Illustration realized in the framework of a collaboration between the Image/Recit option of the HEAD (Haute École d'Art et de Design) - Genève and the Faculty of Sciences of the University of Geneva.

Colorectal cancer figures among the most common and deadly cancers worldwide. With current advances in treatment strategies, a strong need for smarter therapies remains.  Treatment options are especially limited for late-stage colorectal cancer patients, relying mainly on chemotherapy, known as FOLFOXIRI. However, over time this treatment modality induces resistance in most patients, which makes the outcome very poor and the 5-year survival rate of only 15% for late colorectal cancer. It is generally known that the use of one single drug to treat cancer results in loss of efficacy over time, also known as resistance emergence. Therefore, instead of prescribing only one drug, oncologists generally treat their patients with a combination of drugs that target different signaling pathways in the cell, to improve the response. But it is extremely difficult to identify optimal drug mixtures, as the number of combinatorial possibilities is infinite. For example, if we need to test 10 drugs, each at two different doses, we will have 2¹⁰ different possibilities.

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 colorectal cancer cell lines both naïve and chronically treated with the first line chemotherapy treatment, FOLFOXIRI, grown in complex 3D cell models. We validated the activity of the optimized drug combinations on freshly isolated colorectal cancer patient-derived samples that have been diagnosed at the Clinical Pathology Service (Hôpitaux Universitaires de Genève) in collaboration with the department of Clinical Oncology. A new simple culture method was optimized to obtain a single patient-derived organoid at a clinically relevant size of 300-400μm per well, at which size physiological parameters of the tumor are present. Organoids are 3D cellular construct that mimic the architecture, functionality, and genetic features of the corresponding organ they are derived from. Our results showcased a strong inter-patient variable response when the patient-derived organoid from different patients were exposed to the same treatment. Therefore, using our platform we performed a patient-specific drug screen directly on three different patient-derived colorectal cancer material. Using different statistical modeling tools, we identified the most interesting and synergistic interactions between drugs, leading to the optimization of synergistic, low-dose, selective, patient-specific drug mixtures. For each patient, we optimized in terms of drugs and doses a personalized four-drug combination, inhibiting up to 80% of cancer cell viability in the corresponding patient material, while being non-toxic on the healthy colon cells. With our collaborators at the Ecole Polytechnique Fédérale de Lausanne (EPFL), we analyzed the tumor cells – by sequencing the whole exome and RNA - to define the stage of the tumor and the main driver mutations in each patient material, which gave us insight on the mechanism of action of the optimized patient-specific drug mixture. 

We represent an innovative approach in precision oncology for easier, faster, and optimal clinical translation of personalized treatment. This proof-of-concept is based on a “from bench to bedside and back” strategy, which means that the results from our research can be directly translated into the clinics, bringing new therapeutic options to the patients.  This strategy is carried out by fundamental researchers, statisticians, and clinicians, and will lead for the first time to the rapid optimization of synergistic multi-drug combination therapy tailored specifically to individual patients.