Discovery of 3-Substituted 1H-Indole-2-carboxylic Acid Derivatives as a Novel Class of CysLT₁ Selective Antagonists

Abstract

The indole derivative, 3-((E)-3-((3-((E)-2-(7-chloroquinolin-2yl)vinyl)phenyl)amino)-3-oxoprop-1-en-1-yl)-7-methoxy-1H-indole-2-carboxylic acid (17k), was identified as a novel and highly potent and selective CysLT₁ antagonist with IC₅₀ values of 0.0059 ± 0.0011 and 15 ± 4 μM for CysLT₁ and CysLT₂, respectively.

Cysteinyl-leukotrienes (CysLTs) are potent inflammatory lipid mediators synthesized from arachidonic acid.^1,2 CysLTs include leukotriene C₄ (LTC₄), leukotriene D₄ (LTD₄), and leukotriene E₄ (LTE₄). They play established or evolving roles in asthma, allergic rhinitis, and other inflammatory conditions, such as cardiovascular diseases, cancer, and atopic dermatitis.³ CysLTs activate at least two receptors, designated as CysLT₁⁴ and CysLT₂,⁵ which belong to the G protein-coupled receptor (GPCR) superfamily.

CysLT₁ receptors are mainly expressed in the lung, peripheral blood leukocytes, and the spleen.⁴ Most of the pathophysiological effects of CysLTs in asthma are mediated by the CysLT₁ receptor.⁶ Several CysLT₁ selective antagonists have been launched for treating bronchial asthma and allergic rhinitis, such as montelukast, pranlukast, and zafirlukast (Figure 1).⁶

Figure 1.

LTD₄ and launched drugs selectively targeting CysLT₁.

CysLT₂ receptors are highly expressed in the peripheral blood leukocytes, spleen, and lymph nodes and are uniquely expressed in the heart, brain, and adrenal glands.⁵ Recently, Sekioka et al. reported the expression of CysLT₂ receptors in asthmatic lungs and investigated their possible role in bronchoconstriction.⁷ However, the pharmacological roles of CysLT₂ are less well-defined, and there is no specific antagonist being marketed as a therapeutic agent so far.⁶ Wunder et al. reported the identification of the first potent and selective CysLT₂ antagonist, HAMI3379 (Figure 2), which was shown to inhibit the cardiovascular effects mainly through mediation of CysLT₂ receptors.⁸ Meanwhile, a highly similar compound, BayCysLT₂ (Figure 2), was also reported to be a potent and selective CysLT₂ antagonist.⁹

Figure 2.

Selective CysLT receptor antagonists.

Although the marketed CysLT₁ selective antagonists are effective therapeutics for the general treatment of mild to moderate bronchial asthma, it is known that the current CysLT₁ antagonists do not have sufficient effects for some nonresponsive patients. A recent report demonstrated that CysLT₁ may also play a role in multiple sclerosis; blocking this receptor protects the integrity of the blood–brain barrier and reduces infiltration of pathogenic T cells into the central nervous system.¹⁰ More recently, Ludwig et al. reported that antiasthmatic drug montelukast, which antagonizes CysLT₁, reduces neuroinflammation, promotes hippocampal neurogenesis, and improves learning and memory in old animals.¹¹ Therefore, the development of new CysLT antagonists is still of great interest.

A high-throughput screening (HTS) campaign of our compound library yielded indole derivative 1, which showed micromolar CysLT₁ antagonist activity with an IC₅₀ value of 0.66 ± 0.19 μM and exhibited no CysLT₂ antagonist activity (Figure 2). Compound 1 shares same hydrophobic (E)-3-(2-(7-chloroquinolin-2-yl)vinyl)phenyl group as montelukast, but it has a unique and essential indole-2-carboxylic acid moiety, which is different from the corresponding counterparts of known CysLT₁ antagonists. The structural novelty of compound 1 encouraged us to further optimize and develop more potent CysLT₁ antagonists.

First, we investigated the effects of the ester groups and designed compounds 10a–10j. The syntheses of compounds 10a–10j are described in Scheme 1. Ethyl-4,6-dichloro-1H-indole-2-carboxylate 4 was prepared from 3,5-dichloroaniline in a two-step synthetic procedure using a Japp–Klingemann condensation followed by a Fischer indole (aza-Cope) ring closure reaction. Vilsmeier–Haack formylation of 4 afforded aldehyde 5.¹² Subsequent ester hydrolysis of 5 afforded carboxylic acid 6, which upon esterification afforded 7. Compound 7 was reacted with tert-butyl(triphenylphosphoranylidene)acetate in a Horner–Wadsworth–Emmons reaction followed by the chemoselective deprotection of the tert-butyl ester group, which yielded the α,β-unsaturated acid intermediate 8.¹³ Esterification of 8 generated compounds 9a–9j, which were then converted to compounds 10a–10j using TBAF in THF.

Scheme 1. Syntheses of Compounds 10a–10j.

Reagents and conditions: (a) DMF, POCl₃, DCE, reflux, 7 h; (b) EtOH, LiOH·H₂O, 50 °C, 2 h; (c) 2-(trimethylsilyl)ethan-1-ol, EDCI, DMAP, DCE, rt, overnight; (d) tert-butyl (triphenylphosphoranylidene)acetate, toluene, reflux, overnight; (e) 98% HCOOH, rt, overnight; (f) R₁Br, Cs₂CO₃, DMF or R₁OH, DCC/EDCI, DMAP, DMF; (g) TBAF, THF, rt, 2 h.

As shown in Table 1, the results revealed that the (E)-3-(2-(7-chloroquinolin-2-yl)vinyl)phenyl group was necessary for the antagonist activity of the compounds; replacing it with other ester groups eliminated antagonist activity against both CysLT₁ and CysLT₂. Therefore, we attempted to modify the functional groups at other positions of hit compound 1 while retaining the (E)-3-(2-(7-chloroquinolin-2-yl)vinyl)phenyl group to increase its antagonist activity against CysLT₁.

Table 1. Effects of Ester Groups on Activity Profiles.

^aAssay protocols are provided in the Supporting Information. IC₅₀ values were obtained from one experiment with three replicates.

The chlorine atoms in the indole ring of compound 1 were removed, and proposed compound 2 (Table 2) was synthesized following the previously described method. We removed the carboxylic acid group of compound 1 and yielded compound 3, which was synthesized following the method shown in Scheme 2. Decarboxylation of 6 provided aldehyde 11,¹⁴ which was subsequently converted to 12 using a Doebner–Knoevenagel modification reaction;¹⁵ finally, esterification of 12 provided 3.

Table 2. Activity Profiles of Indole Derivatives.

					IC50 (μM)a
Cmpd	R1	R2	X	Y	CysLT1	CysLT2
1	4,6-diCl	–COOH	–CH=CH–	–OCH2–	0.66 ± 0.19	>100
2	H	–COOH	–CH=CH–	–OCH2–	0.12 ± 0.02	>100
3	4,6-diCl	H	–CH=CH–	–OCH2–	31 ± 13	>100
17a	4,6-diCl	–COOH	–CH=CH–	–NH–	0.29 ± 0.14	>100
17b	H	–COOH	–CH=CH–	–NH–	0.0090 ± 0.0043	58 ± 26
19b	H	–COOH		–NH–	0.035 ± 0.005	23 ± 1
21b	H	–COOH	–CH2CH2–	–NH–	0.058 ± 0.014	50 ± 18

^aAssay protocols are provided in the Supporting Information. IC₅₀ values were obtained from one experiment with three replicates.

Scheme 2. Synthesis of 3.

Reagents and conditions: (a) CuCl, quinoline, microwave, 200 °C, 10 min; (b) malonic acid, pyridine, piperidine, 50 °C, 12 h; (c) (E)-2-(3-(bromomethyl)styryl)-7-chloroquinoline, Cs₂CO₃, DMF, rt, overnight.

In order to replace the ester bond of compound 1 with an amide bond, we intended to synthesize (E)-3-(2-(7-chloroquinolin-2-yl)vinyl)benzylamine; unfortunately, we failed to obtain the intermediate after several synthetic attempt. Considering the easy availability of (E)-3-(2-(7-chloroquinolin-2-yl)vinyl)aniline, compound 17a was prepared (Scheme 3). As shown in Table 2, compounds 2 and 17a exhibited slightly improved antagonist activities against CysLT₁ compared to hit compound 1, which suggested that derivatives with no substituents in the indole ring were better than derivatives with chlorine atoms and that amide bonds were superior to ester bonds in the internal chain. However, compound 3, which lacked the indole-2-carboxylic acid moiety, was approximately 47-fold less potent than hit compound 1; this demonstrated that the carboxylic acid group at position 2 of the indole ring was necessary. These results further supported the common features of CysLT₁ antagonists, namely, they all contained a lipophilic region, which incorporates into the lipophilic pocket of the CysLT₁ receptor and an acidic moiety modeling the C1-carboxylic acid of LTD4.¹⁶⁻¹⁸ Furthermore, these results were in accordance with the essential structural elements for CysLT₁ receptor ligands.^16,17,19 The activity results of compounds in Table 1 and compound 3 demonstrated that the indole ring, the carboxylic acid function, and the (E)-3-(2-(7-chloroquinolin-2-yl)vinyl)phenyl group were the three necessary pharmocophores for the novel series, the activity results of compounds 2 and 17a revealed that removal of the chlorine atoms in the indole ring and replacement of the ester bond with the amide function were favorable for improvement of the potency; therefore, compound 17b (Scheme 3) was suggested based on the structure–activity relationships (SARs) in the present study. Satisfactorily, 17b demonstrated significantly better antagonist activity against CysLT₁ than hit compound 1, and its antagonist activity against CysLT₂ remained very low. Subsequently, to explore the effects of the α, β-unsaturated double bond at position 3 of the indole ring of 17b on activity profiles, 17b was modified leading to 19b and 21b (Scheme 3). As shown in Table 2, 19b and 21b were approximately 4- and 6-fold less potent against CysLT₁ than 17b, and they demonstrated stronger antagonist activities against CysLT₂ than 17b, which revealed that the importance of the α, β-unsaturated double bond.

Scheme 3. Syntheses of 17a–17k, 19b, and 21b.

Reagents and conditions: (a) DMF, POCl₃, DCE, reflux, 7 h; (b) tert-butyl (triphenylphosphoranylidene)acetate, toluene, reflux, overnight; (c) TFA, DCM, rt, 3 h; (d) (E)-3-(2-(7-chloroquinolin-2-yl)vinyl)aniline, HATU, DIPEA, DMF, rt, overnight; (e) NaOH aq, EtOH, 50 °C, overnight; (f) 30% H₂O₂, NaClO₂, NaH₂PO₄·2H₂O, CH₃CN, rt, overnight; (g) Pd/C, H₂, EtOH-THF, rt, 7 h.

Finally, we investigated the effects of the substituents of the indole ring on the activity profiles with compounds 17c–17k. The syntheses of 17a–17k, 19b, and 21b are described in Scheme 3. Vilsmeier–Haack formylation of 4 and 13b–13k afforded aldehydes 5 and 14b–14k. The Wittig-type olefination of the latter compounds and the subsequent deprotection of the tert-butyl ester group with TFA afforded compounds 15a–15k. Oxidation of 14b afforded 18b. Reduction of 15b by catalytic hydrogenation afforded 20b. Condensation of 15a–15k, 18b, and 20b with (E)-3-(2-(7-chloroquinolin-2-yl)vinyl)aniline and subsequent hydrolysis provided compounds 17a–17k, 19b, and 21b, respectively.

As shown in Table 3, the fluorine substituted derivatives were more potent than the chlorine substituted derivatives (17d vs 17g and 17h vs 17a (Table 2)); 17c, 17d, 17e, and 17f demonstrated that substitution at position 4 of the indole ring was the least favorable. Moreover, the methoxy group substituted derivatives (17i, 17j, and 17k) exhibited substitution at position 7 of the indole ring was the most favorable. Among these derivatives, compounds 17b, 17g, and 17k showed comparable low nanomolar level potencies against CysLT₁, while some derivatives exhibited weak (17b, 17g, and 17k) to no (17j) CysLT₂ antagonist activities.

Table 3. Effects of the Substituents of the Indole Ring on Activity Profiles.

		IC50 (μM)a
Cmpd	R1	CysLT1	CysLT2
17b	H	0.0090 ± 0.0043	58 ± 26
17c	4-Cl	0.067 ± 0.022	36 ± 13
17d	5-Cl	0.017 ± 0.002	>100
17e	6-Cl	0.022 ± 0.002	87 ± 17
17f	7-Cl	0.016 ± 0.004	>100
17g	5-F	0.0078 ± 0.0013	46 ± 20
17h	4,6-diF	0.038 ± 0.001	76 ± 14
17i	5-MeO	0.025 ± 0.007	46 ± 7
17j	6-MeO	0.012 ± 0.007	>100
17k	7-MeO	0.0059 ± 0.0011	15 ± 4

^aAssay protocols are provided in the Supporting Information. IC₅₀ values were obtained from one experiment with three replicates.

CysLTs have been reported to induce chemotactic activity in eosinophils²⁰ and monocytes.²¹ We previously found that LTD₄ could also induce chemotaxis of splenocytes isolated from mice with experimental autoimmune encephalomyelitis (EAE), an animal model of multiple sclerosis,¹⁰ and this effect could be blocked by montelukast.¹⁰ We then tested compounds 1, 17b, 17g, 17j, and 17k with the chemotaxis assay. We found that these compounds could also block LTD₄-induced chemotaxis of leukocytes isolated from the spleen of EAE-mice in a dose-dependent manner (Figure 3). Compound 1 exhibited similar activity as montelukast (Figure 3A,B), and both compounds displayed ∼50% inhibition of chemotaxis at concentrations of 1 μM. Compounds 17b, 17g, 17k, and 17j showed much better inhibitory effects, and all these compounds displayed ≥50% inhibition at concentrations of 100 nM (Figure 3C–F). In particular, compounds 17g and 17k, which displayed best antagonist activity in calcium assay (Table 3) showed significant inhibition of chemotaxis at 10 nM concentration (Figure 3D,E).

Figure 3.

Selective CysLT₁ antagonists inhibit leukocyte chemotaxis induced by LTD₄. Data are from three independent experiments (means ± SEM). ***p < 0.001, versus vehicle control; ^##p < 0.01, ^###p < 0.001 versus LTD₄ (100 nM) treatment group (Student t test).

In summary, we have discovered a novel class of selective CysLT₁ antagonists. Our results indicated that it is essential that such antagonists possess (E)-3-(2-(7-chloroquinolin-2-yl)vinyl)phenyl and indole-2-carboxylic acid moieties for effective CysLT₁ antagonist activity. Additionally, α, β-unsaturated amide moieties at position 3 of the indole rings were also important factors. The most potent compound (17k, IC₅₀ value of 0.0059 ± 0.0011 μM (CysLT₁)) demonstrated significantly more potent CysLT₁ antagonist activity than the known selective CysLT₁ antagonist, montelukast, both in calcium mobilization assay and chemotaxis assay. The further optimization and development including the pharmacokinetic profile of the most potent antagonists and their pharmacological effects evaluated in relevant animal models are in progress in our lab.

Supporting Information Available

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsmedchemlett.5b00482.

• Synthetic procedures, analytic data, procedures for the biological assays, and NMR spectras of the key compounds (PDF)

Author Contributions

^§ These authors (H.C. and H.Y.) contributed equally to this work.

This work was supported by grants from the Ministry of Science and Technology of China (2015CB964503) and the National Natural Science Foundation of China (81425024).

The authors declare no competing financial interest.