Gene regulatory pattern analysis reveals essential role of core transcriptional factors' activation in triple-negative breast cancer.

Li Min,Jiafei Liu,Chengchao Shou,Guo-Cheng Yuan,Cheng Zhang,Luca Pinello,Like Qu,Jialiang Huang,Lan Jiang

doi:10.18632/oncotarget.15749

Abstract

BackgroundTriple-negative breast cancer (TNBC) is an aggressive breast cancer subtype. Genome-scale molecular characteristics and regulatory mechanisms that distinguish TNBC from other subtypes remain incompletely characterized.ResultsBy combining gene expression analysis and PANDA network, we defined three different TF regulatory patterns. A core TNBC-Specific TF Activation Driven Pattern (TNBCac) was specifically identified in TNBC by computational analysis. The essentialness of core TFs (ZEB1, MZF1, SOX10) in TNBC was highlighted and validated by cell proliferation analysis. Furthermore, 13 out of 35 co-targeted genes were also validated to be targeted by ZEB1, MZF1 and SOX10 in TNBC cell lines by real-time quantitative PCR. In three breast cancer cohorts, non-TNBC patients could be stratified into two subgroups by the 35 co-targeted genes along with 5 TFs, and the subgroup that more resembled TNBC had a worse prognosis.MethodsWe constructed gene regulatory networks in breast cancer by Passing Attributes between Networks for Data Assimilation (PANDA). Co-regulatory modules were specifically identified in TNBC by computational analysis, while the essentialness of core translational factors (TF) in TNBC was highlighted and validated by in vitro experiments. Prognostic effects of different factors were measured by Log-rank test and displayed by Kaplan-Meier plots.ConclusionsWe identified a core co-regulatory module specifically existing in TNBC, which enabled subtype re-classification and provided a biologically feasible view of breast cancer.

Highlights

Breast cancer subtyping was widely used in clinical decisions, such as relapse risk evaluation and treatment selection [1, 2]
We identified a core co-regulatory module existing in triplenegative breast cancer (TNBC), which enabled subtype re-classification and provided a biologically feasible view of breast cancer
Expression data for 63 normal breast (NORM), 445 non-triple-negative breast cancer (nTNBC) and 89 TNBC tissue samples were extracted from The Cancer Genome Atlas (TCGA)

Summary

Introduction

Breast cancer subtyping was widely used in clinical decisions, such as relapse risk evaluation and treatment selection [1, 2]. Multiple molecular characteristics of TNBC have been well identified [8,9,10,11,12], most studies were conducted from the perspective of gene expression, which cannot reflect the whole scope of pathologic mechanisms www.impactjournals.com/oncotarget on gene regulation level, many questions of TNBC remain unanswered [13]. By incorporating multiple sources of data to model biological processes, especially transcriptional factor (TF) -gene regulatory networks, integrative analyses show promising perspective in comprehending of pathophysiologic mechanisms and developing novel and precise therapies [16, 17]. PANDA predicts TF-gene regulatory relationships by integrating information from protein-protein interaction (PPI), gene expression, and TF-sequence-motif data using a messagepassing approach, and it has been successfully used to study several diseases including Chronic Obstructive Pulmonary Disease (COPD) [23] and ovarian cancer [24]. Genome-scale molecular characteristics and regulatory mechanisms that distinguish TNBC from other subtypes remain incompletely characterized

Methods

Results

Conclusion