Chagas disease or American trypanosomiasis is a zoonosis caused by the protozoan Trypanosoma cruzi, discovered by Dr. Carlos Chagas in 1909 (Chagas, 1909). Although parasite transmission may occur through diverse pathways, the most common are vectorial, oral and congenital transmissions (Coura, 2007). The WHO estimates that eight million people are chronically infected with T. cruzi in the world, mostly in Latin America, where the disease is endemic (WHO, 2012). Notwithstanding, several cases have been reported in non-endemic areas such as European countries, North America, Japan and Australia as result of migratory processes (Bern and Montgomery, 2009).

Chagas disease occurs in two phases: acute and chronic. Acute Chagas disease is characterized by high parasitemia, being recognized in few patients due to absence or non-specificity of clinical symptoms (Bern, 2015). Chronic Chagas disease begins about 2 or 3 months after infection and has been divided in the determinate and indeterminate forms. About 70% of patients present positive serology for T. cruzi without detectable anatomic-physiological changes, being classified in the indeterminate form (Coura, 2007). The other amount may develop the cardiac, digestive or cardiodigestive forms (Rassi et al., 2012) along the years (determinate form). Chronic cardiac form is the most expressive manifestation of Chagas disease because of its frequency and severity (Rassi et al., 2009).

Chagas' disease chemotherapy is performed with benznidazole or nifurtimox, drugs that not only are genotoxic but also present severe side effects that might lead to treatment interruption (Rassi et al., 2012). Successful chemotherapy depends on several factors: the stage of disease, patient's age, and biochemical characteristics of the T. cruzi strain. A critical feature related to failure in chemotherapy relies on resistance mechanisms, observed in naïve or in selected strains during the treatment. Resistance to benznidazole and nifurtimox have been reported in a study evaluating 26 strains (Filardi and Brener, 1987). Additionally, the treatment with both drugs can induce resistance in distinct T. cruzi strains in vitro and in vivo, configuring another obstacle to the chemotherapy (Nirde et al., 1995; Dos Santos et al., 2008). Broad-spectrum resistance, named Multidrug Resistance (MDR) phenotype, is determined by cross-resistance to structurally and functionally unrelated drugs, and could be a limiting factor to the treatment of parasitosis. Among the factors related to the onset of the MDR phenotype, the best characterized is the overexpression of membrane transporters, mostly from the ATP-Binding Cassette (ABC) protein superfamily (Gottesman, 2002).

ABC proteins share the well-conserved ABC module and are divided into 8 subfamilies in eukaryotes (ABCA to ABCI) (Dassa, 2011). The “full size” transporter consists of two transmembrane domains (TMD) involved with the attachment and transport of the substrate, and two ABC modules, also known as nucleotide binding domains (NBDs), which bind and hydrolyzes ATP. The most widely studied ABC superfamily protein is ABCB1 (subfamily B, member 1), also known as P-glycoprotein (Pgp), and ABCC1 (subfamily C, member 1), also known as Multidrug Resistance Protein-1 (MRP1). Those proteins were the first cloned human ABC transporters and characterized by their ability to confer MDR phenotype to tumor cells (Juliano and Ling, 1976; Cole et al., 1992). Despite being described in tumor cells, ABC transporters are expressed in prokaryotes as well as eukaryotes, being associated to detoxification (Descoteaux et al., 1995; Raj et al., 2009; Boncoeur et al., 2012), efflux of inflammatory molecules (Leslie et al., 2005), and lipid translocation (Araujo-Santos et al., 2005; Aye et al., 2009).

ABCB1 transports molecules with no chemical, structural nor pharmacological relationship out the cell. In the general, substrates are hydrophobic and present aromatic rings and a positively charged nitrogen (Silva et al., 2015). ABCC1 mediates the efflux of a variety of physiological organic anions alone, in conjugation to glutathione (GSH), glucuronide or sulfate or in cotransport with GSH (Cole, 2014). Multidrug resistant cells present reduced accumulation of fluorescent dyes compared to sensitive cells and the use of MDR inhibitors reverts this phenotype. Rhodamine 123 (Rho 123) (Neyfakh, 1988) and carboxyfluorescein (CFDA) (van der Kolk et al., 1998) were introduced as good fluorescent substrates for ABCB1 and ABCCs respectively due to their chemical natures. Moreover, Verapamil (VP) (Tsuruo et al., 1981), Trifluoperazine (TFP) (Tsuruo et al., 1982), and Cyclosporin A (CsA) (Slater et al., 1986) were described as ABCB1 modulators, although CsA was able to modulate ABC proteins from different subfamilies. Several inhibitors of non-specific organic anion transporters such as probenecid and indomethacin are capable of inhibiting ABCCs, ABCB1 and solute carriers. The quinoline derivative MK-571 is a potent inhibitor of ABCCs proteins, having no effect on ABCB1 (Cole, 2014).

ABC transporters have been described in parasites from the Trypanosomatidae family as well. In the Leishmania spp. genome, four ABCB subfamily genes were identified. Two genes are ABCB1 homologs, encoding typical ABCB2 and ABCB4 proteins (Sauvage et al., 2009). Parasites transfected with the ABCB2 gene showed a reduction in the accumulation of the anticancer drug 5-fluorouracil, suggesting a role in the efflux of xenobiotics (Katakura et al., 2004). ABCB4 transporter contributed in cross-resistance to the leishmanicidal drugs miltefosine and edelfosine (Perez-Victoria et al., 2001). The two other genes from Leishmania show atypical structures and unknown functions (Sauvage et al., 2009). From ABCC subfamily, ABCC3 (MRPA or PGPA) protein was the first identified in Leishmania spp. and participates in the transport of metal-thiol conjugates to vesicles (Legare et al., 2001). Other ABCC subfamily genes described so far are ABCC2 (also known as PGPB), ABCC1 (PGPC), ABCC5 (PGPD), and ABCC4 (PGPE), neither related to the MDR phenotype (Sauvage et al., 2009). The last described member of subfamily ABCC is the ABCC7 transporter, also known as PRP1 (pentamidine resistance protein-1), because of its ability to confer resistance to pentamidine (Coelho et al., 2003). In T. brucei, ABCB genes have not been identified whereas two ABCC genes were functionally described, ABCC2 and ABCC6. The overexpression of these genes suggested that they are involved in resistance to melarsoprol and suramin (Shahi et al., 2002), drugs used in the chemotherapy of sleep disease.

In the T. cruzi genome, 27 ABC genes were identified (Leprohon et al., 2006), including tcpgp1 (Dallagiovanna et al., 1994) and tcpgp2 (Dallagiovanna et al., 1996) the first to be described. Although they belong to the ABC family, their roles in drug resistance are still controversial (Dallagiovanna et al., 1996; Murta et al., 2001; Campos et al., 2013). In 2003, Peelman et al. (2003) identified an ABCA-like transporter (named ABCA3). ABCA3 is a single copy gene expressed along the parasite life cycle, except in infective trypomastigote forms. This protein was located in the plasma membrane, flagellar pocket, and intracellular vesicles, possibly involved in vesicular trafficking (Torres et al., 2004). Recently, an ABCG-like transporter, named ABCG1, was found to be overexpressed in T. cruzi strains naturally resistant to benznidazole. In addition, the transfection of this transporter in CL Brener strain increased drug resistance (Zingales et al., 2015).

Considering that, drug resistance is an obstacle to the treatment of Chagas disease, ABC proteins could be considered as new molecular targets to improve therapy efficacy. Although ABC transporters have been identified in T. cruzi since the 90s, no functional evaluation was performed under physiologic conditions. Therefore, in the present work we evaluated the activities of ABCB1-like and ABCC-like transporters in T. cruzi Y strain, since we confirmed that P-glycoprotein genes belong to the ABCC subfamily in the parasite. Results in this report provide the first description of ABCC-like activity in invertebrate and vertebrate stages of the parasite life cycle and, similar to mammalians, the T. cruzi ABCC-like protein is a thiol transporter.

In T. cruzi, the first ABC genes to be described were tcpgp1 and tcpgp2, found to be similar to ABCB1 in humans and to P-glycoprotein A gene (PGPA) in L. tarentolae. However, the predicted protein sequences showed more identity and similarity with ABCC1, which had been discovered at that time (Dallagiovanna et al., 1994, 1996). Albeit the tcpgp1 gene had only 27.9% identity with ABCB1, it was named as Pgp. Likewise, PGPA, which was described to be similar to ABCB1, was later discovered to be closer to ABCC1. Afterwards, it was demonstrated that PGPA was able to mediate thiol efflux as well (Legare et al., 2001). Following this, it was renamed as MRPA/ABCC3, being recognized as an ABCC subfamily member (Leprohon et al., 2006). As PGPA gene in Leishmania, we showed that the predicted protein sequences from tcpgp1 and tcpgp2 genes correspond to ABCC6 and ABCC1/2 putative proteins described in the T. cruzi genome, respectively. Since ABCC-like activity has not been described in the parasite, we investigated the efflux of carboxyfluorescein (CF), a known fluorescent substrate of ABCC proteins, in live parasites.

MK-571 was the most efficient in inhibiting the CF efflux, as verified by the Δ values. It is important to note that MK-571 is the only modulator having specificity for ABCC subfamily members (Cole, 2014). Probenecid and indomethacin also inhibited the CF transport. In addition, the influence of temperature in the CF efflux was evident when the inhibition index (Δ), which was higher at 37°C regardless of the modulator used, was observed.

Trypomastigote forms of the Y strain also showed ABCC-like activity, which was inhibited by indomethacin and MK-571 in subtoxic concentrations. Although being more efficient, MK-571 inhibited about 30% of trypomastigotes, whereas in epimastigotes the inhibition was of about 90% in the same conditions. Analysis of the Δ values showed that trypomastigotes presented about five-fold less ABCC-like activity than epimastigotes, suggesting that ABCC could be regulated during the life cycle of the parasite.

ATP hydrolysis is a primordial characteristic of the ABC proteins. In the assays, CF efflux was inhibited after preincubation with ATP-depleting agents, although the modulation was lower than the one observed with MK-571. Therefore, it is possible that the concentration of sodium azide or iodoacetic acid was not sufficient to completely deplete the ATP levels, allowing the transport of CF by part of the parasite population. In epimastigotes, glucose is the preferred source for energy generation in culture; however, glucose is not abundant in the gut of the invertebrate host (Bringaud et al., 2006). Moreover, trypanosomatids do not have carbohydrate reserves, being dependent on external energy sources (Maugeri et al., 2011). Consequently, removal of glucose from the assay medium significantly reduced the CF efflux, ruling out the possibility of the transport being exclusively passive. Additionally, solute carriers like OATP1B1 are able to carry organic anions through membranes in humans, though this kind of transport is not dependent of ATP directly.

In our work, modulators employed for inhibiting ABCB1-mediated transport such as VP did not present effect on the CF efflux, while TFP was discarded as an efficient inhibitor in T. cruzi. On the other hand, the broad-spectrum modulator CsA was able to inhibit the ABCC-like activity in T. cruzi as seen in mammal cells.

An important feature of ABCC subfamily members is the ability to transport endo- and xenobiotics in cotransport or in conjugation with GSH. NEM is a sulfhydryl radical alkylating agent, being able to bind covalently to molecules containing non-protein thiols, preventing the participation of these molecules in enzymatic reactions. BSO is the specific inhibitor of γ-glutamylcysteine synthetase, an enzyme of the GSH biosynthesis pathway. In mammalians, GSH is the main non-protein thiol and plays an important role in the biotransformation of molecules and defense against electrophilic compounds (Pisoschi and Pop, 2015). In trypanosomatids, the main non-protein thiol is represented by T(SH)₂, which is formed by two GSH molecules bound by a spermidine chain (Olin-Sandoval et al., 2010). Thus, the use of NEM masks any molecule containing an apparent thiol radical while BSO specifically reduces the synthesis of GSH and, thereafter, T(SH)₂. Preincubation with NEM or BSO reduced non-protein thiol levels of epimastigote forms, as observed by the reduction of the TMF MFI, the thiol-conjugated product of the reaction of CMFDA with a free thiol radical. Considering this, CMFDA dye was used to indirectly measure GSH/T(SH)₂ levels of the parasite. Results indicated that NEM was more effective in reducing non-protein thiol levels, presenting nearly 90% reduction in TMF MFI. Furthermore, the treatment with NEM inhibited the CF efflux, suggesting that part of the CF organic anion transport could be performed via cotransport with GSH or T(SH)₂. Besides that, the TMF efflux was modulated by indomethacin, inhibiting about 35% of the parasites. The lower inhibition of TMF transport compared to CF may be due to the reduced affinity of ABCC members to thiols (Slot et al., 2011). A study carried out on vesicles enriched with different ABCC proteins demonstrated that the transport of dichlorocarboxifluorescein, analogous to CMFDA, differed between ABCC members in tumor cells (Pratt et al., 2006). Likewise, it is known that human ABCC members can transport GSH even if the GSH affinity varies among them. Regardless of the affinity, it is important to focus on that T. cruzi Y strain performed thiol efflux, and this kind of transport is on par with the mechanism described for ABCC members.

In both evolutionary forms neither CsA, VP, nor TFP, all used as ABCB modulators, inhibited Rho 123 efflux, since there was no increase in Rho 123 MFI nor in the percentage of Rho 123+ parasites. Results were similar when temperature was increased from 27 to 37°C or after ATP depletion. In addition, there was no difference in the Rho 123 MFI from accumulation and efflux step (data not shown). These results, together with the predicted protein sequences, suggest that the expression of an ABCB1-like transporter in the cytoplasmic membrane is very low, or even absent.

The presence of higher ABCC-like activity in epimastigote compared to trypomastigote forms could relate to the protection from microenvironment-induced oxidative stress. Hemoglobin degradation in the hematophagous insect gut releases large amounts of heme (Graca-Souza et al., 2006). Heme catalyzes many oxidation processes in biological systems, involved in the cellular respiration, metabolism, growth, and cell differentiation, which are essential processes for survival (Ciccarelli et al., 2007; Paes et al., 2011). Although it is important for T. cruzi growth, heme is toxic due to its ability to generate reactive oxygen species (Kumar and Bandyopadhyay, 2005), and owing to its amphipathic characteristics, is capable to associate with membrane lipids thus altering cellular permeability (Schmitt et al., 1993). Under these conditions, T. cruzi needs to adjust for not only changes in the potentially reducing microenvironment, but also for the increased production of reactive oxygen species that follows heme metabolism. In the most cells, antioxidative machinery relies on GSH as electron source to reduce and inactivate reactive oxygen and nitrogen species, which is regenerated by glutathione reductase (Pisoschi and Pop, 2015). T. cruzi produces both GSH and T(SH)₂ (Krauth-Siegel and Comini, 2008). Notwithstanding, the parasite does not express glutathione reductase, making the antioxidant system exclusively reliant on T(SH)₂ regeneration (Olin-Sandoval et al., 2010). In that context, the ABCC activity could transport oxidized GSH or T(SH)₂, and also metabolites, contributing to the parasite defense system by an alternative route to T(SH)₂ regeneration. In addition, ABCC is involved in the bioavailability of various endobiotics that could contribute for parasite survival in the host, indicating the importance of the study of its physiologic substrates.

There are three ABCC subfamily genes identified in the T. cruzi genome, which may be associated to ABCC-like activity observed in this study. ABCC1/2 is recognized as a pseudogene (Leprohon et al., 2006). Therefore, this transporter is not supposed to codify a protein, although it could be transcribed into mRNA. ABCC6 presents an atypical structure, lacking the sequences corresponding to the 11th and 12th transmembrane helices of TMD and the second NBD, due to the insertion of a retrotransposon (Dallagiovanna et al., 1996). In fact, functional transporters with atypical structures are described in the literature as the case of human ABCG2 (Taylor et al., 2017), and many others in parasites, as Plasmodium falciparum ABCB3 (Sauvage et al., 2009), which can function as an homodimer. The last ABCC subfamily gene is ABCC9, whose function is unknown in T. cruzi. In humans, this protein is not involved in the drug resistance and constitutes the ATP-sensitive potassium channels (Solbach et al., 2006). Conversely, the participation of ABCC9 in the transport of thiol as observed in this study cannot be ruled out, since ABCC9 could exert different functions in the parasite.

To date, there are no studies evaluating the protein expression of neither tcpgp1 nor tcpgp2. Regarding transcript levels, few papers have addressed this question. Northern blot analysis with a tcpgp1 probe revealed that a mRNA molecule that hybridizes to tcpgp1 is constitutively expressed in Y, Tulahuen and Maracay T. cruzi strains (Dallagiovanna et al., 1994). Subsequently, tcpgp2 was discovered in T. cruzi and, again by Northern Blot analysis, its mRNA was found to be expressed in amastigote and epimastigote forms. Despite not being successful in T. cruzi, transfection of the tcpgp2 gene in L. tropica resulted in mRNA expression for tcpgp2, but this was not sufficient to confer cross-resistance to different drugs. Nonetheless, resistance to benznidazole increased in transfected parasites when their EC₅₀ values were compared (Dallagiovanna et al., 1996). In 1998, Murta et al. tested a gene probe for tcpgp in 27 T. cruzi strains sensitive or naturally resistant to benznidazole or nifurtimox. Authors observed the presence of polymorphism in pgp almost exclusively in the sensitive strains. Then, those 27 strains, as well as in vitro induced or in vivo selected benznidazole-resistant strains were analyzed and none presented different transcript levels for tcpgp1 and tcpgp2 (Murta et al., 2001). By differential gene expression, the same group did not observe involvement of ABC genes in the MDR phenotype of different T. cruzi strains (Murta et al., 2008). Campos et al. (2013) then induced resistance to benznidazole and a thiosemicarbazone derivative in the Y strain in vitro. The authors demonstrated rhodamine 123 efflux, which was inhibited by ABCB1 modulators, and increased ATPase activity in the resistant parasites. In addition, authors showed increased transcript levels of tcpgp1 and tcpgp2 in resistant Y parasites compared to the parental strain, although these genes are more related to subfamily C rather than B, as observed in our work. As ABC proteins are involved in drug transport in humans and other trypanosomatids, it is plausible to suggest that ABCC activity could also relate to chemotherapy resistance. Considering this, the ABC protein subfamilies rise as targets for studies aiming to better understand the mechanisms employed by T. cruzi to adapt to microenvironment stress. Further research is pivotal to reveal the role of ABCC transporters in the parasite physiology, during drug resistance and, ultimately, to develop new strategies for the treatment of Chagas disease.

KMC: Designed research and performed the experiments, analyzed, interpreted the results, and drafted the manuscript; RCV: Designed research, contributed to the interpretation of the results, and revised the manuscript; ES: Helped in carrying out MTT reduction assay, interpreted the results, and drafted the manuscript; LBG: Helped in initial analysis of BLAST alignment and revised the manuscript; LFL, LMP, and JOP: Designed research and revised the manuscript. All authors read and approved the final version of the manuscript.