PIERRE NASSOY

Our current efforts focus on optimizing the production of capsules (sphericity, monodispersity) and increasing the versatility of the CCT (shape, stiffness, chemical composition). We aim at designing integrated microfluidics-based lab-on-chips dedicated to the sorting and in situ manipulation, treatments and analysis of 3D cell cultures.

We use cellular capsules to generate iPSC-derived neuronal organoids, artificial liquid tumors comprising B lymphocytes and primary stromal cells, engineered liver modules or in vitro blood vessels, coined vesseloids. This allows us (together with our collaborators) to venture into cell therapy, chemotherapy assays, tumor vascularization, NASH modelling, vessel transplantation.

The alginate capsules can serve as microcompartments to allow two (or more) types of cells to self-organize in a restricted space. Due to the elastic properties of the alginate gel, the capsules can also be viewed as a mechanical tool to apply an external pressure on the growing cell aggregates once they have reached 3D confluence. These compressive cues may lead to a phenotypic switch towards an invasive phenotype, to an enwrapping/spreading transition, to a mechanically-induced differentiation of progenitor cells, or to unusual pulsatile growth dynamics. We always aim for quantitative imaging and theoretical modelling relying on simple concepts inspired from soft matter physics to explain the observed behaviors.

Mathematical models of biological tissues prove increasingly useful in experimental data interpretation when analytical physical models are insufficient. However, these models often rely on a vast number of unknown parameters. In order to increase their predictive value, we pursue a bottom-up approach by usual the cellular capsules as simplified and controlled model tissues. By tuning the cell seeding density, the shape of the capsules, their stiffness, we may help calibrate the mathematical models and bring a reliable contribution to digital twins. This work is based on collaborations with the team of G. Sciumé (for equation-based models) and the one of D. Drasdo (for agent-based approaches).