Selecting targets for the diagnosis of Schistosoma mansoni infection: An integrative approach using multi-omic and...

In order to effectively control and monitor schistosomiasis, new diagnostic methods are essential. Taking advantage of computational approaches provided by immunoinformatics and considering the availability of Schistosoma mansoni predicted proteome information, candidate antigens of schistosomiasis were selected and used in immunodiagnosis tests based on Enzime-linked Immunosorbent Assay (ELISA). The computational selection strategy was based on signal peptide prediction; low similarity to human proteins; B- and T-cell epitope prediction; location and expression in different parasite life stages within definitive host. Results of the above-mentioned analysis were parsed to extract meaningful biological information and loaded into a relational database developed to integrate them. In the end, seven proteins were selected and one B-cell linear epitope from each one of them was selected using B-cell epitope score and the presence of intrinsically disordered regions (IDRs). These predicted epitopes generated synthetic peptides that were used in ELISA assays to validate the rational strategy of in silico selection. ELISA was performed using sera from residents of areas of low endemicity for S. mansoni infection and also from healthy donors (HD), not living in an endemic area for schistosomiasis. Discrimination of negative (NEG) and positive (INF) individuals from endemic areas was performed using parasitological and molecular methods. All infected individuals were treated with praziquantel, and serum samples were obtained from them 30 and 180 days post-treatment (30DPT and 180DPT). Results revealed higher IgG levels in INF group than in HD and NEG groups when peptides 1, 3, 4, 5 and 7 were used. Moreover, using peptide 5, ELISA achieved the best performance, since it could discriminate between individuals living in an endemic area that were actively infected from those that were not (NEG, 30DPT, 180DPT groups). Our experimental results also indicate that the computational prediction approach developed is feasible for identifying promising candidates for the diagnosis of schistosomiasis and other diseases.