CHARACTERIZATION OF A 3D IN VITRO TUMOR AND STROMAL SPHEROID MODEL FOR EVALUATING THE SPATIAL DISTRIBUTION OF STROMAL AND TUMOR CELLS AND THE INFILTRATION OF IMMUNE CELLS IN THE TUMOR MICROENVIRONMENT OF PANCREATIC DUCTAL ADENOCARCINOMA

Abstract

PDA, the most frequent and aggressive subtype of pancreatic cancer, has a stroma-rich TME (̃80% of tumor mass). The characteristically hypoxic TME orchestrates an extensive metabolic reprogramming in immune, stromal and tumor cells, inducing epithelial-mesenchymal transition (EMT), promoting migration to regions better perfused and establishing physical and functional barriers that limit infiltration and affect the distribution and function of immune cells, resulting in a typically immuno-excluded profile associated with resistance to immunotherapies. Here, we sought to establish a 3D spheroid model of PDA's TME to dynamically evaluate spatial distribution of stromal and tumor cells, to phenotypically characterize the distribution of infiltrating immune cells and the contribution of hypoxia to these aspects. To this end, spheroids were generated with fluorescently-labeled cells from PDA (CAPAN-1) and stromal (HS-5) lines. The spatial distribution was evaluated at different times by automated quantitative microscopy (24h, 48h and 72h) or flow cytometry (72h). Fluorescently-labeled blood mononuclear cells (PBMCs) from healthy donors, activated or not with Ionomycin and PMA, were plated together with unlabeled tumor and stromal cells to incorporate into forming spheroids, or after their formation (24h post-plating) for infiltration evaluation. After 48h, the spheroids were enzymatically dissociated and T-cells were immunophenotyped by flow cytometry with CD4, CD8 and CD25 antibodies. The Image-iT Green Hypoxia reagent was used to fluorescently label the cells in the spheroid core hypoxic compartment. After 24h of plating, tumor and stromal cells were well distributed in the formed spheroid; however, after 48h and 72h, tumor cells migrated to spheroid periphery. In agreement, cytometry revealed higher proportion of tumor cells occupied the normoxic compartment, while the hypoxic compartment was enriched for stromal cells. Immunophenotyping showed a enrichment of CD4+ cells (̃70-90%) in relation to CD8+ cells (̃10-30%) in the hypoxic compartment in all experimental conditions. Of note, the CD8+ cells percentage in the hypoxic region was higher in conditions in which PBMCs were activated (̃20-30%), compared to non-activated (̃10%). In turn, normoxic regions had similar proportions of both populations (̃50% of each) in conditions with activated PBMCs, while with non-activated PBMCs a higher CD4+ cells (70%) proportion was observed. For activated PBMCs, in both conditions, we found higher CD25+ cells percentage among the CD4+ and CD8+ populations in the hypoxic compartment (̃90%), compared to normoxic compartment (̃60%). However, for PBMCs infiltrating spheroids already formed, there was a greater activation of lymphocytes, resulting in the presence of CD4+CD25High cells, with a supposedly regulatory profile (Tregs), enriched especially in the hypoxic compartment (̃9%). We successfully developed an experimental approach that allowed us to quantitatively and qualitatively characterize how the spatial distribution of tumor, stromal, and immune cells are affected by the hypoxia in the TME of a PDA model. Our in vitro approach may help to systematize the evaluation of new therapeutic approaches.