Genome-Wide Profiling of AP-1 Regulated Transcription Provides Insights Into the Invasiveness of Triple-Negative Breast Cancer

Home , NFATC2, Transactivation, Transcription factor, ZEB1

Author Manuscript Published OnlineFirst on May 15, 2014; DOI: 10.1158/0008-5472.CAN-13-3396 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

Genome-wide profiling of AP-1 regulated transcription provides insights into the invasiveness of triple-negative breast cancer

Chunyan Zhao1, Yichun Qiao1, Philip Jonsson2, Jian Wang3, Li Xu1, Pegah Rouhi3,

1 3,4,5 2 1,6 Indranil Sinha , Yihai Cao , Cecilia Williams , and Karin Dahlman-Wright

1Department of Biosciences and Nutrition, Novum, Karolinska Institute, S-141 83 Huddinge,

Sweden; 2Center for Nuclear Receptors and Cell Signaling, Department of Biology and

Biochemistry, University of Houston, Houston, TX 77204-5056, USA; 3Department of

Microbiology, Tumor and Cell Biology, Karolinska Institute, 171 77 Stockholm, Sweden;

4Department of Medicine and Health Sciences, Linköping University, 581 83 Linköping,

Sweden; 5Department of Cardiovascular Sciences, University of Leicester, Glenfield Hospital,

Leicester, LE3 9QP, UK; 6Science for Life Laboratory, Karolinska Institute, Solna, Sweden

C. Zhao and Y. Qiao contributed equally to this work.

Corresponding Authors:

Chunyan Zhao, Department of Biosciences and Nutrition, Novum, Karolinska Institute, S-141

83 Huddinge, Sweden. Phone: 46-8-52481126; Fax: 46-8-52481130; E-mail: [email protected]; and Karin Dahlman-Wright, Department of Biosciences and Nutrition,

Novum, Karolinska Institute, S-141 83 Huddinge, Sweden. E-mail: karin.dahlman- [email protected]

Running Title: AP-1 cistrome in triple-negative breast cancer

Keywords: AP-1/ triple-negative breast cancer/cistrome/transcriptome

Word count: 4836

Total number of figures and tables: 7 figures

Disclosure: The authors declare no conflicts of interest. 1

Downloaded from cancerres.aacrjournals.org on October 1, 2021. © 2014 American Association for Cancer Research. Author Manuscript Published OnlineFirst on May 15, 2014; DOI: 10.1158/0008-5472.CAN-13-3396 Author manuscripts have been peer reviewed and accepted for publication but have not yet been edited.

Abstract

Triple-negative breast cancer (TNBC) is an aggressive clinical subtype accounting for up to

20% of all breast cancers, but its malignant determinants remain largely undefined. Here we show that in TNBC the overexpression of Fra-1, a component of the transcription factor AP-1, offers prognostic potential. Fra-1 depletion or its heterodimeric partner c-Jun inhibits the proliferative and invasive phenotypes of TNBC cells in vitro. Similarly, RNAi-mediated attenuation of Fra-1 or c-Jun reduced cellular invasion in vivo in a zebrafish tumor xenograft model. Exploring the AP-1 cistrome and the AP-1-regulated transcriptome, we obtained insights into the transcriptional regulatory networks of AP-1 in TNBC cells. Among the direct targets identified for Fra-1/c-Jun involved in proliferation, adhesion and cell-cell contact, we found that AP-1 repressed expression of E-cadherin by transcriptional upregulation of ZEB2 to stimulate cell invasion. Overall, this work illuminates the pathways through which TNBC cells acquire invasive and proliferative properties.

Introduction

Breast cancer is a biologically and clinically heterogeneous disease. Studies of breast cancers using gene expression profiling have identified several major molecular subtypes, including luminal A, luminal B, HER2 (also known as ERBB2)-overexpressing, normal breast tissue- like, and basal-like (1). In contrast, triple-negative breast cancers (TNBCs) are defined merely by a lack of expression of estrogen receptor (ER) and progesterone receptor (PR) as well as the absence of HER2 overexpression (2, 3). There is substantial overlap between TNBC and the basal-like subtype of breast cancer, approximately 80% of TNBCs are basal-like, and therefore the two terms describe a broadly similar group of cancers (3). TNBCs account for some of the most aggressive types of breast cancers, marked by high rates of relapse, visceral metastases and poor prognosis (2). Clinical advances have led to significant progress in the development of targeted therapies for patients with ER-positive or HER2-positive diseases, for example, endocrine therapies or HER2-directed agents such as trastuzumab (4). However,

TNBC lacks a targeted therapy, has a poor prognosis and represents a major unmet medical need.

The AP-1 transcription factor is a key component of many signal transduction pathways. The AP-1 family of transcription factors consists of dimeric complexes of either homodimers of Jun family members (c-Jun, JunB and JunD) or heterodimers of Jun and Fos family members (c-Fos, FosB, Fra-1 and Fra-2). AP-1 has been shown to regulate the expression of AP-1 target genes by binding to consensus DNA-regulatory elements, known as

12-O-tetradecanoylphorbol-13-acetate (TPA) response elements (TREs) (5). Previous studies showed that Fos/Jun heterodimers are more stable and efficient in driving transcriptional activation than Jun/Jun homodimers, with c-Jun being the most potent transcriptional activator in its group (5). Accumulating evidence has implicated AP-1 in the regulation of a

variety of cellular processes including proliferation, differentiation, growth, apoptosis, cell migration, and transformation (6). For example, AP-1 has been shown to promote proliferation of ER-positive breast cancer cells (7). In the ER-positive breast cancer cell line

MCF-7, upregulated AP-1 activity has been associated with tamoxifen resistance and increased invasiveness (8). However, despite increasing knowledge regarding the physiological functions of AP-1, our understanding of how AP-1 governs transcriptional regulatory networks in breast cancer, and how AP-1 regulates the growth of breast cancer, including TNBC, remains largely unknown. Assaying expression of AP-1 family members in a limited number of breast cancer samples, we previously observed that Fra-1 expression was higher in TNBCs compared to ER-positive breast cancers (9). In the present study, expression pattern of AP-1 members across distinct molecular subtypes, luminal A, luminal B, HER2- enriched, basal-like, and normal-like, reveals the overexpression of Fra-1 in basal-like subtype and its prognostic value. Furthermore, we employ an integrative analysis combining the genomic landscape of the AP-1 components, Fra-1 and c-Jun, with loss-of-function transcriptome analysis to comprehensively decipher the role of AP-1 in TNBC. In addition, we examine the upstream signaling pathway that regulates AP-1 expression and transcriptional activity and identify potential mediators of the proliferative and invasive phenotypes associated with AP-1 signaling.

Materials and Methods

Detailed methods are included in the Supplementary Data.

Cell lines

T47D, MCF-7, MDA-MB-175, ZR-751, SK-BR-3, HCC1569, MDA-MB-453, HCC202,

HCC1954, MDA-MB-231, BT549, Hs578T, Sum159, and MDA-MB-157 cells were obtained

from American Type Culture Collection. Whole cell lysates from MDA-MB-361 and HCC70 were purchased from Biomiga (San Diego, CA).

Analysis of breast tumor microarray data

Four publicly available breast tumor microarray datasets were analyzed. The TCGA RNA-seq data was downloaded from the UCSC Cancer Genomics Browser (10, 11), where gene-level transcript levels are estimated by the RSEM algorithm. The K-M dataset was downloaded from the Kaplan-Meier plotter website (12). Molecular subtypes for this data were determined with the R package genefu

(http://www.bioconductor.org/packages/release/bioc/html/genefu.html), using its robust subtype clustering model for PAM50 subtypes. Data from Wang et al. (13) was downloaded from NCBI’s GEO, accession GSE2034. PAM50 subtypes determined as described above.

Data from Yau et al. (14) was downloaded from the UCSC Cancer Genomics Browser.

Kaplan-Meier plots were generated and log-rank tests performed in R using the package survival.

ChIP and ChIP-Seq

ChIP was performed as previously described (15). The antibodies used in ChIP experiments are anti-Fra-1 (R-20): sc-605 and anti-c-Jun (H-79): sc-1694 from Santa Cruz Biotechnology.

ChIP DNA was prepared into libraries and sequenced using the Genome Analyzer (Illumina) following manufacturer’s protocols. The raw sequencing image data was analyzed by the

Illumina analysis pipeline, mapped to the human reference genome (hg19, GRCh37) using

Bowtie (16). Significantly enriched peak regions were identified in Fra-1 and c-Jun ChIP-seq datasets against both Input and IgG files using the Partek Genomics Suite (Partek® software.

Gene expression microarray analysis

BT549 cells were cultured to 50% confluency and transfected with Fra-1 ONTARGETplus

SMARTpool (ThermoScientific L-004341-00-0010), c-Jun ONTARGETplus SMARTpool

(ThermoScientific L-003268-00-0010), or non-targeting pool (ThermoScientific D-001810-

10-20) using INTERFERinTM siRNA transfection reagent (Polyplus) according to manufacturer’s instructions. Total RNA was isolated 72 hours posttransfection. Total RNA from three biological replicates was hybridized to Affymetrix Human Gene 1.1 ST array. The data were analyzed as previously described (17). Analysis of enrichment of Gene Ontology biological processes was carried out with the Pathway Studio software suite (Ariadne

Genomics, Rockville, MD). The microarray raw data are deposited into GEO (accession number GSE46440).

Western blotting

Western blot analyses were performed as previously described (15). Antibodies used were anti-Fra-1 (R-20): sc-605, anti-c-Jun (H-79): sc-1694 from Santa Cruz Biotechnology, anti-E- cadherin, 610182 from BD Biosciences, anti-FLAG (M2) from Sigma, anti-Akt #9272, anti- phospho-Akt (Ser473) #9271, anti-ERK and anti-phospho-ERK were from Cell Signaling

Technology.

qPCR qPCR was performed as previously described (15).

Cell proliferation assay

Cell proliferation was measured using a WST-1 kit (Roche Applied Science).

Invasion assay

Cells were seeded in the upper chamber of 24-well BD Biocoat growth factor reduced

Matrigel invasion chambers (8.0 μm pore, Becton Dickinson, Bedford, MA, USA) with 0%

FBS media. Media containing 10% FBS was added to the lower chamber. After 24 hours, cells in the upper chamber were removed by scraping. Cells that migrated to the lower chamber were stained and counted under a microscope. ZEB2 siRNA (sc-38641) and CDH1 siRNA (sc-270050) were purchased from Santa Cruz Biotechnology.

Zebrafish metastatic model

Tumor invasion and metastasis experiments in zebrafish were performed according to our previously published methods (18, 19). Briefly, BT549 cells were labeled with 1,1′- dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate (DiI, Fluka, Germany), followed by transfection with siRNAs on the next day. After 24 h, cells were harvested for implantation into zebrafish embryos 2 days post fertilization (dpf). Injected embryos were kept at 28°C and were examined for tumor invasion using a fluorescent microscope (Nikon

Eclipse C1).

Reporter gene assays

BT549, MDA-MB-231 and Hs578T cells were transfected with the 2xTRE luciferase reporter plasmid or the PGL3-promoter vector using Lipofectamine 2000 (Invitrogen Life

Technologies). Luciferase activity was measured using the Dual-Luciferase Reporter Assay

(Promega). To control for transfection efficiency, luciferase activity was normalized to

Renilla luciferase activity.

Results

Fra-1 is overexpressed in basal-like breast cancers and associated with poor prognosis

Gene expression profiling has become an important tool for accurate tumor classification, prognosis and as a guide to therapeutic intervention. We analyzed publically available gene expression profiling datasets from breast tumor samples (11-14) to gain an overall view of the gene expression patterns of AP-1 family members, across the molecular subtypes luminal A, luminal B, HER2-enriched, basal-like, and normal-like, as defined by the 50-gene signature

(PAM50)-based subtype classification (1, 20). Among all AP-1 family members, we observed that Fra-1 (FOSL1) is overexpressed in the basal-like subtype (Fig. 1A). Notably, within the

ER-negative subgroup that contains both basal-like and HER2-enriched tumors, Fra-1 is not highly expressed in the HER2-enriched subtype (Fig. 1A). Moreover, survival analysis of the datasets for which information regarding distant metastasis-free survival (DMFS) was available revealed that Fra-1 was associated with poor clinical outcome (Fig. 1B) (12, 14).

In a panel of 16 breast cancer cell lines, including 5 ER-positive, 5 HER2- positive and 6 basal-like TNBC cell lines, Fra-1 protein was highly expressed in all TNBC cell lines (Fig. 1C), which are models of basal-like breast cancer (21). As Fra-1 requires heterodimerization with Jun family members for efficient transactivation, and c-Jun is the most potent transcriptional activator in its group, we also examined the expression of c-Jun in these cell lines. We observed high levels of c-Jun protein expression in all TNBC cell lines

(Fig. 1C). Moreover, the mRNA expression of both Fra-1 and c-Jun was higher in TNBC cell lines in comparison to ER-positive and HER2-positive cell lines (Fig. 1C). These data

provides a basis for continued studies of the molecular mechanisms of AP-1 signaling in

TNBC cells.

Fra-1 and c-Jun cistromes in TNBC cells

We chose to pursue the cistrome and transcriptome analyses in the BT549 cell line because it belongs to the basal-like subtype based on global gene expression analysis (22) combined with the ability to achieve high transfection efficiency of siRNA in this cell line allowing significant inhibition of Fra-1 or c-Jun expression.

To begin to dissect AP-1 signaling networks in TNBC, we determined genome- wide Fra-1 and c-Jun binding regions in the BT549 cells by ChIP-seq. We identified 11,670

Fra-1 binding regions and 14,201 c-Jun binding regions, with 84% of the Fra-1 cistrome overlapping with the c-Jun cistrome (Fig. 2A), consistent with the notion that Fra-1 binds to

DNA as an obligate heterodimer with Jun proteins. In contrast, the c-Jun cistrome included a large fraction (31%) of unique binding regions (Fig. 2A), in agreement with the ability of c-

Jun homodimers to bind to DNA. Examples of shared and unique binding regions for Fra-1 and c-Jun are shown in Fig. 2B. The Fra-1 and c-Jun cistromes could consistently be verified by independent ChIP-qPCR experiments (Supplementary Fig. S1). Sequence conservation analysis of Fra-1 and c-Jun binding regions across different species showed that sequences near the center of the binding regions tend to be more evolutionarily conserved than flanking sequences (Fig. 2C), suggesting a conserved regulatory role.

We observed that Fra-1 and c-Jun binding regions are distributed throughout the genome with a preponderance of binding regions occurring within intergenic and intronic regions with only 10% of Fra-1 binding sites and 17% of c-Jun binding sites being detected in the proximal 5’ region of the nearest gene (Fig. 2D). To further characterize binding site locations relative to a gene transcription start site (TSS), the distribution of binding sites

relative to the most proximal TSS was determined in a window of 25 kb upstream and downstream of TSS. As shown in Fig. 2E, this distribution reveals that the binding of Fra-1 and c-Jun is highly enriched in close proximity to the TSS. De novo DNA motif analysis revealed several highly enriched motifs. As expected, consensus AP-1 motifs were identified in 58% of Fra-1 binding regions and in 47% of c-Jun binding regions (Fig. 2F), in line with direct DNA binding of AP-1 members to these regions. These results also indicate that AP-1 family members interact indirectly with DNA sequences of some target sites through mechanisms such as tethering.

Fra-1- and c-Jun-regulated transcriptomes in TNBC Cells

To gain further insight into AP-1-regulated transcriptional networks in TNBC cells, we determined global gene expression profiles for TNBC cells upon Fra-1 or c-Jun knockdown.

Protein and mRNA levels of Fra-1 and c-Jun were efficiently depleted by Fra-1 siRNA and c-

Jun siRNA, respectively (Supplementary Fig. S2A and S2B). We identified differential expression of 419 and 690 genes upon Fra-1 and c-Jun knockdown (Fig. 3A), respectively. Of these, 180 genes were induced and 42 genes were repressed by both Fra-1 and c-Jun knockdown (Fig. 3A), indicating that they are regulated by Fra-1/c-Jun heterodimers.

Assigning these genes to functional categories, we found that genes associated with cytokine- mediated signaling, cell adhesion, immune response, cell junction assembly and inflammatory response, among others, were highly enriched (Fig. 3B). This is consistent with the established role of AP-1 in regulating immunity and inflammation (23). We used qPCR to confirm the expression of 10 Fra-1/c-Jun common target genes identified in the gene expression profiling analysis (Supplementary Fig. S2C)

Combining data from the cistrome and transcriptome analysis, 56% (233/419) of

Fra-1-regulated genes were identified as including Fra-1 binding regions and 60% (411/690)

of c-Jun-regulated genes were identified as including c-Jun binding regions within 50 kb upstream or downstream of their TSSs, suggesting that they represent direct AP-1 target genes.

Among these, 101 genes were induced and 24 genes were repressed by both Fra-1 and c-Jun knockdown suggesting that they represent direct target genes of Fra-1/c-Jun heterodimers (Fig.

3C).

Combining data from the cistrome and transcriptome analysis also demonstrates that c-Jun and Fra-1 binding regions were enriched within 50 kb of the TSS of c-Jun- and Fra-

1-regulated genes (Fig. 3D). Furthermore, this analysis showed that c-Jun binding sites were highly enriched near the c-Jun-upregulated genes as compared with its repressed genes (Fig.

3D). Interestingly, it has also been shown that the binding of nuclear receptors such as AR,

ESR2, HNF4A, NR3C1, PGR, and VDR were enriched in regions associated with upregulated genes compared to repressed genes (24).

Fra-1 and c-Jun common target genes regulating cell proliferation

Due to the importance of AP-1 in regulating tumor cell proliferation (25), we assessed the function of Fra-1 and c-Jun in TNBC cell proliferation. Both Fra-1 and c-Jun knockdown markedly inhibited the proliferation of BT549 breast cancer cells (Fig. 3E), implying that AP-

1 increases cell proliferation of TNBC cells. Inhibition of proliferation by Fra-1 and c-Jun siRNA was confirmed using an independent siRNA pool (Supplementary Fig. S3A), suggesting that this effect is specific to Fra-1 and c-Jun silencing. Targeted knockdown of

Fra-1 in an additional TNBC cell line, Hs578T, also led to a significant inhibition of cell proliferation compared with control siRNA, whereas no such effect was observed in MCF-7 breast cancer cell line with low endogenous Fra-1 expression (Supplementary Fig. S3B).

Furthermore, overexpression of Fra-1 in BT549 cells depleted of Fra-1 by siRNA largely reversed the effect of Fra-1 depletion on cell proliferation (Fig. 3F).

We next identify Fra-1- and c-Jun-regulated genes with known regulatory functions in cell proliferation. Using this approach, 20 genes that repress cell proliferation were identified to be upregulated upon Fra-1 and c-Jun knockdown, whereas 7 genes that promote cell proliferation were identified to be downregulated upon Fra-1 and c-Jun knockdown (Fig. 3G). Putative Fra-1 and c-Jun direct target genes were identified from the

Fra-1 and c-Jun cistromes. This analysis identified 16 potential AP-1 direct transcriptional target genes encoding regulators of cell proliferation, including ABI3BP, BIRC3, CLCA2,

DLC1, IL22RA2, ITGB5, KCNIP1, KRT18, NPR3, PAD12, TIMP3, MEF2C, MT1F, NFATC2,

NPNT and NTRK2 (Fig. 3G). Moreover, survival analysis of the K-M dataset (12) revealed that this list of 16 genes has prognostic significance in ER-negative patients (Supplementary

Fig. S4). Among these, CLCA2 (chloride channel accessory 2) was identified as the most upregulated transcript upon Fra-1 and c-Jun knockdown. Interestingly, reduced expression of

CLCA2 was frequently observed in breast cancer (26). In addition, CLCA2 was shown to play a crucial role in inhibition of cancer cell migration and invasion (27, 28). We confirmed by qPCR that CLCA2 expression was substantially elevated after Fra-1 or c-Jun depletion (Fig.

3H). As shown in Fig. 3I, Fra-1 and c-Jun binding regions were observed within intron 3 of the CLCA2 gene. We further validated, by ChIP-qPCR, the binding of Fra-1 and c-Jun to this region in BT549 cells (Fig. 3H). To determine the influence of CLCA2 overexpression on cell proliferation, we generated BT549 cells stably expressing CLCA2. Consistent with previous findings (27), CLCA2-overexpressing cells exhibited reduced cell proliferation, as compared to mock-transfected cells (Fig. 3J and Supplementary Fig. S5A). Collectively, these findings reveal molecular mechanisms by which AP-1 induces cell proliferation.

AP-1 promotes TNBC cell invasion through transcriptional upregulation of the E- cadherin (CDH1) repressor ZEB2

Following our observation of a significant enrichment of genes involved in cell adhesion among AP-1 target genes (Fig. 3B), we examined whether AP-1 may regulate the invasive capacity of TNBC cells. To address this, we performed transwell cell invasion assays for

BT549 cells and observed that knockdown of Fra-1 or c-Jun expression significantly reduced the invasion ability (Fig. 4A). This is consistent with AP-1 promoting cell invasion in TNBC cells.

Screening of the genes involved in the “cell adhesion” gene ontology category revealed significant upregulation of E-cadherin (CDH1) upon Fra-1 and c-Jun knockdown

(Fig. 4B). E-cadherin is widely acknowledged for its role in cell-cell adhesion (29). We validated that knockdown of Fra-1 and c-Jun substantially elevated the mRNA level of E- cadherin using qPCR (Fig. 4B). Western blot analysis showed that knockdown of Fra-1 and c-

Jun significantly increased E-cadherin protein levels (Fig. 4C). Conversely, treatment of cells with TPA, which is a potent activator of AP-1, decreased the level of E-cadherin protein (Fig.

4C). Downregulation of E-cadherin by Fra-1 and c-Jun was further validated in another

TNBC cell line Hs578T (Supplementary Fig. S5B and S5C).

Notably, the Fra-1 and c-Jun cistromes did not reveal any AP-1 binding regions within or proximal to the E-cadherin gene locus. The expression of E-cadherin is regulated by several known transcriptional repressors, including SNAIL, SLUG, ZEB1, ZEB2 and TWIST

(30). Interestingly, a Fra-1 and c-Jun binding region was observed in one intron of the ZEB2 gene (Fig. 4D). A highly conserved sequence with high homology to the AP-1 consensus motif was identified in this binding region (Fig. 4D). This region seems to be an active enhancer, marked by enrichment of H3K4me1, H3K4me4 and H3K27Ac (Fig. 4D). An independent ChIP-qPCR assay confirmed that Fra-1 and c-Jun were directly bound to this genomic region (Fig. 4E). Furthermore, Fig. 4E shows that ZEB2 expression was downregulated in Fra-1 or c-Jun depleted cells, with c-Jun depletion showing a predominant

effect. Analysis of ZEB2 expression in the TNBC cell line Hs578T confirmed downregulation of ZEB2 mRNA following knockdown of Fra-1 or c-Jun (Supplementary Fig. S5D). These data suggest that AP-1 directly activates ZEB2 transcription. Moreover, we noted a statistically significant negative association between c-Jun and E-cadherin expression levels in a published study of breast cancer samples from 286 individuals (13) (Fig. 4F), further supporting the in vivo relevance of the ability of AP-1 to promote tumor invasion in human breast tumors.

ZEB2 mediates downregulation of E-cadherin and promotes cell invasion

To further establish a role of ZEB2 in cell invasion in TNBC cells, we knocked down the expression of ZEB2 in BT549 cells (Fig. 5A). Knockdown of ZEB2 expression greatly inhibited the invasiveness of BT549 cells (Fig. 5B). In line with the suppression of cell invasion, E-cadherin protein levels were increased upon ZEB2 depletion (Fig. 5C). To establish the relationship between ZEB2 expression and an invasive breast cancer phenotype, we examined ZEB2 expression in 9 human breast cancer cell lines, including 3 TNBC cell lines and 6 non-TNBC breast cancer cell lines. High levels of ZEB2 mRNA were detected only in cell lines representing the invasive TNBC subtype (Fig. 5D). These data link ZEB2, a direct target gene of AP-1, to downregulation of E-cadherin and enhanced cell invasion in

TNBC cells (Fig. 7D).

To determine the role of E-cadherin in the regulation of cell invasiveness downstream of Fra-1, we examined whether simultaneous E-cadherin and Fra-1 knockdown would render cells insensitive to Fra-1 knockdown. The reduced invasiveness of BT549 cells upon Fra-1 knockdown was reversed by knockdown of E-cadherin (Fig. 5E and

Supplementary Fig. S5E), demonstrating the importance of the Fra-1 – E-cadherin axis.

Knockdown of AP-1 inhibits invasion of TNBC cells in vivo

To study the role of Fra-1 and c-Jun in regulating TNBC cell invasion and metastasis in vivo, we employed a zebrafish tumor model (18, 19). Tumor-implanted fish embryos were scored for the dissemination of tumor cells at day 4 post injection. Implantation of control-siRNA treated BT549 tumor cells led to broad dissemination of tumor cells in different regions of the fish body, whereas BT549 cells silenced for Fra-1 or c-Jun mostly remained in the perivitelline space (Fig.6A and 6B). Thus, Fra-1 or c-Jun knockdown inhibits the invasiveness of TNBC cells in vivo.

The PI3K/Akt and MAPK/ERK pathways positively regulate Fra-1 and c-Jun expression and are associated with increased AP-1 transcriptional activity in TNBC cells

Our results demonstrated high Fra-1 and c-Jun levels in TNBC cell lines (Fig. 1C). PI3K/Akt signaling has been shown to be enhanced in TNBC compared to other breast tumor subtypes

(31). Furthermore, activation of the ERK signaling pathway induces expression of AP-1 components (32, 33). To determine whether these two signaling pathways mediated upregulation of AP-1 members in TNBC cells, we used two pharmacological inhibitors,

LY294002 and PD98059, to specifically block the PI3K/Akt and MAPK/ERK pathways, respectively. A decrease in phosphorylated Akt upon LY294002 treatment led to a significant downregulation of Fra-1 and c-Jun proteins at 6 and 24 hours (Fig. 7A and Supplementary Fig.

S6A). These results were further validated in two other TNBC cell lines, Hs578T and MDA-

MB-231 (Supplementary Fig. S6B). Similarly, PD98059 treatment led to a significant downregulation of Fra-1 and c-Jun proteins at 2, 6 and 24 hours (Fig. 7B and Supplementary

Fig. S6C). PD98059 inhibition of ERK activity was confirmed by assessing the level of

ERK1/2 phosphorylation (Fig. 7B). These data suggest that both PI3K/Akt and MAPK/ERK signaling can increase the expression of Fra-1 and c-Jun in TNBC cells.

In addition, we assessed whether these two pathways are involved in the enhancement of AP-1 transcriptional activity in TNBC cells. In accordance with the reduced

AP-1 protein levels, the AP-1 transcriptional activity was attenuated by LY294002 and

PD98059 treatment (Fig. 7C). Thus, activation of the PI3K/Akt and MAPK/ERK pathways is associated with increased AP-1 transactivation in TNBC cells.

Discussion

Understanding the molecular basis of TNBC is of pivotal importance in light of the current lack of therapeutic options and the poor clinical outcome for this class of breast cancers. We report that TNBC cell lines and primary tumors express high levels of Fra-1, and that Fra-1 expression is associated with poor prognosis. Based on these observations, we pursue a comprehensive characterization of AP-1 signaling in TNBC and suggest how this contributes to the proliferative and invasive phenotypes. Our results suggest both PI3K/Akt and

MAPK/ERK signaling as important regulators of AP-1 transcriptional activity in TNBC cells and reveal novel mechanisms underlying AP-1-mediated regulation of cell proliferation and invasion in TNBC cells (Fig. 7D).

We demonstrate overexpression of Fra-1 in basal-like breast cancer (Fig. 1A) and basal-like breast cancer cell lines (Fig. 1C), one of the most aggressive types of breast cancer. It is notable that Fra-1 was expressed at a relative high level in luminal MDA-MB-175 cells (Fig. 1C). However, MDA-MB-175 cells have been demonstrated to have much higher pAkt and pERK1/2 expression than other luminal cell lines such as MCF-7 and T47D (34) and Fra-1 has been shown to be activated by PI3K/Akt or MAPK/ERK signaling (Fig. 7A and

7B) (33). Our analysis also reveals that Fra-1 levels correlate with poor outcome in breast tumors (Fig.1B). Consistent with our findings, one recent study showed that the Fra-1-

dependent transcriptome has prognostic potential in human breast cancer (35). It has long been suspected that Fra-1 plays an important role in tumor metastasis. Consistent with this hypothesis, we further show that Fra-1 and c-Jun promote cell invasion both in vitro and in vivo (Fig. 4A and Fig. 6). In line with this, Fra-1 has been shown to be involved in the migratory or invasive capabilities of various cancer cell lines (36, 37). Furthermore, in rodent model systems, Fra-1 was identified as a key regulator of metastasis (35). Based on these data, intervention of Fra-1-mediated signaling may be explored for therapeutic options in basal-like breast cancer.

Our results indicate that high PI3K/Akt or MAPK/ERK activity increases AP-1 signaling and ultimately cell invasion for TNBC cells. In TNBCs, these two pathways are known to be activated and are associated with a poor prognosis (38, 39). A recent study suggested that a combination regimen with MEK/ERK and PI3K inhibitors might effectively be used to treat basal-like breast cancers (40). Our data further suggest that such an approach in combination with intervention of AP-1 signaling could be beneficial in TNBCs.

In recent years, cistrome analyses have shown that AP-1 cooperates with nuclear receptors such as ER (41, 42), LXR (43) and GR (44) in orchestrating gene expression programs. However, the cistromes of AP-1 members have not been described for human breast cancer cells. Using ChIP-seq, 11,670 Fra-1 binding regions and 14,201 c-Jun binding regions were identified in BT549 TNBC cells, the majority of which (about 75%) were distal to the promoter in intergenic and intronic regions, consistent with the pattern for nuclear receptor binding to DNA (42, 45). The majority of de novo Fra-1 or c-Jun binding sites are therefore either not functional or affect transcription through long-range promoter-enhancer interactions. Indeed, long-range chromatin interactions have been proposed as a model for regulation of gene expression by distal ER binding sites (46). It is of interest to note that there are many more Fra-1 and c-Jun binding sites in the genome than differentially expressed

genes, as has been observed for several nuclear receptors (45, 47, 48). This suggests that multiple sites are involved in the regulation of a single gene, and/or that binding alone does not necessarily drive gene expression. In addition, as our analyses were conducted at a single time point, they will provide only a snapshot of the transcriptional consequences of Fra-1 or c-Jun knockdown. Dynamic changes in gene expression, along with potential differences in the kinetics of individual siRNA, may explain some of the discordance between the number of

Fra-1- or c-Jun- regulated genes and common genes regulated by both Fra-1 and c-Jun.

So far, the understanding of molecular events underlying the aggressiveness of

TNBC is limited, but the TNBC subtype is frequently associated with apparent epithelial- mesenchymal transition (EMT). One of the hallmarks of EMT is the downregulation of the cell-cell junction protein E-cadherin (49). The gene encoding E-cadherin is considered as an invasion-suppressor gene and downregulation of E-cadherin expression correlates with malignancy in human cancers. ZEB2 represses E-cadherin transcription through its direct binding to conserved E2 boxes present in the E-cadherin promoter (50). In this study, we have shown that AP-1 directly activates ZEB2 expression through binding to a regulatory region of the ZEB2 gene. This in turn leads to downregulation of E-cadherin. Moreover, analysis of clinical breast tumor data reveals a significant negative correlation between the expression level of c-Jun and E-cadherin. These findings highlight a crucial AP-1-mediated signaling cascade promoting invasion in TNBC.

In summary, we provide the first genome-wide map of AP-1 binding in TNBC cells and demonstrate overexpression of Fra-1 in basal-like breast tumors. Our results reveal novel mechanisms underlying AP-1-mediated regulation of cell proliferation and invasion in

TNBC cells (Fig. 7D). Knockdown of Fra-1 reduces cellular invasion both in vitro and in vivo, suggesting that Fra-1 may be useful as a therapeutic target for patients with invasive breast tumors. One of the future challenges will be to identify specific Fra-1 inhibitors and

demonstrate their efficacy in inhibiting cell proliferation and invasion of TNCB cell lines and relevant animal models.

Disclosure of Potential Confl icts of Interest

No potential conflicts of interest were disclosed.

Authors’ Contributions

Conception and design: C. Zhao, K. Dahlman-Wright, Y. Cao, C. Williams

Development of methodology: C. Zhao, Y. Qiao, P. Jonsson, J. Wang, L. Xu, P. Rouhi

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K. Dahlman-Wright, Y. Cao

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C. Zhao, Y. Qiao, P. Jonsson, J. Wang, L. Xu, I. Sinha, C. Williams

Writing, review, and/or revision of the manuscript: C. Zhao, Y. Qiao, P. Jonsson, Y. Cao,

I. Sinha, C. Williams, K. Dahlman-Wright

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): C. Zhao, Y. Qiao, P. Jonsson, J. Wang, L. Xu, I. Sinha

Study supervision: C. Zhao, K. Dahlman-Wright

Acknowledgements

The authors acknowledge the Bioinformatic and Expression Analysis core facility at the

Karolinska Institute (BEA, www.bea.ki.se) for performing the Affymetrix assays.

Grant Support

This work was supported by grants from the Swedish Cancer Society (Cancerfonden) and the

Center for Biosciences.

References

1. Sorlie T, Tibshirani R, Parker J, Hastie T, Marron JS, Nobel A, et al. Repeated

observation of breast tumor subtypes in independent gene expression data sets. Proc

Natl Acad Sci U S A. 2003;100:8418-23.

2. Elias AD. Triple-negative breast cancer: a short review. Am J Clin Oncol.

2010;33:637-45.

3. Pal SK, Childs BH, Pegram M. Triple negative breast cancer: unmet medical needs.

Breast Cancer Res Treat. 2011;125:627-36.

4. Slamon DJ, Leyland-Jones B, Shak S, Fuchs H, Paton V, Bajamonde A, et al. Use of

chemotherapy plus a monoclonal antibody against HER2 for metastatic breast cancer

that overexpresses HER2. N Engl J Med. 2001;344:783-92.

5. Shaulian E, Karin M. AP-1 as a regulator of cell life and death. Nat Cell Biol.

2002;4:E131-6.

6. Shaulian E. AP-1--The Jun proteins: Oncogenes or tumor suppressors in disguise? Cell

Signal. 2010;22:894-9.

7. Liu Y, Lu C, Shen Q, Munoz-Medellin D, Kim H, Brown PH. AP-1 blockade in breast

cancer cells causes cell cycle arrest by suppressing G1 cyclin expression and reducing

cyclin-dependent kinase activity. Oncogene. 2004;23:8238-46.

8. Smith LM, Wise SC, Hendricks DT, Sabichi AL, Bos T, Reddy P, et al. cJun

overexpression in MCF-7 breast cancer cells produces a tumorigenic, invasive and

hormone resistant phenotype. Oncogene. 1999;18:6063-70.

9. Kharman-Biz A, Gao H, Ghiasvand R, Zhao C, Zendehdel K, Dahlman-Wright K.

Expression of activator protein-1 (AP-1) family members in breast cancer. BMC

cancer. 2013;13:441.

10. Goldman M, Craft B, Swatloski T, Ellrott K, Cline M, Diekhans M, et al. The UCSC

Cancer Genomics Browser: update 2013. Nucleic acids research. 2013;41:D949-54.

11. TCGA. Comprehensive molecular portraits of human breast tumours. Nature.

2012;490:61-70.

12. Gyorffy B, Lanczky A, Eklund AC, Denkert C, Budczies J, Li Q, et al. An online

survival analysis tool to rapidly assess the effect of 22,277 genes on breast cancer

prognosis using microarray data of 1,809 patients. Breast Cancer Res Treat.

2010;123:725-31.

13. Wang Y, Klijn JG, Zhang Y, Sieuwerts AM, Look MP, Yang F, et al. Gene-expression

profiles to predict distant metastasis of lymph-node-negative primary breast cancer.

Lancet. 2005;365:671-9.

14. Yau C, Esserman L, Moore DH, Waldman F, Sninsky J, Benz CC. A multigene

predictor of metastatic outcome in early stage hormone receptor-negative and triple-

negative breast cancer. Breast Cancer Res. 2010;12:R85.

15. Zhao C, Matthews J, Tujague M, Wan J, Strom A, Toresson G, et al. Estrogen receptor

beta2 negatively regulates the transactivation of estrogen receptor alpha in human

breast cancer cells. Cancer Res. 2007;67:3955-62.

16. Langmead B, Trapnell C, Pop M, Salzberg SL. Ultrafast and memory-efficient

alignment of short DNA sequences to the human genome. Genome Biol. 2009;10:R25.

17. Dahlman-Wright K, Qiao Y, Jonsson P, Gustafsson JA, Williams C, Zhao C. Interplay

between AP-1 and estrogen receptor alpha in regulating gene expression and

proliferation networks in breast cancer cells. Carcinogenesis. 2012;33:1684-91.

18. Cao Z, Jensen LD, Rouhi P, Hosaka K, Lanne T, Steffensen JF, et al. Hypoxia-

induced retinopathy model in adult zebrafish. Nat Protoc. 2010;5:1903-10.

19. Lee SL, Rouhi P, Dahl Jensen L, Zhang D, Ji H, Hauptmann G, et al. Hypoxia-induced

pathological angiogenesis mediates tumor cell dissemination, invasion, and metastasis

in a zebrafish tumor model. Proc Natl Acad Sci U S A. 2009;106:19485-90.

20. Parker JS, Mullins M, Cheang MC, Leung S, Voduc D, Vickery T, et al. Supervised

risk predictor of breast cancer based on intrinsic subtypes. J Clin Oncol.

2009;27:1160-7.

21. Kao J, Salari K, Bocanegra M, Choi YL, Girard L, Gandhi J, et al. Molecular profiling

of breast cancer cell lines defines relevant tumor models and provides a resource for

cancer gene discovery. PLoS One. 2009;4:e6146.

22. Neve RM, Chin K, Fridlyand J, Yeh J, Baehner FL, Fevr T, et al. A collection of

breast cancer cell lines for the study of functionally distinct cancer subtypes. Cancer

Cell. 2006;10:515-27.

23. Li MD, Yang X. A Retrospective on Nuclear Receptor Regulation of Inflammation:

Lessons from GR and PPARs. PPAR Res. 2011;2011:742785.

24. Tang Q, Chen Y, Meyer C, Geistlinger T, Lupien M, Wang Q, et al. A comprehensive

view of nuclear receptor cancer cistromes. Cancer Res. 2011;71:6940-7.

25. Eferl R, Wagner EF. AP-1: a double-edged sword in tumorigenesis. Nat Rev Cancer.

2003;3:859-68.

26. Li X, Cowell JK, Sossey-Alaoui K. CLCA2 tumour suppressor gene in 1p31 is

epigenetically regulated in breast cancer. Oncogene. 2004;23:1474-80.

27. Walia V, Ding M, Kumar S, Nie D, Premkumar LS, Elble RC. hCLCA2 Is a p53-

Inducible Inhibitor of Breast Cancer Cell Proliferation. Cancer Res. 2009;69:6624-32.

28. Sasaki Y, Koyama R, Maruyama R, Hirano T, Tamura M, Sugisaka J, et al. CLCA2, a

target of the p53 family, negatively regulates cancer cell migration and invasion.

Cancer Biol Ther. 2012;13:1512-21.

29. Scully OJ, Bay BH, Yip G, Yu Y. Breast cancer metastasis. Cancer Genomics

Proteomics. 2012;9:311-20.

30. Berx G, van Roy F. Involvement of members of the cadherin superfamily in cancer.

Cold Spring Harb Perspect Biol. 2009;1:a003129.

31. Umemura S, Yoshida S, Ohta Y, Naito K, Osamura RY, Tokuda Y. Increased

phosphorylation of Akt in triple-negative breast cancers. Cancer Sci. 2007;98:1889-92.

32. Casalino L, De Cesare D, Verde P. Accumulation of Fra-1 in ras-transformed cells

depends on both transcriptional autoregulation and MEK-dependent posttranslational

stabilization. Mol Cell Biol. 2003;23:4401-15.

33. Doehn U, Hauge C, Frank SR, Jensen CJ, Duda K, Nielsen JV, et al. RSK is a

principal effector of the RAS-ERK pathway for eliciting a coordinate

promotile/invasive gene program and phenotype in epithelial cells. Molecular cell.

2009;35:511-22.

34. Sanchez CG, Ma CX, Crowder RJ, Guintoli T, Phommaly C, Gao F, et al. Preclinical

modeling of combined phosphatidylinositol-3-kinase inhibition with endocrine therapy

for estrogen receptor-positive breast cancer. Breast Cancer Res. 2011;13:R21.

35. Desmet CJ, Gallenne T, Prieur A, Reyal F, Visser NL, Wittner BS, et al. Identification

of a pharmacologically tractable Fra-1/ADORA2B axis promoting breast cancer

metastasis. Proc Natl Acad Sci U S A. 2013;110:5139-44.

36. Belguise K, Kersual N, Galtier F, Chalbos D. FRA-1 expression level regulates

proliferation and invasiveness of breast cancer cells. Oncogene. 2005;24:1434-44.

37. Sayan AE, Stanford R, Vickery R, Grigorenko E, Diesch J, Kulbicki K, et al. Fra-1

controls motility of bladder cancer cells via transcriptional upregulation of the receptor

tyrosine kinase AXL. Oncogene. 2012;31:1493-503.

38. Bartholomeusz C, Gonzalez-Angulo AM, Liu P, Hayashi N, Lluch A, Ferrer-Lozano J,

et al. High ERK protein expression levels correlate with shorter survival in triple-

negative breast cancer patients. Oncologist. 2012;17:766-74.

39. Lopez-Knowles E, O'Toole SA, McNeil CM, Millar EK, Qiu MR, Crea P, et al. PI3K

pathway activation in breast cancer is associated with the basal-like phenotype and

cancer-specific mortality. International Journal of Cancer 2010;126:1121-31.

40. Hoeflich KP, O'Brien C, Boyd Z, Cavet G, Guerrero S, Jung K, et al. In vivo

antitumor activity of MEK and phosphatidylinositol 3-kinase inhibitors in basal-like

breast cancer models. Clinical cancer research 2009;15:4649-64.

41. Lupien M, Meyer CA, Bailey ST, Eeckhoute J, Cook J, Westerling T, et al. Growth

factor stimulation induces a distinct ER(alpha) cistrome underlying breast cancer

endocrine resistance. Genes Dev. 2010;24:2219-27.

42. Zhao C, Gao H, Liu Y, Papoutsi Z, Jaffrey S, Gustafsson JA, et al. Genome-wide

mapping of estrogen receptor-beta-binding regions reveals extensive cross-talk with

transcription factor activator protein-1. Cancer Res. 2010;70:5174-83.

43. Shen Q, Bai Y, Chang KC, Wang Y, Burris TP, Freedman LP, et al. Liver X receptor-

retinoid X receptor (LXR-RXR) heterodimer cistrome reveals coordination of LXR

and AP1 signaling in keratinocytes. J Biol Chem. 2011;286:14554-63.

44. Biddie SC, John S, Sabo PJ, Thurman RE, Johnson TA, Schiltz RL, et al.

Transcription factor AP1 potentiates chromatin accessibility and glucocorticoid

receptor binding. Mol Cell. 2011;43:145-55.

45. Hu S, Yao G, Guan X, Ni Z, Ma W, Wilson EM, et al. Research resource: Genome-

wide mapping of in vivo androgen receptor binding sites in mouse epididymis.

Molecular endocrinology. 2010;24:2392-405.

46. Fullwood MJ, Liu MH, Pan YF, Liu J, Xu H, Mohamed YB, et al. An oestrogen-

receptor-alpha-bound human chromatin interactome. Nature. 2009;462:58-64.

47. Carroll JS, Meyer CA, Song J, Li W, Geistlinger TR, Eeckhoute J, et al. Genome-wide

analysis of estrogen receptor binding sites. Nature genetics. 2006;38:1289-97.

48. Rao NA, McCalman MT, Moulos P, Francoijs KJ, Chatziioannou A, Kolisis FN, et al.

Coactivation of GR and NFKB alters the repertoire of their binding sites and target

genes. Genome Res. 2011;21:1404-16.

49. Talbot LJ, Bhattacharya SD, Kuo PC. Epithelial-mesenchymal transition, the tumor

microenvironment, and metastatic behavior of epithelial malignancies. Int J Biochem

Mol Biol. 2012;3:117-36.

50. Comijn J, Berx G, Vermassen P, Verschueren K, van Grunsven L, Bruyneel E, et al.

The two-handed E box binding zinc finger protein SIP1 downregulates E-cadherin and

induces invasion. Molecular cell. 2001;7:1267-78.

FIGURE LEGENDS

Figure 1. Fra-1 is overexpressed in basal-like breast cancers and associated with poor prognosis. A, Fra-1 (FOSL1) expression in basal-like and luminal A, luminal B, normal-like or HER2-enriched breast cancer molecular subtypes from four published microarray datasets

(11-14). The y-axes show log2-transformed normalized counts for the TCGA data and normalized microarray intensity data for the remaining datasets, according to description available at the data sources. *P < 0.05, **P < 0.01, and ***P < 0.001 by ANOVA with

Tukey's post-hoc test. B, Kaplan-Meier plots showing the probability of DMFS stratified by

Fra-1 (FOSL1) expression, cut off at median expression. P-values were calculated using the log-rank test. C, Western blot analysis of Fra-1 and c-Jun levels in ER+/PR+/HER2- breast cancer cell lines (T47D, MCF-7, MDA-MB-175, ZR-751 and MDA-MB-361), ER-/PR-

/HER2+ breast cancer cell lines (SK-BR-3, HCC1569, MDA-MB-453, HCC202 and

HCC1954) and TNBC cell lines (MDA-MB-231, BT549, Hs578T, Sum159, HCC70 and

MDA-MB-157). β-actin is shown as loading control. Bar graphs show Fra-1 and c-Jun mRNA levels, determined by qPCR, relative to the mRNA expression in T47D cells. mRNA levels are presented as means with SD (n = 3).

Figure 2. Fra-1 and c-Jun cistromes in TNBC cells. A, venn diagram showing the overlap between Fra-1 and c-Jun cistromes. B, representative examples of ChIP-seq results showing binding regions shared by Fra-1 and c-Jun in the CSRNP1 gene, a Fra-1 unique binding region upstream of UBE2D3 and a c-Jun unique region upstream of DUSP16. C, averaged conservation profiles using phastCons scores for Fra-1 and c-Jun binding sites, compared to random regions. D, genomic positions of Fra-1 and c-Jun binding sites relative to the nearest gene. E, kernel density plot illustrating enrichment of Fra-1 and c-Jun peaks close to the TSS

of their most proximal gene. F, top-enriched DNA motifs, prospective families of DNA binding transcription factors and percentage of peaks containing the motif identified by de novo motif analysis of Fra-1 and c-Jun binding sites.

Figure 3. Fra-1- and c-Jun-regulated transcriptomes and target genes related to cell proliferation in TNBC cells. A, overlap of differentially expressed genes upon Fra-1 or c-Jun knockdown, separated by up- and downregulation. B, overrepresented Gene Ontology biological processes for common genes (overlaps from A). C, overlap of genes with Fra-1 or c-Jun binding regions (within 50 kb upstream or downstream of their TSS) also identified as differentially expressed genes by microarray analysis. D, barplot showing fractions of differentially expressed genes with binding sites within 50 kb of their TSS. Black bars represent the same analysis for all genes represented on the microarrays. E, Fra-1 or c-Jun depletion reduces proliferation in BT549 cells. WST-1 assays of cell metabolic activity were carried out at the indicated time points after siRNA transfection. ***P < 0.001 compared with control siRNA (n = 3). F, rescue of siRNA knockdown confirms targeting specificity. BT549 cells were transfected with control or Fra-1 siRNA. Next day, the cells were transfected with either control vector or Fra-1 expression plasmid (Origene) as indicated, and WST-1 assays of cell metabolic activity were carried out at the indicated time points. Western blot confirms knockdown and rescue of Fra-1 levels. β-actin is shown as loading control. G, heatmap showing Fra-1 and c-Jun common target genes involved in cell proliferation. Dots indicate the presence of Fra-1 or c-Jun binding sites within 50 kb of the TSS. H, CLCA2 mRNA levels were determined by qPCR in BT549 cells after transfection with control siRNA or siRNAs against Fra-1 or c-Jun for the indicated times. ChIP-qPCR analysis confirms the recruitment of Fra-1 and c-Jun to a region of the CLCA2 gene. Primers flanking the CLCA2 promoter region were used as a control. Data presented are normalized to input DNA and expressed as

fold enrichment relative to IgG. Values are mean ± SD (n = 3). I, ChIP-seq results show the

Fra-1 and c-Jun binding peaks within the CLCA2 gene locus. J, CLCA2 overexpression reduces BT549 cell proliferation. WST-1 assays of cell metabolic activity were carried out at the indicated time points in pCMV6-Vec and pCMV6-CLCA2 cells. ***P < 0.001 compared with pCMV6-CLCA2 (n = 3).

Figure 4. AP-1 promotes TNBC cell invasion through transcriptional upregulation of the E- cadherin (CDH1) repressor ZEB2. A, Fra-1 or c-Jun depletion reduces BT549 cell invasion.

***P < 0.001 compared with control siRNA (n = 3). B, microarray and qPCR data showing upregulation of E-cadherin (CDH1) mRNA levels following Fra-1 or c-Jun depletion. C,

Western blot analysis showing increased E-cadherin levels upon Fra-1 or c-Jun depletion and decreased E-cadherin levels following TPA treatment. β-actin was used as a loading control.

D, the left panel shows the Fra-1 and c-Jun binding region within the ZEB2 gene locus as identified by ChIP-seq. The right panel shows the histone modification profile and across- species conservation of AP-1 motif in the binding region. E, ChIP-qPCR analysis confirms the recruitment of Fra-1 and c-Jun to ZEB2. ZEB2 mRNA levels were determined by qPCR in

BT549 cells after transfection with control siRNA or siRNAs against Fra-1 or c-Jun for the indicated times. Values are mean ± SD (n = 3). F, scatterplot showing significant correlation between E-cadherin (CDH1) and c-Jun (JUN) levels in the Wang dataset (13), test statistics from Pearson product-moment correlation.

Figure 5. ZEB2 mediates downregulation of E-cadherin and promotes cell invasion. A, knockdown of ZEB2 significantly reduces ZEB2 expression at both the mRNA and protein levels 72 hours after siRNA transfection in BT549 cells. β-actin was used as a loading control.

B, ZEB2 depletion reduces BT549 cell invasion. ***P < 0.001 compared with control siRNA

(n = 3). C, Western blot analysis showing increased E-cadherin (CDH1) levels upon ZEB2 depletion. β-actin was used as a loading control. D, ZEB2 mRNA expression levels in TNBC cell lines compared with non-TNBC breast cancer cell lines. mRNA levels are presented as means with SD (n = 3), relative to the mRNA expression in T47D cells. E, Simultaneous E- cadherin and Fra-1 knockdown rescues the invasion of Fra-1 knockdown cells. ***P < 0.001 compared with Fra-1 siRNA (n = 3).

Figure 6. Knockdown of AP-1 inhibits invasion of TNBC cells in vivo. A, representative micrographs of zebrafish embryos implanted with control-siRNA, Fra-1-siRNA or c-Jun- siRNA transfected BT549 cells at day 0 and 4. Red = DiI-labeled tumor cells; Bar = 100 μm.

Dished lines encircle the implanted primary tumors. Arrowheads point to disseminated tumor cells. B, quantification of disseminated tumor foci in zebrafish implanted with control-siRNA,

Fra-1-siRNA or c-Jun-siRNA transfected BT549 cells (n = 11 embryos/group). (BF, bright field; RF, red field; NS, not significant). ***P < 0.001 compared with control siRNA. The experiment was repeated twice with similar results.

Figure 7. The PI3K/Akt and MAPK/ERK pathways positively regulate Fra-1 and c-Jun expression and are associated with increased AP-1 transcriptional activity in TNBC cells. A,

Western blot analysis of pAkt, Akt, Fra-1, c-Jun, and β-actin (as a loading control) in BT549 cells treated with DMSO or 25 μM LY294002 for 2, 6 and 24 hours. B, Western blot analysis of pERK1/2, ERK1/2, Fra-1, c-Jun, and β-actin (as a loading control) in BT549 cells treated with DMSO or 25 μM PD98059 for 2, 6 and 24 hours. C, BT549 cells were transfected with the 2×TRE reporter plasmid or the PGL3-promoter vector. After transfection, cells were treated with DMSO, LY294002 (25 μM) or PD98059 (25 μM) for 24 hours and assayed for luciferase activity. Data are shown as means with SD. ***P < 0.001 compared to DMSO (n =

3). D, proposed model of AP-1 promoting cell proliferation and invasion in TNBC. Activation of the PI3K/Akt or MAPK/ERK pathway increases expression of AP-1. AP-1 represses

CLCA2 expression through direct binding to its intron region. AP-1 mediates downregulation of E-cadherin through direct transcriptional induction of ZEB2.

Genome-wide profiling of AP-1 regulated transcription provides insights into the invasiveness of triple-negative breast cancer

Chunyan Zhao, Yichun Qiao, Philip Jonsson, et al.

Cancer Res Published OnlineFirst May 15, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-13-3396

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2014/05/15/0008-5472.CAN-13-3396.DC1

Author Author manuscripts have been peer reviewed and accepted for publication but have not yet been Manuscript edited.

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/early/2014/05/15/0008-5472.CAN-13-3396. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.