Introduction

Chagas disease is endemic in 21 Latin American countries, with its prevalence increasing in non-endemic developed countries. It is estimated that over 12,000 people die every year from clinical manifestations of Chagas disease, and more than 75 million people are at risk of acquiring the disease1.The clinical course of Chagas disease begins with an acute phase that is usually asymptomatic and has low morbidity and mortality rates. This is followed by an indeterminate phase that may last a lifetime, characterized by the absence of symptoms. Only 30% of individuals undergoing the indeterminate phase may develop a chronic stage, featured by several clinical manifestations such as dilated cardiomyopathy and/or digestive disorders (megacolon and megaesophagus) and neurological alterations2,3. The causative agent of Chagas disease is the monoflagellate protozoan parasite Trypanosoma cruzi (T. cruzi). Like other protozoan pathogens, the life cycle of T. cruzi involves vertebrate and invertebrate hosts, with intracellular and extracellular parasite stages exposed to adverse environments. Consequently, the parasite undergoes significant morphological and biochemical changes to ensure its survival4.

The main chemotherapy currently available are the nitro-heterocyclic compounds nifurtimox and benznidazole. However, their use is restricted due to the serious side effects that frequently lead to treatment discontinuation. In addition, these drugs have low efficacy in the chronic phase of the disease5, thus making the search of new trypanocidal drugs an urgent issue.

During infection in mammals, parasites are exposed to the oxidative stress generated by macrophages. In this sense, T. cruzi has developed a complex system of antioxidant enzymes and reducing molecules6,7,8. The production of ROS has also been demonstrated during invasion of T. cruzi to mouse HL-1 cardiomyocytes9. Therefore, the capacity of T. cruzi to evade the oxidative response of the host cell relies on the activity of antioxidant enzymes and reducing molecules10. In turn, the multiple components required for the maintenance of the redox balance are in different subcellular compartments11,12,13.

The enzymatic detoxification system of T. cruzi comprises the thioredoxin homologue tryparedoxin (TXN) and the trypanothione (N1, N8-bis(glutathionyl) spermidine) [T(SH)2], which perform roles similar to those of glutathione in mammalian cells. Trypanothione is synthesized by trypanothione synthetase (TryS) and maintained in a reduced state by the NADPH-dependent flavoenzyme trypanothione reductase (TcTR)14. Several peroxidases have been identified in the parasite, each with different subcellular locations and substrate specificity. For example, glutathione peroxidase-I (TcGPXI) is located in the cytosol and glycosome, while TcGPXII is found in the endoplasmic reticulum, both providing resistance against exogenous hydroperoxides. Detoxification of H2O2 and peroxynitrites (ONOO −) is mediated by mitochondrial tryparedoxin peroxidase (mTXNPx), and cytosolic tryparedoxin peroxidase (cTXNPx)15,16,17.

These antioxidant enzymes use tryparedoxin (TXN), as an electrondonor, which is in turn reduced by trypanothione T(SH2). cTcTXNPx and mTcTXNPx present two domains that are common to subgroup 2-Cys, which is present in antioxidant enzymes from the peroxiredoxin family (I and II region). These domains have cysteine residues that are responsible for the peroxidase activity18

Increasing levels of mTXNPx and cTXNPx have been detected in the parasites during the differentiation of epimastigotes to metacyclic trypomastigotes, indicating that these enzymes are crucial to evade the host's immune effector mechanisms7,15,19. In addition, T. cruzi contains a plant-like ascorbate-dependent hemoperoxidase (TcAPX) in the endoplasmic reticulum20 and iron superoxide dismutase (Fe-SODs) which play a crucial role in detoxifying O2− generated within the cytosol, glycosomes, and mitochondria. In contrast, humans possess mitochondrial Mn-SOD and two distinct types of Cu/Zn-SOD21. Glutaredoxins are ubiquitous thiol proteins of the thioredoxin-fold superfamily and they are involved in cellular redox and/or iron sulfur metabolism, closely linked to the glutathione system22.

Since this antioxidant system is only present in parasites, it could be considered as a potential target for new chemotherapeutic agents against T. cruzi23.

Natural products are an interesting source of chemotherapeutic agents, particularly those used to treat infectious diseases. Sesquiterpene lactones (STLs) are C-15 terpenoid secondary metabolites present in higher plants which display a broad spectrum of biological activities, including activity against trypanosomatids24,25,26,27,28,29. Among them, dehydroleucodine (DhL), isolated from Artemisia douglasiana Besser (“matico”), has shown to be active against different stages of T. cruzi with low cytotoxicity on mammalian cells27. In addition, there is evidence that DhL and other sesquiterpene lactones exert leishmanicidal activity by the generation of an oxidative intracellular environment within the parasites26.

Considering these previous studies, it was of interest to determine whether DhL affects the parasite antioxidant defense machinery and if the α-methylene-γ-lactone group is involved in the antiparasitic action. We hypothesized that a pro-oxidative action of DhL might result from the interaction between the α-methylene-γ-lactone moiety of the STL and the sulfhydryl groups of reducing enzymes (2-Cys), or with other reducing molecules, such as trypanothione or glutathione. To test this hypothesis, we studied the effect of DhL on T. cruzi cultures overexpressing different enzymes of the redox defense system. Given that a critical aspect of antiprotozoal drug discovery is to elucidate the mechanism of action of potential candidates and identify their molecular targets, we extended the study to DhL derivatives obtained by chemical substitution in the α-methylene-γ-lactone group (DC-X2–DC-X11) to evaluate the role of this chemical group in the antiparasitic activity. Herein, we present evidence that DhL induces oxidative stress on T. cruzi epimastigotes by interfering with key redox enzymes, and that the α-methylene-γ-lactone group is, in part, involved in the trypanocidal activity.

Results

Antiparasitic effects of dehydroleucodine and derivatives against T. cruzi

Dehydroleucodine (DhL) is a natural sesquiterpene lactone with many biological activities. In this study, we confirmed the antiparasitic activity of DhL against the Dm28c strain of T. cruzi, with an IC50 of 2.179 µM (Table 1 and Fig. 1), lower than on other strain (e.g. Tulahuen). For DhL, the CC50 values and the selectivity index (SI) indicated that this compound was 6.6 and 14.3 times more cytotoxic to parasites than to Vero and macrophages cells respectively, while for BNZ the SI was lower 0.29 (Table 1). It appears that the effect of DhL is rather cytostatic than cytotoxic at concentrations lower than 10 µM (Figs. 1). Subsequently, we aimed at identifying the biologically active group(s) of the DhL, its mechanism of action and the possible targets on the parasite. Presumably, the α-methylene-ɣ-lactone group would be responsible for most of the biological activities of this molecule. To assess this, we carried out a comparative study of the effect of DhL with 10 semi-synthetic derivatives obtained by chemical substitution on the α-methylene-ɣ-lactone group. These derivatives were named DC-X2–DC-X11 (Fig. 2). Among the derivatives studied, only DC-X6 (IC50 = 7.3 µM at 24 h, R2 = 0.9) and DC-X11 (IC50 = 26.1 µM at 24 h; R2 = 0.6) displayed an antiproliferative effect on the parasites, although at a lesser extent than DhL (IC50 = 2.19 µM at 24 h; R2 = 0.87). Benznidazole showed a IC50 value of 486 µM (R2 = 0.96). Considering these results, all subsequent studies were done with DhL and the derivatives DC-X6 and/or DC-X11.

Table 1 Trypanocidal cytotoxic activity and selectivity index on epimastigotes, Vero and macrophage cells of the active sesquiterpenoids.
Full size table
Figure 1
figure 1

The effect of DhL on the growth of T. cruzi epimastigotes. Parasites (2 × 106/mL) were incubated with 1–20 μM of DhL as indicated. Control: no treated parasites, control DMSO: parasites incubated with DMSO < 0.05%. Aliquots were collected every 24 h and the parasites were counted in a hemocytometer. Each point of the curve represents the mean of parasite concentration ± SEM from three independent experiments. (**) (****) indicate significant differences from the controls (p < 0.002 and p < 0.001 respectively).

Full size image
Figure 2
figure 2

Structure of dehydroleucodine and chemical derivatives.

Full size image

The effect of DhL and derivatives on ROS generation

Since the derivatives showed lower activity than the DhL, it could be inferred that the α-methylene-γ-lactone group plays a crucial role in the biological activity of DhL. The antiproliferative effect of DhL was blocked by the sulfhydryl group donor glutathione (GSH) (Fig. 3A, B), but protection against DC-X6 was observed only at 48 h incubation (Fig. 3C). No GSH-induced protection was observed against DC- X11 (data no shown).

Figure 3
figure 3

The reducing agent gluthatione (GSH) protects T. cruzi epimastigotes from DhL and DC-X6 antiproliferative activity. Parasites were incubated with different concentrations of compounds, in the absence or in the presence of 3 mM GSH for 24 h or 48 h. After incubation aliquots were collected and the parasites were counted. Bars represent the means ± SEM of parasite concentration from three independent experiments. (**) (***) and (****) indicate significant differences from the controls (p < 0.002, p < 0.01 and p < 0.001 respectively). GSH alone was also used as control. (A) 24 h, (B) 48 h of DhL. (C) 24 and 48 h of DC-X6.

Full size image

The protective effect of GSH could be due to the scavenging of oxidative oxygen species (ROS), as occurs in other biological models. Assays performed with H2DCFDA showed that the incubation with 41 µM DhL for 12 h induced a significant increase of intracellular ROS (Fig. 4A) while the ROS levels induced by DC-X6 and DC-X11 were of a lower magnitude (Fig. 4B, C). Autofluorescence (in the absence of parasites) of the probe was less than 0.05% (data not shown).

Figure 4
figure 4

The effect of DhL, DC-X6 and DC-X11 on ROS generation. Parasites (3 × 106/mL) were incubated with 20.5 or 41 μM of DhL (A) and in the absence or in the presence of 3 mM GSH; 27.5 or 55 μM of DC-X6 (B) 32 or 98 μM of DC-X11 (C) for 12 h and processed for ROS measurement as described in material and methods. Bars represent the means of the probe fluorescence intensity/106 parasites ± SEM from three independent experiments. (*)(**)(***)(****) indicate significant differences from the controls (p < 0.05, p < 0.002, p < 0.01 and p < 0.001 respectively).

Full size image

The increase of ROS levels could be due to an interference with endogenous reducing molecules [e.g. trypanothione T(SH)2] or a blockade of sulfhydryl groups of crucial reducing enzymes, thereby altering the parasite redox homeostasis. To assess the first hypothesis, the interaction between DhL-GSH, as well as DhL-T(SH)2, was analyzed using unbiased atomistic molecular dynamics simulations with Gromacs in the order of the microsecond. The DhL, T(SH)2 and GSH molecules were parameterized using LigParGen and Charmm-GUI. As shown in Fig. 5 the radial distribution function of the distances between DhL-GSH and/or DhL-T(SH)2did not show significant differences. Similar results were obtained with DC-X11 (data no shown). The maximum of both curves was observed at ~ 0.6 nm, indicating the most frequent distance, and neither simulation evidenced a dominant affinity between the molecule pairs. Quantitatively, we show that the radial distribution function g(r) vanishes for distances below 0.25 nm (Fig. 5a). Additionally, Fig. 5d shows the Potential of Mean Force (PMF) calculations for DhL and GSH as a function of the distance between their respective centers of mass. In good agreement with the radial distribution distances in Fig. 5a, free energy steeply rises as molecules come together, therefore ruling out a potential direct interaction. The results of this analysis would indicate that the oxidative stress induced by DhL would not depend on a direct interaction with glutathione or trypanothione. Since molecular dynamics are limited to mechanical interactions where no chemical reactions are considered, we attempted to evaluate a possible direct interaction between DhL and GSH by 1H-NMR analysis. We observed a diminution of both signals corresponding to the methylene group at 6.051 (d, J = 3.5 Hz) and 5.559 ppm (d, J = 3.5 Hz) compared to the signal of the endocyclic double bound (6.161 ppm, br s). Additionally, an increment of a dd (J = 14.2 and 4.5 Hz) at 3.180 ppm was observed (Fig. 6). This signal is assigned to the new aliphatic CH2 group bounded to the glutathione sulfur and was confirmed by the occurrence of HMBC correlation with carbonyl group of lactone (data not shown).

Figure 5
figure 5

Simulations of atomistic molecular dynamics. The radial distribution function of the distances between (b) DhL and Glutathione [(a)-black curve]; (c) DhL and Trypanothione [(a)-red curve] did not show significant differences. DhL, Trypanothione and Glutathione molecules were parameterized using LigParGen and Charmm-GUI. (d) Potential of mean force (PMF) calculations for DhL and GSH.

Full size image
Figure 6
figure 6

1H-NMR of the reaction between DhL and GSH at T = 0 (blue) and T = 24 h (red). The diminution of signals corresponding to vinyl CH2 group is observed (black arrow) in comparison with the endocyclic double bound (green arrow).

Full size image

DhL affects T.cruzi trypanothione synthetase activity

Trypanothione synthetase (TryS) is a key enzyme to maintain adequate levels of trypanothione in T. cruzi and to deal with oxidative stress. How it is observed in the Fig. 7 DhL treatment induced an increase of TryS activity, similar to that for H2O2. Likewise, this increase was reversed when glutathione was added to the assay.

Figure 7
figure 7

Activity of Trypanothione Synthetase (TryS). Parasites (3 × 107) were incubated with different concentrations of DhL, in the absence or in the presence of 3 mM GSH for 3 h and processed for TryS activity measurement as described in material and methods. Controls were performed with DMSO (< 0.05%) or with GSH alone. We used different concentration of H2O2 (0.5, 1 and 2 mM) as a positive control. Bars represent the specific enzymatic activity of TryS mU/mg protein from three independent experiments. (*)(***) indicate significant differences from each other (p < 0.05 and p < 0.01 respectively, obtained by Turkey–Kramer multiple comparisons test).

Full size image

Mitochondrial tryparedoxin peroxidase overexpression protects parasites against DhL effects

In order to elucidate the mechanism of action of these compounds and identify potential molecular targets, we evaluated the antioxidant enzymes. We used T. cruzi Dm28c strain epimastigotes that stably overexpress mitochondrial tryparedoxin peroxidase (mTXNPx), cytosolic tryparedoxin peroxidase (cTXNPx), tryparedoxin II (Tc-TXN II) and glutaredoxin (Tc-GRx). As shown in Fig. 8, the overexpression of mTXNPx exerted a protective effect against DhL.

Figure 8
figure 8

The effect of DhL on T. cruzi epimastigotes that overexpress reducing enzymes. Dm28c epimastigotes (1.5 × 106 cells/mL) that overexpress; mTXNPx, cTXNPx, Tc-GRx, and Tc-TXN II were incubated with different concentrations of DhL for 72 h. Then, the viability of the parasites was determined by resarzurin assay as described in material and methods. The curves represent the percentages of viability with respect to control (without drug, 100% of viability) at each DhL concentration. Each point on the curves represents the means ± SEM from three independent experiments. Parasites transfected with empty vector (pTEX) were used as controls. (*)(***)(****) indicate significant differences from the control (p < 0.05, p < 0.01and p < 0.001 respectively).

Full size image

Accordingly, a lower amount of ROS was observed in parasites overexpressing mTXNPx in the presence of DhL (Fig. 9). A protective tendency, although not statistically significant, was also observed in cTXNPx overexpressing parasites (data not shown). In turn, parasites transfected with the empty pTEX vector behaved similarly to non-transfected controls.

Figure 9
figure 9

The effect of DhL on ROS generation in overexpressing mTXNPx epimastigotes. Parasites (3 × 106/mL) were incubated with 20.5 µM DhL for 12 h and processed for ROS measurement as described in material and methods. Bars represent the means of the probe fluorescence intensity/106 parasites ± SD from three independent experiments. Treatment with H2O2 was used as positive control. (****) indicate significant differences from the controls (p < 0.001).

Full size image

Ultrastructural changes induced by DhL

Biochemical and molecular alterations caused by drugs are usually accompanied by important ultrastructural changes in the parasites which can be studied by TEM. We observed that 2 μM DhL induced mitochondrial swelling on epimastigotes (48 h), being more evident this effect at 72 h of treatment (70% of parasites were altered), which is indicative of organelle dysfunction induced by the compound (Figs. 10, 11). This effect was accompanied by a slight kinetoplast deformation, but no abnormalities in the plasma membrane were observed. At higher concentrations (4 μM), most of the parasites were totally destroyed at 48 h of treatment (Figs. 10, 11).

Figure 10
figure 10

Effect of DhL on T. cruzi epimastigote ultrastructure. The parasites were incubated in diamond medium alone (A) or with 2 µM of DhL for 48 h (B,C) or 72 h (D, E); or with 4 µM DhL for 48 h (F). N nucleus, K kinetoplast, Mt mitochondrion, FI flagellum.

Full size image
Figure 11
figure 11

Morphometric estimation of treated parasites with DhL. The phenotypes observed more frequently were normal morphology, mitochondrial swelling and broken parasites. Evaluation of 100 cells for each condition from one experiment.

Full size image

On the other hand, no major ultrastructural alterations were observed by treatment with DC-X11, although parasites with some peculiarities such as multiple nuclei, flagella or kinetoplasts were observed (data no shown), suggesting alterations in the cell cycle. Further studies should be carried out to explain this phenomenon.

Discussion

T. cruzi is exposed to several sources of oxidative stress during its digenetic life cycle. Inside the vector, epimastigotes are exposed to the insect’s innate immune response, this includes reactive oxygen intermediates (ROI), reactive nitrogen intermediates (RNI), and pro-apoptotic molecules, among other23. Furthermore, reactive oxygen species (ROS) are produced by its own metabolism, but high ROS levels are also generated as part of the host's immune response30,31. To resist the host’s oxidative stress, the parasite has developed complex defense mechanisms that include ROS detoxification pathways comprising enzymes and reducing molecules different from those found in mammalian cells. For this reason, the antioxidant system of parasites is considered a potential target for new antiparasitic therapies.

In this study we attempted to elucidate the mechanism of action of the sesquiterpene lactone dehydroleucodine (DhL) on T. cruzi epimastigotes focusing on the parasite redox machinery. Since DhL derivatives carrying substitutions on the α-methylene-γ-lactone group displayed less activity than the DhL, we suggest that this group would play a role in the antiparasitic action of DhL. The fact that the antiproliferative effect of DhL is blocked by GSH and that the compound induced an increase in the intracellular ROS levels confirm that DhL interferes with the parasite redox homeostasis. This effect would be accomplished through a direct interaction of the α-methylene-γ-lactone with sulfhydryl groups of reducing molecules, such as trypanothione or GSH, or with enzymes that are crucial for the maintenance of redox homeostasis. As it occurred with other lactones, a direct interaction between GSH and DhL was confirmed by 1H-NMR32. Intriguingly, trypanothione synthetase activity was increased by DhL treatment, similarly to that of H2O2. This enzyme is crucial for maintenance of adequate trypanothione levels, the main thiol redox metabolite in trypanosomatids. We could explain the increased activity of TryS as an immediate response to ROS accumulation due to DhL action, at least at the time point studied. Also a high consume of reduced trypanothione by ROS accumulation may activate both TryS and trypanothione reductase, not evaluated in this work. In experiments carried out with epimastigotes stably overexpressing peroxidase enzymes we observed that the overexpression of mTXNPx and cTXNPx, at a lesser extent, can protect the parasites against the pro-oxidant effects of DhL. The redox enzymes mTXNPx and cTXNPx are considered virulence factors, since their overexpression is associated with increased parasite infectivity on phagocytic and non-phagocytic cells and increased parasite resistance to exogenous H2O2 and ONOO−11,12,19. Effects were not observed in parasites overexpressing either Tc-TXN II or Tc-GRx, indicating that the oxidation products generated by DhL may serve as peroxidase substrates. Tc-TXN II is a tail-anchored protein with a glycosomal, endoplasmic reticulum, or outer mitochondrial membrane distribution with a cytoplasmic orientation of the redox domain. The reducing activity of this enzyme causes the transference of reducing equivalents from trypanothione to various proton acceptors, including cTXNPx and mTXNPx33. On the other hand, TEM has been an important tool to investigate the mechanism of action of trypanocidal drugs. As observed with other drugs, DhL induced some ultrastructural alterations on epimastigotes such as a mitochondrial swelling, which is indicative of organelle dysfunction. These alterations may be attributed to effects on the components in this organelle, such as membrane proteins or enzymes. The mitochondrial swelling can also be attributed to membrane lipid peroxidation induced by the accumulated ROS 34. Trypanosomatids have long been considered to have limited capacity to eliminate ROS such as O2– and H2O2, mainly because they lack catalase and classical selenium-containing glutathione peroxidases6,8,31. However, these parasites have developed antioxidant defenses comprising a wide range of different enzymatic detoxification systems.

A mild vacuolization was also observed in some cells after treatment with DhL, although the integrity of the plasma membrane was preserved.

Taking into account the results presented herein, it can be concluded that one of the mechanisms of action of DhL involves interference with the antioxidant machinery of T. cruzi. Its α-methylene-γ-lactone group plays a role in this effect, either by interacting with sulfhydryls of reducing molecules or enzymes, inducing the accumulation of ROS and causing mitochondrial dysfunction in the parasites. Additionally, we have provided new molecular tools to identify the targets and mechanisms of action for potential new drugs against Chagas disease.

Materials and methods

Extraction and purification of dehydroleucodine (DhL)

The lactone was isolated from Artemisia douglasiana Besser as previously described by Giordano et al.35. The purity of these compounds (> 95%) was confirmed by 13C-NMR, melting point analysis, and optical rotation. The semi-synthetic derivatives were obtained by chemical substitutions in the α- methylene-γ-lactone group (as indicated in the Fig. 2), and they were named DC-X2, DC-X3, DC- X4, DC-X5, DC-X6, DC-X7, DC-X8, DC-X9, DC-X10 and DC-X11. Stock solutions of all compounds were prepared in dimethyl sulfoxide (DMSO) according to their solubility. The purity of these compounds was ~ 95%, as evaluated by NMR. DhL purification and derivatives synthesis were carried out by Dr. O. Giordano, C. Tonn and D. Cifuente at INTEQUI (Chemical Technology Research Institute) of the National University of San Luis.

Parasite culture

T. cruzi epimastigotes, Dm28c strain36 were cultured in diamond liquid medium (0.106 M NaCl, 29 mM KH2PO4, 23 mM K2HPO4, 12.5 g/L tryptone, 12.5 g/L tryptose and 12.5 g/L yeast extract, adjusted to pH 7.4), supplemented with 75 µM hemine, 75 IU/mL penicillin, 75 µg/mL streptomycin and 10% heat-inactivated fetal bovine serum (FBS) at 28 °C. In this work we used parasites stably overexpressing of the following reducing enzymes: mitochondrial tryparedoxin peroxidase (mTXNPx), cytosolic tryparedoxin peroxidase (cTXNPx), glutaredoxin (Tc-GRx) and tryparedoxin II (Tc-TXNII). The overexpression of these enzymes was obtained by transfection of epimastigotes with the pTEX vector (pTEX-mTXNPx; pTEX-cTXNPx, pTEX-Tc-GRx, pTEX-Tc-TXN II)12,37 For Tc-TXNII, primers 5ʹ ggagatctATGCTGCCACGCGTACTTGG3ʹ and 5ʹ ggaagcttCACCGCCAGAATTGATACAG-3ʹ were used (lowercase letters represent non-coding sequences containing the restriction sites Bam HI and HindIII, respectively, and were introduced to facilitate the cloning of PCR products), and the amplified products were cloned into the BamHI–HindIII sites from pTEX, generating the pTEX-Tc-TXNII construct. Transfected parasites were grown in diamond medium with 200 μg/mL of G418. Wild type strains transfected with the empty pTEX were used as control. The overexpression of redox enzymes was confirmed by Western blotting using the corresponding specific antibodies12.

In vitro evaluation of DhL on T. cruzi epimastigotes

Exponentially growing wild type parasites (2 × 106 cells/mL) were incubated at 28 °C either in the absence or in the presence of different concentrations of each compound during 24–72 h. Parasites incubated with DMSO alone (0.05%) were also used as control. Aliquots of epimastigotes were collected every 24 h, fixed with 2% paraformaldehyde (PFA), and counted in a Neubauer hemocytometer27. The overexpressing parasites (1.5 × 106 cells/mL) were incubated under conditions similar to those of the wild type, but DhL was used from 0.3125–41 μM (Fig. 8). The viability of genetically modified parasites was determined with resazurin38. In all cases, IC50 values were calculated, and the subsequent experiments were carried out with those compounds that showed a significant effect on the parasites.

Cytotoxicity on mammalian cells

The cytotoxicity assay was performed on Vero cells and RAW264.7 macrophages growing in DMEM medium supplemented with 10% fetal bovine serum (SFB) at 37 °C in a 5% CO2 humidified incubator. Briefly, 2 × 105 vero cells were placed in a 96-well plate for 24 h in a 5% CO2 incubator at 37 °C. These cells were incubated with serial dilutions of the DhL and BNZ (0.5, 1, 2, 4, 8, 16, and 32 µM for DhL; and 2.5–240 µM for BNZ) at 5% CO2 atmosphere for 24 h and 48 h. Cell viability was evaluated with the MTT assay39. Dose–response curves were plotted and the CC50 values were calculated.

Effect of DhL and DC-X6 in the presence of reduced glutathione

Wild type T. cruzi epimastigotes (2 × 106 parasites/mL) were diluted into Diamond medium, containing or not 3 mM of the reducing agent glutathione (GSH), and then treated either with 2, 4 and 10 µM DhL or 27.5 and 55 µM DC-X6, Controls were performed with DMSO (< 0.05%) or with GSH alone. The parasite concentration was determined at 24 and 48 h by counting the cells in a Neubauer hemocytometer.

Measurement of ROS

The intracellular generation of ROS was measured using the fluorescent probe H2DCFDA40. Briefly, WT, cTXNPx or mTXNPx epimastigotes (3 × 106 cells/mL) were treated with 20.5 µM or 41 µM DhL, 27.5 or 55 µM DC-X6 and 32.6 µM or 98 µM of DC-X11 for 12 h and then incubated with 20 µM of the probe for 1 h at 28 °C in the dark. The fluorescence intensity of H2DCFDA was measured at 492 nm excitation and 527 nm emission wavelengths. To validate the assay, H2O2 (2 mM 30 min) was positive control for ROS generation. We also added a control in the presence of 3 mM of GSH to exclude the possibility that GSH might interact with the probe, (results were standardized as fluorescence units per million parasites). At this condition, toxicity of H2O2 was less than 10%.

Trypanothione synthetase activity (TryS)

Wild type T. cruzi epimastigotes (3 × 107 parasites/mL) were treated with 20, 61 or 102 µM of DhL for 3 h, in the absence or the presence of 3 mM of the reducing agent glutathione (GSH). Controls were performed with DMSO (< 0.05%) or with GSH alone. To validate the assay 0.5, 1 and 2 mM of H2O2 (30 min) were used as a positive control. The specific enzyme activity of TryS was determined using malachite green to measure inorganic phosphate released from ATP during catalysis41. Briefly, in a 96 well plate, a mixture of 100 μL reaction buffer (50 mM HEPES, 5 mM MgCl2, 1 mM DTT, 0.5 mM GSH, 10 mM spermidine, 1 mM ATP, pH 8) and 10 μL protein lysate was incubated 37 °C for 30 min. The reaction was stopped with fresh solution of 0.5 mM malachite green and 34 mM ammonium molybdate in 4 M HCl and 2% Tween 20. Absorbance was measured at 650 nm after 30 min. Values were normalized to total protein content determined by the Lowry method. Enzymatic activity was expressed as units (U) per milligram of total protein. One unit (U) represents the enzyme amount hydrolyzing 1 μmol of ATP per minute under defines conditions.

Simulations of atomistic molecular dynamics

In order to predict a possible interaction of DhL with reducing molecules, we performed unbiased microsecond-length molecular dynamics simulations using the Gromacs-2022.6 software42. For the molecular dynamics simulations, trypanothione T(SH2) and glutathione (GSH) molecules were parameterized with DhL using LigParGen43 and Charmm-GUI44, and the radial distribution function (RDF) of the distances between the molecules were analyzed. The maximum of both curves was observed at ~ 0.6 nm (the most frequent distance).

1H-NMR analysis of DhL-GS adduct

DhL (0.041 mMol, 10 mg) was mixed with GSH (0.204 mMol, 62.9 mg) and stirred for 24 h in 1 mL of 250 μM DTT aqueous solution supplemented with 10% of D2O32. 1H-NMR were recorded at 25 °C in a Bruker Avance NEO 400 MHz equipped with a BBO probe with z gradient. Both experiments were realized at T = 0 and T = 24 h using a zgcppr program with TD = 32 K, 32 scans, SW = 15 ppm and O1P = 4.702 ppm. Spectra were processed with TopSpin 4.4.0 software with exponential multiplication using a LB = 0.3 Hz.

Transmission electron microscopy

All procedures were carried out according to Brengio et al.27. Briefly, parasites were treated with different concentrations of DhL for 48 h or 72 h. Epimastigotes were then centrifuged at 1000 × g for 10 min and fixed with 2.5% glutaraldehyde. Subsequently, they were washed three times with phosphate-buffered saline pH 7.2 (PBS) and post-fixed overnight with 2% OsO4. After washing twice with PBS, cells were stained with 1% uranyl acetate. Samples were dehydrated by treatment with ethanol of increasing grades and acetone and embedded in Epon 812. Sixty nm ultrathin slices were obtained with an automatic Leica-ultracut R ultramicrotome, and analyzed under a Zeiss EM900 microscope.

Statistics and data analysis

All assays were carried out in triplicate. All statistical analyses were performed in the GraphPad PRISM software version 10.1.1, and data expressed as the mean ± SEM of three independent experiments, unless otherwise specified. The statistical significance of results was determined by one-way analysis of variance (ANOVA) followed by Dunnett’s or Turkey–Kramer multiple comparisons test. The compound concentration that inhibited the growth of the tested parasite (IC50) by 50% was obtained by non-linear regression of log(inhibitor) vs. normalized response, or inhibitor vs normalized response.