ED Sciences de la Vie et de la Santé
Targeted lentiviral niche engineering to enhance CAR T-cell efficacy in pancreatic ductal adenocarcinoma
by Julia ROSSI (BoRdeaux Institute of onCology)
The defense will take place at 14h00 - Amphithéâtre BBS Bâtiment BBS, Site Carreire 2 rue du Dr Hoffmann Martinot 33076 Bordeaux
in front of the jury composed of
- Sandrine DABERNAT - Professeure - Université de Bordeaux - Directeur de these
- Melita IRVING - PhD, Team Leader, Senior Lecturer, Maître d'Enseignement et de Recherche de type 1 - University of Lausanne, Suisse - Rapporteur
- Jacob GIEHM MIKKELSEN - Professor - University of Aarhus C, Denmark - Rapporteur
- Alice CARRIER - Directrice de recherche - Université de Marseille - Examinateur
- Nicolas LARMONIER - Professeur - Université de Bordeaux - Examinateur
Pancreatic ductal adenocarcinoma (PDAC) remains one of the most lethal solid malignancies, with limited therapeutic options and poor clinical outcomes. Although immunotherapies have transformed the treatment of several cancers, their efficacy in PDAC remains limited. Among these approaches, adoptive cell transfer using chimeric antigen receptor CAR T cells offers strong cytotoxic potential, but its application to PDAC is constrained by inefficient tumor infiltration, dense stromal barriers, and an immunosuppressive tumor microenvironment. This thesis explores targeted lentiviral niche engineering as a strategy to locally remodel the PDAC microenvironment and improve CAR T-cell efficacy. As transducing every tumor cell is not possible, the proposed approach relies on selective gene delivery to a fraction of antigen-expressing tumor cells, converting them into local sources of diffusible immunomodulatory and stromal-remodeling signals. To achieve tumor-restricted delivery, lentiviral vectors were engineered with retargeted viral glycoproteins designed to recognize tumor-associated surface antigens. A comparative analysis of pseudotyping platforms identified Sindbis virus-derived glycoproteins as a robust and versatile system for antigen-dependent lentiviral transduction of solid tumor cells. This targeted delivery platform will be used to express therapeutic payloads designed to locally reshape the tumor niche, including chemokines to promote immune-cell attraction and decorin to modulate the stromal compartment and facilitate tumor access of endogenous and adoptively transferred cytotoxic T cells, including CAR T cells. Overall, this work supports the concept that targeted lentiviral vectors can be used as local tumor microenvironment engineering platforms. This approach may provide a complementary strategy to overcome key barriers limiting CAR T-cell efficacy in PDAC and other solid tumors.
ED Sciences Physiques et de l'Ingénieur
Super-resolution and undersampling methods for thermal imaging
by Florian CROUAU (I2M - Institut de Mécanique et d'Ingénierie de Bordeaux)
The defense will take place at 9h30 - Amphithéâtre La Rochefoucauld Bât R, Campus Arts et Métiers de Bordeaux-Talence Esplanade des Arts et Métiers 33405 Talence Cedex France
in front of the jury composed of
- Jean-Luc BATTAGLIA - Professeur - Université de Bordeaux - Directeur de these
- Audrey GIREMUS - Professeure - Université de Bordeaux - IMS - UMR 5218 - Examinateur
- Julien LECOMPAGNON - Docteur - Bundesanstalt für Materialforschung und -prüfung (BAM) - Examinateur
- Vincent SCHICK - Professeur - Mines Nancy - Université de Lorraine - Rapporteur
- Thomas PIERRE - Professeur - IRDL - Université de Bretagne Sud - Rapporteur
- Stéphane CHEVALIER - Maître de conférences - Université de Bordeaux - ENSAM - CoDirecteur de these
Infrared imaging allows for non-destructive testing and evaluation of materials from the micro to the macro scale. It gives information about a materials surface just like visible imaging, but also allows for the visualisation of heat distribution as well as the characterisation of thermal properties on the surface but also in depth. Thus, the panel of materials studied range from composites for aerospace applications, to biological tissue, construction materials such as wood or concrete, or semiconductors and microbatteries. However, many of these applications have been faced with physical limitations related to both the nature of infrared light and the current performance of detectors which prevents the best achievable resolution to be lower than the micrometre scale. In this thesis, multiple approaches for achieving super-resolution and under-sampling are explored. In Chapter 2, a model of image acquisition is defined in order to highlight the two main limits to resolution that are the diffraction limit and the sampling limit, with a brief overview of the possible approaches in the literature that overcome, bypass or otherwise push those limits. Then, two well-known infrared point-by-point methods are used, namely Flying Spot (FS) laser thermography for Chapter 3 and Scanning Photothermal Radiometry (SPR) for Chapter 4 in order to first under-sample an object, that is taking only a few points from a classical regularly sampled grid, then reconstruct an equivalent fully sampled image. Various reconstruction methods are compared with a specific focus on compressed sensing, which is a mathematical paradigm that postulates that most non-noise images are very compressible in the sense that they can be expressed with way less data by expressing it in an adapted basis. In both cases, a reduction by at least 80% of the number of samples was achieved. Then, in Chapter 5, super-resolution is explored in both the diffraction and sampling sense. These studies showed that infrared imaging applications in the lab often suffer from the sampling limit first, which has led to the exploration of a compressive sensing based method based on a generalisation of the single pixel camera, in which each pixel is sub-divided thanks to smaller binary masks. A proof of work was made using visible cameras, that allowed to multiply by 16 the effective number of pixels, but multiple limitations still need to be overcome in order to achieve consistent repeatable results in infrared, making this the main perspective development of this work.