c-Myb Enhances Breast Cancer Invasion and Metastasis through the Wnt/β-Catenin/Axin2 Pathway

Increased expression of transcription factor c-Myb is associated with malignant human cancers (1). In colorectal cancer, c-Myb has been reported to cooperate with activated β-catenin to activate the c-Myc promoter and enhance c-Myc expression (2). In addition, c-Myb overexpression, amplification, and dysregulation have been detected in some subtypes of breast cancer (3). In estrogen receptor–positive (ER⁺) cells, c-Myb is induced by estrogen/ER signaling, required for tumorigenesis, and inhibits apoptosis in primary tumor cells (4, 5). c-Myb is also involved in the TGFβ–induced epithelial-to-mesenchymal transition (EMT) process (6). In addition, increasing evidence indicates that c-Myb is found in invasive breast cancer associated with poor prognosis (7). However, the regulation of c-Myb activity and its downstream signaling networks in these processes remain largely unclear.

Wnt proteins are secreted cytokines that play pivotal roles in development, cancer, and other diseases. Canonical Wnt/β-catenin signaling is initiated when Wnt ligands interact with a heterodimeric transmembrane receptor complex, which consists of Frizzled (Fz) plus a low density lipoprotein receptor–related protein (LRP; ref. 8). Upon Wnt binding, the receptor complex, LRP, is phosphorylated, which triggers disruption of the β-catenin “destruction complex.” This enables β-catenin to translocate to the nucleus and form a complex with TCF/LEF (T-cell specific transcription factor/lymphoid enhancer-binding factor) that drives Wnt-induced gene expression (9). In breast cancer, Wnt-induced signaling is a promoter in EMT and invasion/metastasis (10), and components of the canonical Wnt signaling pathway have been identified as important mediators of breast cancer metastasis and as markers of poor prognosis (11). A number of Wnt target genes have been linked to oncogenesis, including Cyclin D1 and Axin2. Activation of Cyclin D1 is associated with poor clinical outcome (12). Axin2 is a Wnt suppressor and mediates negative feedback. However, Axin2 also has tumor-promoting functions; it can, for example, trigger an EMT-like process in breast cancer cells via stabilization of Snail (13). In the current study, we demonstrate a mechanistic link between c-Myb and Wnt/β-catenin signaling in basal breast cancer cells, which plays an important role in invasion and metastasis.

Human MDA-MB-231-luc (BM) cells were obtained as mentioned previously (14). HEK293T and MDA-MB-231 cells were provided by ATCC. 4T1 and MCF10A-M2 cells were a kind gift from Dr. Fred Miller (Barbara Ann Karmanos Cancer Institute, Detroit, MI). The 4T1 cells were used within 20 passages. The human cell lines have been authenticated by short tandem repeat (STR) profiling within 6 months of the last experiment. 4T1, HEK293T, and MDA-MB-231 cells are cultured in DMEM (high glucose with l-glutamine) supplemented with 10% FBS and 100 U/mL penicillin/streptomycin (Gibco). The MCF10A-M2 cells were maintained as described previously (15). All cell lines were maintained at 37°C with 5% CO₂.

Immunoprecipitation and Western blotting was performed as described previously (16). Actin or Tubulin protein levels were used as loading controls. We used the following primary antibodies: anti-N-cadherin (610920, BD Transduction Laboratories), anti-Slug (9585, Cell Signaling Technology), anti-β-catenin (610154, BD Transduction Laboratories), anti-c-Myb (12319s, Cell Signaling Technology), anti-LEF-1 (2230, Cell Signaling Technology), anti-Axin2 (2151, Cell Signaling Technology), anti-tubulin (2148, Cell Signaling Technology), anti-HA (1583816, Roche), anti-Flag (F3165, Sigma), and anti-β-actin (A5441, Sigma). All secondary antibodies were from Sigma.

RNA was isolated using the NucleoSpin RNA II kit (BIOKE) followed by reverse transcription PCR (RevertAid First Strand cDNA Synthesis Kits, Fermentas). The cDNAs were analyzed by real-time quantitative PCR (qPCR), and calculations were performed using the CFX Manager software version 2.0 (Bio-Rad). All mRNA expression levels were analyzed in triplicate and normalized to GAPDH expression. The primer sequences are listed in Supplementary Materials and Methods.

The zebrafish embryonic invasion and metastasis assay was performed as described previously (17). All the zebrafish data are representative of at least two independent experiments and performed according to the guidelines for the use of laboratory animals in Leiden University (Leiden, the Netherlands).

Mice experiments were approved by the Committee for Animal Welfare in Zhejiang University (Zhejiang). For the intracardial injection assay, 5-week-old female BALB/c nude mice were anesthetized with isofluorane and 1 × 10⁵ fresh cells were injected into the left ventricle of the heart. Five weeks after injection, all mice were sacrificed and metastases were analyzed. The mouse mammary fat pad implantation was based on previously published methods (18). Female BALB/c mice were anesthetized and used for this assay. A total of 2 × 10⁵ 4T1 cells were injected through the nipple area into the mammary fat pad. At 21 days after injection, primary tumor was analyzed. Then, the mice were euthanized and analyzed for secondary tumor. All primary/metastatic tumors were detected by bioluminescence imaging (BLI) with the IVIS 100 (Caliper Life Sciences). The BLI signal intensity was quantified as the sum of photons within a region of interest given as the total flux (photons/second).

The results are expressed as the mean ± SD. For analysis, we used Student t test and P < 0.05 was considered to be statistically significant. *, 0.01 <P < 0.05; **, 0.001 <P < 0.01; ***, P < 0.001. Two-way ANOVA was used to analyze the curves of zebrafish embryos displaying tumor progression. The correlations between gene expression were analyzed by Pearson coefficient tests. For analyzing TMA results, the χ² test was used to calculate correlation between c-Myb and Axin2.

To investigate the function of c-Myb in breast cancer progression, we first analyzed copy number alterations of c-Myb in the TCGA Research Network (http://cancergenome.nih.gov) by cBioportal (19, 20). c-Myb is amplified in more than 2% of breast cancer patients, whereas 15% show c-Myb gain (Supplementary Fig. S1A). Notably, c-Myb amplification and gain is associated with a poor clinical outcome (Supplementary Fig. S1B). Oncomine plots show that c-Myb is highly expressed in both invasive ductal and lobular breast cancer, suggesting that c-Myb may play a role in tumor progression (Fig. 1A; refs. 20–23). In addition, elevated c-Myb expression has also been observed in ER⁻ breast cancers, which are frequently invasive and metastatic (24).

To study the function of c-Myb in breast cancer progression, which is associated with enhanced breast cancer migration and invasion (25), we silenced endogenous c-Myb expression in two invasive basal breast cancer cell lines, MDA-MB-231 and MCF10A-M2, in which c-Myb protein was readily detected. Figure 1B showed the c-Myb knockdown efficiency in these cell lines at both mRNA and protein levels. Depletion of c-Myb resulted in a significant decrease in cell invasiveness in a Transwell assay in both cell lines (Fig. 1C). In line with the putative tumor-promoting effect of c-Myb, we found that the mRNA levels of various mesenchymal markers, including Twist, Snail, and Slug, were decreased in these cells after c-Myb knockdown (Fig. 1D). Immunoblotting showed decreased protein expression of N-cadherin, Slug, and Snail in c-Myb–depleted cells (Fig. 1E). In addition, there was a positive correlation between the mRNA expression of c-Myb and mesenchymal makers like N-Cadherin, Snail, and Slug in 189 breast cancer patients in a publicly available microarray database (GSE2990; ref. 26; Fig. 1F). We also examined the prognosis of the patients from this particular dataset and found that high expression of c-Myb is associated with a poorer clinical outcome (Fig. 1G). These clinical results are consistent with the experimental data indicating that c-Myb has a promoting role in breast cancer progression.

We next tested the role of c-Myb on malignant breast cancer in vivo by using a previously published zebrafish embryo xenograft model (15), in which injected red fluorescence–labeled tumor cells can circulate in the blood vessels, extravasate from circulation, and eventually metastasize into the tail fins (Fig. 2A). Embryos transplanted with c-Myb–depleted MDA-MB-231 cells showed much less tumor progression than embryos injected with control cells (Fig. 2B and C). At all time points of the experiment, both the average invasive area and the percentages of embryos exhibiting metastasis were significantly lower in the two groups injected with c-Myb–depleted cells (Fig. 2D and E). Compared with MDA-MB-231, MCF10A-M2 cells show moderate metastatic potential in this zebrafish model and form a cluster of cells between the dorsal aorta and the caudal vein (17). Ectopic expression of c-Myb in MCF10A-M2 cells enhanced migration in vitro (Supplementary Fig. S2A and S2B), in line with the proinvasive function of c-Myb revealed in Fig. 1C. In the zebrafish xenograft model, forced c-Myb expression in MCF10A-M2 cells resulted in increased metastasis and a larger area of invasion (Fig. 2F–I). When the control MCF10A-M2 cells were implanted, 24% of the embryos showed tumor mass formation at 6 days postimplantation. Ectopic expression of c-Myb increased this percentage nearly 2-fold (45% of embryos), further confirming the positive role of c-Myb in breast cancer progression (Fig. 2J).

We next investigated which signaling networks act at downstream of c-Myb to modulate breast cancer progression. For this, we analyzed the association between c-Myb and candidate signaling pathways, including the TGFβ, Wnt, NFκB, and Notch pathways, using the TCGA Research Network (data not shown; ref. 20). We observed a significant positive correlation between c-Myb and three metastasis-associated direct targets of Wnt, namely Axin2, LEF-1, and Cyclin D1 among 534 breast cancer patients (Fig. 3A; ref. 27). Consistent with this notion, ectopic c-Myb expression in HEK293T cells consistently and significantly increased the transcriptional activity of the canonical Wnt reporter TOPFlash-Luc in a dose-dependent manner (Fig. 3B). We subsequently investigated the possibility that c-Myb could function as an activator of the Wnt/β-catenin pathway during breast cancer progression. In line with this hypothesis, the mRNA levels of the Wnt target genes Axin2 and Cyclin D1 were remarkably decreased in c-Myb–deficient MDA-MB-231 and MCF10A-M2 cells upon stimulation with recombinant Wnt3a (100 ng/mL; Fig. 3C). These results suggested that c-Myb is required for the expression of Wnt target genes in these basal breast cancer cells.

The activation of transcription by canonical Wnt signaling is triggered by the binding of nuclear β-catenin to TCF/LEF transcription factors on target promoters. We previously reported that c-Myb can interact with LEF-1 in leukemic cells (28). Interestingly, in HeLa and Cos-7 cells, ectopic c-Myb colocalized in the nucleus with a degradation-resistant β-catenin mutant (SA; Supplementary Fig. S3). Moreover, the ectopically expressed HA-c-Myb was found to strongly interact with Flag–β-catenin in HEK293T cells (Fig. 3D). Coimmunoprecipitation (co-IP) experiments in MDA-MB-231 and MCF10A-M2 cells also showed an interaction between endogenous c-Myb and β-catenin (Fig. 3E). These findings suggested that c-Myb can activate canonical Wnt target genes by binding to β-catenin and/or LEF-1.

Subsequent analysis of the binding of c-Myb to progressive deletion mutants of β-catenin showed that the amino-terminal activating arm (NTAA) of the armadillo repeat (ARM) domain of β-catenin (3–6 repeats) is essential for the interaction (Fig. 3F). This ARM motif is important for the binding of β-catenin cofactors involved in transcriptional activation of Wnt target genes, and in addition for TCF/LEF binding (4–10 repeats; ref. 29). Importantly, we found that the interaction between β-catenin and LEF-1 was reduced when c-Myb was knocked down in MDA-MB-231 cells (Fig. 3G). Moreover, chromatin immunoprecipitations showed that depletion of c-Myb inhibited the recruitment of β-catenin to the promoters of Axin2 and Cyclin D1 (Fig. 3H). Taken together, these data suggest that c-Myb acts as a β-catenin cofactor that stabilizes its interaction with TCF/LEF and thereby can enhance Wnt signaling–induced transcription.

As a direct target gene of Wnt/β-catenin, Axin2 not only mediates negative feedback, but also can enhance EMT and cancer progression (13). In line with the RNA data in Fig. 3, the Axin2 protein level was strongly decreased in c-Myb–depleted MDA-MB-231 cells (Fig. 4A). We therefore examined the role of Axin2 in c-Myb–induced breast cancer invasion/metastasis. For this, we depleted Axin2 in both normal MDA-MB-231 cells and MDA-MB-231 cells ectopically overexpressing c-Myb, in which the endogenous Axin2 protein levels are increased by c-Myb (Fig. 4B). Figure 4B also shows that Axin2 expression was efficiently blocked by Axin2-specific shRNA. These MDA-MB-231 cells were subsequently injected into the duct of Cuvier of zebrafish to analyze their metastatic potential (Fig. 4C and D). Compared with the embryos injected with the control cells, the embryos implanted with c-Myb–overexpressing cells showed a higher tumor area and an increased percentage of tumor progression to the tail fin (70% vs. 53% in controls; Fig. 4E and F). Moreover, the depletion of Axin2 significantly repressed invasion and metastasis in both the control MDA cells and the c-Myb–overexpressing cells (Fig. 4E and F). To confirm that Axin2 is required for the c-Myb–mediated breast cancer progression, we ectopically expressed Axin2 in c-Myb–depleted MDA-MB-231 cells (Supplementary Fig. S4A). The ectopic expression could partly restore tumor progression in the zebrafish xenografts (Supplementary Fig. S4B–S4D). Thus, the c-Myb–induced zebrafish metastasis of these basal breast cancer cells is dependent on Axin2.

Because Axin2 turned out to be a critical target gene of c-Myb in breast cancer cell invasion, we further analyzed the mechanism by which c-Myb activates Axin2 gene expression. As c-Myb is a sequence-specific DNA-binding protein, we examined whether it can bind and regulate the Axin2 promoter directly. Ectopic expression of c-Myb could significantly induce Axin2 promoter-driven luciferase reporter in HEK293T cells (Fig. 4G; ref. 30). We identified three potential c-Myb–binding motifs (TAACt/gG) in the Axin2 promoter at positions -923, -728, and -614 (31). Next, we performed chromatin immunoprecipitations to specifically analyze the binding of c-Myb to the region encompassing the two proximal c-Myb–binding sites (c-Myb binding 1, -745 to -598), and the region encompassing the more distal site (c-Myb binding 2,-959 to -858; Fig. 4H). Enrichment for c-Myb was detected at both c-Myb–binding regions in MDA-MB-231 cells (Fig. 4I), and there was only a slight trend of suppressed binding of c-Myb to the Axin2 promoter upon β-catenin knockdown (Supplementary Fig. S4E). Together, these results indicate that c-Myb can directly bind to the Axin2 promoter to stimulate Axin2 expression and induce an invasive phenotype in MDA-MB-231 cells.

We next examined the role of c-Myb in metastasis in mice. First, we found that both c-Myb and Axin2 are upregulated in a MDA-MB-231-BM (14), a highly metastatic subclone of MDA-MB-231 that preferentially disseminates to bone (Fig. 5A). We subsequently depleted endogenous c-Myb expression in MDA-MB-231 BM cells (Supplementary Fig. S5A) and performed intracardial xenograft experiments in a mouse model for breast cancer metastasis (32). Mice injected with c-Myb–depleted BM cells developed less bone metastases after 36 days (Fig. 5B). Both BLI signal and number of metastatic nodules were significantly decreased in the c-Myb–depleted groups (Fig. 5C and D). Importantly, the c-Myb–depleted cells did not show significant proliferation defects (Supplementary Fig. S5B), in agreement with previously published experiments for ER⁻ breast cancer cells (4). These results suggest that c-Myb is critical for breast cancer bone metastasis. We next examined the role of c-Myb on metastatic dissemination from the primary tumor site to distant organs. For this, we orthotopically transplanted control 4T1 breast cancer cells or c-Myb–depleted 4T1 cells into the mammary fat pad of the BALB/c mice. Consistent with the positive effects of c-Myb on mesenchymal and metastatic breast cancer traits, depletion of c-Myb blocked slug and axin2 expression in 4T1 cells (Supplementary Fig. S5C and S5D). At 21 days after orthotopic implantation, primary tumor formation of 4T1 cells with c-Myb knockdown was similar to that of the control 4T1 cells (Fig 5E). However, the c-Myb–deficient 4T1 cells displayed a significant reduction of both number and size of lung metastasis (Fig 5F–H and Supplementary Fig. S5E). Together, these data show that c-Myb also contributes to breast cancer metastasis in an orthotopic mouse model.

To determine the clinical relevance of the relation between c-Myb and Axin2 shown above, we performed immunohistochemical staining of tissue microarrays (TMA) that included 176 breast cancer samples collected at Leiden University Medical Centre (Leiden, the Netherlands; Supplementary Table S1; Supplementary Fig. S6A). Nuclear expression of c-Myb and cytoplasmic expression of Axin2 were subjectively scored as high staining or low staining (Fig. 6A). As Fig. 6B shows, c-Myb upregulation positively associated with lymph node invasion. Next, we investigated the correlation between c-Myb and Axin2 in the same tissue microarray. Consistent with the result that c-Myb activates Axin2 in breast cancer, we observed that more than 70% of Axin2 high expression samples also have high c-Myb expression (Fig. 6C). Vice versa, the c-Myb levels were significantly positively correlated with Axin2 expression (Fig. 6D). These data support our findings that c-Myb enhances Axin2 expression and that this induces an aggressive phenotype and poor prognosis in human breast cancer.

To search for mechanisms that contribute to c-Myb expression in breast cancer, we tested several cytokines to see whether they regulated c-Myb expression in the MDA-MB-231 breast cancer cell line. The inflammatory cytokine IL1β was found to increase c-Myb mRNA levels (Fig. 7A). Similarly, the c-Myb protein level increased significantly from 1 to 10 hours after 10 ng/mL IL1β treatment, peaking at 3 hours (Fig. 7B). We also found a slight trend of increased Axin2 mRNA expression upon 10-hour IL1β treatment, but c-Myb depletion blocked Axin2 induction in both mock or IL1β–treated cells (Supplementary Fig. S6B). Members of the IL1 family of cytokines are abundantly expressed in the microenvironment of breast cancer and are indispensable for tumor progression (33). To test whether c-Myb is functionally involved in the IL1β response, we examined the effect of c-Myb depletion on IL1β-stimulated cell migration. Interestingly, both basal and IL1β-induced MDA-MB-231 cell migration was effectively blocked by c-Myb knockdown (Fig. 7C). Moreover, EGF-induced migration was also inhibited in the c-Myb–deficient MDA-MB-231 cells (Supplementary Fig. S6C). Axin2 and β-catenin were also required for ILβ-induced cell migration (Supplementary Fig. S6D). These data indicate that one or more (mesenchymal) target genes of c-Myb and β-catenin are required both for basal migration and for migration induced by extracellular stimuli. The stimulation of c-Myb by IL1β prompted us to reanalyze the TCGA dataset by PROGgene (20, 34). For the ER⁻ breast cancer patients, high expression of both IL1β and c-Myb mRNA was significantly associated with poor prognosis (Fig. 7D). Collectively, these results indicate that IL1β is associated with c-Myb activation and leads to poor survival in ER⁻ breast cancer.

Our study suggests a model in which c-Myb is activated by IL1β (and other extracellular cues) and interacts with β-catenin in the nucleus. c-Myb then acts as a cofactor of β-catenin and accelerates expression of genes that are downstream of β-catenin in breast cancer. In addition, c-Myb also directly binds to the Axin2 promoter and drives Axin2 expression enhancing breast tumor progression (Fig. 7E).

c-Myb expression can be induced by ER and high c-Myb predicts good survival in ER⁺ breast cancers that can be efficiently treated (4, 7, 35). The role of Myb in EMT, invasion, and metastasis is less clear, and is likely to depend on the exact genetic changes and pathway defects in the tumor and on the tumor microenvironment. Transcriptional events mediated by estrogen signaling have a distinct role in development and proliferation in the early stages of a tumor, and the ER target GATA3 and can negatively regulate metastasis (36), whereas FOXA1 positively correlates with distant metastasis (37). Moreover, c-Myb can regulate tumor migration in vitro in both ER⁺ and ER⁻ cell lines (6, 38). Previously, Hugo and colleagues concluded that the role of Myb in EMT is most likely context-dependent. These authors found that the EMT-activator Zeb1 can inhibit c-Myb expression, and that c-Myb and Zeb1 are inversely expressed in PMA42 and MDA-MB-231 cells (39). Hugo and colleagues postulated that Myb could drive early EMT by inducing Slug and Snail, but that at later stages Zeb1 could suppress Myb and also Slug and Snail. It should also be noted in this respect that depending on the cellular context excess of c-Myb (e.g., by overexpression) might affect different Myb target genes and functions than Myb depletion. In the current study, we found that c-Myb is specifically required for breast cancer invasion and metastasis in experimental xenograft models in zebrafish and mice. As one of the signature proteins that correlates with ER status, c-Myb has also been implicated in survival in luminal/ER⁺ breast cancer and is linked to the regulation of BCL2 (20, 40). However, we observed no significant proliferation or survival defects after knockdown or ectopic expression of c-Myb/Axin2 in ER⁻ cell lines (Supplementary Fig. S7). This is in line with previous studies that c-Myb knockdown does not influence the proliferation of ER⁻ cells (4).

One of the signaling pathways involved in breast cancer progression is the Wnt pathway, which can induce both EMT and invasion/metastasis (10). Importantly, we found c-Myb to enhance the expression levels of various target genes of Wnt/β-catenin signaling, including Cyclin D1 and Axin2, by functionally interacting with LEF-1/β-catenin and facilitating β-catenin binding to its target promoters. In addition, c-Myb was found to bind to the Axin2 promoter directly, and thereby to induce Axin2 expression and enhance breast tumor progression. High expression of Cyclin D1 promotes local invasion and is associated with bone metastasis (41). Axin2 can fine-tune Wnt signaling and counteract excessive signaling as a Wnt feedback inhibitor, but, more importantly, Axin2 also acts as a chaperone for GSK3β and controls its subcellular location in human breast cancer. As a result of this, Snail is dephosphorylated and stabilized, contributing to EMT and invasion (13). Moreover, we previously demonstrated that Axin2 can form a complex with the E3 ligases Rnf12 and Arkadia and induce Smad7 degradation to promote TGFβ signaling and breast bone metastasis (42). Importantly, we found a positive correlation between c-Myb and Axin2 in clinical samples. We further show that c-Myb expression in breast cancer cells can be activated by IL1β, which suggest that c-Myb may in particular enhance migration and invasion in response to inflammatory signals in the tumor microenvironment. In this respect, it is interesting to note that c-Myb can contribute to TGFβ-induced EMT in ER⁺ breast cancer cells (6). It might therefore be interesting to investigate the IL1β/TGF-β/Wnt interplay in this process.

The exact mechanism by which c-Myb enhances Wnt/β-catenin–inducible genes remains to be established. β-Catenin can interact with both LEF1/TCF and Wnt coactivators/repressors through its ARM domain, which consists of 12 repeats (43). Our results demonstrated that c-Myb binds to the NTAA motif of the ARM domain. Consistent with this model, we observed decreased expression of Wnt target genes and reduced binding affinity of β-catenin to the corresponding promoters when c-Myb was knocked down in MDA-MB-231 cells. c-Myb might therefore act as a β-catenin coactivator. A previous study showed that in colorectal cancer c-Myb can also cooperate with activated β-catenin via independent binding to the c-Myc promoter, which highlights the potential interplay between c-Myb and Wnt activation during cancer development (2).

Our results further suggest that c-Myb positively controls various mesenchymal markers like Twist, Snail, Slug, and N-cadherin to contribute to local invasion and distant metastasis of cancer. Moreover, Wnt signaling can promote EMT by inducing the expression of these mesenchymal markers during breast cancer invasion (11) further supporting our observation that Wnt-β-catenin signaling might contribute to c-Myb–induced cancer progression. Depletion of c-Myb in MDA-MB-231 cells was found to impair breast cancer bone metastasis in a mouse model. Therefore, c-Myb appears to be required for the spreading of advanced metastatic cells. In line with this, we found that silencing of endogenous c-Myb expression in 4T1 cells has no significant effect on primary formation but suppresses distant metastasis. Interestingly, a previous report showed that c-Myb overexpression in 4T1 cells upregulates the expression of matrix metalloproteinases (MMP) 9 and cathepsin D but inhibits MMP1, and promotes migration and Matrigel invasion, but not collagen invasion. In addition, c-Myb overexpression in 4T1 cells was found to delay primary tumor growth and to affect metastasis in an organ-specific manner (38).

Long-term clinical observations and mechanistic studies have shown a strong association of inflammation with cancer development and progression. Inflammation can result in an influx of inflammatory cells and secretion of inflammatory cytokines including TNFα, IL1, IL6, and IL8 into the tumor microenvironment, leading to activation of the NFκB pathway (44). Clinically, high c-Myb and IL1β expression were found to be linked with poor survival in ER⁻ breast cancer. Therefore, our results suggest that levels of c-Myb and IL1β might be important to predict the aggressiveness of specific breast cancers.

No potential conflicts of interest were disclosed.

Conception and design: Y. Li, P. ten Dijke, F. Zhou, L. Zhang

Development of methodology: Y. Li, L. Zhang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Li, K. Jin, G.W. van Pelt, W.E. Mesker, P. ten Dijke, L. Zhang

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Li, G.W. van Pelt, H. van Dam, X. Yu, P. ten Dijke, L. Zhang

Writing, review, and/or revision of the manuscript: Y. Li, G.W. van Pelt, H. van Dam, W.E. Mesker, P. ten Dijke, F. Zhou, L. Zhang

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): K. Jin, G.W. van Pelt, F. Zhou

Study supervision: P. ten Dijke, F. Zhou

The authors thank Dr. Odd Stokke Gabrielsen for the HA-c-Myb plasmid, Dr. Frank Costantini for Axin2 reporter constructs, Dr. Zhijie Chang for the TOPFlash reporter, and all members of ten Dijke laboratory for discussion.

This work is supported by the Cancer Genomics Centre Netherlands, Swedish Cancerfonden (09 0773), Zhejiang Provincial Natural Science Foundation of China (grant number R14C070002), and The Chinese National Natural Science Funds (grant number 31471315).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.