AP-1 Is a Key Regulator of Proinflammatory Cytokine TNFα-mediated Triple-negative Breast Cancer Progression*

Triple-negative breast cancer (TNBC) represents a highly aggressive form of breast cancer with limited treatment options. Proinflammatory cytokines such as TNFα can facilitate tumor progression and metastasis. However, the mechanistic aspects of inflammation mediated TNBC progression remain unclear. Using ChIP-seq, we demonstrate that the cistrome for the AP-1 transcription factor c-Jun is comprised of 13,800 binding regions in TNFα-stimulated TNBC cells. In addition, we show that c-Jun regulates nearly a third of the TNFα-regulated transcriptome. Interestingly, high expression level of the c-Jun-regulated pro-invasion gene program is associated with poor clinical outcome in TNBCs. We further demonstrate that c-Jun drives TNFα-mediated increase of malignant characteristics of TNBC cells by transcriptional regulation of Ninj1. As exemplified by the CXC chemokine genes clustered on chromosome 4, we demonstrate that NF-κB might be a pioneer factor required for the regulation of TNFα-inducible inflammatory genes, whereas c-Jun has little effect. Together, our results uncover AP-1 as an important determinant for inflammation-induced cancer progression, rather than inflammatory response.

Triple-negative breast cancer (TNBC), 3 which lacks expression of estrogen receptor (ER) and progesterone receptor (PR) as well as HER2 amplification, is more aggressive and has poorer survival than other types of breast cancer (1). TNBC is a clinically heterogeneous disease and accordingly, TNBC has recently been classified into five clusters, each harboring a dominant biological function/pathway (2). These are basal-like TNBC with DNA-repair deficiency and growth factor pathway expression, mesenchymal-like TNBC with epithelial-mesenchymal transition (EMT) and cancer stem cell features, immune-associated TNBC, luminal/apocrine TNBC with androgen-receptor (AR) overexpression, and HER2-enriched TNBC. Targeted therapies such as poly-ADP-ribose polymerase (PARP) inhibitors targeting DNA-repair deficiency, AR inhibitors and anti-EGFR therapies are currently under clinical evaluation. However, despite the ongoing development of novel targeted therapies for TNBC, there is a recognized need for additional approaches for each cluster and additionally the clinical success of the approaches above remains to be proven.
The AP-1 family of transcription factors consist of multiple Jun (c-Jun, JunB, and JunD) and Fos (c-Fos, FosB, Fra1, and Fra2) members. AP-1 regulates gene transcription through binding to 12-O tetra-decanoyl-phorbol-13 acetate (TPA)-responsive elements (TRE, 5Ј-TGAG/CTCA-3Ј) or cAMP response elements (CRE, 5Ј-TGACGTCA-3Ј) via a basic leucine zipper (bZIP) domain, which is composed of a DNA-binding basic region and the leucine zipper dimerization region (3). AP-1 has been implicated in regulating various physiological and pathological cellular processes including proliferation, differentiation, growth, apoptosis, cell migration, and transformation (4). AP-1 inhibitors have therefore emerged as an attractive drug class for cancer therapy. AP-1 inhibition-specific retinoids such as SR11302 and MX781 have long been known to have antitumor effects in vitro and in vivo (5)(6)(7)(8). In particular, the retinoid antagonist MX781 showed promising anticancer activities against advanced ER-negative breast cancer (5). The activation of AP-1 can be regulated by multiple mechanisms, including dimer composition, post-translational modifications and interactions with ancillary proteins (9). Our recent studies showed that AP-1 proteins, mainly Fra-1 and c-Jun, are highly expressed in TNBC compared with other types of breast cancer (10,11), underscoring the importance of unraveling the AP-1 signaling pathway in TNBC.
Recent evidence indicates that proinflammatory signaling activated by inflammatory stimuli in the tumor microenvironment facilitates tumorigenesis (12). The inflammatory tumor microenvironment is largely orchestrated by various inflammatory cells, particularly tumor-associated macrophages, which secrete proinflammatory cytokines such as TNF␣, IL-1, IL-6, and TGF␤ (12). It is well recognized that TNF␣ signaling, via ligand binding to TNF receptors, induces a range of inflammatory genes through activation of the nuclear factor-B (NF-B) pathway (12). In addition, it is known that NF-B activation plays critical roles in inflammation-induced tumor growth and metastasis (13). In contrast, much less is known about the nature of proinflammatory cytokine activation of AP-1 pathways.
Here, we define for the first time the inflammatory cistrome for the AP-1 transcription factor c-Jun and demonstrate that AP-1 activation is responsible for inflammation-induced malignant characteristics of TNBC cells, rather than the inflammatory response. We identify a set of c-Jun-regulated pro-invasion genes that are strongly associated with clinical outcomes in TNBCs. In particular, we characterize the Ninj1 gene, which is transcriptionally regulated by c-Jun to drive TNF␣-mediated malignant characteristics of TNBC cells. Our study reveals a critical mechanism underlying inflammation-induced cancer progression and may have important implications for the development of targeted therapies for metastatic breast cancers.

Experimental Procedures
Cell Lines and Reagents-BT549 and Hs578T cells were obtained from the American Type Culture Collection. Human recombinant TNF␣ was purchased from Roche and was used at 10 ng/ml in all experiments.
Gene Expression Microarray Analysis-BT549 cells were cultured to 50% confluency and transfected with control or c-Jun siRNA for 72 h, followed by treatment with or without TNF␣ for 6 h. Total RNA was extracted using the RNeasy Plus Mini Kit (Qiagen). Total RNA from three biological replicates was hybridized to Affymetrix Human Gene 1.1 ST arrays. The data were analyzed as previously described (15). The microarray raw data are deposited into GEO (accession number GSE71977). Gene ontology analysis was performed using the DAVID Functional Annotation Bioinformatics Tool (16,17). A given ontology term within each gene set was considered to be overrepresented on the basis of the p value (Ͻ 0.05), which is calculated by comparing the observed frequency of an annotation term with the frequency expected by chance.
qPCR-Total RNA was extracted, reverse transcribed, and subjected to real-time qPCR using gene-specific primers, as previously described (18). The expression of genes was normalized to the expression of ␤-glucuronidase and 36B4. Similar results were obtained for both reference genes, with only 36B4 normalized values presented.
ChIP and ChIP-Seq-ChIP and ChIP-Seq were performed as previously described (18,19). The antibodies used were as follows: anti-c-Jun (H-79) and normal rabbit IgG from Santa Cruz Biotechnology, anti-p65 (ab7970) and anti-p50 (ab7971) from Abcam. ChIP-Seq data are deposited in GEO (accession number GSE71977).
Invasion Assay-Cells were serum-starved overnight after siRNA transfection and then treated with or without TNF␣ for 48 h. Invasion assays were conducted as previously described (11) but with the following modification. After treatment with TNF␣, cells were trypsinized and stained with propidium iodide and the number of apoptotic cells were determined by flow cytometry. Samples were adjusted to equal numbers of cells, excluding apoptotic cells, before applying cells to transwell chambers.
Cell Death Detection ELISA-Cells were serum-starved overnight after siRNA transfection and then treated with or without TNF␣ for 48 h. Apoptosis was determined with the Cell Death Detection ELISA Plus Kit (Roche) in accordance with the manufacturer's instructions.
Kaplan-Meier Analyses-Kaplan-Meier survival curves were generated within the GOBO breast cancer datasets using the GOBO tool (co.bmc.lu.se/gobo/) (20). Results were stratified into two quintiles based on expression of the c-Jun-regulated pro-invasion gene set.
Statistical Analysis-Data are presented as mean Ϯ S.D. Comparisons were performed by two-tailed Student's t test for single comparison or by one-way ANOVA followed by Bonferroni correction for multiple comparisons as appropriate. p values Ͻ 0.05 were considered as significant.

TNF␣ Signaling Generates a Unique AP-1 Cistrome in TNBC
Cells-We previously demonstrated that the inflammatory cytokine TNF␣ triggered EMT of TNBC cells through AP-1induced ZEB2 up-regulation (14). However, the AP-1 signaling network responsive to TNF␣ remains largely unexplored. To begin to explore the AP-1 signaling network responsive to TNF␣, we performed genome-wide mapping of c-Jun binding regions in BT549 TNBC cells treated with TNF␣ for 3 h. The resulting cistromes correspond to 4,570 and 13,800 binding sites in non-stimulated and TNF␣-stimulated cells, respectively (Fig. 1A). Our analysis reveals that the de novo TNF␣induced c-Jun cistrome corresponds to 9230 binding regions. It has been described that c-Jun regulates its own expression through AP-1 sites in the promoter of the c-Jun gene (21). In line with this, a clear peak was identified in the promoter of the c-Jun gene itself under untreated and TNF␣-stimulated conditions (Fig. 1B). De novo DNA motif search revealed several highly enriched motifs. As expected, the consensus AP-1 motif was the most enriched motif in c-Jun cistromes in the presence and absence of TNF␣ (Fig. 1C). Interestingly, the cistrome in the presence of TNF␣ was enriched for the NF-B motif. Because TNF␣ stimulation could conceivably lead to activation of NF-B (12), these findings suggest that c-Jun might bind sites occupied by NF-B through mechanisms such as tethering. The overall distribution of c-Jun binding regions in relation to gene structures was very similar for non-stimulated and TNF␣stimulated cells, with most binding regions in intergenic and intronic parts of genes (Fig. 1D). To systematically address potential functional consequences of c-Jun binding regions, we used the genomics regions enrichment of annotation tool (GREAT), which is specifically suited to analyze ChIP-Seq data. The GREAT analysis suggests that TNF␣-induced c-Jun DNA binding regulates apoptosis signaling pathways and oxidative stress responses (Fig. 1E).
AP-1 Modulates the TNF␣-regulated Transcriptome in TNBC Cells-We next investigated the role of c-Jun for TNF␣regulated gene expression programs. We suppressed c-Jun expression by RNA interference (11) and investigated the genome-wide effects on TNF␣-mediated transcriptional regulation in BT549 cells. Fig. 2A shows that the expression of 1192 genes in control cells and 940 genes in c-Jun-depleted cells were activated or repressed (at least 1.5-fold), upon TNF␣ stimula- tion ( Fig. 2A). Specifically, treatment with TNF␣ resulted in activation of 684 genes (at least 1.5-fold) in control cells. Assigning the TNF␣-up-regulated genes to functional categories, immune/defense/inflammatory responses were the most frequently represented GO classes, consistent with the known roles of TNF␣. The majority (66%) of TNF␣-regulated genes in control cells remained TNF␣-responsive in c-Jun-depleted cells, but 409 genes were no longer responsive to TNF␣ upon c-Jun depletion. Furthermore, c-Jun depletion exposed a set of novel genes (157) subject to TNF␣ regulation. These results reveal that c-Jun controls more than one-third of the TNF␣regulated transcriptome.
Next, we focused on TNF␣ stimulation. We identified 616 genes, the expression of which were affected by decreased c-Jun expression in the presence of TNF␣ (Fig. 2B). To address which of the 616 genes that are direct AP-1 targets, we extracted the genes that included c-Jun binding regions within 20-kb upstream or downstream of a known transcriptional start site (TSS) in TNF␣-stimulated cells. This revealed 204 direct c-Jun target genes in TNF␣-stimulated cells (Fig. 2B). Using qPCR analysis, we confirmed changes in gene expression, as derived from microarray analysis, for 4 genes (EGR2, ZEB2, MMP9, and TNFAIP8) among the direct c-Jun target genes (Fig. 2C). Assigning the direct target genes to molecular and cellular functions, genes associated with regulation of cell proliferation, phosphorylation, cell adhesion, cell motion, intracellular signaling cascade, gene expression, apoptosis, and cellular homeostasis, among others, were highly enriched (Fig.  2D). Surprisingly, there was no enrichment for genes associated with inflammatory response. Taken together, these data imply that the main biological processes affected by c-Jun in TNF␣-stimulated cells are cancer-related functions, controlling cancer progression, rather than inflammatory responses. TNF␣ is a potent inducer of apoptosis in many tumor cell lines (22,23). Consistently, we found that TNF␣ treatment induced apoptosis in BT549 cells, as judged by induction of the apoptotic marker cleaved caspase-3 (Fig. 3A). Importantly, c-Jun knockdown further sensitized BT549 cells to TNF␣-induced apoptosis (Fig. 3A, compare lanes 4 and 2). These results were confirmed by an ELISA assay for apoptosis (Fig. 3A). The effect of c-Jun in sensitizing cells to TNF␣-induced apoptosis was confirmed in an additional TNBC cell line, Hs578T (data not shown). To examine the effect of c-Jun on TNF␣-induced invasion of TNBC cells, we performed transwell cell invasion assays for BT549 cells and observed that knockdown of c-Jun expression reduced TNF␣-induced cell invasion (Fig. 3B). Our results show that c-Jun knockdown reduced the invasion ability of non-stimulated cells, which is consistent with our previous results (11). These results suggest that c-Jun potentiates TNF␣induced TNBC cell invasion.
We identified c-Jun direct target genes with known regulatory functions in apoptosis and cell invasion. Among 204 putative c-Jun direct target genes (Fig. 2B), 23 anti-apoptotic genes such as SNAI2 were identified to be down-regulated upon c-Jun knockdown, whereas 13 pro-apoptotic genes were identified to be up-regulated upon c-Jun knockdown (Fig. 3C). Moreover, 14 genes that promote cell invasion such as ZEB2 were downregulated by c-Jun knockdown, whereas 5 genes that repress cell invasion were up-regulated by c-Jun knockdown. We further assessed whether the c-Jun-regulated pro-invasion gene set has a broader clinical significance using the Gene Expression-Based Outcome for Breast Cancer Online (GOBO) tool, expressing the outcomes in Kaplan-Meier survival plots. High expression levels of this pro-invasion gene set were shown to be associated with poor outcome in breast cancers, predominantly in basal tumors from 282 individuals (Fig. 3D). Collectively, these results indicate that the c-Jun-regulated gene expression program in response to TNF␣ has the potential to be used as prognostic tool in breast cancer, especially TNBC.
TNF␣/AP-1 Signaling Regulates Apoptosis and Cell Invasion via Ninj1-To determine the mechanisms how c-Jun affects TNF␣-regulated cancer progression, i.e. reduced apoptosis and increased migration we focused on the c-Jun target gene, Nerve injury-induced protein 1 (Ninjurin1, Ninj1) which includes a TNF␣-induced c-Jun binding site in close proximity to its TSS (Fig. 4A). Recently, it has been reported that lack of Ninj1 leads to growth suppression in colon cancer cells, implying that Ninj1 has potential oncogenic functions (24). However, its role in breast tumor growth and progression has not been studied. A highly conserved sequence with high homology to the AP-1 consensus motif was identified in the c-Jun binding region in close proximity to Ninj1 TSS (Fig. 4B). This region seems to be an active enhancer, marked by enrichment of H3K4me1, H3K4me3 and H3K27Ac (Fig. 4B). A ChIP-qPCR assay con-firmed TNF␣-induced c-Jun recruitment to the Ninj1 promoter (Fig. 4C). Furthermore, c-Jun knockdown resulted in reduced TNF␣ induction of Ninj1 mRNA expression (Fig. 4D). This finding was further validated in another TNBC cell line Hs578T (data not shown). These data demonstrate that AP-1 directly activates Ninj1 transcription.
We further analyzed how Ninj1 affects apoptosis and cell invasion in BT549 cells. We found that Ninj1 knockdown enhanced apoptosis under both untreated and TNF␣-stimulated conditions, and further enhanced TNF␣-induced apoptosis (Fig. 4E). The transwell cell invasion assay showed that Ninj1 knockdown greatly reduced the invasiveness of untreated and TNF␣-treated BT549 cells (Fig. 4F). Together, these findings indicate that Ninj1 is a potential oncogene that increases malignant characteristics of TNBC cells, in agreement with previous findings in colon cancer cells (24). To determine the role of Ninj1 in apoptosis and invasion downstream of c-Jun, we examined whether simultaneous c-Jun knockdown and Ninj1 overexpression would rescue cells from the effects of c-Jun depletion. Increased apoptosis of BT549 cells upon c-Jun knockdown was reversed by overexpression of Ninj1 (Fig. 4G). This was further confirmed by an ELISA assay for apoptosis in BT549 cells (Fig. 4G) and Hs578T cells (data not shown). Furthermore, we found that reduced cell invasiveness upon c-Jun knockdown was reversed by overexpression of Ninj1 (Fig. 4H). These results demonstrate that TNF␣/c-Jun signaling regulates apoptosis and cell invasion via Ninj1.

Identification of TNF␣-induced AP-1 Binding Sites Flanking the CXC Chemokine Cluster on Chromosome 4 -TNF␣ is well
known as a central regulator of inflammation. Consistently, our transcriptome analysis identified 62 TNF␣-induced inflammatory response genes. For example, TNF␣ treatment resulted in a striking induction of an array of CC and CXC chemokines involved in inflammatory cell recruitment (25), including Ccl2, Ccl5, Ccl7, and Ccl20, as well as Cxcl1, Cxcl2, Cxcl3, and Cxcl6. We identified a region which showed strong TNF␣-induced c-Jun binding at the promoters of IL8, CXCL1, 2, and 3 (Fig. 5A, marked by solid boxes), within a genomic region spanning 0.4 MB on chromosome 4. Quantitative ChIP-PCR experiments were performed to confirm TNF␣-induced c-Jun recruitment to the promoters of IL8, CXCL1, 2, and 3 (Fig. 5B). We further analyzed whether the TNF␣ stimulated recruitment of c-Jun leads to changes in expression of the associated genes. Surprisingly, knockdown of c-Jun only slightly impaired TNF␣-induced expression of IL8, CXCL1, 2, and 3 mRNAs (Fig. 5C), suggesting that c-Jun is not essential for chemokine gene expression.
NF-B Regulation of the CXC Chemokine Cluster-Because NF-B is an important regulator of proinflammtory gene expression, and we observed an enrichment of NF-B motifs in the c-Jun cistrome upon treatment with TNF␣, we hypothesized that NF-B, rather than c-Jun, might be a master regulator of TNF␣-inducible expression of chemokine genes. ChIP-qPCR analysis demonstrated that the NK-B family members, p65 and p50, were bound to the CXC chemokine cluster and that this binding was enhanced upon TNF␣ treatment (Fig. 6A). The binding of p65 to these sites was more pronounced than p50 (Fig. 6A). Furthermore, siRNA-mediated knockdown of p65 and p50 dramatically reduced the expression of IL8, CXCL1, 2, and 3 mRNAs in response to TNF␣ (Fig. 6, B and C). We further investigated the relationship between NF-B and c-Jun recruitment to the promoters of IL8, CXCL1, 2, and 3 loci in the presence of TNF␣. Knockdown of p65 and p50 prevented the recruitment of c-Jun to the promoters of IL8, CXCL1, 2, and 3 (Fig. 6D). In contrast, knockdown of c-Jun increased rather than decreased the recruitment of p65 and p50 (Fig. 6E). The increased binding of NF-B upon c-Jun knockdown could result from reduced competition between these factors for chromatin binding. Together, these findings indicate that NF-B might be a pioneer factor for c-Jun recruitment to this   3). C, mRNA levels of CXC chemokine genes were determined by qPCR in BT549 cells after transfection with control or c-Jun siRNA for 72 h, and followed by TNF␣ treatment for the indicated times. p values were generated by Student's t test. MARCH 4, 2016 • VOLUME 291 • NUMBER 10 gene cluster and is required for the regulation of TNF␣-inducible inflammatory genes, whereas c-Jun has little effect.

Discussion
Proinflammatory signaling in cancers activated by inflammatory stimuli such as TNF␣ in the tumor microenvironment facilitates both tumor development and metastatic progression (12). Although it has long been known that AP-1-mediated gene expression in response to inflammatory cytokines is important for the pathogenesis of a range of diseases including arthritis, septic shock, and inflammatory bowel disease (26), the role of AP-1-mediated proinflammatory signaling in tumor growth and progression remains largely unknown. In this study, we have defined molecular mechanisms that underlie TNF␣/ AP-1-mediated proinflammatory signaling, its functional consequences and potential clinical relevance in TNBCs. We demonstrate that the AP-1 transcription factor c-Jun is a mediator of TNF␣-regulated transcriptional events in TNBC cells. Highlevel expression of c-Jun-regulated pro-invasion gene set is associated with poor outcome in breast cancer, especially TNBC. Importantly, we show that the main biological processes affected by c-Jun in the presence of TNF␣ in TNBC cells are cancer-related functions, controlling cancer progression, rather than inflammatory responses.
To uncover AP-1-regulated genes in response to TNF␣ signaling, we report the first genome-wide map of c-Jun binding sites in TNF␣-stimulated TNBC cells. We identified 13,800 c-Jun binding regions in the presence of TNF␣, of which 9230 binding sites represented a de novo TNF␣-induced c-Jun cistrome. The dominant binding motif was an AP-1 motif in both the non-stimulated and TNF␣-stimulated c-Jun cistromes, supporting the high reliability of our c-Jun cistromes. Our data revealed enrichment of NF-B binding motifs within the TNF␣-stimulated c-Jun cistrome, suggesting NF-B as a pioneer factor providing points of chromatin access for the recruitment of AP-1 and other collaborating transcription factors.
It has been described that AP-1 plays a fundamental role in mediating tumor promotion induced by TNF␣ in mice through induction of a specific subset of AP-1 responsive genes (27), yet its target genes in human cells are not well characterized. To systematically address the role of c-Jun in TNF␣-regulated gene expression programs, we used c-Jun knockdown experiments in BT549 cells. BT549 cells have been classified as basal-B, specifically as mesenchymal, claudin-low, and highly invasive (28). We showed that c-Jun controls around one-third of the TNF␣regulated transcriptome, suggesting that c-Jun is an important regulator of TNF␣-driven transcriptional events. Integration of transcriptome and cistrome data identified 204 direct c-Jun target genes in the presence of TNF␣. Analyzing cellular functions of these genes based on gene ontology indicates that c-Jun, in this system, affects genes involved in cancer-related functions such as regulation of cell proliferation, cell adhesion, cell motion, and apoptosis. Taken together, these results suggest that the biological processes affected by c-Jun can influence the progression of established breast cancer.
TNF␣ is a potent cytokine that induces apoptosis by several pathways. The most widely accepted pathway involves TNFR1 interacting with TRADD, which serves as a platform for recruitment of additional mediators, including FADD and TRAF2 (29). Caspase-8 is next recruited to this protein complex, and the active caspase-8 initiates a caspase cascade, which results in apoptosis. At the same time, TNF␣ signaling also activates the NF-B survival pathway. NF-B activation antagonizes TNF␣-induced apoptosis by inducing the transcription of anti-apoptotic genes, including members of the Bcl-2 family (30). So far, the effect of AP-1 activation on TNF␣-induced apoptosis in cancer cells has not been well characterized. The role of c-Jun in TNF␣-induced apoptosis in mouse fibroblasts remains controversial. It was reported that c-Jun cooperates with NF-B to prevent apoptosis induced by TNF␣ (31). Contrary to this is the observation that c-Jun does not affect TNF␣induced apoptosis (32). Our study showed for the first time that suppression of AP-1 signaling can potentiate TNF␣-induced apoptosis in TNBC cells, which is analogous to the function of NF-B in TNF␣-induced apoptosis. TNF␣-induced apoptosis has been previously observed in other breast cancer cell lines, including the estrogen receptor-positive MCF-7 (33) and HER2 overexpressing SK-BR-3 (23).
More and more evidence indicate that inflammation plays a critical role in tumor progression, including angiogenesis, invasion and metastasis (12). Indeed, many cancers arise from sites of infection, chronic irritation and inflammation (34). In this regard, recent studies have shown that proinflammatory signaling is associated with more aggressive breast cancers (35). More recent data suggest that, depending on the breast cancer subtypes, proinflammatory signaling can be associated with either good or poor clinical outcomes (36). We show here that, in highly invasive breast cancer cells, AP-1/proinflammatory signaling potently increases cell invasion. Consistent with this observation, we previously reported that TNF␣/AP-1 signaling induces EMT in TNBC cells (14). Our results indicate a critical role for AP-1 signaling potentiates TNF␣-driven tumor aggressiveness. Furthermore, we have identified a unique transcriptome, which consists of 14 target genes directly regulated by c-Jun upon proinflammatory signaling, that offers to be used as tool to predict the outcome of breast cancer, especially TNBC.
The involvement of Ninj1 in mediating the effects of AP-1 on apoptosis and cell invasion is intriguing. Ninj1 is a small adhesion molecule that is involved in cell-cell interactions (37). It is broadly expressed not only in the injured nervous system but in all tissues of epithelial origin (37). Recent evidence indicates that Ninj1 has oncogenic potential. In line with this, up-regulation of Ninj1 expression was observed in hepatocellular carcinoma (38) and acute lymphoblastic leukemia (39). Cho et al. reported that knockdown of Ninj1 suppresses cell proliferation and enhances apoptosis in human RKO colon cancer cells (24). They further demonstrated that silencing of Ninj1 increases p53 expression through modulating p53 mRNA translation in RKO cells that express wild-type p53. Here, we provide evidence that Ninj1 silencing enhances TNF␣-induced apoptosis and inhibits cell invasion in TNBC cells, supporting that Ninj1 acts as a tumor promoter.
TNF␣ is best known as a central regulator of inflammation. Consistently, our transcriptome analysis identified 62 TNF␣induced inflammatory response genes. For example, TNF␣ stimulation resulted in a striking induction of an array of CC and CXC chemokines involved in inflammatory cell recruitment (25), including Ccl2, Ccl5, Ccl7, and Ccl20, as well as Cxcl1, Cxcl2, Cxcl3, and Cxcl6. TNF␣ triggers activation of both NF-B and AP-1 signaling pathways. While the critical role of the transcription factor NF-B for the TNF␣-induced inflammatory response is well established, the role of AP-1 signaling in cytokine/chemokine expression is not fully explored. Here, we identify a TNF␣-inducible c-Jun binding region flanking a cluster of four highly TNF␣-regulated chemokine genes. This region has been found to harbor p65 binding sites induced by interleukin-1 (IL-1) in KB cells (40). By quantitative ChIP-PCR, we confirmed the binding of both AP-1 and p65 at this CXC locus. However, our c-Jun and NF-B knockdown experiments reveal that c-Jun is not an essential component governing chemokine gene expression, whereas NF-B appears to be an important regulator. Accordingly, there was no enrichment for genes associated with inflammatory response in the c-Jun direct target transcriptome in the presence of TNF␣ (Fig. 2D). However, few c-Jun target genes, such as TNFAIP8L2 (41) and MMP9 (42) have been shown to be involved in the inflammatory response. Of relevance to our reported data, NF-B activation was essential for induction of IL-6 and IL-8 mRNA expression in rheumatoid arthritis synovial fibroblasts, whereas c-Jun has no effect (43). On the contrary, induction of IL8 by TNF␣ plus interferon ␥ has been suggested to be mediated through NF-B in cooperation with AP-1 (44). It is therefore possible that c-Jun may contribute to the expression of IL-8 in other cell types or when cells are subject to alternative stimulation. Mechanistically, we speculate that NF-B might be a pioneer transcription factor that functions to maintain an open chromatin architecture, because its absence diminishes c-Jun binding. In contrast, the absence of c-Jun did not affect, but rather increased NF-B binding, which may be due to reduced competition between these two factors for chromatin binding. Indeed, recent studies have suggested a role of NF-B as a pioneer factor that promotes an open chromatin in response to proinflammatory signaling (36,45). The role of AP-1 binding but no activation at this CXC chemokine locus remains to be determined. It is possible that the AP-1 exerts further functions such as regulating the kinetics of TNF␣-mediated gene expression or mediating long-range chromatin interactions.
In conclusion, our results revealed a novel mechanism by which the inflammatory microenvironment stimulates breast cancer progression. We propose that inhibition of AP-1 function may have therapeutic potential in aggressive breast cancer.  MARCH 4, 2016 • VOLUME 291 • NUMBER 10 JOURNAL OF BIOLOGICAL CHEMISTRY 5077