Abstract
Hepatocellular carcinoma (HCC) is the main primary liver cancer. Its prognosis is poor, notably due to limited therapeutic options for the advanced stages of the disease. Genome-wide analyses of human tumor samples identified the common oncogenic drivers and allowed the molecular classification of HCC. They also highlighted a complex genetic landscape of HCC, that results in extensive inter-tumor heterogeneity. My thesis project was focused on another aspect of tumor heterogeneity: the study of oncogenic cooperation and functional intra-tumor heterogeneity led me to discover quantitative subclonal differences in the growing tumors. This quantitative heterogeneity originates from the variations of an oncogenic signal intensity that modulates the interactions of cancer cells with their microenvironment.To study population dynamics and intercellular interactions during tumor growth and dissemination, we have developed a murine model of hepatocellular carcinoma that combines intrahepatic injection of cells and lineage tracing. Orthotopic allografts of fluorescently labeled cells generate hepatic tumors, whose growth was analyzed, notably in terms of clonal composition, invasion and metastasis. My results revealed a novel aspect of clonal competition in a growing tumor, namely a selective advantage afforded to tumor cells by an optimal intensity of activation of a major signal transduction pathway: the MAPK pathway. I showed that the optimal activation intensity, compatible with the tumor growth, is defined both by the intrinsic, cell-autonomous mechanisms and by the interactions with the microenvironment. In-depth analysis of the stromal component, which differs between the primary and the metastatic tumor locations, allowed me to identify stromal cell types that appear to modulate the selection and to propose a model of molecular interactions involved in this process.