A Personalized Therapeutics Approach Using an In Silico Drosophila Patient Model Reveals Optimal Chemo- and Targeted Therapy Combinations for Colorectal Cancer

Abstract

In silico models of biomolecular regulation in cancer, annotated with patient-specific gene expression data, can aid in the development of novel personalized cancer therapeutic strategies. Drosophila melanogaster is a well-established animal model that is increasingly being employed to evaluate such preclinical personalized cancer therapies. Here, we report five Boolean network models of biomolecular regulation in cells lining the Drosophila midgut epithelium and annotate them with colorectal cancer patient-specific mutation data to develop an in silico Drosophila Patient Model (DPM). We employed cell-type-specific RNA-seq gene expression data from the FlyGut-seq database to annotate and then validate these networks. Next, we developed three literature-based colorectal cancer case studies to evaluate cell fate outcomes from the model. Results obtained from analyses of the proposed DPM help: (i) elucidate cell fate evolution in colorectal tumorigenesis, (ii) validate cytotoxicity of nine FDA-approved CRC drugs, and (iii) devise optimal personalized treatment combinations. The personalized network models helped identify synergistic combinations of paclitaxel-regorafenib, paclitaxel-bortezomib, docetaxel-bortezomib, and paclitaxel-imatinib for treating different colorectal cancer patients. Follow-on therapeutic screening of six colorectal cancer patients from cBioPortal using this drug combination demonstrated a 100% increase in apoptosis and a 100% decrease in proliferation. In conclusion, this work outlines a novel roadmap for decoding colorectal tumorigenesis along with the development of personalized combinatorial therapeutics for preclinical translational studies.

Navigating Multi-Scale Cancer Systems Biology Towards Model-Driven Clinical Oncology and Its Applications in Personalized Therapeutics

“…In particular, data-drive computational models are now acting as an enabling technology for precision medicine and personalized treatment of cancer. To date, several single and multi-scale software platforms have been reported to model biomolecules (171,(221)(222)(223)(224)(225)(226)(227)(228), cellular environments (229), as well as cell-level (230), and multi-scale (231) information into coherent in silico cancer models (Figure 3).…”

“…This approach has been widely adopted as a tool to model gene, transcript, protein, and metabolite regulatory networks. Numerous mathematical and computational cancer models have been developed using this representation (171,(221)(222)(223)(224)(225). To facilitate the Boolean model development and analysis process, several platforms have been devised (234)(235)(236) (Table S8).…”

“…TISON aims to develop single or multi-scale models for designing personalized cancer therapeutics. To exemplify the use case for TISON, Gondal et al employed TISON to model colorectal tumorigenesis in Drosophila melanogaster’ s midgut towards evaluating efficacious combinatorial therapies for individual colorectal cancer patients ( 225 ).…”

Combined Cytotoxic Effect of Inhibitors of Proteostasis on Human Colon Cancer Cells

“…Despite the apparent diversity of drugs and treatment regimens, there are a number of problems associated with these approaches. The main disadvantage of using monotherapeutic agents for cancer treatment is tumor heterogeneity, which leads to the selection of resistant cells, in addition to the fact that they act specifically on individual pathways [ 4 ]. Furthermore, when cancer cells are exposed to chemotherapy drugs, the surrounding healthy cells and tissues also suffer [ 5 ].…”

CanSeer: A Method for Development and Clinical Translation of Personalized Cancer Therapeutics

“…The proposed annotation of network models with somatic mutations remained deficient in construing a comprehensive molecular signature of individual patients. Moreover, the coverage of druggable and clinically actionable targets remained limited during exhaustive screening for predicting efficacious targets [74], [75]. To date, a multi-factorial patient data-integrative approach to simulate drug effects using drug scores, followed by quantification of therapy response remains unassessed.…”