Contrasting Ex Vivo Efficacies of “Reversed Chloroquine” Compounds in Chloroquine-Resistant Plasmodium falciparum and P. vivax Isolates

Abstract

Chloroquine (CQ) has been the mainstay of malaria treatment for more than 60 years. However, the emergence and spread of CQ resistance now restrict its use to only a few areas where malaria is endemic. The aim of the present study was to investigate whether a novel combination of a CQ-like moiety and an imipramine-like pharmacophore can reverse CQ resistance ex vivo. Between March to October 2011 and January to September 2013, two “reversed chloroquine” (RCQ) compounds (PL69 and PL106) were tested against multidrug-resistant field isolates of Plasmodium falciparum (n = 41) and Plasmodium vivax (n = 45) in Papua, Indonesia, using a modified ex vivo schizont maturation assay. The RCQ compounds showed high efficacy against both CQ-resistant P. falciparum and P. vivax field isolates. For P. falciparum, the median 50% inhibitory concentrations (IC₅₀s) were 23.2 nM for PL69 and 26.6 nM for PL106, compared to 79.4 nM for unmodified CQ (P < 0.001 and P = 0.036, respectively). The corresponding values for P. vivax were 19.0, 60.0, and 60.9 nM (P < 0.001 and P = 0.018, respectively). There was a significant correlation between IC₅₀s of CQ and PL69 (Spearman's rank correlation coefficient [r_s] = 0.727, P < 0.001) and PL106 (r_s = 0.830, P < 0.001) in P. vivax but not in P. falciparum. Both RCQs were equally active against the ring and trophozoite stages of P. falciparum, but in P. vivax, PL69 and PL106 showed less potent activity against trophozoite stages (median IC₅₀s, 130.2 and 172.5 nM) compared to ring stages (median IC₅₀s, 17.6 and 91.3 nM). RCQ compounds have enhanced ex vivo activity against CQ-resistant clinical isolates of P. falciparum and P. vivax, suggesting the potential use of reversal agents in antimalarial drug development. Interspecies differences in RCQ compound activity may indicate differences in CQ pharmacokinetics between the two Plasmodium species.

INTRODUCTION

Malaria remains one of the most important infectious diseases in the world (1). A key component of malaria control programs is the ability to provide early diagnosis and treatment with effective antimalarial drugs. Chloroquine (CQ) has been the most commonly used drug for the treatment and prevention of malaria since the 1940s; it is cheap, well tolerated, and widely available. However, within a decade of its deployment, the first reports of CQ resistance were noted in Southeast Asia and South America. By the 1980s, CQ was no longer effective in much of these regions, and reports began to emerge of declining efficacy in Africa (2). It is estimated that more than 80% of the world's malaria is now resistant to CQ (3), leading the World Health Organization (WHO) to recommend that the use of CQ be restricted to non-falciparum malarias and the treatment of falciparum malaria in a few specific areas only (4).

CQ resistance in P. falciparum is a multigenic process (5), the main molecular determinant being mutations in the gene encoding PfCRT (Plasmodium falciparum chloroquine resistance transporter), a putative channel involved in enhanced CQ efflux from the parasite's digestive vacuole (DV) (6).

Several resistance reversal agents, also known as chemosensitizers, have been identified and shown to inhibit CQ efflux in CQ-resistant parasites (7). Many of these agents are licensed for alternative medical conditions, such as antidepressants (desipramine), antihistamines (chlorpheniramine), and calcium channel blockers (verapamil). However, the doses required to achieve adequate “reversal activity” in vivo can lead to poor tolerability and adverse side effects for the patient (8).

One of the strategies to improve both pharmacokinetic and pharmacodynamic properties of CQ resistance reversal agents is to combine the reversal agent with CQ (8, 9). In 2006, Burgess et al. reported innovative hybrid compounds, “reversed CQ” (RCQ) molecules, combining a CQ-like moiety with a CQ resistance reversal agent closely related to imipramine (10). In vivo drug susceptibility testing showed that these molecules are potent in vitro growth inhibitors of CQ-susceptible and CQ-resistant P. falciparum laboratory strains, with 50% inhibitory concentrations (IC₅₀s) in the low-nanomolar range (10). Subsequent studies in animals have confirmed antimalarial efficacy with no overt signs of toxicity (11). These compounds result in enhanced uptake and accumulation in the parasite's digestive vacuole of CQ-resistant parasites and inhibition of hemozoin production in a manner analogous to, but more potent than, that of CQ (11). However, the efficacy of RCQs against P. falciparum field isolates has not been reported, and there are no data on the efficacy of RCQ compounds against P. vivax.

The aims of this study were to (i) investigate the ex vivo susceptibility profiles of two RCQ compounds, PL69 and PL106 (Fig. 1A), against drug-resistant clinical isolates of P. falciparum and P. vivax in Papua, Indonesia, (ii) to examine the stage-specific action of these compounds, and (iii) to investigate cross-susceptibility patterns with conventional antimalarials.

FIG 1.

(A) Chemical structure of “reversed chloroquine” (RCQ) compounds PL69 (compound 22 in reference 11) and PL106. (B) Ex vivo drug susceptibility (median IC₅₀s) to standard antimalarials and the RCQ compounds PL69 and PL106 in P. falciparum (closed symbols) and P. vivax (open symbols) clinical isolates. Significance of P values was determined by Wilcoxon rank sum test. (C and D) Ex vivo drug susceptibilities (median IC₅₀s) for PL69 (C) and PL106 (D) according to species tested and for paired ring-stage (closed symbols) versus trophozoite-stage (open symbols) parasites. Significance of P values was determined by Wilcoxon rank sum test.

MATERIALS AND METHODS

Study area and field sample collection.

The studies were conducted in Timika, Papua Province, Indonesia, between March and October in 2011 and January and September in 2013. Patients presenting to the Rumah Sakit Mitra Masyarakat (RSMM) Hospital were enrolled into the study if diagnosed with a microscopically confirmed P. falciparum or Plasmodium vivax monoinfection and a parasitemia of between 2,000 and 80,000 parasites/μl. Isolates of Plasmodium spp. were processed only if at least 70% of asexual forms were at the ring stage (12). Patients treated with CQ in the last 30 days or any other antimalarial treatments in the last 2 weeks or hemoglobin levels below 5 g/dl were excluded. With written informed consent, 5 ml of venous blood was collected by venipuncture and processed immediately. Host white blood cells were removed using cellulose columns as previously described (13), and packed infected red blood cells (iRBC) were used for the ex vivo drug susceptibility assay.

Drug compounds and drug plate preparation.

Each drug plate (96 wells) contained 11 serial concentrations (2-fold dilutions) in duplicate of four antimalarials. The following antimalarial drugs were assayed: chloroquine (CQ), amodiaquine (AQ), piperaquine (PIP), mefloquine (MFQ), and artesunate (AS) (all provided by the WWARN QA/QC Reference Material Programme [14]), as well as the “reversed CQ” (RCQ) compounds PL69 and PL106 (DesignMedix, Inc., Portland, OR) (Fig. 1A). Predosed drug plates were made by adding 25 μl of drug dilutions to each well, followed by lyophilization and storage at 4°C. The following maximum concentrations were used: 2,993 nM for CQ, 158 nM for AQ, 1,029 for PIP, 338 nM for MFQ, 49 nM for AS, and 295 nM and 349 nM for the RCQ compounds PL69 and PL106, respectively.

Ex vivo drug susceptibility assay.

Drug susceptibilities of Plasmodium isolates were measured using a modified schizont maturation assay as described previously (15, 16). Drug susceptibility profiles were presented as inhibition of parasite growth from the ring to schizont stage. Two hundred microliters of a 2% hematocrit blood medium mixture (BMM), consisting of RPMI 1640 medium supplemented with 10% AB⁺ human serum (for P. falciparum) or McCoy's 5A medium supplemented with 20% AB⁺ human serum (for P. vivax), was added to each well of the predosed 96-well drug plates. The parasites were cultured in a candle jar at 37.0°C for 34 to 50 h. Incubation was stopped when more than 40% of ring-stage parasites had reached the mature schizont stage in the drug-free control well.

The plates were harvested by preparing thick blood films from each well of the plates. Thick blood films were stained with 5% Giemsa solution for 25 min and examined microscopically. Differential counts of 200 asexual parasites in the test slides were classified into ring stages, trophozoites, and mature schizonts. To reduce parasite classification error and ease parasite identification, only schizonts with at least 5 or more well-defined chromatin dots were classified as schizonts at harvest. Free merozoites and gametocytes were not included in the count. To determine the effect of each antimalarial drug, the number of schizonts was determined for each drug concentration and then normalized to that of the corresponding drug-free control well.

To investigate the stage-specific drug susceptibility, Plasmodium isolates with greater than 90% rings were set up (i) in culture in the presence of drug directly for 24 h and after washing out of the drug for another 20 to 24 h and (ii) after culture in the absence of drug for 20 to 24 h to achieve 90% trophozoites that were then drug exposed for 20 to 24 h (17).

Data analysis.

The raw dose-response data were analyzed using nonlinear regression analysis (WinNonLn 4.1; Pharsight Corporation), and the IC₅₀s were derived using an inhibitory sigmoid E_max model. Ex vivo IC₅₀ data were used only from predicted curves where the E_max (maximum effect) and E₀ (minimum effect) were within 15% of 100 or 0, respectively. Data analysis was performed using the STATA (version 13.1; Stata Corp., College Station, TX) and GraphPad Prism (version 6) software packages. The Mann-Whitney U test, Wilcoxon signed-rank test, and Spearman's rank correlation were used for nonparametric comparisons and correlations.

Ethical approval.

Ethical approval for this project was obtained from the Human Research Ethics Committee of the NT Department of Health & Families and Menzies School of Health Research, Darwin, Australia (HREC 2010-1396) and the Eijkman Institute Research Ethics Commission, Jakarta, Indonesia (EIREC-47) in September 2010.

RESULTS

Antimalarial susceptibility.

Ex vivo drug susceptibility profiles were attempted in 86 clinical isolates; baseline characteristics of these isolates are presented in Table 1. Adequate growth for harvest was achieved in 81% (33/41) of P. falciparum isolates and 76% (34/45) of P. vivax isolates. The median IC₅₀s for all clinical isolates successfully cultured are shown in in Fig. 1B and Table 2. PL69 IC₅₀ estimates could not be derived for four P. falciparum isolates (MIC, >295 nM) and one P. vivax isolate (MIC, >349 nM).

TABLE 1.

Baseline characteristics of isolates for which ex vivo assays were accomplished

Baseline characteristic	P. falciparum (n = 41)	P. vivax (n = 45)
Isolates reaching harvest, no. (%)	33 (80.5)	34 (75.6)
Median delay from venipuncture to start of culture, min (range)	153 (85–280)	180 (100–330)
Median duration of assay, h (range)	46 (30–49)	48 (43–50)
Geometric mean parasitemia, no. of asexual parasites/μl (95% CI)a	29,342 (19,286–44,639)	20,936 (14,293–30,666)
Median initial amt of parasites at ring stage, % (range)	100b	95 (71–99)
Mean schizont count at harvest, % (95% CI)	46 (16–76)	43 (25–70)

^aCI, confidence interval.

^bNo range given. (All values were 100%.)

TABLE 2.

Overall ex vivo susceptibility for each drug according to the species tested

Drug	P. falciparum				P. vivax clinical field isolates		Pc
	Laboratory line (IC50, nM)a		Clinical field isolates		No. of assaysb	Median IC50, nM (range)
	FC27 (CQs)	K1 (CQr)	No. of assaysb	Median IC50, nM (range)
Chloroquine	24.09	140.48	33	79.36 (10.92–158.66)	34	60.86 (8.96–334.55)	0.990
Amodiaquine	20.27	26.89	33	23.76 (4.56–71.38)	33	22.63 (8.80–49.16)	0.211
Piperaquine	31.72	47.21	33	27.30 (10.62–121.16)	34	21.69 (3.50–51.86)	0.013
Mefloquine	53.77	13.74	33	16.16 (3.23–53.29)	34	16.82 (4.83–61.48)	0.082
Artesunate	10.12	7.70	33	4.58 (0.83–25.90)	34	2.12 (0.88–13.54)	0.001
PL69	28.49	41.72	29	23.24 (3.18–85.54)	33	18.95 (1.70–114.87)	0.250
PL106	52.56	47.87	27	26.55 (2.18–228.26)	29	59.89 (10.45–248.41)	0.021

^aMean IC₅₀s (derived from 3 independent experiments) were assessed by in vitro schizont maturation quantified by microscopy. CQ^s, chloroquine-sensitive laboratory strain; CQ^r, chloroquine-resistant laboratory strain.

^bTotal number of assays with acceptable drug assay results.

^cSignificant difference in median drug IC₅₀ between species.

Drug susceptibility differed significantly between species, with IC₅₀s being higher in P. falciparum than in P. vivax for PIP (median IC₅₀ = 27.3 nM versus 21.7 nM; P = 0.012) and artesunate (median IC₅₀ = 4.6 nM versus 2.1 nM; P = 0.001) but lower for PL106 (median = 26.6 versus 59.9 nM; P = 0.021). In P. falciparum isolates, IC₅₀s were significantly lower for PL69 and PL106 than for CQ. In P. vivax, IC₅₀s for PL69 (median = 19.0 nM; P < 0.001) but not PL106 (median = 59.9 nM) were significantly lower than those for CQ (median = 60.9 nM).

Stage-specific drug susceptibility.

Stage-specific drug susceptibility was investigated in 5 P. falciparum and 7 P. vivax isolates (Table 3 and Fig. 1C and D). P. falciparum isolates assayed at trophozoite stage had higher IC₅₀s than with drug exposure at the ring stage for PIP (median = 128.2 nM versus 21.9 nM) but lower IC₅₀s for PL69 (median = 50.7 nM versus 70.2 nM), although this was of borderline significance (P = 0.044 and 0.047, respectively). In contrast, P. vivax isolates assayed at trophozoite stage had significantly higher IC_50s for CQ (median = 778.2 nM versus 88.9 nM; P = 0.003), AS (median = 7.2 nM versus 1.9 nM; P = 0.054), PL69 (median = 130.2 nM versus 17.6 nM; P = 0.018), and PL106 (median = 172.5 nM versus 91.3 nM; P = 0.025).

TABLE 3.

Ex vivo susceptibilities for paired isolates tested with 24 h of drug exposure at ring and trophozoite stage

Drug	P. falciparum				P. vivax
	na	Median IC50, nM (range)		Pb	na	Median IC50, nM (range)		Pb
		Rings	Trophozoites			Rings	Trophozoites
Chloroquine	5	107.87 (60.17–163.58)	142.71 (25.50–163.64)	0.855	8	88.87 (41.85–190.17)	778.18 (99.88–4,033.31)	0.003
Amodiaquine	5	16.23 (4.86–201.77)	17.96 (7.76–25.43)	0.570	8	21.61 (5.69–39.47)	53.22 (7.57–89.76)	0.062
Piperaquine	5	21.85 (7.78–33.28)	128.19 (13.80–359.25)	0.044	8	20.88 (9.76–69.60)	30.34 (18.08–205.52)	0.076
Mefloquine	5	15.26 (4.74–46.88)	17.26 (4.60–30.56)	0.570	8	17.87 (8.17–41.00)	37.06 (11.88–85.49)	0.062
Artesunate	5	2.10 (0.74–5.11)	3.87 (1.28–8.13)	0.584	8	1.91 (1.21–4.57)	7.20 (0.83–19.32)	0.054
PL69	5	90.39 (6.61–142.06)	23.55 (7.25–92.82)	0.465	7	17.59 (5.42–49.13)	130.22 (10.26–268.15)	0.018
PL106	5	70.17 (62.24–110.05)	50.67 (27.06–71.36)	0.047	6	91.25 (54.35–126.84)	172.48 (86.76–314.27)	0.025

^an, number of paired isolates.

^bComparison of drugs tested at ring and trophozoite stage, respectively (Wilcoxon rank sum test).

Cross-susceptibility patterns.

Spearman's rank correlation coefficients (r_s) are presented in Table 4. PL69 showed significantly positive correlations in both P. falciparum and P. vivax for AQ (r_s = 0.512 and 0.457, respectively), PIP (r_s = 0.631 and 0.566), and MFQ (r_s = 0.480 and 0.556). However, the correlations of PL69 were only significant with AS in P. falciparum (r_s = 0.446, P = 0.015) and CQ in P. vivax (r_s = 0.727, P < 0.001). Contrasting correlation patterns between species were observed for PL106 with significant positive correlations in P. falciparum only apparent with PIP (r_s = 0.456, P = 0.019). In P. vivax, PL106 was positively correlated with CQ (r_s = 0.830, P < 0.001), AQ (r_s = 0.505, P = 0.005), and MFQ (r_s = 0.401, P = 0.031) and negatively correlated with AS (r_s = −0.470, P = 0.010).

TABLE 4.

Correlation of ex vivo antimalarial susceptibilities in P. falciparum and P. vivax clinical field isolates

Antimalarial combination	P. falciparum			P. vivax
	rsa	P	dfb	rsa	P	dfb
PL69-CQ	0.097	0.618	29	0.727	<0.001	33
PL69-AQ	0.512	0.005	29	0.457	0.008	33
PL69-PIP	0.631	<0.001	29	0.566	0.001	33
PL69-MFQ	0.480	0.008	29	0.556	0.001	33
PL69-AS	0.446	0.015	29	−0.124	0.492	33
PL106-CQ	0.113	0.582	26	0.830	<0.001	29
PL106-AQ	−0.036	0.862	26	0.505	0.005	29
PL106-PIP	0.456	0.019	26	0.332	0.079	29
PL106-MFQ	0.220	0.281	26	0.401	0.031	29
PL106-AS	−0.108	0.601	26	−0.470	0.010	29
PL106-PL69	0.472	0.023	23	0.694	<0.001	29

^ar_s, Spearman rank correlation coefficient.

^bdf, degrees of freedom.

DISCUSSION

This is the first report of the ex vivo drug susceptibilities of two novel RCQ compounds, PL69 and PL106, against multidrug-resistant clinical field isolates of P. falciparum and P. vivax. The IC₅₀s of PL69 were 3-fold lower than those of CQ in both P. falciparum and P. vivax and similar to those of AQ and PIP in both species. While PL106 was also 3-fold more potent than CQ against P. falciparum, it was only moderately more potent than CQ in P. vivax. However, 5 isolates (four of P. falciparum and one of P. vivax) had elevated PL69 MICs, with the range of drug concentrations assessed precluding derivation of reliable IC₅₀s. The significance of these findings warrants further molecular analysis.

Using our schizont maturation assay, the RCQ IC₅₀s in the laboratory strains FC27 and K1 ranged from 28.5 to 41.7 nM for PL69 and 47.9 to 52.6 nM for PL106. These values are significantly higher than the unpublished data derived using a SYBR green I assay (range, 0.9 to 1.8 nM and 1.7 to 2.7 nM, respectively) (11). This may reflect differences in the assay methodology. Previous data for these RCQ compounds were derived by proliferation-based growth inhibition assays of culture-adapted laboratory strains in which in vitro growth was assessed by quantification of DNA using a SYBR green I fluorescence-based assay. This methodology generally produces similar results to radiotracer methods such as [³H]hypoxanthine incorporation (18) but is known to produce significantly lower IC₅₀s than the field-based schizont maturation assay used in the present study (19, 20).

Although the IC₅₀s of the standard antimalarials were similar between species, the IC₅₀s of PL106 in P. vivax were almost 2-fold higher than those observed in P. falciparum. However, since both compounds (albeit PL106 to a lesser extent) have higher activity than CQ in isolates of both species, this suggests an overall lower CQ resistance reversal activity of PL106 than PL69 in P. vivax.

The two RCQ compounds showed notable differences in the correlation patterns between and within the two Plasmodium species (Table 4). Whereas statistically significant positive correlations were found between PL69 and the quinoline-based antimalarials AQ, PIP, and MFQ in both species, a strongly positive correlation with CQ was observed in P. vivax only (r_s = 0.727, P < 0.001). The same was observed with PL106, which showed a strong and significant positive correlation with CQ in P. vivax (r_s = 0.830, P < 0.001) but not in P. falciparum. This observation most likely reflects the expected positive correlation between the CQ moiety of the RCQ compounds and CQ due to their shared modes of uptake, partition, and action; this action is abolished by the potent reversal activity of these compounds in CQ-resistant P. falciparum isolates. However, the fact that the same phenomenon was not observed in P. vivax isolates indicates again that both RCQ compounds have greater CQ resistance reversal potential in P. falciparum than in P. vivax. This is further corroborated by the strong and significant correlation that was observed between PL69 and PL106 in P. vivax (r_s = 0.694, P < 0.001) but not in P. falciparum.

Our results confirm previous observations demonstrating the marked stage specificity of chloroquine (12, 17) in P. vivax, with trophozoites being almost completely resistant to CQ. IC₅₀s for P. vivax trophozoite stages were also higher for AS, PL69, and PL106 than for the ring stages of the same isolates, but this was not apparent in P. falciparum (Table 3). Although P. vivax trophozoites showed significantly higher median IC₅₀s for PL69 (7-fold) and PL106 (2-fold) than the ring stages, the values of the RCQ compounds against P. vivax trophozoites were almost a log fold lower than those of CQ, suggesting that some resistance reversal activity was exerted against both stages.

Our results demonstrating ex vivo IC₅₀s of the RCQ compounds against CQ-resistant P. falciparum field isolates in the low- and mid-nanomolar ranges are in line with the data obtained with P. falciparum laboratory strains. A previous study of P. vivax isolates showed no significant difference between CQ IC₅₀s of isolates exposed to CQ alone compared to isolates exposed to CQ plus verapamil (21). In the present study, RCQ IC₅₀s in P. vivax were lower than those of CQ, which suggests that the resistance reverser moiety of the RCQ compounds does have an effect on CQ activity in P. vivax isolates. The mechanisms of both CQ resistance and CQ resistance reversal activity of various chemosensitizers have been investigated in P. falciparum and revealed pfcrt to be a key determinant of resistance and a target for chemosensitizers; however, both processes are likely to be mediated by multiple factors (5, 22, 23). The knowledge of these processes in P. vivax is still elusive. By using heterologous expression systems, Sa and colleagues have shown that the pfcrt orthologue pvcrt-o plays a role in CQ accumulation and that increased accumulation could be reverted by verapamil (24). Our data demonstrating the CQ chemosensitizing effect for the imipramine-like reversal agent in P. vivax suggests that CQ resistance in this species may also be mediated by decreased drug accumulation; however, further investigations are needed to better understand CQ pharmacokinetics and its genetic determinants in P. vivax.

The RCQ compounds demonstrate significant enhancement of ex vivo activity over CQ against CQ-sensitive and CQ-resistant clinical isolates of both P. falciparum and P. vivax and highlight the potential approach of reversal agents in antimalarial drug development. Although the action of the RCQ compounds against P. vivax has yet to be elucidated further, the current variation of RCQ compounds between P. falciparum and P. vivax supports a growing body of work suggesting that CQ uptake and metabolism, as well as CQ resistance mechanisms, in these two species may be different (25).