Abstract
The risk behaviors underlying the most prevalent chronic noncommunicable diseases (NCDs) encompass alcohol misuse, unhealthy diets, smoking and sedentary lifestyle behaviors. These are all linked to the altered function of the mesocorticolimbic (MCL) system. As the mesocorticolimbic circuit is central to the reward pathway and is involved in risk behaviors and mental disorders, we set out to test the hypothesis that these pathologies may be approached therapeutically as a group. To address these questions, the identification of novel targets by exploiting knowledge-based, network-based and disease similarity algorithms in two major Thomson Reuters databases (MetaBase™, a database of manually annotated protein interactions and biological pathways, and Integrity(SM), a unique knowledge solution integrating biological, chemical and pharmacological data) was performed. Each approach scored proteins from a particular approach-specific standpoint, followed by integration of the scores by machine learning techniques yielding an integrated score for final target prioritization. Machine learning identified characteristic patterns of the already known targets (control targets) with high accuracy (area under curve of the receiver operator curve was ~93%). The analysis resulted in a prioritized list of 250 targets for MCL disorders, many of which are well established targets for the mesocorticolimbic circuit e.g., dopamine receptors, monoamino oxidases and serotonin receptors, whereas emerging targets included DPP4, PPARG, NOS1, ACE, ARB1, CREB1, POMC and diverse voltage-gated Ca(2+) channels. Our findings support the hypothesis that disorders involving the mesocorticolimbic circuit may share key molecular pathology aspects and may be causally linked to NCDs, yielding novel targets for drug repurposing and personalized medicine.