Chagas disease (American Trypanosomiasis), a parasitic protozoal infection caused by Trypanosoma cruzi, is a neglected tropical illness of high medical relevance, that affects at least six million people, mostly in the endemic regions of Latin America including Brazil, where there are currently more than 1.1 million chronically infected individuals (WHO, 2015[35]). In recent years, the number of T. cruzi infections acquired through congenial and unscreened blood transmissions and organ transplants have been increasing in non-endemic areas worldwide, including North America and Europe, as a result of international migration (Soriano-Arandes et al., 2016[32]). An increasing number of Chagas disease cases has also been reported in both endemic and non-endemic regions due to oral transmission via consumption of food and beverages contaminated with infected triatomine bugs and their fecal matter (Coura and Viñas, 2010[18]; Jackson et al., 201[24]4; Montgomery et al., 2014[27]).

Two currently available chemotherapies for the treatment of Chagas disease are benznidazole and nifurtimox, nitroimidazole-derived drugs that are well known for their substantial toxicity and poor efficacy during the chronic phase of the disease (Bern et al., 2007[5]). Therefore, the search for safer and more reliable trypanocidal agents, such as plant-derived secondary metabolites, is urgently needed.

Aiouea trinervis Meisn. is a shrub of the Lauraceae family which grows in the “Cerrado” of Mato Grosso do Sul, Brazil. The biological properties of members from this genus have been scarcely reported and refer to the evaluation of the genotoxic and/or cytotoxic potentials of only two species, namely A. costaricensis and A. trinervis. A weak cytotoxic activity against neoplastic cell lines was reported for the essential oils from the leaves and branches of A. costaricensis (Chaverri et al., 2010[12]), while in a previous study of the roots, xylopodium and leaves of A. trinervis we reported the isolation of four butanolides (epilitsenolide C1, epilitsenolide C2, isoobtusilactone A and obtusilactone A) and evaluation of the cytotoxic activities of the first three compounds against Hep2 cells (Garcez et al., 2005[22]). The genotoxicity of obtusilactone A and isoobtusilactone A isolated from A. trinervis was also assessed by using the somatic mutation and recombination test (SMART) on Drosophila melanogaster and/or the comet assay on CHOK1 and HTC mammalian cells (Garcez et al., 2005[22]; Guterres et al., 2014[23]). The ability of isoobtusilactone A obtained from Cinnamomum kotoense to induce apoptosis in Hep G2 cells and its apoptotic mechanisms of action were described by Chen et al. (2008[13], 2012[14]). In addition, this same compound isolated from Persea americana was reported to have a nematicidal potential (Dang et al., 2010[21]). Cytotoxic and anti-HIV activities have also been described for structurally similar butanolides isolated from other lauraceous species, such as Litsea verticillata, L. akoensis, Lindera communis, and Cinnamomum subavenium (Chen et al., 1998[15]; Shen et al., 2011[31]; Tsai et al., 2002[33]; Zhang et al., 2005[36]).

As part of our ongoing project aimed at the discovery of new natural product-based antitrypanosomal agents from plants occurring in the Cerrado ecosystem of midwest Brazil, we found that an EtOH extract of the xylopodium of Aiouea trinervis Meisn. (Lauraceae) showed in vitro activity against epimastigote forms of Trypanosoma cruzi (IC₅₀ = 6.77 µg mL^-1). Herein, we describe the isolation and structural characterization of three butanolides from this bioactive extract as well as an assessment of their in vitro antiprotozoal activity and cytotoxicity in mammalian Vero cells. In addition, the most active compound was evaluated against amastigote cells, for cell cycle alterations in epimastigote forms, and subjected to in silico prediction for ADMET properties, and detection of PAINS substructures.

After partitioning of the bioactive EtOH extract from the xylopodium of Aiouea trinervis, the resulting hexane, dichloromethane and EtOAc phases were assessed for their in vitro effects against Trypanosoma cruzi epimastigotes. The results revealed that the antitrypanosomal activity resided in the hexane and dichloromethane phases, with IC₅₀ values of 3.30 µgmL^-1 and 8.18 µgmL^-1, respectively, while the EtOAc phase was devoid of any activity.

Further trypanocidal activity-guided fractionation of the bioactive hexane and dichloromethane phases led to the isolation of the polyketide-type butanolides isoobtusilactone A (1) from the hexane phase and epilitsenolide C2 (2) and epilitsenolide C1 (3) from both the hexane and dichloromethane phases (Figure 1(Fig. 1)). In a previous study, we reported the isolation of butanolides 1-3 from the xylopodium and roots of Aiouea trinervis (Garcez et al., 2005[22]). In the present investigation, the structures of 1-3 were determined based on their ¹H and ¹³C NMR, and HRESIMS data, which were in agreement with those formerly reported (Anderson et al., 1992[1]; Chen et al., 2000[16]; Garcez et al., 2005[22]; Rollinson et al., 1981[29]), as well as by comparison with authentic samples.

For this bioprospective study on anti-T. cruzi compounds from plant species of the Brazilian Midwest, we opted for the parasite epimastigote form (strain Dm28c), because it is a clonal population, whose axenic cultivation is well established in vitro, thus guaranteeing homogeneity of the assays (Camargo, 1964[9]). Therefore, the anti-trypanosomal evaluation against epimastigote forms may be employed, not only for assessment of the anti-T. cruzi potential of isolated compounds, but also as a screening for new promising trypanocidal agents in crude extracts. Therefore, following isolation, compounds 1-3 were evaluated in vitro for their activity against T. cruzi epimastigotes, and the results are shown in Table 1(Tab. 1) (the raw data can be found in the Supplementary data, Supplementary Tables 1 and 2).

Isoobtusilactone A (1) was the most active butanolide, with significant IC₅₀ values, at both 24 and 72 h (2.75 and 3.32 µmol L^-1, respectively), followed by epilitsenolide C1 (3) (IC₅₀/24 h = 7.77 and IC₅₀/72 h = 11.12 µmol L^-1). In contrast, epilitsenolide C2 (2) was inactive at the tested concentrations even after a 72 h exposure. The anti-T. cruzi activities of butanolides 1 and 3 were stronger than that of the reference drug benznidazole, and anti-T. cruzi activity of 1 was about 9-fold higher than that of benznidazole at 24 h, and about 4-fold higher at 72 h. Similarly, albeit evidenced to a lesser degree, the activity of butanolide 3 was approximately 68 % and 24 % higher than that of benznidazole at 24 and 72 h, respectively.

The foregoing results revealed some possible structure-activity relationships for the isolated compounds. Based on the IC₅₀ values of 2 and 3, it can be inferred that the (Z)-geometry of the conjugated double bond in the α-alkylidene-γ-lactone skeleton is an important feature for the trypanocidal activity of 3, since its corresponding (E)-stereoisomer 2 showed no activity against T. cruzi epimastigotes. Therefore, with all other things being equal, an improvement in the activity was observed by changing the stereochemistry of the side chain from (E) to (Z). In contrast, butanolide 1, which bears the same (E)-configuration on the side chain as 2, but differs from the latter only by the presence of a Δ^4,5 unsaturation, showed the strongest activity against T. cruzi. These results suggest that the activity of compound 1 stems from the presence of an exocyclic double bond functionality at C-4. This structural feature seems to exert a stronger effect on T. cruzi epimastigotes than the orientation of the side chain at C-6, since compound 1 was the most active compound, even though it has an (E)-configuration of the conjugated double bond. Therefore, it can be assumed that both the geometry of the conjugated double bond at C-2, and particularly the presence of an exocyclic double bond at C-4 play key roles in the activity of this class of compounds against T. cruzi epimastigotes.

When assessing the cytotoxicity of butanolides 1 and 3 against mammalian Vero cells to determine their toxicity and selectivity index (SI) values, compounds 1 and 3 showed CC₅₀ values >156 µmol L^-1 after a 24 h exposure (Table 1(Tab. 1)). The obtained SI values revealed that isoobtusilactone A and epilitsenolide C1 were respectively 56.8 and 28.6 times more selective for the epimastigote forms of T. cruzi than for mammalian cells. According to the Drugs for Neglected Diseases initiative (DNDi), a candidate drug for in vivo trials should have an SI > 10 (Chatelain, 2015[11]). It is also noteworthy that isoobtusilactone A proved more selective than benznidazole, which has an SI of 46.5.

To verify the effect of isoobtusilactone A (1) on the multiplication of T. cruzi epimastigotes, their growth was analyzed in the presence of different concentrations of compound 1 for 72 h. As depicted in Figure 2(Fig. 2), a drastic reduction in the number of parasites was observed following incubation with 8 and 16 μmol L^-1 of isoobtusilactone A, suggesting that it has trypanocidal activity at concentrations as low as 8 μmol L^-1. Data obtained at 72 h indicated that treatment with 4 μmol L^-1 led to a 70 % decrease in growth when compared to the untreated control, while treatment with 16 μmol L^-1 led to the death of more than 80 % of the parasites (the raw data of Figure 2 can be found in the Supplementary data, Supplementary Table 3).

Given the evident alteration in the growth of isoobtusilactone A-treated parasites, the percentage of epimastigotes in each phase of the cell cycle (G1, S and G2/M) was determined by flow cytometric analysis. Based on the cell growth curve data (Figure 2(Fig. 2)), two concentrations were then selected (2.75 and 24.0 μmol L^-1). Cell cycle analysis revealed no difference between the parasites treated with the former concentration of isoobtusilactone A and control parasites, which might indicate that this compound does not affect this cellular mechanism. However, when a higher concentration of the butanolide was applied, a reduction of population in G2 phase of the cycle was observed (from 34 % to 22.3 %), as well as an increase in epimastigotes percentage in both G1 (45.5 % to 52.2 %) and S phase (18.6 % to 23.8 %). This alteration might represent a secondary damage effect caused by isoobtusilactone A to the parasite. It is worth mentioning that cell cycle arrest at the G2-M phase of cancer cell lines has been previously reported for isoobtusilactone A (Kuo et al., 2007[25]). Further studies are therefore needed so that the mechanisms of action of this butanolide can be fully clarified.

Based on the observed cell cycle changes of epimastigote forms treated with the highest concentration of isoobtusilactone A, an intracellular replicative form growth inhibition assay was performed on the vertebrate cell host. It was shown that the same concentration able to induce cell cycle arrest of epimastigote forms also inhibited more than 70 % in growth of amastigote forms after a 24 h treatment. These data indicate the potential of this butanolide also against the replicative form of T. cruzi which is present in humans.

Given the foregoing results, a computational approach was conducted for prediction of some pharmacokinetics properties and toxicity (ADMET) of isoobtusilactone A (1)- the most active butanolide-, using the on line platforms pkCSM, Tox-prediction, STopTox and Pred-hERG. The compliance of 1 to the Lipinski's rule of five calculation, as well as pan-assay interference compounds (PAINS) analysis were also performed, using the pkCSM and PAINS-remover web servers, respectively. In silico estimation approaches of absorption, metabolism, excretion and toxicity (ADMET) and other drug-like properties, such as the Lipinski's rule of five and PAINS parameters, have to be considered at an early stage on evaluation of potential candidates for drug development (Baell and Nissink, 2018[3]; Lipinski et al., 2012[26]; Pires et al., 2015[28]). Therefore, according to Table 2(Tab. 2), butanolide 1 is expected to be non-carcinogenic, non-mutagenic, non-hepatotoxic, and non-hERG blocker (non-cardiotoxic), as well as having no rat toxicity by the oral route. It is also likely to be metabolized by CYP3A4 and an inhibitor of CYP2C19 enzymes, but probable neither substrate nor inhibitor of the other main cytochrome P450 isoforms analyzed. The evaluation of absorption parameters revealed that 1 has a good prediction of human intestinal permeability and absorption and a relatively low probability of skin permeability. As for blood-brain barrier (BBB) and central nervous system (CNS) permeabilities, this butanolide is predicted to have a poor CNS permeability, but a favorable BBB penetration. Compound 1 was also not recognized as a pan-assay interference compound (PAINS), which lowers the probability of its biological properties be considered as artifacts (Baell and Nissink, 2018[3]). These results, in addition to no violation of Lipinski's rule of five, support the drug-like characteristics of isoobtusilactone A.

Lauraceous species are known to be sources of secondary metabolites with trypanocidal activity (Uchiyama et al., 2002[34]; Da Silva Filho et al., 2004[19]; Setzer and Setzer, 2006[30]; Cabral et al., 2010[8]; Conserva et al., 2019[17]). Among these bioactive compounds, only two butanolides (isolinderadolide D and isolinderadolide E) and a seco-butanolide (seco-subamolide A) from Nectandra opposittifolia have hitherto been reported (Conserva et al., 2019[17]). Although compounds 1-3 were obtained in our previous investigation of Aiouea trinervis (Garcez et al., 2005[22]), this is the first report on the isolation of the bioactive butanolides isoobtusilactone A (1) and epilitsenolide C1 (3), and the inactive epilitsenolide C2 (2) using in vitro trypanocidal activity-guided fractionation against epimastigote forms of T. cruzi. Given the IC₅₀ and SI values of compound 1, and its good drug-like characteristics revealed by the results obtained in the in silico analyses, this butanolide can be considered as a promising candidate for further investigation of its mechanisms of action against the parasitic stages of T. cruzi, as well as studies of structural modifications in order to potentiate its anti-T. cruzi properties. The bioactive butanolides from A. trinervis and others from the Lauraceae may thus be regarded as a promising class of compounds in the search for new trypanocidal agents.

The authors gratefully acknowledge the Fundação de Apoio ao Desenvolvimento do Ensino, Ciência e Tecnologia do Estado de Mato Grosso do Sul (FUNDECT-grant numbers 027159 and 021683) for financial support, and the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES, Finance Code 001), for the grants awarded to F.O.N. and J.M.A. Dr. Arnildo Pott is acknowledged for his assistance in the identification of the plant material.

The authors declare that they have no conflict of interest.

Supplementary data (¹H and ¹³C NMR spectra and HRESIMS spectra of compounds 1-3, and raw data of Table 1 and Figure 2) can be found here.