Identification, design and biological evaluation of bisaryl quinolones targeting Plasmodium falciparum type II NADH:quinone oxidoreductase (PfNDH2).

A program was undertaken to identify hit compounds against NADH:ubiquinone oxidoreductase (PfNDH2), a dehydrogenase of the mitochondrial electron transport chain of the malaria parasite Plasmodium falciparum. PfNDH2 has only one known inhibitor, hydroxy-2-dodecyl-4-(1H)-quinolone (HDQ), and this was used along with a range of chemoinformatics methods in the rational selection of 17 000 compounds for high-throughput screening. Twelve distinct chemotypes were identified and briefly examined leading to the selection of the quinolone core as the key target for structure-activity relationship (SAR) development. Extensive structural exploration led to the selection of 2-bisaryl 3-methyl quinolones as a series for further biological evaluation. The lead compound within this series 7-chloro-3-methyl-2-(4-(4-(trifluoromethoxy)benzyl)phenyl)quinolin-4(1H)-one (CK-2-68) has antimalarial activity against the 3D7 strain of P. falciparum of 36 nM, is selective for PfNDH2 over other respiratory enzymes (inhibitory IC(50) against PfNDH2 of 16 nM), and demonstrates low cytotoxicity and high metabolic stability in the presence of human liver microsomes. This lead compound and its phosphate pro-drug have potent in vivo antimalarial activity after oral administration, consistent with the target product profile of a drug for the treatment of uncomplicated malaria. Other quinolones presented (e.g., 6d, 6f, 14e) have the capacity to inhibit both PfNDH2 and P. falciparum cytochrome bc(1), and studies to determine the potential advantage of this dual-targeting effect are in progress.

Introduction

Drug resistance to currently deployed, established antimalarials such as chloroquine is driving the rise in global mortality due to malaria.[1] Malaria is responsible for roughly one million deaths annually,[2] and as such novel inhibitors active against new parasite targets are urgently required in order to sustain and develop treatments against malaria.[3] To this end, a program was undertaken to identify hit compounds active against the electron transport chain (ETC) of Plasmodium falciparum and specifically against NADH:ubiquinone oxidoreductase (PfNDH2).

PfNDH2 is a single subunit 52 kDa enzyme involved in the redox reaction of NADH oxidation with subsequent quinol production.[4] Localized in the mitochondrion, PfNDH2 is a principal elctron donor to the ETC, linking fermentative metabolism to the generation of mitochondrial electrochemical membrane potential (Δψm), an essential function for parasite viability (Figure 1).[4] Targeting the electron transport chain of the mitochondrion is a proven drug target as demonstrated by the drug atovaquone, targeting the cytochrome bc1 complex.[5]

Figure 1

Mitochondrial electron transfer chain and the role of PfNDH2 and bc1. Schematic representation of the respiratory chains of P. falciparum and M. tuberculosis. The chain components are (i) P. falciparum: PfNDH2 – type II NADH:quinone oxidoreductase, DHODH – dihydroorotate dehydrogenase, G3PDH – glycerol-3-phosphate dehydrogenase, MQO – malate:quinone oxidoreductase, SDH – succinate dehydrogenase, bc1 – cytochrome bc1 complex, c – cytochrome c, aa3 – cytochrome c oxidase and the F1Fo-ATPase (Complex V).

In order to identify hit compounds, we employed a range of ligand-based chemoinformatics methods in the rational selection of approximately 17 000 compounds that were predicted to possess activity against PfNDH2. The chemoinformatics approach were initiated from the identity of only one inhibitor of the target, hydroxy-2-dodecyl-4-(1H)-quinolone (HDQ)[6] and used molecular fingerprints,[7] turbo similarity,[8] principal components analysis, Bayesian modeling,[9] and bioisosteric[10] replacement in order to select compounds for high-throughput screening (HTS). The compounds were selected from a commercial library of ∼750 000 compounds (Biofocus DPI) and were predicted to possess favorable absorption, distribution, metabolism, excretion, and toxicity (ADMET) characteristics.[11] The selected compounds were subject to a sequential high-throughput screening methodology using an in vitro assay against recombinant PfNDH2 as described previously.[6] Hit confirmation and potency determination revealed over 40 compounds with IC50 values ranging from below 50 nM to 40 μM. Analysis of these hits revealed that only two of the compounds were selected by more than one chemoinformatic method, justifying the use of several virtual screening approaches. Seven distinct chemotypes were identified from the hit compounds and were thus primed for development as new agents against malaria (see Supporting Information). All 12 distinct chemotypes were briefly examined and key compounds were synthesized, and this led to the selection of the quinolone core as one of the main target chemotypes for structure–activity relationship (SAR) development due to its HDQ-like structure (Figure 2).

Figure 2

Mono 2-aryl quinolones emerging from quinolone hits identified in high-throughput screen and initial SAR performed on template.

Quinolones identified from the HTS were not considered appropriate for further optimization (see CDE204758 and CDE264055) but given the high potency of hit CDE021056, versus PfNDH2, we selected 2-substituted monoaryl quinolones as a core template with potential for SAR development (Note that several low micromolar saturated quinolones, e.g., CDD038715, were identified in this screen). The rationale for selection of the 2-aryl quinolone pharmacophore was to introduce additional lipophilicity in a region where HDQ contains the flexible aliphatic side chain. Subsequently, further extension of the side chain was performed, so it is more HDQ-like, while incorporating functionality to impart metabolic stability within the analogue series, and this approach led to eventual identification of early lead compounds for this series. In terms of SAR, the nature of the group at 3-position, the electronic/steric effect of substituents placed at the 5, 6, and 7 positions, the presence of a nitrogen in the A ring of the quinolone core, and changing from NH to NOH (as in HDQ) were all examined (Figure 2).

Results

Having identified mono aryl quinolones as hits against the target PfNDH2 (ca. 50–250 nM, e.g., 14a and 15a) with moderate activity in the whole cell phenotypic screen, our efforts were initially concentrated on the synthesis of a small number of additional analogues to see if activity could be improved further. The first structural alteration was to introduce a methyl substituent at the three position (e.g., 6a); this manipulation twists the 2-aryl side-chain, altering the torsion angle (Figure 3) leading to a subsequent reduction in aggregation. Aggregation via π-stacking of aromatic ring systems leads to higher melting points,[12] which has been shown to be closely related to solubility.[13] Molecular modeling was performed in order to analyze the relationship between melting point and the conformational effect of introducing a methyl or chloro group at the three position of the quinolone. Monte Carlo simulations were performed in order to sample the thermally accessible conformations and calculate the Boltzmann weighted average torsion angle that best described the planarity of the 2-position aryl ring with respect to the quinolone ring (see Figure 3 for torsional angle and Supporting Information for computational details).

Figure 3

The torsion angle that best represents the planarity of the 2-aryl group with respect to the quinolone core.

The melting point and computed Boltzmann weighted average torsion angle were examined for four pairs of compounds: 6a and 14a, 6b and 14a, 14c and 15a, and finally 6d and 14e; compounds within a pair are close analogues of each other with one compound incorporating a hydrogen substituent at the 3-position and the other either a methyl or chloro group. Higher melting points within a pair were found to correlate with a hydrogen substituent at the 3-position; conversely, lower melting points correlated with the presence of more bulky methyl or chloro groups. Hydrogen substituted compounds were found to have lower computed thermally accessible torsional angles than their methyl-substituted counterparts, as exemplified by compounds 6a and 14a (Figure 4, see Supporting Information for information for all pairs of compounds). This analysis supports the hypothesis that the solubility is related to the planarity/π-stacking propensity of 2-aryl substituted quinolones.

Figure 4

Lowest energy conformations for compounds 6a and 14a. Carbon, hydrogen, nitrogen, oxygen, and fluorine atoms are depicted in dark gray, off-white, blue, red, and pale-yellow respectively. The accompanying table shows the corresponding melting points and computed Boltzmann weighted average torsional angles. Images were produced in the Spartan ’08 Version 1.0.0 (Wavefunction Inc., Irvine, CA, USA).

Other structural modifications investigated for the mono aryl series were the presence of a nitrogen within the A ring of the quinolone core, altering the phenyl substituent and using H or Cl at the 3-position. From preliminary testing against the 3D7 strain of P. falciparum, it rapidly became apparent that activities below 500 nM versus the 3D7 strain could not be achieved. The chemistry employed to synthesize these compounds and their structures are covered in Schemes 1–4 and Tables 1–4 which will be described in detail for the subsequent bisaryl compounds.

Scheme 1

Synthesis of Quinolones 6a–w

Scheme 4

Synthesis of Quinolones 14a–m

Table 1

Yields for the Synthesis of Compounds 6a–w

compound	R	X	% yield 2	% yield 3	% yield 5	% yield 6
6a	-PhpCF3	H			45	32
6b	-PhpOCF3	H			45	32
6c	-PhpCH2Ph	H	72	94	45	42
6d	-PhpCH2PhpOCF3	H	99	99	45	42
6e	-PhmCH2PhpOCF3	H	78	82	45	30
6f	-PhpCH2PhpF	H	97	97	45	20
6g	-PhpCH2PhpOMe	H	71	90	45	32
6h	-PhpCH2PhpOCF3	6-CF3	99	99	51	52
6i	-PhpCH2PhpOCF3	7-CF3	99	99	55	24
6j(CK-2-68)	-PhpCH2PhpOCF3	7-Cl	99	99	58	30
6k	-PhpCH2PhpOCF3	6-Cl, 7-F	99	99	41	34
6l	-PhpCH2PhpOCF3	6-F, 7-Cl	99	99	30	27
6m	-PhpCH2PhpOCF3	5-OMe	99	99	98	8
6n	-PhpCH2PhpOCF3	6-OMe	99	99	48	28
6o	-PhpCH2PhpOCF3	7-OMe	99	99	21	29
6p	-PhpCH2PhpOCF3	8-OMe	99	99	30	27
6q	-PhmCH2PhpOCF3	6-Cl	78	82	47	20
6r	-PhmCH2PhpOCF3	7-Cl	78	82	58	30
6s	-PhpCH2PhpF	7-Cl	97	97	58	16
6t	-Ph2FpCH2PhpOCF3	7-Cl	88	88	58	35
6u	-PhpOPhpOCF3	H		63a	45	28
6v	-PhpOPhpOCF3	7-Cl		63a	58	32
6w	-PhpOPhpCl	H	84	91	45	8.5

Alternative route please see Supporting Information.

Table 4

Yields for the Synthesis of Compounds 15a–c

compound	R	% yield 15
15a	OCF3	60
15b	OMe	53
15c	CH2PhpOCF3	64

Being cognizant of the HDQ inhibitory activity against PfNDH2, compounds were designed with an extended side chain. In order to avoid the metabolically unstable HDQ side chain, a bisaryl group was chosen to mimic this side chain but maintain metabolic stability. The first series of compounds contains a methyl group at the 3-position (Scheme 1 and Table 1). Bisaryl compounds with a CH2 and O linker were investigated with the nature of the terminal phenyl substituent being varied along with the position of the linker. The presence of additional substituents around the A ring of the quinolone core was also looked at in detail.

The synthesis of these compounds was achieved in 4–6 steps from commercially available, inexpensive starting materials. Aldehyde 1 was utilized in a Grignard reaction to give alcohol 2 in 70–99% yields. Where aldehyde 1 was not commercially available, the aldehydes were synthesized in house (see Supporting Information). Alcohol 2 was oxidized using PCC to give ketone 3 in 80–99% yields. Oxazoline 5 was prepared from the respective isatoic anhydride 4 in yields of 40–60%. In the majority of compounds, the isatoic anhydrides were commercially available; when this was not the case they were synthesized (see Supporting Information). Reaction of oxazoline 5 with ketone 3 in the presence of PTSA gave the desired quinolones 6a–w in 20–85% yields.[14]

A selection of methoxy quinolones 6n–p were then demethylated using BBr3 to give hydroxy quinolones 7a–c in 51–69% yields. 7c was then acetylated using triethylamine and acetyl chloride to give quinolone 8 in 70% yield (Scheme 2).

Scheme 2

Synthesis of Quinolones 7a–c and 8

Where yields of 3-methyl quinolones were very low in the final step, the methodology depicted in Scheme 3 was employed. This route was also the highest yielding for compounds containing a nitrogen within the A ring of the quinolone core. Ketone 3 was converted to dimethoxy acetal 9 in 40–90% yield using trimethyl orthoformate and PTSA. Reaction of diacetal 9 with various anthranilic acids 10 by refluxing in Dowtherm A gave quinolones 11a–h in 43–66% yields (Scheme 3 and Table 2).

Scheme 3

Synthesis of Quinolones 11a–h

Table 2

Yields for the Synthesis of Compounds 11a–h

compound	R	X	Y	% yield 9	% yield 11
11a	-PhpCF3	H	N	45	46
11b	-PhpOCF3	H	N	89	56
11c	-PhpOMe	H	N	58	43
11d	-PhpBr	H	N	55	66
11e	-PhpCH2PhpOCF3	H	N	89	44
11f	-PhpCH2PhpOCF3	7-F	CH	89	44
11g	-PhpCH2PhpOCF3	6-F, 7-F	CH	89	39
11h	-PhpCH2PhpF	H	N	76	42

Analogues with a hydrogen at the 3-position were also synthesized (Scheme 4 and Table 3). In this case ketone 3 was reacted with diazo ethyl malonate to give diketylester 12 in 58–68% yield. Reaction with a variety of anilines gave amine 13 in 52–78% yield. Heating amine 13 in Dowtherm A gave the desired quinolones 14a–k in 48–70% yield. Crystals of quinolone 14e were grown and its structure was confirmed by X-ray crystallography (see Figure 5, CCDC 833920). The effect of a hydroxyl group both in the A ring of the quinolone core and at the terminal end of the side chain was explored. To this end it was necessary to treat 7-OMe quinolone 14h with BBr3 to give 7-OH quinolone 14l in 60% yield. Treatment of 14k containing a side chain with a terminal methyl ester with LAH gave 14m with a terminal methyl alcohol in 78% yield.

Table 3

Yields of Quinolones 14a–k

Figure 5

X-ray crystal structure of quinolone 14e.

It has been shown from GSK’s pyridone series that the presence of a chlorine at the 3-position was well tolerated.[15] With this in mind a selection of the 3-H compounds were treated with sodium dichloroisocyanurate and sodium hydroxide to give 3-chloro quinolones 15a–c in 53–64% yields.

While we believe the 2-bisaryl 3-methyl quinolones to be optimal for antimalarial activity and PfNDH2 selectivity, it was a logical progression to investigate how interchanging the two substituents would influence both activity and selectivity. The first compounds synthesized were the 3-aryl variants of 6d and 6u. Ketone 16 was reacted with oxazoline 17 to give the 3-monoaryl quinolone 18 in 27% yield. Reaction of quinolone 18 with the boronic ester 19 gave the desired 3-bisaryl quinolone 20 in 89% yield (see Figure 6). For the 6u variant quinolone 18 was reacted with phenol 21 in 30% yield to give the 3-bisaryl quinolone 22 with an oxygen linked side chain. The synthesis of hydroxymethyl quinolone 25 was undertaken to see if a hydroxymethyl group was tolerated in the molecule[15] in order to provide a handle for the synthesis of appropriate pro-drugs such as phosphates[16] or carbamates.[17] Reaction of quinolone 22 with ethyl chloroformate gave ester 23 in 70% yield, and subsequent reaction with selenium dioxide in dioxane gave a 98% yield of aldehyde 24. This was followed by conversion to the alcohol 25 in 69% yield. Synthesis of 2-H, 3-bisaryl quinolone 29 was achieved by carrying out a Suzuki reaction on chloro, bromo quinoline 26 in 45% yield. A further Suzuki reaction was then undertaken to give chloro quinoline 28 in 50% yield. Conversion to quinolone 29 was achieved using formic acid in 94% yield.

Figure 6

Phosphate pro-drug of anticancer drug CHM-1-PNa.

Further investigation into the nature of the group tolerated at the 3-position was carried out. Quinolones 32a–c containing an ethyl ester at the 3-position were synthesized by reacting isatoic anhydride 30 with β-keto ester 31 in the presence of NaH and DMF in 30–35% yield (see Figure 7). The presence of a methyl alcohol at the 3-position could then be achieved using LAH to convert the esters to 3-methyl alcohol quinolones 33a and 33b in 46% and 48% yields.

Figure 7

2-Aryl quinolones vs 3-aryl quinolones.

As there are several examples of naturally occurring 1-hydroxy-4(1H)-quinolones that are known inhibitors of respiratory and photosynthetic electron transport chains,[18] it was logical to explore the effect of an N–OH variant of our template on antimalarial activity and PfNDH2 activity. Synthesis of the N–OH compounds was achieved by reacting the quinolone with ethyl chloroformate to give carbonate 34 in 60–99% yield. Synthesis of the N-hydroxy analogues via the carbonate intermediate was advantageous as the carbonates themselves are possible pro-drugs and so subsequently were also tested for antimalarial activity. This was then oxidized using m-CPBA to give the N-oxide 35 which was used crude in the final step. Reaction with KOH gives the desired N-hydroxy compound 36 in 80–98% yield (Scheme 8).[19]

Scheme 8

Synthesis of N–OH Quinolones 36a–e

Optimization of the side chain to improve solubility and drug delivery is key to the successful development of these hits, and there are several strategies that we have adopted to date. Further modifications to the side chain have included extending the terminal group using an oxy-linked alkyl morpholine to provide the opportunity for developing molecules that can be formulated as salts. This type of approach has been applied in the development of kinase inhibitors where incorporation of cyclic amine groups such as morpholine has transformed highly insoluble compounds into candidates with excellent drug-like properties.[20] To incorporate the oxyl-linked morpholine side chain bisaryl aldehyde 37 was converted to the ethyl ketone 39 using chemistry described previously. BBr3 was then used to demethylate 39 to give alcohol 40 in 50–70% yields. Addition of the ethyl morpholine subunit was achieved using potassium carbonate to give side chain 41 in 86% yields. Reaction with oxazoline 5a in the presence of triflic acid gave quinolones 42 a–c in 30–55% yields (Scheme 9).

Scheme 9

Synthesis of Extended Side Chain Ethoxy Morpholine Quinolones 42a–c

While our primary focus was to use medicinal chemistry manipulation of the core template to maximize solubility and activity, pro-drug approaches were also briefly examined. Pro-drugs have been successfully adopted by GSK in their antimalarial pyridone (GSK932121) program. Impressive in vivo antimalarial activity and exposure profiles have been achieved with pyridone-based phosphate pro-drugs.[16] Phosphate ester pro-drugs are highly ionized at physiological pH, highly soluble in water, are chemically stable and enzymatic cleavage at the gut wall by membrane-bound alkaline phosphatases produces high concentrations of the parent drug in the systemic circulation. Phosphate pro-drugs have also been successfully developed for the 2-arylquinolone series of anticancer agents developed by Chou et al. where CHM-1-PNa was developed as a novel water-soluble drug candidate (Figure 6).[21] Morpholine carbamate pro-drugs were also investigated.[17]

Compound 6j was used for the basis of our pro-drug work as it exhibited good in vitro antimalarial activity and selectivity against PfNDH2 (see below). Quinolone 6j was reacted with tetrabenzyl pyrophosphate in the presence of NaH to give the phosphonate ester 43 in 87% yield. Hydrogenation using Pd/C gave phosphate pro-drug 44 in 80% yield (Scheme 10).

Scheme 10

Synthesis of Phosphate Pro-Drug 44

Morpholine carbamate pro-drug 45 was made by reacting quinolone 6j with morpholine carbonyl chloride in the presence of potassium tert-butoxide to give the pro-drug in 66% yield (Scheme 11).

Scheme 11

Synthesis of Morpholine Pro-Drug 45

Further strategies including the use of polar heterocycles in the side chain, use of other protonatable groups within the side chain, extending the terminal group using polar heterocycles, and the placement of a polar group centrally in the side chain with a lipophilic group at the terminal end are covered in the subsequent paper.

Antimalarial Activity

Tables 6, 7, and 8 show the antimalarial activity of all quinolones synthesized against the 3D7 strain of P. falciparum. Table 6 shows activity for monoaryl analogues and while activity of these compounds is generally poor, a few key points can be taken from these results in terms of SAR. In the case of monocyclic compounds, there is a definite trend toward better activity when CF3 groups are present in the side chain and when a chlorine atom is present at the 3-position. Nitrogen within the A ring of the quinolone core results in reduced activity.

Table 6

In Vitro Antimalarial Activities of Monocyclic Quinolones versus 3D7 P. falciparum

compound	X	Y	R1	R2	IC50 (nM) 3D7 ± SD/(IC50 (nM) PfNDH2)
6a	H	CH	Me	CF3	752 ± 7.8/(88.5)
6b	H	CH	Me	OCF3	>1000
11a	H	N	Me	CF3	>1000
11b	H	N	Me	OCF3	>1000
11c	H	N	Me	OMe	>1000
11d	H	N	Me	Br	>1000
14a	H	CH	H	CF3	579 ± 120/(52.90)
14b	H	CH	H	OCF3	675 ± 80/(47.9)
14c	H	CH	H	OMe	>1000
14d	H	CH	H	OH	>1000
15a	H	CH	Cl	OCF3	513 ± 134/(253)
15b	H	CH	Cl	OMe	560 ± 110/(1670)

Table 7

In Vitro Antimalarial Activities of Bicyclic Quinolones versus 3D7 P. falciparum*

First aromatic ring attached to quinolone core has a 2-F substituent.

Pfbc1 IC50 data (nM): 6d = 37.5, 6e = 219, 6f = 25.5, 6j = 9800, 14e = 13.9.

Table 8

In Vitro Antimalarial Activities of Structurally More Diverse Bicyclic Quinolones versus 3D7 P. falciparuma

compound	core	X	Y	R1	R2	IC50 (nM)3D7 ± SD/(IC50 (nM) PfNDH2)
20	A	H	H	-PhpCH2PhpOCF3	Me	36 ± 6/(492)
22	A	H	H	-PhpOPhpOCF3	Me	10 ± 1.2/(190)
25	A	H	H	-PhpOPhpOCF3	CH2OH	91 ± 21/(>56 μM)
29	A	H	H	-PhpOPhpOCF3	H	797 ± 130
32a	A	H	H	CO2Et	-PhpCH2PhpOCF3	39.6 ± 6/(268)
32b	A	7-Cl	H	CO2Et	-PhpCH2PhpOCF3	26 ± 1
33a	A	7-Cl	H	CH2OH	-PhpCH2PhpOCF3	63 ± 5
33b	A	7-Cl	H	CH2OH	-PhpOPhpOCF3	200 ± 22
34a	B	H	H	Me	-PhpCH2PhpOCF3	27 ± 4.3
34c	B	H	H	-PhpCH2PhpOCF3	Me	27 ± 4.4
34e	B	H	H	Me	-PhpOPhpOCF3	60 ± 12
36a	A	H	OH	Me	-PhpCH2PhpOCF3	22 ± 0.4/(55.2)
36b	A	7-Cl	OH	Me	-PhpCH2PhpOCF3	149 ± 40
36c	A	H	OH	-PhpCH2PhpOCF3	Me	175 ± 80/(13.5)
36d	A	H	OH	H	-PhpCH2PhpOCF3	263 ± 64/(71)
36e	A	H	OH	Me	-PhpOPhpOCF3	35 ± 9/(108)
42a	A	H	H	Me	-PhpOPhpO(CH2)2N(CH2CH2)2O	>1000
42b	A	H	H	Me	-PhpOPhmO(CH2)2N(CH2CH2)2O	719 ± 87
42c	A	H	H	Me	PhpCH2PhmO(CH2)2N(CH2CH2)2O	355 ± 60/(279)

Numbers in parentheses are IC50 (nM) PfNDH2.

Table 7 shows the antimalarial activities of quinolones 6c–w, 11e–h, 14e–14m, and 15c. Clear trends are seen in the nature of the A ring substituent X. Generally the presence of Cl and F on the A ring is well tolerated and often enhances activity as seen when comparing 6d (117 nM) to 6j (36 nM), 6k (70 nM), and 6l (38 nM). Larger A ring substituents such as CF3 as in the case of 6h (654 nM) and 6i (212 nM) and piperazine (14i, 430 nM and 14j, 443 nM) are less well tolerated with a 10-fold drop in activity seen. The presence of an OMe group on the A ring is tolerated with substitution at the 7-position greatly enhancing activity. 6o has activity of 8 nM activity whereas all other regioisomeric OMe compounds exhibit antimalarial activity of >350 nM (6m, n and p). Substitution at the 7-position is also favorable when looking at OH substitution (7b, 139 nM versus 7a, 465 nM and 7c, 819 nM). Nitrogens within the A ring are also not tolerated well as seen with 11e (407 nM) and 11h (506 nM). Of the three substituents examined at the 3-position all are well-tolerated. A hydrogen at the 3-position (R1) does seem to offer a small advantage in terms of activity when comparing 14e (48 nM) to 6d (117 nM) and 14f (16 nM) to 11g (24 nM); however, this small increase in activity is far outweighed by the decrease seen in solubility. When comparing 15c (19 nM) with 6d (117 nM) the presence of a chlorine atom greatly enhances activity, and this observation will be employed in future lead optimization campaigns in this area.

Looking in detail at the side chain, linker A variants para-CH2, meta-CH2, and para-O are all well tolerated with activity effects being determined by other areas of the molecule. The effect of the side chain terminal substituent is highly dependent on other functionality within the molecule, but as a general rule OCF3 is the optimal terminal group as demonstrated by the comparison of 6r (34 nM) to 6s (105 nM) and 6u (26 nM) to 6w (230 nM). Large electron withdrawing groups are less well tolerated as seen with 14k (272 nM). Alcohol groups both on the A ring and at the terminal end of the side chain results in a decrease in activity as demonstrated by 14l and 14m.

Table 8 shows the antimalarial activities of the more structurally diverse bicyclic quinolones. From the small number of 3-aryl compounds synthesized, the effect of altering R2 can be seen. For this series of compounds Me > CH2OH > H in terms of antimalarial activity. For a comparison of 3-aryl compounds vs 2-aryl compounds across the full range of in vitro data, see Figure 7. Other comparisons that can be made from the table include the effect of having an ethyl ester at the 3-position. Ester 32a (39.6 nM) can be compared to its methyl equivalent 6c (107 nM) and likewise ester 32b (26 nM) to 6j (36 nM). In both cases the ester does offer a slightly better activity. The presence of an alcohol at the 3-position does however reduce activity slightly. Core B compounds tested demonstrate good antimalarial activity, but there is no definite trend when compared to their core A counterparts.

The general trend when N-hydroxy compounds are compared to the NH variants (36b (149 nM) cf 6j (36 nM), 36c (175 nM) cf 20 (36 nM), 36d (263 nM) cf 14e (48 nM), and 36e (35 nM) cf 6u (26 nM)) is a reduction in activity, although 36a is an exception to this. Generally, the addition of an ethoxy morpholine group leads to a drop in 3D7 activity. This would concur with our previous observations that larger terminal substituents on the side chain are not well tolerated.

Having established the whole cell activity of all quinolone compounds, they were then tested against the PfNDH2 enzyme. Because of the time-consuming nature of the assay[22] and large volume of parasites needed, only a small selection of the most active compounds were then tested against parasite bc1 in order to establish the selectivity of the compounds against PfNDH2 (see footnote, Table 7). A large number of the quinolones tested demonstrate nanomolar activity against PfNDH2 and some selectivity against parasite bc1.

From these compounds a selection was tested against the atovaquone resistant TM90C2B strain of P. falciparum (IC50 for atovaquone is 12 μM in this strain).

Additionally a more select range of compounds were tested against the chloroquine resistant strain of P. falciparum, W2. The SAR trends identified from the 3D7 data largely hold true for the W2 data with the presence of a 7-methoxy (6o, 13 nM) and 7-Cl (6j, 17 nM) groups enhancing activity when compared to unsubstituted 6d (26 nM).

A direct comparison of 3-aryl and 2-aryl quinolones can be made from the two pairs of compounds depicted in Figure 7. This clearly depicts a loss of PfNDH2 activity when moving from 2-aryl to 3-aryl examples with 6u having PfNDH2 activity of 10 nM and its 3-aryl counterpart 22 having activity of 190 nM. 6d has 20 nM PfNDH2 activity with this droping to 492 nM for 3-aryl quinolone 20. The most extreme example of this being 3-aryl quinolone 25 which shows no PfNDH2. This trend is also observed with the W2 P. falciparum data. All analogues depicted in Figure 7 demonstrate good levels of 3D7 antiparasitic activity.

A selection of the most active quinolones were tested for in vivo activity using Peters’ Standard 4-day test (Table 11).[23] Some solubility problems were encountered with the use of SSV (in most cases compounds had to be dosed as suspensions), but the use of DET (compounds fully dissolved) is proof of concept that 6j (CK-2-68) clears the parasite in vivo with 100% parasite kill being achieved at 20 mg/kg. The pro-drug of 6j, compound 44 was successfully dosed in a sodium carbonate solution and 100% parasite kill was also seen at 20 mg/kg. 6d was also potent by oral route in the mouse model with 100% clearance at 20 mg/kg in this model. In the cases where parasite clearance did not reach 100%, we believe this to be a solubility issue as from the table it is clearly vehicle dependent.

Table 11

In Vivo Peters’ Standard 4 Day Testa

	% parasite clearance on day 4 (20 mg/kg po)
	vehicle
compound	SSV	DET	Na2CO3
atovaquone	100	100	ND
6d	100	100	ND
6j	59	100	ND
6u	100	95.4	ND
44	ND	ND	100
45	ND	ND	100

Day 4 suppressive activity of key compounds in male CD-1 mice infected with Plasmodium berghei. Mice were exposed to the infection via intraperitoneal injection and then orally dosed with the relevant compound. Data were obtained from 5 mice per group.

Because of 6j having excellent in vitro activity and selectivity against PfNDH2, it was selected as the lead compound for further investigation.

Cytotoxicity

No significant cytotoxicity was observed for 6j at any concentration (CC50 > 50 μM) in HEPG2 cells. Cytotoxicity data established a selectivity index (CC50/IC50) > 1388.

Human Liver Microsomal Incubations

6j was incubated at a concentration of 1 μM with human liver microsomes (1 mg/mL) in the presence of NADPH for 0, 10, 30, and 60 min. After 60 min, 80% of 6j remained. The in vitro half-life for 6jwas shown to be 226 min, with an intrinsic clearance value of 0.76 mL/min/kg.

Conclusions

To conclude, a 4–6 step synthesis of a range of bisaryl quinolones with potent antimalarial activity both in vitro and in vivo has been reported. Several compounds within this series have been proven to be selectively active against the PfNDH2 enzyme. Lead compounds within this series have antimalarial activity against the 3D7 strain of P. falciparum and PfNDH2 activity in the low nanomolar region and for the most selective quinolone, 6j, a PfNDH2/Pfbc1 selectivity ratio of up to 600-fold. It is important to note that additional quinolones in this series have the ability to inhibit both PfNDH2 and bc1 in the low nanomolar range and this dual targeting of two key mitochondrial enzyme targets may prove to be an advantage over single-targeting inhibitors with respect to drug efficacy and delaying the onset of parasite drug resistance.

Representative quinolones and their phosphate pro-drugs also have proven to be effective at clearing parasitic infection at 20 mg/kg in a murine model of malaria, and further work is in progress to optimize the solubility and ADMET properties of this series.

Experimental Section

Chemistry

All reactions that employed moisture sensitive reagents were performed in dry solvent under an atmosphere of nitrogen in oven-dried glassware. All reagents were purchased from Sigma Aldrich or Alfa Aesar chemical companies, and were used without purification. Thin layer chromatography (TLC) was carried out on Merck silica gel 60 F-254 plates and UV inactive compounds were visualized using iodine or anisaldehyde solution. Flash column chromatography was performed on ICN Ecochrom 60 (32–63 mesh) silica gel eluting with various solvent mixtures and using an air line to apply pressure. NMR spectra were recorded on a Bruker AMX 400 (1H, 400 MHz; 13C, 100 MHz) spectrometer. Chemical shifts are described in parts per million (δ) downfield from an internal standard of trimethylsilane. Mass spectra were recorded on a VG analytical 7070E machine and Fisons TRIO spectrometer using electron ionization (EI) and chemical ionization (CI). All compounds were found to be >95% pure by HPLC unless specified below. See Supporting Information for experimental and data on all intermediates.

General Procedure for the Synthesis of Quinolones 6

The appropriately substituted oxazoline 5 (4 mmol, 1.0 equiv) was added to a solution of ketone 3 (4 mmol, 1.0 equiv) and para-toluenesulfonic acid (20 mol %) in n-butanol (10 mL). The reaction mixture was heated to 130 °C under nitrogen and stirred for 24 h. The solvent was removed under a vacuum and water (20 mL) was added. The aqueous solution was extracted with EtOAc (3 × 20 mL), dried over MgSO4, and concentrated under a vacuum. The product was purified by column chromatography (eluting with 20–80% EtOAc in n-hexane) to give quinolone 6.

6d: White powder (Yield 36%); mp 212–214 °C; 1H NMR (400 MHz, DMSO) δH 8.98 (s, 1H, NH), 8.27 (d, J = 8.3 Hz, 1H), 7.60 (d, J = 8.1 Hz, 1H), 7.56 (dt, J = 1.4 Hz, 8.3 Hz, 1H), 7.9 (d, J = 8.1 Hz, 2H), 7.26 (dt, J = 1.5 Hz, 8.1 Hz, 1H), 7.20 (d, J = 8.0 Hz, 2H), 7.16 (d, J = 8.6 Hz, 2H), 7.11 (d, J = 8.1 Hz, 2H), 3.96 (s, 2H), 2.01 (s, 3H); 13C NMR (100 MHz, DMSO), δC 178.7, 149.0, 142.4, 139.5, 133.8, 132.0, 130.6, 129.4, 126.4, 123.8, 121.5, 118.0, 116.6, 41.3, 12.9; MS (ES+), [M + H]+m/z 410.1, HRMS calculated for 410.1368 C24H19NO2F3, found 410.1348.

6j: White solid (Yield 30%); mp 240–242 °C; 1H NMR (400 MHz, MeOD) δ 8.27 (d, J = 8.8 Hz, 1H), 7.62 (s, 1H), 7.52–7.45 (m, 5H), 7.43–7.34 (m, 3H), 7.24 (d, J = 7.9 Hz, 1H), 4.15 (s, 2H), 2.05 (s, 3H); MS (ES+) m/z 444 [M + H]+ Acc mass found: 444.0962, calculated 444.0978 for C24H18NO2F3Cl.

6u: White solid (Yield 28%); mp 207–208 °C; 1H NMR (400 MHz, CDCl3) δ 8.24 (d, J = 8.2 Hz, 1H), 7.61 (d, J = 8.3 Hz, 1H), 7.54 (t, J = 7.5 Hz, 1H), 7.42 (d, J = 8.5 Hz, 2H), 7.31 – 7.17 (m, 3H), 7.03 (dd, J = 8.6, 6.9 Hz, 4H), 2.02 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 179.14, 158.36, 155.06, 148.33, 145.44, 139.67, 131.94, 130.92, 130.58, 125.83, 123.79, 123.76, 123.15, 120.76, 118.57, 118.17, 116.35, 12.76; MS (ES+) m/z 412 [M + H]+ Acc mass found: 412.1175, calculated 412.1161 for C23H17NO3F3.

Procedure for the Synthesis of Phosphate Pro-Drug 44

A suspension of phosphate 43 (0.18 mmol, 1.0 equiv) in anhydrous methanol (10 mL) was subjected to hydrogenation in the presence of 10% Pd/C (50 mg) at room temperature for 10 min. The catalysts and any precipitates were filtered off and the methanol portion was analyzed by TLC. The solvent was removed in vacuo to give the desired phosphate pro-drug 44 and no further purification was required. White solid (Yield 80%); mp 201–203 °C; 1H NMR (400 MHz, CDCl3) δ 11.82 (s, 1H), 11.62 (s, 1H), 8.32 (d, J = 8.2 Hz, 1H), 8.26 (d, J = 8.0 Hz, 1H), 8.12 (d, J = 8.4 Hz, 1H), 8.06 (s, 1H), 7.91 (t, J = 8.2 Hz, 1H), 7.82 (t, J = 8.1 Hz, 1H), 7.68 (d, J = 8.0 Hz, 1H), 7.56 (d, J = 8.2 Hz, 2H), 7.52 (d, J = 8.0 Hz, 2H), 7.46–7.32 (m, 12H), 4.12 (s, 2H), 4.09 (s, 2H), 2.40 (s, 3H), 2.37 (s, 3H). 31P NMR (162 MHz, CDCl3) δ −5.027, −5.396; MS (ES–) m/z 522 [M – H]− Acc mass found: 522.0471, calculated 522.0485 for C24H17NO5F3PCl.

Procedure for the Synthesis of Morpholine Pro-Drug 45

Quinolone 6j (0.31 mmol) in anhydrous THF was added tBuOK (52.7 mg, 0.47 mmol) at room temperature. The mixture was stirred for 1/2 h. 4-Morpholinecarbonyl chloride (0.05 mL, 0.41 mmol) was added. The mixture was stirred for a further 2 h (followed by TLC). The reaction was quenched with brine and was extracted with ethyl acetate, dried over Na2SO4, filtered, and concentrated to an oil. The crude product was purified by column chromatography using 20% ethyl acetate in hexane to give 43 as a white solid (Yield 66%); mp 148–150 °C; 1H NMR (400 MHz, CDCl3) δ 8.12 (s, 1H), 7.74 (d, J = 8.9 Hz, 1H), 7.53 (d, J = 8.1 Hz, 2H), 7.49 (d, J = 8.9 Hz, 1H), 7.30 (d, J = 8.1 Hz, 2H), 7.24 (d, J = 8.9 Hz, 2H), 7.14 (d, J = 8.8 Hz, 2H), 4.06 (s, 2H), 3.87–3.83 (m, 6H), 3.65 (brs, 2H), 2.30 (s, 3H); MS (ES+) m/z 557 (M + H)+ Acc mass found: 557.1443, calculated 557.1455 for C29H25N2O4F3Cl.

Biology

Parasite Culture

Plasmodium blood stage cultures[24] and drug sensitivity[25] were determined by established methods. IC50s (50% inhibitory concentrations) were calculated by using the four-parameter logistic method (Grafit program; Erithacus Software, United Kingdom)

High-Throughput Screening (HTS)

PfNDH2 activity was measured using an end-point assay in a 384-well plate format. Final assay concentrations used were 200 μM NADH, 10 mM KCN, 1 μg/mL F571 membrane,[6] and 20 μM decylubiquinone (dQ). A pre-read at 340 nm was obtained prior to the addition of dQ to initiate the reaction followed by a post-read at 1 min. HDQ was used as positive control at 5 μM. The agreed QC pass criteria was Z′ > 0.6 and signal/background >10. Compounds were selected by the described chemoinformatics algorithms from the Biofocus DPI compound library (Galapagos Company).

Enzymology

P. falciparum cell-free extracts were prepared from erythrocyte-freed parasites as described previously,[22] and recombinant PfNDH2 was prepared from the Escherichia coli heterologous expression strain F571.[6] PfNDH2 and bc1 activities were measured as described previously.[6,22]

Pharmacology

In vivo efficacy studies were measured against P. berghei in the standard 4-day test.[23] All in vivo studies were approved by the appropriate institutional animal care and use committee and conducted in accordance with the International Conference on Harmonization (ICH) Safety Guidelines.

Table 5

Yields of Carbamates 34a–d and N-Hydroxy Quinolones 36a–d

compound	R1	R2	X	% yield 34	% yield 36
36a	Me	-PhpCH2PhpOCF3	H	98	98
36b	Me	-PhpCH2PhpOCF3	7-Cl	67	80
36c	-PhpCH2PhpOCF3	Me	H	99	91
36d	H	-PhpCH2PhpOCF3	H	88	98
36e	Me	-PhpOPhpOCF3	H	60	83

Table 9

In Vitro Antimalarial Activities of Selected Quinolones versus TM90C2B

compound	IC50 (nM) TM90C2B ± SD/(IC50 (nM) 3D7 ± SD)	compound	IC50 (nM) TM90C2B ± SD (IC50 (nM) 3D7 ± SD)
6c	416 ± 74 (107 ± 14)	14e	251 ± 22 (48 ± 7)
6d	122 ± 26 (117 ± 27)	14f	626 ± 69 (16 ± 4)
6e	65 ± 11(26 ± 2)	15c	328 ± 48 (19 ± 6)
6f	273 ± 35 (83 ± 9)	32a	1400 ± 57 (39.6 ± 6)
6g	577 ± 43 (37 ± 7)	32b	92 ± 2 (26 ± 1)
6j	178 ± 9 (36 ± 5)	33a	330 ± 58 (63 ± 5)
6q	31 ± 7 (8.4 ± 0.4)	34a	6.8 ± 3.5 (27 ± 4.3)
6r	94 ± 3 (34 ± 6)	34c	224 ± 47 (27 ± 4.4)
6s	552 ± 35 (105 ± 15)	34e	406 ± 74 (60 ± 12)
6u	92 ± 2 (26 ± 1)	36a	217 ± 18 (2.2 ± 0.4)
6v	274 ± 58 (73 ± 19)	36b	>1000 (149 ± 40)
6w	797 ± 34 (230 ± 43)	36c	670 ± 24 (175 ± 80)
11e	1880 ± 150 (407 ± 30)	36d	403 ± 38 (263 ± 64)
11f	191 ± 35 (69 ± 11)	36e	566 ± 35 (35 ± 9)
11g	207 ± 43 (24 ± 6)

Table 10

In Vitro Antimalarial Activities of Selected Quinolones versus W2

compound	IC50 (nM) W2 ± SD/(IC50 (nM) 3D7 ± SD)	compound	IC50 (nM) W2 ± SD (IC50 (nM) 3D7 ± SD)
6d	26 ± 1.2 (117 ± 27)	15c	22 ± 2.5 (19 ± 6)
6j	17 ± 0.6 (36 ± 5)	20	42 ± 1.3 (36 ± 6)
6o	13 ± 0.9 (8 ± 2)	22	34 ± 3.4 (10 ± 1.2)
14e	36 ± 0.5 (48 ± 7)	36a	8 ± 0.7 (22 ± 0.4)