A Transcriptional Link between HER2, JAM-A and FOXA1 in Breast Cancer

Abstract

Overexpression of the human epidermal growth factor receptor-2 (HER2) is associated with aggressive disease in breast and certain other cancers. At a cellular level, the adhesion protein Junctional Adhesion Molecule-A (JAM-A) has been reported to regulate the expression of HER3 via a transcriptional pathway involving FOXA1. Since FOXA1 is also a suggested transcription factor for HER2, this study set out to determine if JAM-A regulates HER2 expression via a similar mechanism. An integrated tripartite approach was taken, involving cellular expression studies after targeted disruption of individual players in the putative pathway, in silico identification of relevant HER2 promoter regions and, finally, interrogation of cancer patient survival databases to deconstruct functionally important links between HER2, JAM-A and FOXA1 gene expression. The outcome of these investigations revealed a unidirectional pathway in which JAM-A expression transcriptionally regulates that of HER2 by influencing the binding of FOXA1 to a specific site in the HER2 gene promoter. Moreover, a correlation between JAM-A and HER2 gene expression was identified in 75% of a sample of 40 cancer types from The Cancer Genome Atlas, and coincident high mean mRNA expression of JAM-A, HER2 and FOXA1 was associated with poorer survival outcomes in HER2-positive (but not HER2-negative) patients with either breast or gastric tumors. These investigations provide the first evidence of a transcriptional pathway linking JAM-A, HER2 and FOXA1 in cancer settings, and support potential future pharmacological targeting of JAM-A as an upstream regulator of HER2.

Keywords: FOXA1; HER2; JAM-A; bioinformatics; breast cancer; gastric cancer; patients; survival; tight junction; transcription factor.

Figure 1

JAM-A and HER2 gene expression positively correlates in multiple cancer types and their high mean expression is associated with poor survival in HER2-positive breast cancer patients. (A) Co-expression of the JAM-A (F11R) and HER2 (ERBB2) genes was analyzed in multiple cancer types using the web resource TIMER 2.0: Tumor IMmune Estimation Resource (http://timer.cistrome.org/ [29,30], accessed on 3 December 2021), which incorporates > 10,000 patient samples across 40 cancer types from The Cancer Genome Atlas (TCGA) database. The heatmap represents the purity-adjusted partial Spearman’s rho value as the degree of gene expressional correlation between F11R and ERBB2 in various cancer types. (B–D) Mean mRNA expression of the JAM-A (F11R) and HER2 (ERBB2) genes in breast cancer patients was correlated with distant metastasis-free survival (DMFS) in breast cancer patient samples using the online resource kmplot.com [31], (accessed on 2 December 2021) using only JetSet probes. The cutoff between high versus low expression was auto-selected by the online tool.

Figure 2

JAM-A gene silencing reduces HER2 mRNA and protein expression in HER2-positive breast cancer cell lines. SK-BR-3 and MCF7-HER2 cells were transfected with control siRNA (siNEG) or a pool of JAM-A siRNA (siJAM1 and siJAM2 combined) and harvested 72 h later for qRT-PCR and immunoblot analysis. Results show qRT-PCR analysis of JAM-A and HER2 mRNA expression after JAM-A silencing in (A) MCF7-HER2 and (C) SKBR3 cells. Representative Western blot images and densitometric analysis for HER2 and JAM-A protein normalized to actin expression after JAM-A knockdown in (B) MCF7-HER2 and (D) SKBR3 cells. Experiments were performed three times and data represent mean ± s.e.m, compared using two-tailed, equal variance Student’s t-tests, ** p < 0.01, *** p < 0.001.

Figure 2

JAM-A gene silencing reduces HER2 mRNA and protein expression in HER2-positive breast cancer cell lines. SK-BR-3 and MCF7-HER2 cells were transfected with control siRNA (siNEG) or a pool of JAM-A siRNA (siJAM1 and siJAM2 combined) and harvested 72 h later for qRT-PCR and immunoblot analysis. Results show qRT-PCR analysis of JAM-A and HER2 mRNA expression after JAM-A silencing in (A) MCF7-HER2 and (C) SKBR3 cells. Representative Western blot images and densitometric analysis for HER2 and JAM-A protein normalized to actin expression after JAM-A knockdown in (B) MCF7-HER2 and (D) SKBR3 cells. Experiments were performed three times and data represent mean ± s.e.m, compared using two-tailed, equal variance Student’s t-tests, ** p < 0.01, *** p < 0.001.

Figure 3

JAM-A does not regulate the HER2 transcriptional repressor YBX3/ZONAB in SK-BR-3 cells. (A,B) SKBR3 cells were transfected with control siRNA (siNEG) or a pool of JAM-A siRNA and harvested 72 h later for qRT-PCR (A) and immunoblot/densitometric (B) analysis of three independent experiments. Data represent mean ± s.e.m, compared using two-tailed, equal variance Student’s t-tests, *** p < 0.001. (C) SK-BR-3 breast cancer cells were transfected with 25 nM of control siRNA or a pool of JAM-A siRNA. Cells were harvested 72 h later and non-nuclear/nuclear fractions were prepared for Western blot analysis. The two bands for ZONAB reflect parallel detection of its long form and an alternately spliced shorter form. (D) SKBR3 cells were transfected with control siRNA or ZONAB Smartpool siRNA. After 72 h, RNA was extracted for RT-qPCR analysis. Experiments were performed three times and data represent mean ± s.e.m, compared using two-tailed, equal variance Student’s t-tests, *** p < 0.001.

Figure 4

FOXA1 knockdown reduces HER2 mRNA and protein expression in breast cancer cells. SK-BR-3 and MCF7-HER2 cells were transfected with control siNEG or FOXA1 Smartpool siRNA. RNA and protein were extracted for qRT-PCR and immunoblot analysis. Results show qRT-PCR analysis of HER2 mRNA expression after FOXA1 knockdown in (A) SK-BR-3 and (B) MCF7-HER2 cells. Representative Western blot images and densitometric analysis for FOXA1 and HER2 protein normalized to actin expression after FOXA1 knockdown in (C) SK-BR-3 and (D) MCF7-HER2 cells. Experiments were performed three times and data represent mean ± s.e.m, compared using equal variance, unpaired Student’s t-tests, *** p < 0.001.

Figure 5

FOXA1 binds to a specific sequence within the HER2 gene promoter and its binding activity is influenced by JAM-A expression. (A) The ENCODE project (https://www.encodeproject.org/, accessed on 15 December 2021), TRANSFAC and the TFFFIND search tool from the Piptools package were used to identify segments of the HER2 proximal gene promoter in breast cancer cells that contained binding sequences for FOXA1. Four such sequences were identified (sequence logo for the FOXA1 consensus motif shown, accessed on 15 December 2021), of which two were synthesized as oligonucleotides for binding assays. (B) Nuclear proteins were extracted from untreated MCF7-HER2 cells and those transfected with non-targeting control siRNA or JAM-A siRNA, and used in protein–DNA binding assays. Nuclear protein binding activity to two different oligonucleotide sequences representing the HER2 gene promoter. (C) Relative optical densities of FOXA1 binding to the HER2 gene promoter in JAM-A knockdown compared to control siNEG conditions. Experiments were performed three times and data represent mean ± s.e.m, compared using equal variance, unpaired Student’s t-tests, * p < 0.05, *** p < 0.001.

Figure 5

FOXA1 binds to a specific sequence within the HER2 gene promoter and its binding activity is influenced by JAM-A expression. (A) The ENCODE project (https://www.encodeproject.org/, accessed on 15 December 2021), TRANSFAC and the TFFFIND search tool from the Piptools package were used to identify segments of the HER2 proximal gene promoter in breast cancer cells that contained binding sequences for FOXA1. Four such sequences were identified (sequence logo for the FOXA1 consensus motif shown, accessed on 15 December 2021), of which two were synthesized as oligonucleotides for binding assays. (B) Nuclear proteins were extracted from untreated MCF7-HER2 cells and those transfected with non-targeting control siRNA or JAM-A siRNA, and used in protein–DNA binding assays. Nuclear protein binding activity to two different oligonucleotide sequences representing the HER2 gene promoter. (C) Relative optical densities of FOXA1 binding to the HER2 gene promoter in JAM-A knockdown compared to control siNEG conditions. Experiments were performed three times and data represent mean ± s.e.m, compared using equal variance, unpaired Student’s t-tests, * p < 0.05, *** p < 0.001.

Figure 6

FOXA1 or HER2 gene silencing do not affect JAM-A expression in breast cancer cells. MCF7-HER2 cells were transfected with 25 nM of control siNEG, FOXA1 or HER2 Smartpool siRNA, and RNA and protein extracted for qRT-PCR and immunoblot analysis. Representative Western blot images and densitometric analysis are shown for expression of individual proteins normalized to actin expression after FOXA1 or HER2 gene silencing. FOXA1 and JAM-A mRNA (A) and protein (B) expression in MCF7-HER2 cells after FOXA1 gene silencing. mRNA (C) and protein (D) expression of FOXA1 or JAM-A in MCF7-HER2 cells after HER2 gene silencing. mRNA (E) and protein (F) expression of FOXA1 or JAM-A in SK-BR-3 cells after gene silencing of HER2. All experiments were performed three times, and data represent mean ± s.e.m, compared using equal variance, unpaired Student’s t-tests, *** p < 0.001.

Figure 7

High average gene expression of JAM-A, HER2 and FOXA1 correlates with poorer survival prospects in HER2-positive but not HER2-negative breast cancer patients. Mean mRNA expression of the JAM-A (F11R), HER2 (ERBB2) and FOXA1 genes in breast cancer patients was correlated with distant metastasis-free survival (DMFS; A–C) or recurrence-free survival (RFS; D–F) in breast cancer patient samples using the online resource kmplot.com [31], (accessed on 8 December 2021) using only JetSet probes. The cutoff between high and low expression was automatically selected by the online tool.

Figure 8

Proposed model linking JAM-A and HER2 in breast cancer. High JAM-A expression in the cell membrane of certain cancer cells leads to upregulated expression of the transcription factor FOXA1, by a pathway potentially involving β-catenin (dotted line). FOXA1 is capable of translocating to the nucleus and turning on the transcription of HER2, whereupon a subsequent upregulation of HER2 protein expression would be predicted to drive tumorigenic signaling via pathways including the activation of AKT and ERK.