Colorectal cancer (CRC) figures among the most common and deadly cancers. Although cancer treatment has improved over the past decades, there remains an unmet medical need for smarter therapies. Specifically for late-stage CRC patients, where treatment options are limited, relying mainly on chemotherapy. However, over time this treatment modality induces resistance in most patients, which makes the prognosis very poor and the survival dismal for many patients. It is generally known that the combination of drugs can be beneficial. However, it is extremely difficult to identify optimal drug mixtures, as the number of combinatorial possibilities is infinite.

In this thesis, I present various technological and scientific development performed over last four years to advance the optimization of drug combinations that could be further used for treatment of CRC patients in the future.

I revisited drug-drug interactions and activity of clinically used chemotherapy for treatment of CRC patients, known as FOLFOXIRI. This drug combination consists of folinic acid, 5-fluorouracil (5-FU), oxaliplatin and irinotecan. We chronically exposed the CRC cells to induce acquired chemoresistance to FOLFOXIRI, to therefore mimic tumor response to the treatment in the clinical setting. Furthermore, we overcame this acquired FOLFOXIRI resistance by administration of tyrosine kinase inhibitors mixture.

I furthermore worked on an improved method to select an optimized drug combination therapy. Using the proprietary platform developed at the Molecular Pharmacology Group, the Therapeutically Guided Multidrug Optimization (TGMO), we initiated a screen for optimized drug-combinations using complex CRC-based 3-dimentional models. The activity of the optimized drug combinations was validated on freshly isolated patient-derived organoids. Furthermore, the TGMO-based screen was performed directly on the patient-derived CRC material that has been previously diagnosed, classified, and molecularly stratified at the Clinical Pathology Service (HUG) in collaboration with Clinical Oncology (HUG). Using different mathematical modeling tools, including second order linear regression analysis and penalized regression, the most interesting and synergistic interactions between drugs are identified, leading to the optimization of synergistic, low-dose, selective, patient-specific drug mixtures. Using whole exome sequencing and RNA sequencing data, we obtained insight in the mechanism of action of optimized drug mixtures.

This thesis represents a proof-of-concept of a platform for designing precision treatment strategies for a facilitated translation of treatment to be tailored specifically to individual patients. The strength of the proposed approach is based on a “from bench to bedside and back” strategy carried out by specialists from fundamental researchers and clinicians. The proposed approach will, for the first time, lead to the rapid optimization of synergistic multi-drug combination therapy tailored specifically to individual patients. By selecting drugs for which clinical data are already available, a rapid progression to phase I/II clinical trials for the optimized drug mixture is envisioned.