Cross-regulation between FOXA1 and ErbB2 signaling in estrogen receptor-negative breast cancer

Abstract

Molecular apocrine is a subtype of estrogen receptor-negative (ER.) breast cancer, which is characterized by a steroid-response gene signature that includes androgen receptor, FOXA1, and a high frequency of ErbB2 overexpression. In this study, we demonstrate that there is a strong association between the overexpression of FOXA1 and ErbB2 in ER- breast tumors. This has led us to identify a cross-regulation network between FOXA1 and ErbB2 signaling in ER- breast cancer. We present two mechanisms to explain the association between FOXA1 and ErbB2 overexpression in molecular apocrine cells. In one process, ErbB2 signaling genes CREB1 and c-Fos regulate FOXA1 transcription, and in another process, AP2α regulates the expression of both FOXA1 and ErbB2. Moreover, we demonstrate that FOXA1, in turn, regulates the transcription of ErbB2 signaling genes. This includes a core gene signature that is shared across two molecular apocrine cell lines. Importantly, the most upregulated (RELB) and downregulated (PAK1) genes in this signature are direct FOXA1 targets. Our data suggest that FOXA1 acts as a dual-function transcription factor and the repressive function of FOXA1 on RELB can be explained by the recruitment of its binding partner corepressor TLE3. It is notable that a group of FOXA1-regulated genes vary across molecular apocrine cell lines leading to the differences in the functional effects of FOXA1 on extracellular signal-regulated kinase phosphorylation and cell viability between these lines. This study demonstrates that there is a cross-regulation network between FOXA1 and ErbB2 signaling that connects FOXA1 to some of the key signaling pathways in ER-breast cancer.

Figure 1

The association between FOXA1 and ErbB2 expression in ER-negative breast tumors and cell lines. FOXA1 IHC in a breast tumor with ErbB2 overexpression (ErbB2+; A) and in an ErbB2- breast tumor (B). Magnifications, x10. (C) Percentage of FOXA1+ staining using IHC in ErbB2+/ER- and ErbB2-/ER- breast tumors. *P < .01 is for ErbB2+ versus ErbB2- groups. Error bars, ±2 SEM. (D) Receiver operating characteristic analysis to test ErbB2 status as a predictor for FOXA1 overexpression (FOXA1+) in ER- breast tumors. Dashed line is a diagonal reference line. (E) Scatterplot to demonstrate the correlation between the relative expression of FOXA1 and ErbB2 in ER- cell lines. Regression fitted line using the quadratic method is depicted. (F) Scatterplot to demonstrate the correlation between relative expression of FOXA1 and AP2a as described in E.

Figure 2

The regulation of FOXA1 expression by ErbB2- ERK signaling. (A) Relative expression of FOXA1 using RT-PCR in HCC-1954 cell line after treatment with CI-1040 (MEK Inh.) at 10 µM concentration for 24 hours. Expression is relative to that of vehicle-treated cells (CTL). *P < .01 is for CI-1040 versus control. (B) FOXA1 protein levels using Western blot analysis in MDA-MB-453 cell line after treatment with CI-1040 as described in A. (C) Putative transcription factor binding sites for CREB1, c-Jun, c-Fos, AP2α, and Elk1 in 1-kb promoter region of FOXA1. P1 (primer set 1) and P2 (primer set 2) are regions of amplification for ChIP assays. (D) Luciferase reporter assay. The transcriptional activation of FOXA1 promoter by CREB1, c-Jun, c-Fos, AP2α, and Elk1 expression constructs was assessed using dual-luciferase assays in MCF-7 cells, and relative response ratios are reported. Cotransfection with the FOXA1 reporter vector and an empty pcDNA vector was used as a control. *Compared with the control group. (E) ChIP assay using CREB1, AP2α, and c-Fos antibodies and primer set 1 in MDA-MB-453 cell line. The results of end point RT-PCR amplification for ChIP assay are demonstrated with primer set 1 for FOXA1 promoter. Input indicates input chromatin at a dilution of 1:50; Neg. CT, nonspecific antibody. The relative copy number changes to control are shown as -Log2 values. *P < .01 is for each antibody versus Neg. CT. (F) ChIP assay using CREB1, AP2α, and c-Fos antibodies and primer set 2 as described in E. All error bars, ±2 SEM.

Figure 3

FOXA1-regulated gene signature in ErbB2 signaling. (A) Western blot analysis to show FOXA1 protein level after FOXA1-KD (FOX-KD) in MDA-MB-453 and HCC-1954 cell lines. Fold changes (RR) in band densities were measured relative to the control (CTL). (B) RT-PCR to demonstrate FOXA1-KD efficiencies in MDA-MB-453 and HCC-1954 cell lines. FOXA1 expression after KD was assessed relative to nontargeting siRNA control (CTL), and fold change is shown for each cell line and replicate experiment. Rep1 indicates replicate 1; Rep2, replicate 2. (C) Heat map of FOXA1-regulated gene signature in ErbB2 signaling generated using RT-PCR data. Heat map is constituted of genes with more than two-fold change in expression after FOXA1-KD (KD) in both MDA-MB-453 and HCC-1954 cell lines. Mean fold change in expression is shown for each gene. Red and green depict up-regulation and down-regulation, respectively. Bar indicates the range of fold changes in gene expression.

Figure 4

ErbB2 signaling genes that are differentially regulated by FOXA1 between molecular cell lines. (A) RT-PCR to demonstrate the relative folds of gene expression to control after FOXA1-KD in MDA-MB-453 cell line. Box plots show genes with more than two-fold change in expression after FOXA1-KD, which were unique to MDA-MB-453 cell line. (B) RT-PCR to demonstrate the relative folds of gene expression to control after FOXA1-KD in HCC-1954 cell line. Box plots to show genes with more than two-fold change in expression after FOXA1-KD, which were unique to HCC-1954 cell line.

Figure 5

The effect of FOXA1 KD on ERK phosphorylation and cell viability. (A) Western blot analysis to measure the level of phosphorylated ERK (ph-ERK), total ERK, and ErbB2 after FOXA1-KD (FOX-KD) in MDA-MB-453 cell line. Fold change was assessed relative to nontargeting siRNA control (CTL). (B) MTT assay to measure cell viability after FOX-KD in MDA-MB-453 cell line. CTL indicates nontargeting siRNA control. *P < .03 is for FOX-KD versus CTL. (C) Western blot analysis to measure the level of phosphorylated ERK (ph-ERK) and total ERK after FOX-KD in HCC-1954 cell line. (D) MTT assay to measure cell viability after FOX-KD in HCC-1954 cell line. All error bars, ±2 SEM.

Figure 6

FOXA1 binding to RELB and PAK1 promoters and TLE3 regulation of RELB. (A) Putative FOXA1 binding sites in 1 kb promoter regions of RELB and PAK1. P1 (primer set 1) and P2 (primer set 2) are regions of amplification for ChIP assays. (B) ChIP assay for RELB promoter using FOXA1 antibody in the MDA-MB-453 cell line. The results of end point RT-PCR amplification for ChIP assay are shown with two sets of primers for RELB promoter. Input indicates input at a dilution of 1:50; Neg. CT, nonspecific antibody. The relative copy number changes to control are shown as -Log2 values. *P < .01 is for FOXA1 Ab versus Neg. CT. (C) ChIP assay for PAK1 promoter using FOXA1 antibody. The results of end point RT-PCR amplification for ChIP assay are shown with two sets of primers for PAK1 promoter as described in B. (D) RT-PCR to demonstrate TLE3-KD efficiencies with two sets of siRNA duplexes (KD1 and KD2) in MDA-MB-453 cell line. TLE3 expression after KD was assessed relative to nontargeting siRNA control (CTL), and the fold change is shown for each duplex. (E) RT-PCR to show RELB expression levels after TLE3-KD with two sets of siRNA duplexes as described in D. (F) ChIP assay for RELB promoter using TLE3 antibody. The results of end point RT-PCR amplification for ChIP assay are shown with two sets of primers for RELB promoter as described in B. *P < .01 is for TLE3 Ab versus Neg. CT. All error bars, ±2 SEM.

Figure 7

Schematic diagram of a cross-regulation network between FOXA1 and the ErbB2 signaling pathway in ER-negative breast cancer. Red and black arrows denote stimulatory and inhibitory effects, respectively.