Engineering the tumor microenvironment of colorectal cancer

Abstract

The tumor microenvironment contributes significantly to the clinical prognosis, especially in CRC. Infiltration by cells from the adaptive immune system (e.g. CD8) has been linked to a better survival independently of TNM-stage. The complex tumor microenvironment is not well mirrored by conventional in vitro tumor assays where response to drug or treatment is evaluated on tumor cell lines monolayers. In has been proposed that 3D tumor models could at least to some extent better represent the in-vivo tumor microenvironment. In this work we aimed therefore not only at the characterization of the CRC- microenvironment in regard to innate and adaptive immunity but also to engineer innovative in vitro models of CRC better reflecting in-vivo tumor response to treatment. Using tumor microarray (TMA) technology we found that an infiltration by cells with neutrophil phenotype expressing high MPO positivity was associated with a favorable prognostic effect and correlated with CD8 infiltration. As tumoricidal activity of CD8-T-cells has been described to be influenced by immunological checkpoints engagement we evaluated in another TMA study the expression of PD-L1 on colorectal cancer specimens. PD-L1 expression on CRC correlated with CD8 infiltration but, interestingly, was associated with a good survival. This intriguing role of the microenvironment composition on the survival of colorectal cancer patients lead us then to start working on an innovative perfused 3D tumor model to better mimic the complexity of cancer tissue in-vivo. By using cancer cell lines we could establish a model, which had phenotypical but as well functional properties similar to xenografted tumors. Most importantly the response to standard treatments, were also similar. As this perfused tumor model only partially mimics the complex tumor microenvironment, an ultimate goal remains the culture of primary colorectal cancer specimens. By using the perfused u-tube-bioreactor and adaption of protocol using collagen-sandwich we could with a high efficiency keep tumor tissue alive and proliferating. In our studies on CRC microenvironment we could observe that activated neutrophils, somehow unknown and unappreciated players in tumor immunology, contributed significantly to a better survival. On the other hand the immune-regulatory checkpoint ligand PD-L1 has been associated with improved survival in CRC in contrast to melanoma or breast-cancer. This could be regarded as marker of an ongoing-immune response because we observed a correlation between CD8 cells and IFN-γ gene expression. Engineering the complex tumor microenvironment by a perfused 3D bioreactor able to generate large tumor-tissues in short time could be helpful as tool for further phenotypical and functional screenings. Additionally, we could observe a significant impact of perfusion on survival and proliferation of tumor cells within primary tumor fragments.