Research Papers:
Six novel immunoglobulin genes for breast cancer recurrence identified by gene coexpression network analysis
Huan-Ming Hsu1,2, Jyh-Cherng Yu3, Chien-Ting Chen4, Chen-En Jian4, Chia-Yi Lee4, Yu-Jia Chang5, Chi-Ming Chu4 and Yu-Tien Chang4,6
1Department of Surgery, Songshan Branch of Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan
2Division of General Surgery, Department of Surgery, Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan
3Division of General Surgery, Department of Surgery, Tri-Services General Hospital, National Defense Medical Center, Taipei, Taiwan
4Division of Biostatistics and Informatics, Department of Epidemiology, School of Public Health, National Defense Medical Center, Taipei, Taiwan
5Graduate Institute of Clinical Medicine, College of Medicine, Taipei Medical University, Taipei, Taiwan
6Graduate of Medical Sciences, National Defense Medical Center, Taipei, Taiwan
Correspondence to:
Chi-Ming Chu, email: [email protected]
Yu-Tien Chang, email: [email protected]
Keywords: co-expression analysis; co-expression gene networks; integrated microarray analysis; NCBI GEO; breast cancer
Received: August 20, 2017 Accepted: March 19, 2018 Published:
ABSTRACT
Purpose: Gene co-expression network analysis (GCNA) can detect alterations in regulatory activities in case/control comparisons. We propose a framework to detect novel genes and networks for predicting breast cancer recurrence.
Results: We applied the framework to four comparison groups according to node (+/–) and recurrence (+/–). We identified a sub-network containing two candidate genes (LST1 and IGHM) and six novel genes (IGHA1, IGHD, IGHG1, IGHG3, IGLC2, and IGLJ3) related to B cell-specific immunoglobulin. These novel genes were correlated with recurrence under the control of node status and were found to function as tumor suppressors; higher mRNA expression indicated a lower risk of recurrence (hazard ratio (HR) = 0.87, p = 0.001). We created an immune index score by performing principle component analysis and divided the genes into low and high groups. This discrete index significantly predicted recurrence-free survival (RFS, high: HR = 0.77 p = 0.019, low: control). Using the public tool KM Plotter, we found that all of the genes except IGHD were significantly correlated with RFS and distant metastasis-free survival (DMFS) in triple-negative breast cancer (TNBC).
Materials and Methods: Thirty-four prognosis candidate genes were selected based on a literature review. Four Gene Expression Omnibus Series (GSE) microarray datasets (n = 920) were used to create gene co-expression networks based on these candidates.
Conclusions: We identified and validated six genes related to immune function as potential biomarkers of recurrence for both general breast cancer and TNBC. Our results suggest that GCNA can effectively and efficiently detect novel prognostic biomarkers of breast cancer.