3D imaging of immune cell reaction in pancreatic cancer organoids: the response to a nanovaccine therapy

This PhD thesis explored the use of patient-derived organoids (PDOs) obtained from tumor tissue, to address research limitations and therapeutic challenges in pancreatic ductal adenocarcinoma (PDAC), characterized by a remarkable inter- and intra-tumoral heterogeneity, by an immunosuppressive and immune evasive microenvironment and resistance to conventional cytostatic drugs and immunotherapy. Two separate but complementary in vitro studies were conducted applying immune cell co-cultures based on PDAC PDOs to innovate PDAC research at the preclinical level and provide a platform to monitor therapeutic approaches. The first research priority outlines cell live imaging with OrganoIDNet, a platform for real-time characterization and monitoring of PDAC organoid response to both chemotherapy and immunotherapy. PDAC organoids, which closely mimic in vivo tumor behavior, were cultured either alone or in co-culture with immune cells, allowing assessment over time of not only cytostatic but also immunotherapeutic effects in individual patient-derived PDAC organoids. Initially, OrganoIDNet, a deep-learning-based algorithm capable of analyzing bright-field images of murine and human PDAC organoids acquired with live-cell imaging was established to evaluate the effects of gemcitabine, a nucleoside analog considered the gold standard in systemic chemotherapy. This study reveals distinct responses to gemcitabine in mouse and human PDAC organoids, emphasizing the platform sensitivity to minor variations. Real-time monitoring provided organoid morphological parameters such as eccentricity, in response to gemcitabine. Using OrganoIDNet, organoid healthy status was assessed using pixel intensity analysis, which classifies organoids as healthy or unhealthy based on brightness values. While gemcitabine-induced reduction in organoid count and area are significant, pixel intensity analysis depicts size-dependent effects of drugs on organoids and distinguishes between healthy and unhealthy status, both parameters undetectable by conventional endpoint viability assays, underlying OrganoIDNet's superiority in analyzing organoid behavior, such as size-dependent therapy effects over time. To account for organoid size heterogeneity, OrganoIDNet is capable of categorizing PDAC PDOs into five size bins, revealing size-dependent responses to gemcitabine. Smaller organoids increase in number, while larger ones decrease in response to treatment, highlighting the platform's capacity to identify size-based effects in organoid cultures in response to therapy. OrganoIDNet's accuracy in depicting anti-tumor efficacy was validated through the endpoint assay CellTiter-Glo, confirming its reliability in assessing organoid viability. Most importantly, the platform's potential to improve our understanding of personalized drug responses is underlined by the dynamic insights into organoid responses that cannot be captured by conventional endpoint assays. In addition, a new organoids/PBMCs sandwich-based co-culture protocol enabled longitudinal analysis of organoid responses to immunotherapy with the PD-L1 inhibitor Atezolizumab. The use of OrganoIDNet and imaging of live cells revealed an increased potency of PBMCs tumor-killing in an organoid-individual manner when Atezolizumab was added. The second part focused on investigating the potential of Mesovac, a mesothelin (MSLN)-based nanovaccine formulation alone or in combination with chemotherapy or the PD-L1 antibody Atezolizumab as an immunotherapeutic strategy for PDAC. This was achieved in multiple experimental stages, including in vitro T-cell stimulation assays, ex vivo T-cell expansion of reactive T cells, and immune cell co-culture experiments based on PDAC PDOs and stimulated T-cells. First, to assess Mesovac ability to induce an immune response, in vitro T-cell-based peptide stimulation was conducted using unfractionated peripheral blood mononuclear cells (PBMCs) from healthy donors. Although modest, an increase in IFN-γ+ T cells suggests the Mesovac potential to activate T-cell responses in vitro. To overcome variability and expand Mesovac-stimulated T cells, artificial antigen-presenting cells (aAPCs) were employed. These aAPCs, coupled with MSLN4 and adjuvants, enhance the reactivity of PDAC patient-derived T cells when exposed to Mesovac components. Further investigations explore the effect of Mesothelin-peptide-stimulated T cells on PDO growth and proliferation. The combination of Mesothelin-peptide-stimulated T cells with gemcitabine shows promising results in attenuating PDO expansion, particularly in synergistic effect with gemcitabine. Additionally, Mesothelin-peptide-stimulated T cells combined with Atezolizumab and FOLFIRINOX results in an increased efficacy in targeting PDAC organoids. Mesothelin-peptide-stimulated T cells treatment reduces the overall tumor cell population in PDAC PDO co-cultures as well as cells expressing PDAC-associated tumor markers, measured by flow cytometry. Moreover, cells expressing cancer stem markers, such as CD24, EpCAM or CD133, show an overall reduction upon treatment with FOLFIRINOX. This effect is even more evident with the addition of Atezolizumab. In summary, this PhD thesis provides a novel platform for evaluating therapeutic effects on organoids in co-cultures with immune cells. OrganoIDNet, an innovative tool that integrates artificial intelligence with real-time imaging, offers insights into PDAC organoid responses, enhancing personalized drug response assessments. Simultaneously, the co-culture approach using activated T-cells with PDAC PDOs demonstrates its potential to evaluate individual responses of organoids to chemo- and immunotherapeutic on the used case of Mesothelin peptides, emphasizing the importance of personalized strategies in combating PDAC. These insights collectively contribute to developing more effective and tailored therapies for this challenging malignancy.