The Redox Cycler Plasmodione Is a Fast-Acting Antimalarial Lead Compound with Pronounced Activity against Sexual and Early Asexual Blood-Stage Parasites

The lead benzylmenadione plasmodione and the derivatives benzylMD 1a and 1g were prepared as previously described (3). The synthesis and chemical analysis of benzylMD analogues 1h, 1i, and 1j will be reported elsewhere. The compounds allopurinol, amodiaquine (dihydrochloride and dihydrate), artemisinin, chloroquine (diphosphate salt), cycloguanil (HCl), 5-fluoroorotic acid hydrate, fosmidomycin (sodium salt hydrate), methylene blue (trihydrate), proguanil (HCl), and pyrimethamine were purchased from Sigma-Aldrich. Quinine (HCl) was purchased from Serva (Heidelberg, Germany). Atovaquone was obtained from GlaxoSmithKline (Evreux, France) (laboratory C) or from Sigma-Aldrich (laboratory A). Dihydroartemisinin was purchased from Euromedex (Souffelweyersheim, France) (laboratory A) or from Sigma-Aldrich (laboratory C). Cytochalasin D and ferroquine were gifts from Freddy Frischknecht (University Hospital Heidelberg, Germany) and Jacques Brocart (Lille University, France), respectively. The cysteine protease inhibitor E64 was obtained from Sigma-Aldrich Chemie S.a.r.l. (Saint-Quentin Fallavier, France), and Plasmion was obtained from Fresenius Kabi (Sèvres, France).

In general, compound stock solutions were prepared in dimethyl sulfoxide (DMSO), with the following exceptions. In laboratories A and B, methylene blue, chloroquine, and fosmidomycin were prepared in pure water, quinine was dissolved in 70% ethanol (vol/vol), and proguanil was dissolved in 50% ethanol. In laboratory C, quinine, dihydroartemisinin, and atovaquone were first dissolved in methanol and then diluted in water. All stock solutions were stored as aliquots at −20°C.

The following strains of P. falciparum (common laboratory strains or culture-adapted strains obtained from patient isolates and grown in culture for an extended period) were used in this study: Dd2 (Indochina), 3D7 (Africa), D6 (Sierra Leone), IMTSSA 8425 (Senegal), IMT Vol (Djibouti), IMTSSA L1 (Niger), PA (Uganda), IMTSSA Bres (Brazil), FCR3 (Gambia), W2 (Indochina), FCM29 (Cameroon), and IMTSSA K2 and IMTSSA K14 (Cambodia). “IMTSSA” strains were isolated at the Institut de Médecine Tropicale du Service de Santé des Armées (now called the Institut de Recherche Biomédicale des Armées). All strains are clonal, and clonality was verified using PCR genotyping of the polymorphic genetic markers msp1 and msp2 and microsatellite loci (23, 24). The chemosusceptibility profiles of these lines are reported in Table S2 in the supplemental material. P. falciparum gametocytogenesis was induced from the luciferase-expressing strain pfs16-GFP-Luc (NF54 background) (17).

Intraerythrocytic stages of P. falciparum were cultured according to standard protocols (25) under individual routine conditions as previously described for each laboratory (see below). Cultures were maintained under the following controlled atmospheric conditions: for laboratory A, an incubator with a fixed atmosphere of 5% O₂, 3% CO₂, 92% N₂, and 95% humidity at 37°C (8); for continuous parasite cultures in laboratory B, a fixed atmosphere of 6% O₂, 3% CO₂, and 91% N₂; for short-term experiments in laboratory B, a candle jar with an atmosphere of approximately 17% O₂, 3% CO₂, and 80% N₂ at 37°C; for laboratory C, an incubator with a fixed atmosphere of 10% O₂, 5% CO₂, and 85% N₂ at 37°C and a humidity of 95%; and for laboratory D, a fixed atmosphere of 5% O₂, 5% CO₂, and 90% N₂. Synchronization of cultures over a large time window (0 to 20 h postinvasion [p.i.]) was established by sorbitol treatment (26). Synchronization over a short time window (0 to 4 h p.i.) was achieved using successive treatments with Plasmion (27, 28) and sorbitol.

Inhibition of intraerythrocytic parasite development by benzylmenadione derivatives and control agents was determined by microtiter tests performed according to standard protocols (see below). In vitro antiplasmodial activity is expressed as the 50% inhibitory concentration (IC₅₀) or 90% inhibitory concentration (IC₉₀) for inhibition of parasite development. Activities of plasmodione and chloroquine against the P. falciparum 3D7 and Dd2 strains (as presented in Table 1 and in Table S1 in the supplemental material) were determined in laboratory A by using the SYBR green I assay as described before (8, 29). Briefly, synchronous ring-stage parasites were incubated for 72 h in the presence of decreasing drug concentrations in microtiter plates (final parasitemia, 0.5%; final hematocrit, 1.5%). Each inhibitor was analyzed in 3-fold serial dilutions in duplicate and in at least three independent experiments. Parasite replication was assessed by fluorescent SYBR green staining of parasitic DNA (30) as previously described (29). IC₅₀ and IC₉₀ values were calculated using Prism (GraphPad) [log(inhibitor) versus normalized response − variable slope].

The antiplasmodial activities of the lead compound plasmodione, five benzylmenadione derivatives (1a, 1g, 1h, 1i, and 1j), chloroquine, quinine, monodesethylamodiaquine, and mefloquine against 12 selected P. falciparum strains (as presented in Fig. 2 and in Tables S3 and S4 in the supplemental material) were determined in laboratory C by using the [³H]hypoxanthine incorporation assay (31). Briefly, synchronous parasites at the ring stage (final parasitemia, 0.5%; final hematocrit, 1.5%) were incubated for 48 h in the presence of different drug concentrations in 96-well plates. All strains were synchronized twice with sorbitol before use. Inhibitors were analyzed in 2-fold serial dilutions in duplicate and in 3 to 5 independent repeats. Parasite growth was assessed by adding [³H]hypoxanthine at time zero (31, 32).

Assessment of cross-resistance of plasmodione with clinically used antimalarials was estimated with the Spearman nonparametric correlation test as previously described (33, 34). Normality was tested with the D'Agostino and Pearson omnibus and Shapiro-Wilk normality tests.

The sensitivity of P. falciparum gametocytes to plasmodione or the control agent methylene blue was determined by measuring the luciferase activity in luciferase-expressing gametocytes after a 72-h exposure to diverse drug concentrations according to the protocol developed by Adjalley and colleagues (17), with minor modifications. The pfs16-GFP-Luc (NF54 background) parasite strain (17) was maintained at 1% to 10% parasitemia in RPMI 1640 culture medium supplemented with A+ erythrocytes (2% hematocrit), 10% decomplemented human serum, 9 mM (0.16%) glucose, 0.2 mM hypoxanthine, 2.1 mM l-glutamine, and 20 μg/ml gentamicin under the routine culture conditions of laboratory D. Asexual blood-stage parasites were first synchronized by 5% (wt/vol) d-sorbitol treatment for 10 min at 37°C for two or more successive growth cycles. Synchronous parasites at 3% hematocrit were cultured in 10-cm petri dishes to 10% parasitemia (ring stage), at which point gametocytogenesis was induced by feeding cultures a mixture of conditioned and fresh media (1:1). The next day, trophozoite cultures were diluted 4-fold. N-Acetylglucosamine (NAG; Sigma-Aldrich) was added to a final concentration of 50 mM after all schizonts had ruptured. NAG treatment was maintained for the next two or three cycles of reinvasion to eliminate all asexual forms from the culture. Immature gametocyte stages (I and II) were purified on a Percoll gradient (35) and incubated directly in 96-well plates at 4.5 to 6% gametocytemia (150 μl/well). Gametocyte cultures were exposed for 72 h to dilutions of plasmodione (ranging from 0.125 nM to 600 μM), methylene blue (Sigma) (ranging from 0.021 nM to 100 μM), DMSO (0.17%), or medium starting at day 3, 8, or 11 after gametocytogenesis induction. For this purpose, plasmodione (6 mM) and methylene blue (10 mM) were dissolved in 100% DMSO and H₂O, respectively, diluted 1:10 in RPMI medium, and further diluted in RPMI medium to the appropriate concentrations. One hundred fifty microliters of each solution was added to the gametocyte culture in triplicate. Following drug exposure, parasites were cultivated for two additional days in normal medium before centrifugation and freezing at −80°C. The cells were lysed at room temperature in 1× luciferin lysis buffer (5× luciferase cell culture lysis reagent; Promega), and luciferase activity (luciferase assay system; Promega) was measured on a luminescence plate reader (Mithras LB940 luminometer; Berthold Technologies). The mean of the luciferase activity was calculated for each technical triplicate and normalized as follows: (Luc_x − Luc_RBC)/(Luc_solvent − Luc_RBC), where Luc_x and Luc_solvent are the mean luciferase activities for gametocytes incubated with compound x and with solvent (DMSO for plasmodione and phosphate-buffered saline [PBS] for methylene blue), respectively, and Luc_RBC is the mean luciferase activity of noninfected erythrocytes (background). IC₅₀ and IC₉₀ values were calculated using Prism (GraphPad) [log(inhibitor) versus normalized response − variable slope].

Cytotoxicity against hepatic HepG2 cells was determined using the trypan blue exclusion method and flow cytometry according to standard protocols (36, 37). Detailed procedures are presented in the supplemental material. Genotoxicity was assessed by CEREP (Redmond, WA; now Eurofins Panlabs Inc.), using the Ames fluctuation assay on three Salmonella enterica serovar Typhimurium strains and an in vitro micronucleus assay on CHO-K1 cells according to previously described procedures (38, 39). The data are presented in Tables S6 to S9 in the supplemental material. The cardiotoxicity assay was performed by Eurofins Panlabs Inc., using a patch-clamp assay on human ether-a-go-go-related gene (hERG) potassium channels as further described in the supplemental material.

Hemolysis rate and glutathione (GSH) depletion experiments were performed with G6PD-deficient erythrocytes from at least 3 male individuals hemizygous for G6PD-Med (Mediterranean variant; 1 to 3% residual G6PD activity) (40 – 43) according to standard protocols (44, 45). Detailed procedures are presented in the supplemental material.

In the following experiments, drug concentrations were chosen based on multiples of the IC₅₀ of a given drug as determined under the conditions of each laboratory, as described above.

The stage specificity of drug action was assessed by exposing highly synchronous parasites to successive short drug incubation times along the parasite cycle and subsequently analyzing the completion of the parasite's life cycle (i.e., reinvasion). Under the culture conditions of laboratory B, where the assay was performed, the intraerythrocytic development cycle of strain 3D7 was found to last 48 h. Accordingly, the cycle was divided into six time frames of 8 h each. To obtain a highly synchronous culture, blood-stage parasites were synchronized over a 4-h time window by successive Plasmion and sorbitol treatments. Parasites were split into four culture flasks, each containing erythrocytes and plasma from different donors, and grown for 48 h under normal culture conditions. Another round of Plasmion and sorbitol treatments was then performed in order to fine-tune the 0- to 4-h synchronization of ring-stage parasites. After sorbitol treatment, the cultures were immediately adjusted to 0.5% parasitemia and 2% hematocrit and then used for the experiment (i.e., time zero of the experiment). At 0 h (parasites aged 0 to 4 h), 4 h (parasites aged 4 to 8 h), 12 h (parasites aged 12 to 16 h), 20 h (parasites aged 20 to 24 h), 28 h (parasites aged 28 to 32 h), and 36 h (parasites aged 36 to 40 h), aliquots from the initial culture were distributed into a 96-well microtiter plate and incubated in normal medium (control) or in medium with plasmodione or chloroquine at 1× IC₅₀, 5× IC₅₀, or 10× IC₅₀ (final concentration; final hematocrit, 1.5%). The microplate was then put in a candle jar at 37°C for 4 h (youngest rings [0 to 4 h old]) or 8 h (older parasites). Upon incubation, cells were washed twice with 2 volumes of complete medium, resuspended in fresh medium, and returned to the candle jar. Controls for plasmodione and chloroquine effects on the whole parasite cycle were prepared by 48 h of incubation of the parasites with the drugs. After washing, parasites were allowed to complete their life cycle. Smears were made from all samples at 62 h postinvasion (p.i.) to ensure that reinvasion was fully complete. Smears were stained with Diff-Quik, and parasitemia was determined by optical counting of 2,000 erythrocytes per smear. The reinvasion rate (percentage) was established for each culture and incubation period, with the parasitemia of the corresponding untreated control culture considered to be 100%. Parasite development in the initial culture flasks was followed over time by optical examination of Diff-Quik-stained smears.

The effects of prolonged drug application on early (ring) and late (trophozoites and schizonts) stages were assessed similarly. Rings (aged 0 to 4 h) or trophozoites (aged 24 to 28 h) were incubated for 24 h in normal medium (control) or in medium with plasmodione or chloroquine at 1× IC₅₀, 5× IC₅₀, 10× IC₅₀, or 50× IC₅₀ (final concentration) and then washed and returned to normal culture conditions until reinvasion was complete. Parasitemia upon reinvasion was determined from Diff-Quik-stained smears.

Drug effects on parasite egress or invasion were also assessed by exposing highly synchronous mature schizonts (segmenters) to plasmodione for 8 h in order to allow the parasite's egress and invasion processes to proceed under drug treatment (detailed methods are presented in the supplemental material).

The speeds of action of plasmodione, artemisinin, and atovaquone were assessed on the P. falciparum 3D7 strain according to the in vitro protocol developed recently by Sanz and colleagues (18). The principle of the assay is based on determination of parasite viability in response to drug incubation for increasing periods. Briefly, cultures of asynchronous parasites (2% hematocrit, 0.5% parasitemia) were exposed to a fixed concentration of 5× IC₅₀ of the drug in 3D7, which was 250 nM for plasmodione, 5 nM for atovaquone, and 50 nM for artemisinin. Medium containing fresh drug was renewed daily. Samples of untreated (time zero) and treated (every 24 h for up to 144 h p.i.) parasites were withdrawn from the treated culture and washed with fresh medium. Cells were diluted in 3-fold serial dilutions (12 points) in 96-well plates to achieve maximal dilution as described in the original protocol. Parasites were cultured for a total of 28 days under the routine conditions of laboratory A, medium was renewed twice a week, and fresh erythrocytes were added once a week. On days 21 and 28, parts of the cultures were transferred to black 96-well plates, and parasite growth was assessed using the SYBR green protocol as described above. The number of viable parasites, the resulting parasite reduction ratio (PRR), the lag phase, and the 99.9% parasite clearance time (PCT_99.9%) were calculated as described in the original protocol (18).

The potential of plasmodione to induce resistance in cultured blood-stage parasites (P. falciparum strain Dd2) was assessed according to the protocol proposed by Rathod and colleagues (46), with minor modifications. Two different lethal concentrations of plasmodione were chosen, i.e., 5× IC₅₀ (≈2× IC₉₀) and 10× IC₅₀ (≈4× IC₉₀). Control cultures were performed with 100 nM 5-fluoroorotate (corresponding to 20× IC₅₀) according to the original protocol. Briefly, a series of 10-ml cultures at 2% hematocrit were inoculated with synchronized ring stages at various inoculum concentrations (10⁵, 10⁶, 10⁷, or 10⁸ infected erythrocytes). A control culture with an inoculum of 10 infected erythrocytes was maintained without drug. For each condition, two independent cultures were established from individual stock cultures and maintained for 12 weeks under the standard conditions used in laboratory A (as described above). Medium containing fresh drug was changed three times a week, and fresh erythrocytes were added regularly. A blood smear was prepared with every medium change to check for the emergence of parasites.

In vitro interactions between plasmodione and clinical antimalarials were investigated using the fixed-ratio isobologram method established by Fivelman et al. (47). These assays were performed in laboratory A (allopurinol, fosmidomycin, cycloguanil, and proguanil) or laboratory B (amodiaquine, chloroquine, quinine, ferroquine, artemisinin, dihydroartemisinin, atovaquone, methylene blue, and pyrimethamine). Briefly, asynchronous cultures of P. falciparum strain 3D7 or K14 were incubated for 48 h in the presence of decreasing drug concentrations in a microtiter plate (1% parasitemia, 1.5 to 2% hematocrit). Drugs were applied alone or in combination at fixed concentration ratios (A:B ratios [vol/vol] of 1:4, 2:3, 2.5:2.5, 3:2, and 4:1; the ratio of 2.5:2.5 was used in laboratory A only) and analyzed in 3-fold serial dilutions. IC₅₀s of drugs A and B alone were determined on the same plate to ensure accurate calculation of drug interactions. The inherent variability of the experimental setup was estimated by combining two individual working solutions of the plasmodione lead. Parasite growth was assessed in laboratory A by using the SYBR green protocol as described above and in laboratory B by using the [³H]hypoxanthine incorporation assay (32), adding [³H]hypoxanthine after 24 h of incubation, and the remainder of the experiment was conducted as previously described (48).

In vitro drug interactions were interpreted based on arithmetical and graphical analyses of the results. For the arithmetical analysis, the fractional inhibitory concentration (FIC) of each drug in a given combination was calculated according to the following definitions: FIC A = IC₅₀ of A in combination/IC₅₀ of A alone, and FIC B = IC₅₀ of B in combination/IC₅₀ of B alone. The sum of FIC A and FIC B (∑FIC) was established for each combination, and then the mean value for the four or five ∑FICs was calculated. A mean ∑FIC of ≤0.7 was assumed to reflect a synergistic interaction between the tested drugs and a mean ∑FIC of ≥1.3 an antagonistic interaction. Interactions with a mean ∑FIC value of 0.7 to 1.3 were considered indifferent (in other words, there was no interaction between the drugs, leading to what was formerly described as an additive effect). The graphical presentation (isobolograms) was established by plotting pairs of FIC values for drug A and drug B for each combination. A straight line indicates an indifferent interaction, a curve toward the origin of the axes represents synergy, and a curve in the opposite direction indicates antagonism.

Pictures of Giemsa-stained blood smears of untreated parasite cultures were taken at different time points by use of an Olympus DP72 camera coupled to an Olympus BX63 motorized microscope and using cellSens (v 1.9) software (Olympus, France).

The antiplasmodial activity of plasmodione against the asexual intraerythrocytic stages was evaluated against two common P. falciparum laboratory strains: the chloroquine-sensitive strain 3D7 and the chloroquine-resistant strain Dd2. Plasmodione effectively inhibited multiplication of both parasite strains, with half-maximal inhibitory concentrations (IC₅₀s) in the lower nanomolar range (46 nM to 60 nM) (Table 1). Chloroquine resistance of the Dd2 strain did not affect the activity of plasmodione. In parallel, we determined the IC₅₀s of chloroquine and methylene blue. These values were in the ranges previously reported for these two antimalarial drugs (Table 1) (49 – 51). The IC₉₀ values for plasmodione and control agents are given in Table S1 in the supplemental material. We next investigated the activity of plasmodione against gametocytes. Gametocytes are the sexual stages of the parasite that mediate the transmission of the pathogen from the human host to an Anopheles mosquito during blood feeding. To determine whether plasmodione acts on sexual stages and whether it has transmission-blocking potential, we exposed gametocytes of different maturities to increasing concentrations of this compound for 72 h. Methylene blue, an established gametocytocidal agent (17), was used as a positive control and was found to be effective against all sexual stages at nanomolar concentrations, consistent with previous reports. Plasmodione exerted a strong activity against early (II-III)-stage gametocytes, with IC₅₀s in the low nanomolar concentration range (20.8 nM) (Table 1), and a weak activity against mid-late (IV) and mature (V) gametocytes, with IC₅₀s in the micromolar concentration range (2.4 μM and 11 μM, respectively) (Table 1). The IC₉₀ values are given in Table S1.

Plasmodione's in vitro antiplasmodial efficacy was further evaluated against a range of drug-resistant P. falciparum laboratory strains and field isolates of different geographic origins. This included chloroquine-resistant strains from Southeast Asia (W2, K14, K2, and Dd2), Africa (FCM29, FCR3, PA, L1, and Voll), and Latin America (Bres); pyrimethamine-resistant strains (W2, FCM29, K14, K2, and Dd2); strains with reduced quinine responsiveness (W2, Dd2, FCM29, FCR3, PA, Bres, K14, K2, and L1) and/or mefloquine responsiveness (3D7, W2, D6, FCM29, FCR3, PA, Bres, 8425, K14, K2, L1, Voll, and Dd2); and monodesethylamodiaquine-resistant strains (W2, FCM29, PA, Bres, K14, K2, L1, and Voll). The chemosusceptibility profiles of these P. falciparum strains for 11 antimalarial drugs are compiled in Table S2 in the supplemental material.

Plasmodione was active against all drug-resistant strains investigated irrespective of their resistance profile. We did observe strain-dependent variations in the IC₅₀s of plasmodione (approximately 5-fold; ranging from 41 nM [FCM29] to 215 nM [W2]) (Fig. 2; see Table S3 in the supplemental material). Interestingly, plasmodione showed the highest activity against FCM29, the strain with the highest chloroquine IC₅₀ (Fig. 2). A statistical correlation analysis of the drug susceptibility profiles confirmed the absence of in vitro cross-resistance between plasmodione and chloroquine, quinine, monodesethylamodiaquine, or mefloquine (Table 2). In comparison, reduced chloroquine and quinine responsiveness was positively correlated (rs = 0.6503; P < 0.05), as was reduced susceptibility to chloroquine and monodesethylamodiaquine (rs = 0.9161; P < 0.0001) (Table 2) (33, 34). These results are consistent with plasmodione having a mode of action and/or mechanism of resistance distinct from that of other antimalarial drugs. A table listing all IC₅₀ and IC₉₀ values for plasmodione, chloroquine, quinine, monodesethylamodiaquine, and mefloquine is given in Table S3 in the supplemental material.

In parallel to the lead benzylmenadione, plasmodione, we tested two recently described derivatives (benzylMD 1a and 1g) (3) along with three new analogues (1h, 1i, and 1j) for antiplasmodial activities against the 12 drug-resistant P. falciparum strains. All benzylmenadione derivatives presented potent activities, with the majority of IC₅₀s ranging from 100 nM to 200 nM (IC₅₀ and IC₉₀ values are listed in Table S4 in the supplemental material). As in the case of plasmodione, the lowest IC₅₀s of the compounds were measured against the highly chloroquine-resistant FCM29 strain (18 to 50 nM). The plasmodione lead appeared to be the most active compound against 8 of the 12 tested strains. Together, these data supported our selection of plasmodione as a lead agent for further detailed studies.

Depending on their mode of action, most antimalarials express particular stage specificity by acting at different time points during the intraerythrocytic development of the parasite. To reveal the possible target stage(s) of plasmodione, the P. falciparum strain 3D7 was synchronized within a 0- to 4-h time window. Highly synchronized cultures were subsequently incubated with different drug concentrations (i.e., 1× IC₅₀, 5× IC₅₀, and 10× IC₅₀) for 8 h at consecutive 8-h intervals throughout the 48-h intraerythrocytic life cycle (with the exception of the first 4-h period). Parasite development was measured by determining the reinvasion rate and was compared to the effects of chloroquine treatment.

Plasmodione demonstrated the strongest activity against ring stages (0 to 16 h p.i.), with parasite development being reduced by 60% and 80% at drug concentrations corresponding to 5 times and 10 times the IC₅₀, respectively (Fig. 3a, plot II). Remarkably, a short (4 h) treatment of very young ring stages (incubation from 0 to 4 h p.i. to 4 to 8 h p.i.) also strongly inhibited parasite development, by ∼40% and ∼55% at 5× IC₅₀ and 10× IC₅₀, respectively (Fig. 3a, plot I). The potent activity of plasmodione against ring stages is supported by previous morphological studies, which revealed that ring-stage parasites develop a characteristic pyknotic morphology upon drug treatment (3, 8).

Although mature parasite stages did not respond to plasmodione treatment in the assay described above, the in vitro activity of plasmodione is not restricted to ring stages. Full arrest of trophozoite and schizont development was observed after incubating highly synchronized cultures at the ring (0 h to 4 h p.i.) and trophozoite (20 h to 24 h p.i.) stages with plasmodione for an extended period of 24 h (Fig. 3b). These results suggest that plasmodione inhibits parasite development throughout the intraerythrocytic life cycle, but in a dose- and time-dependent manner and depending on the developmental stage of the parasite (Fig. 3b). For instance, 50 nM plasmodione (1× IC₅₀) was sufficient to inhibit ring-stage development by ≈60% (Fig. 3b, plot I), whereas a concentration of 250 nM (5× IC₅₀) was necessary to induce a comparable effect on later stages (Fig. 3b, plot II). Similarly, a 24-h application of 10× IC₅₀ of plasmodione almost fully inhibited trophozoite and schizont development (Fig. 3b, plot II), while the same concentration applied for only 8 h had no effect on these stages (Fig. 3a, plots IV and V). Further, a 24-h application of plasmodione (Fig. 3b, plot I) induced more inhibition of ring development than an 8-h application (Fig. 3a, plots I to III). In comparison to plasmodione, chloroquine did not act against early rings but instead exerted its antiplasmodial activity against trophozoites and schizonts (24 to 40 h p.i.) (Fig. 3a, plots IV and V), consistent with previous reports.

We next examined whether plasmodione interferes with parasite egress from the host cell and/or invasion of an erythrocyte (see Fig. S1 in the supplemental material). To this end, a culture enriched in mature schizonts was incubated with drug concentrations corresponding to 5 times and 10 times the IC₅₀ of plasmodione for 8 h, and parasite reinvasion was assessed by determining the numbers and proportions of remaining schizonts and newly invaded ring stages at the end of the incubation period. The cysteine protease inhibitor E64 and the mycotoxin cytochalasin D, an inhibitor of actin polymerization, were used as positive controls for inhibition of parasite egress and invasion, respectively (20). The numbers and proportions of schizonts and ring stages in plasmodione-treated cultures were comparable to those in the controls, indicating that plasmodione did not affect parasite egress or reinvasion.

The efficiency of antimalarials to decrease blood parasitemia and cure the infection in patients is mediated in part by their speed of action. We explored plasmodione's potency by assessing parasite viability as a function of drug exposure time, based on limiting serial dilutions of treated parasites and regrowth monitoring, according to the protocol described by Sanz et al. (18). From the resulting viability time course profiles, the following killing rate-defining parameters can be obtained: the lag time between drug addition and onset of action; the logarithm of the parasite reduction rate determined for a period of 48 h of drug exposure in the phase of maximal drug action (log PRR); and the 99.9% parasite clearance time (PCT_99.9%), which is the time needed to clear 99.9% of the initial parasite population. For comparative reasons, the fast-acting drug artemisinin and the slow-acting drug atovaquone were analyzed in parallel. Each drug was applied at a fixed concentration of 5 times the IC₅₀ for strain 3D7.

Artemisinin quickly reduced the parasite load, without a lag phase, with a log PRR of 5, and with a PCT_99.9% of 30 h (Fig. 4), consistent with previous reports (18). For comparison, atovaquone was substantially less potent, as indicated by a lag time of 48 h, a log PRR of 1.6, and a PCT_99.9% of 112 h (Fig. 4) (18). Importantly, plasmodione behaved like artemisinin with regard to the speed of action. Plasmodione acted rapidly and efficiently, with an immediate onset of action (no lag phase), a log PRR of 4.8, and a PCT_99.9% of 33 h (Fig. 4). Thus, plasmodione decreased the number of viable parasites by a factor of 10^4.8 within the first 48 h, and it killed 99.9% of the initial parasite load within 33 h.

The rapid evolution of resistance to many antimalarial drugs highlights the importance of assessing the risk of de novo resistance mechanisms as early as possible during a drug development program. The intrinsic potential of plasmodione to induce parasite resistance was assessed by continuously exposing different inocula of infected erythrocytes (10⁸, 10⁷, 10⁶, and 10⁵ erythrocytes infected with the Dd2 strain) to lethal drug concentrations corresponding to 5 and 10 times the IC₅₀. This experimental setup aims to determine the minimum number of parasites necessary for the development of resistant mutants (referred to as the frequency of resistance selection) (46). Parasites were completely cleared within the first 5 days of treatment with both concentrations of plasmodione. Even after 12 weeks of continuous culture, no resistant parasites emerged in any of the cultures. In comparison, parasites resistant to 5-fluoroorotate emerged in all culture flasks after 5 weeks of treatment. The experiment was repeated and confirmed. These results suggest a very low resistance-inducing potential of plasmodione.

Drug combination therapies are currently preferred over monotherapies for the treatment of malaria in order to prevent the emergence and spread of drug resistance. In search of a possible future combination partner for plasmodione, we investigated the in vitro interactions of our lead compound with clinically used antimalarials from different chemical classes and with different modes of action. These included artemisinin, its active metabolite dihydroartemisinin, quinine, chloroquine, amodiaquine, proguanil, and atovaquone. In addition, we investigated combinations of plasmodione with several experimental antimalarial drugs, including ferroquine. Interaction profiles were assessed using the fixed-ratio isobologram method (47) and were evaluated based on the arithmetical and graphical analyses of results. Synergistic interactions were classified as having a mean ∑FIC of ≤0.7, and antagonistic interactions as having a mean ∑FIC of ≥1.3, and any value between was considered to show an indifferent (formerly known as additive) interaction. To validate the robustness of our experimental setup, we initially investigated plasmodione combined with itself by using two independently prepared drug solutions. The expected mean ∑FIC of 1.0 was obtained (Fig. 5). In addition, we validated the assay by using the two partner drugs of the antimalarial coformulation Malarone (atovaquone and proguanil). These two drugs are known to act synergistically, which we confirmed in our assay (mean ∑FIC of 0.2) (Fig. 5).

Plasmodione revealed a synergism in combination with dihydroartemisinin, the active metabolite of artemisinin. This synergistic effect was observed for both the chloroquine-sensitive P. falciparum strain 3D7 and the chloroquine-resistant strain K1 (mean ∑FICs of 0.6 and 0.7, respectively) (Fig. 5). In comparison, the plasmodione-artemisinin combination revealed only a tendency to synergism rather than true synergism (mean ∑FIC of 0.8 for strain 3D7). The plasmodione-quinine combination was slightly synergistic on strain 3D7 (mean ∑FIC of 0.7) but was indifferent on strain K14 (mean ∑FIC of 1.1). The isobologram data for dihydroartemisinin and quinine combinations determined for P. falciparum strain 3D7 show the concave shape characteristic of synergistic interactions (Fig. 5).

The combination of plasmodione with either chloroquine, amodiaquine, methylene blue, or ferroquine was indifferent (Fig. 5; see Fig. S2 in the supplemental material for graphical representations). Antagonistic interactions with plasmodione were found for the antibiotic fosmidomycin, the purine analogue allopurinol, and the antifolate proguanil (see Table S5 and Fig. S2). Other antimetabolite compounds tested (pyrimethamine and cycloguanil) showed a trend toward antagonism, with a weakly convex isobologram (see Table S5 and Fig. S2). Plasmodione and the structurally related antimalarial atovaquone were indifferent in combination (mean ∑FIC of 1.0) (Fig. 5; see Fig. S2), consistent with these drugs having different modes of action (8).

An initial toxicological profiling revealed that plasmodione exhibits low toxicity against human buccal carcinoma cells (KB cells; IC₅₀ of 80.3 μM) and human lung fibroblasts (MRC-5 cells; IC₅₀ > 32 μM) (3). We extended these toxicity studies to include human hepatocytes. Hepatocytes were chosen for further profiling because these cells host the malaria parasite during early infection following transmission of the parasite by the bite of an Anopheles mosquito. No significant cytostatic or cytotoxic effects were observed in the liver carcinoma cell line HepG2 after a 24-h treatment with plasmodione or the reference antimalarial drug chloroquine at concentrations of up to 64 μM (see Fig. S3a and b in the supplemental material). Prolonged treatment of HepG2 cells for 48 h also revealed no indication of cytotoxicity at concentrations of up to 48 μM. At a concentration of 64 μM, only a moderate cytotoxic effect was observed using the trypan blue assay, resulting in slightly increased cell mortality (40.0 × 10³ ± 5.7 × 10³ dead cells versus 15.6 × 10³ ± 5.7 × 10³ dead cells for the untreated control). However, chloroquine was more toxic (increase of dead cells from 22.4 × 10³ ± 5.4 × 10³ to 140.4 × 10³ ± 13.6 × 10³) than plasmodione (see Fig. S3a to c) under the same conditions. With a prolonged incubation time of 48 h, we observed a cytostatic effect of plasmodione on HepG2 cells, at all concentrations examined, which was comparable to that seen for chloroquine.

We next explored the cardiac and mutagenic profiles of plasmodione, which are generally part of advanced standardized ADME-Tox (absorption, distribution, metabolism, and excretion - toxicity in pharmacokinetics) studies. Cardiac toxicity was evaluated using a patch-clamp assay of hERG K⁺ channels. No effects on hERG-mediated potassium conductance were observed at a concentration of 1 μM (see Table S6 in the supplemental material). The concentration of 1 μM is in the range of the maximal concentration (C_max) of plasmodione in mice when the drug was administered at 1 mg/kg of body weight per os (C_max was 0.2 μM after 15 min) or at 30 mg/kg intraperitoneally (C_max was 2.6 μM after 5 min) (E. Davioud-Charvet, unpublished data). This result indicates that no initial liabilities were identified at 1 μM, but more extensive testing would be necessary during further preclinical development of the compound or optimized analogues. Plasmodione (up to 100 μM) scored negative in a bacterial genotoxicity test, an Ames test, and a micronucleus assay, indicating a low mutagenic potential as well as a lack of genetic toxicity and genetic damage (i.e., clastogenic/aneugenic activity) (see Tables S7 to S9 in the supplemental material). In summary, plasmodione presents a suitable safety profile for a novel early antimalarial lead compound.

A number of antimalarial drugs exert their antiparasitic action via production of reactive oxygen species or other oxidant radicals, and some of those drugs (e.g., primaquine and dapsone) have been shown to elicit hemolytic anemia in G6PD-deficient subjects (42, 43, 52, 53). Erythrocyte destruction in vivo is the sum of two processes: (i) intravascular (hemo)lysis (i.e., rupture) and (ii) elimination of oxidatively stressed/modified erythrocytes by phagocytosis (extravascular hemolysis). Extravascular hemolysis is the vastly predominant process occurring in G6PD-deficient patients upon drug treatment with primaquine or dapsone (52 – 54), the consequence of which is anemia. Considering the redox-cycling behavior of benzylmenadiones, a specific aim of the present study was to evaluate plasmodione's capacity to trigger intra- or extravascular hemolysis in G6PD-deficient erythrocytes (Mediterranean variant, from hemizygous males). With respect to the risk of intravascular hemolysis, plasmodione treatment of G6PD-deficient erythrocytes at 4× to 100× IC₅₀ did not induce cell lysis with liberation of intracellular hemoglobin into the incubation fluid. The 4 experiments demonstrated a <2 to 4% difference in hemolysis between treated and untreated erythrocytes during an 8-h incubation period, without any dose relationship (see Fig. S4 in the supplemental material for hemolysis rates in comparison to the reference compounds diamide and phenylhydrazine). Furthermore, plasmodione at 100× IC₅₀ was previously shown to induce no lysis of infected or uninfected G6PD-sufficient erythrocytes (3, 11). With respect to extravascular hemolysis, a typical early effect of oxidant drugs on erythrocytes is a fast and permanent disappearance of glutathione (GSH), the main intracellular antioxidant protector. A peculiarity of G6PD-deficient erythrocytes is their inability to regenerate reduced GSH; thus, permanent GSH oxidation inevitably leads to erythrocyte phagocytosis and anemia in affected patients (40, 43, 55). Treatment with the control agents diamide and phenylhydrazine resulted in a fast and permanent depletion of intracellular GSH. In contrast, intracellular GSH levels of G6PD-deficient erythrocytes after treatment with plasmodione (4× and 10× IC₅₀) for up to 8 h were comparable to those of untreated cells, thereby demonstrating that the lead compound caused no oxidative stress on the particularly vulnerable G6PD-deficient erythrocytes (see Fig. S4). In addition to these results, we previously investigated oxidative damage in plasmodione-treated G6PD-deficient erythrocytes by measuring (i) membrane-bound hemichromes, which are highly sensitive markers of oxidative stress; and (ii) phagocytosis of oxidatively stressed cells by human phagocytes (THP-1 cells). Similar to the results presented here, we observed no hemichrome deposition on nonparasitized G6PD-deficient erythrocytes and no phagocytosis upon plasmodione treatment at high concentrations (100× IC₅₀) (11). In conclusion, plasmodione did not show particular hemolytic potential in G6PD-deficient cells.

In conclusion, the findings presented here highlight plasmodione's very attractive in vitro antimalarial activity profile. Together with its favorable safety profile (cytotoxicity) established using human cell lines, in particular in vulnerable G6PD-deficient erythrocytes, our data strongly support further development of redox-active benzylmenadiones as antimalarial agents. Given plasmodione's activity against parasite blood stages and early sexual stages (gametocytes), it will be of interest to test its activity in blocking transmission to mosquito vectors and to further screen for more potent analogues. Notably, plasmodione's promising antiplasmodial activity, demonstrated here against cultured P. falciparum parasites, contrasts with the moderate in vivo activity observed in P. berghei-infected mice (3). Our preliminary data suggest that this is due to low bioavailability. Before emphasizing detailed in vivo efficacy studies (for instance, using P. falciparum-infected humanized mice) (22), it will thus be important to better understand the pharmacokinetic and pharmacodynamic properties and metabolism of the lead compound to provide key information on the development of better derivatives. Our continuous work on the chemical optimization of benzylmenadione derivatives has already led to the synthesis of plasmodione analogues with potent antiplasmodial activities against P. falciparum blood-stage parasites and with potentially superior in vivo stability properties. Detailed drug profiling and toxicological studies in relevant animal species will be essential to fully evaluate the risk of off-target effects.