Enone- and chalcone-chloroquinoline hybrid analogs: In silico-guided design, synthesis, antiplasmodial activity, in vitro metabolism and mechanistic studies

Abstract

Analogs of the previously reported antimalarial hybrid compounds 8b and 12 were proposed with the aim of identifying compounds with improved solubility and retained antimalarial potency. In silico characterization predicted improved solubilities of the analogs, particularly at low pH; they retained acceptable predicted permeability properties, but were predicted to be susceptible to hepatic metabolism. These analogs were synthesized and found to exhibit notable in vitro antimalarial activity. Compounds 25 and 27 were the most active of the analogs. In vitro metabolism studies indicated susceptibility of the analogs to hepatic metabolism. There was also evidence of primary glucuronidation for analogs 24 – 27. Presumed cis - trans isomerism of 12, 22 and 23 under in vitro metabolism assay conditions was also observed, with differences in the nature and rates of metabolism observed between isomers. Biochemical studies strongly suggested that inhibition of hemozoin formation is the primary mechanism of action of these analogs.

Introduction

The World Health Organization (WHO) estimates that over 200 million episodes of malaria occur annually.^1;2 These result in between 0.6 – 1.2 million deaths a year, nearly all caused by Plasmodium falciparum, with ~85% of these deaths being of children below the age of 5 years in sub-Saharan Africa.¹ P. falciparum, the most virulent species of the malaria parasite, has developed varying degrees of resistance to many classes of drugs,^3;4 including possibly artemisinin-derived antimalarials.^5;6 Thus, the development of new, highly efficacious drugs to treat malaria remains a key priority.⁷

Antimalarial drug discovery, and indeed drug discovery in general, faces a myriad of challenges. For example, it is widely acknowledged that challenging physicochemical and absorption, distribution, metabolism and excretion (ADME) properties significantly hamper the development of many promising compounds into leads and drug candidates.^8;9 This has led to an increased interest in the early determination (and/or prediction) of ADME properties, with simultaneous filtering in vitro for potency.^8;10 The use of large compound and data sets has dramatically increased the ability to model biological processes and relate these to easily measured or calculated physicochemical and structural features of the compounds.^11;12 In silico (computational) ADME prediction models and tools therefore complement in vitro and in vivo ADME experiments by facilitating the characterization of compounds prior to their synthesis,^9;11 thereby focusing synthetic and experimental evaluation efforts on a smaller number of particularly promising compounds. However, the results from such predictions should, whenever possible, be corroborated by in vitro and/or in vivo ADME assays.

For drugs to treat uncomplicated malaria, oral administration is practically mandatory, as huge numbers of patients across the developing world must be treated in facilities with limited resources or infrastructure.^10;13 The proportion of an orally administered drug that is delivered into the systemic circulation (i.e. its oral bioavailability) is heavily dependent on three primary factors: solubility, permeability and metabolic stability.^14;15 The in silico prediction and/or in vitro determination of these three factors is therefore highly valuable in antimalarial drug design.

In this study, in silico prediction tools were applied in the characterization of a series of analogs proposed for synthesis. These analogs were based on the promising antimalarial chalconechloroquinoline hybrid compound 8b and the lower molecular weight intermediate 12) (Figure 1), both of which were identified and reported previously.¹⁶ One of the main concerns about compound 8b was its poor aqueous solubility, which necessitated the use of up to 10% DMSO as co-solvent and extended periods of sonication and agitation during the initial dissolution of the compounds in the preparation of the stock solutions for in vitro evaluation. The compound was thus anticipated to have poor oral bioavailability.

Figure 1.

Core structural features of 8b and 12

The core structural features thought to be responsible for the observed antimalarial activity of 8b and 12, and which were therefore to be retained in the analogs, were identified as the chloroquinoline moiety and either the phenyl-butenone moiety (for 12) or the chalcone moiety (for 8b) (Figure 1) - the chloroquinoline moiety for its role in inhibition of hemozoin formation,¹⁷ and the chalcone and phenyl-butenone moieties for their possible dual roles in malarial cysteine protease inhibition¹⁸ and facilitating inhibition of hemozoin formation through π-π stacking interactions.

A small series of analogs of 8b and 12 was therefore proposed based on the replacement of the triazole linker. These proposed analogs 20 – 27 (Figure 2) contained amino-alkyl or piperazinyl-based linkers, thereby introducing the 4-amino-substitution on the chloroquinoline sub-unit and resulting in more polar compounds with potentially enhanced (pH-dependent) solubility. In addition, the piperazinyl-based linkers possess an additional protonatable nitrogen, which should further enhance the ionization of the analogs bearing this type of linker. Analogous features are present in chloroquine and amodiaquine, both of which are established chloroquinoline antimalarials that possess a 4-amino substitution and a second protonatable nitrogen, and which have been successfully formulated as water-soluble salts.

Figure 2.

Structures of analogs of 8b and 12 derived from replacement of the triazole linker

The in silico characterization of the proposed analogs and their subsequent synthesis and biological evaluation is reported and discussed herein.

Results and discussion

In silico characterization of 8b, 12 and the proposed analogs

Three software packages, with varying but complementary capabilities, were applied for the in silico prediction of various parameters of interest. VolSurf+ was applied to predict physicochemical and ADME parameters of the compounds, such as aqueous solubility (at various pHs), Caco2 permeability and metabolic stability.^19;20 MetaSite was applied to predict sites of metabolism by hepatic cytochrome P450 enzymes and the structures of possible metabolites.^21;22 MoKa was used to predict the pKa of ionizable compounds.²³ These three software packages are available from Molecular Discovery (http://www.moldiscovery.com/software.php).

The chloroquinoline-based antimalarials chloroquine and amodiaquine were selected as reference compounds and were characterized with and compared to the study compounds. A deliberate attempt was made not to over-interpret the prediction results, but rather to use them as a guide in determining whether or not the proposed compounds were likely to be improvements on the parent compounds 8b and 12, and to highlight any potential liabilities.

Figure 3A shows the logarithm of the predicted aqueous solubility (log Sol.) plotted against the predicted n-Octanol-water partition coefficient (logD) at approximately physiological pH (pH 7.5) and at pH 5.0 for the compounds of interest. Chloroquine and amodiaquine showed significantly superior predicted aqueous solubility at both pHs when compared to 8b, 12 and the proposed analogs. At both pH 5.0 and 7.5, 8b showed the lowest predicted solubility; 12 was predicted to be more soluble than 8b, most likely due to its lower molecular weight. As expected, a similar trend was seen for the proposed compounds, whereby analogs of 12 were predicted to be more soluble than the corresponding analogs of 8b.

Figure 3.

A: Plots of predicted solubility (expressed as the logarithm, log Sol.) against the predicted n-Octanol-water partition coefficient (also expressed as the logarithm, log D) at pH 5.0 and at pH 7.5; B: Plots showing the proposed analogs projected onto the PLS models used to predict Caco2 permeability and metabolic stability – the black dots represent the compounds comprising the models’ training data sets while the yellow dots represent the test compounds; the blue regions indicate acceptable predicted permeability/metabolic stability while the red zones indicate poor predicted properties.

At pH 7.5, the predicted solubilities of the proposed analogs were only slightly higher than those of the parent compounds. However, as anticipated, there was a significant influence of pH on the predicted solubilities, which were significantly higher at pH 5.0 than 7.5. This difference is attributable to the fact that these basic compounds would exist predominantly in their ionized state at lower pH levels, rendering them more soluble. Analysis with MoKa predicted that compounds 20 – 27 were >90% ionized at pH 5.0.

The number of protonatable centers was also predicted to influence pH-dependent solubility. For example, among the analogs of 12, the piperazinyl analogs 22, 24 and 26 (which have two protonatable centers) had higher predicted solubilities at pH 5.0 than the amino-ethoxy analog 20, despite the fact that these three piperazinyl derivatives have higher molecular weights than 20. A similar trend was observed for the analogs of 8b, and both observations can be attributed to the positive influence on solubility of the higher charge-to-mass ratio afforded by the di-protonation of the piperazinyl analogs.

Prediction of Caco2 permeability was qualitative. When projected onto the two-dimensional PLS plot of the database of compounds used to generate the Caco2 permeation model, all the compounds (along with chloroquine and amodiaquine) showed acceptable predicted Caco2 permeation, falling within the blue (permeable) zone of the plot (Figure 3B).

Amodiaquine and chloroquine were predicted to have intermediate metabolic stability, falling somewhat between the blue (stable) and red (unstable) zones on the PLS plot (Figure 3B); they were predicted to be more stable to metabolism than the test compounds. 8b was predicted to be less stable to hepatic metabolism than 12. The proposed analogs were also shown to have relatively low predicted metabolic stabilities, particularly compounds 22, 23 and 27. Site-of-metabolism analysis was carried out on MetaSite in an attempt to shed light on these findings (Figure 4).

Figure 4.

Output from MetaSite showing the predicted sites of metabolism on the various analogs

The main route of metabolism for 8b was predicted to be O-demethylation of the para-methoxy group on ring B, as indicated by the blue ring highlighting this group; this group is absent in 12. O-demethylation on ring A and N-oxidation of the quinoline ring were the main routes of metabolism predicted for 12, but these may not be as efficient as O-demethylation on ring B, and this is probably why this compound was predicted to be more stable to metabolism than 8b.

N-dealkylation of the nitrogen-containing side-chain was the principal route of metabolism for chloroquine and amodiaquine predicted by MetaSite. This result is in agreement with reported experimental findings, with their major metabolites identified as N-desethylchloroquine and N-desethylamodiaquine, respectively.^24;25 N-oxidation of the quinoline nitrogen, although highlighted by MetaSite as a possible site of metabolism, is not a favored route of metabolism for this class of compounds.

Unsurprisingly, N-dealkylation of the nitrogen-containing linkers was predicted as a possible route of metabolism for the proposed analogs. The piperazinyl-based linkers present in 22 – 27 were predicted as being particularly susceptible to N-dealkylation, possibly explaining why these compounds were predicted to be relatively unstable. This factor and the additional possibility of O-demethylation on ring B most likely explains why compound 23 was predicted to be the least stable of the analogs. The aminoethoxy linker present in 20 and 21 appears to be somewhat less susceptible to N-dealkylation, and was not predicted as the primary target of metabolism for these compounds, which were predicted to be the most stable of the studied analogs.

In summary, compounds 8b and 12 showed poor predicted solubility profiles, suggesting limitations in oral bioavailability. They had acceptable predicted Caco2 permeation when compared to the reference compounds chloroquine and amodiaquine, though their metabolic stability was predicted to be lower. The proposed analogs had improved predicted solubilities relative to their parent compounds, particularly at lower pH, and this was a feature that could possibly be utilized during the solubilization of these analogs for in vivo pharmacokinetic (PK) studies. The analogs also maintained acceptably high predicted Caco2 permeability. On the basis of these predictions, the analogs were synthesized for in vitro antimalarial evaluation to determine if they retained antimalarial activity.

The main alert for the analogs was predicted metabolic instability, primarily due to N-dealkylation of the nitrogen-containing linkers or O-demethylation of ring B of the chalcone derivatives.

Synthesis

The hydroxylated enones (15a and 15b) and chalcones (16a and 16b) were intermediates in the synthesis of these compounds; the synthesis of these intermediates is depicted in Scheme 1, and involved the acid-catalyzed Claisen-Schmidt reaction.

Scheme 1.

Reagents and conditions: (i) acetone, 5M HCl, rt, 24 hrs; (ii) 2’,4’-dimethoxy acetophenone, 5M HCl, MeOH, 50 °C, 24 hrs.

Vanillin or 4-hydroxybenzaldehyde was stirred in acetone (as reagent and solvent) in the presence of aqueous 5M HCl at room temperature for 24 hours. The crude product of the reaction was purified by silica column chromatography to furnish the pure 15a and 15b in good yields. The synthesis of the chalcones 16a and 16b involved the reaction of vanillin or 4-hydroxybenzaldehyde with 1.1 equivalents of 2’,4’-dimethoxy-acetophenone in methanol, in the presence of aqueous HCl as catalyst. The crude products were then purified by silica gel column chromatography to furnish the pure chalcone intermediates in average yields.

Scheme 2 shows the synthesis of the target analogs 20 – 23. The required hydroxylated quinoline intermediates 17 and 18 were afforded by typical aromatic substitution reactions using the (modified) method reported by Singh et al. 1971.²⁶ Briefly, 4,7-dichloroquinoline was heated with excess ethanolamine or with 1 equivalent of 1-(2-hydroxyethyl)piperazine in the presence of potassium carbonate to yield the target intermediates in moderate to good yield after purification by crystallization. The intermediates 17 and 18 were then coupled to either 15a or 16a to yield the target compounds 20 – 23. This coupling was achieved by applying the Mitsunobu reaction, a convenient, one-step S_N2 reaction using 1.2 equivalents of triphenylphosphine (PPh₃) and 1.2 equivalents of diisopropyl azodicarboxylate (DIAD) as described by Winssinger et al. 2007.²⁷

Scheme 2.

Reagents and conditions: (i) ethanolamine, 130 °C, 5 hrs; (ii) 1-(2-hydroxyethyl)piperazine, K₂CO₃, DMF, 80 °C, 24 hrs; (iii) 5.1, 5.2, 5.3 or 5.4, PPh₃, DIAD, DCM, 0 °C – rt, 6 hrs.

The synthesis of the proposed analogs 24 – 27 (Scheme 3) began with the synthesis of the piperazinyl intermediate 19 via an aromatic substitution reaction of 4,7-dichloroquinoline and 6 equivalents of piperazine, in the presence of potassium carbonate, using the (modified) method reported by Singh et al. 1971.²⁶ Purification by silica gel column chromatography afforded 19 in good yield. The Mannich reaction was then applied to furnish the target analogs using the (modified) method described by Chibale et al. 2007.²⁸ This involved the refluxing of the appropriate phenolic intermediate (15 or 16) and 1 equivalent of 19 in ethanol in the presence of 10 equivalents of aqueous formaldehyde for 12 hours. The crude products were then purified by silica gel column chromatography to afford the pure target compounds in average yields, which could be partly attributed to the fact that these reactions did not proceed to completion even with the use of a large excess of formaldehyde or after extended refluxing.

Scheme 3.

Reagents and conditions: (i) piperazine, K₂CO₃, NMP, 135 °C, 4 hrs; (ii) 5.1, 5.2, 5.3 or 5.4, HCHO, EtOH, reflux at 110 °C, 12 hrs.

All compounds were characterized by ¹H-NMR, ¹³C-NMR, Low Resolution Mass Spectrometry, IR Spectroscopy, Elemental analysis and melting point determination.

In vitro antiplasmodial assay results and discussion

The intermediates and target compounds were evaluated for their in vitro antiplasmodial activity against the chloroquine sensitive strain D10 and chloroquine resistant strains Dd2 and W2 of P. falciparum (Table 1).

Table 1.

In vitro antiplasmodial activities against the D10, Dd2 and W2 strains of P. falciparum (IC₅₀, µM)

Compound	IC50, µMa
	D10	Dd2	W2
15a	>20	>20	>10
16a	5.3	7.3	>10
17	2.6	1.0	0.2
18	5.2	3.8	1.0
19	0.7	0.9	0.3
20	1.0	1.5	0.1
21	0.8	1.1	0.4
22	1.2	1.0	0.5
23	1.2	0.5	0.6
24	1.0	1.7	0.5
25	0.5	0.5	0.4
26	1.7	3.1	0.7
27	0.6	0.5	0.3
12	0.8	0.7	ND
8b	0.04	0.07	0.09
Chloroquine	0.017	0.097	0.069

^aaverage values from two independent determinations, each carried out in duplicate

The hydroxylated enone intermediate 15a showed negligible antiplasmodial activity in vitro across all three strains of P. falciparum. However, the chalcone intermediate 16a showed modest antiplasmodial activity against two of the three tested strains. The chloroquinoline-based intermediates 17 – 19 exhibited more potent antiplasmodial activity in vitro, with 19 showing sub-micromolar IC₅₀ values against all three strains.

All the target analogs exhibited notable antiplasmodial activity, with IC₅₀ values in the lower micromolar to mid-nanomolar range. The related analogs 25 and 27 were the most active of the analogs; both are chalcone derivatives with piperazinyl linkers, and had IC₅₀ values between 0.3 – 0.6 µM against all three strains of P. falciparum. Compound 23 was also quite active, with mid-nanomolar IC₅₀ values against the Dd2 and W2 P. falciparum strains.

Intermediate 19 was predicted by MetaSite as the main metabolite of 24 – 27 (arising from N-dealkylation of the linker), and the fact that it exhibited notable in vitro antiplasmodial potency implies that this intermediate/metabolite may contribute to in vivo antimalarial activity of these analogs, and thereby offset any potential liabilities of these compounds arising from susceptibility to metabolism via this route.

In summary, the replacement of the triazole linker in 8b and 12 with the selected nitrogen-containing linkers was associated with retention of antiplasmodial activity, though none were as active as 8b.

In vitro Metabolic Stability

The metabolic stability assay was performed by separately incubating test compounds at 37°C (in duplicate, substrate concentrations 1 µM) with human or mouse liver microsomes (0.4 mg/mL protein concentration). The reaction was initiated by the addition of an NADPH-regenerating system (containing 1 mg/mL NADP, 1 mg/mL glucose-6-phosphate, 1 U/mL glucose-6-phosphate dehydrogenase) and quenched at various time points over the incubation period by the addition of acetonitrile. . NADPH is the cofactor required for cytochrome P450-mediated metabolism. Additional samples with the dual cofactors NADPH and UDPGA (the latter the cofactor for glucuronidation) were included in the incubations for qualitative assessment of the potential for glucuronide formation. Control samples (containing neither NADPH nor UDPGA) were also included to monitor for potential degradation in the absence of cofactors. The concentrations of the substrates after the various periods of incubation were quantified against standard curves prepared in blank microsomal matrix by LC-MS using a Waters/Micromass ZQ mass spectrometer coupled to a Waters Alliance 2975 HPLC and an Ascentis Express C18 column maintained at a temperature of 40°C. Data processing was performed using Quanlynx software.

The results of the in vitro metabolism studies are summarized in Table 2. Two peaks were observed for 12, 22 and 23 during chromatographic LC-MS analysis, most likely indicating cis - trans isomerism occurring under assay conditions. The rates of degradation for each individual peak were therefore determined.

Table 2.

Results from the in vitro metabolism studies

Compound		Degradation half-life (min)		Microsome-Predicted Extraction ratio (EH)		Metabolites detected*
		Human	Mouse	Human	Mouse
8b		51.1	56.8	0.65	0.57	P+16
12	Peak 1	7.4	10.8	0.93	0.88	Not Detected
	Peak 2	9.1	6.2	0.91	0.92
20		10.5	13.4	0.90	0.85	P−218 P+16 (x2), P+32 (x2), P−14, P+16+176 P−14, P−218
21		17.5	12.7	0.85	0.86
22	Peak 1	14.2	6.7	0.87	0.92
	Peak 2	9.8	5.6	0.91	0.93	P+16 (x4), P−14 (x2) P−340
23	Peak 1	11.5	9.8	0.89	0.89
	Peak 2	93.2	149.8	0.51	0.34
24		16.9	14.0	0.85	0.84	P+16, P−204, P+176 P−14 (x2), P−326, P+176 P+16, P−174, P+176 P+16 (x2), P−14, P−296 P+176
25		37.0	11.8	0.72	0.87
26		7.5	15.0	0.93	0.84
27		58.6	45.4	0.62	0.63
Chloroquine		>250	177.0	<0.28	0.31	Not Undertaken

^*as represented by fragments detected after Mass Spectroscopy analysis, and which had m/z values corresponding to putative metabolites; P represents the molecular mass of the parent compound; the interpretation of the results is contained in-text in the section ‘Putative Metabolite Identification’.

With respect to metabolic stability, the compounds exhibited varying degradation half-lives in human and mouse microsomes; however, the calculated hepatic extraction ratios (E_H values) were similar between the species. E_H values take into account the different hepatic blood flow rates between the species. The similarities in these values suggest no significant inter-species differences in the relative rates of metabolism.

Most of the compounds exhibited high E_H values (>0.7), suggesting notable susceptibility to hepatic metabolism. This is particularly apparent when compared to the relatively low values for chloroquine (≤0.31), and is in agreement with the prediction that this set of compounds would be less stable than chloroquine. Notably, 8b and 27 had intermediate E_H values, as did the isomer of 23 corresponding to peak 2, indicating better metabolic stability than that of the other tested compounds.

The majority of the test compounds showed no measurable degradation in the microsomal matrix devoid of cofactors (i.e. controls), suggesting that there was no major non-cofactor dependent metabolism contributing to their rate of metabolism in liver microsomes. The only exceptions were 12 (peak 2) and 22 (peak 2), both of which demonstrated apparent non-cofactor dependent degradation (which was not observed for the other isomer), potentially indicating a stereo-specific process.

As noted above, the two peaks observed for 12, 22 and 23 during chromatographic LC-MS analysis suggested that there was possible cis - trans isomerism of these compounds under assay conditions. No obvious structural features specific to these compounds which would selectively predispose them to isomerism could be identified. The results suggest that one isomer of a compound could be more susceptible than the other to metabolism (e.g. compound 23) or to additional routes of metabolism (e.g. peak 2 of compounds 12 and 22 to non-cofactor dependent hepatic metabolism). This non-uniform but notable influence of isomerism on biological predisposition, and the fact that such isomerism may be random and unpredictable even among closely related compounds, underscored the need for close attention when studying compounds containing systems potentially capable of isomerization, such as the α-β unsaturated system present in enones and chalcones.

Putative Metabolite Identification

The MS scan data from the metabolic stability assay of each compound was screened for the presence of fragments corresponding to putative metabolites; data acquisition was conducted in MS scan mode (spectra acquired from 100 to 1000 Daltons). It should be noted that, in the absence of authentic metabolites, the chromatographic conditions in each case were optimized for each parent compound and the identities of the putative metabolites were not confirmed by dedicated MS-MS experiments.

With respect to routes/sites of metabolism, the metabolite profiles of all the examined compounds were found to be similar across the two species tested. Oxygenation (P+16, P+32), O-demethylation (P−14) and N-dealkylation (P−174, P−218, P−240, P−296, P−326, P−326) appeared to be the predominant pathways, where P is the molecular mass of the parent compound.

This was consistent with the site-of-metabolism predictions derived from MetaSite (Figure 4), where these three were the major predicted routes of metabolism. However, there were differences between the predicted routes of metabolism and detected metabolites for some of the compounds; for example, no predicted O-demethylation metabolites were detected for 8b and 20, and no N-dealkylation metabolite was detected for 21.

For compounds with the piperazine linkers (22 – 27) a common metabolite with m/z of 248 Daltons was found (i.e. P – 218 [22]; P – 340 [23]; P – 204 [24]; P – 326 [25]; P – 174 [26]; and P – 296 [27]). This was consistent with N-dealkylation of the side chain from the piperazine, and was confirmation of this route of metabolism as predicted by MetaSite for all the piperazinyl-based analogs. For 20, N-dealkylation of the amino-ethoxy linker was observed, resulting in the generation of a metabolite with m/z 179 Daltons (i.e. P – 218); this was not observed for 21, which also bears an amino-ethoxy linker. The putative metabolites from these N-dealkylation reactions are shown in Figure 5.

Figure 5.

Figure showing the structures of the putative metabolites of 22 – 27 (A) and 20 (B)

The potential of the test compounds to undergo primary glucuronidation was assessed in the microsomal incubation by the addition of the cofactor UDPGA. There was no major increase in the rate of degradation of 8b, 20, 21, 23, or chloroquine in microsomal samples containing NADPH and UDPGA relative to those containing NADPH alone, suggesting that these compounds were not susceptible to primary glucuronidation. For 12 and 22, the impact of UDPGA on the overall rate of metabolism could not be assessed, as the rate of NADPH-dependent degradation alone was too rapid. However, for these two compounds, there were no putative metabolites detected with molecular weights consistent with glucuronidation.

In contrast to the results discussed above, there was evidence of primary glucuronidation for 24 – 27 in the microsomal assay, where there was an increase in the relative rates of degradation of test compounds in the presence of UDPGA and/or the detection of a putative metabolite with a molecular weight consistent with glucuronide conjugation [P + 176]. Each of these compounds bear a free phenolic –OH group, and it is therefore not surprising that they are substrates for glucuronidation. This route of metabolism was not predicted by MetaSite, which only predicts cytochrome P450-mediated phase I metabolism.

Physicochemical properties – solubility and partition coefficients

The partition coefficients (logD) of the test compounds were estimated using a modification of the method described by Lombardo and co-workers²⁹ (Table 3). The series exhibited moderate to high partition coefficients, with measured logD_7.4 values ranging from 3.0 to 5.3. Under acidic conditions (pH 3.0), the majority of compounds showed a reduction in logD values, consistent with ionization of the weakly basic groups; however, the logD values of the neutral 8b and 12 were unchanged across the pH range tested. The above predictions were consistent with the in silico predicted results. Compounds 24 and 26 had notably lowed logD_3.0 values, reflecting the dual ionization of the piperazinyl and quinoline nitrogens under acidic conditions, as well as their lower molecular weights (relative to 25 and 27, respectively). The partition coefficients for chloroquine were lower than those observed for the novel compounds.

Table 3.

Results from the experimental determination of solubility and partition coefficients

Compound	LogDa		Solubility (µg/mL)
	pH 3.0	pH 7.4	pH 2.0	pH 6.5
8b	4.3	4.3	< 1.6	< 1.6
12	3.5	3.6	3.1 – 6.3	3.1 – 6.3
20	1.8	3.0	50.0 – 100.0	6.3 – 12.5
21	2.8	4.2	> 100	3.1 – 6.3
22	1.2	4.2	> 100	1.6 – 3.1
23	2.2	5.3	> 100	< 1.6
24	0.9	3.9	> 100	1.6 – 3.1
25	2.1	5.0	> 100	< 1.6
26	0.8	4.2	> 100	< 1.6
27	2.0	5.3	> 100	< 1.6
Chloroquine	0.4	2.0	> 100b	>100b

^aValue measured using the chromatographic gLogD technique.

^bStock solution of compound prepared in water at 10 mg/mL concentration

The solubility of the test compounds was estimated using the turbidimetric method described by Bevan and Lloyd,³⁰ applying standard test buffers (pH 2.0 and 6.5). As predicted, the solubility characteristics of the series were relatively poor at pH 6.5, with only 20 showing moderate solubility under neutral conditions (Table 3). Under acidic conditions, compounds 20 –27 showed improved solubility consistent with ionization of the basic centres within the structures. 8b and 12 showed no increase in solubility under acidic conditions, consistent with the neutral characteristics of these triazole substituted compounds. Chloroquine showed good solubility under both neutral and acidic conditions.

Mechanistic studies

A limited investigation was carried out on the synthesized analogs in an attempt to elucidate their mechanisms of action. The mechanisms considered for investigation were based on literature accounts of the possible mechanisms by which chalcones and quinoline-containing compounds exert their antimalarial activities.

Inhibition of hemozoin formation is an attractive target for antimalarial drugs, and quinoline-based antimalarials such as chloroquine and amodiaquine are proposed to exert their antimalarial activity via this mechanism.^17;31;32 The plasmodial cysteine proteases involved in hemoglobin catabolism, such as falcipain 2, have also been proposed as potential antimalarial targets.^33;34 By possessing α-β unsaturated carbonyl groups and thereby able to act as Michael acceptors, chalcones could potentially irreversibly inhibit the thiol-containing malarial cysteine proteases;¹⁸ several quinolinyl chalcones which are moderate inhibitors of falcipain 2 have been identified.³⁵ Some chalcones have also been shown to inhibit new permeability pathways (NPPs), which are parasite-induced solute channels in erythrocyte membranes.³⁶ Considering these potential targets, a selection of intermediates and analogs were evaluated for inhibition of β-hematin formation, falcipain 2 activity, and NPP function (sorbitol-induced hemolysis of infected erythrocytes) (Table 4).

Table 4.

Results from the mechanistic studies

Compd	β-hematin formation (IC50)a	Falcipain 2 (IC50;µM)b	Inhibition of sorbitol- induced hemolysis (%)c
15a	>5	>50	21.1
15b	>5	>50	ND
16a	>5	>50	27.9
16b	1.9	>50	ND
17	4.9	>50	0.9
18	3.9	>50	ND
19	>5	>50	5.8
20	0.6	>50	12.3
21	0.3	22.9	ND
22	0.5	>50	ND
23	0.2	11.7	ND
24	0.5	>50	16.8
25	0.3	9.1	3.9
26	0.5	29.9	11.6
27	0.2	10.0	ND
8b	1.6	10.8	15.4
Chloroquine	1.91	-	-
E64	-	0.055	-

^aThe IC₅₀s are averages of triplicate determinations and are reported as equivalents of the compound (relative to hematin) that inhibit the formation of β-hematin by 50%.

^bAverages of duplicate determinations.

^cInhibition of sorbitol-induced lysis at a concentration of 10 µM.

ND: Not determined

Compound 8b, the most potent compound in vitro, was only slightly more potent (IC₅₀ 1.6 equiv.) than chloroquine in inhibiting β-hematin formation. Of the analogs tested, compounds 20 – 27 were all very potent (more potent than chloroquine) inhibitors of β-hematin formation. Compounds 23 and 27 (which were also some of the more active of these analogs) had the lowest values of 0.2 equiv. Thus, inhibition of hemozoin formation may be a primary mechanism of antimalarial action of these compounds. The intermediates 17 – 19 showed only moderate inhibition of β-hematin formation. Interestingly, the hydroxylated chalcone 16b, which showed activity comparable to that of chloroquine, while the chalcone and enone intermediates 15a, 15b and 16a showed negligible β-hematin inhibition activity.

The analogs showed moderate to poor inhibition of falcipain 2 relative to E64, a known, potent inhibitor of the enzyme; only compound 8b and the analogs 21, 23, 25, 26 and 27 had lower micromolar IC₅₀ values. The chalcone/enone intermediates 15 – 16 and the quinoline intermediates 17 – 19 did not show appreciable falcipain 2 inhibition even at a concentration of 50 µM. The relatively low potencies of compounds 20, 22 and 24 as falcipain 2 inhibitors suggested that this was not the primary mechanism of antimalarial action of these compounds. However, the most potent analogs against falcipain 2, 23, 25, and 27, were among the most potent against cultured parasites, suggesting that protease inhibition may play a role in the mechanism of action of these compounds.

All analogs tested were poor inhibitors of sorbitol-induced lysis, with 24, the most active compound, with only 17% inhibition. This result discounts inhibition of NPPs as a mechanism of action for these analogs.

Summary and conclusions

8b and 12 were selected for structural modification, with the aim of identifying analogs with improved solubility and which retained antimalarial potency. Eight analogs were proposed, and these were characterized in silico for selected physicochemical and ADME parameters. These analogs were predicted to have improved solubilities relative to the parent compounds, particularly at low pH; they retained acceptable predicted Caco2 permeability, but were identified as possible substrates for metabolism, principally N-dealkylation by hepatic cytochromes.

The proposed analogs were then synthesized and tested, and were found to exhibit notable in vitro antiplasmodial activity, though none were as active as 8b. Compounds 25 and 27 were the most active of the analogs. 19, an intermediate in the synthesis (and a putative metabolite) of some of the analogs was also found to exhibit notable antiplasmodial activity.

There was no evidence of significant inter-species differences in the relative rates of microsomal metabolism of these compounds between mice and humans. However, most of the compounds exhibited high E_H values (>0.7), suggesting considerable susceptibility to hepatic metabolism. Oxygenation, O-demethylation and N-dealkylation appeared to be the predominant metabolic pathways, which was consistent with the site-of-metabolism predictions derived from MetaSite. There was also evidence of primary glucuronidation for 24 – 27. The majority of the compounds showed no susceptibility to metabolism by enzymes other than hepatic cytochromes and transferases.

Inhibition of hemozoin formation was found to be a likely mechanism by which the tested analogs exert their antiplasmodial activity. Analogs 20 – 27 were all more potent than chloroquine in inhibiting β-hematin formation. Inhibition of falcipain 2 could also contribute to the observed antiplasmodial activities of some of these compounds, while inhibition of parasite-induced transport channels did not appear to be a notable mechanism of action.

The proposed analogs were by no means exhaustive, but represented compounds that did not represent significant deviations from the molecules known to be active, and were therefore more likely to retain activity. They consisted of a small range of compounds that that could be conveniently synthesized, and which offered possible improvements on certain physicochemical and/or ADME properties of the parent compounds, particularly solubility. In addition, they opened up possibilities for further derivatization in future.

The more promising compounds 25 and 27 were evaluated for their cytotoxicity against the Chinese Hamster Ovarian (mammalian) cell line. No cytotoxicity was observed at the highest concentration (100 µM) tested, an observation that was consistent with the findings previously reported for compound 8b.¹⁶

Experimental

To guide in the interpretation of the ¹H-NMR data presented below, the complete (and numbered) structures of all the intermediates and target compounds can be found in the Supporting Information. This material is available free of charge via the Internet at http://pubs.acs.org. Purity was determined by combustion analysis, and all compounds were confirmed to have > 95% purity.

7.1 General Method A for the preparation of compounds 15a and 15b

Vanillin or 4-hydroxybenzaldehyde (20 mmol) was dissolved in 30 mL acetone, and 5 mL of 5 M HCl was then added gradually to the stirring mixture. The reaction mixture was then stirred at room temperature for 24 hours. The reaction mixture was then neutralized with 2.5 M NaOH, after which the acetone in the mixture was removed at reduced pressure. The remaining compound mixture was then taken up in 100 mL EtOAc, washed three times with water, dried over anhydrous Na₂SO₄, concentrated in vacuo and purified by column chromatography (EtOAc/Hexane) to yield the pure target compound.

7.1.1 4-(4-Hydroxy-3-methoxy-phenyl)-but-3-en-2-one 15a

Pale yellow solid (3.2 g, 84%); mp 124 – 125°C (from EtOAc/Hex); R_f (EtOAc:Hex 4:6) 0.30; IR ν_max (KBr)/cm⁻¹ 3261 (Ar O-H), 3000 (Ar C-H), 1668 (C=O), 1581 (Ar C=C); δ_H (400 MHz, CDCl₃) 7.46 (1H, d, J 16.2, H-a), 7.09 (1H, dd, J 1.9 and 8.2, H-2), 7.07 (1H, d, J 1.9, H-1), 6.94(1H, d, J 8.1, H-3), 6.60 (1H, d, J 16.2, H-b), 5.98 (1H, s, H-5), 3.95 (3H, s, H-4), 2.37 (3H, s, H-b’); δ_C (100 MHz, CDCl₃) 198.3, 150.4, 146.9, 143.5, 126.5, 125.0, 123.8, 114.9, 109.4, 55.8, 24.5; LRMS (EI) m/z 193.1 (MH⁺); Anal. (C₁₁H₁₂O₃) C H.

7.1.2 4-(4-Hydroxy-phenyl)-but-3-en-2-one 15b

Dark yellow solid (2.4 g, 76%); mp 81 – 82°C (from EtOAc/Hex); R_f (EtOAc:Hex 4:6) 0.40; IR ν_max (KBr)/cm⁻¹ 3150 (Ar C-H), 1666 (C=O), 1590 (Ar C=C); δ_H (400 MHz, CDCl₃) 7.50 (1H, d, J 16.2, H-a), 7.42 (2H, d, J 8.7, H-1, 2), 6.88 (2H, d, J 8.7, H-3, 4), 6.58 (1H, d, J 16.2, H-b), 2.35 (3H, s, H-b’); δ_C (100 MHz, CDCl₃) 200.0, 159.1, 144.8, 130.4 (2C), 126.5, 124.3 (2C), 116.2, 27.2; LRMS (EI) m/z 163.3 (MH⁺); Anal. (C₁₀H₁₀O₂) C H.

7.2 General Method B for the preparation of compounds 16a and 16b

Vanillin or 4-hydroxybenzaldehyde (20 mmol) and 2,4-dimethoxyacetophenone (4.0 g, 22 mmol) were dissolved in 20 mL MeOH. 5 mL of 5 M HCl was then added gradually to the stirring mixture. The reaction mixture was then stirred at 50°C for 24 hours. The reaction mixture was then allowed to cool to ambient temperature and neutralized with 2.5 M NaOH, after which the MeOH in the mixture was removed under reduced pressure. The remaining compound mixture was then taken up in 100 mL EtOAc, washed three times with water, dried over anhydrous Na₂SO₄, concentrated in vacuo and purified by column chromatography (EtOAc/Hexane) to yield the pure target compound.

7.2.1 1-(2,4-Dimethoxy-phenyl)-3-(4-hydroxy-3-methoxy-phenyl)-propenone 16a

Yellow solid (4.3 g, 68%); mp 92 – 94°C (from EtOAc /Hex); R_f (EtOAc:Hex 4:6) 0.30; IR ν_max (KBr)/cm⁻¹ 3506 (Ar O-H), 3113 (Ar C-H), 1640 (C=O), 1598 (Ar C=C); δ_H (400 MHz, CDCl₃) 7.74 (1H, d, J 8.6, H-3’), 7.61 (1H, d, J 15.7, H-a), 7.35 (1H, d, J 15.7, H-b), 7.17 (1H, dd, J 1.9 and 8.3, H-2), 7.10 (1H, d, J 1.9, H-1), 6.94 (1H, d, J 8.3, H-3), 6.58 (1H, dd, J 2.3 and 8.6, H-2’), 6.51 (1H, d, J 2.3, H-1’), 5.93 (1H, s, H-5), 3.94 (3H, s, H-4’), 3.91 (3H, s, H-5’), 3.88 (3H, s, H-4); δ_C (100 MHz, CDCl₃) 190.8, 163.9, 160.2, 147.8, 146.7, 142.6, 132.6, 128.0, 125.1, 122.8, 122.5, 114.8, 110.2, 105.1, 98.8, 55.9, 55.8, 55.5; LRMS (EI) m/z 314.0 (M⁺); Anal. (C₁₈H₁₈O₅.) C H.

7.2.2 1-(2,4-Dimethoxy-phenyl)-3-(4-hydroxy-3-methoxy-phenyl)-propenone 16b

Yellow solid (3.2 g, 59%); mp 92 – 94°C (from EtOAc /Hex); R_f (EtOAc:Hex 4:6) 0.27; IR ν_max (KBr)/cm⁻¹ 3204 (Ar O-H), 1635 (C=O), 1600(Ar C=C); δ_H (400 MHz, CDCl₃) 7.53 (2H, d, J 8.7, H-1, 2), 7.51 (1H, d, J 8.7, H-3’), 7.43 (1H, d, J 15.8, H-a), 7.28 (1H, d, J 15.7, H-b), 6.78 (2H, d, J 8.5, H-3, 4), 6.62 (1H, d, J 2.1, H-1’), 6.58 (1H, dd, J 2.0 and 8.6, H-2’), 3.83 (3H, s, H-4’), 3.79 (3H, s, H-5’); δ_C (100 MHz, CDCl₃) 189.9, 164.1, 160.4, 160.3, 142.4, 132.3, 130.8, 126.3, 124.3, 122.3, 116.4, 106.4, 99.2, 56.4, 56.0; LRMS (EI) m/z 284.0 (M₊); Anal. (C₁₇H₁₆O₄)- C H.

7.3 2-(7-Chloro-quinolin-4-ylamino)-ethanol 17

4,7-dichloroquinoline (2 g, 10mmol) was heated with stirring in neat ethanolamine (10 mL) at 130 °C for 5 hours. The mixture was then allowed to cool to room temperature, during which the compound 17 precipitated. This crude mixture was then suspended in water, filtered and crystallized from hot MeOH to yield pure 17 as a white solid (1.97 g, 89%). Mp 215 – 216°C (from MeOH) (lit. 214 °C)³⁷ Rf (MeOH/EtOAc, 1:9) 0.20; δ_H (400 MHz; DMSO-d₆) 8.41 (1H, d, J 5.4, H2), 8.30 (1H, d, J 9.0, H5), 7.81 (1H, d, J 2.2, H3), 7.46 (1H, dd, J 2.2 and 8.9, H4), 7.26 (1H, t, J 5.2, H-6), 6.52 (1H, d, J 5.5, H1), 4.85 (1H, s, H-9), 3.70 (2H, t, J 5.9, H8), 3.38 (2H, q, J 5.8, H7).

7.4 2-[4-(7-Chloro-quinolin-4-yl)-piperazin-1-yl]-ethanol 18

4,7-dichloroquinoline (2 g, 10mmol) was dissolved in 10 mL anhydrous DMF, after which K₂CO₃ (2.1 g, 15 mmol) and 1-(2-hydroxyethyl)piperazine (1.3 g, 10 mmol) were added. This reaction mixture was then stirred at 80 °C for 24 hrs, upon which TLC indicated completion of the reaction. The mixture was then diluted in 100 mL EtOAc, washed with water (2 × 30 mL), dried over anhydrous Na₂SO₄ and concentrated under reduced pressure. The resultant oily product crystallized on standing, and these crystals were washed with hexane and dried to yield 18 as pale yellow crystals (1.83 g, 63%). M.p. 111 – 112°C (from EtOAc); R_f (MeOH/EtOAc, 1:9) 0.24; IR ν_max (CHCl₃)/cm⁻¹ 3222 (O-H), 2903 (C-H), 1677 (C=O), 1569 (Ar C=C); δ_H (400 MHz, DMSO-d₆) 8.66 (1H, d, J 5.0, H-2), 7.99 (1H, d, J 9.0, H-5), 7.95 (1H, d, J 2.2, H-3), 7.51 (1H, dd, J 2.2 and 9.0, H-4), 6.96 (1H, d, J 5.1, H-1), 4.33 (1H, t, J 5.3, H-12), 3.55 (2H, q, J 6.0, H-11), 3.17 (4H, t, J 4.8, H-6, 9), 2.70 (4H, t, J 4.8, H-7, 8), 2.51 (2H, t, J 6.2, H-10); δ_C (100 MHz, CDCl₃) 156.2, 151.9, 149.6, 133.4, 127.9, 125.9, 125.5, 121.3, 109.1, 60.0, 58.5 (2C), 52.8 (2C), 51.7; LRMS (EI) m/z 290.5 (M⁺); Anal. (C₁₅H₁₈ClN₃O) C H N.

7.5 7-Chloro-4-piperazin-1-yl-quinoline 19

4,7-dichloroquinoline (2 g, 10mmol) was dissolved in 15 mL anhydrous NMP, after which K₂CO₃ (2.8 g, 20 mmol) and piperazine (5.2 g, 60 mmol) were added. This reaction mixture was then stirred at 135 °C for 4 hrs, upon which TLC indicated completion of the reaction. The mixture was then diluted in 100 mL EtOAc, washed with water (2 × 30 mL), dried over anhydrous Na₂SO₄, concentrated under reduced pressure and purified by column chromatography (NH₄OH/MeOH/EtOAC, 1:9:90) to yield 19 as an off-white solid (1.9 g, 78%). M.p. 111 – 112°C (lit. 113 – 115 °C)²⁶; δ_H (400 MHz, DMSO-d₆) 8.68 (1H, d, J 5.0, H-2), 8.01 (1H, d, J 9.0, H-3), 7.89 (1H, d, J 2.0, H-5), 7.52 (1H, dd, J 2.1 and 9.0, H-4), 6.95 (1H, d, J 5.0, H-1), 3.05 – 3.11 (4H, m, H-6, 9), 2.92 – 2.97 (4H, m, H-7, 8), 3.19 (1H, br s, H-10); δ_C (100 MHz, DMSO-d₆) 157.4, 152.6, 150.2, 133.9, 128.5, 126.6, 121.9, 109.7, 53.7 (2C), 45.9 (2C).

7.6 General Method C for the preparation of compounds 20 – 23

Hydroxylated chloroquinoline intermediate 17 or 18 (1 eq.), enone 15a or chalcone 16a (1 eq.) and triphenylphosphine (1.2 eq.) were dissolved in anhydrous DCM under a N₂ atmosphere. The solution was cooled to 0°C, and di-isopropyl azodicarboxylate (DIAD, 1.2 eq.) was added slowly over 5 min. The stirred reaction mixture was then allowed to warm to ambient temperature, and was further stirred for 6 hrs. The mixture was then concentrated under reduced pressure, and the target compound purified by silica column chromatography (5% MeOH in EtOAc).

7.6.1 4-{4-[2-(7-Chloro-quinolin-4-ylamino)-ethoxy]-3-methoxy-phenyl}-but-3-en-2-one 20

Yellow solid (289 mg, 73%); mp 143 – 144°C (from MeOH/EtOAc); R_f (MeOH/EtOAc 1:9) 0.23; IR ν_max (KBr)/cm⁻¹ 3420 (N-H), 2972 (Ar C-H), 1686 (C=O), 1584 (Ar C=C); δ_H (400 MHz, CDCl₃) 8.57 (1H, d, J 5.3, H-9), 7.99 (1H, d, J 1.8, H-10), 7.76 (1H, d, J 8.9, H-12), 7.46 (1H, d, J 16.2, H-a), 7.39 (1H, dd, J 1.9 and 8.9, H-11), 7.13-7.11 (2H, m, H-1, 2), 6.98 (1H, d, J 8.0, H-3), 6.63 (1H, d, J 16.2, H-b), 6.49 (1H, d, J 5.3, H-8), 5.70 (1H, s, H-7), 4.39 (2H, t, H-5.0, H-5), 3.92 (3H, s, H-4), 3.73 (2H, q, J 5.0, H-6), 2.37 (3H, s, H-a’); δ_C (100 MHZ, CDCl₃) 198.2, 151.7, 150.1, 149.8, 148.9, 142.9, 135.1, 128.9, 125.9, 125.5, 122.6, 121.2, 117.4, 114.7, 110.5, 99.3, 67.7, 55.8, 42.3, 27.4, 21.9; LRMS (EI) m/z 395.3 (M - H)⁺; Anal. (C₂₂H₂₁ClN₂O₃) C H N.

7.6.2 3-{4-[2-(7-Chloro-quinolin-4-ylamino)-ethoxy]-3-methoxy-phenyl}-1-(2,4-dimethoxy-phenyl)-propenone 21

Pale yellow solid (310 mg, 60%); mp 164 – 166°C (from MeOH/EtOAc); R_f (MeOH:EtOAc 1:9) 0.27; IR ν_max (KBr)/cm⁻¹ 3406 (N-H), 2942 (Ar C-H), 1646 (C=O), 1584 (Ar C=C); δ_H (400 MHz, CDCl₃) 8.57 (1H, d, J 5.3, H-9), 7.98 (1H, d, J 2.1, H-10), 7.78-7.74 (2H, m, H-3’, 12), 7.61 (1H, d, J 15.7, H-a), 7.39 (1H, d, J 15.7, H-b), 7.38 (1H, dd, J 2.1 and 8.9, H-11), 7.17 (1H, dd, J 1.7 and 8.2, H-2), 7.14 (1H, d, J 1.7, H-1), 6.96 (1H, d, J 8.2, H-3), 6.57 (1H, dd, 2.2 and 8.6, H-2’), 6.52 (1H, d, J 2.2, H-1’), 6.48 (1H, d, J 5.3, H-8), 5.78 (1H, t, J 5.5, H-7), 4.39 (2H, t, J 5.2, H-5), 3.92 (3H, s, H-4’), 3.90 (3H, s, H-5’), 3.88 (3H, s, H-4), 3.72 (2H, q, J 5.2, H-6); δ_C (100 MHZ, CDCl₃) 190.5, 164.1, 160.3, 152.0, 150.1, 149.6, 149.4, 149.2, 141.8, 134.9, 132.7, 130.2, 128.9, 126.1, 125.4, 122.3, 122.2, 121.2, 117.5, 115.0, 111.2, 105.2, 99.3, 98.8, 67.8, 55.9, 55.7, 55.5, 42.3; LRMS (EI) m/z 517.1 (M - H)⁺; Anal. (C₂₉H₂₇ClN₂O₅) C H N.

7.6.3 4-(4-{2-[4-(7-Chloro-quinolin-4-yl)-piperazin-1-yl]-ethoxy}-3-methoxy-phenyl)-but-3-en-2-one 22

Pale yellow (316 mg, 68%); mp 61 – 62°C (from MeOH/EtOAc); R_f (MeOH/EtOAc 1:9) 0.26; IR ν_max (KBr)/cm⁻ 2977 (Ar C-H), 1714 (C=O), 1594 (Ar C=C); δ_H (300 MHz, CDCl₃) 8.72 (1H, d, J 5.0, H-12), 8.04 (1H, d, J 2.0, H-13), 7.93 (1H, d, J 8.9, H-15), 7.46 (1H, d, J 16.1, H-a), 7.41 (1H, dd, J 2.2 and 9.0, H-14), 7.12 (1H, dd, J 1.9 and 8.3, H-2), 7.09 (1H, d, J 1.9, H-1), 6.92 (1H, d, J 8.2, H-3), 6.83 (1H, d, J 5.0, H-11), 6.61 (1H, d, J 16.1, H-b), 4.25 (2H, t, J 5.9, H-5), 3.90 (3H, s, H-4), 3.28-3.25 (4H, m, H-8, 9), 2.99 (2H, t, J 5.9, H-6), 2.91-2.88 (4H, m, H-7, 10), 2.36 (3H, s, H-a’); δ_C (75 MHZ, CDCl₃) 198.0, 156.9, 151.9, 150.5, 150.1, 149.7, 143.2, 134.8, 128.8, 127.8, 126.1, 125.4, 125.1, 122.7, 113.0, 110.3, 108.9, 67.0, 56.8, 55.9, 53.4, 52.9, 52.1, 27.3, 21.9; LRMS (EI) m/z 464.2 (M - H)⁺; Anal. (C₂₆H₂₈ClN₃O₃) C H N.

7.6.4 3-(4-{2-[4-(7-Chloro-quinolin-4-yl)-piperazin-1-yl]-ethoxy}-3-methoxy-phenyl)-1-(2,4-dimethoxy-phenyl)-propenone 23

Yellow solid (435 mg, 74%); mp 62 – 63°C (from MeOH/EtOAc); R_f (MeOH/EtOAc 1:9) 0.30; IR ν_max (KBr)/cm⁻¹ 2938 (C-H), 1648 (C=O), 1600 (Ar C=C); δ_H (300 MHz, CDCl₃) 8.66 (1H, d, J 5.0, H-12), 8.01 (1H, d, J 9.0, H-15), 7.95 (1H, d, J 2.1, H-13), 7.56 (1H, d, J 8.5, H-3’), 7.53 (1H, dd, J 2.3 and 9.0, H-14), 7.46 (1H, d, J 15.8, H-a), 7.37 (1H, d, J 15.8, H-b), 7.30 (1H, d, J 1.9, H-1), 7.23 (1H, dd, J 1.9 and 8.4, H-2), 7.05 (1H, d, J 8.4, H-3), 6.99 (1H, d, J 5.1, H-11), 6.67 (1H, d, J 2.2, H-1’), 6.62 (1H, dd, J 2.2 and 8.6, H-2’), 4.19 (2H, t, J 5.8 H-5), 3.87 (3H, s, H-4’), 3.84 (3H, s, H-5’), 3.83 (3H, s, H-4), 3.19 (4H, t, J 4.8, H-8, 9), 2.85 (2H, t, J 5.8, H-6), 2.79 (4H, t, J 4.8, H-7, 10); δ_C (75 MHZ, CDCl₃) 189.7, 163.5, 159.9, 156.2, 151.9, 150.1, 149.6, 149.2, 141.6, 133.4, 131.5, 127.9, 125.8, 125.5, 125.1, 122.3, 121.7, 121.2, 113.3, 111.4, 109.2, 105.8, 98.7, 65.6, 56.3, 55.8 (2C), 52.8 (2C), 51.7 (2C); LRMS (EI) m/z 586.0 (M - H)⁺; Anal. (C₃₃H₃₄ClN₃O₅) C H N.

7.7 General Method D for the preparation of compounds 24 – 27

The appropriate hydroxylated chalcone or enone (15 or 16) (1 eq.) and the piperazinylquinoline intermediate, 19 (1 eq) were dissolved in EtOH. Formaldehyde (37% solution in water, 10 eq.) was then added to the stirring solution, which was then refluxed at 110°C for 12 hrs. The reaction mixture was then allowed to cool to ambient temperature, after which it was concentrated under reduced pressure and purified by column chromatography (EtOAc) to yield the pure target compounds.

7.7.1 4-{3-[4-(7-Chloro-quinolin-4-yl)-piperazin-1-ylmethyl]-4-hydroxy-5-methoxyphenyl}-but-3-en-2-one 24

Yellow solid (58%); mp 91 – 92°C (from EtOAc); R_f (EtOAc) 0.15; IR ν_max (KBr)/cm⁻¹ 2943 (C-H), 1662 (C=O), 1575 (Ar C=C); δ_H (300 MHz, CDCl₃) 8.73 (1H, d, J 5.0, H-11), 8.06 (1H, d, J 2.1, H-12), 7.89 (1H, d, J 9.0, H-14), 7.43 (1H, d, J 16.1, H-a), 7.40 (1H, dd, J 2.1 and 9.0, H-13), 7.03 (1H, d, J 1.5, H-2), 6.91 (1H, d, J 1.5, H-1), 6.82 (1H, d, J 5.0, H-10), 6.58 (1H, d, J 16.1, H-b), 3.91 (3H, s, H-3), 3.89 (2H, s, H-5), 3.28 – 3.34 (4H, m, H-7, 8), 2.90 – 2.96 (4H, m, H-6, 9), 2.35 (3H, s, H-a’); δ_C (75 MHz, CDCl₃) 198.1, 156.4, 151.6, 149.9, 149.6, 148.4, 143.5, 135.2, 128.8, 126.5, 125.8, 124.9, 124.8, 122.5, 121.7, 120.9, 110.1, 109.1, 60.7, 55.9, 52.4 (2C), 51.9 (2C), 27.4; LRMS (EI) m/z 450.7 (M - H)⁺; Anal. (C₂₅H₂₆ClN₃O₃.H₂O) C H N.

7.7.2 3-{3-[4-(7-Chloro-quinolin-4-yl)-piperazin-1-ylmethyl]-4-hydroxy-5-methoxyphenyl}-1-(2,4-dimethoxy-phenyl)-propenone 25

Deep yellow solid (66%); mp 95 – 97°C (from EtOAc); R_f (EtOAc) 0.20; IR ν_max (KBr)/cm⁻¹ 2938 (C-H), 1648 (C=O), 1597 (Ar C=C); δ_H (400 MHz, CDCl₃) 8.71 (1H, d, J 5.1, H-11), 8.10 (1H, d, J 1.9, H-12), 7.89 (1H, d, J 9.0, H-14), 7.72 (1H, d, J 8.6, H-3’), 7.58 (1H, d, J 15.8, H-a), 7.44 (1H, dd, J 1.9 and 8.9, H-13), 7.33 (1H, d, J 15.7, H-b), 7.09 (1H, d, J 1.8, H-2), 6.97 (1H, d, J 1.8, H-1), 6.84 (1H, d, J 5.1, H-10), 6.57 (1H, dd, J 1.7 and 8.6, H-2’), 6.51 (1H, d, J 1.8, H-1’), 3.92 (3H, s, H-4’), 3.91 (5H, s, H-5, 5’), 3.87 (3H, s, H-3), 3.32 – 3.55 (4H, m, H-7, 8), 2.90 – 2.96 (4H, m, H-6, 9); δ_C (100 MHz, CDCl₃) 190.9, 163.9, 160.3, 156.7, 151.2, 149.4, 149.2, 148.3, 142.5, 135.5, 132.6, 128.4, 126.6, 124.9, 124.8, 122.7, 122.6, 121.6, 120.7, 110.8, 110.5, 109.0, 105.2, 98.8, 60.6, 56.0, 55.8, 55.5, 52.3 (2C), 51.8 (2C); LRMS (EI) m/z 572.7 (M - H)⁺; Anal. (C₃₂H₃₂ClN₃O₅.H₂O) C H N.

7.7.3 4-{3-[4-(7-Chloro-quinolin-4-yl)-piperazin-1-ylmethyl]-4-hydroxy-phenyl}-but-3-en-2-one 26

Pale yellow solid (45%); mp 83 – 85°C (from EtOAc); R_f (EtOAc) 0.35; IR ν_max (KBr)/cm⁻¹ 2823 (C-H), 1663 (C=O), 1578 (Ar C=C); δ_H (300 MHz, CDCl₃) 8.71 (1H, d, J 5.1, H-11), 8.07 (1H, d, J 2.1, H-12), 7.89 (1H, d, J 9.0, H-14), 7.44 (1H, d, J 16.4, H-a), 7.43 (1H, dd, J 2.1 and 9.0, H-13), 7.41 (1H, dd, J 2.3 and 8.4, H-2), 7.30 (1H, d, J 2.3, H-1), 6.86 (1H, d, J 8.4, H-3), 6.85 (1H, d, J 5.1, H-10), 6.58 (1H, d, J 16.3, H-b), 3.88 (2H, s, H-5), 3.28 – 3.35 (4H, m, H-7, 8), 2.85 – 2.93 (4H, m, H-6, 9), 2.34 (3H, s, H-a’); δ_C (75 MHz, CDCl₃) 198.2, 160.1, 156.5, 151.5, 149.7, 143.2, 135.3, 129.8, 129.1, 128.7, 126.5, 126.0, 124.8, 124.6, 121.7, 121.1, 116.9, 109.1, 61.2, 52.4 (2C), 51.9 (2C), 27.4; LRMS (EI) m/z 420.7 (M - H)⁺; Anal. (C₂₄H₂₄ClN₃O₂.H₂O) C H N.

7.7.4 3-{3-[4-(7-Chloro-quinolin-4-yl)-piperazin-1-ylmethyl]-4-hydroxy-phenyl}-1-(2,4-dimethoxy-phenyl)-propenone 27

Yellow solid (49%); mp 85 – 87°C (from EtOAc); R_f (EtOAc) 0.40; IR ν_max (KBr)/cm⁻¹ 2935 (C-H), 1649 (C=O), 1598 (Ar C=C); δ_H (400 MHz, CDCl₃) 8.68 (1H, d, J 5.0, H-11), 8.07 (1H, d, J 1.8, H-12), 7.89 (1H, d, J 9.0, H-14), 7.69 (1H, d, J 8.6, H-3’), 7.57 (1H, d, J 15.7, H-a), 7.46 (1H, dd, J 1.9 and 8.5, H-2), 7.41 (1H, dd, J 1.8 and 8.9, H-13), 7.32 (1H, d, J 15.7, H-b), 7.24 (1H, d, J 1.9, H-1), 6.83 (1H, d, J 8.3, H-3), 6.82 (1H, d, J 4.9, H-10), 6.52 (1H, dd, J 2.0 and 8.6, H-2’), 6.47 (1H, d, J 2.0, H-1’), 3.87 (3H, s, H-4’), 3.85 (2H, s, H-5), 3.83 (3H, s, H-5’), 3.28 – 3.34 (4H, m, H-7, 8), 2.84 – 2.90 (4H, m, H-6, 9); δ_C (100 MHz, CDCl₃) 190.6, 163.9, 160.2, 159.6, 156.7, 151.1, 149.3, 142.1, 135.5, 132.6, 129.6, 129.3, 128.4, 127.1, 126.6, 124.8, 124.6, 122.5, 121.5, 120.8, 116.9, 108.9, 105.1, 98.7, 61.55, 55.7, 55.4, 52.4 (2C), 51.8 (2C); LRMS (EI) m/z 542.8 (M⁺); Anal. (C₃₁H₃₀ClN₃O₄.H₂O) C H N.

7.8 Evaluation of in vitro antiplasmodial activity

The in vitro antiplasmodial activities of the compounds were evaluated against D10, Dd2 and W2 strains of P. falciparum.

The parasites were maintained in continuous in vitro culture by standard methods.³⁸ The cultured parasites were, whenever necessary, synchronized by treatment with 5% D-sorbitol (Sigma) at the ring stage.³⁹

For the studies on the D10 and Dd2 strains, the assays were carried out as previously reported by Guantai et al. 2010.¹⁶ The assays on the W2 strain were conducted as previously reported (Shenai et al. 2003)⁴⁰ with parasites cultured in standard medium including 10% human serum, except that parasite development was assessed with a FACS-based assay, as previously described (Sijwali and Rosenthal, 2004).⁴¹ The final concentration of DMSO that the parasites were exposed to did not exceed 0.5%, which has no measurable effect on parasite viability.

7.9 LogD Measurement

Partition coefficient values (LogD) of the test compounds were estimated by correlation of their chromatographic retention properties against the characteristics of a series of standard compounds with known partition coefficient values. The method employed was a gradient HPLC based derivation of the method developed by Lombardo and co-workers, 2001.²⁹

7.10 Solubility Estimates

Compound in DMSO was spiked into either pH 6.5 phosphate buffer or 0.01M HCl (approximately pH 2.0) with the final DMSO concentration being 1%. Samples were then analysed via nephelometry to determine a solubility range, as described by Bevan and Lloyd, 2000.³⁰

7.11 Evaluation of β-hematin inhibition

β-hematin inhibition assays were conducted as previously reported.⁴²

7.12 Evaluation of Falcipain 2 inhibition

Falcipain 2 inhibition assays were carried out as previously described, except that the proteolytic substrate was benzyloxycarbonyl-Phe-Leu-7-amino-4-methyl-coumarin.⁴³

7.13 Evaluation of inhibition of sorbitol-induced lysis

The method described by Go et al. 2004³⁶ was applied for this assay, with modifications. Briefly, chloroquine-sensitive P. falciparum D10 was cultured and trophozoite-stage infected erythrocytes (10% parasitemia) were harvested by centrifugation (750 r.p.m., 3 min), washed twice with a solution of NaCl (150 mM) – HEPES (20 mM) (pH 7.4; 304 mosM), and suspended in the same solution to give a hematocrit of about 50%. Test compounds were dissolved in dimethylsulfoxide (DMSO) to give 10 mM stock solutions, which were diluted further with a solution comprising 300 mM sorbitol – 20 mM HEPES (pH 7.4; 345 mosM) to give the test solutions of 100 µM.

In 10 mL screw-cap centrifuge tubes, 100 µL of the test solutions were separately combined with an additional 700 µL of the sorbitol-HEPES buffer and 200 µl of the parasitized cell suspension in NaCl-HEPES buffer. The final concentration of the test compounds was 10 µM. For the negative control, 800 µL of the sorbitol-HEPES buffer and 200 µl of the parasitized cell suspension in NaCl-HEPES buffer were transferred into a 10 mL screw-cap centrifuge tube. For the blank, 800 µL of the sorbitol-HEPES buffer and 200 µl of non-parasitized erythrocyte suspension (50% hematocrit) in NaCl-HEPES buffer were transferred into a 10 mL screw-cap centrifuge tube.

The mixtures were incubated at 37°C for 15 min and centrifuged, and aliquots of the supernatant (100 µL) were dispensed into 96-well plates. The absorbances at 560 µm were then determined and used to estimate the amount of hemoglobin released and, by extension, the degree of sorbitol-induced hemolysis. The absorbances of the control and blank wells were determined concurrently, and corrections to the absorbances measured in the test wells were made as appropriate. The percentage inhibition of sorbitol-induced lysis was then calculated for each test compound.

List of abbreviations

ADME

absorption, distribution, metabolism and excretion

DIAD

diisopropyl azadicarboxylate

NADP(H)

Nicotinamide Adenine Dinucleotide Phosphate (reduced)

PK

pharmacokinetic

PPh₃

triphenylphosphine

UDPGA

Uridine Diphosphate Glucuronic Acid

WHO

World Health Organization