Machine learning developed a CD8+ exhausted T cells signature for predicting prognosis, immune infiltration and drug sensitivity in ovarian cancer

Abstract

Abstract CD8+ exhausted T cells (CD8+ Tex) played a vital role in the progression and therapeutic response of cancer. However, few studies have fully clarified the characters of CD8+ Tex related genes in ovarian cancer (OC). The CD8+ Tex related prognostic signature (TRPS) was constructed with integrative machine learning procedure including 10 methods using TCGA, GSE14764, GSE26193, GSE26712, GSE63885 and GSE140082 dataset. Several immunotherapy benefits indicators, including Tumor Immune Dysfunction and Exclusion (TIDE) score, immunophenoscore (IPS), TMB score and tumor escape score, were used to explore performance of TRPS in predicting immunotherapy benefits of OC. The TRPS constructed by Enet (alpha = 0.3) method acted as an independent risk factor for OC and showed stable and powerful performance in predicting clinical outcome of patients. The C-index of the TRPS was higher than that of tumor grade, clinical stage, and many developed signatures. Low TRPS score indicated a higher level of CD8+ T cell, B cell, macrophage M1, and NK cells, representing a relative immunoactivated ecosystem in OC. OC patients with low risk score had a higher PD1&CTLA4 immunophenoscore, higher TMB score, lower TIDE score and lower tumor escape score, suggesting a better immunotherapy response. Moreover, higher TRPS score indicated a higher score of cancer-related hallmarks, including angiogenesis, EMT, hypoxia, glycolysis, and notch signaling. Vitro experiment showed that ARL6IP5 was downregulated in OC tissues and inhibited tumor cell proliferation. The current study constructed a novel TRPS for OC, which could serve as an indicator for predicting the prognosis, immune infiltration and immunotherapy benefits for OC patients.

Keywords

• CD8+ Tex
• Machine learning
• Ovarian cancer
• Prognostic signature
• Immunotherapy