Atovaquone Tolerance in Plasmodium falciparum Parasites Selected for High-Level Resistance to a Dihydroorotate Dehydrogenase Inhibitor

Abstract

Atovaquone is a component of Malarone, a widely prescribed antimalarial combination, that targets malaria respiration. Here we show that parasites with high-level resistance to an inhibitor of dihydroorotate dehydrogenase demonstrate unexpected atovaquone tolerance. Fortunately, the tolerance is diminished with proguanil, the second partner in Malarone. It is important to understand such “genetic cross talk” between respiration and pyrimidine biosynthesis since many antimalarial drug development programs target these two seemingly independent pathways.

TEXT

Malarone is an antimalarial drug combination, often prescribed for travelers as a prophylactic. but its high cost so far has limited its general use in countries where malaria is endemic. The patent for this drug expired in 2013, and its global usage could soon surge. Atovaquone and proguanil are the synergistic pair of antimalarials that make up this effective, well-tolerated cocktail. While the mechanism of action for proguanil remains uncertain (1, 2), it is well know that atovaquone targets the cytochrome bc₁ complex of the Plasmodium falciparum respiratory chain (3, 4). One function of this mitochondrial pathway is to generate a proton gradient that is required for maintaining the membrane potential for important processes such as protein synthesis, heme biogenesis, and the citric acid cycle (reviewed in reference 2). Importantly, the P. falciparum cytochrome bc₁ complex has been implicated in the direct regeneration of oxidized ubiquinone for dihydroorotate dehydrogenase (DHODH), a key enzyme in the biosynthesis of pyrimidine nucleotides (5–7). Thus, DHODH is physically and functionally tied to the respiratory chain.

Recently, P. falciparum parasites resistant to DSM1, a potent inhibitor of DHODH (8, 9), were generated at two levels (333 nM for round 1 and 3 or 10 μM for round 2) (10). Amplification of the genome segments that encompassed the DHODH gene were responsible for the observed phenotype. Whole-genome sequence methods ruled out resistance-conferring mutations elsewhere, and no alterations in the sensitivities to several unrelated antimalarials were detected. However, here we report the unexpected development of tolerance to atovaquone in parasites resistant to higher levels of the DHODH inhibitor DSM1 (round 2).

Following DSM1 selection and subcloning (10), atovaquone sensitivity was tested. Round 1 parasites (∼3-fold DSM1 resistance) exhibited wild-type sensitivity to atovaquone in a 72-h assay (Fig. 1A and Table 1). Long-term exposure to low levels of atovaquone confirmed the sensitivity of these parasites: 10 nM atovaquone completely prevented the proliferation of parental Dd2, as well as round 1 parasites within 2 days, and continued to do so over an extended period of 8 days (Fig. 1C and D, red lines). We further assessed the survival of Dd2 and round 1 clone C by exposing small clonal populations of parasites to atovaquone before extensive washing and replating in the absence of drug and following the number of viable colonies over 60 days (11). In this clonal viability assay, with the exception of a slightly increased survival of clone C at the lowest exposure level (24 h, 10 nM), we observed the full cidal activity of atovaquone on these parasite clones (average survival of 0% compared with that of untreated controls) (Table 2).

FIG 1.

Atovaquone tolerance in high-level DSM1-resistant parasites. (A and B) SYBR Green-based dose responses of various P. falciparum clones to atovaquone are shown in the absence of DSM1 (combined from 3 independent experiments). (A) DSM1-sensitive clonal Dd2 (black filled circles) compared with partial DSM1-resistant round 1 clones (C, green open squares; and D, green open triangles) (10). (B) DSM1-sensitive clonal Dd2 (black filled circle) compared with high-level DSM1-resistant round 2 clones (C53-1, blue squares; C710-1b, blue circles; C710-2a, blue inverted triangles; and D73-1, blue triangles) (10). The decrease in proliferation in the presence of atovaquone is calculated as a percentage of activity from dimethyl sulfoxide (DMSO) controls. The lines on the plots show the nonlinear curve fits of data points from individual clones. EC₅₀s, where measurable, are listed in Table 1. (C to E) Growth of different P. falciparum clones in the presence of continuous levels of 0 (black), 10 (red), or 100 nM (purple) atovaquone over 8 days (see Table 1 for the respective DSM1 concentrations). (C) Clonal Dd2 (single values). (D) Partially DSM1-resistant round 1 clones (means of clone C and clone D with standard deviations). (E) Highly DSM1-resistant round 2 clones (means of C53-1, C710-1b, C710-2a, and D73-1 clones with standard deviations). Cumulative % parasitemia was calculated by normalizing parasitemia values (measured using SYBR Green-based flow cytometry method) to the starting parasitemia and calculating the percentage of the maximum value achieved over 8 days.

TABLE 1.

EC₅₀s of DSM1-resistant clones for various antimalarials^a

Antimalarial(s)	Method	Dd2	EC50s (± confidence interval) for:
			Round 1		Round 2
			C	D	C53-1	D53-1	D73-1	C710-1a	C710-1b	C710-2a	C710-2b
DSM1 selection (μM)		0	0.3	0.3	3	3	3	10	10	10	10
Atovaquone (nM)	SYBR	1.2 (0.2)	1.0 (0.4)	0.5 (0.4)	Tolb	—c	Tol	—	Tol	Tol	—
Atovaquone (nM) + 1 μM proguanile	SYBR	0.3 (0.1)	—	—	—	—	0.6 (0.5)	—	0.2 (NDd)	0.4 (0)	—
Myxothiazol (μM)f	Hypo	0.3 (0.1)	0.2 (0.1)	0.5 (0.3)	1.5 (1.1)	—	0.4 (0.6)	—	1.1 (0.7)	0.7 (0.7)	—
Antimycin A (μM)f	Hypo	0.4 (0.07)	0.2 (0.1)	0.5 (0.4)	1.1 (1.4)	—	0.3 (0.6)	—	0.9 (0.4)	0.6 (0.7)	—
DSM1 (μM)g	SYBR	0.1 (0.02)	0.4 (0.04)	0.3 (0.04)	—	—	11.1 (6.2)	—	4.7 (2.3)	5.8 (1.6)
Atovaquone (nM)	Hypo	5.9 (0.3)	8.9 (1.1)	4.6 (0.3)	Tol	Tol	—	Tol	—	—	Tol
Proguanil (μM)	Hypo	14.1 (3.6)	14.8 (3.6)	14.8 (3.4)	10.3 (4.4)	7.0 (2.1)	—	10.3 (2.7)	—	—	17.6 (3.9)
Artemisinin (nM)	SYBR	12.5 (ND)	8.3 (0.8)	7.4 (1.0)	12.5 (ND)	—	4.7 (0.5)	—	5.0 (ND)	5.4 (0.4)	—
Artemisinin (nM)	Hypo	5.2 (0.6)	5.3 (1.3)	5.2 (1.3)	7.5 (2.1)	—	6.6 (3.2)	—	6.4 (0.8)	6.2 (6.1)	—

^aThe development of infected erythrocytes exposed to antimalarials was measured in triplicate by flow cytometry using SYBR green fluorescence (SYBR) or by hypoxanthine uptake assays (Hypo). EC₅₀s were calculated as previously described (10). Values listed are from a single representative experiment.

^bTolerance (Tol) is assigned when the EC₅₀ could not be determined due to incomplete inhibition, but parasites grew at attenuated rates (see Fig. 1B and E).

^c—, experiment not performed.

^dND, could not be determined.

^eWe have also performed this experiment with 100 and 500 nM proguanil. Similar to what we observed for 1 μM proguanil (included in the table), the lower nM proguanil concentrations shift the EC₅₀s of atovaquone to Dd2 levels (our unpublished observation).

^fDose responses of various P. falciparum clones to these two compounds are presented in Fig. 2.

^gDSM1 EC₅₀s were also measured by the hypoxanthine uptake assay and published in reference 10.

TABLE 2.

Survival of DSM1-resistant cell lines after exposure to lethal atovaquone concentrations^a

Condition	Average % survivalb (range of days to detectionc) of clonal population:
	Dd2	C12	C53-1	C710-2a
1-day exposure, 10 nM ATOd	0 (NAe)	14 (16–18)	86 (9–13)	88 (11–13)
5-day exposure, 10 nM ATO	0 (NA)	0	29 (21)	63 (16–18)
1-day exposure, 100 nM ATO	0 (NA)	0	57 (25–28)	38 (23–25)
5-day exposure, 100 nM ATO	0 (NA)	0	14 (34)	25 (30–37)

^aThis method is derived from our earlier work (11). Starting from freshly thawed parasites of low parasitemia, serial dilutions were performed to isolate clonal populations. Triplicate T25 flasks were seeded with 10 asynchronous infected erythrocytes and incubated for 1 or 5 days with solvent only (control) or 10 nM or 100 nM atovaquone (tests). Each culture was treated at approximately the same time, and at the end of incubation, the erythrocyte pellets were washed three times with complete medium and transferred to 24 wells of a flat-bottom 96-well plate. The plated cultures were grown in inhibitor-free medium under standard culture conditions with three medium changes per week out to 60 days. Viable populations were identified using flow cytometry-detected SYBR green fluorescence after each medium change.

^bAverage % survival = 100 × (average positive wells [test] ÷ average positive wells [control]).

^cRecovered parasites were observable around the same time as the solvent-only controls (our unpublished data).

^dATO, atovaquone.

^eNA, not applicable.

In contrast, 3 or 10 μM level resistance to DSM1 (round 2, ∼20- to 50-fold resistance) (Table 1) conferred atovaquone “tolerance” (Fig. 1B and Table 1). This phenotype is not like the full resistance seen in traditional studies (12–14) because proliferation stalls at ∼50%, and parasites persist at a low level even at atovaquone concentrations as high as 10 μM in a 72-h study (as measured by SYBR Green DNA labeling or by hypoxanthine uptake assays) (Table 1 and Fig. 1B). During longer exposures, unlike parental Dd2 clones and round 1 low-level DSM1-resistant clones, these round 2 parasites were not inhibited and continued to proliferate in the presence of 10 nM or 100 nM atovaquone. After 5 days, parasite growth slowed, but viable parasites continued to persist in culture even longer (Fig. 1E). When the relative growth is quantified from the later part of this experiment (from 3 to 8 days), round 2 DSM1-resistant clones continued to increase by ∼40% in the presence of 10 nM atovaquone, while parental Dd2 and round 1 clones did not proliferate at all. Further supporting the atovaquone tolerance phenotype in round 2 clones, clonal viability assays on two round 2 clones displayed marked increases in survival under all conditions (average survival of up to ∼90% compared with that of untreated controls) (Table 2).

Atovaquone tolerance in round 2 DSM1-resistant parasites is diminished following the addition of the synergistic partner drug proguanil: the 50% effective concentration (EC₅₀) of round 2 clones shifted to Dd2 levels in a 72-h assay (Table 1). In this way, the tolerance phenotype is similar to that in previous reports of atovaquone resistance (5, 13, 15). Occasionally, proliferation of round 2 clones did not disappear completely in these assays, indicating that viable parasites may still be present (our unpublished observation). Interestingly, tolerance was also observed when round 2 parasites were exposed to other cytochrome bc₁ complex inhibitors such as myxothiazol and antimycin A (Fig. 2 and Table 1). These two inhibitors target different sites of the enzyme, and, therefore, cross-tolerance indicates that there may be a global perturbation of the complex that alters the binding of both inhibitors (3, 16).

FIG 2.

Tolerance to alternative cytochrome bc₁ inhibitors in high-level DSM1-resistant parasites. Hypoxanthine uptake dose response of various P. falciparum clones to antimycin A (A) or myxothiazol (B) in the absence of DSM1. DSM1-sensitive clonal Dd2 (black filled circle) compared to partial DSM1-resistant round 1 clones (C, green open square; and D, green open triangle) and high-level DSM1-resistant round 2 clones (C53-1, blue square; C710-1b, blue circle; C710-2a, blue inverted triangle; and D73-1, blue triangle). Proliferation is calculated as a percentage of activity from dimethyl sulfoxide (DMSO) controls. Lines on the plots show the nonlinear curve fits of the data points from individual clones. EC₅₀s are listed in Table 1.

At first it is puzzling to understand how high-level resistance to a DHODH inhibitor (DSM1) confers tolerance to inhibitors of the cytochrome bc₁ complex. The present phenomenon is clearly different from traditional atovaquone resistance. While mutations in the mitochondrially encoded cytochrome b gene are commonly observed in parasites that are resistant to antimalarials that target the cytochrome bc₁ complex (3, 17–22), such parasites remain susceptible to DHODH inhibitors (23). Furthermore, whole-genome sequencing of round 1 and round 2 DSM1-resistant parasites failed to show causal mutations in the cytochrome bc₁ complex or anywhere in the nuclear or mitochondrial genome (10). This was also the case in a previous study where parasites developed resistance against a broad range of cytochrome bc₁ complex inhibitors; the genetic cause of this resistance was not identified (13). Since it had been proposed that the sole function of the respiratory chain was to regenerate ubiquinone for DHODH (5), the authors had speculated that DHODH had been uncoupled from the electron transport chain through an alternate source of ubiquinone oxidation. Despite the potential differences between atovaquone resistance and tolerance, a cytochrome bc₁“bypass” could also explain the current observations made in DSM1-resistant clones. Fumarate reductase activity, for example, had previously been proposed as an alternate way to regenerate ubiquinone for the use of DHODH, but this has not been biochemically investigated (24).

An alternative theory to explain atovaquone tolerance might involve some contribution of the neighboring upregulated genes besides DHODH. In the DSM1-selected parasites, round 2 resistance always involved high-level amplification of at least 35 kb of sequence surrounding the DHODH gene on chromosome 6. Intriguingly, the smallest conserved amplicon (8 genes) includes PlasmoDB identifier no. PF3D7_0603200, which is annotated as a putative mitochondrial chaperone BCS-1. In both bacteria and mitochondria, this gene product is required for the translocation of an essential subunit of the cytochrome bc₁ complex from the mitochondrial matrix to the intermembrane space (25). This protein has not yet been investigated in P. falciparum, but it is possible that amplification of the BSC-1 gene, along with DHODH, might confer enhanced functions to the complex and somehow increase atovaquone tolerance.

There are several antimalarials in the drug development pipeline that target pyrimidine biosynthesis (23, 26), and ones that target the respiratory chain (17, 27–32). Some of the new cytochrome bc₁ inhibitors are effective against atovaquone-resistant parasites (33), but it would be of interest to determine if they also participate in the cross talk described here. Since we routinely observe cross-resistance between DHODH inhibitors (our unpublished observations), the present studies raise important questions about the mechanisms of action and of cross-resistance directed at mitochondrial functions among antimalarials. This alert must be tempered by the fact that atovaquone tolerance is only seen at very high-level amplifications of the DHODH locus for cell lines that had undergone two sequential rounds of selection and that this tolerance is lessened with the addition of proguanil.