Identification of systems biomarkers and candidate drugs in prostate adenocarcinoma

Abstract

Deciphering the alterations in the protein interactome is mandatory to reach a systems-level understanding of tumorigenesis, since physical interactions among proteins influence cellular pathways, and mediate various physiological processes in all living organisms [1]. Also, the elucidating of the molecular mechanisms underlying cancer and the identification of efficacious biomarkers is crucial for accurate diagnosis and prognosis of cancers as well as the prevention of tumorigenesis. However, it is a challenging task to develop highly accurate and robust biomarkers considering the complexity of the molecular biology behind these pathologies. On the other hand, the discovery and production of therapeutic agents for cancer treatment need investments in point of money, time and labor and thus it causes waste of money and time. Increasing studies and improvements on omic technologies and computational analysis provide opportunities for drug repostioning which is a useful approach to find out already approved drugs with high confidence and good pharmacokinetic properties on new diseases [2]. The present study highlights the concept of systems biomarkers with special focus on prostate adenocarcinoma using our differential interactome approach and provides the presenting of new drug candidates for prostate adenocarcinoma. To that end, differential interactome algorithm [3] was developed and applied to gene expression profile of prostate adenocarcinoma having 550 samples (52 normal - 498 tumor samples) by using the human protein interactome data in order to find significant protein-protein interactions differentiated at tumor state (dPPI). Our results show that 183 dPPIs among 194 differentially interacting proteins (DIPs) taking roles in dPPIs were significant in prostate adenocarcinoma, which 19 of these dPPIs were repressed, while 164 were activated in tumor phenotype. In addition, the features of DIPs were investigated and it was found that 77 of DIPs were druggable, 49 DIPs were tumor suppressor proteins and 50 DIPs were oncogenes. Also, the gene set enrichment analysis were carried out for DIPs. Moreover, we found various significant modules consisting of DIPs and their diagnostic and prognostic features were examined. Several modules were determined as having diagnostic properties with sensitivity and specificity values > 0.9 and as having prognostic properties with hazard ratio>7, p-value<0.05. Finally, some candidate therapeutic agents were submitted for prostate adenocarcinoma by the aid of drug repositioning approach. This study will pave the way for further studies integrating with systems-level analyses of cancers and provide an insight for clinic studies to design diagnostic kit for early diagnosis of cancer and to investigate repositioned cancer drugs.