Patient-Specific Models of the Immune-Competent Tumor and its Microenvironment with Physiological Perfusion

Introduction:: Many exciting models have been developed for in vitro oncology studies in recent years, but two major challenges that remain in the field are building immunocompetent models that will enable reliable evaluation of immunotherapies and vascularization of models in order to recapitulate physiologically relevant barriers to delivery of drugs and cell therapies. The microphysiological model of the tumor microenvironment described here is generated from cells and tissue derived from individual cancer resections, in combination with engineered vasculature to enable patient-specific studies of immune cell trafficking. Endothelial cells were engineered to eliminate expression of major histocompatibility complex class I (MHC I) in order to facilitate co-culture with cytotoxic T cells from any donor without generating an allogeneic response. This platform allows T cells to be introduced and re-circulated through vascular networks in microfluidic devices mimicking the tumor microenvironment, which will facilitate our understanding of mechanisms underlying heterogeneous patient-specific responses to immunotherapy.


Materials and Methods:: Patient-specific models were generated using a combination of intact tumor fragments and isolated cells from small tumor resection specimens of head and neck cancers. After collecting organotypic tumor spheroids to retain the tumor microenvironment, cancer associated fibroblasts (CAF) and T cells were isolated and expanded. Human umbilical vein endothelial cells (HUVEC) were engineered to eliminate MHC I expression by knocking out β2-microglobulin using CRISPR/cas9, thus shielding them from T cell recognition. Endothelial cells and fibroblasts were suspended in fibrin solution and injected into microfluidic devices, allowing microvascular networks to self-assemble in the polymerized gel. Organotypic spheroids were similarly introduced in a small port surrounded by the vascular bed, and then T cells were fluorescently labeled and perfused through the microvasculature. T cell extravasation from vasculature and migration toward tumor spheroids was imaged with confocal microscopy. The angiogenic potential of CAF was evaluated in a different microfluidic device, by introducing them in a fibrin gel and applying a monolayer of fluorescently labeled endothelial cells on the gel surface adjacent to the media channel. Secreted factors were quantified using multiplex cytokine assays and ELISA. Image analysis was performed in FIJI, and statistics were performed in R.


Results, Conclusions, and Discussions:: Vascularized tumor-on-chip devices were created with engineered endothelial cells in combination CAF and organotypic tumor spheroids. Once fully perfusable lumens had developed after 5-7 days, autologous T cells were introduced and recirculated at physiological flow rates using a small, custom PDMS pump, and microcopy revealed the fate of fluorescently labeled T cells. Within hours, T cells bound to the endothelium and began to extravasate, and within 2 days, T cells were observed migrating through the matrix between vessels and surrounding the organotypic tumor spheroids. Compared to tissue-matched normal fibroblasts, patient-derived CAF revealed striking effects on the tumor microenvironment. In the angiogenesis assay, CAF elicited significantly greater endothelial sprouting and permeability. Analysis of the secretome revealed overexpression of several targets with angiogenic and inflammatory properties, notably vascular endothelial growth factor (VEGF) and chemokine (C-C motif) ligand 5 (CCL5). Blocking these targets alone or in combination reduced angiogenesis in the model. The presence of CAF alone was sufficient to increase immune cell recruitment, which was investigated by perfusing labeled primary monocytes and T cells from blood through the vasculature, and quantifying the number of cells bound and retained in the tissue-chip model.


This work represents a new tumor-on-chip model using autologous cells and tissue in combination with engineered vasculature. T cells were perfused through the system to image extravasation and migration toward tumor spheroids. We found that CAF promoted neovascularization and provoked vascular leakiness, likely through the production of secreted factors including VEGF and CCL5, which ultimately resulted in the recruitment of immune cells. Future studies may determine if these immune cells subsequently become polarized to tumor-supportive phenotypes such as regulatory T cells and tumor associated macrophages. This model of the tumor microenvironment will advance our understanding of how the tumor microenvironment drives growth, metastasis, and resistance to therapy at the individual patient level. Furthermore, the creation of an immunocompetent model that supports the incorporation of T cells without non-specific activation allows us to perform precision immunotherapy testing with in vitro technology.


Acknowledgements (Optional): : This work was supported by K00CA212227, U01CA214381 and U54CA261694.

References (Optional): :