2-dimethyl-2,3-dihydrobenzofuran-7-yl)cyclopropyl)methanamine as a Selective Serotonin 2C Agonist

Abstract

A conformationally restricted analog of a selective cyclopropane-bearing serotonin 2C agonist was designed and synthesized. A 2,2-dimethyl-2,3-dihydrobenzofuran scaffold was investigated as a constrained variant of a biologically active isopropyl phenyl ether. Construction of the required dimethyl-2,3-dihydrobenzofuran intermediate began using a procedure that relied on a microwave-assisted alkylation reaction. The synthesis of the designed compound as its HCl salt is reported in a total of 12 steps and 17% overall yield. Biological evaluation revealed the constrained analog to be a selective serotonin 2C agonist with modest potency.

Introduction

The serotonin 2C (5-HT_2C) receptor has been found to represent a promising drug target in the search for new treatments for a variety of disorders, including obesity and various mental diseases, such as schizophrenia, depression, and anxiety.^1,2,3 One of the many advantages of the 5-HT_2C receptor as a central nervous system (CNS) drug target stems from the fact that it is almost exclusively found in the CNS,^4,5 and thus compounds that selectively activate this receptor should have limited impact on peripheral tissues. However, the activation of two other closely related 5-HT₂ receptor subtypes, namely the 5-HT_2A and 5-HT_2B receptors, has been reported to be associated with hallucinatory effects and life threatening cardiac valvulopathy, respectively.^6,7,8 Therefore, the identification of ligands possessing exquisite selectivity for the 5-HT_2C versus the 5-HT_2A and 5-HT_2B receptor subtypes is a key requirement for therapeutic advancement of 5-HT_2C agonists.³ However, the achievement of such selectivity has proven to be challenging because of the high amino acid sequence homology shared by members of the 5-HT₂ subfamily.

During our studies aimed at designing selective 5-HT_2C agonists for possible use in CNS diseases, compound (+)-1 was found to be a relatively potent and selective 5-HT_2C agonist (5-HT_2C: EC₅₀ = 71 nM, E_max = 100% (of 5-HT); 5-HT_2B: EC₅₀ = 682 nM, E_max = 57%; 5-HT_2A: EC₅₀ = 7190 nM, E_max = 17%). Based on the sum total of the structure activity relationship (SAR) data we had assembled for this lead candidate, we considered the possibility of exploring the activity of a dimethyl-2,3-dihydrobenzofuran analog 2, which represents a conformationally constrained version of the isopropyl phenyl ether 1. Conformational restriction is, of course, a well-recognized medicinal chemistry strategy for enhancing the binding affinity as well as the target selectivity of small molecules.⁹ This particular structure was also of interest as the heterocyclic benzofuran system is present in a number of drugs.^10,11 For example, the 2,2-dimethyl-2,3-dihydrobenzofuran scaffold is found in the drug zatosetron (3),¹² which is a 5-HT₃ antagonist that is able to induce antinausea effects without affecting gastrointestinal motility.¹³ Zatosetron was also found to be an effective anxiolytic in both animal studies and human trials.¹⁴ The similarity of compound 2 with 3, and the fact that both compounds 1 and 3 act through serotonin receptors thus encouraged us to synthesize compound 2 for biological evaluation.

Result and discussion

The retro-synthetic analysis of compound (±)-2 is outlined in Figure 2. Benzaldehyde 4 was selected as a key intermediate, as the cyclopropylmethylamine side chain of 2 can be installed from this aldehyde, using methods similar to those that we have reported previously.¹⁵ This benzaldehyde could in turn be prepared from its corresponding benzoate 5.^{16, 17} As compound 5 has previously served as an intermediate in the synthesis of zatosetron (3), a synthetic approach to this intermediate starting from methyl 5-chloro-2-hydroxybenzoate (6) has been reported.^12,18 We note that no spectroscopic data have been reported previously for the intermediates prepared in the synthesis of 5, and that this compound had previously been obtained in only very low yield.

Figure 2.

Retrosynthetic approach of compound (±)-2.

The synthesis of intermediate 5 is depicted in Scheme 1. The alkylation of the starting material methyl 5-chloro-2-hydroxybenzoate (6) was accomplished using 3-chloro-2-methylprop-1-ene under reflux conditions in acetone, however, this led to only a 29% yield of product 7.¹² In order to improve the yield of this alkylation reaction, 3-bromo-2-methylprop-1-ene was used as the alkylating reagent and the reaction was carried out under microwave conditions (80 °C). This method gave an improved yield (81%) of the product 7, and the reaction time was reduced from overnight to 1.5 h. The Claisen rearrangement of intermediate 7 turned out to be sluggish. Heating of 7 in NMP at 200 °C for 8 h provided a 65% yield of product 8, with a 24% recovery of the starting material. The subsequent cyclization of compound 8 was brought about by refluxing in 95% formic acid, which gave an excellent yield of intermediate 5. With the assistance of microwave conditions, intermediate 5 was prepared in 52% yield from 6 in three steps, compared to an overall 20% yield as reported previously.¹²

Scheme 1.

Synthesis of benzaldehyde 5. Reagents and conditions: (a) 3-bromo-2-methylprop-1-ene, K₂CO₃, DMF, microwave, 80 °C, 1.5h, 81%; (b) NMP, 200 °C, 8h, 65%; (c) formic acid (95%), reflux, 3h, 98%.

The ester 5 was then subjected to a sequence of lithium aluminum reduction and pyridinium dichromate (PDC) oxidation to give the benzaldehyde 4 in good yield. Wittig reaction of 4 with a commercially available reagent N-methoxy-N-methyl(triphenylphosphoranylidene)acetamide in dichloromethane at room temperature provided the Weinreb amide 10 in high yield. The double bond was obtained in solely the E form, without the observation of its Z isomer. The coupling constant of the two protons attached to the double bond is 16.0 Hz, which is consistent with the data reported for similar compounds.¹⁵ Subsequent Corey-Chaykovsky cyclopropanation of the E double bond, with the sulfur ylide which was generated from trimethylsulfoxonium iodide upon treatment with sodium hydride, gave the cyclopropane 11 as its trans isomer. The Weinreb amide was then reduced with diisobutylaluminium hydride at low temperature to give aldehyde 12, which was subsequently reduced to the corresponding alcohol 13 in excellent yield using sodium borohydride. The primary amine 2 was prepared from its alcohol 13 through a sequence of steps involving a Mitsunobu reaction employing phthalimide to afford the Gabriel imide 14, followed by de-protection with hydrazine. Finally, the hydrochloride salt of 2 was prepared using HCl in diethyl ether. This sequence of transformations provided the target compound in 9 steps from 5, with an overall yield of 33%. While preparative chiral HPLC methods were investigated for the separation of the optically pure enantiomers of the racemic Boc derivative of compound (±)-2, no separation was observed using either OD or OJ columns employing various solvent systems. Therefore, the biology activity of compound 2 was evaluated using its racemate.

Functional activity at the 5-HT_2A, 5-HT_2B and 5-HT_2C receptors was determined using the calcium flux assay (for details of the assay, see supplementary data) for the racemic form of compound 2. Moderate potency was observed for compound 2 at the 5-HT_2C receptors (EC₅₀ = 600 nM) where it acts as a partial agonist (E_max = 66% of 5-HT). No activity was observed at either the 5-HT_2A or 5-HT_2B receptors. The absence of any 5-HT_2B activity for compound 2 is beneficial, as this would preclude any undesirable cardiac toxicity as noted above. The decreased activity of this compound at 5-HT_2C receptors may be a result of its increased steric size compared to compound 1, as previous SAR studies have shown that increasing the length of the ether chain by one carbon led to a significant decrease in 5-HT_2C activity.¹⁵

Conclusion

In this letter we report the design and synthesis of a selective 5-HT_2C agonist compound 2 bearing a dimethyl-2,3-dihydrobenzofuran moiety. An improved synthesis of the dimethyl-2,3-dihydrobenzofuran scaffold is reported that makes use of a microwave-assisted alkylation reaction followed by a Claisen rearrangement and a formic acid promoted cyclization. The 2-phenylcyclopropylmethylamine side chain of the target compound was synthesized using a sequence of steps involving a Wittig reaction, Corey-Chaykovsky cyclopropanation, reduction of a Weinreb amide, and a Mitsunobu reaction followed by deprotection of the thus formed Gabriel imide. The HCl salt of the target compound was prepared in 12 steps in an overall yield of 17%. The functional activity of this compound at the 5-HT_2C, 5-HT_2A and 5-HT_2B receptors was then determined using a calcium flux assay, and compound 2 was found to be a selective 5-HT_2C partial agonist of moderate potency. The design of other analogs of 2 would thus appear to be valuable to explore.

Figure 1.

Structures of compounds (+)-1, (±)-2 and zatosetron (3).

Scheme 2.

Synthesis of target compound 2. (a) LiAlH₄, THF, 0 °C to rt, 2h, 92%; (b) PDC, CH₂Cl₂, rt, 2h, 65%; (c) Ph₃P=CHC(O)N(OMe)Me, CH₂Cl₂, rt, overnight, 92%; (d) Me₃S⁺(O)I⁻, NaH, DMSO, rt, overnight, 98%; (e) DIBAL-H, THF, −78 °C to rt, 95%; (f) NaBH₄, MeOH, 0 °C to rt, 96%; (g) Phthalimide, PPh₃, DEAD, THF, rt, 85%; (h) N₂H₄–H₂O, EtOH, reflux, 3h, 97%; (i) 2M HCl in Et₂O, rt, 2h, 81%.

Table 1.

Pharmacological profiles of compound (+)-1 and (±)-2.^a

Compound	EC50, nM (Emax; % 5-HT)
	5-HT2C	5-HT2B	5-HT2A
5-HT	0.20 (100%)	1.3 (100%)	2.1 (100%)
(+)-1 (TFA salt)	71 (100%)	682 (57%)	7190 (17%)
(±)-2 (HCl salt)	600 (66%)	NA	NA

^aFunctional data was acquired with recombinantly expressed human serotonin receptors in HEK-293 (5-HT_2A and 5-HT_2B) and PO1C (5-HT_2C) cell lines, fluorescence imaging plate reader (FLIPR) assay; “NA”, no activity at 10 µM.