Poster S20 - Modeling tumor stromal matrix structure and its impact on T cell immunotherapy

Introduction: Tumor stromal microenvironment, such as the dense matrix fiber network, poses a major obstacle for T cell infiltration. The fibrotic tumor stroma, remodeled by cancer-associated fibroblasts (CAFs), forms matrix fiber patterns known as tumor-associated collagen signatures (TACS), which have distinct mechanical and structural properties. The factors that regulate the formation of the fibrotic tumor stroma and their impact on T cell infiltration are unknown. Recently, we created a tumor-stroma model that closely mimics the collagen signatures observed in actual tumor tissue. We characterized the structural properties of the stromal matrix and compared them with those of the actual tumor tissue. The infiltration of cytotoxic T cells in this model was also reported.

Materials and

Methods: Soft lithography techniques were used to fabricate square PDMS wells. Pre-formed microspheres composed of NY-ESO antigen-positive A549 cells were placed in the PDMS well prior to seeding cancer-associated fibroblasts (CAFs). Once a mixture of CAFs and collagen type I matrix was placed in the PDMS well, microtissues formed through cell-mediated self-assembly around the micropillars, encapsulating the A549 microspheres. Fibronectin (FN) staining was used to indicate the distribution of matrix fibers in the tumor stroma, including the collagen fibers. Second Harmonic Generation (SHG) images of collagen structure in ductal carcinoma were obtained from the "HE-vs-MPM" collection of The Cancer Imaging Archive (TCIA). Matrix fiber properties were analyzed using a custom Matlab code modified from CurveAlign V5.0. NY-ESO+ CD8+ T cells were activated and dropped onto the microtissue surface at a 4:1 T cell:fibroblast ratio. Collagen signatures and T cell distribution relevant to A549 spheres in the tissue were examined and compared to tumor slices from patients.

Results, Conclusions, and Discussions: Analysis of the stromal matrix structure of the ductal carcinoma showed stretched collagen fiber bundles aligning along the edge of the tumor core, representing tumor-associated collagen signature 2 (TACS2) (Fig. 1A,B). The density of fibers peaks at the tumor-stroma interface, then progressively decreases with increasing distance from the edge of the tumor (Fig. 1B). Confocal imaging and analysis of the matrix structure of the tumor-stroma model showed elongated bundles of matrix fibers aligned along the perimeter of the A549 tumor core (Fig. 1D-E). The density of these fibers diminished from the tumor-stroma boundary to areas farther from the tumor edge (Fig. 1E). The observed matrix structures resembled those found in ductal carcinoma, indicating the model's success in replicating tumor-associated matrix structures. Upon introducing TCR-T cells into the tumor-stroma model, the T cells displayed a regionally varied distribution within the tissue. The concentration of T cells was highest in areas of dense matrix fiber at the tumor-stroma boundary, as shown in Figure 1G. The number of T cells sharply decreased both towards the interior of the tumor and in regions distant from the tumor-stroma boundary. This pattern of distribution closely mirrors that of T cells around solid tumors observed in animal studies and clinical specimens. These findings imply that the infiltration of T cells into the tumor core is hindered by the dense matrix fibers encasing the tumor core. The developed tumor-stroma model offers a valuable tool for exploring interactions between T cells and the stroma in the context of cancer immunotherapy.
This 3D co-cultured model system encapsulates CAF-tumor interaction and tumor-associated collagen signatures (TACS), providing a novel platform for investigating the mechanisms of T cell infiltration. It may also serve as a future platform for studying metastasis and testing T cell immunotherapy.

Acknowledgements (Optional):