Inhibition of Glutathione Biosynthesis Sensitizes Plasmodium berghei to Antifolates

Abstract

Glutathione plays a central role in maintaining cellular redox homeostasis, and modulations to this status may affect malaria parasite sensitivity to certain types of antimalarials. In this study, we demonstrate that inhibition of glutathione biosynthesis in the Plasmodium berghei ANKA strain through disruption of the γ-glutamylcysteine synthetase (γ-GCS) gene, which encodes the first and rate-limiting enzyme in the glutathione biosynthetic pathway, significantly sensitizes parasites in vivo to pyrimethamine and sulfadoxine, but not to chloroquine, artesunate, or primaquine, compared with control parasites containing the same pyrimethamine-resistant marker cassette. Treatment of mice infected with an antifolate-resistant P. berghei control line with a γ-GCS inhibitor, buthionine sulfoximine, could partially abrogate pyrimethamine and sulfadoxine resistance. The role of glutathione in modulating the malaria parasite's response to antifolates suggests that development of specific inhibitors against Plasmodium γ-GCS may offer a new approach to counter Plasmodium antifolate resistance.

INTRODUCTION

Malaria causes approximately 438,000 deaths from 214 million cases annually (1). As no effective vaccine is yet available, chemotherapy is relied upon to combat the disease. However, Plasmodium falciparum has become resistant to all front-line antimalarial drugs in current use, such as chloroquine (2), combination sulfadoxine-pyrimethamine (Fansidar) (3, 4) and, most worrying, artemisinin and its analogs (5–9). Thus, new strategies are needed to counter drug resistance in Plasmodium.

Among the various metabolic pathways in Plasmodium, antioxidant systems are likely to be involved in drug sensitivity as they maintain redox homeostasis and counter oxidant stress triggered by antimalarials (10, 11). Glutathione (GSH) is a major antioxidant in Plasmodium (12, 13). De novo GSH biosynthesis consists of two steps: (i) production of γ-glutamylcysteine (γ-GC) from glutamic acid and cysteine by the rate-limiting γ-glutamylcysteine synthetase (γ-GCS) and (ii) production of GSH from γ-GC and glycine by glutathione synthetase (14, 15). Levels of intracellular GSH content and expression of GSH biosynthetic genes have been shown to correlate with artemisinin and chloroquine sensitivity (10, 16–18). However, knocking out or overexpressing the γ-gcs gene in Plasmodium berghei does not alter chloroquine or artemisinin sensitivity but rather affects the ability of the parasite to recrudesce after artemisinin treatment (19).

The effect of GSH on modulating Plasmodium sensitivity to other antimalarials is unclear. In this study, we investigated whether inhibiting GSH biosynthesis by genetically disrupting γ-gcs or by using a γ-GCS inhibitor would alter P. berghei sensitivity to other standard antimalarials, such as pyrimethamine, sulfadoxine, and primaquine.

MATERIALS AND METHODS

P. berghei infection of mice.

Six- to 10-week-old female BALB/c mice, purchased from the National Laboratory Animal Center, Mahidol University, Thailand, were used for P. berghei infection. Mice were infected with 1 × 10⁶ infected red blood cells (iRBCs) intraperitoneally and were given access to drinking water with or without 70 μg/ml pyrimethamine. Parasitemia was determined from tail vein blood either by manual counting of Giemsa-stained parasites in thin blood smears on microscope slides or by detection of parasites expressing green fluorescent protein (GFP) using a flow cytometer (Cytomics FC 500 MPL; Beckman Coulter), with 488-nm excitation and detection using a 525/40 band-pass filter. The P. berghei ANKA strain expressing GFP (PbGFP) used as the parental strain in this study was a gift of Andrew Waters and Chris Janse, Leiden University Medical Center, The Netherlands. Wild-type P. berghei ANKA was obtained from the Malaria Research and Reference Reagent Resource Center (MR4) (www.beiresources.org). All animal experiments were approved by the BIOTEC Institutional Animal Care and Use Committee, National Science and Technology Development Agency, Thailand.

Plasmid construction.

Plasmid pL0017 (a gift from Andrew Waters and Chris Janse) (20) was used for construction of pL0017-Δgcs for deletion of P. berghei γ-gcs. In brief, the 5′ and 3′ regions of γ-gcs were amplified from P. berghei ANKA genomic DNA using primer pairs 37 (5′ CTACTGAAGCTTCCATGGCGTACATGTACGCATATATTATACA 3′) and 38 (5′ CTACTGAAGCTTCAGGAGTCCAATGGGATCC 3′) and 39 (5′ CTACTGCTTAAGGGAAAAGGTGAAATTGCTCCTC 3′) and 40 (5′ CTACTGGGCGCCGGTGTGTATATACCAAACCGTTTC 3′), respectively. The 5′ γ-gcs amplicon, containing the region upstream of the gene and part of the coding sequence, was cloned into the HindIII site of pL0017, while the 3′ γ-gcs amplicon, containing the coding sequence and the region 3′ of the gene, was cloned into the AflII and KasI sites such that these regions flanked the drug-selectable marker cassette containing the pyrimethamine-resistant Toxoplasma gondii dihydrofolate reductase-thymidylate synthase (Tgdhfr-ts) gene.

Transfection of P. berghei.

P. berghei transfection was performed according to published protocols (21, 22). In short, P. berghei-infected blood was cultured in RPMI 1640 medium supplemented with 0.2% (wt/vol) NaHCO₃, 20% (vol/vol) fetal bovine serum, and 50 μg/ml neomycin and incubated under a 94% N₂, 1% O₂, and 5% CO₂ gas mixture overnight at 37°C with shaking. Schizonts were purified on a Nycodenz gradient, washed, and pelleted. Approximately 10 μg of linearized (using NcoI and KasI) pL0017-Δgcs or linearized pL0017 (using ApaI and SacII) was added to Nucleofector solution (Lonza) and transfected into purified PbGFP or P. berghei ANKA schizonts, respectively, using program U-33 of the Amaxa Nucleofector device (Lonza). Transfected parasites were injected intravenously into mice, which were treated with 70 μg/ml pyrimethamine in drinking water starting from day 1 postinfection. The presence of transgenic parasites was confirmed by PCR. Transgenic parasites were cloned into naive mice by intravenous injection of serially diluted infected mouse blood. Serial dilution was done in 200 U/ml heparin in phosphate-buffered saline. Integration of transfection plasmid into the correct locus was determined by PCR using primer pairs 51 (5′ GTAGCACTTGAAGAAAATATACG 3′) and 56 (5′ CGCATTATATGAGTTCATTTTACAC 3′) and 52 (5′ GCGATTCTATGTATGCACTC 3′) and 57 (5′ TGCAGCCCAGCTTAATTCTT 3′) for proper insertion of linearized pL0017-Δgcs at the 5′ and 3′ termini of γ-gcs, respectively, primer pair 13 (5′ CTACTGGGATCCATGGGTTTTCTAAAAATTGGAACT 3′) and 14 (5′ CTACTGTCTAGACTATGCACTGAATTGATACAAC 3′) for the presence of endogenous γ-gcs, and primer pairs 111 (5′ TTTGGATATTTTCATATATG 3′) and 112 (5′ TTTCCCAGTCACGACGTTG 3′) and 113 (5′ GTTGAAAAATTAAAAAAAAAC 3′) and 114 (5′ CTAAGGTACGCATATCATGG 3′) for the presence of pL0017 at the 5′ and 3′ termini of ssu, respectively (20).

RT-PCR of Pbdhfr-ts, Tgdhfr-ts, and Pbγ-gcs.

Levels of expression of the P. berghei dihydrofolate reductase-thymidylate synthase (Pbdhfr-ts), T. gondii dhfr-ts (Tgdhfr-ts), and Pbγ-gcs genes were assessed in each P. berghei line by reverse transcription-PCR (RT-PCR). Pbdhfr-ts was amplified with primer 268 (5′ CTGAATTTATACATGTATTGGG 3′) and primer 55 (5′ GGGGGCAGTTATAAATACAATCAATTGG 3′). Tgdhfr-ts was amplified with primer 269 (5′ ACGTCTGCAACCTAAAACCT 3′) and primer 270 (5′ CTAGACAGCCATCTCCATCT 3′). Pbγ-gcs was amplified with primer 5 (5′ CTACTGCTCGAGGTTATAAGCCCACATTCTAG 3′) and primer 6 (5′ CTACTGGAATTCCAATTACAGCTAGCTGATC 3′). Infected mice were sacrificed using CO₂ asphyxiation, and infected blood was collected by cardiac puncture and eluted through a cellulose column (Sigma-Aldrich) to remove white blood cells, and red cells were lysed with 0.2% (wt/vol) saponin. RNA was extracted from the washed parasite pellet using an RNeasy minikit (Qiagen) and digested with Turbo DNase (Ambion, Life Technologies) according to the manufacturers' instructions. RNA (1 μg) was converted to cDNA using Improm-II reverse transcriptase (Promega) and oligo(dT)₂₁ primer (25 nmol). cDNA was purified using phenol-chloroform extraction and ethanol precipitation and dissolved in 50 μl of distilled water. A 1-μl aliquot of cDNA was used as a template in a 15-μl PCR solution containing the previously mentioned primers and GoTaq DNA polymerase (Promega) according to the manufacturer's protocol.

In vivo drug sensitivity assay.

Drug sensitivity was determined using a 4-day suppressive test (23). In short, BALB/c mice (n = 5 to 10) were injected intravenously with 1 × 10⁷ iRBCs and treated orally with 200 μl of antimalarials at 4, 24, 48, and 72 h postinfection. Buthionine sulfoximine (BSO), chloroquine diphosphate, primaquine, pyrimethamine, and sulfadoxine were from Sigma-Aldrich, and artesunate was a gift from Dafra Pharma, Belgium. Stocks of antifolates were dissolved in dimethyl sulfoxide (DMSO) and diluted in hydroxypropylmethylcellulose (HPMC) suspension solution (0.5% HPMC, 0.4% Tween 80, 0.5% benzoyl alcohol) to the desired concentrations (24, 25). Four days postinfection, parasitemia was determined by detection of GFP-expressing iRBCs using a flow cytometer (Cytomics FC 500 MPL; Beckman Coulter). Mean parasitemia of the mock-treated control group was used as 0% inhibition. Data were fitted to a variable slope sigmoidal dose-response curve equation, and statistical significance (P < 0.05) was determined using GraphPad Prism 4.0 software (GraphPad Software, Inc.).

RESULTS

Construction of the γ-gcs-disrupted PbGFP strain and a P. berghei ANKA control line containing the Tgdhfr-ts cassette.

GSH has been implicated in modulating antimalarial sensitivity (17, 26–28). In order to evaluate this notion, the γ-gcs-disrupted PbGFP (Δgcs) strain was constructed by integrating linearized pL0017-Δgcs (Fig. 1A) into γ-gcs (Fig. 1B and C). Transfected parasites were cloned in vivo by limiting dilution in the absence of drug selection to obtain a Δgcs line. PCR was employed to verify that γ-gcs was disrupted by linearized pL0017-Δgcs at the 5′ (Fig. 1D) and 3′ (Fig. 1E) extremities and that intact γ-gcs was absent (Fig. 1F). In addition, a P. berghei ANKA line in which a pyrimethamine-resistant Tgdhfr-ts cassette (Fig. 2A) inserted into the nonessential ssu locus (Fig. 2B) was generated (WT+pL17) (Fig. 2C) in a similar fashion. PCR verified proper integration (Fig. 2D and E).

FIG 1.

Construction of the Δgcs line of P. berghei ANKA. (A) Diagram of linearized pL0017-Δgcs. The regions 5′ and 3′ of γ-gcs flank the Tgdhfr-ts cassette. (B) Diagram of the γ-gcs locus. Locations of primers for amplifying full-length γ-gcs are as indicated. (C) Diagram of the disrupted Δgcs locus, resulting from double homologous recombination between linearized pL0017-Δgcs and the native γ-gcs locus. Locations of primers for verifying the integration of linearized pL0017-Δgcs are as indicated. PCR product indicating integration of pL0017-Δgcs into the γ-gcs locus at the 5′ end (using primers p51 and p56) (D) and at the 3′ end (using primers p57 and p52) (E). PCR product indicating the presence of full-length γ-gcs and modified Δgcs locus (using primers p13 and p14) (F). pL17-Δgcs, pL0017-Δgcs plasmid; PbGFP, P. berghei ANKA strain expressing GFP; Δgcs, P. berghei ANKA strain expressing GFP with γ-gcs disrupted.

FIG 2.

Construction of wild-type P. berghei ANKA containing the Tgdhfr-ts cassette. (A) Diagram of the linearized pL0017 plasmid, containing the ssu sequence, a Tgdhfr-ts cassette, and a GFP cassette. (B) Diagram of the ssu locus. (C) Diagram of the ssu locus with pL0017 integrated. Locations of primers for verifying the integration of pL0017 into the ssu locus at the 5′ and 3′ ends are as indicated. (D) PCR product indicating the integration of pL0017 into the ssu locus at the 5′ end (using primers p111 and p112). (E) PCR product indicating the integration of pL0017 into the ssu locus at the 3′ end (using primers p113 and p114). pL17, pL0017; WT ANKA, wild-type P. berghei ANKA; WT+pL17, wild-type P. berghei ANKA with pL0017 integrated into the ssu locus.

Growth of the WT+pL17, PbGFP, and Δgcs strains in the absence and presence of pyrimethamine.

The growth of the Δgcs strain was significantly slower than those of both the PbGFP and WT+pL17 strains, while the WT+pL17 and PbGFP strains grew at a comparable rate (Fig. 3A), indicating that loss of γ-gcs, and not the integration of the Tgdhfr-ts cassette, was responsible for the growth retardation. These results confirm that GSH biosynthesis was important for optimal growth of P. berghei ANKA in vivo (29).

FIG 3.

Growth and specific gene expression of the WT+pL17, PbGFP, and Δgcs strains. (A) Growth of the PbGFP, Δgcs, and WT+pL17 strains in the absence of pyrimethamine. A total of 1 × 10⁶ iRBCs were injected intraperitoneally into BALB/c mice. Parasitemia was monitored using a flow cytometer detecting GFP-positive red blood cells. Data are shown as means ± standard errors of the means (SEM). Six to eight mice were used per group. Repeated-measures ANOVA with Bonferroni's multiple-comparison test was used to determine statistical significance. *, statistically significant at P < 0.05. (B) Growth of the PbGFP, Δgcs, and WT+pL17 strains in the presence of pyrimethamine. BALB/c mice were infected with 1 × 10⁶ WT+pL17, PbGFP, or Δgcs strain iRBCs and were given 70 μg/ml pyrimethamine in drinking water. Data are shown as means ± SEM. Three mice were used per group. Repeated-measures ANOVA with Bonferroni's multiple-comparison test was used to determine statistical significance. **, statistically significant at P < 0.01. (C) γ-gcs expression in the WT+pL17, PbGFP, and Δgcs strains was determined by RT-PCR (using primers p5 and p6). Oligo(dT) was used as a primer in the cDNA synthesis step. Primers binding within γ-gcs were used in subsequent PCR. (D) Tgdhfr-ts and Pbdhfr-ts expression was determined by RT-PCR (using primers p269 and p270 for Tgdhfr-ts and p268 and p55 for Pbdhfr-ts). PbGFP, P. berghei ANKA expressing GFP; Δgcs, P. berghei ANKA expressing GFP with γ-gcs disrupted; WT+pL17, wild-type P. berghei ANKA with pL0017 integrated into the ssu locus; RT, reverse transcriptase; +, with addition of RT; −, without addition of RT.

When infected mice were given 70 μg/ml pyrimethamine in drinking water, WT+pL17 grew well, but the PbGFP strain, which contained wild-type dhfr-ts, did not (Fig. 3B). The Δgcs strain, containing a pyrimethamine-resistant Tgdhfr-ts cassette, survived and grew to a detectable level on day 11 postinfection, but its growth remained low (Fig. 3B). These results indicate that the Δgcs strain was more sensitive to pyrimethamine than the WT+pL17 strain, to a level close to that of the PbGFP strain.

Expression of γ-gcs, Tgdhfr-ts, and Pbdhfr-ts in the WT+pL17, PbGFP, and PbGFPΔgcs strains.

RT-PCR assays were employed to demonstrate that the WT+pL17 and PbGFP strains expressed γ-gcs as expected, but the Δgcs strain did not (Fig. 3C). Pbdhfr-ts was expressed in all parasite lines, while Tgdhfr-ts was expressed in both the WT+pL17 and Δgcs strains but not in the PbGFP strain (Fig. 3D).

Pyrimethamine and sulfadoxine sensitivity of the Δgcs strain compared with the WT+pL17 and PbGFP strains.

The above results suggested that the lack of γ-gcs may make parasites more sensitive to pyrimethamine. Using a 4-day suppressive test, low doses (0.1 and 0.5 mg/kg) of pyrimethamine did not inhibit the growth of the WT+pL17 strain but inhibited 23% and 33% of the Δgcs strain's growth, respectively (Fig. 4A). Pyrimethamine sensitivity of the Δgcs strain was comparable to that of the PbGFP strain in a 4-day suppressive test where 0.5 mg/kg of pyrimethamine inhibited 42% of the PbGFP strain's growth (Fig. 4A). The difference between growth inhibition of the WT+pL17 and Δgcs strains was statistically significant at P < 0.01 at 0.1 mg/kg pyrimethamine and P < 0.001 at 0.5 mg/kg pyrimethamine, respectively (Fig. 4A). The difference between rates of growth inhibition of the WT+pL17 and PbGFP strains at 0.5 mg/kg pyrimethamine was statistically significant at P < 0.001 (Fig. 4A). Higher doses (2 mg/kg or 10 mg/kg) of pyrimethamine are required to significantly inhibit the growth of the WT+pL17 strain, becoming comparable to that of the Δgcs strain (Fig. 4A). This phenomenon also was observed with another antifolate, sulfadoxine, which inhibits dihydropteroate synthase (DHPS) in the folate biosynthetic pathway (30). Besides being resistant to pyrimethamine, the WT+pL17 strain was highly resistant to sulfadoxine compared with the PbGFP strain, showing no growth inhibition after treatment with up to 1 mg/kg sulfadoxine (Fig. 4B). In contrast, the growth of the Δgcs strain was significantly inhibited by 0.1, 0.5, and 1 mg/kg of sulfadoxine, to a level similar to that of the PbGFP strain (P < 0.001) (Fig. 4B).

FIG 4.

Disruption of γ-gcs sensitizes P. berghei ANKA to pyrimethamine and sulfadoxine. A total of 1 × 10⁷ WT+pL17, PbGFP, or Δgcs strain RBCs were injected intravenously into BALB/c mice. Infected mice were given daily doses of pyrimethamine (A) or sulfadoxine (B) for 4 days. Parasitemia was determined 4 days postinfection. Percentage of inhibition was calculated. Data are shown as means ± SEM. Five to 10 mice were used per group. Two-way ANOVA with Bonferroni's posttest was used to determine statistical significance. Asterisks indicate statistical significance: **, P < 0.01; ***, P < 0.001. Black asterisks show statistically significant differences between the WT+pL17 and PbGFP strains. Gray asterisks show statistically significant differences between the WT+pL17 and Δgcs strains. WT+ pL17, P. berghei ANKA expressing TgDHFR-TS and GFP; PbGFP, P. berghei ANKA expressing GFP; Δgcs, PbGFP strain with γ-gcs disrupted.

Effects of BSO on WT+pL17 strain sensitivity to pyrimethamine and sulfadoxine.

As deletion of γ-gcs increased pyrimethamine and sulfadoxine sensitivity in P. berghei ANKA (Fig. 4), the effects of BSO, a general γ-GCS inhibitor, were evaluated in a 4-day suppression test alone or in combination with pyrimethamine and sulfadoxine. The antifolate doses used were 0.5 mg/kg of pyrimethamine and 1 mg/kg of sulfadoxine, which were the doses that gave the biggest difference between the WT+pL17 and Δgcs strains among the doses tested (Fig. 4A and B). Treatment of WT+pL17 strain-infected mice with 0.5 mg/kg pyrimethamine alone or 100 mg/kg of BSO alone had no effect on parasite growth, but a combination of both resulted in 21% reduction in growth with significant difference between the WT+pL17 strain treated with 0.5 mg/kg pyrimethamine alone and the WT+pL17 strain treated with 0.5 mg/kg pyrimethamine and 100 mg/kg BSO with P < 0.05 (Fig. 5A). A combination of 0.5 mg/kg pyrimethamine and 50 mg/kg BSO resulted in 9% growth inhibition (Fig. 5A). Treatment of Δgcs strain-infected mice with 0.5 mg/kg resulted in 33% growth inhibition, while the same treatment of PbGFP-infected mice resulted in 42% growth inhibition (Fig. 5A). Similarly, treatment of WT+pL17 strain-infected mice with 1 mg/kg of sulfadoxine alone or 50 mg/kg of BSO alone did not affect parasite growth, but when combined, 37% growth inhibition was obtained (Fig. 5B). These results indicated that BSO can potentiate the effect of pyrimethamine and sulfadoxine and partially revert the resistance phenotype.

FIG 5.

BSO partially reverts pyrimethamine and sulfadoxine resistance. A total of 1 × 10⁷ WT+pL17, PbGFP, or Δgcs strain iRBCs were injected intravenously into mice. Infected mice were given drug orally for a total of four daily doses of pyrimethamine (A) or sulfadoxine (B) with or without the addition of BSO. Parasitemia was determined on day 4 postinfection. The percentage of inhibition of each dose of drug was calculated relative to parasitemia of the untreated group. Data are shown as means ± SEM. The unpaired t test was used to determine statistical significance. Asterisks indicate statistical significance: *, P < 0.05; **, P < 0.01; ***, P < 0.001. Four to seven mice per group were used in panel A, and six to nine mice per group were used in panel B.

Chloroquine, artesunate, and primaquine sensitivity of the Δgcs strain compared with the WT+pL17 strain.

Recently published work shows that knocking out γ-gcs did not alter chloroquine or artemisinin sensitivity in a 4-day suppressive test (19). In this study, we confirmed that the chloroquine and artesunate sensitivities of the Δgcs strain were comparable to those of the WT+pL17 and PbGFP strains (Fig. 6A and B). Moreover, the Δgcs and WT+pL17 strains showed comparable sensitivities to primaquine (Fig. 6C), even though primaquine's mode of action involves generation of radicals (31), suggesting that γ-gcs does not play a role in primaquine sensitivity in P. berghei ANKA.

FIG 6.

Disruption of γ-gcs does not alter chloroquine, artesunate, or primaquine sensitivity in P. berghei ANKA. BALB/c mice were infected intravenously with 1 × 10⁷ WT+pL17, PbGFP, or Δgcs strain iRBCs. Chloroquine (A), artesunate (B), or primaquine (C) was given to mice orally once a day for 4 days. Parasitemia was determined 4 days postinfection. Percentage of inhibition compared with the untreated control was calculated. Data are shown as means ± SEM. Four to six mice per group were used for chloroquine, five to eight mice per group were used for artesunate, and four to five mice per group were used for primaquine. Two-way ANOVA with Bonferroni's posttest was used to determine statistical significance. A gray asterisk indicates a statistically significant difference between the WT+pL17 and Δgcs strains at P < 0.05. WT+ pL17, P. berghei ANKA expressing TgDHFR-TS and GFP; PbGFP, P. berghei ANKA expressing GFP; Δgcs, PbGFP strain with γ-gcs disrupted.

DISCUSSION

In this study, we confirmed that γ-gcs can be knocked out in P. berghei, similar to a previous report (29). Interestingly, γ-gcs cannot be knocked out in P. falciparum (32). The discrepancy may come from differences in the two species' ability to transport glutathione or glutathione-related metabolites. Patzewitz et al. 2012 shows that P. falciparum does not significantly uptake glutathione from the environment, making γ-gcs essential in P. falciparum (32). No study has yet addressed whether P. berghei can uptake glutathione or γ-glutamylcysteine. If P. berghei can import glutathione, the ability to transport glutathione may be a reason why γ-gcs is nonessential in P. berghei. Other possible explanations why γ-gcs is nonessential in P. berghei include (i) P. berghei can upregulate other antioxidant pathways, such as thioredoxin and plasmoredoxin pathways, to compensate for the loss of γ-gcs, while P. falciparum cannot, and (ii) the difference between the cellular tropisms of P. falciparum and P. berghei. P. falciparum infects mature erythrocytes, while P. berghei infects reticulocytes, which contain richer metabolites and higher antioxidant levels than mature erythrocytes (33). Therefore, P. berghei might live in a friendlier environment than P. falciparum, resulting in the ability of Δgcs P. berghei to survive while Δgcs P. falciparum cannot.

Levels of GSH and its biosynthesis gene expression have been shown to be associated with chloroquine resistance in P. berghei (17, 18). GSH has been implicated in artemisinin action and resistance (34, 35). Primaquine acts by generating free radicals (31). Nevertheless, disruption of γ-gcs failed to sensitize P. berghei ANKA to chloroquine, artemisinin, or primaquine. A recent study employing a similar strategy in rodent malaria parasites reported similar findings with chloroquine and artemisinin (19).

We sometimes observe that low doses of drug increase parasitemia, as in the case of the WT+pL17 strain treated with 0.1 and 0.5 mg/kg of pyrimethamine (Fig. 4A), The Δgcs strain treated with 0.1 mg/kg artesunate (Fig. 6B), and the WT+pL17 strain treated with 0.5 mg/kg primaquine (Fig. 6C). Antifolate treatment is shown to increase DHFR-TS protein expression (36). Therefore, there may be a balance between an increase of enzyme activity by drug-induced expression and the activity-inhibitory effect of the drug. A low drug dose may induce more protein expression than it could inhibit, leading to an increased DHFR-TS level and thus higher parasite growth. A similar phenomenon with other drugs may occur, leading to higher parasitemia when parasites are treated with low doses of drug.

We demonstrate that disruption of P. berghei ANKA γ-gcs induced sensitivity to antifolates (pyrimethamine and sulfadoxine) in transgenic parasites carrying pyrimethamine-resistant Tgdhfr-ts. An equivalent effect (enhanced sensitivities to pyrimethamine and sulfadoxine) was obtained by treating P. berghei ANKA with intact γ-gcs but carrying pyrimethamine-resistant Tgdhfr-ts with BSO in a dose-dependent manner. BSO is an inhibitor of both host and parasite γ-GCS; nonetheless, higher concentrations of BSO have been used in vivo without apparent toxicity in animals (37, 38).

The mechanism linking GSH and antifolate sensitivity is unclear. As γ-gcs disruption leads to a malaria parasite growth defect (29) and sensitivity to antifolate, GSH biosynthesis may be involved in DNA synthesis or repair. GSH deficiency in other organisms, such as Escherichia coli, Saccharomyces cerevisiae, and Acinetobacter baylyi, leads to hypersensitivity to mutagens, similar to that in organisms with deficiency in DNA replication and repair (39–41). In mammalian cells, depletion of glutathione by BSO or other chemicals leads to DNA damage and prevents DNA repair, while the addition of glutathione or glutathione precursor results in a rescue of DNA repair (42–44). While antifolates directly inhibit folate and hence DNA biosynthesis and induce chromosome damage in vitro (45), GSH has been shown to play a positive role in cell proliferation, chromosome structure formation, and nuclear protein function (46–50). Depletion of GSH in P. berghei ANKA may make the parasites more susceptible to DNA damage and hence more sensitive to the effects of antifolates. Further studies are needed to test this notion.

In conclusion, this study shows that in a rodent malaria model, depletion of parasite GSH serves a dual purpose: limiting parasite proliferation in vivo and sensitizing parasites to the action of antifolates. This suggests that drug development targeting Plasmodium-specific γ-GCS, the rate-limiting enzyme in the GSH biosynthesis pathway, may help resurrect the use of Fansidar.

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.