Thesis, COLLÉGIALITÉ, FRANZEN Rachelle

Abstract

Triple-negative breast cancer (TNBC) is one of the most aggressive and heterogeneous subtypes of the disease, characterised by the absence of specific therapeutic targets. Given the limited effectiveness of current treatments, the development of new therapeutic strategies is a priority. The screening of anti-tumour compounds is mainly based on the use of simplified in vitro models, almost exclusively consisting of tumour cell monocultures. However, these models only partially reflect the biological complexity of tumours in vivo, particularly in terms of cell-cell and cell-extracellular matrix interactions. This limitation contributes in part to the frequent failure of treatments once they have been applied in vivo in the preclinical or clinical phases.

The aim of this work was to design more representative cell models by integrating immortalised fibroblasts co-cultured with tumour cells derived from TNBC lines, in two-dimensional and three-dimensional systems in the form of spheroids. The aim of these configurations is to reproduce more faithfully certain characteristics of the tumour microenvironment, while enabling the dynamic study of cell viability and response. Two compounds were explored: etoposide, a classic chemotherapy agent, and cannabidiol (CBD), a molecule with emerging anti-tumour properties. Their impact, alone or in combination, was assessed using live cell imaging. The potential synergy of these treatments was then analysed using appropriate computer tools. The overall aim of this approach is to propose more robust in vitro models for screening new therapies, with the ambition of better predicting the responses observed in vivo and, ultimately, helping to optimise therapeutic strategies against TNBC.

Comparison of heterocellular 2D models, with varying proportions of fibroblasts, revealed a protective effect of the latter against CBD-induced cytotoxicity. Conversely, their presence did not attenuate the effect of etoposide, suggesting a differential response depending on the type of treatment. The combination of the two compounds mainly led to an additive effect, with no marked synergy. In addition, a large number of heterospheroids were successfully generated using microfluidic technology. However, due to time constraints and contamination problems during the spheroid recovery phase, CBD and etoposide treatments could not be tested on these 3D models.

This project highlights the relevance of real-time imaging on living cells and microfluidic technologies for the development of complex models that more accurately represent the tumour microenvironment. The approach developed here offers an adaptable experimental framework for studying anti-tumour compounds, considering the cellular interactions that can influence the response to treatments.