Exploring novel roles of tumor pericytes

Sammanfattning: Tumor biology has been extensively studied over the last few decades, with a principal focus on how neoplastic cells obtain cellular immortality. A number of oncogenes and tumor suppressor genes have been uncovered that regulate cancerous transformation. However, tumors are now believed to be complex tissues consisting of various kinds of tumor stromal cells as well as transformed cancer cells. The tumor stroma is mainly comprised of tumorinfiltrating leukocytes, cancer-associated fibroblasts, vascular endothelial cells, lymphatic endothelial cells, tumor pericytes, and extracellular matrix. These cells have an extensive interplay with one another or with cancer cells, from the initiation stage of tumor development to its metastatic dissemination. The aim of this thesis was to investigate the tumor pericytes—one of the tumor stroma constituents that has not been widely explored—in order to determine their novel roles in tumor malignancy. The use of a pericyte-deficient mouse model (pdgfbret/ret), in paper I, confirmed that the myeloid-derived suppressor cells (MDSCs), one of the most aggressive types oftumor-infiltrating leukocytes, are significantly increased both at the tumor site and in the peripheral blood in B16 melanoma and Lewis lung carcinoma (LLC) subcutaneous mouse models ofpdgfbret/ret, compared to their littermate controls. The increase in the MDSC number was dependent on expression oftumor-derived IL-6, induced by the hypoxic tumor microenvironment in pericyte-deficient B16 and LLC tumors. Analysis of gene expression in human samples (253 breast cancer patients of an Uppsala dataset) showed an inverse correlation between human pericyte-related genes and human MDSC markers and a subsequent relevance to the survival rate ofbreast cancer patients. The relevance of the tumor pericytes to other tumor stroma cells was studied in paper II, which revealed a comparable abundance of PDGFRα- expressing perivascular cells in pericyte-deficient B16 melanomas. A further investigation identified the PDGFRα-expressing perivascular cells as “specialized myofibroblasts” with gene signature features of both fibroblast-related (Fap, Pdgfra, Hgf) and pericyte-related (Cspg4, Pdgfrb, Asma) gene sets. Moreover, pericyte-deficient, B16-melanoma-bearing mice showed an elevated level of the serum S100B protein, which is widely considered to be a distinctive prognostic marker for malignant melanoma patients. The B16 melanoma tumor cell-derived S100B was then confirmed to pass into the peripheral blood. Presumably “activated endothelial cells” in pdgfbret/ret mice would facilitate more rapid transport of S100B protein via endothelial caveola-mediated transcytosis. In conclusion, tumor pericytes directly interact with adjacent endothelial cells, thereby controlling the tumor vasculature and further changing the tumor microenvironment. Tumor pericytes favor tumor immunogenicity by blocking systemic MDSC bursts in experimental mouse models (B16, LLC) and are also negatively involved in recruitment of the perivascular myofibroblasts. However, the biological relevance ofthe perivascular myofibroblasts should be further investigated.