In this study, two assay formats as a means for determining the timing and severity of drug-induced stress in Plasmodium falciparum parasite cultures over short incubation periods were explored. A panel of six anti-malarial drug compounds with different modes and, presumably, rates of action were used to interrogate and characterize the utility of the ATP and luciferase assays described above. The observed responses of trophozoite-stage parasite ATP content to drug exposure over a 10 hour incubation period can be broadly classified into three phenotypes: little or no change relative to untreated controls (chloroquine and DFMO); marked and sustained increase in ATP levels (mefloquine and artemisinin); rapid depletion of ATP content (ritonavir and gramicidin). These classifications may conceivably be interpreted as signifying, respectively: absence of significant drug-induced stress; altered (increased) metabolic activity to counteract drug-induced stress; severely compromised metabolic status due to drug action. This interpretation correlates with the results obtained with subsidiary assays. Morphological evaluation of the drug-treated parasites revealed mild abnormalities, mostly limited to growth retardation, in the ATP non-responders, and (with the exception of chloroquine) competence to recover from a 6h drug exposure. By contrast, increased ATP levels correlated with earlier appearance of growth-inhibited parasites and additional aberrant morphologies and a 44%-54% reduction in recovery following 6h drug exposure, while rapid ATP depletion was accompanied by the early appearance and preponderance of pyknotic (presumably irretrievably compromised) parasite forms and a significantly greater inhibition (79%-89%) of parasite recovery following 6h drug exposure. Interestingly, the timing and magnitude of the ATP changes, whether increased or decreased, also correlated with the timing of appearance of abnormal morphologies and ability to recover from brief drug exposure, although arguably more compounds will need to be assessed to determine the validity of this trend. Based on these results, the rate and severity of drug stress can be tentatively ranked as ritonavir>gramicidin>artemisinin>mefloquine>>chloroquine, DFMO.
The lack of a notable ATP response with DFMO agrees with its accepted mode of action, which is cytostatic rather than cytocidal. It inhibits ornithine decarboxylase, a key enzyme in polyamine biosynthesis, which slows down growth and eventually blocks parasite cell cycle progression in the late trophozoite stage in a manner which is reversible by exogenous addition of polyamines to the medium
. Thus, a 10h incubation with DFMO starting in the early trophozoite stage may generate insufficient stress to result in a notable disruption of ATP homeostasis. The consensus view is that chloroquine becomes ionized and trapped in the low pH environment of the parasite food vacuole, where it disrupts haemozoin formation and causes an accumulation of toxic free haem and chloroquine-haem complexes
. The results of this study suggest that a 10h incubation with chloroquine from the early trophozoite stage does not generate sufficient haem complexes to exert a significant effect on parasite ATP levels and/or haem-induced toxicity is slow-acting. Interestingly, the failure of parasites to recover from the 6h chloroquine incubation in the recovery assay may provide additional evidence for the irreversible entrapment of chloroquine in the food vacuole, where it probably continues to cause haem accumulation and toxicity despite the washing away of exogeneous chloroquine in the medium. Thus, the trapping phenomenon likely contributes greatly to the effectiveness of chloroquine as an anti-malarial drug, despite the apparently slow rate at which it induces metabolic stress compared to the other compounds. An alternative explanation is that haem accumulating in the parasite during the 6h chloroquine treatment remains parasite associated after chloroquine washout and continues to exert toxicity in the following 48h incubation.
Despite decades of clinical use, the mode of action of mefloquine is still uncertain. A general assumption is that it shares chloroquines effect on haemozoin formation and causes a toxic accumulation of haem or haem complexes
. However, cell biological evidence e.g. the differential effects of the two drugs on haemoglobin endocytosis, trafficking and digestion suggests otherwise
[19, 20]. In the present study, the rapid and marked effect of mefloquine on ATP levels and luciferase activity is not mirrored by chloroquine and thus strongly implies a different primary mode of action for mefloquine. As for mefloquine, the mode of action of artemisinin is uncertain and has been suggested to involve inter alia inhibition of the parasite equivalent of the sarcoplasmic reticulum Ca2+-ATPase (PfATP6), depolarization of mitochondrial membranes, haem adduct formation, generation of reactive oxygen species in conjunction with haem and protein alkylation
. Nonetheless, the appeal of artemisinin as an anti-malarial is partly based on its perceived rapid action against parasites
, which is supported by the observations from this study that ATP and luciferase activity levels in treated parasites are markedly affected by the drug at the first 2h time-point of exposure. Uncertainty regarding the modes of action of mefloquine and artemisinin make it difficult to fully explain the cause for increased ATP levels in treated parasites, other than to make a general assumption that it reflects increased metabolic activity by the parasite as part of a cellular stress response to overcome detrimental drug effects. This likely necessitates increased production of ATP to fuel synthesis and activities of enzymes, substrates and co-factors involved in e.g. antioxidant defence and protein chaperone systems.
The very rapid and profound depletion of ATP in ritonavir-treated parasites was supported by the early preponderance of pyknotic parasite morphologies and highly compromised ability to recover from a 6h drug exposure. This was surprising, given that ritonavir is an HIV protease inhibitor and was proposed to act against parasites by inhibiting aspartyl proteases responsible for haemoglobin digestion
. Arguably, inhibition of this process would result in a more protracted growth inhibition of parasites due to amino acid starvation, not the rapid and lethal effect observed here. This argues for a different mode of action of ritonavir, which was also proposed in a study reporting the anti-malarial interactions of HIV protease inhibitors with hemoglobin protease inhibitors, mefloquine and chloroquine
. The rapid depletion of parasite ATP by gramicidin, however, is consistent with its probable mode of action. Gramicidins are lipophilic, linear peptides that form channels in membranes that are permeable to monovalent cations
. The rapid disruption of cellular sodium, potassium and proton gradients through these channels should have immediate pleiotropic consequences for parasite metabolism, which may also be reflected by the extreme potency of gramicidin against parasites (IC50 < 0.1nM – Additional file
The results obtained with the ATP assay suggest that it could represent a sensitive, quantitative means for detecting the earliest time-points of drug-induced stress to inform and complement drug mode of action studies. However, the question remains whether it could also be a useful tool for unambiguously determining the rate and extent to which parasite viability is irrevocably compromised by a particular drug. In principle, a total depletion of ATP could have been regarded as a signpost for irreversible parasite lethality. However, this is not entirely the case, as evidenced by the ability of ritonavir and gramicidin-treated parasites to recover from a 6h treatment, albeit severely limited, despite an apparent complete reduction in ATP in 2–4 hours. Conversely, artemisinin and mefloquine treated parasites actually display increased ATP levels at 6h, despite the fact that their recovery from a 6h treatment is inhibited by approximately 50%. The fact that treatment with the panel of six drugs produces three distinct phenotypes of ATP responses (increased, decreased and unchanged ATP levels) may further complicate a detailed interpretation of ATP responses to experimental drug pressure, necessitating an exploration of ATP responses with a larger drug panel before considering scale-up of the assay. In addition, 5xIC50 drug concentrations were used in this study as a compromise to obtain sufficient drug pressure on the parasite without excessive off-target effects unrelated to the primary drug mode of action which may be prevalent at higher concentrations. However, it should be cautioned that 5xIC50 does not necessarily equate to lethal dose to the same extent for all compounds. Therefore, rate of killing studies may best be performed at lethal dose concentrations. Characterization of the assay, perhaps in conjunction with the recently described parasite recovery rate (PRR) assay
, would be required to more firmly define the correlates of ATP levels and irreversible parasite lethality. Nonetheless, the proposed ranking of the test drugs based on ATP responses, as discussed above, suggests that the assay in its current form may be used to assess the rate of parasite viability inhibition of experimental compounds relative to each other and standard “benchmark” drugs.
Serendipitously, it was found that luciferase activity in transgenic parasites responds rapidly and markedly to drug exposure. A highly attractive advantage of the luciferase assay is that it is entirely multiwell plate-based, requires minimal liquid handling steps and provides an extremely sensitive and robust read-out, thus making it potentially amenable to high-throughput formats. The overall trend in the results was similar to that obtained with the ATP assay. DFMO and chloroquine produced a slow, mild decrease in luciferase activity, while artemisinin, mefloquine and ritonavir profoundly compromised luciferase activity within 2 hours. The rapid decreases in luciferase activity during drug exposure could be construed as a cellular stress response in which proteolysis and amino acid release and/or selective translational inhibition is used to alter the proteome of the parasite
[25, 26]. Luciferase is particularly known to be susceptible to proteolytic degradation
. Nonetheless, the activity decrease was not affected by proteasome inhibitors, even though proteasomes are principally responsible for cytoplasmic protein turnover and homeostasis in mammalian cells
. Interestingly, the proteasome inhibitors on their own also produced marked loss of luciferase activity in 6 hours, suggestive of parasite stress experienced by the inhibition of their protein turnover ability. The rapid loss of luciferase activity is also not shared by all parasite cytoplasmic proteins. In contrast to luciferase, parasite pLDH activity showed only mild changes after 6 hours incubation with all the drugs, which hints that luciferase is particularly sensitive to cellular stress conditions. Caution should be exercised in interpreting the luciferase results, however. The luciferase assay requires the overexpression of a foreign reporter protein in the parasite. Conceivably, this could cause subtle alterations that may obscure drug-specific effects in subsequent drug mode of action studies (e.g. by transcriptomic, proteomic or metabolomic analyses). In addition, transgene overexpression might alter parasite sensitivity and the rate of drug-induced stress by certain compounds. For example, the detailed ranking of compound rate of action in the luciferase assay would differ from that obtained by the ATP assay. Artemisinin produced more profound and rapid changes in luciferase activity than the other drugs, while gramicidin had a milder effect than expected from the earlier assays (ATP, morphology and parasite recovery). While mindful of these caveats, the ease of the assay still suggests that it could be used as a preliminary screen for rapid vs. slow acting compounds over a 4–6 hour incubation period, especially when a large number of compounds need to be assessed.