Trypanosoma cruzi (T. cruzi) is the intracellular parasite that causes Chagas disease, which is considered a neglected disease and remains a public health problem in Latin America, estimated to affect approximately 6 million people and with 70 million at risk of infection (Pérez-Molina and Molina, 2018). Trypanosoma cruzi has a complex life cycle that involves two hosts, the insect vector and the vertebrate host, and four stages of the parasite. The replicative, non-infective epimastigote stage is found in the insect midgut and, during its migration to the hindgut, differentiates to the infective, non-replicative metacyclic trypomastigotes that can invade different vertebrate cells. Inside cells, trypomastigote transforms into the infective, replicative intracellular amastigote. Following several rounds of replication, the infective blood trypomastigote escapes and disseminates the infection to other tissues. During the life cycle, parasite survival depends on its rapid and efficient adaptation to the distinct environments faced, which is accomplished by extensive biochemical and morphological changes. Trypanosoma cruzi is a heterogeneous species, with large genetic variability and different strains that circulate between the vertebrate host and the insect vector. This heterogeneity may explain the variation in the clinical manifestation of Chagas disease and the geographical differences in morbidity and mortality (Laurent et al., 1997; Da Silva Manoel-Caetano and Silva, 2007; McCall and McKerrow, 2014).

Activation of professional phagocytes (macrophages and neutrophils) with the concomitant generation of oxidant species is a medullar innate immune process for the control of acute T. cruzi infection. Recent data reinforce the hypothesis that parasites more prepared to deal with the host-oxidative assault are more efficient for the establishment of the infection (Piacenza et al., 2009a; Alvarez et al., 2011; Martinez et al., 2014; Estrada et al., 2018; Piacenza et al., 2019). Nevertheless, T. cruzi contains a complex system of antioxidant enzymes and related proteins that allows it to overcome the oxidants, produced by the host immune response (Wilkinson et al., 2000; Mateo et al., 2008; Piacenza et al., 2009b; Wilkinson et al., 2002a; Wilkinson et al., 2002b; Wilkinson et al., 2002c). This antioxidant system has multiple components distributed in all parasite subcellular compartments, where they maintain the redox balance playing a key role in parasite virulence and persistence in tissues (Piacenza et al., 2013; Estrada et al., 2018; Piacenza et al., 2019).

Peroxide metabolism in T. cruzi relies on five peroxidases located in different subcellular compartments: heme-containing cytochrome c ascorbate peroxidase, two glutathione-like peroxidases and two peroxiredoxins, ubiquitous peroxidases that depend on cysteine residues to reduce peroxides (Wilkinson et al., 2000; Wilkinson et al., 2002a; Wilkinson et al., 2002b; Wilkinson et al., 2002c; Hugo et al., 2017). Both the cytosolic (CPX) and the mitochondrial peroxiredoxin (MPX) are two cysteine-typical peroxiredoxins which use tryparedoxin (TXN) as reducing substrate. The oxidized TXN is then reduced by trypanothione [T(SH)₂], a low molecular weight thiol unique to trypanosomatids, at the expense of NADPH. Trypanosoma cruzi peroxiredoxins are able to detoxify H₂O₂, peroxynitrite (ONOOH), and short-chain organic hydroperoxides (ROOH), and its level of expression has been correlated with the virulence of different strains (Wilkinson et al., 2000; Trujillo et al., 2004; Piñeyro et al., 2008; Piacenza et al., 2009b). In most two cys-typical peroxiredoxins, such as MPX and CPX, the basic functional unit is the dimer, but generally, they are able to establish higher-order structures specifically as octameric, decameric, and dodecameric toroids (Cao and Lindsay, 2017). In recent years, it has been described in many organisms that peroxiredoxins, besides its function as peroxidases, have other activities such as holdase and as part of signaling pathways (Rhee et al., 2012; Perkins et al., 2015; Sobotta et al., 2015). Holdases are part of the molecular chaperone family which prevents aggregation of proteins in an ATP-independent manner (Folgueira and Requena, 2007; Kim et al., 2013). Holdase activity is crucial for cellular protection during different stress conditions that induces protein damage and unfolding with the possible formation of protein aggregates. The interplay between the peroxidase and holdase function of peroxiredoxins is regulated by several environments in the cell such as pH, temperature, ions, posttranslational modifications, and redox balance (Jang et al., 2006; Saccoccia et al., 2012; Morais et al., 2017a). Moreover, holdase activity seems to be related to the oligomeric state of the enzyme, taking into consideration that holdase activity is mainly reported in decameric or higher molecular weight oligomers, which means that conditions that favor the said structure will stimulate this activity in peroxiredoxins. It is proposed that oligomerization dynamics and, therefore, holdase function are dependent on the redox state of the cell, given that oxidized forms of peroxiredoxins, which have mainly dimeric conformations, do not exhibit holdase activity, whereas their reduced or overoxidized forms, which form decamers and/or higher molecular weight oligomers, display holdase activity (Wood et al., 2002; Wood et al., 2003; Rhee and Woo, 2011; Perkins et al., 2015). However, there are reports that also show that oxidized peroxiredoxins are able to acquire high molecular weight conformations, and thus, the precise relationship between redox state and holdase activity will depend on the specific peroxiredoxin under study (Yewdall et al., 2016). Regarding these novel functions, it was found that T. cruzi CPX has the ability to prevent aggregation of malate dehydrogenase in vitro independently of its redox state (Piñeyro et al., 2019). Furthermore, the crystal structure of the CPX is available and shows that it is a toroid-shaped decamer in its active and reduced form (Piñeyro et al., 2005), like most peroxiredoxins that act as holdases. It was shown that this peroxiredoxin does not disassemble into dimers upon oxidation. Nonetheless, it has not been yet proven in vivo that T. cruzi peroxiredoxins have holdase activity. In Leishmania infantum, another trypanosomatid with long-lasting residence in the macrophage phagosome, it has been reported that mitochondrial peroxiredoxin (LiPrx1m) presents holdase activity in vitro and that a deficiency in this enzyme yields parasites more sensitive to a rise in temperature (Castro et al., 2011). These deficient parasites also presented a decreased survival in macrophages, but interestingly, they could be rescued by the expression of the LiPrx1m in which the peroxidatic cysteine (Cp) was absent, indicating that other activity than the classical peroxidase (i.e., holdase) is involved in protecting amastigotes to macrophage-derived stressors (Castro et al., 2011; Teixeira et al., 2015).

Treatment of Chagas disease relies mainly on two old and poorly specific chemotherapeutic agents: benznidazole (BZ) and nifurtimox (NFX). These drugs have several disadvantages; they are only effective in the acute phase of the disease, treatment may last for months with important side effects, and there are resistant strains (Rassi et al., 2010). Both of these drugs are nitroheterocycle compounds: NFX is a nitrofurane derivative, while BZ is a nitroimidazole. The mode of action of NFX is still elusive and two major mechanisms have been proposed, which agree in the fact that NFX has to be metabolized to act as an anti-trypanosomatid drug. First, NFX toxicity was explained by NFX redox cycling with superoxide radical (O₂ ^•−) and H₂O₂ production (Docampo and Stoppani, 1979; Docampo and Moreno, 1984), with a concomitant increase in oxidative stress within the parasite. The drug was thought to be activated by one electron reduction catalyzed by type II nitroreductase (NTR-II), producing nitrofurane radical, that in aerobic conditions gives rise to a futile cycle that results in regeneration of the nitro group and production of O₂ ^•− which dismutate to H₂O₂ and O₂ catalyzed by superoxide dismutases (SODs) in a diffusion-controlled reaction (McCord and Fridovich, 1969). This postulation is based on the observation that after NFX treatment, parasite extracts showed an increased oxygen consumption and production of O₂ ^•− as well as in the fact that Trypanosoma brucei strains deficient in cytosolic iron SOD (TbSODB1) showed more sensitivity toward NFX (Roberto Docampo and Stoppani, 1979; Prathalingham Radhika et al., 2007). Moreover, it was reported that treatment with NFX leads to depletion of the principal low molecular weight thiols of the parasite, although it was discussed that this could be explained by redox cycling of the drug or by conjugation of thiols with drug metabolites (Maya et al., 1997).

The second mechanism relies on the activation of NFX by a parasite NADH-dependent mitochondrial type I nitroreductase (NTR-I), insensitive to O₂, that has FMN as cofactor and catalyzes sequential 2-electron reduction steps to yield unsaturated open chain nitriles that have trypanocidal activities with no significant generation of oxidative stress (Boiani et al., 2010; Hall et al., 2011). Trypanosoma cruzi-resistant strains have only one copy of TcNTR-I gene, and genetic modification of parasites to give knockout or overexpressers resulted in resistant or hypersensitive parasites to NFX, respectively. This shows a connection between NTR-I activity and nifurtimox toxicity (Wilkinson et al., 2008; Cerecetto and González, 2011).

Taking into consideration the abovementioned results related to the peroxiredoxins of trypanosomatids and its different functions within the cell, we evaluated the ability of T. cruzi mitochondrial peroxiredoxin to act as holdase and studied the relative weight of both activities (peroxidase and holdase) toward macrophage and NFX-derived toxicity. The results presented herein contribute to the understanding of the role of T. cruzi mitochondrial peroxiredoxin in response to different stress, gaining insights into the redox and holdase function of this peroxidase.

The importance of enzymes of the antioxidant network for parasite protection toward exogenous or endogenously derived reactive oxygen and nitrogen species has been previously documented (Piacenza et al., 2007; Piacenza et al., 2013), and in particular, the role of peroxiredoxins in the protection of the parasite against macrophage-derived oxidants has been established (Piacenza et al., 2008; Alvarez et al., 2011; Piñeyro et al., 2011). However, little is known regarding the possibility that, as in other organisms, T. cruzi peroxiredoxins have other roles in dealing with different kinds of stress. Herein, we showed the involvement of MPX in protecting parasites from NFX toxicity. The observations that increased levels of MPX help parasites to overcome the toxicity of NFX ( Figures 1A and 2A ) and that incubation in these conditions causes overexpression of MPX in wild-type parasites ( Figure 3A ) but not that of the cytosolic CPX ( Figure 3C ) clearly indicate that MPX is involved in the parasite protective response toward the drug. Overexpression of MPX was also observed in BZ-treated parasites ( Supplementary Figure S4 ), which may be explained by the fact that BZ metabolism relies on the same mitochondrial nitro reductase NTR-I (as for NFX activation), which means that the mitochondria is the first organelle affected by the action of these drugs (although each drug may have different targets). However, other drugs like ketoconazole and azasterol (McCabe et al., 1984; Goad et al., 1989; Urbina et al., 1995; de Souza and Rodrigues, 2009) that inhibit T. cruzi growth by affecting sterol synthesis did not affect the MPX content in parasites ( Supplementary Figure S4 ). This result points to the conclusion that MPX overexpression is a specific protective response that occurs when the mitochondria is faced with stress.

Since different hypotheses were built in order to explain NFX toxicity, we first assessed the possibility that the ability of MPX to detoxify peroxides was the basis for the observed protection. Nonetheless, we did not observe any evidence of a significant increase in reactive oxygen species production and/or mitochondrial superoxide generation in wild-type parasites treated with NFX ( Figures 1B , 2B ), so we could not conclude that peroxidase activity explains the protection conferred by MPX toward NFX in this experimental condition. These results are in agreement with previous reports that showed that NFX and BZ are activated by mitochondrial NTR-I in a sequential 2-electron reduction steps that does not involve the generation of reactive oxygen species (Wilkinson et al., 2008; Cerecetto and González, 2011).

The protective role of L. infantum mitochondrial peroxiredoxin (LimTXNPx) during thermal stress and also during macrophage infections was previously proposed. This protective role was fulfilled by the expression of a peroxidase inactive version of the enzyme (Castro et al., 2011). Taking this result into consideration and knowing the ability of T. cruzi CPX to prevent protein aggregation in vitro (Piñeyro et al., 2019), we consider the possibility that the protection conferred by MPX toward NFX may involve the alternative functions of peroxiredoxins, such as the yet unexplored holdase activity.

In this sense, we evaluated the ability of recombinant MPX to aid in GFP refolding and established that in vitro this enzyme increases the rate by which unfolded GFP regains its native structure ( Figure 4A ). Knowing that holdase activity of several peroxiredoxins is proposed to be modulated by the redox state of the protein, we evaluated MPX holdase activity in its reduced and oxidized state as well as for the catalytic mutant ( Supplementary Figure S5 ). Similar holdase activity was observed in both redox states of the enzyme ( Figure 4B ), in agreement with what was previously shown for T. cruzi cytosolic CPX (Piñeyro et al., 2019). The observation that recombinant MPX showed holdase activity ( Figure 4 ) prompted us to analyze whether MPX is able in vivo to acquire high molecular weight (HMW) oligomers that are related to holdase activity in other peroxiredoxins, including T. cruzi CPX (Wood et al., 2002; Piñeyro et al., 2019). We found that MPX can acquire this oligomeric structure in the parasite, and interestingly, HMW aggregates are significantly increased in NFX-treated parasites ( Figure 5A ). These same MPX aggregates were observed in parasites under thermal stress, which suggest a connection between NFX and temperature challenge ( Figure 5A ).

After establishing that MPX has holdase activity in vitro and that in parasites it can acquire HMW oligomers that increase its abundance in response to NFX, we aimed to study if this is also accompanied by an increase in holdase activity. For this purpose, we measured holdase activity in the HMW protein extracts (>100 kDa) of parasites not treated and treated with NFX. For WT parasites in which an increase in MPX content was observed following NFX treatment, a significant increase in holdase activity with respect to untreated parasites was observed ( Figure 5B ). This result first establishes MPX holdase activity in vivo and, second, could indicate that parasites increase and activate anti-aggregation systems in response to NFX, suggesting that the drug mitochondrial metabolism may generate a rise in damaged and/or unfolded proteins. Importantly, the MPX-overexpressing parasites with enhanced MPX content did not show an increase in holdase activity after NFX treatment suggesting that they can deal with the NFX unfolded protein damage response without the need to further increase MPX expression ( Figure 5C ).We have yet to establish the oligomerization state for MPX and the minimum level of MPX that allows the oligomerization needed to function as a holdase, considering that peroxiredoxins of trypanosomatids seem to have holdase activity in its decameric form (Teixeira et al., 2015; Piñeyro et al., 2019).

To evaluate if NFX causes the unfolding or misfolding of proteins tagging them for degradation, we search for protein ubiquitination and increase in protease activity after NFX treatment of WT parasites. We found an increase in protein ubiquitination and enhancement in proteosome-related protease activity following NFX treatment ( Figures 6A, B ), which may link NFX toxicity with the increase of MPX expression in WT parasites and the need to gain holdase activity to counteract the toxic effects. The unfolding of proteins could be a consequence of the production of reactive nitriles generated by NFX metabolism by parasite mitochondrial NTR-I, an enzyme that is not found in the vertebrate host (Wilkinson et al., 2008; Cerecetto and González, 2011). This hypothesis is in agreement with the observation that MPX expression increases in parasites grown under thermal stress as previously reported (Pérez-Morales et al., 2012) and as also observed in this work ( Figures 3B , 5A ). Trypanosoma cruzi life cycle entails multiple changes in environmental conditions such as temperature, pH, and oxidative stress. Parasites have to adapt to these varying environments in order to survive and be able to establish the infection. Besides the antioxidant enzyme system described by our group and others (Piacenza et al., 2008; Piñeyro et al., 2008; Piacenza et al., 2009b; Martínez et al., 2019), the parasite has evolved other mechanism to deal with stress such as molecular chaperones, which highlights the importance of maintaining protein homeostasis (Shonhai et al., 2011). The observations reported herein could indicate that NFX alters protein homeostasis in T. cruzi causing damage to the mitochondria and that the increase in MPX holdase activity could be part of the response to protein unfolding. The fact that thermal stress, known to require a chaperone response (Sanchez and Lindquist, 1990; Hübel et al., 1997; Jäättelä, 1999), induces MPX overexpression as previously reported (Pérez-Morales et al., 2012) and shown herein supports this hypothesis. Further work is required to characterize the alteration in proteostasis caused by NFX and if this is linked to the toxicity produced by the drug.

Taking into consideration the MPX holdase function described herein, it was considered whether this activity may be of relevance in the context of macrophage infection in which reactive oxygen and nitrogen species are being generated (Alvarez et al., 2011).

In an attempt to disclose the relevance of both MPX activities (peroxidase vs. holdase) and knowing that T. cruzi knockout parasites are extremely difficult to obtain, we evaluate the ability of CA as a possible inhibitor of peroxidase but not holdase activity, since it reacts with the peroxidatic cysteine that is not necessary for holdase activity (Teixeira et al., 2015; Piñeyro et al., 2019). First, we assessed if recombinant MPX reacts with CA which would generate the reported adducts inhibiting peroxidase but not holdase activity. MPX-CA treatment generates the non-reducible protein adducts and inhibits peroxidase activity, but this treatment does not affect MPX holdase activity ( Figure 7 ).

Previous studies from our group have shown the ability of MPX to cope with exogenous or endogenously produced H₂O₂ and peroxynitrite (Piacenza et al., 2008), but until now, the potential protection conferred by MPX in macrophage infections is not studied. First, it was proven that MPX contents are important for parasite survival on both naive (H₂O₂ production) and immune-stimulated macrophages for peroxynitrite generation ( Figure 8A ). This result is in agreement with MPX peroxidase activity, since inside the phagosome the parasite is exposed to peroxynitrite and H₂O₂ that can diffuse inside the internalized parasite reaching the parasite mitochondria (Piacenza et al., 2008; Alvarez et al., 2011), or alternatively, it can be generated inside the parasite by means of the diffusion of its precursors nitric oxide and superoxide radicals (Piacenza et al., 2009a; Martínez et al., 2019). Then, in order to assess the possibility to inhibit MPX peroxidase activity in the parasite, and whether this inhibition may influence T. cruzi ability to establish the infection, we first evaluated if CA generates MPX adducts in vivo. As shown in Figure 8B , CA also generates non-reducible covalent adducts of MPX in vivo similar to the effect observed in the recombinant protein ( Figure 7A ), indicating that CA is able to act in vivo. Unfortunately, CA acts in both T. cruzi peroxiredoxins, and thus, the unique role of MPX in macrophage infections is not disclosed ( Figure 8C ). Nevertheless, the role of the peroxidase and/or holdase activities of both peroxiredoxins can be addressed using this inhibitor. We found that CA-treated parasites have a diminished infection index ( Figure 8D ), showing the importance of peroxidase activity to overcome macrophage-derived oxidants in divergence to what was previously observed for Lm mitochondrial peroxiredoxin (Castro et al., 2011). This result does not discard the fact that MPX holdase activity could be relevant in the infection process. Ongoing experiments are focused on addressing the relevance of MPX in the context of NFX treatment during macrophage infections.

With the observation that MPX is overexpressed in wild-type parasites when exposed to NFX and BZ ( Figure 3A and Supplementary Figure S4 ) and knowing that the mitochondrial metabolism of these drugs gives rise to different toxic and reactive intermediates (Cerecetto and González, 2011; Hall and Wilkinson, 2012; Trochine et al., 2014), we propose that the increase in MPX expression observed may be caused by an enhancement in the modified/damaged macromolecules at the mitochondrial compartment. To further understand if MPX is a direct target for metabolites derived from NFX or if its overexpression is induced by other effects caused by the drug, further experiments need to be performed including the characterization of the adducts observed in Figure 3A . It was previously shown that there is a distinct MPX content among natural occurring strains and that it has a positive correlation with parasite virulence (Piacenza et al., 2009b). Taking into account that MPX overexpression protects parasites from NFX toxic effect ( Figure 1 ), we proposed that the MPX content may be one of the factors that contributes to the differences in NFX sensitivity reported for different T. cruzi strains (Murta et al., 1998).

In conclusion, our results demonstrate that MPX is able to function not only as an efficient peroxidase but also as a holdase, preventing protein aggregation. We observed that peroxidase function is crucial for macrophage infection, but MPX holdase activity may be relevant in other conditions, such as thermal or drug-related stress. We have yet to unravel how these two functions are modulated and which conditions could favor holdase activity. There are many factors proposed to impact on holdase activity of peroxiredoxins, such as redox state, ion concentration, and thermal activation (Wood et al., 2002; Morais et al., 2017; Teixeira et al., 2019). Our results point to the fact that holdase activity of MPX is not related to its redox state. However, given that we observed overexpression and high molecular weight aggregates after thermal stress, it would be interesting to deepen our understanding of the effect of temperature on MPX function during the infection process.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

GS: cell culture, infection experiments, enzyme activities, flow cytometry and HPLC experiments, data collection, analysis, interpretation, and writing of the article. DE: flow cytometry and HPLC experiments. RR: data interpretation and critical revision of the article. LP: conception and design of the work, data interpretation, and writing of the article. All authors contributed to the article and approved the submitted version.

GS and DE were funded by fellowships from the Comisión Académica de Posgrado, Universidad de la República and Agencia Nacional de Investigación e Innovación, respectively. This work was supported by grants from Universidad de la República: Espacio Interdisciplinario (to RR), Comisión Sectorial de Investigación Científica, I+D-2017 (to LP and GS), I+D-2020 (to LP, DE and GS), and Grupos 2018 (to RR). Additional support was obtained from Programa de Desarrollo de Ciencias Básicas (PEDECIBA, Uruguay) and the Uruguayan National System of Researchers (SNI) from ANII.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.