The role of the immune system in the peritoneal spreading of epithelial ovarial cancer

A specific characteristic of high grade serous ovarian cancer (HGSOC) is the local tumor spread in the peritoneum, contributing to the high mortality. The underlying mechanisms enabling peritoneal tumor spread seem to be fundamentally different than mechanisms involved in distant metastatis via the circulatory system. Current data suggests that in the former the immune system in the peritoneum is a key player, whereas in the latter the epithelial-mesenchymal transition (EMT) and vice-versa the mesenchymal-epithelial transition (MET) are important mechanisms. In this project we plan to molecularly define two clinically distinguishable types of peritoneal tumor spread in HGSOC: i) miliary, i.e. many small millet sized tumor implants, versus ii) non-miliary, i.e. few larger and bulky tumor implants (or even only retroperitoneal lymph node metastases). The main aim is the integrative analysis of multiple omics data (long RNAs, miRs, and exome) generated by the new gold standard method Next Generation Sequencing (RNA-seq and whole exome seq). A multitude of tissues (blood, ascites, ovarian and metastatic peritoneal tumor tissue) and tissue sub-fractions (immune cells as well as tumor cells) will be analyzed. Complementary, the cellular parts of the tissue sub-fractions will be assessed by flow cytometry and immunofluorescent staining; the subcellular fractions of blood and ascites by multiplexed Luminex cyto-/chemokine analysis and metabolomics. All these results will be integrated to characterize the two peritoneal tumor spread types, miliary and non-miliary. First preliminary data from tumor cells and cyto/chemokines revealed the following three characteristics for patients with miliary tumor spread compared to non-miliary: i) Tumor cells are more epithelial like. ii) The inflammatory status in the peritoneum seems to be more sedated. And, iii) patients showed a significant worse prognosis, even corrected for all relevant clinicopathologic characteristics. These data raise the question of alternative treatment strategies for these patients. The unique selling point of this project is the establishment of integrative analyses strategies typical for Systems Medicine approaches, combining omics and low dimensional data with the focus on biological network information. We address the clinically relevant question of differences between the two apparent tumor spread types possibly identifying new targets for specific treatment strategies. We think that this approach will be the future of translational research in cancer biology.

During this project the immunogenic impact on two different tumor spread types in high grade serous ovarian cancer (HGSOC), the most abundant ovarian cancer type with a low 5-year overall survival rate of about 30%, was analyzed. The first characterized by many small sized tumor nodules throughout the peritoneal cavity, called miliary, and the second characterized by only a few, large and exophytically growing tumor nodules. Tumor cells of the miliary type are more epithelial in their gene expression signature, derive more likely from the fallopian tube secretory cells and induce more local and systemic inflammation. Their more epithelial phenotype allows them to adhere more easily on the peritoneal wall and build the numerous millet sizes small tumor nodules. This so called peritoneal carcinosis is a very bad prognostic feature of HGSOC. Tumor cells of the non-miliary type are more mesenchymal in their gene expression signature, derive more likely from the ovarian surface epithelial cells, and show a more pronounced PD-L1 positive tumor cell indicated adaptive immune reaction. Higher PD-1 expression on tumor cells is correlated to a better prognosis in HGSOC cancer, contrary to many other cancer entities like malignant melanoma or lung cancer. The latter are characterized by a high mutational load and therefore also a high load on tumor specific antigens. In HGSOC high PD-1 expression seems to indicate an immune reaction against the tumor. Interestingly, low PD-1 expression is correlated to low expression of MHC-I genes, the complex responsible for the presentation of self-proteins on all cells. Thus, low expression of MHC-I indicates an immunogenic silent tumor but probably also a more malignant phenotype, correlating to unfavorable outcome. Nevertheless, our immune system is also primed to recognize cells with low MHC-I expression and this could be the cause for the more inflamed situation. Patients with a high number of tumor infiltrating leukocytes (TILs) with high PD-1 expression have a better prognosis and patients with a high number of TILs with high PD-L1 expression have an unfavorable prognosis. There is a correlation of the non-miliary tumor type and the activity of neutrophil immune cells, especially with one of their functions, the release of Neutrophil Extracellular Traps (NETs). Together with the release of NETs, S100A8 is released and we showed that a high ratio of S100A8 to the inflammatory marker, C-reactive protein, CRP, is associated with the non-miliary tumor type and with favorable survival. It was shown recently, that the NETs associated DNA (NET-DNA) is associated with cancer metastasis in mouse models, fitting to the fact that metastases of non-miliary tumors seem to be associated with blood or lymph vessels, whereas the small tumor nodules from the miliary type are typical local metastases, independent from the vessel systems.