Combination With Tomatidine Improves the Potency of Posaconazole Against Trypanosoma cruzi

Azoles such as posaconazole (Posa) are highly potent against Trypanosoma cruzi. However, when tested in chronic Chagas disease patients, a high rate of relapse after Posa treatment was observed. It appears that inhibition of T. cruzi cytochrome CYP51, the target of azoles, does not deliver sterile cure in monotherapy. Looking for suitable combination partners of azoles, we have selected a set of inhibitors of sterol and sphingolipid biosynthetic enzymes. A small-scale phenotypic screening was conducted in vitro against the proliferative forms of T. cruzi, extracellular epimastigotes and intracellular amastigotes. Against the intracellular, clinically relevant forms, four out of 15 tested compounds presented higher or equal activity as benznidazole (Bz), with EC50 values ≤2.2 μM. Ro48-8071, an inhibitor of lanosterol synthase (ERG7), and the steroidal alkaloid tomatidine (TH), an inhibitor of C-24 sterol methyltransferase (ERG6), exhibited the highest potency and selectivity indices (SI = 12 and 115, respectively). Both were directed to combinatory assays using fixed-ratio protocols with Posa, Bz, and fexinidazole. The combination of TH with Posa displayed a synergistic profile against amastigotes, with a mean ΣFICI value of 0.2. In vivo assays using an acute mouse model of T. cruzi infection demonstrated lack of antiparasitic activity of TH alone in doses ranging from 0.5 to 5 mg/kg. As observed in vitro, the best combo proportion in vivo was the ratio 3 TH:1 Posa. The combination of Posa at 1.25 mpk plus TH at 3.75 mpk displayed suppression of peak parasitemia of 80% and a survival rate of 60% in the acute infection model, as compared to 20% survival for Posa at 1.25 mpk alone and 40% for Posa at 10 mpk alone. These initial results indicate a potential for the combination of posaconazole with tomatidine against T. cruzi.


INTRODUCTION
Chagas disease (CD), a vector-borne anthropozoonosis endemic in the American continent, is caused by the protozoan parasite Trypanosoma cruzi (Chagas, 1909). The triatomine vector of CD is spread from the southern United States to the south of Argentina. Due to increasing global migration, CD has spread to other continents through a diversity of other transmission routes such as blood transfusion, organ transplantation, and mother-to-child (Gascon et al., 2010;Peŕez-Molina and Molina, 2018). Also, oral transmission due to beverages contaminated with the feces or with infected triatomines currently represents a serious challenge in many endemic areas such as Brazil (Coura et al., 2014;Dias, 2017). This neglected disease presents a short acute phase with patent parasitemia, which is usually asymptomatic or oligosymptomatic with "flu-like" symptoms (Prata, 2001;Rassi et al., 2010). After six to nine weeks, parasite proliferation is controlled due to a competent immune response, and infected individuals enter a second stage, the chronic phase, with most of them remaining in an indeterminate form. However, after years or even decades, about 30% of the patients in the chronic phase develop progressive cardiac or gastrointestinal injuries (Ribeiro et al., 2012;Malik et al., 2015).
The front-line drugs for CD are two nitroderivatives, benznidazole (Bz) and nifurtimox. Both are far from ideal, with the occurrence of naturally resistant strains, lack of efficacy in the later chronic phase, and severe side effects that led to 10-30% therapy withdrawals (Molina et al., 2015). These limitations highlight an urgent need for novel, potent, and safer drugs for CD, and many strategies have been followed, including drug repurposing and drug combinations (Coura, 2009;Miranda and Saye, 2019).
Drug combination may tackle more than one target simultaneously, allowing reduced doses, costs, time of drug administration, and reducing the risk of parasite drug resistance, providing increased efficacy and selectivity (Sun et al., 2016). These approaches have been largely explored in experimental models of Chagas disease Diniz et al., 2013) as well as in clinical trials with chronic chagasic patients (Morillo et al., 2017).
A possible explanation for the disappointing outcome of the clinical trials with azole-type CYP51 inhibitors is, that these molecules fail to kill every single parasite, i.e., they have high EC 99 values. This notion is supported by in vivo (Francisco et al., 2015) and in vitro Cal et al., 2016;Fesser et al., 2020) models of pharmacodynamics. Nevertheless, azole-type CYP51 inhibitors have nanomolar EC 50 values against T. cruzi and a high selectivity index, and they are well tolerated by the treated patients (Morillo et al., 2017). As a strategy to overcome the limitation of CYP51 inhibitors, we have proposed to combine them with a partner drug that either acts in the same pathway, sterol biosynthesis, or that inhibits a functionally linked pathway, sphingolipid synthesis (Fügi et al., 2015). Both rationales are supported by genetic interaction data from yeast (Eisenkolb et al., 2002;Guan et al., 2009;Fügi et al., 2015). Sphingolipids are a major class of lipids and ubiquitous constituents of eukaryotic membranes, playing also a role as bioactive signaling molecules involved in the regulation of cell growth, differentiation, senescence, and death (Pruett et al., 2008), as well as in virulence and survival of pathogens upon interaction with the host, including T. cruzi (Goldston et al., 2012;Guan and Mäser, 2017).
In the present work, we have assembled a panel of fifteen drugs and experimental compounds that interfere either with sterol synthesis or with sphingolipid metabolism. The compounds, their target enzymes, and their medical use (if any) are described in Table  1. All compounds were phenotypically assayed against the multiplicative forms of T. cruzi. Identified hit compounds were further combined with Posa and reference drugs in in vitro and in vivo models of parasite experimental infection.

Activity on T. cruzi Epimastigotes
Epimastigotes (10 7 parasites/ml) were incubated for 72 h at 28°C with serially diluted compound concentrations (eleven 1:3 dilutions) in supplemented LIT medium. Parasite viability and motility were evaluated by direct observation by light microscopy and fluorometric assays performed with resazurin (12.5 mg resazurin dissolved in 100 ml distilled water). After 2-4 h of incubation with resazurin solution, plates were read in Spectramax Gemini XS microplate fluorometer (Molecular Devices Cooperation, USA) using wavelengths of 536 nm (excitation) and 588 nm (emission) (Räz et al., 1997). Growth was expressed as percentage of the values of solvent-treated controls. The graphics program Softmax Pro (Molecular Devices) was used to construct dose-response curves and calculate EC 50 (half maximal inhibitory concentration) values. Bz was used as reference drug.

Activity on T. cruzi Intracellular Amastigotes
Rat skeletal myoblasts (L6 cell lines) were seeded in 96-well plates (10 4 cells/well) in 100 ml RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) and 2 mM L-glutamine. The medium was removed after 24 h incubation and replaced by 100 ml of fresh medium containing 10 5 LacZ trypomastigotes (Tulahuen DTU VI). After 48 h, the medium was again removed and replaced with or without a serial of compound concentrations (eleven three-fold dilution steps). After 96 h of compound exposure, CPRG/Nonidet (50 ml) substrate was added and the reading performed after 2-6 h at 540 nm in Versamax microplate reader (Molecular Devices Cooperation, USA). Growth was expressed as percentage of the values of solvent-treated controls. The graphics program Softmax Pro (Molecular Devices) was used to construct dose-response curves and calculate EC 50 values. Bz was used as reference drug.

Cytotoxicity Against L6 Cells
Rat skeletal myoblasts (L6 cell lines) were seeded in 96-well plates (4 × 10 4 cells/well) in 100 ml RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) and 2 mM L-glutamine. After 24 h incubation the medium was replaced with 100 ml of fresh medium with or without a serial dilution of compound concentrations (eleven three-fold dilution steps).
After 72 h of compound exposure, fluorescent dye resazurin (10 ml, 12.5 mg resazurin dissolved in 100 ml water) was added for 2 h and the readings performed at Spectramax Gemini XS microplate fluorometer (Molecular Devices Cooperation, USA), with excitation wavelength of 536 nm and an emission wavelength of 588 nm. Growth was expressed as percentage of the values of solvent-treated controls. The graphics program Softmax Pro (Molecular Devices) was used to construct doseresponse curves and calculate EC 50 values. Podophyllotoxin (a microtubule destabilizing agent) was used as positive control.

Animals
Male Swiss Webster mice (18-23 g) were obtained from the Instituto de Ciencia e Tecnologia em Biomodelos (ICTB/ Fiocruz) (Rio de Janeiro, Rio de Janeiro, Brazil). Five mice were housed per cage, kept in a conventional room at 20-24°C under 12 h/12 h light/dark cycle. Sterilized water and chow were provided ad libitum. The animals were acclimated for 7 days before being used in the different assays. All procedures were done following Biosafety Guidelines in compliance with the Fiocruz and all animal procedure approved by the Committee of Ethics for the Use of Animals (CEUA L-38/17).

Mouse Infection and Efficacy Studies
Male mice (n = 5 per group) were infected i.p. with 10 4 bloodstream trypomastigotes of T. cruzi (Y strain). Only mice with positive parasitemia at day 5 post infection (dpi) were included in the studies. T. cruzi-infected mice were treated (p.o.) for ten consecutive days, from 5 to 14 dpi, with Posa (10 and 1.25 mg/kg body weight (mpk) corresponding to optimal and suboptimal doses of Posa for parasitemia suppression, respectively), TH (5-0.5 mpk) and combos of Posa plus TH, using the suboptimal dose of Posa (1.25 mpk) in different proportions, nearby those with best in vitro outcomes as follows: Posa 1.25 mpk + TH 5 mpk (ratio 1:4), Posa 1.25 mpk + TH 3.75 mpk (ratio 1:3) and Posa 1.25 mpk + TH 0.5 mpk (ratio 2.5:1)). Uninfected and T. cruzi-infected mice treated only with vehicle (aqueous solution of 0.5% carboxymethylcellulose) were agematched and housed under identical conditions and used as controls (Simões-Silva et al., 2017). All compound formulations were freshly prepared before every administration.

Parasitemia, Mortality Rates, and Endpoint
All animals were individually checked for circulating blood parasitemia by counting the number of parasites in 5 µl of blood taken from the tail vein and investigated under the microscope (Brener, 1962). Parasitemia was checked till 30 dpi, while mortality was checked daily up to 30 days after the administration of the last dose. Mortality was given as percentage of cumulative mortality (CM) (Simões-Silva et al., 2017).

Statistical Analysis
All experiments were performed in triplicate in three independent experimental sets. The citotoxicity and antitrypanosomal activity were analyzed by ANOVA/Dunnet test using GraphPad Prism 5.01 software. P values of 0.05 or lower were assumed as significant.

RESULTS
The in vitro activity of the fifteen compounds ( Table 1) was assessed on the multiplicative forms of T. cruzi: epimastigotes (strain STIB 980) and intracellular amastigotes (Tulahuen C2C4 strain expressing the b-galactosidase gene LacZ) ( Table 2). In parallel, the cytotoxicity of the compounds was evaluated on mammalian host cells ( Table 2). Against T. cuzi epimastigotes, two compounds (FTY720 and Ro48-8071) were promising, displaying similar potency as Bz (7.55, 11.6 and 13.9 µM, respectively; both of which were not significantly different to Bz, with p values >0.05 in comparison to Bz), and showing about 2.5-8-fold lower EC 90 values than the reference drug (both with p < 0.05) ( Table 2). Against the intracellular amastigotes, four compounds (FTY720, RO48-8071, tomatidine hydrochloride (TH) and TMP-153) displayed EC 50 values ≤1 µM, lower than Bz (2.2 µM) (the four compounds presenting p <0.05 in comparison to Bz) ( Table 2). Most of the tested compounds showed quite relevant toxicity towards mammalian host cells, leading to low selectivity indices (SIs), except for Ro48-8071 and TH, which presented promising SIs of 12 and 115, respectively ( Table 2).
Based on their high activity against intracellular amastigote T. cruzi and good selectivity towards the mammalian host cells, Ro48-8071 and tomatidine were moved to in vitro combination assays with the reference drug for CD (Bz) and two others that displayed efficacy in in vitro and in vivo assays of T. cruzi experimental infection: the imidazole CYP51 inhibitor Posa and the nitroimidazole Fexinidazole (Fexi) ( Table 3). Of the six combinations tested, only that between TH and Posa had a synergistic profile with mean SFICI values below 0.5, based on their EC 50 (Table 3, Figure 1). These results encouraged us to follow up with in vivo studies. Posa or TH did not show any signs of toxicity when administered to female mice p.o. up to 200 mpk (data not shown).
Before moving to co-administration schemes, TH and Posa alone were administered in mouse models of acute T. cruzi infection (Figure 2). Posa at 10 mpk suppressed parasitemia on the peak (8 dpi), providing 40% survival of mice, while all vehicletreated mice died until the endpoint (Figure 2). A suboptimal dose of Posa (1.25 mpk) decreased the parasitemia peak (about 80%), but only provided a mild protection against mortality (20% of animal survival) (Figure 2). On the other hand, all tested doses of TH (up to 5 mpk) alone resulted in a maximum reduction of only 28% in blood parasitemia on the peak and were unable to protect against the mortality induced by the infection since all animals died by 20 dpi (Figure 2).
The co-administration of TH (variable doses from 0.5 to 5 mpk) plus Posa at the suboptimal dose of 1.25 mpk led to a parasitemia drop ranging from 60 to 80%, and cumulative death ranging from 40 to 100% (Figure 3). The best effect as evaluated by the concomitant reduction in peak parasitemia (80%) and increased animal survival (60%) was achieved with the combo Posa 1.25 mpk + TH 3.75 mpk (ratio 1:3) (Figure 3), which corroborated the best in vitro ratio of combination (Table 3).

DISCUSSION
Drug repurposing and drug combination are pre-clinical strategies used in experimental pharmacology to tackle many diseases, TABLE 2 | Activity (EC 50 and EC 90 , µM, n = 3) and selectivity of lipid biosynthesis inhibitors on Trypanosoma cruzi epimastigotes (STIB 980 strain) and intracellular amastigotes (Tulahuen strain in rat L6 myoblasts).

Epimastigotes (mean ± SD) (µM)
Amastigotes ( reaching quite effective results when applied to clinical trials and off-label use (Sbaraglini et al., 2016). Successful examples from neglected tropical diseases include the repositioning of nifurtimox administered in combination with eflornithine for Human African trypanosomiasis (Priotto et al., 2007), the association of miltefosine and paromomycin, and sodium stibogluconate plus paromomycin, for visceral leishmaniasis (Atia et al., 2015;Rahman et al., 2017;Alves et al., 2018). Regarding Chagas disease, fungicidal inhibitors of CYP51 enzymes have been assayed in clinical trials (e.g., Posa and E1224 in association with Bz), but unfortunately had high rates of therapeutic failure (Morillo et al., 2017;Torrico et al., 2018). On the other hand, the BENEFIT trial demonstrated that, although effective to reduce parasite load in chronic chagasic patients, Bz did not impair the progression of cardiac damages, reinforcing the need to search for new therapeutic alternatives for CD (Rassi et al., 2017). Inhibitors of sterol biosynthetic enzymes and sphingolipid metabolism and signaling had been proposed as combination partners for Posa (Fügi et al., 2015). Selected inhibitors ( Table 1) were phenotypically assessed against T. cruzi. FTY720 and Ro48-8071 were as active as Bz against epimastigotes. Against the therapeutically relevant intracellular forms, four compounds had similar or even higher potency than Bz: FTY720, Ro48-8071, tomatidine hydrochloride and TMP-153. The different EC 50 values of the studied compounds (including the reference drug) against epimastigotes and amastigotes can be explained by the distinct cellular metabolism of these proliferative forms that face different environments and hosts. Although we cannot exclude variabilities in drug susceptibility among the different strains (Zingales et al., 2014;Zingales, 2018), our data confirm the importance of using the intracellular amastigote form for drug discovery (Romanha et al., 2010;de Castro et al., 2011).
Based on their promising activity and selectivity against T. cruzi amastigotes, Ro48-8071 and TH were selected for in vitro combination testing using fixed-ratio proportions as reported (Simões-Silva et al., 2016). Regarding the choice of partner drugs, Bz as one of the standard drugs for CD was an obvious candidate (Coura, 2009). Posa and Fexi are very potent anti-T. cruzi agents in vitro and in vivo (Urbina, 2015) that were moved to clinical trials for CD (Bahia et al., 2012;Morillo et al., 2017;DNDi Portfolio, 2020). None of the combos made with Ro48-8071 showed synergistic activity and were therefore not further investigated. Combos of Bz and Fexi with TH were also additive. However, the association of TH plus Posa was essentially synergistic, displaying ∑FICI = 0.2.
A synergistic interaction for TH has already been reported with aminoglycoside antibiotics, being more effective in inhibiting colony growth of S. aureus clinical isolates as compared to  standard monotherapies (Mitchell et al., 2011;Mitchell et al., 2012;Soltani et al., 2017). Also, the combination of TH with fluconazole exhibited synergistic interaction against a C. albicans azoleresistant strain (Robbins et al., 2015), thus confirming the potential of TH for drug combination protocols. Based on these in vitro results, Posa + TH was moved to in vivo assays using a well-established mouse model of acute T. cruzi infection (Romanha et al., 2010). TH alone did not present antiparasitic activity in vivo. However, it is important to note that the lack of in vivo efficacy may be due to the poor solubility of TH. Previous studies reported that, as TH is a highly hydrophobic sterol-like molecule, many vehicles including DMSO, ethanol or cyclodextrin failed to demonstrate efficacy in in vivo models of C. albicans infection, except for the use of a nanoparticle-based formulation that allowed successful reduction of fungal burden in infected mice (Dorsaz et al., 2017). Thus, the exploration of other formulations for TH is desirable to better assess its potential against T. cruzi in vivo. When the combos were assayed, the best results in terms of reduction of parasitemia and mortality were obtained with the proportion of Posa 1.25 mpk + TH 3.75 mpk, which correlates to the most synergistic combo in vitro (drug ratio 1:3). The combination of Posa at 1.25 mpk plus TH at 3.75 mpk displayed a survival rate of 60% in the acute infection model as compared to 20% for Posa at 1.25 mpk alone, and 40% for Posa at 10 mpk alone.
Thus, our finding that only the combo of Posa plus TH gave a synergistic profile in vitro was further corroborated by our in vivo assays demonstrating a reduction in parasite load and animal death rates. These results possibly are due to the simultaneous action on enzymes (lanosterol 14-a demethylase and sterol 24-C-methyltransferase) to the sterol biosynthetic route, impacting the fitness profile and metabolism of the intracellular, clinically relevant form of T. cruzi.
TH is a natural compound, originally found in unripe tomatoes, that has a wide array of bioactivities including antioxidant, anticarcinogenic and antimicrobial effects (Friedman, 2013). TH exerts antifungal and antitrypanosomatid effects by inhibition of C-24 sterol methyltransferase (Medina et al., 2015;Dorsaz et al., 2017). The finding that TH synergistically interacts with Posa encourages further studies with this class of compound and reinforces the potential of drug repurposing and combination protocols. These approaches represent reduced cost and time in the search for better treatments for CD, which is clearly relevant considering the shortage of resources in benefit of the poor population around the world affected by neglected tropical diseases such as Chagas disease. Although Posa at 10 mpk and the combo Posa 1.25 mpk + FIGURE 2 | Parasitemia and cumulative mortality of mice infected with T. cruzi (Y strain) treated with vehicle alone, posaconazole (Posa 10 and 1.25 mpk), or tomatidine hydrochloride (TH 5, 3.75 and 0.5 mpk) administrated for ten consecutive days (dpi 5 to 14). TH 3.75 mpk suppressed/highly reduced the parasitemia, neither therapeutic scheme was able to reach 100% animal survival and induce sterile cure. Thus, further studies will need to address the efficacy against dormant forms of T. cruzi as recent data suggest the existence of an adaptive difference between parasite strains to generate dormant cells, and that homologous recombination in T. cruzi may be important for dormancy stages (Sańchez-Valdeź and Padilla, 2019; Resende et al., 2020).

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
All procedures followed the guidelines in compliance with the Fiocruz Committee of Ethics for the Use of Animals (CEUA L-38/17).

AUTHOR CONTRIBUTIONS
MR-H performed the in vitro and in vivo studies, data analysis, and drafted the manuscript. PM, MK, and MS (corresponding author) obtained the funding, conceived the study, performed data analysis, and drafted the manuscript. XG helped in drafting the manuscript and in study design. GO, AD, LF, and RP contributed to the in vivo studies. AF, MC, and RR contributed to the in vitro studies. All authors contributed to the article and approved the submitted version.