Andrographolide, a diterpene from Cymbopogon schoenanthus, was identified as a new hit compound against Trypanosoma cruzi through machine learning method and experimental evaluation
Chagas disease, a potentially fatal illness caused by the protozoan Trypanosoma cruzi, has limited options for treatment. Natural products, including those from plants, offer a wide variety of structurally complex metabolites with biological activities, including those with antiparasitic potential. The discovery and development of new prototypes from plants frequently displays multiple phases which could be facilitated by use of machine learning techniques in order to provide a fast and efficient form of selecting new hits and leads candidates. Using random forest model, we constructed a model capable of predicting biological activity of new natural products provided from plant against intracellular amastigotes of T. cruzi. From a virtual screening, andrographolide was identified as a promising hit and, in order to prove the obtained results, it was obtained from the hexane extract of Cymbopogon schoenanthus leaves and was chemically characterized by analysis of NMR and MS spectral data. Andrographolide was evaluated against trypomastigote and amastigote forms of T. cruzi, showing IC50 values of 29.4 and 2.9 µM, respectively, while the standard drug benznidazole displayed IC50 of 17.7 and 5.0 µM, respectively. Additionally, the isolated compound displayed reduced cytotoxicity (CC50 = 92.8 µM) against mammalian cells, and a selectivity index against amastigote forms of 32.0, similar to that calculated for benznidazole, with SI = 39.3. Using in silico analysis, it was possible to conclude that andrographolide fulfills many requirements implemented by DNDi to be a lead compound. Therefore, this work was successful in obtaining a machine learning model capable of predicting the activity of compounds against intracellular forms of T. cruzi.