Tumor clusters with divergent inflammation and human retroelement expression determine the clinical outcome of patients with serous ovarian cancer.

Abstract

High-grade serous ovarian carcinoma (HGSOC) associates with the worst patient outcome. Understanding the tumor environment in terms of quantifying endogenous retroviruses (ERVs) and LINE-1 expression and their correlations with inflammation genes, checkpoint inhibitors and patient survival is needed. Analysis of 102 treatment-naïve HGSOC and control tissues for ERVs, LINE-1, inflammation and immune checkpoints identified five clusters with diverse patient recurrence-free survivals. One cluster termed Triple-I with the best patient survival showed the highest number of tumor infiltrating lymphocytes along with 22 overexpressed genes, including CXCL9 and AIM2. However, Triple-I associated with the lowest ERV/LINE-1 expression. The tumor cluster with the second-best patient survival had both high ERV/LINE-1 expression and inflammation. Multiplex-immunohistochemistry revealed CD28 protein solely on immune cells, without co-expression of the inhibitory CTLA4 receptor. The largest tumor cluster with high ERV/LINE-1 expression but low inflammation showed a significant low gene expression of the dsRNA sensors MDA5 and RIG-I supporting an aberrant block in IFN signaling. Our study represents an intrinsic 'molecular and immunological snapshot' of the HGSOC tumor environment important for understanding retroelements and inflammation for clinical relevance.