Abstract:
BACKGROUND: The merits of ethnomedicine-led approach
to identify and prioritize anticancer medicinal plants have
been challenged as cancer is more likely to be poorly
understood in traditional medicine practices. Nonetheless, it
is also believed that useful data can be generated by
combining ethnobotanical findings with available scientific
studies. Thus, this study combined an ethnobtanical study
with ligand based in silico screening to identify relevant
medical plants and predict their anticancer potential based on
their phytoconstiutents reported in scientific literatures.
METHODS: First, relevant medicinal plants were identified
through an ethnobotanical survey. A list of phytochemicals was
prepared based on literature review of articles which reported on
the natural products of identified medicinal plants. Then, their
phytochemicals were subjected to in silico evaluation, which
included a hybrid score similarity measure, rule of five, Ghose–
Viswanadhan–Wendoloski (GVW)-indices and structural features
criteria, to predict their anticancer activity and drugability.
RESULTS: A total of 18 medicinal plants and 265
phytoconstituents were identified. The natural product pool
constituted 109(41.13%) terpenoids, 67(25.28%) phenolics,
29(10.94%) simple and functionalized hydrocarbons,
26(9.81%) alkaloids, 25(9.43%) glycosides and 9(3.40%)
compounds belonging to different phytochemical classes. The
similarity measure using CDRUG identified 34(12.73%)
phytochemicals with high (p-Value < 0.05) and 35(13.21%)
with moderate possibility (p-Value < 0.1) of anticancer
activity. In fact, three of the predicted compounds had the
same structure with known anticancer compounds
(HSCORE=1). The 80% GVW-indices based antineoplastic
drugabilityranges were all mate by 25 of the predicted
compounds. Predicted compounds were also shown to have
ring structures and functional groups deemed important for
anticancer activity.
