Katherine R Miller1, Jai N Patel1, Qing Zhang2, Eric J Norris1, James Symanowski2, Chad Michener3, Jalid Sehouli4, Ioana Braicu4, Darla D Destephanis1, Ashley P Sutker1, Wendell Jones5, Chad A Livasy6, Charles Biscotti7, Ram N Ganapathi8, David L Tait9, Mahrukh K Ganapathi10. 1. Department of Cancer Pharmacology, Levine Cancer Institute, Carolinas HealthCare System, Charlotte, NC, USA. 2. Department of Biostatistics, Levine Cancer Institute, Carolinas HealthCare System, Charlotte, NC, USA. 3. Women's Health and Obstetrics/Gynecology Institute, Cleveland Clinic, Cleveland, OH, USA. 4. Department of Gynecology, Charité Medical University of Berlin, Berlin, Germany. 5. Bioinformatics and Clinical Systems, Q(2) Solutions - EA Genomics, 5927 S. Miami Blvd., Suite 100, Morrisville, NC 27560, USA. 6. Carolinas Pathology Group, Carolinas HealthCare System, Charlotte, NC, USA. 7. Department of Anatomic Pathology, Cleveland Clinic, Cleveland, OH, USA. 8. Department of Cancer Pharmacology, Levine Cancer Institute, Carolinas HealthCare System, Charlotte, NC, USA. Electronic address: ram.ganapathi@carolinashealthcare.org. 9. Division of Gynecologic Oncology, Levine Cancer Institute, Carolinas HealthCare System, Charlotte, NC 28204, USA. Electronic address: david.tait@carolinashealthcare.org. 10. Department of Cancer Pharmacology, Levine Cancer Institute, Carolinas HealthCare System, Charlotte, NC, USA. Electronic address: mahrukh.ganapathi@carolinashealthcare.org.
Abstract
OBJECTIVES: Aberrant homeobox (HOX) gene expression is reported in high-grade serous ovarian carcinoma (HGSOC), however, its prognostic significance remains unclear. METHODS: HOX genes associated with progression-free survival (PFS) in a discovery cohort of primary HGSOC samples with RNA sequencing data, and those previously reported to be associated with clinical outcomes, were selected for qPCR testing in an independent training cohort of primary HGSOC samples (n=71). A prognostic model for PFS was developed using univariate and multivariate Cox regression. Patients were stratified into risk groups that optimized the test statistic. The model was tested in an independent HGSOC cohort from The Cancer Genome Atlas (TCGA) (n=320). The effect of selected HOX genes on drug sensitivity and reactive oxygen species (ROS) accumulation was examined in vitro. RESULTS: Of 23 HOX genes tested in the training cohort, HOXA4 (HR=1.20, 95% CI=1.07-1.34, P=0.002) and HOXB3 (HR=1.09, 95% CI=1.01-1.17, P=0.027) overexpression were significantly associated with shorter PFS in multivariate analysis. Based on the optimal cutoff of the HOXA4/HOXB3 risk score, median PFS was 16.9months (95% CI=14.6-21.2months) and not reached (>80months) for patients with high and low risk scores, respectively (HR=8.89, 95% CI=2.09-37.74, P<0.001). In TCGA, the HOXA4/HOXB3 risk score was significantly associated with disease-free survival (HR=1.44, 95% CI=1.00-2.09, P=0.048). HOXA4 or HOXB3 overexpression in ovarian cancer cells decreased sensitivity to cisplatin and attenuated the generation of cisplatin-induced ROS (P<0.05). CONCLUSIONS: HOXA4/HOXB3 gene expression-based risk score may be useful for prognostic risk stratification and warrants prospective validation in HGSOC patients.
OBJECTIVES: Aberrant homeobox (HOX) gene expression is reported in high-grade serous ovarian carcinoma (HGSOC), however, its prognostic significance remains unclear. METHODS: HOX genes associated with progression-free survival (PFS) in a discovery cohort of primary HGSOC samples with RNA sequencing data, and those previously reported to be associated with clinical outcomes, were selected for qPCR testing in an independent training cohort of primary HGSOC samples (n=71). A prognostic model for PFS was developed using univariate and multivariate Cox regression. Patients were stratified into risk groups that optimized the test statistic. The model was tested in an independent HGSOC cohort from The Cancer Genome Atlas (TCGA) (n=320). The effect of selected HOX genes on drug sensitivity and reactive oxygen species (ROS) accumulation was examined in vitro. RESULTS: Of 23 HOX genes tested in the training cohort, HOXA4 (HR=1.20, 95% CI=1.07-1.34, P=0.002) and HOXB3 (HR=1.09, 95% CI=1.01-1.17, P=0.027) overexpression were significantly associated with shorter PFS in multivariate analysis. Based on the optimal cutoff of the HOXA4/HOXB3 risk score, median PFS was 16.9months (95% CI=14.6-21.2months) and not reached (>80months) for patients with high and low risk scores, respectively (HR=8.89, 95% CI=2.09-37.74, P<0.001). In TCGA, the HOXA4/HOXB3 risk score was significantly associated with disease-free survival (HR=1.44, 95% CI=1.00-2.09, P=0.048). HOXA4 or HOXB3 overexpression in ovarian cancer cells decreased sensitivity to cisplatin and attenuated the generation of cisplatin-induced ROS (P<0.05). CONCLUSIONS:HOXA4/HOXB3 gene expression-based risk score may be useful for prognostic risk stratification and warrants prospective validation in HGSOC patients.
Authors: Sophie Elena McGrath; Nicola Annels; Thumuluru K Madhuri; Anil Tailor; Simon A Butler-Manuel; Richard Morgan; Hardev Pandha; Agnieszka Michael Journal: BMC Cancer Date: 2018-10-03 Impact factor: 4.430