Small RNAs and the competing endogenous RNA network in high grade serous ovarian cancer tumor spread

High grade serous ovarian cancer (HGSOC) is among the most deadly malignancies in women, frequently involving peritoneal tumor spread. Understanding molecular mechanisms of peritoneal metastasis is essential to develop urgently needed targeted therapies. We described two peritoneal tumor spread types in HGSOC apparent during surgery: miliary (numerous millet-sized implants) and non-miliary (few big, bulky implants). The former one is defined by a more epithelial-like tumor cell characteristic with less immune cell reactivity and with significant worse prognosis, even if corrected for typical clinicopathologic factors. 23 HGSOC patients were enrolled in this study. Isolated tumor cells from fresh tumor tissues of ovarian and peritoneal origin and from ascites were used for ribosomal RNA depleted RNA and small RNA sequencing. RT-qPCR was used to validate results and an independent cohort of 32 patients to validate the impact on survival. Large and small RNA sequencing data were integrated and a new gene-miRNA set analysis method was developed. Thousands of new small RNAs (miRNAs and piwi-interacting RNAs) were predicted and a 13 small RNA signature was developed to predict spread type from formalin-fixed paraffin-embedded tissues. Furthermore, integrative analyses of RNA sequencing and small RNA sequencing data revealed a global upregulation of the competing endogenous RNA network in tumor tissues of non-miliary compared to miliary spread, i.e. higher expression of circular RNAs and long non-coding RNAs compared to coding RNAs but unchanged abundance of small RNAs. This global deregulated expression pattern could be co-responsible for the spread characteristic, miliary or non-miliary, in ovarian cancer.