Abstract 5708: HOXA4 and HOXB3 gene expression signature as a biomarker of tumor recurrence in patients with high-grade serous ovarian cancer (HGSOC) following primary cytoreductive surgery and first-line adjuvant chemotherapy

Abstract

Abstract Background: Aberrant HOX gene expression has been observed in various malignancies; however, their role as a biomarker for recurrence in patients with HGSOC is unknown. We tested the hypothesis that a HOX gene signature is a biomarker for disease-free survival (DFS) in HGSOC patients following primary cytoreductive surgery and adjuvant platinum-based chemotherapy. Methods: A panel of HOX genes associated with DFS in a discovery cohort of 19 HGSOC patients with available RNA-Seq data, and/or previously reported as associated with survival endpoints in the literature, were selected for testing by qPCR using the Fluidigm platform in an independent training cohort of primary tumors from 73 HGSOC patients. Cox proportional hazards model was used to identify HOX genes significantly associated with DFS. A prognostic gene signature was developed based on the combined linear predictor from the final multivariable Cox model. Patients were stratified into risk groups using the optimal cutoff of the linear predictor risk score, whereby the optimal cutoff was defined as the value that maximized the log-rank test statistic. Risk group stratification was further tested for association with DFS in a larger HGSOC dataset (N=414) with similar clinical characteristics from The Cancer Genome Atlas (TCGA). The role of the identified HOX genes in influencing drug sensitivity was examined by overexpressing these genes in a HGSOC cell line, PEO1. Results: Of 23 HOX genes selected for testing in the training cohort, overexpression of HOXA4 (HR=1.21, 95% CI=1.08-1.35, P=0.001) and HOXB3 (HR=1.09, 95% CI=1.02-1.17, P=0.016) were significantly associated with decreased DFS in multivariate analysis. The median DFS in patients with a HOXA4/HOXB3 risk score ≤ the optimal cutoff was > 80 months (not yet reached), whereas the median DFS in patients with a HOXA4/HOXB3 risk score > the optimal cutoff was 16.9 months (HR=7.85, 95% CI=4.19-14.69, P<0.001). The HOXA4/HOXB3 risk score remained significantly associated with DFS in the HGSOC dataset from the TCGA (P=0.02). In vitro studies demonstrated that PEO1 cells overexpressing either HOXA4 or HOXB3 were significantly (P<0.05) more resistant to cisplatin and carboplatin, but not to paclitaxel, as compared to empty vector transfected cells. Conclusion: The HOXA4/HOXB3 risk score was significantly associated with DFS in a training cohort of HGSOC patients, and remained significantly associated in a larger independent HGSOC dataset from the TCGA. Further, in vitro studies suggest that HOXA4 and HOXB3 overexpression is associated with platinum resistance. Thus, overexpression of HOXA4 and HOXB3 in primary HGSOC may serve as a potential biomarker for recurrence and the risk score model developed warrants prospective validation in HGSOC patients. Citation Format: Jai N. Patel, Katherine Miller, James Symanowski, Jalid Sehouli, Chad Michener, Ioana Braicu, Darla Destephanis, Ashley P. Sutker, Eric J. Norris, David Tait, Wendell Jones, Qing Zhang, Chad Livasy, Charles Biscotti, Ram N. Ganapathi, Mahrukh K. Ganapathi. HOXA4 and HOXB3 gene expression signature as a biomarker of tumor recurrence in patients with high-grade serous ovarian cancer (HGSOC) following primary cytoreductive surgery and first-line adjuvant chemotherapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 5708. doi:10.1158/1538-7445.AM2017-5708