Amino-Substituted 3-Aryl- and 3-Heteroarylquinolines as Potential Antileishmanial Agents

Abstract

Leishmaniasis, a disease caused by protozoa of the Leishmania species, afflicts roughly 12 million individuals worldwide. Most existing drugs for leishmaniasis are toxic, expensive, difficult to administer, and subject to drug resistance. We report a new class of antileishmanial leads, the 3-arylquinolines, that potently block proliferation of the intramacrophage amastigote form of Leishmania parasites with good selectivity relative to the host macrophages. Early lead 34 was rapidly acting and possessed good potency against L. mexicana (EC₅₀ = 120 nM), 30-fold selectivity for the parasite relative to the macrophage (EC₅₀ = 3.7 μM), and also blocked proliferation of Leishmania donovani parasites resistant to antimonial drugs. Finally, another early lead, 27, which exhibited reasonable in vivo tolerability, impaired disease progression during the dosing period in a murine model of cutaneous leishmaniasis. These results suggest that the arylquinolines provide a fruitful departure point for the development of new antileishmanial drugs.

Introduction

Leishmaniasis is a neglected tropical disease whose causative organism possesses a complex life cycle that presents a difficult challenge for drug development.¹⁻⁴ Leishmaniases are caused by the protozoan parasites from the genus Leishmania carried between mammalian hosts by sandflies of the Phlebotomus and Lutzomyia genera. All Leishmania species have a digenetic life cycle that includes motile promastigotes that reside in the gut of the sand fly vector and nonmotile amastigotes that live in the phagolysosomal vesicles of mammalian host macrophages.⁵ Leishmaniasis, endemic throughout the tropics, subtropics, and the Mediterranean basin, places an estimated 350 million people at risk and causes 1.5 million annual cases.^3,6−8 Leishmaniasis presents with a spectrum of symptoms and the WHO characterizes infections into three broad categories: the self-healing but disfiguring cutaneous leishmaniasis;⁹ disfiguring mucocutaneous leishmaniasis;¹⁰ and potentially fatal visceral leishmaniasis (VL).¹¹ In visceral disease, the parasites attack the patient’s internal organs including the liver, spleen, and bone marrow. The annual morbidity of VL ranges as high as 100,000 people per year and more than 90% of cases occur in India, Bangladesh, Sudan, Ethiopia, and Brazil. Even when treated, VL has a mortality estimated at 10 to 20%.^7,12 If left untreated, the mortality rate of VL approaches 95%.⁶

The treatment of leishmaniasis depends on several factors including the type of disease, concomitant pathologies, parasite species, and the geographic location.^1,6,7,13,14 Current antileishmanial drugs include the antimonides¹⁵ (e.g., sodium stibogluconate and meglumine antimonate), the bisamidines¹⁶ (e.g., pentamidine), liposomal amphotericin B,¹⁷ paromomycin,¹⁸ and miltefosine.¹⁹⁻²² Ideal new drugs for leishmaniasis should be orally bioavailable, active against all relevant species and strains, highly effective, minimally toxic, and inexpensive.²³⁻²⁶ Most current drug discovery efforts target either cutaneous²⁷ or visceral²⁸ disease and although existing medicines find widespread application in afflicted populations, none possess the desired properties.^6,8 Only miltefosine can be administered orally, the ideal route for countries with rural populations with limited health care access.¹⁹ All current drugs require lengthy treatment courses, and most are poorly tolerated or outright toxic.^1,7,29 Finally, resistance to the first-line antimonial drugs is common and resistance is emerging for the other classes of drugs.^30,31 Compounding these challenges, the different species and strains of Leishmania parasites exhibit distinct pathologies and differential susceptibility to novel chemotherapeutics.⁵ Therefore, we embarked on a drug discovery campaign focused on identifying novel early lead compounds for leishmaniasis possessing potential to be developed into an orally bioavailable and efficacious drug.³²

We previously reported a phenotypic high-throughput screen to identify new antileishmanial leads.²⁴ We have also pursued target-based approaches and the presence of microtubule-interacting³³ and vimentin-like proteins³⁴ in Leishmania parasites prompted our examination of the vimentin-targeting 3-arylquinolines (2, Figure 1), which were previously pursued as oncology leads.^35,36 We tested more than 100 existing 3-arylquinolines, at concentrations of 10 and 1 μM for 96 h, to determine if they affected proliferation of luciferase expressing L. mexicana intracellular amastigotes, a causative strain of cutaneous leishmaniasis, cultured in immortalized J774A.1 macrophages.³⁷ After lysis, host macrophage toxicity was assessed using the nucleic acid-binding dye SYBR Green I.³⁸ Inhibition of proliferation of nontransformed mammalian fibroblast cells (BJ fibroblast cells) was carried out in parallel (Supporting Information Table S1). After identifying hits, defined as > 70% inhibition at 1 μM, we carried out dose–response experiments (1 nM to 10 μM) to establish the potency of inhibiting proliferation (EC₅₀) of both the L. mexicana intracellular amastigotes and the host macrophages using the same test system. This study revealed several potent 3-arylquinolines (EC₅₀ < 250 nM) with significantly less potency against host macrophages (EC₅₀ > 3,000 nM), suggesting at least a 10-fold selectivity index (SI).³² These data allowed us to define preliminary structure–activity relationships (SAR) within the series and 3-[3-(N-methylindolyl)-N⁷,N⁷-dimethyl]-2,7-quinolinediamine (34) emerged as a potent, rapidly acting early lead also possessing significant potency against several patient-derived, antimony-resistant, strains of Leishmania donovani, a causative strain of VL. Next, we profiled selected analogs for in vitro ADME/Tox and in vivo pharmacokinetics in mice. Finally, we showed that compound 27, which exhibited reasonable in vivo tolerability, significantly impaired disease progression in a murine model of cutaneous leishmaniasis. Together, these results suggest that the arylquinolines may provide a novel and promising starting point for the development of orally bioavailable antileishmanial drugs.

Figure 1.

General synthetic route. A Friedländer condensation between aromatic ortho-aminobenzaldehydes 1 and aryl or heteroaryl acetonitriles in the presence of potassium tert-butoxide in N,N-dimethylformamide (90 °C for 1–3 h) afforded 3-arylquinolines 2 in good to excellent yields.

Chemistry

Synthesis of 3-Arylquinolines

Friedländer condensation³⁹ between various ortho-aminobenzaldehydes (1) and either arylacetonitriles or heteroarylacetonitriles (Figure 1), using potassium tert-butoxide in N,N-dimethylformamide at 90 °C for 1–3 h, produced 3-arylquinolines in reasonable yields. The poorly stable 2-amino-4-(N,N-dimethylamino)benzaldehyde was produced directly prior to use by the reduction of commercially available 4-(N,N-dimethylamino)-2-nitrobenzaldehyde with iron powder in hydrochloric acid. For the arylquinolines bearing C-7 heterocyclic groups, nucleophilic substitution reactions of 4-fluoro-2-nitrobenzaldehyde with either pyrrolidine, piperidine, and morpholine or N-methylpiperazine, followed by reduction with iron in hydrochloric acid, secured the required heterocyclic-substituted 2-aminobenzaldehydes. In all cases, the resulting free bases were treated with ethereal hydrochloric acid to afford more water-soluble salts than the corresponding free bases. All compounds utilized for biological assays had a structure confirmed by ¹H NMR, ¹³C NMR, and mass spectrometry (MS) data and purity greater than 95% as confirmed by ultraperformance liquid chromatography (UPLC)–MS and/or combustion analyses.

Results and Discussion

SARs: Initial Modifications

Our hit-to-lead strategy was to identify the pharmacophore and key structural drivers for potency (SAR), understand structural drivers for physiochemical properties such as solubility structure–property relationships (SPR), and evaluate metabolic and toxic liabilities inherent in the scaffolds. In exploring SAR, we adopted a strategy of systematically modifying the structural features embodied within 2 (Figure 2) in the following order: (1) the core quinoline system itself and (2) the pendent substituents at C-2, C-3, and C-7 positions on the quinoline ring. We tested the analogs to establish their growth inhibitory potency (EC₅₀) for both intracellular amastigotes and macrophages. To inspire confidence that efficacious plasma concentrations could be maintained without inducing gross toxicity in murine models, we initially sought to identify an arylquinoline with EC₅₀ < 250 nM for inhibiting proliferation of L. mexicana intracellular amastigotes and with a selectivity ratio against macrophages exceeding 25.³²

Figure 2.

Modifications of the quinoline pharmacophore (Supporting Information Table S2).

We replaced the core quinoline ring system with other related heterocycles (Figure 2 and Supporting Information Table S2), beginning with the C-3 ortho-fluoroarylquinoline 16, based on our previous work.^35,36 Our explorations (Figure 2 and Supporting Information Table S2) included introducing additional nitrogen atoms into the quinoline framework, as in 1,5-naphthyridine (3), 1,6-naphthyridine (4), and quinazoline (5); replacing the quinoline ring with another bicyclic heterocycle in 2-methyl-2H-pyrazolo[3,4-b]pyridine (6); and fusing an additional heterocyclic ring to the quinoline in imidazo[1,2-a]quinoline (7). Analogs 3, 4, and 5 exhibited little activity (≤25% inhibition at 10 μM). Analogs 6 and 7 possessed some activity but less potency than 2 (EC₅₀ > 1 μM). Therefore, we focused the remainder of our studies on the quinoline core.

SARs of 3-Arylquinolines: Modifications at C-2

Replacement of the C-2 amino group on the quinoline ring with a hydrogen (8), chloro (9), thiomethyl (10), morpholin-1-yl (11), or 4-methylpiperazin-1-yl (12) group afforded weakly potent (EC₅₀ > 1 μM) inhibitors of amastigote proliferation (Figure 3 and Supporting Information Table S2). Likewise, replacement of the C-2 amino group in the quinoline ring with either a carbonyl (13) or thiocarbonyl (14) group, as in quinolin-2(1H)-ones or quinoline-2(1H)-thiones, respectively, also gave weakly potent analogs. From these data, a requirement for the hydrogen bond-donating and electron-rich amine at the C-2 position was identified.

Figure 3.

Modifications of the C-2 quinoline substituent (Supporting Information Table S2).

SARs of 3-Arylquinolines: Modifications at C-3

Substitution of the C-3 aryl group with differentially substituted aryl groups and heteroaryl groups (Table 1) revealed minimal differences in potency (less than 2-fold) regardless of the electronic characteristics of the substituent. Likewise, we observed minimal changes in the SI (ratio of macrophage potency divided by intracellular amastigote potency) and none of the analogs met our target of 25-fold selectivity. Given the potential reduction of oxidative metabolism from halogenated analogs, we prioritized examining the chlorinated analogs that might afford a good balance between electronics and sterics without significantly compromising molecular weight. Subsequent studies showed that a chlorine atom could be placed in any position around the aryl ring without compromising potency. The meta-Cl (21) afforded the best SI (SI = 35). We selected 3-(3,5-dichlorophenyl)-N⁷,N⁷-dimethylquinoline-2,7-diamine (27), the first compound found with an EC₅₀ value below 250 nM and a good selectivity ratio (SI = 17) for further optimization.

Table 1. Modifications of the C-3 Aryl Group^a.

^aData represented as the mean of three replicates with errors reported as the standard deviation. Note: an error of 0.0 indicates identical values were obtained for each replicate.

Although reasonably potent and selective in cell-based assays, compound 27 was relatively hydrophobic (cLogP = 4.86), a finding that could limit its potential for in vivo studies.⁴⁰ Compounds with cLogP values >4 often exhibit rapid metabolic turnover, poor aqueous solubility, high plasma protein binding, and tissue accumulation.^41,42 They are also prone to receptor promiscuity and toxicity.⁴³ To address this issue, we tested whether the C-3 dichlorophenyl ring could be replaced by heterocycles while retaining potency and selectivity (Table 2). Smaller, five-membered heterocycles (28, 29, and 30) were significantly less potent than compounds 21 or 27. Isosteric six-membered heterocycles retained or slightly improved potency. For example, the unsubstituted 2-pyridyl analog (31) was equipotent to 27 and predicted to be significantly less hydrophobic (cLogP = 2.82). More sterically encumbered bicyclic heterocycles were also well tolerated, with the methyl indole (34) proved to be potent and selective (EC₅₀ = 120 nM; SI = 30). This SAR suggested that the target possesses a relatively deep and flexible hydrophobic pocket enveloping the C-3 substituent.

Table 2. Heterocyclic Modifications of the C-3 Aryl Group^a.

^aData represented as the mean of three replicates with errors reported as the standard deviation. Note: an error of 0.0 indicates identical values were obtained for each replicate. ND = not determined.

SARs of 3-Arylquinolines: Modifications at C-7

Replacement of the C-7 N,N-dimethylamino group allowed probing steric and electronic tolerances at that position while leaving the C-3 substituent fixed as either the 3,5-dicholoro or the N-methyl indole (Table 3). A secondary goal was to reduce crystallinity and hopefully improve kinetic aqueous solubility by incorporating a higher percentage of sp³ hybridized carbons.^44,45 Our early pharmacophore studies (Figures 2 and 3) suggested the presence of a polar amino substituent at C-7 was critical for potency. Incorporation of either a pyrrolidine or piperidine afforded roughly equivalent potency. However, incorporation of the N-methylpiperazine or morpholine groups reduced potency 2- to 4-fold, suggesting the pocket was slightly less tolerant of distal polarity (Table 3). Overall, the range of potencies exhibited was narrow (<3-fold) and the trends exhibited were mirrored by both C-3 substituents. This led to the conclusion that the C-7 substituent could be chosen to improve physiochemcial properties without reducing potency and selectivity.

Table 3. Modification of the C-7 Group^a.

^aData represented as the mean of three replicates with errors reported as the standard deviation. ND = not determined.

In Vitro Efficacy of Arylquinolines against Drug-Resistant L. donovani Strains

Arylquinoline 34, possessing good intracellular amastigote activity (EC₅₀ = 120 nM) and a 30-fold SI, is a representative, early lead. We tested compound 34 against several patient-derived strains of L. donovani, a causative strain of VL: strain BPK282 that is antimonial-sensitive; strain BPK275 that is moderately antimonial-resistant; and strain BPK173 that is highly antimonial-resistant.³⁰ Arylquinoline 34 retained activity against all resistant strains (Table 4 and Supporting Information Figure S1). Therefore, 34 is active against one of the causative strains of VL and does not appear to be subject to cross-resistance to antimony-based drugs.

Table 4. Assessment of Potency (EC₅₀) Against Antimony-Resistant L. Donovani Parasites^a.

	antimony potassium tartrate	amphotericin B	miltefosine	arylquinoline 34
cell line	EC50 (μM)	EC50 (μM)	EC50 (μM)	EC50 (μM)
L. donovaniBPK282	9.5 ± 5.4*	0.030 ± 0.026	6.2 ± 0.49	0.86 ± 0.25
L. donovani BPK275	18 ± 11*	0.027 ± 0.021	12 ± 2.1	0.71 ± 0.27
L. donovaniBPK173	350 ± 240*	0.023 ± 0.017	16 ± 0.71	0.66 ± 0.28
L. mexicana	1.2 ± 0.53	0.23 ± 0.14	2.8 ± 1.2	0.12 ± 0.090

^aAsterisks represent dose–response curves that are significantly different from each other for the three BPK lines, as determined using the sequential sum of squares F-test. Data represented as the mean of three replicates with errors reported as the standard deviation (Supporting Information Figure S1).

Time of Effect Experiments

A metric for assessing the in vitro efficacy of antiparasitic compounds is to measure how quickly parasite numbers decline after drug treatment.³² Rapidly acting drugs have the potential to relieve symptoms quickly and minimize the time-window for resistant parasite selection. To determine the time necessary to inhibit parasite growth, we exposed L. mexicana intracellular amastigotes to a range of concentrations of arylquinoline 34 for varying lengths of time (0.5, 2, 8.5, 24, 48, 72, and 96 h, Figure 4). Arylquinoline 34 rapidly inhibited the growth of intracellular amastigotes with an EC₅₀ of 1.0 μM after 0.5 h of incubation and appeared to approach its maximum effect by roughly 24 h of drug exposure.

Figure 4.

In vitro time of effect profiling of arylquinoline 34. Time-dependent EC₅₀ values were determined by exposing L. mexicana intracellular amastigotes to a range of concentrations of arylquinoline 34 for varying lengths of time (0.5, 2, 8.5, 24, 48, 72, and 96 h). Data represented as the mean of four replicates with errors reported as the standard deviation.

Mouse Liver Microsomal Metabolism

In order to finalize the selection of an early lead compound for in vivo studies, we characterized the metabolic stability of 34 and several related compounds using in vitro murine liver microsomes. These studies revealed that most analogs were rapidly metabolized (half-life, t_1/2 < 15 min, Table 5). In order to understand the primary sites of metabolism, we carried out UPLC–MS analysis of compounds 27 and 34 after microsomal incubation. These studies revealed that oxidative demethylation of the C-7 N,N-dimethylamino group to the corresponding monomethyl counterparts was the major site of metabolism. Further exploration with other analogs suggested the monomethyl compounds (39 and 44) were the most stable, having t_1/2 > 1 h.

Table 5. Mouse Microsomal Stability Studies with Selected Arylquinolines^a.

^aData represented as the mean of two replicates with errors reported as the standard deviation.

In Vivo Pharmacokinetics

To evaluate the potential of the early lead arylquinolines (27 and 34) in vivo and begin to establish an in vitro to in vivo correlation, we performed preliminary single intravenous dose pharmacokinetic studies in mice to determine circulating plasma concentrations of both the parent (27 and 34) and monomethyl compounds (39 and 44). Following a single intravenous (IV) administration of arylquinoline 27 (10 mg/kg) to mice (Figure 5), the plasma concentration reached a peak (C_max) of 0.74 μM, had an elimination half-life (t_1/2) of 2.3 h, and an AUC of 3.91 μM*h (Table 6). Compound 27 concentrations in plasma remained above its in vitro EC₅₀ of 0.22 μM for approximately 2 h. As the concentrations of the dimethyl amino (27) dwindled, the plasma concentration of the monomethyl metabolite (39) increased, suggesting it remained a primary metabolite in vivo and that the microsomal models were faithfully predicting metabolism. Following a single intravenous (IV) administration of arylquinoline 34 (10 mg/kg) in mice (Figure 5), the plasma concentration peaked (C_max) at 2.84 μM, the elimination half-life (t_1/2) was 1.29 h, and the AUC was 5.95 μM*h (Table 6). Compound 34 remained above its EC₅₀ of 0.12 μM for approximately 4 h. Again, as concentrations of the dimethyl parent (34) decreased, the plasma concentration of the monomethyl metabolite (44) increased. However, the relatively lower exposure of the monomethyl metabolite 44 suggests other routes of metabolism are also at play.

Figure 5.

In vivo pharmacokinetic profiling of arylquinolines 27 and 34. Murine pharmacokinetic studies for arylquinolines delivered intravenously at 10 mg/kg.

Table 6. Pharmacokinetic Parameters for Compounds 27 and 34 Based on Intravenous, Oral, and IP Administration in Mice^a.

arylquinoline	dose (mg/kg)	t1/2 (h)	Cmax (μM)	tmax (h)	AUC (μM*h)	CL (L/h/kg)	Vd (L/kg)
27 (IV)	10	2.32	0.741	NA	3.91	7.7	25.9
34 (IV)	10	1.29	2.84	NA	5.95	5.31	9.88
34 (PO)	10	0.68	1.00	0.1	0.921	33.4	33.6
34 (IP)	10	0.88	3.17	0.1	3.81	7.92	10.1

^aLegend: t_1/2 is the compound half-life in plasma; C_max is the maximum concentration; t_max is the time the compound takes to achieve the maximum plasma concentration; AUC is the area-under-the-curve; CL is the clearance; and V_d is the volume of distribution at the steady state for IV, and apparent volume of distribution for other routes.

Among the current standard of care agents for the treatment of leishmaniasis, only miltefosine is orally bioavailable. To begin to understand the bioavailability of the arylquinolines, we conducted single oral dose studies with compound 34 using a highly solubilizing formulation (10/10/40/39 EtOH/PG/PEG400/phosphate buffered saline (PBS) (pH 7.4) and 1% w/v HβCD). Following a single oral dose of 34 (10 mg/kg, Figure 5 and Table 6) to mice, compound 34 showed rapid absorption (t_max ∼ 0.1 h) with a C_max of 1.0 μM, a t_1/2 of 0.68 h, and an AUC of 0.921 μM*h. The plasma concentration remained above the EC₅₀ of 0.12 μM for about 2 h. The apparent oral bioavailability (15%) was lower than optimal and did not reach concentrations expected to deliver the desired efficacy. Therefore, intraperitoneal injection (IP) was also explored. IP injection achieved a much higher exposure than oral administration and C_max (3.17 μM) and overall AUC (3.81 μM*h) values were much closer to those observed after IV dosing.

Given the better exposures from the IP route than the oral route, we evaluated 10 sequential daily IP injections of 34 in a murine model. Unfortunately, daily IP administration of 34 in a 100% dimethyl sulfoxide (DMSO) solution led to significant adverse events including: rough hair, hunched posture, distended abdomen, lethargy, seizures, and bloody stool. These effects were not observed after repeated IP injection of compound 27. Thus, the hydrochloride salt of 27 was chosen for in vivo efficacy testing. The improved water solubility of this salt permitted dosing as a solution of 50% PEG400 in isotonic PBS rather than dosing as a DMSO solution.

In Vivo Pharmacodynamics

Arylquinoline 27 was tested for efficacy using an in vivo footpad murine model of cutaneous leishmaniasis.⁴⁶ Briefly, the footpad of BALB/c mice was injected with 10⁶L. mexicana parasites on day zero. Four weeks after inoculation, a palpable lesion was observed, and drug treatment was initiated. Cohorts of five animals were treated either with arylquinoline 27 (10 mg/kg) for 10 consecutive days administered by IP injection (50% PEG400 in PBS_7.4), or with vehicle alone as a negative control. Lesion size (determined by caliper) was measured for six additional weeks postdrug treatment (Figure 6). For vehicle-treated mice, the lesions grew steadily for 10 weeks, at which time the mice were euthanized. During the administration period, arylquinoline 27 fully inhibited the progression of lesion size, maintaining dimensions matching those at the time of initial dosing (i.e., at 4 weeks postinjections with L. mexicana). However, after drug treatment was withdrawn, the footpad lesions began to increase in size, eventually growing to a size roughly equivalent to the size seen for vehicle treatment.

Figure 6.

In vivo efficacy for controlling cutaneous lesion progression in the mouse. Mice (5 per cohort) were infected with L. mexicana promastigotes on day 0; by week 4 after infection, cutaneous lesions had grown to ∼0.1 mm width. Vehicle or arylquinoline 27 (10 mg/kg) was delivered daily by IP injection for 10 consecutive days. Measurements are plotted as the mean ± standard deviation. *p < 0.01.

Conclusions

Leishmaniasis comprises of a spectrum of diseases with significant unmet clinical need and the development of novel therapeutics for leishmaniasis remains particularly challenging. The current target product profile for new antileishmanial drugs requires oral bioavailability, activity against drug-resistant strains, excellent potency, high efficacy, minimal toxicity, and low cost.²³ We investigated whether the 3-arylquinolines³⁶ could provide a new starting point for the development of antileishmanial drugs. We tested more than 100 arylquinoline analogs at 1 μM and 10 μM for their ability to block growth of intracellular amastigotes, the disease-causing stage of the Leishmania parasite. We determined growth inhibition potencies (EC₅₀) against both the amastigotes and the host macrophages in which they reside for the most active compounds. These studies revealed several 3-arylquinolines with potent (EC₅₀ < 250 nM) antileishmanial activity and weaker potency (EC₅₀ > 3,000 nM) against macrophages, giving at least a 10-fold SI.

Initial studies defining the pharmacophore involved large structural perturbations to the quinoline core and demonstrated that roughly isosteric 6,6- and 6,5- heteroaryl cores were significantly less potent (Figure 2). We systematically explored the substituents at C-2, C-3, and C-7 positions on the quinoline ring. The aryl substituent at the C-3 position apparently occupies a relatively deep and flexible hydrophobic pocket on the target that can accommodate a range of electron-donating and electron-withdrawing substituents and six-membered heterocycles (Tables 2 and 3). However, there is a requirement for a reasonable steric bulk because smaller, five-membered heterocycles were significantly less potent (Table 3).

Examining physiochemical characteristics and metabolic stability of a small number of analogs revealed the current generation of arylquinolines has a high cLogP (>5) and are rapidly cleared in microsomal models of oxidative metabolism (Cl_int > 80 ml/min/kg) (Table 5). Metabolite identification studies after microsomal incubation revealed that a single demethylation of the N,N-dimethylamino group was the primary metabolic event. Fortuitously, the monomethyl analogs of 27 and 34 retained their parent’s amastigote growth inhibition potency. To mitigate the metabolism of the labile N,N-dimethylamino substituent, we prepared a series of C-7 related compounds containing amines of varying basicity and steric bulk, while attempting to increase solubility by introducing nonplanar cyclic systems to confer flexibility and decrease the propensity to form crystal lattices (Table 3). These studies revealed that the N,N-dimethylamino group could be replaced by the piperidine or pyrrolidine with retention of potency but unfortunately without improved microsomal stability. Therefore, arylquinoline 34 was chosen as a representative early lead with promising rapid intracellular amastigote activity (EC₅₀ = 120 nM) and a 30-fold selectivity over macrophage toxicity (EC₅₀ = 3.7 μM). Compound 34 was also shown to retain potency in both moderately and significantly drug-resistant, patient-derived L. donovani parasites (Table 4), demonstrating activity against causative strains of both cutaneous and VL.

To establish an in vitro to in vivo correlation, we selected two potent analogs, 27 and 34, for preliminary in vivo PK studies. A single low-dose (10 mg/kg) murine intravenous (IV) PK study of either 27 or 34 showed no gross toxicity but revealed rapid clearance and the formation of significant quantities of the monomethyl metabolites (Figure 5). Complementary single oral dosing studies suggest 34 has poor bioavailability (F < 20%). One explanation for the observed low bioavailability is that the compounds are rapidly metabolized by the liver during first-pass metabolism, effectively limiting the amount of drug that can reach systemic circulation. Although repeated IP injection of 34 proved toxic, delivery of 27 at 10 mg/kg IP daily for 10 days proved tolerable and led to partial control of an in vivo infection with L. mexicana. The sizes of cutaneous lesions, however, returned to those of untreated animals following cessation of treatment.

In conclusion, the 3-arylquinolines feature a compact size, good potency, a clear SAR, and synthetic accessibility that together with the reasonable murine exposure and partial control of disease progression exhibited by arylquinoline 27 suggest they can be considered viable early leads for leishmaniasis therapy. However, future studies must determine if the arylquinolines can mechanistically clear an infection or if the apparent stasis is a function of achieved compound concentrations. Future studies should also focus on understanding the mechanistic drivers of both amastigote growth inhibition and of the in vivo toxicity observed after repeated administration of 34, suppressing oxidative metabolism using microsomal models, and improving physiochemical properties and in vivo pharmacokinetics and pharmacodynamics.

Experimental Section

Chemistry

General Chemistry

Chemicals were purchased from Sigma-Aldrich (St. Louis, MO) or Fisher Scientific (Pittsburgh, PA) unless otherwise noted or were synthesized according to literature procedures. Solvents were obtained from commercial vendors without further purification unless otherwise noted. Nuclear magnetic resonance spectra were obtained on Varian instruments (¹H, 400 or 500 MHz; ¹³C, 100 or 126 Mz). Chemical shifts (δ) are reported in parts per million (ppm) with internal CHCl₃ (δ 7.26 ppm for ¹H and 77.0 ppm for ¹³C), internal DMSO (δ 2.50 ppm for ¹H and 39.5 ppm for ¹³C), or internal TMS (δ 0.0 ppm for ¹H and 0.0 ppm for ¹³C) as the reference. ¹³C NMR data are reported as follows: chemical shift, multiplicity (s = singlet, bs = broad singlet, d = doublet, t = triplet, q = quartet, p = pentet, sext = sextet, sep = septet, m = multiplet, dd = doublet of doublets, dt = doublet of triplets, td = triplet of doublets, and qd = quartet of doublets), coupling constant(s) (J) in Hertz (Hz), and integration. Low-resolution mass spectra were obtained using an Agilent 1100 (atmospheric pressure and chemical ionization) instrument. High-resolution mass data were obtained by direct infusion electrospray ionization MS using a LTQ-Orbitrap mass spectrometer coupled with a heated electrospray ionization (HESI-II) Probe (Thermo Fisher Scientific, Waltham, MA) and an FT analyzer at a resolution of 100,000. The reported m/z mass was a mean of 20 scans. Melting points were determined in open capillarity tubes with a Buchi B-535 apparatus and are uncorrected. Compounds were purified by chromatography on preparative layer Merck silica gel F254 unless otherwise noted. Combustion analyses were conducted at Atlantic Microlabs (Norcross, GA). All compounds were characterized for structural assignments and purity by a combination of ¹H and ¹³C NMR, MS data, and combustion analyses. All compounds utilized for biological assays had a confirmed structure and purity greater than 95%.

General Procedure for the Synthesis of Arylquinolines^35,36

To a solution of 2.38 mmol (1.3 equiv) of the appropriate benzyl cyanide in 3 mL of anhydrous N,N-dimethylformamide at 0 °C was added 2.38 mmol (1.3 equiv) of potassium tert-butoxide. The mixture was stirred for 15 min, and 1.83 mmol of appropriate 2-aminobenzaldehyde dissolved in 1 mL of anhydrous N,N-dimethylformamide was added dropwise at 0 °C. The mixture was warmed to 25 °C and stirred for 3–4 h at 90 °C. After cooling, the mixture was quenched by pouring into water to afford a precipitate that was collected and purified by recrystallization and/or chromatography as noted for individual compounds described below.

3-[3,5-Dichlorophenyl]-N⁷,N⁷-dimethylquinoline-2,7-diamine (27)

Purified by chromatography with methanol: dichloromethane (1:10) as the eluent three times: yield 88%, mp 169–171 °C. ¹H NMR, 400 MHz (DMSO-d₆): δ ppm 7.72 (s, 1H), 7.58 (t, J = 1.8 Hz, 1H), 7.52–7.49 (m, 3H), 6.86 (dd, J = 8.8 and 2.4 Hz, 1H), 6.59 (d, J = 2 Hz, 1H), 5.95 (s, 2H, NH₂), 3 (s, 6H). ¹³C NMR, 100 MHz, (DMSO-d₆): δ ppm 151.5, 149.1, 142, 137.1, 134.3, 128.4, 127.5, 126.7, 117.3, 115.4, 111.3, 103.3, 40. HRMS (ESI) calcd for C₁₇H₁₆³⁵Cl³⁵ClN₃ [MH⁺], 332.0716; found, 332.0708; calcd for C₁₇H₁₆³⁵Cl³⁷ClN₃ [MH⁺], 334.0747; found, 332.0661; calcd for C₁₇H₁₆³⁷Cl³⁷ClN₃ [MH⁺], 336.0629; found, 336.0657.

3-[3-(N-Methylindolyl)-N⁷,N⁷-dimethyl]-2,7-quinolinediamine (34)

Purified by chromatography using methanol: dichloromethane (1:10) as the eluent and recrystallized from acetonitrile: yield 45%, mp 152–155 °C. ¹H NMR, 400 MHz (DMSO-d₆): δ ppm 7.76 (s, 1H), 7.58 (s, 1H), 7.55 (d, J = 7.6, 1H), 7.52 (d, J = 8 Hz, 1H), 7.5 (d, J = 8.4 Hz, 1H), 7.23 (t, J = 8 Hz, 1H), 7.1 (t, J = 8 Hz, 1H), 6.86 (dd, J = 8.8 and 2.4 Hz, 1H), 6.65 (d, J = 2.4 Hz, 1H), 5.84 (s, 2H, NH₂), 3.86 (s, 3H), 3 (s, 6H). ¹³C NMR, 100 MHz, (DMSO-d₆): δ ppm 156.6, 151, 148, 136.8, 135.7, 128.7, 127.8, 126.5, 121.6, 119.4, 119.3, 115.8, 113.4, 111, 110.5, 110.1, 103.9, 40.3, 32.6. HRMS (ESI) calcd for C₂₀H₂₁N₄ [MH⁺], 317.1761; found: 317.1748. Anal. Calcd for C₂₀H₂₀N₄: C, 75.92; H, 6.37. Found: C, 75.77; H, 6.26.

Biology

Assays for Growth Inhibition of Leishmania Intracellular Amastigotes

RLuc L. mexicana and Hluc L. donovani-BPK stationary phase promastigotes were used to infect J774A.1 macrophages at a multiplicity of infection of 10:1 and 5:1, respectively. After overnight 16 h incubation, infected macrophages were washed three times in PBS to remove extracellular promastigotes and transferred to 96-well plates at a seeding density of 1 × 10⁴ parasites per well in 0.2 mL Minimal Essential Media (Gibco). Compounds (2 μL volumes in DMSO) were added to the parasites using serial 3-fold dilutions to cover a range of concentrations from about 10 μM to 1 nM. Cultures were incubated at 37 °C under a humidified 5% CO₂ atmosphere for 96 h for RLuc L. mexicana parasites or 72 h for L. donovani Hluc-BPK cells, unless otherwise noted for time of effect studies. Growth of intracellular amastigotes was measured using the Renilla luciferase assay system (Promega), as detailed previously³⁷ for RLuc parasites and ONE-Glo (Promega) for Hluc-BPK cells.

Cytotoxicity of compounds to J774.A1 macrophages was quantified separately; 10 μL of 100 X stock SYBR Green I (Sigma-Aldrich) in 10% Triton in PBS was added to 100 μL of lysed cells and fluorescence measured (Ex 497 nm; Em 520 nm) after 1 h incubation in the dark using a Wallac 1420 Victor2 Microplate Reader. Data were log transformed and EC₅₀ values were determined using GraphPad Prism 8 (GraphPad Software). In the absence of growth inhibitors or DMSO, the macrophages increased in number ∼6-fold over 96 h in Minimum Essential Medium, employed for both macrophage infections and the toxicity assays.

Animal Studies Statement

All animal studies carried out to support this work were executed under approved protocols governed by the respective Institutional Animal Care and Use Committees.

Efficacy Studies Using the Murine Model of Cutaneous Leishmaniasis

Female BALB/c mice (∼20 g) were injected in one hind foot pad with 1 × 10⁶ stationary phase promastigotes suspended in 25 μL of PBS. Four weeks after infection, when a small cutaneous lesion was visible in the injected footpad, cohorts of five mice were treated with either compound or vehicle alone (90 μL), delivered daily for 10 consecutive days by IP injection using a 20-gauge × 30 mm disposable plastic needle. Vehicle consisted of 50% PEG400 in isotonic PBS. The width of the footpad (top to bottom) was measured with calipers before injection of parasites (day 0) and weekly from weeks 4–10. The width of the uninfected contralateral footpad was also measured each week, and its width was subtracted from that of the infected footpad to determine the lesion size.

Glossary

Abbreviations

ADME

absorption, distribution, metabolism, and excretion

UPLC

ultraperformance liquid chromatography

EC₅₀

half-maximal proliferation inhibitory concentration

WHO

World Health Organization

SAR

structure–activity relationship

SPR

structure–property relationships

VL

visceral leishmaniasis

CL

cutaneous leishmaniasis

SI

selectivity index

L

Leishmania

t-BuOK

Potassium tert-butoxide

DMF

N,N-dimethylformamide

DMSO

dimethyl sulfoxide

t_1/2

apparent compound half-life in plasma

C_max

maximum concentration

t_max

time the compound takes to achieve the maximum plasma concentration

AUC

area-under-the-curve

CL

cutaneous leishmaniasis

V_d

volume of distribution at steady state for IV and apparent volume of distribution for other routes

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jmedchem.1c00813.

• Molecular formula strings and growth inhibition data for L. mexicana intracellular amastigotes, J774A.1 macrophages, and normal fibroblasts (BJ) (CSV)

• NMR (¹H & ¹³C) characterization, melting point, mass spectra, and elemental analysis data and experimental details regarding biological and in vivo studies (PDF)

Author Contributions

All authors participated in writing and gave their approval to the final version of the manuscript.

The authors declare no competing financial interest.