INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

With more than 300 million cases and 2 million deaths estimated to occur annually, malaria continues to be a major problem in the world. The emergence and spread of drug-resistant parasites further exacerbates this serious situation. Quinoline derivatives and parasite dihydrofolate reductase (DHFR) inhibitors have been the drugs most commonly used against malaria during the last 50 years, but resistance to these is now widespread (27, 39). Hence, drugs that target different metabolic features of the malaria parasites are clearly needed, and identification of unique metabolic features of the parasites is a priority in efforts to control malaria. Atovaquone {2-[trans-4-(4'-chlorophenyl)cyclohexyl]-3-hydroxy-1,4-naphthoquinone}, a hydroxynaphthoquinone, was developed during the last decade as a potential antimalarial (19-21). Hydroxyquinones were long known to have possible antimitochondrial activity (40) and were explored as antimalarials as early as the 1940s (38). Problems with the toxicity, pharmacokinetics, and metabolic instability of these compounds, however, dampened enthusiasm for them as antimalarial drugs (20). Atovaquone, on the other hand, was found to be well tolerated, metabolically stable, and a very effective agent with a broad-spectrum antiparasite activity (10-21). It is currently used against Pneumocystis carinii pneumonia and toxoplasmosis in patients with AIDS (1, 22, 23). Unfortunately, in clinical trials against Plasmodium falciparum malaria, atovaquone as a single agent met with approximately 30% treatment failure and rapid emergence of resistant parasites (7, 25). A search for drugs with potential synergy with atovaquone identified proguanil [1-(p-chlorophenyl)-5-isopropylbiguanide], an established antimalarial in its own right, as a candidate (4). Clinical trials with an atovaquone-proguanil combination have provided very encouraging results, with cure rates approaching 100% and little evidence of the emergence of resistance (25, 29). This combination (trademark, Malarone) is being registered as an antimalarial, with plans for its controlled distribution in countries where malaria is endemic.

The effectiveness of the atovaquone-proguanil combination, however, is puzzling. Because a major role of the mitochondrial electron transfer chain in malaria parasites is to serve as an electron sink for dihydroorotate dehydrogenase (15, 16), a critical enzyme in the requisite pyrimidine biosynthesis by the parasites, proguanil could act in synergy by inhibiting DHFR, another enzyme important in pyrimidine synthesis. However, proguanil by itself does not have any effect on parasite DHFR; it needs to be metabolized by the host to a cyclic triazine molecule, cycloguanil, the compound that inhibits the parasite DHFR (5, 8). This oxidative conversion is controlled by certain isoforms of cytochrome P450 (2, 17, 41), and about 20% of the Asian and African population are deficient in this metabolic step (18, 37). Furthermore, in Thailand, where the initial atovaquone-proguanil combination clinical trials were conducted, about 90% of the patients with falciparum malaria failed to respond to proguanil alone (25), which was likely due to widespread resistance-imparting point mutations in the parasite DHFR (12, 28, 30). Hence, the synergistic effect of proguanil on atovaquone has been suspected to involve mechanisms other than the inhibition of parasite DHFR.

In their initial studies to investigate the mechanism of atovaquone action, Fry and Pudney (14) used cholate-lysed mitochondria from P. falciparum and Plasmodium yoelii, supplemented with 100 µM heterologous cytochrome c, to assay electron transport through the cytochrome bc₁ complex of the parasite respiratory chain. They found that atovaquone inhibited electron transport in this assay. Because of significant technical difficulties associated with obtaining workable quantities of functional mitochondria from malaria parasites, we previously developed a flow cytometry assay to measure electropotential across the inner mitochondrial membrane (_m) in live intact parasites and showed that atovaquone collapsed _m in malaria parasites within minutes (34). The compound was also shown to inhibit electron transport as measured by the respiration rate of the intact parasites (34). We have now used these assays to analyze the effects of the atovaquone-proguanil combination on _m and electron transport in malaria parasites. Our results suggest that proguanil enhances _m collapse by atovaquone but has no effect on electron transport inhibition by atovaquone. Parasite DHFR inhibitors, on the other hand, have no effect on atovaquone action.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Mice. BALB/cByJ mice, 6 to 8 weeks old, were purchased from Jackson Laboratories (Bar Harbor, Maine) and maintained in our American Association of Laboratory Animal Care-accredited animal facility until they were used for various experiments.

Parasites. P. yoelii 17XL was maintained in vivo in male or female BALB/cByJ mice by repeated passage. After leukocytes and platelets were removed over a microcrystalline cellulose column (10), infected erythrocytes were enriched for schizonts and trophozoites by centrifugation over a Percoll discontinuous density gradient as described earlier (26). The fractions containing schizonts and trophozoites were pooled, washed twice with RPMI 1640 medium containing 1% fetal bovine serum, and evaluated for leukocyte contamination by Giemsa staining of thin blood smears. These purified schizonts and trophozoites were used in all the experiments described here.

Inhibitors. The antimalaria compound atovaquone was a gift from Glaxo Wellcome, Research Triangle Park, N.C. Proguanil and cycloguanil were kindly provided by Wilbur Milhous, Experimental Therapeutics, Walter Reed Army Institute of Medical Research, Washington, D.C. Mitochondrial respiratory chain inhibitors, uncoupler carbonyl cyanide m-chlorophenylhydrazone (CCCP), pyrimethamine, and chloroquine were purchased from Sigma Chemical Co. (St. Louis, Mo.).

Flow cytometric assay for _m. Flow cytometric assays for measuring the effects of various compounds either alone or in different combinations were carried out with a FACScan (Becton Dickinson Cellular Imaging) essentially as previously described in detail (34). Briefly, 5 × 10⁶ parasitized erythrocytes were incubated with 2 nM 3,3' dihexyloxacarbocyanine iodide [DiOC6(3)] (Molecular Probes, Eugene, Oreg.) for 20 min at 37°C. At the end of the incubation period, these cells were aliquoted into different tubes and again incubated for 20 min with the compound to be tested. At the end of the incubation period the parasitized cells were subjected to fluorescence-activated cell sorter (FACS) analysis. In each experiment fluorescence intensity measurements were carried out in the presence and absence of probe as well as in the presence of the protonophore CCCP. The results were expressed as percent inhibition of the fluorescence intensity, using the measurement in the presence of CCCP as a reference for 100% _m collapse.

Rate of respiration by P. yoelii-infected erythrocytes. The rate of O₂ consumption was measured in a closed system with Clark's oxygen electrode and K-lC Oxygraph (Gilson Medical Electronics Inc., Middleton, Wis.), following the method of Chance and Williams (6) exactly as described elsewhere in detail (34). The rate of oxygen consumption was measured in a reaction volume of 1.5 ml containing 10⁸ parasitized erythrocytes/ml at 37°C and was calculated as nanoatoms of oxygen (nAO) consumed per 10⁸ infected erythrocytes per minute. The rate of respiration ranged from 15 to 20 nAO/10⁸ infected erythrocytes/min for these experiments. The difference in the rate of O₂ consumption by the infected erythrocytes in the presence versus in the absence of the compound was calculated as the measure of respiration inhibition. Various concentrations of each compound were tested individually in separate sets of experiments.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Effect of proguanil on _m and respiration. The effect of proguanil alone on _m and respiration by live intact parasitized erythrocytes was examined. As shown in Fig. 1A, proguanil concentrations up to 12 µM had no effect on _m, but a 10-fold-higher concentration (120 µM) reduced DiOC6(3) fluorescence of the parasites. This effect at such high concentration is likely to be irrelevant in pharmacological terms. Hence, we conclude that proguanil by itself has no significant effect on parasite _m. In comparison, Fig. 1A also shows the effect of atovaquone on _m. Atovaquone collapsed ca. 70% of _m at submicromolar concentrations, achieving an EC₅₀ (concentration at which 50% of the maximal effect was observed) of 1.5 nM. The effect of proguanil alone on respiration was also measured (Fig. 1B), and no effect was observed up to a 20 µM concentration of the compound. Atovaquone, as reported earlier (34), inhibited respiration, with an EC₅₀ of 75 nM (Fig. 1B). We conclude that proguanil by itself does not affect the mitochondrial physiology of malaria parasites at pharmacologically relevant concentrations.

View larger version (16K): [in this window] [in a new window]   FIG. 1.   Concentration-dependent effect of atovaquone and proguanil on m and respiration. (A) Concentration-dependent effect of atovaquone and proguanil on m. Fluorescence intensity was quantitated by FACS analysis in the presence of various concentrations of atovaquone () and proguanil (). The results are presented as percent inhibition of fluorescence intensity, using CCCP-mediated inhibition as 100%. (B) Concentration-dependent effect of atovaquone and proguanil on parasite respiration. The rate of O2 consumption by P. yoelii-infected erythrocytes was measured in the presence of various concentrations of proguanil () and atovaquone () independently. The results are plotted as percent inhibition of respiration. The error bars indicate standard deviations.

Effect of atovaquone-proguanil combination on _m. Various concentrations of proguanil, ranging from 1.2 × 10¹⁰ to 1.2 × 10⁵ M, were tested for their effects on atovaquone-mediated collapse of _m in P. yoelii. As shown in Fig. 2, starting with the concentration of 7 × 10⁷ M, proguanil enhanced the ability of atovaquone to collapse _m. The EC₅₀ of atovaquone was reduced approximately sevenfold (from 15 to 2 nM) in the presence of 3.5 × 10⁶ M proguanil. More significantly, perhaps, the magnitude of atovaquone-mediated _m collapse increased from ca. 70 to 85% by inclusion of 3.5 × 10⁶ M proguanil. Pharmacokinetic studies have shown that this level of proguanil concentration in plasma is achieved within 3.5 h following therapeutic administration of the drug (18, 24). Hence, the proguanil enhancement seen here appears to approximate therapeutically relevant conditions.

View larger version (26K): [in this window] [in a new window]   FIG. 2.   Potentiation of atovaquone-mediated collapse of m by proguanil. Fluorescence intensity was quantitated by FACS analysis in the presence of various concentrations of atovaquone either alone () or in combination with proguanil at various concentrations (3.5 × 107 M [], 7.0 × 107 M [], 1.3 × 106 M [], 3.5 × 106 M [], 7.0 × 106 M [], and 1.3 × 105 M []). Proguanil concentrations below 3.5 × 107 M were also tested but had no effect on m collapse by atovaquone. Therefore, the results are not included in the figure for clarity. The error bars indicate standard deviations.

Effect of proguanil on myxothiazole- and antimycin-mediated _m collapse. We have previously shown that the well-known cytochrome bc₁ complex inhibitors myxothiazole and antimycin collapsed _m of malaria parasites in a dose-dependent manner (34). Therefore, to test whether proguanil potentiated the effects of all electron transport inhibitors, we examined the effect of proguanil in combination with myxothiazole or antimycin. As shown in Fig. 3, the profiles of _m collapse by myxothiazole (Fig. 3A) and antimycin (Fig. 3B) were not significantly altered by inclusion of proguanil at 1 µM or higher. A slight increase by inclusion of proguanil in the overall magnitude (70 to 80%) of _m collapse by the highest concentrations of antimycin (Fig. 3B) is well within the range of magnitude seen at such high inhibitor concentrations and is not likely to be relevant or significant. These results suggest that proguanil enhancement of _m collapse is specific for its combination with atovaquone and does not occur with other electron transport inhibitors that also work on the cytochrome bc₁ complex.

View larger version (19K): [in this window] [in a new window]   FIG. 3.   Effect of standard mitochondrial inhibitors in combination with proguanil on m. (A) Fluorescence intensity was quantitated by FACS analysis in the presence of various concentrations of myxothiazole either alone () or in combination with the two highest concentrations of proguanil, i.e., 1.3 × 106 () and 1.3 × 105 M (). (B) The effect of antimycin on m was determined either alone () or in combination with 1.3 × 106 () and 1.3 × 105 () M proguanil. The error bars indicate standard deviations.

Other DHFR inhibitors do not enhance atovaquone-mediated _m collapse. Proguanil is metabolically converted to cycloguanil, which in turn inhibits parasite DHFR, thereby imparting therapeutic value to this antimalarial (5, 8). Although parasites are not believed to possess the cytochrome P450 enzyme needed for cyclization of proguanil, the fact that proguanil was synergistic with atovaquone in P. falciparum cultures (4) raises the possibility that a small amount of cycloguanil could be formed by the parasites themselves, sufficient to inhibit parasite DHFR. Hence, to investigate whether DHFR inhibition was responsible for the potentiation of the atovaquone effect, we used two authentic parasite DHFR inhibitors, cycloguanil and pyrimethamine, to observe their effects on atovaquone-mediated _m collapse. As shown in Fig. 4, neither of these compounds, over a wide range of concentrations, had any significant effect on atovaquone-mediated _m collapse. These experiments rule out the possibility that DHFR inhibition was responsible for the enhancement of atovaquone effect.

View larger version (18K): [in this window] [in a new window]   FIG. 4.   Effect of parasite DHFR inhibitors on atovaquone-mediated collapse of m. (A) Concentration-dependent effect of cycloguanil-atovaquone combination on m. Fluorescence intensities were quantitated by FACS analysis in the presence of atovaquone, either alone () or in combination with a series of cycloguanil concentrations (1.6 × 1010 M [], 1.6 × 109 M [] 1.6 × 108 M [], 1.6 × 107 M [], 1.6 × 106 M [], and 1.6 × 105 M []). (B) Concentration-dependent effect of pyrimethamine-atovaquone combination on m. The effect of atovaquone on m was determined either alone () or in combination with three concentrations of pyrimethamine, i.e., 1.6 × 107 M (), 1.6 × 106 M (), and 1.6 × 105 M (). The error bars indicate standard deviations.

Chloroquine-proguanil combination does not affect _m. To assess the possibility that proguanil augments m collapse in a dying parasite, we tested the chloroquine-proguanil combination for its effect on m. Similarly, the atovaquone-chloroquine combination was also tested for the possible effect of chloroquine on atovaquone-mediated collapse of _m. As shown in Fig. 5, the proguanil-chloroquine combination had no effect on the _m of P. yoelii and chloroquine did not affect the profile of atovaquone-mediated _m collapse.

View larger version (15K): [in this window] [in a new window]   FIG. 5.   Effect of chloroquine-proguanil and chloroquine-atovaquone combinations on m. (A) Fluorescence intensity was quantitated by FACS analysis in the presence of various concentrations of chloroquine either alone () or in combination with the two highest concentrations of proguanil, i.e., 1.3 × 106 M () and 1.3 × 105 M (). (B) Effects of various concentrations of atovaquone on m either alone () or in combination with the highest concentration of chloroquine (1.25 × 105 M []).

Proguanil has no significant effect on respiration inhibition by atovaquone. We tested the effect of the atovaquone-proguanil combination on mitochondrial electron transport as judged by the respiration rate of the intact parasitized erythrocytes. As shown in Fig. 6, inclusion of proguanil did not alter the profile of atovaquone-mediated respiration inhibition. Concentrations of proguanil (>10⁶ M) that significantly enhanced _m collapse by atovaquone (Fig. 2) had no effect on respiration inhibition by atovaquone. This suggests that inhibition of electron transfer by atovaquone is unaffected by the presence of proguanil.

View larger version (24K): [in this window] [in a new window]   FIG. 6.   Concentration-dependent effect of atovaquone-proguanil combination on parasite respiration. The rate of O2 consumption was measured in the presence of various concentrations of atovaquone either alone () or in combination with the two highest concentrations of proguanil (2.3 × 106 M [] and 2.3 × 105 M []). The error bars indicate standard deviations.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

At first glance, the collapse of _m by atovaquone in malaria parasites could be interpreted as a consequence of electron transport inhibition, especially since a lack of ATP synthase (13) may prevent restoration of _m. However, the results of the experiments described in this paper raise the possibility that atovaquone-mediated _m collapse is not entirely a consequence of electron transport inhibition. Proguanil by itself had minimal effects on electron transport and _m of P. yoelii, but it enhanced the ability of atovaquone to collapse membrane potential without affecting electron transport inhibition. This suggests that the respiration-inhibitory property of atovaquone can be dissociated from its uncoupling ability. A large number of metabolic processes relegated to mitochondria (35) are dependent upon metabolic transport across the inner membrane, which in turn is dependent upon maintenance of _m. Hence, collapsed _m may have more far-reaching consequences than electron transport inhibition. In this regard, the site-specific protonophoric activity of the atovaquone-proguanil combination suggested by our results may prove to be a novel mechanism of action.

Our results clearly show that proguanil acts to enhance the atovaquone effect as a biguanide rather than as its metabolite cycloguanil. Biguanides have a long history of use as agents that affect cellular physiology (32). Drugs such as metformin are currently used as hypoglycemic agents for the treatment of insulin-independent diabetes (9). It was initially suggested that these compounds are uncouplers of oxidative phosphorylation (31); however, the uncoupling effect is seen at millimolar concentrations, which are not pharmacologically relevant. The puzzle of atovaquone-proguanil efficacy against malaria in areas where a large number of individuals do not metabolize proguanil, and where proguanil-resistant parasites are widespread, can now possibly be explained by the results described here. The slow uptake of atovaquone and its high lipophilicity (21) may result in a relatively prolonged period of parasite exposure to suboptimal concentrations of the drug when it is used as a single agent. Under these conditions, atovaquone-resistant parasites appear to emerge frequently (7, 25). The inclusion of proguanil with atovaquone will effectively lower the in vivo 50% inhibitory concentration of atovaquone, thereby resulting in parasite demise at atovaquone concentrations which otherwise would have been suboptimal. The net result will be a much lower incidence of treatment failure and resistance emergence, which is what has been observed in clinical trials (25, 29). Pharmacokinetic studies have shown that proguanil concentrations required for the enhancement of atovaquone effect shown here are achieved in an adult within 3.5 h after an oral dose of 200 mg (18, 24).

The molecular basis for proguanil enhancement is unclear. Fidock and Wellems (11) have suggested that proguanil may have intrinsic activity independent of its metabolic activation and DHFR inhibition. Indeed, at concentrations higher than 12 µM, proguanil does seem to have an effect on _m in malaria parasites, although the specificity of this effect is unclear, since a number of compounds can affect mitochondrial physiology at high concentrations. The concentrations of proguanil used by Fidock and Wellems (11) are significantly higher (50% inhibitory concentration, 50-75 µM) than those achieved in vivo (18, 24). Thus, the relevance of the intrinsic in vitro activity of proguanil to the clinical situation remains unclear. The synergistic activity of proguanil in our experiments was observed at pharmacologically achievable drug concentrations.

Atovaquone appears to block electron transfer from ubiquinol to cytochrome c (14). Atovaquone binds the cytochrome bc₁ complex at or around the ubiquinol oxidation (Q_o) site, a site believed also to be occupied by standard Q_o site inhibitors, such as myxothiazole (3). The crystal structure of the bovine cytochrome bc₁ complex revealed that myxothiazole binds in a pocket within the cytochrome b that lies between the low-potential heme and the iron-sulfur cluster of the Rieske protein (42). Indeed, characterization of atovaquone-resistant malaria parasites has identified a discrete cavity within the parasite cytochrome b where atovaquone is likely to bind (33). The specificity of this binding appears to be dictated by the unique structural features of the parasite cytochrome b compared to its mammalian homologue. This binding could explain the electron transport inhibition by atovaquone. The rapidity of the _m collapse, however, suggests an additional direct effect of atovaquone acting as a site-specific protonophore. This suggestion is consistent with the observation that proguanil is able to enhance _m collapse without affecting electron transport inhibition. We would like to speculate that, in addition to electron transport inhibition, atovaquone acts to destabilize the cytochrome bc₁ complex, causing proton leakage to occur through this site, and that proguanil, through as yet unclear means, enhances this destabilization. This would then result in proton leakage at lower concentrations of atovaquone. Examination of other compounds, especially other biguanides, for their ability to act in a similar synergistic manner could provide additional leads in devising drug combinations that target mitochondrial physiology in malaria parasites.