Optimizing drug combination and detecting mechanism of drug combination based on pathway crosstalk in pan-cancer

Abstract The occurrence and development of cancer is a complex process involving multi-factor crosstalk. The low efficiency of drugs is an important problem for clinical treatment. Combination therapy is currently the most effective way to improve cancer treatment, but currently approved drug combinations result mainly from empirical clinical experience. Therefore, it is meaningful to present an approach to discover new drug combinations for cancer treatment. In this work, we integrated the pan-cancer and drugs high-throughput data to construct the miRNA-mediated crosstalk network between pathways of cancer and further construct the therapeutic networks for drug. Totally, 15416 crosstalks between 268 pathways in 21 cancers were identified. Base on the crosstalk inhibition, we developed a computational network biological approach to identify the effective drug which inhibition risk pathway crosstalk of cancer, and then filtrated and optimized the drug combination for cancer treatment. Ultimately, we obtained 687 optimized drug combinations of 83 first-line anticancer drugs in pan-cancers. Next, we analyzed the mechanism of drug combination, and confirmed that the targets of cancer-specific crosstalk network in drug combination were closely related to cancer prognosis by survival analysis. In addition, we developed a user-friendly web interface ODrugCP (http://bio-bigdata.hrbmu.edu.cn/oDrugCP/) that can be used to explore 687 optimized drug combination strategies related to 21 cancers. In conclusion, our study mined the key cascade pathways of drug action and analyzed the mechanism of drug combination. It can provide an effective method for screening precise drug combinations for various cancer treatments, which may have important scientific significance and clinical application value for revealing the regulatory mechanism of tumor pathogenesis and exploring new tumor treatment strategies.