ICI 56,780 Optimization: Structural-Activity and Relationship Studies of 7-(2-phenoxyethoxy)-4(1H)-quinolones with Antimalarial Activity

Abstract

Malaria is estimated to have caused 584,000 deaths and 198 million cases of the disease globally in 2013. Though mortality rates are down 47% globally since 2000 and significant progress has been made in the quest for eradication, reported occurrences of resistance against current therapeutics threaten to reverse that progress. Recently, antimalarials which were once considered unsuitable therapeutic agents have been revisited to improve physicochemical properties and efficacy required for selection as a drug candidate. One such compound is 4(1H)-quinolone ICI 56,780, which is known to be a causal prophylactic that also displays blood schizontocidal activity against P. berghei. Rapid induction of parasite resistance in rodent malaria models however, stalled its further development. We have completed a full structure-activity relationship study on ICI 56,780 focusing on the reduction of cross-resistance with atovaquone for activity against the clinical isolates W2 and TM90-C2B. The optimization work focusing primarily on the 3-, 6-, and 7-positions of the 4(1H)-quinolone core. The best compound 16c, a 3-bromo-substituted 4(1H)-quinolone, was found to possess low nanomolar EC₅₀ values against the blood stages of the P. falciparum strains W2 and TM90-C2B, as well as low digit nanomolar inhibitory constant against P. berghei liver stages. Furthermore, this bromo compound 16c was also shown in an in vivo efficacy study to reduce the parasitemia by 61 % on day 6 post-exposure, whereas the original 4(1H)-quinolone displayed no inhibition.

Graphical abstract

Introduction

Malaria is considered endemic in 104 countries and it is estimated that 3.4 billion people are at risk for contracting malaria.^1–3 Although the number of deaths caused by malaria has decreased by 42% since the turn of the century, resistance to current treatments is a mounting problem.³ Therefore, there is an immediate need to develop new antimalarial agents, which ideally should be active against all developmental stages of the parasite within the host and within the mosquito vector. In humans, malaria is caused by P. falciparum, P. knowlesi, P. malariae, P. ovale, and P. vivax, of which P. falciparum is the most common cause for infections. Malaria begins its life cycle in a host when an infected female Anopheles mosquito takes a blood meal from a host. Sporozoites are injected from the salivary gland of the mosquito into the human host and first infect liver cells, which mature into schizonts. Next, schizonts rupture and release merozoites, which rapidly infect red-blood cells causing the clinical symptoms of the disease. In stark contrast to the most prevalent P. falciparum infections, P. vivax parasites can infect and stay dormant in liver cells and reemerge weeks, months or even years later causing a new infection.^4–6 Currently, the only effective treatment for the dormant liver stage of the parasite is primaquine, which is also used as a causal prophylactic agent.^7,8

Effective drug treatment remains the cornerstone of malaria control,¹ nevertheless WHO states that without new therapeutics, all the strides made in reducing the deaths from the disease could be reversed owing to resistance of parasite strains to many of the common antimalarials and artemisinin combination therapies (ACTs).³ Due to the limited chemotypes active against malaria, researchers have begun to optimize old antimalarial agents or drugs, evaluating these in current preclinical efficacy models and assessing these for proper physicochemical properties.^7,9,10 This approach has been shown to be effective for endochin, a 3-substituted 4(1H)-quinolone,^11–13 floxacrine, a dihydroacridinedione,^7,14–17 chloroquine,^10,18 and other chemotypes.⁷ 4(1H)-quinolone ester ICI 56,780^14,19,20 (6b) is an example of an old scaffold that is primed for recycling. This analogue was discovered in 1970 by Ryley and Peters to have blood schizontocidal activity against P. berghei (Pb) and prophylactic activity against P. cynomolgi infections.^20,21 Compound 6b was found to produce radical cures at 15 mg/kg however, resistance was observed after only one passage in Pb infected mice and the compound was eventually abandoned.²⁰ Studies by Manetsch, Kyle, Guy, Ward and O’Neill have shown that considerable optimization can be done to offset some of the liabilities of the 4(1H)-quinolone compound class.^22–25 However, these failed to address the fundamental questions about why 6b has such broad range anti-malarial activity or how structural modifications of 6b may reduce the propensity to induce resistance. Given the promise of this scaffold, detailed structure-activity relationship studies against the blood and liver stages of the parasite were conducted in order to gain a deeper understanding of this promising scaffold.

Results and Discussion

Synthetic Chemistry

Previous efforts to optimize 6b using Suzuki-Miyaura cross couplings yielded compounds with reduced antimalarial activities, suboptimal atovaquone cross-resistance indices, and poor physicochemical properties.²⁴ In our efforts to improve the bioavailability as well as to better understand the structural reasons for the broad spectrum activity of compound 6b, initially a focused structure-activity relationship (SAR) study was undertaken with a small number of analogues (Table 1). The goal was to individually assess the 3-ester, the 6-butyl and the 7-phenoxyethoxy groups’ contribution to the overall activity of 6b. Compound 6b was synthesized by an eight-step reaction sequence previously reported.²⁴ Analogue 6a was synthesized by the alkylation of acetamide 1a followed by acetyl deprotection and Gould-Jacobs cyclization. Compounds 6c and 6d were both prepared using Gould-Jacobs cyclization of the appropriate aniline along with malonate 4a (Scheme 1), while 8a-c, were synthesized by the removal of the ester moiety of their corresponding 4(1H)-quinolones 6a–c by a two-step reaction sequence of first refluxing in 10% sodium hydroxide then heating the intermediate acid in toluene at 270 °C for 5 minutes (Scheme 2). Due to limited commercial availability of highly substituted anilines, optimization of the 6- and 7-positions were conducted in parallel using two individual compound series focusing either on the 6-position or 7-position (see Tables 2 and 3). Compounds 6k–v and 8a were accessed through commercially available amines as opposed to 4(1H)-quinolones 6e–j, which were synthesized via a reaction sequence similar to the one of compound 6a. Nucleophilic substitution of N-acetyl-protected phenols 1 provided acetamides 2, which under basic conditions were deacetylated to anilines 3. Treatment of anilines 3 with malonate 4a followed by microwave-assisted cyclization furnished 4(1H)-quinolone esters 6 (Scheme 1).

Table 1.

Sequential removal of ICI 56,780 substituents^a

Compound	R1	R2	R3	EC50 W2 (nM)	EC50 TM90-C2B (nM)	EC50 Pb (nM)	EC50 J774 (μM)	RIb	LBIc	LBId
6b	(CH2)3CH3	O(CH2)2OPh	CO2CH3	0.039	7.89	0.052	>20.0	202	0.75	152
8b	(CH2)3CH3	O(CH2)2OPh	H	328	430	5.82	>20.0	1.31	56.4	73.9
8c	(CH2)3CH3	H	H	352	> 12400	395	>20.0	>35.2	0.89	>31.4
6c	(CH2)3CH3	H	CO2CH3	1.58	222	0.819	>20.0	141	1.93	271
8a	H	O(CH2)2OPh	H	5960	7680	>354	>20.0	1.29	<16.8	<21.7
6a	H	O(CH2)2OPh	CO2CH3	2.80	73.6	0.286	>20.0	26.3	9.79	257
6d	H	H	CO2CH3	200	6590	3.65	>20.0	33.0	54.8	1810

^aChloroquine (CQ), atovaquone (ATO), and dihydroartemisinin (DHA) are internal controls for each in vitro assay: CQ, 421 nM W2, 229 nM TM90-C2B and 47.23 mM J774; ATO, 1.4 nM W2, 18.4 nM TM90-C2B and 23.83 mM J774; DHA, 5.5 nM W2, 5.9 nM TM90-C2B and 1.53 mM J774.

^bRI = TM90-C2B/W2;

^cLBI = W2/Pb;

^dLBI = TM90-C2B/Pb.

Scheme 1.

Synthesis of 4(1H)-quinolones 6a-v

Scheme 2.

Synthesis of decarboxylated 4(1H)-quinolones 8a-c

Table 2.

Structure activity relationship of the 7-position^a

Compound	R1	EC50 W2 (nM)	EC50 TM90-C2B (nM)	EC50 Pb (nM)	EC50 J774 (μM)	RIb	LBIc	LBId
6d	H	200	6590	3.65	>20.0	33.0	54.8	1810
6a	O(CH2)2OPh	2.80	73.6	0.286	>20.0	26.3	9.79	257
6j	O(CH2)3Ph	0.711	>784	5.26	1.48	>1100	0.14	>149
6i	O(CH2)2Ph	139	7740	NDe	17.2	55.7	ND	ND
6h	OCH2Ph	1830	8080	ND	18.9	4.42	ND	ND
6k	OPh	16.6	1950	110	9.12	117	0.15	17.7
6m	Ph	2560	11300	ND	>20.0	4.41	ND	ND
6l	Cl	2360	5800	ND	>20.0	2.46	ND	ND
6n	F	26.2	395	>100	>20.0	15.1	<0.262	<3.95

^aChloroquine (CQ), atovaquone (ATO), and dihydroartemisinin (DHA) are internal controls for each in vitro assay: CQ, 421 nM W2, 229 nM TM90-C2B and 47.23 mM J774; ATO, 1.4 nM W2, 18.4 nM TM90-C2B and 23.83 mM J774; DHA, 5.5 nM W2, 5.9 nM TM90-C2B and 1.53 mM J774.

^bRI = TM90-C2B/W2;

^cLBI = W2/Pb;

^dLBI = TM90-C2B/Pb.

^eND: not determined.

Table 3.

Structure activity relationship of the 6-position^a

Compound	R1	EC50 W2 (nM)	EC50 TM90-C2B (nM)	EC50 Pb (nM)	EC50 J774 (μM)	RIb	LBIc	LBId
6d	H	200	6590	3.65	>20.0	33.0	54.8	1810
6c	(CH2)3CH3	1.58	222	0.819	>20.0	140	1.93	271
6o	OCH3	482	>10700	79.6	6.18	>22.2	6.05	>134
6p	OCH2CH3	292	>10100	ND	>20.0	>34.6	ND	ND
6q	CH(CH3)2	354	>10200	ND	>20.0	>28.8	ND	ND
6r	Cl	614	82200	46.6	>20.0	134	13.2	1760
6s	O(p-F)Ph	117	5710	ND	>20.0	48.8	ND	ND
6t	CH2CH3	145	>10800	ND	>20.0	>74.5	ND	ND
6u	Ph	2090	7180	>179	>20.0	3.44	<11.7	<40.1
6v	OPh	381	7060	ND	>20.0	18.5	ND	ND
6e	OCH2Ph	11.6	1100	127	7.59	94.8	0.091	8.66
6f	O(CH2)2Ph	24.6	3050	6.97	2.20	124	3.53	438
6g	O(CH2)3Ph	14.4	3670	11.1	17.1	255	1.30	331

^aChloroquine (CQ), atovaquone (ATO), and dihydroartemisinin (DHA) are internal controls for each in vitro assay: CQ, 421 nM W2, 229 nM TM90-C2B and 47.23 mM J774; ATO, 1.4 nM W2, 18.4 nM TM90-C2B and 23.83 mM J774; DHA, 5.5 nM W2, 5.9 nM TM90-C2B and 1.53 mM J774.

^bRI = TM90-C2B/W2;

^cLBI = W2/Pb;

^dLBI = TM90-C2B/Pb.

^eND: not determined.

To explore the optimal length of the alkyl group in 6-position of compound 6b, a series of 4(1H)-quinolones was prepared whose 6-position was substituted with a variety of groups ranging from one hydrogen to a pentyl chain. Compounds 6w-y were synthesized from 1-bromo-4-nitro-2-(2-phenoxyethoxy)benzene (11a) via 1-bromo-2-methoxy-4-nitrobenzene (9). Phenol 10a was synthesized from anisole 9 via a demethylation with BBr₃. Compound 10a was alkylated with (2-bromoethoxy)-benzene to give 11a, the needed starting material for the next reaction steps.²⁴ The alkynyl nitrobenzenes 12 were synthesized via Sonogashira coupling of the available alkynyl starting materials (Scheme 2). Nitrobenzene 11a was first subjected to a Sonogashira-cross coupling using TMS-acetylene with PdCl₂(PPh₃)₂, CuI, and TEA, then finally deprotected with TBAF to yield 12a.²⁶ To access the nitro alkyne 12b, a microwave-assisted deprotection and coupling reaction using trimethyl(prop-1-yn-1-yl)silane, Pd(OAc)₂, and TBAF was employed to give the desired product in low yields.²⁷ Lastly, analogue 12c was obtained in good yields using 1-pentyne, PdCl₂(PPh₃)₂, CuI, PPh₃, and diethylamine in a microwave reactor.²⁸ Each alkyne 12 was reduced using hydrogenation conditions in order to reduce both the alkyne and nitro groups to the corresponding alkyl-substituted anilines 3, which were subjected to standard Gould-Jacobs cyclization conditions furnishing the 4(1H)-quinolones 6w-y (Scheme 3).

Scheme 3. Synthesis of 4(1H)-quinolones 6w-y with varying alkyl group in 6-position.

^aR = H: i) TMS acetylene, PdCl (PPh 2 3)2, CuI, triethylamine, 85° C, 1 h; ii) TBAF, THF, rt, 5 min; R = methyl: trimethyl(prop-1-yn-1-yl)silane, Pd(OAc)₂, TBAF, THF, MW 120 °C, 15 min; R = propyl: 1-pentyne, PdCl₂(PPh₃)₂, CuI, PPh₃, diethylamine, THF, MW 120 °C, 25 min.

Next, a small set of 4(1H)-quinolones with electron donating or electron withdrawing groups were synthesized. This chemistry allowed access to 4(1H)-quinolones 6z, 6aa, 6ab, and 6ac substituted in 6-position with a bromo, methoxy, chloro, or methyl group. These compounds were synthesized starting from the alkylation of corresponding nitrophenols 10 followed by the reduction of the nitrobenzenes 11a–d to give anilines 3z–ac which were cyclized to give the desired 4(1H)-quinolones 6z–ac (Scheme 4).

Scheme 4.

Synthesis of 4(1H)-quinolones 6z-ac probing 6-position electronics

As the ester group in 3-position of compound 6b was consider to be a potential metabolic liability, 4(1H)-quinolone analogues were prepared in which the methyl ester group was replaced by amides, halides, a nitrile, a keto, an ethyl, or other esters. The amide functionalities were introduced by the aminolysis of 6b using trimethyl aluminum and the appropriate amine in dry benzene²⁹ to yield compounds 13a-e in moderate to low yield after purification using preparative HPLC (Scheme 5).

Scheme 5.

Synthesis of 3-amide substituted 4(1H)-quinolones 13a–e

The 3-cyano-4(1H)-quinolone was synthesized from aniline 3b and the commercially available ethyl-2-cyano-3-ethoxyacrylate (4b). Standard Gould-Jacobs cyclization of intermediate N-aryl enamine 5ad afforded the cyano-substituted 4(1H)-quinolone 6ad in low yield (Scheme 6).

Scheme 6.

Synthesis of 3-cyano substituted 4(1H)-quinolone 6ad

A three-step reaction sequence was utilized for the preparation of 3-ethyl-substituted 4(1H)-quinolone 6ae. First, ethyl butyrate was reacted with diethyl oxalate (14) to give diester 15,³⁰ which was used as the starting material along with the aniline 3b to furnish intermediate 5ae followed by cyclization to give 3-ethyl-substituted 4(1H)-quinolone 6ae (Scheme 7). 3-ethyl ester substituted 4(1H)-quinolone 6af was cyclized from enamine 5af following a reaction of aniline 3b with malonate 4c. (Scheme 8). Ketone 6ag was synthesized through a standard Conrad-Limpach cyclization (Scheme 9). 3-Iodo-4(1H)-quinolone 16b was an intermediate for the preparation of 3-aryl-substituted 4(1H)-quinolones previously reported.²⁴ The remaining halides 16a and 16c-16e were synthesized from 4(1H)-quinolone 8b or 2-methyl-4(1H)-quinolone 8d using NBS, NCS or NIS in DMF (Scheme 9).

Scheme 7.

Synthesis of 3-ethyl substituted 4(1H)-quinolone 6ae

Scheme 8.

Synthesis of 3-ethyl ester substituted 4(1H)-quinolone 6af

Scheme 9.

Synthesis of keto 4(1H)-quinolone 6ag

Antimalarial Activity and Cytotoxicity

All synthesized compounds were routinely tested against the clinically relevant multidrug resistant malarial strains W2 (pyrimethamine and chloroquine resistant strains) and TM90-C2B (mefloquine, chloroquine, atovaquone, pyrimethamine resistant strains) as previously reported.^11–13 Due to the emergence and rapid acquisition of cross resistance,³¹ each compound was also evaluated on its resistance index (RI), which is the ratio of the effective concentrations needed to kill 50% of the population of parasites (EC₅₀) for TM90-C2B and W2 strains (RI=EC₅₀ TM90-C2B/EC₅₀ W2). Ideally, the RI of a compound should lie between 0.3 and 3 in order to avoid rapidly inducing resistance in the parasite. The target for 4(1H)-quinolones is cytochrome b of the mitochondria; TM90-C2B has a Y268S mutation in cytochrome b that confers high grade resistance to atovaquone. Therefore, the RI values with TM90-C2B reflect potential cross-resistance to atovaquone. Selected compounds were also tested for in vitro liver stage activity using P. berghei sporozoites expressing luciferase, harvested from mosquito salivary glands and allowed to infect HEPG2 hepatoma cells in order to assess if the compounds possessed causal prophylactic activity.¹² Similar to the RI, liver blood indices (LBI) were assessed in order to relate activity against P. berghei with activity against W2 or TM90-C2B (LBI=EC₅₀ W2/EC₅₀ Pb or LBI= EC₅₀ TM90-C2B/EC₅₀ Pb). This allowed a simplified method for comparing the different assays being simultaneously run. Additionally, each compound was also tested for cytotoxicity using mammalian J774 cell lines in a 96 well plate format.^11–13,15

Structural Activity Relationships

The original compound 6b was found to have an excellent EC₅₀ for W2 at 39.0 pM, however it suffered from a two orders of magnitude drop in activity for the atovaqone resistant strain TM90-C2B with an EC₅₀ value of 7.89 nM. Compound 6b also showed excellent in vitro liver stage activity in P. berghei infected cells with an EC₅₀ of 52.0 pM. Initially, each substituent in 3-, 6-, and 7-position was removed sequentially to give compounds 6a, 6c and 8b. The removal of the 6-butyl or 7-(2-phenoxyethoxy) group left moderately active compounds 6a and 6c with EC₅₀ values against W2 of 2.80 and 1.58 nM, respectively, whereas both maintained subnanomolar liver stage activity. In contrast, removal of the ester functionality in 3-position severely affected the potency, dropping the EC₅₀ constants of W2, TM90-C2B and P. berghei to 328 nM, 430 nM, and 5.82 nM, respectively. These data suggest the importance of the ester in liver stage activity. Despite the severe potency reduction for compound 8b, it was noted that the resistance index equaled 1.31, which stands in stark contrast to the large RI values for compounds 6a and 6b. Subsequently, compounds 6d, 8a, and 8c were tested, in which two of the 3-, 6- and 7-substituents were simultaneously removed. Both 6d and 8c retained respectable potency against W2, however both were devoid of any activity against TM90-C2B. Similar to compound 8b, when the ester group was removed in compounds 8a, inhibition against W2 was lost, however the RI value converged towards 1. Importantly, it was also observed that compounds 8a, 8b, and 8c, lacking the ester group in 3-position, were the least potent analogues against P. berghei of the entire subseries (Table 1).

Due to the scarcity and costs of commercially available, di- and tri-substituted anilines along with the awareness that 3-ester-substituted 4(1H)-quinolones 6a and 6c retained moderate blood stage activity, a set of compounds were synthesized, in which the 7- or 6-position were probed (Tables 2 and 3). Considering possible metabolic instabilities of the 2-phenoxyethoxy substituent in 4(1H)-quinolone 6b, analogue 6j was designed in which one of the oxygens in the 7-(2-phenoxyethoxy) moiety was replaced by a methylene unit. Compound 6j, which in comparison to analogue 6a was 4 times more active against W2 with an EC₅₀ of 0.711 nM, displayed a greater than ten-fold drop in activity against TM90-C2B. This was mirrored in the P. berghei activities where there was an 18-fold drop in the activity of 6j as compared to the reference 6a, from EC₅₀s of 0.286 nM in 6a to 5.26 nM in 6j. The length of the alkyl chain linker was also shortened in hopes of improvement similar to 4(1H)-quinolone 6j, however, analogues 6i, 6h, and 6k were all less active than the reference 6a. Finally, in order to test electronic and steric effects, 4(1H)-quinolones substituted in the 7-position with phenyl, chloro and fluoro functionalities were tested. 7-Chloro and 7-phenyl substituents were not tolerated, as compounds 6l and 6m were devoid of antimalarial activity, whereas 7-fluoro-4(1H)-quinolone 6n was active against W2 and TM90-C2B with EC₅₀s of 26.2 nM and 395 nM, respectively.

A similar study was conducted investigating possible moieties in the quinolone’s 6-position. The reference compound 6c, which maintains the 6-butyl substituent, proved to be the most active compound of this set, with EC₅₀s of 1.58 nM for W2, 222 nM for TM90-C2B, and 0.819 nM for P. berghei. Among the other compounds tested, substituents derived from the 2-phenoxyethoxy group were shown to be the most potent analogues. Compound 6e, containing the benzyloxy substituent in the 6-position, was the most active analogue against W2, TM90-C2B and P. berghei with EC₅₀ constants equaling 11.6 nM, 1100 nM, and 127 nM, respectively. Compounds 6f and 6g, analogues which increase the alkyl chain linker length in comparison to 6e, were the only other analogues of this subseries displaying EC₅₀ with double digit nM activity against W2. Overall, with the exception of 6-phenyl-substituted 4(1H)-quinolone 6u, which was devoid of activity, probing of the six position gave mediocre compounds leading to the belief that it was overall less important than the 7-position.

In a subseries of close analogues, the optimal length of the 6-alkyl residue was investigated, leading to the design of compounds 6ac, 6w, 6x and 6y. 7-Methyl-substituted 4(1H)-quinolone 6ac had modest activity with an EC₅₀ of 2.01 nM for W2, an EC₅₀ of 499 nM for TM90-C2B, and Pb EC₅₀ of 152 nM. Antimalarial potency increased as the number of methylene units were added until the alkyl chain length reached four carbons and decreased as the chain length grew past four carbons. In general, all compounds in this subseries were potent with most displaying subnanomolar EC₅₀ values for W2, and with the 6-butyl-substituted 4(1H)-quinolone being the most potent.

In the same vein, analogues in Table 5 were designed to probe the same steric and electronic effects of the 6-position. 6-Chloro and 6-bromo compounds, 6ab and 6z, possessed moderate activity with 6ab having an EC₅₀ of 9.93 nM and 617 nM for W2 and TM90-C2B while 6z had EC₅₀s of 4.47 nM and 847 nM for W2 and TM90-C2B. Both 6-chloro and 6-bromo compounds, 6ab and 6z, also displayed subnanomolar P. berghei EC₅₀s of 0.063 nM and 0.336 nM, respectively. Interestingly, 6-methoxy-7-phenoxyethoxy-substituted 4(1H)-quinolone 6aa had an RI of 1.60 with good activities for both blood and liver stages. It was also observed by the direct comparison pairs between compound 6aa and its 7-phenoxythoxy-omitted analogue 6o, and between 6-chloro 4(1H)-quinolone 6ab and its 7-phenoxythoxy-omitted analogue 6r that the addition of the 7-(2-phenoxyethoxy) group increases the activity for W2 by >60-fold and >130-fold for TM90-C2B.

Table 5.

Structural activity relationship of 6-position using electron withdrawing and donating groups^a

Compound	R1	EC50 W2 (nM)	EC50 TM90-C2B (nM)	EC50 Pb (nM)	EC50 J774 (μM)	RIb	LBIc	LBId
6ac	CH3	2.01	499	152	>20.0	248	0.013	3.28
6aa	OCH3	6.93	11.1	17.5	>20.0	1.60	0.396	0.634
6ab	Cl	9.93	617	0.063	>20.0	62.1	158	9790
6z	Br	4.47	847	0.336	>20.0	189.5	13.3	2520

^aChloroquine (CQ), atovaquone (ATO), and dihydroartemisinin (DHA) are internal controls for each in vitro assay: CQ, 421 nM W2, 229 nM TM90-C2B and 47.23 mM J774; ATO, 1.4 nM W2, 18.4 nM TM90-C2B and 23.83 mM J774; DHA, 5.5 nM W2, 5.9 nM TM90-C2B and 1.53 mM J774.

^bRI = TM90-C2B/W2;

^cLBI = W2/Pb;

^dLBI = TM90-C2B/Pb.

Following the 6- and 7-positions, the 3-position was probed. A possible metabolite of ester 6b is acid 7b. Therefore it was synthesized in order to see whether it could be the active form of the compound. With activities of 64.0 nM for W2, 1110 nM for TM90-C2B and 0.267 nM for Pb, it was not as active as its parent, however, all activity was not lost. Ethyl ester 6af showed potent activity, with a W2 EC₅₀ constant of 0.143 nM and was active against P. berghei with a subnanomolar EC_50. Converting the ester to an amide dropped activity slightly. Methylamide 13a was the most active at 13.3 nM for W2 and 3,980 nM for TM90-C2B. As the alkyl chain increased in size, activity decreased with isopropyl amide 13e being completely inactive with EC₅₀ constants in the μM range. Also, when each amide is compared to its ester counterpart, the ester is in all cases significantly more active. Interestingly, 2-cyano-substituted compound 6ad had W2 activity in the high double digit nM range. When the ester was converted to ketone 6ag, the activity dropped by almost 250-fold for W2 moving to 9.67 nM, the change for TM90-C2B was not as pronounced, however, with only a 48-fold difference. Lastly, complete removal of the ketone leaving just an alkyl chain, as with compound 6ae, was potent, having an EC₅₀ of 0.440 nM for W2, 1.92 nM for TM90-C2B, and 2.65 for P. berghei. With an RI of 4.36, it showed once again that removal of the ester left compounds with improved RIs (Table 6).

Table 6.

Structural activity relationship of the 3-position using ester isosteres^a

Compound	R1	EC50 W2 (nM)	EC50 TM90-C2B (nM)	EC50 Pb (nM)	EC50 J774 (μM)	RIb	LBIc	LBId
8b	H	328	430	5.82	>20.0	1.31	56.4	73.9
7b	CO2H	64.0	1110	0.267	>20.0	17.3	240	4160
6b	CO2CH3	0.039	7.89	0.052	>20.0	202	0.75	152
6af	CO2CH2CH3	0.143	100	0.007	>20.0	699	19.6	14300
6ae	CH2CH3	0.440	1.92	2.65	0.360	4.36	0.166	0.725
6ad	CN	97.2	3170	6.09	>20.0	32.6	15.7	521
13a	CONHCH3	13.3	3980	2.87	>20.0	299	4.63	1390
13b	CON(CH3)2	56.7	>6120	122	15.3	>108	0.46	>27.7
13c	CONHCH2CH3	206	>6120	ND	>20.0	>29.7	ND	ND
13d	CONH(CH2)2OH	253	5890	12.8	>20.0	23.3	19.8	460
13e	CONHCH(CH3)2	3340	5140	ND	>20.0	1.54	ND	ND
6ag	COCH3	9.67	378	ND	ND	39.1	ND	ND

^aChloroquine (CQ), atovaquone (ATO), and dihydroartemisinin (DHA) are internal controls for each in vitro assay: CQ, 421 nM W2, 229 nM TM90-C2B and 47.23 mM J774; ATO, 1.4 nM W2, 18.4 nM TM90-C2B and 23.83 mM J774; DHA, 5.5 nM W2, 5.9 nM TM90-C2B and 1.53 mM J774.

^bRI = TM90-C2B/W2;

^cLBI = W2/Pb;

^dLBI = TM90-C2B/Pb.

^eND: not determined.

3-Iodo-substituted 4(1H)-quinolone 16b was synthesized with the intention of replacing the 2-methylester substituent with a polarizable group. Interestingly, its testing gave an active compound having EC₅₀s of 1.23 nM for W2 and 6.04 nM for TM90-C2B with an RI of 4.91. Other 3-halo analogues were made, with and without a 2-methyl group. 3-Iodo-2-methyl-4(1H)-quinolone 16a was approximately 2-fold more potent than its reference 16b. The 3-bromo 4(1H)-quinolone 16d and its 2-methyl analogue 16c were slightly less potent than their 3-iodo counterparts with W2 activities of 4.64 nM and 2.60 nM, respectively, and TM90-C2B activities of 48.7 nM and 12.2 nM, respectively. 3-Chloro-2-methyl analogue 16e had EC₅₀s of 6.92 nM for W2 and 67.7 nM for TM90-C2B with an RI of 9.79. Overall, the 3-halo-substituted subseries followed the trend that the 2-methyl substituted analogues were approximately two-fold more active than the 2-hydrogen analogues and with RIs of 4.69 to 10.5, all five 3-halo substituted PEQs 16a–16e consistently had the best RI values identified thus far.

Correlation Between Blood Stage Activity Liver Stage Activity

In general, 7-phenoxyethoxy-substituted 4(1H)-quinolones display potent in vitro liver stage activity with low-nanomolar to sub-nanomolar EC₅₀ constants. The most potent 4(1H)-quinolone esters 6a, 6b, 6c, 6y, and 6af, as well as the medium potent compounds follow an activity order similar to the ranking observed for W2 and TM90-C2B, whereby the EC₅₀ values for TM90-C2B are approximately 20-200-fold higher than the EC₅₀ constants for W2 or in vitro liver stage activity. This observation is also supported by the LBI values, which are 0.1-10 for W2 and 2-2000 for TM90-C2B. In contrast, the 3-halo-substituted 7-(2-phenoxyethoxy)-4(1H)-quinolones 16a-16e with low nanomolar EC₅₀ constants for W2, TM90-C2B, and in vitro liver stage activity, provide RI and LBI values in the ranges of 1–10.

In vivo Efficacy Evaluation of Selected Compounds

The 12 4(1H)-quinolones with potent in vitro activity against blood stages of P. falciparum were chosen to be screened for in vivo efficacy according to a previously reported P. berghei infected mouse model (Table 8).¹¹ Criteria such as in vitro liver stage activity, RIs, and LBIs were considered for the selection of the screening candidates 6b, 6e, 6f, 6g, 6j, 6k, 6i, 6ab, 6ae, 13d, 16c and 16b. The screening was performed by treating two infected mice with a single 50 mg/kg oral dose of test compound suspended in PEG400 on day 1 post-exposure (PE) followed by assessing parasitemia on days 3 and 6 PE. Compounds with a parasitemia reduction of greater than 50% on days 3 and 6 PE were considered to be active. The original compound 6b was one of the most active candidates in the group with 76.9% inhibition on day 3 PE and 49% inhibition on day 6 PE (Table 8). Quinolone ester 6j and 3-bromo-4(1H)-quinolone 16c were the only other two compounds that displayed a parasitemia reduction similar to the original compound 6b with 47.1% and 49.5% inhibition on day 6 PE. Analogues 6g and 6ae with >65% inhibition on day 3 PE were considered promising, nevertheless the lack of significant parasitemia reduction on day 6 PE suggested that these compounds may suffer from insufficient bioavailability. A similar outcome has been observed for the previously reported 3-aryl-substituted 6-chloro-7-methoxy-4(1H)-quinolones compound series, for which a long half-life has been detrimental for the development of curative antimalarials.¹¹

Table 8.

Results of the In Vivo Efficacy Scout Screening^a

Compound	% Inhibition Day 3 PE	% Inhibition Day 6 PE	Compound	% Inhibition Day 3 PE	% Inhibition Day 6 PE
6b	76.9	49.0	6i	60.0	<1
6e	46.2	<1	6ab	<1	<1
6f	7.69	<1	6ae	69.2	17.7
6g	76.9	<1	13d	40.0	<1
6j	69.2	47.1	16c	100	49.5
6k	40.0	29.8	16b	53.9	19.6
Amodiquine	95.5	99.9	Atovaquone	96.3	99.8

^apercent inhibition compared to untreated animals

The two frontrunner compounds, 6b and 16c, displaying the best in vivo efficacy were subjected to a more rigorous in vivo efficacy testing in a modified Thompson test.^11–13,15 Using mice infected with 1 × 10⁶ Plasmodium berghei-GFP parasites, compounds were dosed orally on days 3, 4, and 5 at a 10 mg/kg concentration of compound suspended or dissolved in HEC/Tween. On days 3, 6, 9, 13, 21, and 30 PE, parasitemia was monitored by Giemsa-stained blood smears. Both compounds were evaluated by the reduction of parasitemia on day 6 PE and the survival up to day 30 PE. Compound 6b showed no inhibition on day 6 PE and mice were euthanized on day 13, the same day as the controls. This result is most likely due to the poor solubility and poor metabolic stability of 6b. The second compound, 16c, showed 61% inhibition on day 6 PE. In comparison to the original compound, 6b, the bromo substituent in the 3-position of 4(1H)-quinolone 16c probably increased the in vivo stability improving the exposure and thus the in vivo efficacy. Nevertheless, parasitemia rapidly rebounded so that mice dosed with compound 16c were euthanized on day 13 PE.

Conclusion

Due to the promising results of past recycling of old antimalarials and the promise of 3–4(1H)-quinolone esters, we synthesized a library of 46 4(1H)-quinolones with a variety of substituents focusing primarily on the 3-, 6-, and 7-positions. All compounds were tested in vitro against the two clinically relevant P. falciparum strains W2 and TM90-C2B. Furthermore, selected compounds with promising activity against W2 and TM90-C2B were evaluated for in vitro liver stage activity against P. berghei. Overall, this 4(1H)-quinolone series displays potent erythrocytic and exoerythrocytic activity with many compounds displaying low nanomolar EC₅₀ constants. The vast majority of the 4(1H)-quinolones were also shown to be nontoxic in J774 cytotoxicity assays with EC₅₀s > 20 μM. However, most compounds and their in vitro activity were not in vivo efficacious most likely due to poor aqueous solubility and low oral bioavailability.

There was, however, invaluable data gained from the analogues synthesized as it was shown that the 3-ester group is very important in addressing atovaquone cross resistance since 3-halide compounds 16a – 16e had excellent RI values. We also determined that the seven position was the most important for blood stage activities, however, there could be some synergy between the 6- and 7-positions due to the combination of the 6- and 7-substituted compounds being much more active than the 6- or 7-substituted compounds separately. Compound 6b and 16c were both tested in vivo using a modified Thompson test. There was a marked increase in the parasite inhibition for compound 16c over 6b on day 6 PE showing significant improvements over the original compound 6b. Although bromo-substituted 4(1H)-quinolone 16c showed less in vitro blood and in vitro liver stage activity than the predecessor 6b, the in vivo activity of 16c on day 6 PE clearly proved its superiority overall. The blood and liver stage activities of these quinolone esters along with the discovery of the 3-halo-4(1H)-quinolones lead us to postulate that there is still much potential in this scaffold and further optimization of solubility and stability could lead to orally bioavailable, curative agents with activity across the full spectrum of malaria life cycle stages.

Figure 1.

Antimalarial structures primed for recycling

Scheme 10.

Synthesis of 3-halide substituted 4(1H)-quinolones 16a-e

Table 4.

Optimal aliphatic chain length in 6-position^a

Compound	R1	EC50 W2 (nM)	EC50 TM90-C2B (nM)	EC50 Pb (nM)	EC50 J774 (μM)	RIb	LBIc	LBId
6a	H	2.80	73.6	0.286	>20.0	26.3	9.79	257
6ac	CH3	2.01	499	152	>20.0	248	0.013	3.28
6w	CH2CH3	0.461	65.3	0.737	>20.0	142	0.626	88.6
6x	(CH2)2CH3	0.157	42.4	<0.060	>20.0	270	>2.62	>707
6b	(CH2)3CH3	0.039	7.89	0.052	>20.0	202	0.75	152
6y	(CH2)4CH3	0.183	68.7	<0.56	>20.0	375	>0.33	>123

^aChloroquine (CQ), atovaquone (ATO), and dihydroartemisinin (DHA) are internal controls for each in vitro assay: CQ, 421 nM W2, 229 nM TM90-C2B and 47.23 mM J774; ATO, 1.4 nM W2, 18.4 nM TM90-C2B and 23.83 mM J774; DHA, 5.5 nM W2, 5.9 nM TM90-C2B and 1.53 mM J774.

^bRI = TM90-C2B/W2;

^cLBI = W2/Pb;

^dLBI = TM90-C2B/Pb.

Table 7.

3-Halo substituted 4(1H)-quinolones with and without 2-methyl groups^a

Compound	R1	R2	EC50 W2 (nM)	EC50 TM90-C2B (nM)	EC50 Pb (nM)	EC50 J774 (μM)	RIb	LBIc	LBId
16a	I	CH3	0.867	5.87	ND	ND	6.77	ND	ND
16b	I	H	1.23	6.04	2.55	ND	4.91	0.482	2.37
16c	Br	CH3	2.60	12.2	2.12	ND	4.69	1.23	5.75
16d	Br	H	4.64	48.7	18.7	ND	10.5	0.248	2.60
16e	Cl	CH3	6.92	67.7	11.2	ND	9.79	0.616	6.03

^aChloroquine (CQ), atovaquone (ATO), and dihydroartemisinin (DHA) are internal controls for each in vitro assay: CQ, 421 nM W2, 229 nM TM90-C2B and 47.23 mM J774; ATO, 1.4 nM W2, 18.4 nM TM90-C2B and 23.83 mM J774; DHA, 5.5 nM W2, 5.9 nM TM90-C2B and 1.53 mM J774.

^bRI = TM90-C2B/W2;

^cLBI = W2/Pb;

^dLBI = TM90-C2B/Pb.

^eND: not determined.

Table 9.

Results of Thompson Test

Compound	Dose (mg/kg)	Inhibition (%) Day 6 PEa	Avg. Day of Death
6b	10	<1	13
16c	10	61	13
atovaquone	10	99	30

^apercent inhibition as compared to untreated control animals

Abbreviations

Ac

acetyl

ACT

artemisinin combination therapy

DMF

N,N-dimethylformamide

EC₅₀

half maximal effective concentration

ED₅₀

half maximal effective dose

HPLC

high performance liquid chromatography

LBI

liver blood index

NBS

N-bromosuccinimide

NCS

N-chlorosuccinimide

ND

not determined

NIS

N-iodosuccinimide

Pb

P. berghei

PE

post exposure

RI

resistance index

r.t

room temperature

SAR

structure-activity relationship

SPR

structure-property relationship

TBAF

tetra-n-butylammonium fluoride

TEA

triethylamine

TMS

trimethylsilyl

WHO

World Health Organization