Repression of p63 and induction of EMT by mutant Ras in mammary epithelial cells

Kathryn E. Yoh; Kausik Regunath; Asja Guzman; Seung-Min Lee; Neil T. Pfister; Olutosin Akanni; Laura J. Kaufman; Carol Prives; Ron Prywes

doi:10.1073/pnas.1613417113

Contributed by Carol Prives, August 23, 2016 (sent for review December 30, 2015; reviewed by Nabeel M. Bardeesy, Leif Ellisen, and Robert A. Weinberg)

Significance

The oncogenes Harvey Rat Sarcoma Virus GTPase (H-RAS) and phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA) are well known for altering cell growth and morphology. We show here that they are also able to modify the differentiation state of mammary epithelial cells by inducing an epithelial-to-mesenchymal transition (EMT). This transition leads to greater invasiveness, a hallmark of the progression of tumors toward metastasis. Expression of p63, a protein required for the development of mammary epithelial cells, is strongly repressed by these oncogenes. In turn, loss of p63, which occurs at the transcriptional level, causes a shift in microRNAs and transcription factors that control EMT. Targeting specific genes within this Ras-to-EMT axis may be useful as a therapy to block breast cancer progression.

Abstract

The p53-related transcription factor p63 is required for maintenance of epithelial cell differentiation. We found that activated forms of the Harvey Rat Sarcoma Virus GTPase (H-RAS) and phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA) oncogenes strongly repress expression of ∆Np63α, the predominant p63 isoform in basal mammary epithelial cells. This regulation occurs at the transcriptional level, and a short region of the ∆Np63 promoter is sufficient for repression induced by H-RasV12. The suppression of ∆Np63α expression by these oncogenes concomitantly leads to an epithelial-to-mesenchymal transition (EMT). In addition, the depletion of ∆Np63α alone is sufficient to induce EMT. Both H-RasV12 expression and ∆Np63α depletion induce individual cell invasion in a 3D collagen gel in vitro system, thereby demonstrating how Ras can drive the mammary epithelial cell state toward greater invasive ability. Together, these results suggest a pathway by which RAS and PIK3CA oncogenes induce EMT through regulation of ∆Np63α.

H-Ras, a member of the Ras family of GTPases, was originally identified as the transforming protein encoded by the Harvey rat sarcoma retrovirus (1). Activating mutations in H-Ras and its family members, K-Ras and N-Ras, were identified in a variety of cancers, and nearly 30% of all cancers have mutations in one of the Ras genes (2, 3). The EGFR2 receptor, also known as HER2/neu, is an important upstream activator of Ras and is amplified in about 20–30% of breast cancers (4, 5). Although many HER2⁺ cancers are initially responsive to treatment with the monoclonal antibody Herceptin (trastuzumab), resistance usually develops (6, 7). Therefore, it is of great importance to understand signaling pathways that are downstream of HER2 and Ras, to identify key factors responsible for their tumor-promoting effects. It may also be possible to target such factors, in combination with HER2 or Ras inhibitors, to achieve greater clinical efficacy.

A major effector downstream of Ras is the phosphatidylinositol 3-kinase (PI3K) pathway. The catalytic subunit of PI3K, PIK3CA, is frequently found mutated in cancers, especially those cancers originating in the breast (8, 9). Like mutant Ras, expression of mutant PIK3CA can cause nontumorigenic cells to undergo transformation and gain invasive abilities (10, 11). Carcinoma cells, in particular, may gain increased invasive abilities by undergoing an epithelial-to-mesenchymal transition (EMT), where adherens junctions formed by E-cadherin are disrupted (12, 13). Canonical transcription factors that repress E-cadherin include Twist, Snail, Slug, and Zeb1, although this network has grown more complex in recent years (14, 15).

The transcription factor p63 not only induces transcription of canonical p53 targets but is also a master regulator of epithelial cells (16, 17). Mice losing both alleles of p63 display complications due to the loss of epithelial stratification, including the absence of mammary glands (18, 19). They also have significant craniofacial and limb abnormalities, indicating that p63 plays a key role in embryonic development. There are many isoforms of p63, including the major types TAp63 and ΔNp63, which are transcribed from alternative start sites (20). There is also alternative splicing at the 3′ end, resulting in the isoforms α, β, and γ (20). ΔNp63 was determined to be responsible for the aforementioned characteristics in mice, because isoform-specific knockdown led to similar epidermal defects (21, 22).

Interestingly, the role of p63 in cancer is controversial, possibly due to the different activities of its various isoforms and/or tissue specificity. In head and neck squamous cell carcinomas, the p63 locus is often amplified, suggesting an oncogenic role (23). However, in other types of tumors, basal epithelial markers like p63 and keratin 14 are lost (24 ⇓–26), and ∆Np63 has been found to suppress EMT in prostate and bladder cancer cells (27, 28). Surprisingly, in one report, ∆Np63α and ∆Np63β were found to inhibit, whereas ∆Np63γ promoted, EMT in MCF10A mammary epithelial cells (29). These conflicting results make it important to determine the effects of p63 on cell growth, differentiation, and invasiveness in different cell types.

We endeavored to study gene and network changes downstream of Ras in mammary epithelial cells. Our analysis of these changes indicates that H-RAS and PIK3CA oncogenes can induce EMT via repression of p63.

Results

We were originally interested in how the H-Ras and p53 pathways might interact to regulate gene expression. For this purpose, we used a pair of isogenic MCF10A cell lines, one with wild-type (WT) p53 and another with a homozygous deletion of p53’s second exon leading to the loss of functional p53 protein (termed “p53-del” here; clone 1A from ref. 30). MCF10A is a nontransformed mammary epithelial cell line that was spontaneously immortalized after derivation ex vivo from a healthy woman who underwent reduction mammoplasty (31). This line is thought to derive from myoepithelial cells because they express p63, keratin 5, and keratin 14 (32). Activated H-RasV12 or the empty vector (hereafter, Vector) were introduced into both p53 WT and p53-del MCF10A cell lines by retroviral transduction. Expression of H-Ras was confirmed by immunoblots and quantitative RT-PCR (qPCR) (Fig. 1A and Fig. S1A). The morphology of the H-Ras cells was altered, as previously seen for the comparable cell line MCF10AT (33) (Fig. S1B). Although the Vector lines remained cobblestone-like when confluent, characteristic of epithelial cells, the two lines expressing mutant H-Ras displayed an elongated shape with loss of cell-to-cell attachments (Fig. S1B). We also found that the H-Ras–infected cell lines, but not the Vector-infected cell lines, formed large colonies in soft agar, demonstrating anchorage-independent growth (Fig. S1B), indicating their transformed properties.

Fig. 1.

Differential expression in H-Ras–transformed MCF10A cells. (A) Western blot of MCF10A cells with WT-p53 or p53-del background, with H-RasV12 (R) or Vector (V). (B) Venn diagram is shown for differentially expressed genes (DEG) identified by RNA-Seq in H-RasV12 compared with Vector cells in either background. Top GO categories are shown for genes down-regulated in H-Ras cells in both WT and p53-del settings. (C) Top 10 down-regulated genes from the overlap seen in B (values from WT-p53 set); P values shown are always adjusted for multiple testing using the Benjamini–Hochberg procedure. (D) Validation of ΔNp63 mRNA levels by qPCR. Error bars are the SD from three experiments. (E) Levels of p63 pre-mRNA assayed by qPCR using intron-specific primers.

Fig. S1.

Characterization of MCF10A lines with and without H-RasV12. (A) H-Ras mRNA levels were measured in transduced Vector (V) or H-RasV12 (R) MCF10A cells by qPCR. (B, Top) DIC images of V and R cells. (Scale bar: 100 μM.) (B, Bottom) Colony growth of WT-p53 cells in soft agar after 11 d, representative of three independent experiments. (Scale bar: 200 μM.) (C) Top 10 H-RasV12 up-regulated genes from the overlapping gene set of p53-del and WT-p53 cells, sorted by adjusted P value. (D) Top GO biological pathways associated with genes up-regulated after H-RasV12 expression. (E) Levels of E-cadherin (E-cadh) and N-cadherin (N-cadh) mRNA measured by qPCR.

To assess gene expression changes and identify novel targets downstream of H-Ras and p53, we performed RNA-sequencing (RNA-Seq) for the four cell lines. A Venn diagram of genes significantly regulated by H-RasV12 (fold change > 2 and P < 0.05) is shown for the p53-del and WT cell lines (Fig. 1B). Interestingly, the total number of significantly changed targets was smaller in the WT-p53 set of cell lines than in the p53-del set of cell lines, suggesting that the presence of p53 has a dampening effect on H-Ras–activated signaling pathways.

There was an overlap of 821 genes up- or down-regulated by H-Ras in both the p53-del and WT cell lines, which are listed in Dataset S1A. Genes uniquely regulated by H-Ras in the p53-del and WT cell lines are listed in Dataset S1 B and C. In addition to H-Ras, which was elevated due to its introduction into these cell lines, a number of mesenchymal genes, including Zeb1, Zeb2, Vimentin (VIM), N-cadherin (CDH2), and Fibronectin (FN1), were within the top 100 significantly up-regulated targets (Fig. S1C and Dataset S1A). This finding suggested that the cells had undergone EMT. Correspondingly, genes associated with the epithelial state were significantly down-regulated, including E-cadherin (CDH1), miR-205 host gene (MIR205HG), keratin 5 (KRT5), grainyhead-like transcription factor 2 (GRHL2), and epithelial splicing regulatory protein 1 (ESRP1) (Fig. 1C and Dataset S1A). Interestingly, TP63, a transcription factor associated with the epithelial state, was the second and sixth most significantly down-regulated target in WT-p53 and p53-del cells, respectively. We performed a gene ontology (GO) cluster analysis and found that mutant H-Ras–down-regulated genes were strongly associated with epidermis and ectoderm development as well as epithelial cell differentiation, further suggesting that H-Ras causes changes in epithelial differentiation (Fig. 1B). There was a weaker, but significant, association of Ras–up-regulated genes with “response to wounding” and vasculature development (Fig. S1D). Due to this regulation of epithelial-related genes in both the p53-del and WT-p53 lines, we chose to focus on this p53-independent aspect of H-Ras signaling. (For future reference, Dataset S1 B and C provides lists of genes regulated by H-Ras uniquely in either the WT-p53 or p53-del background.)

We validated by qPCR that the level of p63, specifically its ΔNp63α isoform, is strongly reduced in the H-Ras cells and found that the ΔNp63α isoform is the predominant p63 isoform expressed in MCF10A cells, as has been described previously (34) (Fig. 1D and Dataset S1D). Furthermore, using intronic primers to measure pre-mRNA in the nucleus, we showed that p63 pre-mRNA was similarly repressed by H-RasV12, indicating that ΔNp63α expression is regulated by Ras at the transcriptional level (Fig. 1E).

To investigate more precisely how ΔNp63α expression is repressed by H-RasV12, we constructed promoter reporter genes with various lengths of an upstream p63 genomic sequence, starting with a construct containing −3,043 to +139 of the ΔNp63α promoter (35). These reporters were transfected into the MCF10A cells (WT-p53) expressing Vector or H-Ras. We found that there was two- to threefold higher expression of the promoter constructs in Vector vs. H-Ras cells, after normalizing to the internal control of an SV40 promoter-Renilla luciferase plasmid (pRLSV40P) (Fig. 2A). Progressive deletion from the 5′ end resulted in reduced expression; however, each construct down to −100 showed similar differential expression between Vector and H-Ras cells. As controls, we found that the promoter activity of a p53 response element plasmid was not altered, whereas a minimal E-cadherin promoter was significantly lowered in Ras cells as expected, given the lower E-cadherin expression in these cells as discussed below (Fig. 2C).

Fig. 2.

Analysis of the ∆Np63 promoter. (A) Luciferase reporter assays were performed in the V and R cell lines (WT-p53 background) with ∆Np63 promoter constructs with the indicated sequence upstream of the TSS. (B, Left) Sequence of the −83 to −44 ∆Np63 promoter region with the indicated transcription factor binding sites and mutations. (B, Right) Reporter assays of −83 to −44 region on a Fos minimal promoter and indicated mutants. (C) Luciferase reporter assays of control E-cadherin (E-cadh) and p53 response element (p53 RE) reporters. These results are the means of at least three experiments with the SD. (D) Chromatin immunoprecipitation of the ∆Np63 and TK1 promoters. The indicated antibodies were used to detect transcription factor or RNA polymerase II phospho-S5 binding to the ∆Np63 TSS (Left) or control Thymidine Kinase 1 (TK1) TSS (Right) in Vector of H-RasV12 cells. Nonspecific IgG was used as a control antibody. The error bars indicate the SD of three experiments. The x axis is arbitrary units (A.U.) normalized to input DNA for each cell line.

To separate the effects of the upstream promoter from the effects of the sequence surrounding the transcriptional start site (TSS), we tested the −100 to −30 region (just upstream of a TATA box) using a reporter with a minimal c-fos promoter. This sequence was similarly preferentially expressed in Vector cells (Fig. S2A). We also tested a trimer of the −100 to −30 region and found even stronger Vector-specific activity (Fig. S2B). The −83 to −30 region was also sufficient for this regulation; however, deletion of either the 5′ or 3′ part of this sequence caused a loss of promoter activity, suggesting that multiple sequences within this region are required in concert for proper regulation of expression (Fig. S2A).

Fig. S2.

Luciferase reporter assays. (A) Indicated regions of the ΔNp63 promoter were cloned upstream of the c-fos minimal promoter and assayed for luciferase activity in V or R MCF10A cells. (B) The 3× (−100 to −30) promoter sequence was cloned upstream of the c-fos minimal promoter and assayed for luciferase activity as in A (A and B are the average of three experiments). (C) Indicated mutants (described in Fig. 2B) were constructed in the −500 to +139 background, replicating the results in Fig. 2B (average of two experiments).

We were able to narrow down the region required for differential expression further to −83 to −44 (Fig. 2B). Examination of the sequence of this region showed a CCAAT box and potential SP1 site. When we made specific mutations along the −83 to −44 sequence (Fig. 2B, Left), altering CCAAT Box and Sp1 sites, promoter activity was completely abolished, whereas mutations in the other sites had only partial effects on expression (Fig. 2B, Right). We also tested the CCAAT Box and Sp1 mutations in the context of a longer promoter sequence, −500 to +139, and these mutations also displayed strongly decreased promoter activity (Fig. S2C). Collectively, these results suggest the importance of the entire −83 to −44 region for regulation, and point to the CCAAT box and SP1 sites as being particularly critical.

Given the strong effects of the CCAAT box and SP1 site mutations, we tested whether the main factors associated with these sites bind to them in MCF10A cells using chromatin immunoprecipitation followed by qPCR. For the CCAAT box, a trimer of NF-YA/B/C is the primary binding factor (36, 37), whereas SP1 binds to GC-rich boxes (38). These factors were previously found to bind to the ΔNp63 promoter in mouse keratinocytes (39). Although there were no differences in the expression levels of NF-YA or SP1 from immunoblotting in our MCF10A cell lines, we were able to detect their binding by chromatin immunoprecipitation to a sequence near the TSS of a known target, Thymidine Kinase 1 (Fig. 2D, Right). However, no binding was detected at the TSS of the ΔNp63α promoter above the binding of a nonspecific IgG control (Fig. 2D, Left). These results argue against an involvement of these factors in ΔNp63α regulation. It is possible that factors other than Sp1 and NF-YA bind to the above-mentioned sequences in the ΔNp63 promoter in contrast to the case in mouse keratinocytes. Importantly, when we measured binding of activated RNA polymerase II at the ΔNp63α TSS using antibodies to phosphoserine 5 of the C-terminal domain, there was a strong signal in Vector cells, but greatly decreased binding in H-Ras cells that was close to the background of the IgG control (Fig. 2D). These results further confirm that ΔNp63α is regulated by H-RasV12 at the transcriptional level, affecting RNA polymerase II activity and leading to repression of ΔNp63α expression.

Given the changes in EMT-related genes seen by RNA-Seq, we tested by qPCR and Western blotting whether these genes were, in fact, regulated at both the protein and mRNA levels. The hallmark epithelial marker E-cadherin was suppressed, whereas the mesenchymal marker N-cadherin was activated by H-Ras at both levels (Fig. 3A and Fig. S1E). Similar to the mRNA levels, ∆Np63α protein was suppressed, whereas mesenchymal proteins Fibronectin and Vimentin were increased in H-Ras cells (Figs. 1D and 3A), demonstrating that these cells have undergone EMT.

Fig. 3.

Regulation of EMT factors in H-Ras cells. (A) Protein levels of p63 and EMT markers. Western blots were performed with the indicated antibodies. The results are representative of at least three experiments. (B) Zeb1 and Zeb2 mRNA levels were measured by qPCR. (C) Western blots of Vector and Ras cells were performed with the indicated antibodies. (D) Expression of the indicated miRNA levels were measured by qPCR. The results are normalized to 1.0 in the p53-del Vector cell line for each miRNA.

Although many reports have suggested that the transcription factors Twist, Snail, and Slug are regulators of the cadherin switch, it is likely they are not EMT inducers in this model due to the lack of RNA regulation (although we cannot rule out posttranscriptional regulation). Other known regulators driving EMT and E-cadherin repression are the ZEB1 and ZEB2 transcription factors (15, 40, 41). In contrast to the first three inducers, these factors were strongly up-regulated in the H-Ras cells, suggesting that they are the likely regulators of EMT in these cells (Fig. 3 B and C).

A well-known mechanism of suppression of the ZEB factors is by miR-205 and members of the miR-200 family (including miR-200a, miR-200b, miR-200c, miR-141, and miR-429; henceforth, miR-200f) (42, 43). These microRNAs (miRNAs) have also been reported to be regulated by p63 (27, 28, 44), suggesting a pathway for ∆Np63α regulation of ZEB1/2. We validated that expression of miR-200b, miR-200c, and miR-205 was suppressed by H-Ras in the four cell lines, whereas there was no change in expression of a control miRNA, miR-375 (Fig. 3D). These results are consistent with p63 regulation of ZEB1/2 through these miRNAs, although additional mechanisms are possible.

To test whether the effect of H-RasV12 was specific to the cell line used, we expressed H-RasV12 in an independent mammary epithelial cell line, MCF12A. This cell line is an immortalized, nontumorigenic breast epithelial cell line from a different patient than MCF10A, who also underwent reduction mammoplasty. The MCF12A H-RasV12 line similarly underwent EMT, displaying the same loss of E-cadherin and ∆Np63α, along with induction of N-cadherin, and Fibronectin (Fig. S3 A and B). Therefore, the downstream effects of H-Ras activation may be generalizable to multiple mammary epithelial cell lines.

Fig. S3.

Induction of EMT after H-RasV12 was transduced into MCF12A cells. (A) Western blot with the indicated antibodies. Vec, Vector. (B) ΔNp63 and E-cadherin mRNA levels were determined by qPCR. Error bars represent the SD of three experiments. (C) Total Ras levels in MCF10A set of cell lines, compared with duplicate plates of Hs-578T and HT29 cells (representative of two experiments). A pan-Ras antibody was used to detect total H-, K- and N-Ras levels. Ras quant., quantification of Ras normalized to actin.

Although mutant H-Ras is significantly overexpressed in the MCF10A cells, it is also possible that endogenous H-Ras is expressed at low levels in MCF10A cells. We therefore compared the levels of total Ras proteins using a pan-Ras antibody. In this case, the overexpression of Ras in our MCF10A cells is only three- to fourfold (Fig. S3C). We also found that the levels of mutant Ras expressed in our engineered cell lines were only 2.5- to sevenfold higher than the level in a breast adenocarcinoma line Hs578T containing mutant H-Ras. The level of mutant Ras was also only onefold to 4.4-fold higher than the level in a colorectal carcinoma line HT-29 containing WT Ras (Fig. S3C). These results suggest that the expression of mutant H-Ras in the MCF10A lines is within a range of physiological expression and that the observed effects cannot be attributed to grossly overexpressed Ras oncoprotein.

To understand more about the kinetics of the effects of Ras, we engineered a stable cell line with doxycycline-inducible H-RasV12, termed “TR-Hras.” We found that it took about 5 d of H-Ras induction by doxycycline to cause both ∆Np63 and E-cadherin repression and to induce the mesenchymal markers N-cadherin, Fibronectin, and Zeb1 (Fig. S4 A and B). These results further confirm the regulation of p63 and EMT by H-Ras. Also, these effects of H-RasV12 activity are not immediate; rather, they necessitate several days, as is common for induction of an EMT (45).

Fig. S4.

EMT is induced in a “Tet-on” cell line, TR-Hras. (A, Left) TR-Hras cells with and without doxycycline treatment for the indicated days were lysed and immunoblotted with the indicated antibodies. (A, Right) qPCR analysis for the indicated genes after 4 d of doxycycline treatment. (B) Indicated mesenchymal gene expression was measured by qPCR after 4 d of doxycycline (Dox) treatment.

Because activation of PI3K and loss of the phosphoinositide phosphatase phosphatase and tensin homolog (PTEN), which leads to hyperactive PI3K, are more common in breast cancers than Ras mutations, we expressed mutant H1047R of PIK3CA in MCF10A cells by retroviral transduction, establishing two clonal cell lines, P1 and P2 (Fig. 4A). Hotspot mutations in the PI3K catalytic subunit PIK3CA, like H1047R, are found in nearly a quarter of breast cancers, and H1047R, in particular, was shown to lead, in turn, to strong activation of the Akt kinase, driving mammary tumorigenesis (8, 10, 46). Clones P1 and P2 showed induced phosphorylation of Akt1, demonstrating its activation downstream of PI3K in both cell clones (Fig. 4A). Note that the levels of exogenous PIK3CA H1047R expression are similar to the levels of the endogenous PIK3CA protein. The cells displayed fibroblast-like morphology, along with repression of ∆Np63α and E-cadherin, similar to the H-Ras cell lines (Fig. 4 A and B). Mesenchymal genes were again up-regulated in these clones (Fig. 4B). These results show that PIK3CA, a common breast cancer oncogene, can induce EMT and ΔNp63α repression and suggest that the effects of H-Ras may be due to its activation of PI3K/Akt signaling.

Fig. 4.

Activation of the PI3K pathway causes EMT. (A) Phase contrast (Left) and Western blots with the indicated antibodies (Right) of MCF10A cells infected with V or PI3KCA mutant H1047R (P1 and P2). Scale bars (Left): 100 μm. The endogenous and 3× Flag-tagged p110α can be seen separately in the P1 and P2 cells. (B) Western blots and qPCR of p63 and EMT markers. For qPCR, the results are normalized to 1.0 in V cells for E-cadherin and ∆Np63 and in P1 cells for N-cadherin, Zeb1, and Zeb2; error bars show the SD of three experiments.

To determine whether repression of ΔNp63α is sufficient to induce EMT, we made stable cell lines with shRNA to inhibit ΔNp63α expression. We established two clonal cell lines with independent shRNAs targeting the unique N terminus of ΔNp63. These clones displayed near-total inhibition of ∆Np63 mRNA and protein levels (Fig. 5A). Concordantly, they exhibited characteristics of EMT, including loss of E-cadherin and gain of N-cadherin, Vimentin, Fibronectin, and Zeb1 expression (Fig. 5A and Fig. S5A). The epithelial-associated miRNAs showed strongly reduced expression, indicating that ΔNp63 is a key regulator (Fig. 5B).

Fig. 5.

Depletion of ∆Np63 induces an EMT phenotype. (A) MCF 10A cells stably expressing shRNA (#1, #2) to ∆Np63 or a control shRNA (V) were immunoblotted with the indicated antibodies (Left) or analyzed for mRNA expression by qPCR (Right). (B) Levels of the indicated miRNAs in the control and sh-∆Np63 cell lines were determined by qPCR. (C) Venn diagram is shown for genes identified by RNA-Seq that are differentially expressed in H-RasV12 and ∆Np63 shRNA cell lines compared with control cells. Top GO processes are shown for genes down-regulated in both H-Ras and ∆Np63 shRNA cells.

Fig. S5.

Characterization of the ΔNp63 shRNA cell lines. (A) Indicated mesenchymal genes were measured by qPCR. (B) Expression of genes in the ΔNp63 shRNA and control cell lines was determined by RNA-Seq. Selected GO categories that are associated with genes significantly down-regulated in the ΔNp63 shRNA cell lines are shown.

We also performed RNA-Seq on the control (LMP vector) and sh-ΔNp63 lines, and identified 2,255 genes that were significantly altered by both ΔNp63 shRNAs (Dataset S2A). GO analysis found that the same epithelial gene signatures as in H-Ras cells were prominently altered, along with additional signatures of cell migration and adhesion (Fig. S5B). Upon comparison of the sh-ΔNp63–regulated genes with the mutant Ras-regulated genes, 380 overlapping targets were identified (Fig. 5C and Dataset S2B), indicating that p63 regulates an appreciable subset of the genes that H-Ras controls. We performed GO analysis on the genes down-regulated by both perturbations and again found many processes related to changes in epithelial differentiation and development (Fig. 5C). Epithelial-specific genes, such as MIR205HG, KRT5, CDH1, ESRP1, and GRHL2, were significantly down-regulated following both alterations. The loss of these genes further demonstrated that the induction of EMT in these cells is due to circuits controlled by Ras and p63.

When we analyzed several transcription factors among genes uniquely regulated by H-Ras (Dataset S2C) in preliminary experiments, we did not identify any one factor as being sufficient to control p63 or induce EMT. It is possible that several factors coordinately regulate this pathway, or it is feasible that other targets are regulated posttranscriptionally to control EMT.

Collagen I is a major component of extracellular matrix (ECM) in human breast tissue and is known to support growth of MCF10A mammospheres (47). Because H-Ras–controlled EMT has been associated with metastasis (48, 49), we examined 3D cell invasiveness in 3D collagen I gels. For these assays, we embedded spheroids consisting of 5,000 cells into collagen I matrices and then monitored any outward cellular invasion into the fibrillar matrix 24 h later. Under these conditions, cell proliferation is minimal. We found that Vector cells only spread from the spheroid in a sheet-like manner, exhibiting a closed front characteristic of epithelial growth (Fig. 6A). In contrast, H-RasV12 and sh-ΔNp63α cells demonstrated extensive invasion of individual cells into the ECM (Fig. 6A). We quantified the change in invasive behavior by counting the number of single cells that left the spheroid. Although few cells invaded beyond the spheroid core formed by the Vector cells, a significant number left the core of H-RasV12 and sh-ΔNp63α spheroids (Fig. 6B). This invasive behavior was consistent with a shift to a more mesenchymal phenotype.

Fig. 6.

Invasiveness of H-RasV12 and p63-depleted cells. (A) Multicellular spheroid invasion was measured in 3D collagen I gels. After cells invaded for 24 h, spheroids were fixed and stained with fluorescently labeled phalloidin. (Scale bars: 200 μM.) (B) Individual cell invasion was quantified by counting single cells beyond the spheroid core (at least 20 spheroids were measured from two independent experiments; ***P < 0.0001 from a one-tailed t test). Vec, Vector.

Adhesion and force generation on ECM via integrin receptors are prerequisites for mesenchymal locomotion in 3D environments, such as the collagen I matrices used here (50). Radial alignment of collagen fibers at the spheroid surface is a robust indicator of integrin-dependent interaction with the ECM (51, 52). To exclude the possibility that invasion differences were caused by variations in the cells’ abilities to establish contacts with the collagen matrix, we imaged them soon after embedment (t = 2 h) with confocal reflectance microscopy, which allows for the visualization of collagen fibers. In each case, we observed collagen fibers radially aligned to the spheroid surface, indicating comparable cell-matrix attachments (Fig. S6). Thus, the observed differences in invasive behavior cannot be attributed to insufficient surface presentation among the MCF10A-derived lines.

Fig. S6.

Collagen attachment of spheroids. Confocal reflectance microscopy (CRM) was used to assess cell–ECM contact and visualize collagen fibers in MCF10A-derived cell lines 2 h after the spheroids were embedded in the matrix. Collagen fibers radially aligned to the spheroid surface (arrows) indicate integrin-mediated cell-matrix adhesion. (Scale bars: 50 μm.) The bright spot in the images is an optical artifact associated with CRM.

To determine whether p63 levels in tumors might be similarly regulated as in Ras-transformed MCF10A cells, we looked at expression of the p53 family in paired normal/tumor samples from The Cancer Genome Atlas Breast Cancer dataset (9). In fact, p63 expression was significantly lower in tumor samples relative to their normal tissue counterparts, whereas there was little change in p53 and p73 expression (Fig. 7A). This finding suggests that although p63 is not frequently mutated in cancers, its expression is commonly lowered by tumorigenic mechanisms.

Fig. 7.

p63 expression in human breast cancers. (A) p53 family expression in breast tumors. TP53, TP63, and TP73 gene expression in paired normal and tumor samples (n = 110) from the TCGA Breast Cancer dataset is shown as normalized counts using the RSEM software package. The P value for the difference in TP63 expression between normal and tumor cells is shown. (B) Breast tumors (n = 795) were stratified by molecular subtype using the PAM50 gene set and compared with normal samples, from above (n = 110). HER2-Enr, HER2-enriched tumors. The normalized expression for ΔNp63α is shown (*** P < 0.0001 using Welch’s t test for basal compared with normal). (C) Model of signaling from activated Ras leading to ∆Np63α repression and EMT.

It was also possible that because p63 is expressed in basal but not luminal epithelial cells (20), the low p63 expression was simply due to a predominance of nonbasal tumors. However, after stratifying a larger breast cancer dataset of tumors by molecular subtype according to the PAM50 (Predication analysis of microarrays 50) gene set (53), p63 expression in basal-like breast cancers remained close to zero and was therefore significantly lower than that in normal tissue (Fig. 7B). In fact, expression of p63 in basal tumors was comparable to expression of p63 in luminal-type tumors, which would not be expected to express p63. Finally, we examined a tissue microarray (TMA) of triple-negative breast cancer (TNBC) tissues for p63 protein expression. Various studies have shown that the majority of basal-like breast cancers fall within the TNBC subtype, and that these cancers express basal cytokeratins (54, 55). However, in 94% of these patient samples, p63 was expressed in ≤5% of tumor nuclei (Dataset S3). Together, these results show that p63 expression is greatly reduced in a preponderance of breast cancers and support our model that oncogene down-regulation of this gene is important for tumor progression. We cannot rule out, however, that basal tumors originate in luminal progenitors as has been reported for basal tumors from BRCA1 carriers (56, 57).

Discussion

We have found that ectopic expression of mutant H-Ras and PIK3CA in MCF10A cells leads to down-regulation of ∆Np63 and transdifferentiation of these cells to a more mesenchymal phenotype. This process was accompanied by a distinct change in invasive behavior, suggesting that the transdifferentiation process, as orchestrated by the EMT, leads to greater metastatic potential. Depletion of ∆Np63 with shRNA also caused EMT, demonstrating that this depletion is also sufficient to induce the transition. Along with other evidence that ∆Np63 is required for epithelial stratification (16, 18), these results confirm that ∆Np63 is a master regulator of the epithelial state. ∆Np63 activates expression of epithelial-specific miRNAs that are repressors of the mesenchymal-inducing transcription factors Zeb1 and Zeb2. These factors directly repress E-cadherin expression, suggesting a pathway for induction of an EMT from Ras to ∆Np63 to E-cadherin (Fig. 7C).

There are many afferent signaling pathways that can induce an EMT (39). TGF-β was one of the first characterized inducers of this transition, and signaling from Ras and TGF-β can synergistically drive cells toward a more mesenchymal state (58, 59). We have not observed activation of the TGF-β pathway in our MCF10A-Ras cells, suggesting that they are driven to EMT independent of TGF-β. Because we found that mutant PIK3CA phenocopied the H-Ras cells, we postulate that a factor(s) within the PI3K/Akt signaling pathway is responsible for direct repression of the ΔNp63 promoter and induction of EMT, independent of the TGF-β pathway.

We were concerned that the MCF10A cell lines we generated markedly overexpress mutant H-Ras compared with the endogenous H-Ras in the control cell lines and that this overexpression might cause nonphysiological effects. However, the levels of total Ras proteins (H-, K- and N-Ras) were more modestly overexpressed, and there was one- to sevenfold higher expression than endogenous Ras compared with two human cancer cell lines that were tested (Fig. S3C). Although we cannot rule out spurious effects of high levels of ectopic mutant H-Ras in MCF10A cells, the PIK3CA cell lines we generated expressed p110α protein at levels similar to the endogenous protein. Because PIK3CA mutations are common in breast cancer, our results suggest that physiological levels of mutant PIK3CA can have effects on EMT during breast cancer progression and that the similar effects of H-Ras overexpression in our cell lines on EMT are physiological. Further work will be required in mammary tumor models to test the importance of oncogene repression of p63 and activation of EMT.

We mapped a short region of the ∆Np63 promoter required for differential expression in Ras-transformed cells. Although the entire region of −83 to −44 is required, the CCAAT box and SP1 sites were particularly critical. A similar requirement for these sites for ∆Np63 expression was also found in mouse keratinocytes (60). However, although Romano et al. (60) detected binding of NF-Y and Sp1 to the ∆Np63 promoter in these cells, we did not detect binding by chromatin immunoprecipitation in MCF10A mammary epithelial cells, suggesting that other factors are required. A number of transcription factors have been reported to regulate ∆Np63 expression, including negative regulation by both WT and mutant p53 (61). For instance, mutant p53 can form a complex with p63 and inactivate it (62, 63). Further, depending on the context, ∆Np63 levels can be either up- or down-regulated via intersection with Notch, Wnt, and other signaling pathways. More work will be required to determine which transcription factors directly regulate the ∆Np63 promoter in mammary cells, and how Ras and PI3K signaling modulate their activity.

Among all of the canonical EMT regulators, Zeb1 and Zeb2 expression was induced in both H-Ras and sh-∆Np63 cells. One mechanism for ∆Np63 regulation of Zeb1/2 is through miR-200f and miR-205 (27, 28, 42, 43) (Fig. 7C). We found that these miRNAs are strongly repressed by H-Ras and ∆Np63 depletion in MCF10A cells, consistent with this pathway leading to regulation of E-cadherin and EMT.

Consistent with our data linking repression of ∆Np63 with the occurrence of EMT, analysis of the The Cancer Genome Atlas (TCGA) breast cancer dataset shows that breast tumors contain low ∆Np63 expression compared with normal tissue. In particular, our analyses revealed that basal-like cancers and TNBCs have low ∆Np63 expression. In addition, it should be noted that there have been many other reports of low p63 expression in breast cancers (32, 64, 65). There are, however, cases where high p63 expression is seen in breast tumors (56, 65, 67). We propose that these cases represent a distinct type of breast carcinoma similar to esophageal squamous carcinomas, which have high p63 (57, 68). These tumors appear to have a more stratified epithelial nature and may use a separate pathway to tumorigenesis.

The cell of origin of basal-like breast cancers from BRCA1 mutation carriers may actually be a luminal progenitor cell; such cells have similar gene expression patterns to patient basal-like tumors (69,70). It is currently unknown whether all basal-like breast cancers derive from luminal progenitors and not just those breast cancers with BRCA1 alterations. Indeed, there may be several differentiation pathways for cells of origin for breast cancers. Multipotent and unipotent mammary cell progenitors have been characterized and would provide alternative stem cell-like cells for breast cancer initiation (reviewed in ref. 71). Some basal myoepithelial cells also have mammary stem cell properties in that they are able to repopulate a mammary gland in vivo (72), raising the possibility that basal myoepithelial cells may also be able to function as cancer stem cells and as a cell of origin for basal-like tumors. It is possible that oncogenes, such as H-RAS and PIK3CA, could participate in transformation of these basal-like mammary stem cells, and subsequently induce their gain of mesenchymal properties. The further determination of cells of origin for different types of breast cancer is key to understanding breast cancer development and subsequent differentiation of breast cancer cells by EMT or to other fates.

There is considerable evidence that EMT contributes to cancer progression and metastasis, as well as inducing stem cell properties (73). EMT is also known to induce individual cell migration, which is a property we observed after Ras pathway activation and ∆Np63 depletion (13, 73, 74). However, two recent papers have concluded rather provocatively that EMT is not required for metastasis, although it increases chemoresistance (75, 76). It is unclear if these conclusions hold true for human breast cancer, so further manipulation of EMT will be required to determine its role in different cancer types.

In conclusion, the strong repression of ∆Np63 by H-RAS and PIK3CA and induction of EMT suggest that this process is critical for mammary tumorigenesis. Future studies must determine how important this pathway is to human breast cancer progression and how p63 regulation influences pathogenesis. A more detailed understanding of this pathway may lead to an alternate approach to block this extremely prevalent disease.

Materials and Methods

A comprehensive methods section is available in Supporting Information. A brief summary follows.

Retroviral vectors were used to express H-RasV12, PI3KCA-H1047R, and FLAG-∆Np63α in MCF10A cells with or without p53 deletion (30). Tet-on H-RasV12 cells were made in the MCF10A-TR cell line (77) using a lentiviral H-RasV12 vector. The shRNA cell lines were made with short hairpin sequences to ∆Np63 (unique to the N-terminal coding region).

Protein extraction, Western blotting, chromatin immunoprecipitation (ChIP), and qPCR, to measure mRNA and miRNA levels, were performed by standard methods. Luciferase assays were performed using the Dual-Luciferase Reporter Assay (Promega) with the indicated firefly luciferase reporters and an internal control of the Renilla luciferase vector pRLSV40P. Sequences of oligonucleotides used for qPCR, ChIP, and shRNAs are shown in Dataset S4.

For the invasion assay, spheroids were formed using a previously described centrifugation method (78) that results in multicellular spheroids surrounded by a shell of basement membrane. Single spheroids were collected, and each was added to a collagen solution followed by collagen gelation. For assessment of cell invasion, live cell imaging was performed. At least 20 spheroids were imaged for each condition from two independent experiments. Only cells completely distinct from the spheroid core at t = 24 h were counted. Collagen was imaged via confocal reflectance microscopy with the 488-nm laser at t = 2 h after spheroid embedding.

Bioinformatic analysis of TCGA datasets are described in Supporting Information. For RNA-Seq, mRNA was isolated from total RNA using oligo-dT beads (Thermo Fisher). RNA-Seq was carried out by C&K Genomics (Vector and H-Ras cells) and Macrogen (LMP and sh-∆Np63 cells). The H-Ras RNA-Seq data were deposited in the Gene Expression Omnibus (www.ncbi.nlm.nih.gov/geo/) under accession number GSE81593. Data from the sh-∆Np63 RNA-Seq are hosted at a Columbia University website (biology.columbia.edu/privesngs). Differentially expressed genes were filtered using the criteria of fold change > 2.0 and P < 0.05 (adjusted using the Benjamini–Hochberg procedure) (79). Only genes above a threshold for expression in the positive cell lines were included.

Three TNBC TMAs were constructed under Columbia University Medical Center Institutional Review Board protocol AAAB2667 (Research with Human Surgical Specimens). Invasive breast cases negative for estrogen and progesterone receptor expression and Her2-negative for overexpression were collected with informed consent. Three cores of invasive breast carcinoma, and, in many cases, three cores of adjacent normal (nonneoplastic) breast tissue were represented per case in the array. The ΔNp63 antibody from Santa Cruz Biotechnology (sc-71825) was used at a ratio of 1:100.

Plasmids, Cell Culture, and Generation of Stable Cell Lines

The plasmid pBabe puro HA PIK3CA H1047R was a gift from Jean Zhou, Harvard Medical School, Boston (plasmid 12524; Addgene) (10). The MSCV (Murine Stem Cell Virus) LTRmiR30-PIG (LMP) vector, pBabepuro, pBabepuro-H-RasV12 (plasmid 18744; Addgene), and Phoenix-Ampho cell line were generous gifts from Scott Lowe, Memorial Sloan Kettering Cancer Center, New York. The −3,000 ∆Np63 promoter construct was a gift from Antonio Costanzo, Humanitas University, Milan, Italy, and contains −3,043 to +139 of the human ∆Np63 gene in the pGL3-basic luciferase plasmid (35). The −1,500, −500, −165, and −100 constructs were made by PCR from these bases to +139 and cloning the fragments between the KpnI and HindIII sites of pGL3-basic. The Cdh1 (E-cadherin) reporter contains −178 to +91 of the E-cadherin promoter and was a gift from Kumiko Ui-Tei, University of Tokyo (80). The c-fos minimal promoter, pOF-GL3, was previously described (81). The fragments of the ∆Np63 promoter were made by PCR and cloned upstream of the c-fos minimal promoter into pOF-GL3 containing −54 to +49 of the human c-fos promoter using KpnI and HindIII sites. The 3× (−100 to −30) construct was made by synthesizing the 3× fragment (Genewiz) before cloning into pOF-FL3. The −83 to −44 fragment and mutations were made by synthesizing oligonucleotides (Integrated DNA Technologies) that were annealed before cloning into pOF-GL3. The p53 response element (p53 RE) promoter plasmid contains the p53 RE from the p21 promoter (−2.3 kB) cloned in front of the c-fos minimal promoter of pOF-GL3.

MCF10A and MCF10A clone 1A (p53^−/−) were a gift from David Weber, University of Maryland School of Medicine, Baltimore (30). We received the MCF12A cells from Jose Silva, Mount Sinai School of Medicine, New York. These breast cell lines and their derivatives were grown in Dulbecco's modified Eagle's medium (DME)/F-12 media (11765; Thermo Fisher) with 5% (vol/vol) horse serum (16050122; Thermo Fisher), 0.5 μg/mL hydrocortisone (H0888; Sigma–Aldrich), 10 μg/mL insulin (91077C; Sigma–Aldrich), and 20 ng/mL EGF (AF-100-15; Peprotech) in a 37 °C incubator with 5% CO₂. Hs-578T cells were a generous gift from Timothy Bestor, Columbia University, New York. Both Hs-578T and HT29 are grown in DME with 10% (vol/vol) FBS (900108H; Gemini). For stable expression lines, the Phoenix-Ampho packaging cell line was grown in DME with 10% (vol/vol) FBS. To package virus, cells were transfected with 10 μg of plasmid with Lipofectamine 2000 (Thermo Fisher). After 3 d, retroviral supernatant was filtered (0.45 μm) and added to MCF10A or MCF12A cells with 5 μg/mL polybrene (107689; Sigma–Aldrich). After 2 d, cell lines were selected with 2 μg/mL puromycin (InvivoGen), and individual clones were picked (for PIK3CA) or surviving cells were pooled (for H-RasV12). After selection was completed (about 7 d), all cells were kept in 1 μg/mL puromycin for maintenance.

MCF10A-TR cells (containing the tetracycline repressor TetR) were from Rosalie Sears, Oregon Health and Sciences University, Portland, OR (77), and pLenti CMV/TO-RasV12-Puro vector was from Eric Campeau, University of Massachusetts Medical School, Worcester, MA (plasmid 22262; Addgene). For the TR-Hras cell line, the pLenti CMV/TO-RasV12-Puro vector was transfected into 293 cells with the lentiviral packaging plasmids (p∆8.9 and pVSVG), and the viral supernatant was then used to infect MCF10A-TR cells. The infected cells were selected with 5 μg/mL blasticidin (Thermo Fisher).

For the OEΔN line, the FLAG-∆Np63α sequence was taken from pCMV7.1-∆Np63 as previously described (82) and inserted in pBabebleo at the BamHI site to make the pBabebleo-Flag-∆Np63α vector. pBabebleo was a gift from Hartmut Land, University of Rochester, Rochester, NY and Jay Morgenstern, Warp Drive Bio, Cambridge, MA (plasmid 1766; Addgene) (83). The same transduction procedure was used as above for pBabepuro-H-RasV12, except that these cells were selected with 5 μg/mL phleomycin (Invivogen) and maintained in 2 μg/mL blasticidin.

For the shRNA cell lines, short hairpin sequences to ∆Np63 (unique to the N-terminal coding region) were designed as previously described (84) (Dataset S4). These sequences were subcloned into the LMP vector. As above, Phoenix packaging cells were transfected using Lipofectamine 2000 with the vector of interest. Retroviruses were harvested 48–60 h posttransfection, and the viral supernatant was filtered before it was added to plates of MCF10A cells in the presence of 8 μg/mL polybrene (Sigma–Aldrich). Cells were grown for 1 d before selection was begun using 1 μg/mL puromycin (Sigma–Aldrich). Antibiotic selection was complete 24 h after the last cell on the untransduced plate was eliminated, and cells were grown without any further antibiotic use.

Protein Extraction and Western Blotting

Cells grown on 10-cm plates to 80–90% confluency were washed twice with cold PBS and scraped into cold TEGN buffer [10 mM Tris (pH 7.9), 1 mM EDTA, 10% (vol/vol) glycerol, 40 mM NaCl, 0.5% Nonidet P-40, and 100 μM DTT with protease inhibitors (10 μM benzamidine, 70 nM leupeptin, 10 μg/mL α₂-macroglobulin, 0.7 μM bacitracin, and 0.5 μM PMSF)]. The lysates were kept on ice for 30 min with vortexing every 10 min. They were centrifuged at 14,000 × g for 10 min, after which the supernatant was isolated. Finally, the protein concentrations of the lysates were quantitated using Bradford assays (Bio-Rad). These whole-cell lysates (40 μg) were run on 7–15% (wt/vol) polyacrylamide gels and transferred to nitrocellulose filters, followed by blocking for 1 h at room temperature (RT) in 5% (wt/vol) nonfat dry milk in PBS with 0.05% Tween-20). Antibodies used include anti-ΔNp63 (polyclonal rabbit serum raised from peptide sequence MLYLENNAQTQFSEP; Covance), anti–N-cadherin (561553; BD Biosciences), anti-Flag (F1804; Sigma–Aldrich), anti-Fibronectin (F3648; Sigma–Aldrich), anti-Ras (3965; Cell Signaling), and anti–β-actin (A5316; Sigma–Aldrich). The following antibodies were from Santa Cruz Biotechnology: p63α (H-129), Zeb1 (H-102), Vimentin (C-20), E-cadherin (H-108), H-Ras (F235), pS473-Akt1 (11E6), Akt1 (5C10), p110α (H-201), NF-YA (H-209 X), and Sp1 (PEP2 X).

qRT-PCR

For qPCR, RNA was extracted from cell pellets using TRIzol reagent according to the manufacturer’s protocol (Thermo Fisher) and quantitated by UV spectrophotometry. cDNA was created from 1 μg of total RNA using the ImProm-II Reverse Transcription System (Promega), following the protocol for random hexamer priming. PCR was then performed using a dNTP mix (Roche) and Sybr Green Master Mix (Life Technologies) on an Applied Biosystems Step One Plus instrument. Conditions for linear amplification were established through template and cycle curves. The cycling conditions were as follows: a denaturation step at 95 °C for 10 min followed by 40 cycles at 95 °C for 15 s, 60 °C for 1 min, and 72 °C for 30 s, with a final extension at 72 °C for 7 min. The cDNA quantity was normalized to GAPDH or 18s rRNA levels. Primer sets are listed in Dataset S4. The following miScript Primer Assays were used to quantitate miRNA levels following the Qiagen protocol: Hs_miR-205_1, Hs_miR-200b_3, and Hs_miR-200c_1, all of which were normalized to RNU6B using Hs_RNU6-2_11.

Luciferase Reporter Assays

Cells were plated in 24-well plates in triplicate for each point, and transfected at 50% confluency with 400 ng of promoter plasmid and 200 ng of pRLSV40P following the protocol for Lipofectamine LTX with Plus reagent in 50% (vol/vol) Opti-MEM/media (Thermo Fisher). After 1 d, cells were washed in PBS and then gently swirled with 1× passive lysis buffer (Promega) for 30 min at RT. Firefly and Renilla luciferease activites were measured following the protocol for Dual-Luciferase Reporter Assays (Promega) using a Turner 20/20 Luminometer (TD-20/20). Background readings for Firefly and Renilla luciferase in the untransfected control cells were subtracted from experimental readings; then, ratios were quantitated and averaged for at least three experiments.

Invasion and Soft Agar Assays

For the invasion assay, cells were grown to 80–90% confluency, Accutase (MP Biomedicals) was used to detach cells from the plates, and spheroids were formed for 24 h at 37 °C and 5% CO₂ using a previously described centrifugation method (78) that results in multicellular spheroids surrounded by a shell of basement membrane. Spheroids were immersed in Cell Recovery Solution (Corning) to remove the basement membrane shell. Single spheroids were collected, and each was added to collagen solution prepared from pepsin-treated bovine collagen I (Advanced BioMatrix). This solution (1.0 mg/mL collagen) was prepared by diluting the stock solution with DMEM (Sigma–Aldrich), 2.5% (wt/vol) Hepes buffer (Invitrogen), 2.5% (wt/vol) sodium bicarbonate (Sigma–Aldrich), and double-distilled (dd) H₂O. Just before addition of the spheroid, the collagen solution was neutralized with NaOH to adjust the pH to 7.4. The samples were then transferred to an incubator at 37 °C and 5% CO₂. After 1 h, which allowed for complete collagen gelation, the gels were overlaid with 50 μL of growth medium. For immunocytochemical staining, cells were fixed in neutrally buffered 4% (vol/vol) formalin solution (Fisher Scientific) for 20 min at 24 °C and washed extensively with PBS. For staining of the actin cytoskeleton, cells were permeabilized with 0.5% Triton X-100 and washed with PBS; fluorescently labeled phalloidin (Invitrogen) was then added to the samples for 16 h at 4 °C. This solution was then removed, and spheroids were rinsed several times with PBS. For assessment of cell invasion, live cell imaging was performed using a custom-built microscope incubation chamber and objective heater to keep cells at 37 °C and 5% CO₂. A 10× air objective (N.A. = 0.4) was used during this imaging. At least 20 spheroids were imaged for each condition from two independent experiments. Only cells completely distinct from the spheroid core at t = 24 h were counted. For more detailed characterization of cell morphology and collagen structure, a 60× oil objective (N.A. = 1.42) was used. For all imaging, an inverted confocal laser-scanning microscope (Fluoview 300; Olympus) was used in transmitted light or confocal fluorescence/reflectance mode. For fluorescence imaging of cells, a 543-nm HeNe laser was used for excitation and detection was performed through a long-pass 570-nm filter. Black and white inverted images are shown for maximum clarity. Collagen was imaged via confocal reflectance microscopy with the 488-nm laser at t = 2 h after spheroid embedding.

For soft agar plates, 4% (wt/vol) agarose (Invitrogen) was diluted to 1.2% and mixed with 2× DME/F-12 to make a solution of 0.6% agarose DME/F-12 media. To each 35-mm plate, 1 mL of this solution was added and allowed to harden for at least 30 min at 4 °C. Cells (n = 500) were plated in a final concentration of 0.3% agarose in DME/F-12 media (1 mL). Complete (1 mL) medium was added, and this medium was changed every 3 d. Differential interference contrast (DIC) micrographs were taken 11 d after seeding on a Nikon Diaphot 300 microscope. Morphology pictures of passaging cells were taken in DIC or phase contrast mode.

Bioinformatics/RNA-Seq

TCGA datasets were downloaded directly from the TCGA data portal (cancergenome.nih.gov/). We used the Breast Invasive Carcinoma (BRCA) (9) TCGA Provisional dataset for analysis (February 2014). Data analysis was performed using MATLAB (MathWorks) (85). The RNA-Seq V2 (RNA-SeqV2) dataset was downloaded and analyzed to determine the expression levels of genes of interest. In the TCGA portal, the RNA-SeqV2 dataset includes the normalized gene expression of all genes and isoforms as estimated by an upper quartile normalization procedure using the RSEM software package (86). The median gene expression was calculated for each of the genes of interest following sample stratification based on metadata (normal tissue or tumor sample) and plotted using the box plot function. The statistical significance of the findings was determined by Welch’s t test (87), which is a modification in which the two samples may have unequal variances (heteroscedastic). The results were corrected for multiple testing by using the false discovery rate procedure of Benjamini and Hochberg to obtain the adjusted P values (88, 89). The box plots (90, 91) in the figures were plotted in MATLAB and are standard box plots with a notch to show the confidence intervals of the median of gene expression. In the plots, asterisks are used to denote statistical significance (***P < 0.0001). The accuracy of the analytical procedure was verified by corroborating several key genes with the results obtained from the cBioPortal website (92). Then, the samples were further stratified on the basis of the PAM50 classifier of 50 genes (53, 93) based on the datasets downloaded from the UCSC Cancer Genome Browser (94, 95). The calculation of the median of gene expression and statistical significance of the findings were determined as described for the previous analysis. Some of the computationally intensive analyses were carried out using the high-performance computing systems of the Extreme Science and Engineering Discovery Environment network (96).

For RNA-Seq, mRNA was isolated from total RNA using oligo-dT beads (Thermo Fisher). Construction of the RNA-Seq library was carried out based on Illumina HiSeq2000 protocols to generate 101 paired-end RNA-Seq reads. RNA-Seq was carried out by C&K Genomics (Vector and H-Ras cells) and Macrogen (LMP and sh-∆Np63 cells). The H-Ras RNA-Seq data were deposited in the Gene Expression Omnibus (www.ncbi.nlm.nih.gov/geo/) under accession number GSE81593. Data from the sh-∆Np63 RNA-Seq are hosted at a Columbia University website (biology.columbia.edu/privesngs). The raw data quality was checked using FastQC, and adapter sequences were trimmed using Trimmomatic before further analysis (97, 98). All quality-filtered reads were aligned to the human reference genome (GRCh37) from the Ensembl database using Tophat (99). Aligned reads were sorted using Samtools, and the number of reads for each gene were counted using HTSeq (100, 101). Differentially expressed genes were filtered using the criteria of fold change > 2.0 and P < 0.05 (adjusted using the Benjamini–Hochberg procedure) (79). Only genes above a threshold for expression in the positive cell lines were included. These genes were categorized for GO with the DAVID v6.7 program restricting to categories of biological processes (https://david.ncifcrf.gov) (102, 103). BioVenn was used to generate area-proportional Venn diagrams (www.cmbi.ru.nl/cdd/biovenn/) (104).

Chromatin Immunoprecipitation

Cell lines were grown to 80–90% confluence in 10-cm plates, washed twice with PBS, and cross-linked with 1% formaldehyde in PBS for 15 min at RT. Glycine was added to 0.125 M for 5 min to quench the cross-linker; then, cells were washed twice with PBS. Cells were scraped into PBS with protease inhibitors (same as above) and centrifuged at 8,000 × g for 5 min at 4 °C. Cells were then resuspended in lysis buffer [50 mM Tris⋅HCl (pH 8.0), 150 mM NaCl, and 1% Nonidet P-40 with the same protease inhibitors as above]. Batches of cell lysates (2 mL) were sonicated for 4 min each, consisting of 20-s pulses at 24 W, with a Misonix Sonicator 3000. Debris was removed by centrifugation for 10 min at 4 °C at 14,000 × g. Supernatant was isolated, and protein levels were quantitated using Bradford reagent and normalized to 1.5 mg/mL aliquots. An aliquot (500 μL) was saved as input for each cell line, and was frozen with liquid nitrogen and stored at −80 °C. Aliquots for chromatin immunoprecipitation were precleared with 1 μg/mL BSA and salmon sperm DNA (Invitrogen) for 2 h. Previously, Protein A-Protein G beads (Thermo Fisher, GE Healthcare) were washed three times, resuspended in the same lysis buffer as above, and then rotated overnight at 4 °C with control IgG or specific antibodies, as described above for Western blots and with anti-RNA polymerase II p-Ser5 CTD (ab5408; Abcam). For the chromatin immunoprecipitation, 100 μL of antibody-bound beads (1 μg of antibody) was added to precleared aliquots of sonicated lysates and rotated overnight at 4 °C. The following day, beads were washed five times with radioimmunoprecipitation assay buffer [150 mM sodium chloride, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, and 50 mM Tris (pH 8.0)] and then once with TE [10 mM Tris⋅Cl (pH 7.5) and 1 mM EDTA], with a 5-min rotation between each wash. Beads were resuspended in 100 μL of TE, and 200 μL 1.5× Talianidis Elution Buffer [105 mM Tris⋅Cl (pH 8.0), 1.5 mM EDTA (pH 8.0), and 2.25% (wt/vol) SDS] was added, along with 1 μg/mL RNase A (Qiagen). Beads were left to elute at RT for 1 h, and then for 10 min at 65 °C. At this point, NaCl was added to 0.2 M to both inputs and samples, and they were incubated at 65 °C for 2 h to reverse cross-linking. Proteinase K (20 μg/mL) was added for the last hour. Finally, samples and inputs were subjected to ethanol precipitation and resuspended in 300 μL ddH₂O. Eluates (8 μL) were analyzed by qPCR as above (primer sets in Dataset S4). Normalization was set to 2.5 ng of input DNA for each sample.

Tissue Microarrays

Three TNBC TMAs were constructed under Columbia University Medical Center (CUMC) IRB protocol AAAB2667 (Research with Human Surgical Specimens). Cases were selected for inclusion from the clinical records at CUMC matching the following criteria: invasive breast cases negative for estrogen and progesterone receptor expression (<10% nuclear staining) by immunohistochemistry (IHC) and Her2-negative for overexpression by IHC (HercepTest score of 0, 1, or 2, with the latter also FISH- or SISH (Silver in situ hybridization)-negative for Her2 amplification). Three cores of invasive breast carcinoma (1 mm in diameter) and, in many cases, three cores of adjacent normal (nonneoplastic) breast tissue were represented per case in the array. Regarding staining, a 10 mM citrate buffer (pH 6.0), adapted from Brown and Chirala (105), was used for heat-induced antigen retrieval. The ΔNp63 antibody from Santa Cruz Biotechnology (sc-71825) was used at 1:100 for 90 min at RT. The secondary antibody was mouse EnVision (Dako), and hematoxylin was counterstained.

Acknowledgments

We thank Drs. Ella Freulich, Yan Zhu, Wen Zhou, and members of the C.P. and R.P. laboratories for assistance and advice. We thank Dr. Scott Lowe for the pBabepuro, pBabepuro-H-RasV12, and LMP vectors and Phoenix-Ampho cells. We thank Dr. Rosalie Sears for the MCF10A-TR cells and Dr. Kumiko Ui-Tei for the E-cadherin minimal reporter. We thank Dr. Antonio Costanzo for the gift of the −3,000 ∆Np63 promoter construct. We acknowledge the Texas Advanced Computing Center at The University of Texas at Austin for providing access to the high-performance computing system Stampede and the Pittsburgh Supercomputing Center for the Blacklight system. This work was supported by NIH Grant P01 CA87497. This work used the Extreme Science and Engineering Discovery Environment (XSEDE), which is supported by National Science Foundation Grant ACI-1053575. The XSEDE computational grant (to C.P. and K.R.) helped support data analysis carried out on the Stampede and Blacklight supercomputers.

Footnotes

↵¹Present address: College of Physicians and Surgeons, Columbia University, New York, NY 10032.
↵²To whom correspondence may be addressed. Email: clp3{at}columbia.edu or mrp6{at}columbia.edu.

Author contributions: K.E.Y., C.P., and R.P. designed research; K.E.Y., A.G., S.-M.L., N.T.P., O.A., and R.P. performed research; N.T.P. contributed new reagents/analytic tools; K.E.Y., K.R., A.G., S.-M.L., L.J.K., C.P., and R.P. analyzed data; and K.E.Y., C.P., and R.P. wrote the paper.
Reviewers: N.M.B., Harvard Stem Cell Institute; L.E., Harvard Medical School; and R.A.W., Whitehead Institute for Biomedical Research; Department of Biology, Massachusetts Institute of Technology; Ludwig MIT Center for Molecular Oncology, Cambridge.
The authors declare no conflict of interest.
Data deposition: The H-Ras RNA-sequencing data were deposited in the Gene Expression Omnibus (GEO) database, www.ncbi.nlm.nih.gov/geo/ (accession no. GSE81593). Data from the sh-∆Np63 RNA-sequencing are hosted at a Columbia University website (biology.columbia.edu/privesngs).
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1613417113/-/DCSupplemental.

References

↵
1. DeFeo D, et al.
(1981) Analysis of two divergent rat genomic clones homologous to the transforming gene of Harvey murine sarcoma virus. Proc Natl Acad Sci USA 78(6):3328–3332.
.
OpenUrl Abstract/FREE Full Text
↵
1. Forbes SA, et al.
(2011) COSMIC: Mining complete cancer genomes in the Catalogue of Somatic Mutations in Cancer. Nucleic Acids Res 39(Database issue):D945–D950.
.
OpenUrl Abstract/FREE Full Text
↵
1. Prior IA,
2. Lewis PD,
3. Mattos C
(2012) A comprehensive survey of Ras mutations in cancer. Cancer Res 72(10):2457–2467.
.
OpenUrl Abstract/FREE Full Text
↵
1. Slamon DJ, et al.
(1987) Human breast cancer: Correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science 235(4785):177–182.
.
OpenUrl Abstract/FREE Full Text
↵
1. Varley JM,
2. Swallow JE,
3. Brammar WJ,
4. Whittaker JL,
5. Walker RA
(1987) Alterations to either c-erbB-2(neu) or c-myc proto-oncogenes in breast carcinomas correlate with poor short-term prognosis. Oncogene 1(4):423–430.
.
OpenUrl PubMed
↵
1. Narayan M, et al.
(2009) Trastuzumab-induced HER reprogramming in “resistant” breast carcinoma cells. Cancer Res 69(6):2191–2194.
.
OpenUrl Abstract/FREE Full Text
↵
1. Gajria D,
2. Chandarlapaty S
(2011) HER2-amplified breast cancer: Mechanisms of trastuzumab resistance and novel targeted therapies. Expert Rev Anticancer Ther 11(2):263–275.
.
OpenUrl CrossRef PubMed
↵
1. Samuels Y, et al.
(2004) High frequency of mutations of the PIK3CA gene in human cancers. Science 304(5670):554.
.
OpenUrl FREE Full Text
↵
1. Cancer Genome Atlas Network
(2012) Comprehensive molecular portraits of human breast tumours. Nature 490(7418):61–70.
.
OpenUrl CrossRef PubMed
↵
1. Zhao JJ, et al.
(2005) The oncogenic properties of mutant p110α and p110β phosphatidylinositol 3-kinases in human mammary epithelial cells. Proc Natl Acad Sci USA 102(51):18443–18448.
.
OpenUrl Abstract/FREE Full Text
↵
1. Kang S,
2. Bader AG,
3. Vogt PK
(2005) Phosphatidylinositol 3-kinase mutations identified in human cancer are oncogenic. Proc Natl Acad Sci USA 102(3):802–807.
.
OpenUrl Abstract/FREE Full Text
↵
1. Perl AK,
2. Wilgenbus P,
3. Dahl U,
4. Semb H,
5. Christofori G
(1998) A causal role for E-cadherin in the transition from adenoma to carcinoma. Nature 392(6672):190–193.
.
OpenUrl CrossRef PubMed
↵
1. Krakhmal NV,
2. Zavyalova MV,
3. Denisov EV,
4. Vtorushin SV,
5. Perelmuter VM
(2015) Cancer invasion: Patterns and mechanisms. Acta Naturae 7(2):17–28.
.
OpenUrl PubMed
↵
1. De Craene B,
2. Berx G
(2013) Regulatory networks defining EMT during cancer initiation and progression. Nat Rev Cancer 13(2):97–110.
.
OpenUrl CrossRef PubMed
↵
1. Comijn J, et al.
(2001) The two-handed E box binding zinc finger protein SIP1 downregulates E-cadherin and induces invasion. Mol Cell 7(6):1267–1278.
.
OpenUrl CrossRef PubMed
↵
1. Koster MI,
2. Kim S,
3. Mills AA,
4. DeMayo FJ,
5. Roop DR
(2004) p63 is the molecular switch for initiation of an epithelial stratification program. Genes Dev 18(2):126–131.
.
OpenUrl Abstract/FREE Full Text
↵
1. Barbieri CE,
2. Pietenpol JA
(2006) p63 and epithelial biology. Exp Cell Res 312(6):695–706.
.
OpenUrl CrossRef PubMed
↵
1. Yang A, et al.
(1999) p63 is essential for regenerative proliferation in limb, craniofacial and epithelial development. Nature 398(6729):714–718.
.
OpenUrl CrossRef PubMed
↵
1. Mills AA, et al.
(1999) p63 is a p53 homologue required for limb and epidermal morphogenesis. Nature 398(6729):708–713.
.
OpenUrl CrossRef PubMed
↵
1. Yang A, et al.
(1998) p63, a p53 homolog at 3q27-29, encodes multiple products with transactivating, death-inducing, and dominant-negative activities. Mol Cell 2(3):305–316.
.
OpenUrl CrossRef PubMed
↵
1. Koster MI, et al.
(2007) p63 induces key target genes required for epidermal morphogenesis. Proc Natl Acad Sci USA 104(9):3255–3260.
.
OpenUrl Abstract/FREE Full Text
↵
1. Romano R-A, et al.
(2012) ΔNp63 knockout mice reveal its indispensable role as a master regulator of epithelial development and differentiation. Development 139(4):772–782.
.
OpenUrl Abstract/FREE Full Text
↵
1. Hibi K, et al.
(2000) AIS is an oncogene amplified in squamous cell carcinoma. Proc Natl Acad Sci USA 97(10):5462–5467.
.
OpenUrl Abstract/FREE Full Text
↵
1. Urist MJ, et al.
(2002) Loss of p63 expression is associated with tumor progression in bladder cancer. Am J Pathol 161(4):1199–1206.
.
OpenUrl CrossRef PubMed
↵
1. Dairkee SH,
2. Blayney C,
3. Smith HS,
4. Hackett AJ
(1985) Monoclonal antibody that defines human myoepithelium. Proc Natl Acad Sci USA 82(21):7409–7413.
.
OpenUrl Abstract/FREE Full Text
↵
1. Nylander K, et al.
(2002) Differential expression of p63 isoforms in normal tissues and neoplastic cells. J Pathol 198(4):417–427.
.
OpenUrl CrossRef PubMed
↵
1. Tucci P, et al.
(2012) Loss of p63 and its microRNA-205 target results in enhanced cell migration and metastasis in prostate cancer. Proc Natl Acad Sci USA 109(38):15312–15317.
.
OpenUrl Abstract/FREE Full Text
↵
1. Tran MN, et al.
(2013) The p63 protein isoform ΔNp63α inhibits epithelial-mesenchymal transition in human bladder cancer cells: Role of MIR-205. J Biol Chem 288(5):3275–3288.
.
OpenUrl Abstract/FREE Full Text
↵
1. Lindsay J,
2. McDade SS,
3. Pickard A,
4. McCloskey KD,
5. McCance DJ
(2011) Role of DeltaNp63γ in epithelial to mesenchymal transition. J Biol Chem 286(5):3915–3924.
.
OpenUrl Abstract/FREE Full Text
↵
1. Weiss MB, et al.
(2010) Deletion of p53 in human mammary epithelial cells causes chromosomal instability and altered therapeutic response. Oncogene 29(33):4715–4724.
.
OpenUrl CrossRef PubMed
↵
1. Soule HD, et al.
(1990) Isolation and characterization of a spontaneously immortalized human breast epithelial cell line, MCF-10. Cancer Res 50(18):6075–6086.
.
OpenUrl Abstract/FREE Full Text
↵
1. Barbareschi M, et al.
(2001) p63, a p53 homologue, is a selective nuclear marker of myoepithelial cells of the human breast. Am J Surg Pathol 25(8):1054–1060.
.
OpenUrl CrossRef PubMed
↵
1. Miller FR, et al.
(1993) Xenograft model of progressive human proliferative breast disease. J Natl Cancer Inst 85(21):1725–1732.
.
OpenUrl Abstract/FREE Full Text
↵
1. Carroll DK, et al.
(2006) p63 regulates an adhesion programme and cell survival in epithelial cells. Nat Cell Biol 8(6):551–561.
.
OpenUrl CrossRef PubMed
↵
1. Lanza M, et al.
(2006) Cross-talks in the p53 family: DeltaNp63 is an anti-apoptotic target for deltaNp73alpha and p53 gain-of-function mutants. Cell Cycle 5(17):1996–2004.
.
OpenUrl CrossRef PubMed
↵
1. Dorn A,
2. Bollekens J,
3. Staub A,
4. Benoist C,
5. Mathis D
(1987) A multiplicity of CCAAT box-binding proteins. Cell 50(6):863–872.
.
OpenUrl CrossRef PubMed
↵
1. Sinha S,
2. Maity SN,
3. Lu J,
4. de Crombrugghe B
(1995) Recombinant rat CBF-C, the third subunit of CBF/NFY, allows formation of a protein-DNA complex with CBF-A and CBF-B and with yeast HAP2 and HAP3. Proc Natl Acad Sci USA 92(5):1624–1628.
.
OpenUrl Abstract/FREE Full Text
↵
1. Dynan WS,
2. Tjian R
(1983) The promoter-specific transcription factor Sp1 binds to upstream sequences in the SV40 early promoter. Cell 35(1):79–87.
.
OpenUrl CrossRef PubMed
↵
1. Kalluri R,
2. Weinberg RA
(2009) The basics of epithelial-mesenchymal transition. J Clin Invest 119(6):1420–1428.
.
OpenUrl CrossRef PubMed
↵
1. Guaita S, et al.
(2002) Snail induction of epithelial to mesenchymal transition in tumor cells is accompanied by MUC1 repression and ZEB1 expression. J Biol Chem 277(42):39209–39216.
.
OpenUrl Abstract/FREE Full Text
↵
1. Vandewalle C, et al.
(2005) SIP1/ZEB2 induces EMT by repressing genes of different epithelial cell-cell junctions. Nucleic Acids Res 33(20):6566–6578.
.
OpenUrl Abstract/FREE Full Text
↵
1. Gregory PA, et al.
(2008) The miR-200 family and miR-205 regulate epithelial to mesenchymal transition by targeting ZEB1 and SIP1. Nat Cell Biol 10(5):593–601.
.
OpenUrl CrossRef PubMed
↵
1. Korpal M,
2. Lee ES,
3. Hu G,
4. Kang Y
(2008) The miR-200 family inhibits epithelial-mesenchymal transition and cancer cell migration by direct targeting of E-cadherin transcriptional repressors ZEB1 and ZEB2. J Biol Chem 283(22):14910–14914.
.
OpenUrl Abstract/FREE Full Text
↵
1. Knouf EC, et al.
(2012) An integrative genomic approach identifies p73 and p63 as activators of miR-200 microRNA family transcription. Nucleic Acids Res 40(2):499–510.
.
OpenUrl Abstract/FREE Full Text
↵
1. Grünert S,
2. Jechlinger M,
3. Beug H
(2003) Diverse cellular and molecular mechanisms contribute to epithelial plasticity and metastasis. Nat Rev Mol Cell Biol 4(8):657–665.
.
OpenUrl CrossRef PubMed
↵
1. Bachman KE, et al.
(2004) The PIK3CA gene is mutated with high frequency in human breast cancers. Cancer Biol Ther 3(8):772–775.
.
OpenUrl CrossRef PubMed
↵
1. Krause S,
2. Maffini MV,
3. Soto AM,
4. Sonnenschein C
(2008) A novel 3D in vitro culture model to study stromal-epithelial interactions in the mammary gland. Tissue Eng Part C Methods 14(3):261–271.
.
OpenUrl CrossRef PubMed
↵
1. Thiery JP,
2. Acloque H,
3. Huang RYJ,
4. Nieto MA
(2009) Epithelial-mesenchymal transitions in development and disease. Cell 139(5):871–890.
.
OpenUrl CrossRef PubMed
↵
1. Zheng H,
2. Kang Y
(2014) Multilayer control of the EMT master regulators. Oncogene 33(14):1755–1763.
.
OpenUrl CrossRef PubMed
↵
1. Friedl P,
2. Wolf K
(2010) Plasticity of cell migration: A multiscale tuning model. J Cell Biol 188(1):11–19.
.
OpenUrl Abstract/FREE Full Text
↵
1. Schiro JA, et al.
(1991) Integrin alpha 2 beta 1 (VLA-2) mediates reorganization and contraction of collagen matrices by human cells. Cell 67(2):403–410.
.
OpenUrl CrossRef PubMed
↵
1. Friedl P, et al.
(1997) Migration of highly aggressive MV3 melanoma cells in 3-dimensional collagen lattices results in local matrix reorganization and shedding of alpha2 and beta1 integrins and CD44. Cancer Res 57(10):2061–2070.
.
OpenUrl Abstract/FREE Full Text
↵
1. Parker JS, et al.
(2009) Supervised risk predictor of breast cancer based on intrinsic subtypes. J Clin Oncol 27(8):1160–1167.
.
OpenUrl Abstract/FREE Full Text
↵
1. Bertucci F, et al.
(2008) How basal are triple-negative breast cancers? Int J Cancer 123(1):236–240.
.
OpenUrl CrossRef PubMed
↵
1. Cheang MCU, et al.
(2008) Basal-like breast cancer defined by five biomarkers has superior prognostic value than triple-negative phenotype. Clin Cancer Res 14(5):1368–1376.
.
OpenUrl Abstract/FREE Full Text
↵
1. Sailer V,
2. Lüders C,
3. Kuhn W,
4. Pelzer V,
5. Kristiansen G
(2015) Immunostaining of ∆Np63 (using the p40 antibody) is equal to that of p63 and CK5/6 in high-grade ductal carcinoma in situ of the breast. Virchows Arch 467(1):67–70.
.
OpenUrl CrossRef PubMed
↵
1. Hu H, et al.
(2002) Elevated expression of p63 protein in human esophageal squamous cell carcinomas. Int J Cancer 102(6):580–583.
.
OpenUrl CrossRef PubMed
↵
1. Oft M, et al.
(1996) TGF-beta1 and Ha-Ras collaborate in modulating the phenotypic plasticity and invasiveness of epithelial tumor cells. Genes Dev 10(19):2462–2477.
.
OpenUrl Abstract/FREE Full Text
↵
1. Janda E, et al.
(2002) Ras and TGF[beta] cooperatively regulate epithelial cell plasticity and metastasis: dissection of Ras signaling pathways. J Cell Biol 156(2):299–313.
.
OpenUrl Abstract/FREE Full Text
↵
1. Romano R-A,
2. Birkaya B,
3. Sinha S
(2006) Defining the regulatory elements in the proximal promoter of DeltaNp63 in keratinocytes: Potential roles for Sp1/Sp3, NF-Y, and p63. J Invest Dermatol 126(7):1469–1479.
.
OpenUrl CrossRef PubMed
↵
1. Yoh K,
2. Prywes R
(2015) Pathway regulation of p63, a director of epithelial cell fate. Front Endocrinol (Lausanne) 6:51.
.
OpenUrl PubMed
↵
1. Adorno M, et al.
(2009) A Mutant-p53/Smad complex opposes p63 to empower TGFbeta-induced metastasis. Cell 137(1):87–98.
.
OpenUrl CrossRef PubMed
↵
1. Muller PAJ, et al.
(2009) Mutant p53 drives invasion by promoting integrin recycling. Cell 139(7):1327–1341.
.
OpenUrl CrossRef PubMed
↵
1. Stefanou D,
2. Batistatou A,
3. Nonni A,
4. Arkoumani E,
5. Agnantis NJ
(2004) p63 expression in benign and malignant breast lesions. Histol Histopathol 19(2):465–471.
.
OpenUrl PubMed
↵
1. Hanker L, et al.
(2010) Clinical relevance of the putative stem cell marker p63 in breast cancer. Breast Cancer Res Treat 122(3):765–775.
.
OpenUrl CrossRef PubMed
1. Leong C-O,
2. Vidnovic N,
3. DeYoung MP,
4. Sgroi D,
5. Ellisen LW
(2007) The p63/p73 network mediates chemosensitivity to cisplatin in a biologically defined subset of primary breast cancers. J Clin Invest 117(5):1370–1380.
.
OpenUrl CrossRef PubMed
↵
1. Li H, et al.
(2007) Nestin is expressed in the basal/myoepithelial layer of the mammary gland and is a selective marker of basal epithelial breast tumors. Cancer Res 67(2):501–510.
.
OpenUrl Abstract/FREE Full Text
↵
1. Reis-Filho JS, et al.
(2003) Distribution of p63, cytokeratins 5/6 and cytokeratin 14 in 51 normal and 400 neoplastic human tissue samples using TARP-4 multi-tumor tissue microarray. Virchows Arch 443(2):122–132.
.
OpenUrl CrossRef PubMed
↵
1. Lim E, et al., kConFab
(2009) Aberrant luminal progenitors as the candidate target population for basal tumor development in BRCA1 mutation carriers. Nat Med 15(8):907–913.
.
OpenUrl CrossRef PubMed
↵
1. Molyneux G, et al.
(2010) BRCA1 basal-like breast cancers originate from luminal epithelial progenitors and not from basal stem cells. Cell Stem Cell 7(3):403–417.
.
OpenUrl CrossRef PubMed
↵
1. Fu N,
2. Lindeman GJ,
3. Visvader JE
(2014) The mammary stem cell hierarchy. Curr Top Dev Biol 107:133–160.
.
OpenUrl CrossRef PubMed
↵
1. Prater MD, et al.
(2014) Mammary stem cells have myoepithelial cell properties. Nat Cell Biol 16(10):942–950, 1–7.
.
OpenUrl CrossRef PubMed
↵
1. Wan L,
2. Pantel K,
3. Kang Y
(2013) Tumor metastasis: Moving new biological insights into the clinic. Nat Med 19(11):1450–1464.
.
OpenUrl CrossRef PubMed
↵
1. Friedl P,
2. Alexander S
(2011) Cancer invasion and the microenvironment: Plasticity and reciprocity. Cell 147(5):992–1009.
.
OpenUrl CrossRef PubMed
↵
1. Fischer KR, et al.
(2015) Epithelial-to-mesenchymal transition is not required for lung metastasis but contributes to chemoresistance. Nature 527(7579):472–476.
.
OpenUrl CrossRef PubMed
↵
1. Zheng X, et al.
(2015) Epithelial-to-mesenchymal transition is dispensable for metastasis but induces chemoresistance in pancreatic cancer. Nature 527(7579):525–530.
.
OpenUrl CrossRef PubMed
↵
1. Farrell AS, et al.
(2013) Pin1 regulates the dynamics of c-Myc DNA binding to facilitate target gene regulation and oncogenesis. Mol Cell Biol 33(15):2930–2949.
.
OpenUrl Abstract/FREE Full Text
↵
1. Ivascu A,
2. Kubbies M
(2006) Rapid generation of single-tumor spheroids for high-throughput cell function and toxicity analysis. J Biomol Screen 11(8):922–932.
.
OpenUrl Abstract/FREE Full Text
↵
1. Anders S,
2. Huber W
(2010) Differential expression analysis for sequence count data. Genome Biol 11(10):R106.
.
OpenUrl CrossRef PubMed
↵
1. Mazda M,
2. Nishi K,
3. Naito Y,
4. Ui-Tei K
(2011) E-cadherin is transcriptionally activated via suppression of ZEB1 transcriptional repressor by small RNA-mediated gene silencing. PLoS One 6(12):e28688.
.
OpenUrl CrossRef PubMed
↵
1. Wang Y,
2. Prywes R
(2000) Activation of the c-fos enhancer by the erk MAP kinase pathway through two sequence elements: The c-fos AP-1 and p62TCF sites. Oncogene 19(11):1379–1385.
.
OpenUrl CrossRef PubMed
↵
1. Li Y,
2. Peart MJ,
3. Prives C
(2009) Stxbp4 regulates DeltaNp63 stability by suppression of RACK1-dependent degradation. Mol Cell Biol 29(14):3953–3963.
.
OpenUrl Abstract/FREE Full Text
↵
1. Morgenstern JP,
2. Land H
(1990) Advanced mammalian gene transfer: High titre retroviral vectors with multiple drug selection markers and a complementary helper-free packaging cell line. Nucleic Acids Res 18(12):3587–3596.
.
OpenUrl Abstract/FREE Full Text
↵
1. Zuber J, et al.
(2011) Toolkit for evaluating genes required for proliferation and survival using tetracycline-regulated RNAi. Nat Biotechnol 29(1):79–83.
.
OpenUrl CrossRef PubMed
↵
1. Sobie EA
(2011) An introduction to MATLAB. Sci Signal 4(191):tr7.
.
OpenUrl Abstract/FREE Full Text
↵
1. Li B,
2. Dewey CN
(2011) RSEM: Accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics 12:323.
.
OpenUrl CrossRef PubMed
↵
1. Jeanmougin M, et al.
(2010) Should we abandon the t-test in the analysis of gene expression microarray data: A comparison of variance modeling strategies. PLoS One 5(9):e12336.
.
OpenUrl CrossRef PubMed
↵
1. Hochberg Y,
2. Benjamini Y
(1990) More powerful procedures for multiple significance testing. Stat Med 9(7):811–818.
.
OpenUrl CrossRef PubMed
↵
1. Benjamini Y,
2. Hochberg Y
(1995) Controlling the false discovery rate: A practical and powerful approach to multiple testing. J R Stat Soc Series B Stat Methodol 57(1):289–300.
.
OpenUrl
↵
1. Krzywinski M,
2. Altman N
(2014) Visualizing samples with box plots. Nat Methods 11(2):119–120.
.
OpenUrl CrossRef PubMed
↵
1. Streit M,
2. Gehlenborg N
(2014) Bar charts and box plots. Nat Methods 11(2):117.
.
OpenUrl CrossRef PubMed
↵
1. Gao J, et al.
(2013) Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Sci Signal 6(269):pl1.
.
OpenUrl Abstract/FREE Full Text
↵
1. Bastien RRL, et al.
(2012) PAM50 breast cancer subtyping by RT-qPCR and concordance with standard clinical molecular markers. BMC Med Genomics 5:44.
.
OpenUrl CrossRef PubMed
↵
1. Zhu J, et al.
(2009) The UCSC Cancer Genomics Browser. Nat Methods 6(4):239–240.
.
OpenUrl CrossRef PubMed
↵
1. Goldman M, et al.
(2015) The UCSC Cancer Genomics Browser: Update 2015. Nucleic Acids Res 43(Database issue):D812–D817.
.
OpenUrl Abstract/FREE Full Text
↵
1. Towns J, et al.
(2014) XSEDE: Accelerating scientific discovery. Comput Sci Eng 16(5):62–74.
.
OpenUrl CrossRef
↵
1. Andrews S
(2010) FastQC: A quality control tool for high throughput sequence data. Available at www.bioinformatics.babraham.ac.uk/projects/fastqc/. Accessed February 15, 2015.
.
↵
1. Bolger AM,
2. Lohse M,
3. Usadel B
(2014) Trimmomatic: A flexible trimmer for Illumina sequence data. Bioinformatics 30(15):2114–2120.
.
OpenUrl Abstract/FREE Full Text
↵
1. Trapnell C,
2. Pachter L,
3. Salzberg SL
(2009) TopHat: Discovering splice junctions with RNA-Seq. Bioinformatics 25(9):1105–1111.
.
OpenUrl Abstract/FREE Full Text
↵
1. Li H, et al., 1000 Genome Project Data Processing Subgroup
(2009) The Sequence Alignment/Map format and SAMtools. Bioinformatics 25(16):2078–2079.
.
OpenUrl Abstract/FREE Full Text
↵
1. Anders S,
2. Pyl PT,
3. Huber W
(2015) HTSeq--a Python framework to work with high-throughput sequencing data. Bioinformatics 31(2):166–169.
.
OpenUrl Abstract/FREE Full Text
↵
1. Huang W,
2. Sherman BT,
3. Lempicki RA
(2009) Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 4(1):44–57.
.
OpenUrl CrossRef PubMed
↵
1. Huang W,
2. Sherman BT,
3. Lempicki RA
(2009) Bioinformatics enrichment tools: Paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res 37(1):1–13.
.
OpenUrl Abstract/FREE Full Text
↵
1. Hulsen T,
2. de Vlieg J,
3. Alkema W
(2008) BioVenn - a web application for the comparison and visualization of biological lists using area-proportional Venn diagrams. BMC Genomics 9:488.
.
OpenUrl CrossRef PubMed
↵
1. Brown RW,
2. Chirala R
(1995) Utility of microwave-citrate antigen retrieval in diagnostic immunohistochemistry. Mod Pathol 8(5):515–520.
.
OpenUrl PubMed

Article Alerts

Email Article

Citation Tools

Request Permissions

Article Classifications

Biological Sciences
Cell Biology

Proceedings of the National Academy of Sciences: 113 (41)

Table of Contents

Submit

[1] ↵
DeFeo D, et al.
(1981) Analysis of two divergent rat genomic clones homologous to the transforming gene of Harvey murine sarcoma virus. Proc Natl Acad Sci USA 78(6):3328–3332.
.
OpenUrl Abstract/FREE Full Text

[2] DeFeo D, et al.

[3] ↵
Forbes SA, et al.
(2011) COSMIC: Mining complete cancer genomes in the Catalogue of Somatic Mutations in Cancer. Nucleic Acids Res 39(Database issue):D945–D950.
.
OpenUrl Abstract/FREE Full Text

[4] Forbes SA, et al.

[5] ↵
Prior IA,
Lewis PD,
Mattos C
(2012) A comprehensive survey of Ras mutations in cancer. Cancer Res 72(10):2457–2467.
.
OpenUrl Abstract/FREE Full Text

[6] Prior IA,

[7] Lewis PD,

[8] Mattos C

[9] ↵
Slamon DJ, et al.
(1987) Human breast cancer: Correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science 235(4785):177–182.
.
OpenUrl Abstract/FREE Full Text

[10] Slamon DJ, et al.

[11] ↵
Varley JM,
Swallow JE,
Brammar WJ,
Whittaker JL,
Walker RA
(1987) Alterations to either c-erbB-2(neu) or c-myc proto-oncogenes in breast carcinomas correlate with poor short-term prognosis. Oncogene 1(4):423–430.
.
OpenUrl PubMed

[12] Varley JM,

[13] Swallow JE,

[14] Brammar WJ,

[15] Whittaker JL,

[16] Walker RA

[17] ↵
Narayan M, et al.
(2009) Trastuzumab-induced HER reprogramming in “resistant” breast carcinoma cells. Cancer Res 69(6):2191–2194.
.
OpenUrl Abstract/FREE Full Text

[18] Narayan M, et al.

[19] ↵
Gajria D,
Chandarlapaty S
(2011) HER2-amplified breast cancer: Mechanisms of trastuzumab resistance and novel targeted therapies. Expert Rev Anticancer Ther 11(2):263–275.
.
OpenUrl CrossRef PubMed

[20] Gajria D,

[21] Chandarlapaty S

[22] ↵
Samuels Y, et al.
(2004) High frequency of mutations of the PIK3CA gene in human cancers. Science 304(5670):554.
.
OpenUrl FREE Full Text

[23] Samuels Y, et al.

[24] ↵
Cancer Genome Atlas Network
(2012) Comprehensive molecular portraits of human breast tumours. Nature 490(7418):61–70.
.
OpenUrl CrossRef PubMed

[25] Cancer Genome Atlas Network

[26] ↵
Zhao JJ, et al.
(2005) The oncogenic properties of mutant p110α and p110β phosphatidylinositol 3-kinases in human mammary epithelial cells. Proc Natl Acad Sci USA 102(51):18443–18448.
.
OpenUrl Abstract/FREE Full Text

[27] Zhao JJ, et al.

[28] ↵
Kang S,
Bader AG,
Vogt PK
(2005) Phosphatidylinositol 3-kinase mutations identified in human cancer are oncogenic. Proc Natl Acad Sci USA 102(3):802–807.
.
OpenUrl Abstract/FREE Full Text

[29] Kang S,

[30] Bader AG,

[31] Vogt PK

[32] ↵
Perl AK,
Wilgenbus P,
Dahl U,
Semb H,
Christofori G
(1998) A causal role for E-cadherin in the transition from adenoma to carcinoma. Nature 392(6672):190–193.
.
OpenUrl CrossRef PubMed

[33] Perl AK,

[34] Wilgenbus P,

[35] Dahl U,

[36] Semb H,

[37] Christofori G

[38] ↵
Krakhmal NV,
Zavyalova MV,
Denisov EV,
Vtorushin SV,
Perelmuter VM
(2015) Cancer invasion: Patterns and mechanisms. Acta Naturae 7(2):17–28.
.
OpenUrl PubMed

[39] Krakhmal NV,

[40] Zavyalova MV,

[41] Denisov EV,

[42] Vtorushin SV,

[43] Perelmuter VM

[44] ↵
De Craene B,
Berx G
(2013) Regulatory networks defining EMT during cancer initiation and progression. Nat Rev Cancer 13(2):97–110.
.
OpenUrl CrossRef PubMed

[45] De Craene B,

[46] Berx G

[47] ↵
Comijn J, et al.
(2001) The two-handed E box binding zinc finger protein SIP1 downregulates E-cadherin and induces invasion. Mol Cell 7(6):1267–1278.
.
OpenUrl CrossRef PubMed

[48] Comijn J, et al.

[49] ↵
Koster MI,
Kim S,
Mills AA,
DeMayo FJ,
Roop DR
(2004) p63 is the molecular switch for initiation of an epithelial stratification program. Genes Dev 18(2):126–131.
.
OpenUrl Abstract/FREE Full Text

[50] Koster MI,

[51] Kim S,

[52] Mills AA,

[53] DeMayo FJ,

[54] Roop DR

[55] ↵
Barbieri CE,
Pietenpol JA
(2006) p63 and epithelial biology. Exp Cell Res 312(6):695–706.
.
OpenUrl CrossRef PubMed

[56] Barbieri CE,

[57] Pietenpol JA

[58] ↵
Yang A, et al.
(1999) p63 is essential for regenerative proliferation in limb, craniofacial and epithelial development. Nature 398(6729):714–718.
.
OpenUrl CrossRef PubMed

[59] Yang A, et al.

[60] ↵
Mills AA, et al.
(1999) p63 is a p53 homologue required for limb and epidermal morphogenesis. Nature 398(6729):708–713.
.
OpenUrl CrossRef PubMed

[61] Mills AA, et al.

[62] ↵
Yang A, et al.
(1998) p63, a p53 homolog at 3q27-29, encodes multiple products with transactivating, death-inducing, and dominant-negative activities. Mol Cell 2(3):305–316.
.
OpenUrl CrossRef PubMed

[63] Yang A, et al.

[64] ↵
Koster MI, et al.
(2007) p63 induces key target genes required for epidermal morphogenesis. Proc Natl Acad Sci USA 104(9):3255–3260.
.
OpenUrl Abstract/FREE Full Text

[65] Koster MI, et al.

[66] ↵
Romano R-A, et al.
(2012) ΔNp63 knockout mice reveal its indispensable role as a master regulator of epithelial development and differentiation. Development 139(4):772–782.
.
OpenUrl Abstract/FREE Full Text

[67] Romano R-A, et al.

[68] ↵
Hibi K, et al.
(2000) AIS is an oncogene amplified in squamous cell carcinoma. Proc Natl Acad Sci USA 97(10):5462–5467.
.
OpenUrl Abstract/FREE Full Text

[69] Hibi K, et al.

[70] ↵
Urist MJ, et al.
(2002) Loss of p63 expression is associated with tumor progression in bladder cancer. Am J Pathol 161(4):1199–1206.
.
OpenUrl CrossRef PubMed

[71] Urist MJ, et al.

[72] ↵
Dairkee SH,
Blayney C,
Smith HS,
Hackett AJ
(1985) Monoclonal antibody that defines human myoepithelium. Proc Natl Acad Sci USA 82(21):7409–7413.
.
OpenUrl Abstract/FREE Full Text

[73] Dairkee SH,

[74] Blayney C,

[75] Smith HS,

[76] Hackett AJ

[77] ↵
Nylander K, et al.
(2002) Differential expression of p63 isoforms in normal tissues and neoplastic cells. J Pathol 198(4):417–427.
.
OpenUrl CrossRef PubMed

[78] Nylander K, et al.

[79] ↵
Tucci P, et al.
(2012) Loss of p63 and its microRNA-205 target results in enhanced cell migration and metastasis in prostate cancer. Proc Natl Acad Sci USA 109(38):15312–15317.
.
OpenUrl Abstract/FREE Full Text

[80] Tucci P, et al.

[81] ↵
Tran MN, et al.
(2013) The p63 protein isoform ΔNp63α inhibits epithelial-mesenchymal transition in human bladder cancer cells: Role of MIR-205. J Biol Chem 288(5):3275–3288.
.
OpenUrl Abstract/FREE Full Text

[82] Tran MN, et al.

[83] ↵
Lindsay J,
McDade SS,
Pickard A,
McCloskey KD,
McCance DJ
(2011) Role of DeltaNp63γ in epithelial to mesenchymal transition. J Biol Chem 286(5):3915–3924.
.
OpenUrl Abstract/FREE Full Text

[84] Lindsay J,

[85] McDade SS,

[86] Pickard A,

[87] McCloskey KD,

[88] McCance DJ

[89] ↵
Weiss MB, et al.
(2010) Deletion of p53 in human mammary epithelial cells causes chromosomal instability and altered therapeutic response. Oncogene 29(33):4715–4724.
.
OpenUrl CrossRef PubMed

[90] Weiss MB, et al.

[91] ↵
Soule HD, et al.
(1990) Isolation and characterization of a spontaneously immortalized human breast epithelial cell line, MCF-10. Cancer Res 50(18):6075–6086.
.
OpenUrl Abstract/FREE Full Text

[92] Soule HD, et al.

[93] ↵
Barbareschi M, et al.
(2001) p63, a p53 homologue, is a selective nuclear marker of myoepithelial cells of the human breast. Am J Surg Pathol 25(8):1054–1060.
.
OpenUrl CrossRef PubMed

[94] Barbareschi M, et al.

[95] ↵
Miller FR, et al.
(1993) Xenograft model of progressive human proliferative breast disease. J Natl Cancer Inst 85(21):1725–1732.
.
OpenUrl Abstract/FREE Full Text

[96] Miller FR, et al.

[97] ↵
Carroll DK, et al.
(2006) p63 regulates an adhesion programme and cell survival in epithelial cells. Nat Cell Biol 8(6):551–561.
.
OpenUrl CrossRef PubMed

[98] Carroll DK, et al.

[99] ↵
Lanza M, et al.
(2006) Cross-talks in the p53 family: DeltaNp63 is an anti-apoptotic target for deltaNp73alpha and p53 gain-of-function mutants. Cell Cycle 5(17):1996–2004.
.
OpenUrl CrossRef PubMed

[100] Lanza M, et al.

[101] ↵
Dorn A,
Bollekens J,
Staub A,
Benoist C,
Mathis D
(1987) A multiplicity of CCAAT box-binding proteins. Cell 50(6):863–872.
.
OpenUrl CrossRef PubMed

[102] Dorn A,

[103] Bollekens J,

[104] Staub A,

[105] Benoist C,

[106] Mathis D

[107] ↵
Sinha S,
Maity SN,
Lu J,
de Crombrugghe B
(1995) Recombinant rat CBF-C, the third subunit of CBF/NFY, allows formation of a protein-DNA complex with CBF-A and CBF-B and with yeast HAP2 and HAP3. Proc Natl Acad Sci USA 92(5):1624–1628.
.
OpenUrl Abstract/FREE Full Text

[108] Sinha S,

[109] Maity SN,

[110] Lu J,

[111] de Crombrugghe B

[112] ↵
Dynan WS,
Tjian R
(1983) The promoter-specific transcription factor Sp1 binds to upstream sequences in the SV40 early promoter. Cell 35(1):79–87.
.
OpenUrl CrossRef PubMed

[113] Dynan WS,

[114] Tjian R

[115] ↵
Kalluri R,
Weinberg RA
(2009) The basics of epithelial-mesenchymal transition. J Clin Invest 119(6):1420–1428.
.
OpenUrl CrossRef PubMed

[116] Kalluri R,

[117] Weinberg RA

[118] ↵
Guaita S, et al.
(2002) Snail induction of epithelial to mesenchymal transition in tumor cells is accompanied by MUC1 repression and ZEB1 expression. J Biol Chem 277(42):39209–39216.
.
OpenUrl Abstract/FREE Full Text

[119] Guaita S, et al.

[120] ↵
Vandewalle C, et al.
(2005) SIP1/ZEB2 induces EMT by repressing genes of different epithelial cell-cell junctions. Nucleic Acids Res 33(20):6566–6578.
.
OpenUrl Abstract/FREE Full Text

[121] Vandewalle C, et al.

[122] ↵
Gregory PA, et al.
(2008) The miR-200 family and miR-205 regulate epithelial to mesenchymal transition by targeting ZEB1 and SIP1. Nat Cell Biol 10(5):593–601.
.
OpenUrl CrossRef PubMed

[123] Gregory PA, et al.

[124] ↵
Korpal M,
Lee ES,
Hu G,
Kang Y
(2008) The miR-200 family inhibits epithelial-mesenchymal transition and cancer cell migration by direct targeting of E-cadherin transcriptional repressors ZEB1 and ZEB2. J Biol Chem 283(22):14910–14914.
.
OpenUrl Abstract/FREE Full Text

[125] Korpal M,

[126] Lee ES,

[127] Hu G,

[128] Kang Y

[129] ↵
Knouf EC, et al.
(2012) An integrative genomic approach identifies p73 and p63 as activators of miR-200 microRNA family transcription. Nucleic Acids Res 40(2):499–510.
.
OpenUrl Abstract/FREE Full Text

[130] Knouf EC, et al.

[131] ↵
Grünert S,
Jechlinger M,
Beug H
(2003) Diverse cellular and molecular mechanisms contribute to epithelial plasticity and metastasis. Nat Rev Mol Cell Biol 4(8):657–665.
.
OpenUrl CrossRef PubMed

[132] Grünert S,

[133] Jechlinger M,

[134] Beug H

[135] ↵
Bachman KE, et al.
(2004) The PIK3CA gene is mutated with high frequency in human breast cancers. Cancer Biol Ther 3(8):772–775.
.
OpenUrl CrossRef PubMed

[136] Bachman KE, et al.

[137] ↵
Krause S,
Maffini MV,
Soto AM,
Sonnenschein C
(2008) A novel 3D in vitro culture model to study stromal-epithelial interactions in the mammary gland. Tissue Eng Part C Methods 14(3):261–271.
.
OpenUrl CrossRef PubMed

[138] Krause S,

[139] Maffini MV,

[140] Soto AM,

[141] Sonnenschein C

[142] ↵
Thiery JP,
Acloque H,
Huang RYJ,
Nieto MA
(2009) Epithelial-mesenchymal transitions in development and disease. Cell 139(5):871–890.
.
OpenUrl CrossRef PubMed

[143] Thiery JP,

[144] Acloque H,

[145] Huang RYJ,

[146] Nieto MA

[147] ↵
Zheng H,
Kang Y
(2014) Multilayer control of the EMT master regulators. Oncogene 33(14):1755–1763.
.
OpenUrl CrossRef PubMed

[148] Zheng H,

[149] Kang Y

[150] ↵
Friedl P,
Wolf K
(2010) Plasticity of cell migration: A multiscale tuning model. J Cell Biol 188(1):11–19.
.
OpenUrl Abstract/FREE Full Text

[151] Friedl P,

[152] Wolf K

[153] ↵
Schiro JA, et al.
(1991) Integrin alpha 2 beta 1 (VLA-2) mediates reorganization and contraction of collagen matrices by human cells. Cell 67(2):403–410.
.
OpenUrl CrossRef PubMed

[154] Schiro JA, et al.

[155] ↵
Friedl P, et al.
(1997) Migration of highly aggressive MV3 melanoma cells in 3-dimensional collagen lattices results in local matrix reorganization and shedding of alpha2 and beta1 integrins and CD44. Cancer Res 57(10):2061–2070.
.
OpenUrl Abstract/FREE Full Text

[156] Friedl P, et al.

[157] ↵
Parker JS, et al.
(2009) Supervised risk predictor of breast cancer based on intrinsic subtypes. J Clin Oncol 27(8):1160–1167.
.
OpenUrl Abstract/FREE Full Text

[158] Parker JS, et al.

[159] ↵
Bertucci F, et al.
(2008) How basal are triple-negative breast cancers? Int J Cancer 123(1):236–240.
.
OpenUrl CrossRef PubMed

[160] Bertucci F, et al.

[161] ↵
Cheang MCU, et al.
(2008) Basal-like breast cancer defined by five biomarkers has superior prognostic value than triple-negative phenotype. Clin Cancer Res 14(5):1368–1376.
.
OpenUrl Abstract/FREE Full Text

[162] Cheang MCU, et al.

[163] ↵
Sailer V,
Lüders C,
Kuhn W,
Pelzer V,
Kristiansen G
(2015) Immunostaining of ∆Np63 (using the p40 antibody) is equal to that of p63 and CK5/6 in high-grade ductal carcinoma in situ of the breast. Virchows Arch 467(1):67–70.
.
OpenUrl CrossRef PubMed

[164] Sailer V,

[165] Lüders C,

[166] Kuhn W,

[167] Pelzer V,

[168] Kristiansen G

[169] ↵
Hu H, et al.
(2002) Elevated expression of p63 protein in human esophageal squamous cell carcinomas. Int J Cancer 102(6):580–583.
.
OpenUrl CrossRef PubMed

[170] Hu H, et al.

[171] ↵
Oft M, et al.
(1996) TGF-beta1 and Ha-Ras collaborate in modulating the phenotypic plasticity and invasiveness of epithelial tumor cells. Genes Dev 10(19):2462–2477.
.
OpenUrl Abstract/FREE Full Text

[172] Oft M, et al.

[173] ↵
Janda E, et al.
(2002) Ras and TGF[beta] cooperatively regulate epithelial cell plasticity and metastasis: dissection of Ras signaling pathways. J Cell Biol 156(2):299–313.
.
OpenUrl Abstract/FREE Full Text

[174] Janda E, et al.

[175] ↵
Romano R-A,
Birkaya B,
Sinha S
(2006) Defining the regulatory elements in the proximal promoter of DeltaNp63 in keratinocytes: Potential roles for Sp1/Sp3, NF-Y, and p63. J Invest Dermatol 126(7):1469–1479.
.
OpenUrl CrossRef PubMed

[176] Romano R-A,

[177] Birkaya B,

[178] Sinha S

[179] ↵
Yoh K,
Prywes R
(2015) Pathway regulation of p63, a director of epithelial cell fate. Front Endocrinol (Lausanne) 6:51.
.
OpenUrl PubMed

[180] Yoh K,

[181] Prywes R

[182] ↵
Adorno M, et al.
(2009) A Mutant-p53/Smad complex opposes p63 to empower TGFbeta-induced metastasis. Cell 137(1):87–98.
.
OpenUrl CrossRef PubMed

[183] Adorno M, et al.

[184] ↵
Muller PAJ, et al.
(2009) Mutant p53 drives invasion by promoting integrin recycling. Cell 139(7):1327–1341.
.
OpenUrl CrossRef PubMed

[185] Muller PAJ, et al.

[186] ↵
Stefanou D,
Batistatou A,
Nonni A,
Arkoumani E,
Agnantis NJ
(2004) p63 expression in benign and malignant breast lesions. Histol Histopathol 19(2):465–471.
.
OpenUrl PubMed

[187] Stefanou D,

[188] Batistatou A,

[189] Nonni A,

[190] Arkoumani E,

[191] Agnantis NJ

[192] ↵
Hanker L, et al.
(2010) Clinical relevance of the putative stem cell marker p63 in breast cancer. Breast Cancer Res Treat 122(3):765–775.
.
OpenUrl CrossRef PubMed

[193] Hanker L, et al.

[194] Leong C-O,
Vidnovic N,
DeYoung MP,
Sgroi D,
Ellisen LW
(2007) The p63/p73 network mediates chemosensitivity to cisplatin in a biologically defined subset of primary breast cancers. J Clin Invest 117(5):1370–1380.
.
OpenUrl CrossRef PubMed

[195] Leong C-O,

[196] Vidnovic N,

[197] DeYoung MP,

[198] Sgroi D,

[199] Ellisen LW

[200] ↵
Li H, et al.
(2007) Nestin is expressed in the basal/myoepithelial layer of the mammary gland and is a selective marker of basal epithelial breast tumors. Cancer Res 67(2):501–510.
.
OpenUrl Abstract/FREE Full Text

[201] Li H, et al.

[202] ↵
Reis-Filho JS, et al.
(2003) Distribution of p63, cytokeratins 5/6 and cytokeratin 14 in 51 normal and 400 neoplastic human tissue samples using TARP-4 multi-tumor tissue microarray. Virchows Arch 443(2):122–132.
.
OpenUrl CrossRef PubMed

[203] Reis-Filho JS, et al.

[204] ↵
Lim E, et al., kConFab
(2009) Aberrant luminal progenitors as the candidate target population for basal tumor development in BRCA1 mutation carriers. Nat Med 15(8):907–913.
.
OpenUrl CrossRef PubMed

[205] Lim E, et al., kConFab

[206] ↵
Molyneux G, et al.
(2010) BRCA1 basal-like breast cancers originate from luminal epithelial progenitors and not from basal stem cells. Cell Stem Cell 7(3):403–417.
.
OpenUrl CrossRef PubMed

[207] Molyneux G, et al.

[208] ↵
Fu N,
Lindeman GJ,
Visvader JE
(2014) The mammary stem cell hierarchy. Curr Top Dev Biol 107:133–160.
.
OpenUrl CrossRef PubMed

[209] Fu N,

[210] Lindeman GJ,

[211] Visvader JE

[212] ↵
Prater MD, et al.
(2014) Mammary stem cells have myoepithelial cell properties. Nat Cell Biol 16(10):942–950, 1–7.
.
OpenUrl CrossRef PubMed

[213] Prater MD, et al.

[214] ↵
Wan L,
Pantel K,
Kang Y
(2013) Tumor metastasis: Moving new biological insights into the clinic. Nat Med 19(11):1450–1464.
.
OpenUrl CrossRef PubMed

[215] Wan L,

[216] Pantel K,

[217] Kang Y

[218] ↵
Friedl P,
Alexander S
(2011) Cancer invasion and the microenvironment: Plasticity and reciprocity. Cell 147(5):992–1009.
.
OpenUrl CrossRef PubMed

[219] Friedl P,

[220] Alexander S

[221] ↵
Fischer KR, et al.
(2015) Epithelial-to-mesenchymal transition is not required for lung metastasis but contributes to chemoresistance. Nature 527(7579):472–476.
.
OpenUrl CrossRef PubMed

[222] Fischer KR, et al.

[223] ↵
Zheng X, et al.
(2015) Epithelial-to-mesenchymal transition is dispensable for metastasis but induces chemoresistance in pancreatic cancer. Nature 527(7579):525–530.
.
OpenUrl CrossRef PubMed

[224] Zheng X, et al.

[225] ↵
Farrell AS, et al.
(2013) Pin1 regulates the dynamics of c-Myc DNA binding to facilitate target gene regulation and oncogenesis. Mol Cell Biol 33(15):2930–2949.
.
OpenUrl Abstract/FREE Full Text

[226] Farrell AS, et al.

[227] ↵
Ivascu A,
Kubbies M
(2006) Rapid generation of single-tumor spheroids for high-throughput cell function and toxicity analysis. J Biomol Screen 11(8):922–932.
.
OpenUrl Abstract/FREE Full Text

[228] Ivascu A,

[229] Kubbies M

[230] ↵
Anders S,
Huber W
(2010) Differential expression analysis for sequence count data. Genome Biol 11(10):R106.
.
OpenUrl CrossRef PubMed

[231] Anders S,

[232] Huber W

[233] ↵
Mazda M,
Nishi K,
Naito Y,
Ui-Tei K
(2011) E-cadherin is transcriptionally activated via suppression of ZEB1 transcriptional repressor by small RNA-mediated gene silencing. PLoS One 6(12):e28688.
.
OpenUrl CrossRef PubMed

[234] Mazda M,

[235] Nishi K,

[236] Naito Y,

[237] Ui-Tei K

[238] ↵
Wang Y,
Prywes R
(2000) Activation of the c-fos enhancer by the erk MAP kinase pathway through two sequence elements: The c-fos AP-1 and p62TCF sites. Oncogene 19(11):1379–1385.
.
OpenUrl CrossRef PubMed

[239] Wang Y,

[240] Prywes R

[241] ↵
Li Y,
Peart MJ,
Prives C
(2009) Stxbp4 regulates DeltaNp63 stability by suppression of RACK1-dependent degradation. Mol Cell Biol 29(14):3953–3963.
.
OpenUrl Abstract/FREE Full Text

[242] Li Y,

[243] Peart MJ,

[244] Prives C

[245] ↵
Morgenstern JP,
Land H
(1990) Advanced mammalian gene transfer: High titre retroviral vectors with multiple drug selection markers and a complementary helper-free packaging cell line. Nucleic Acids Res 18(12):3587–3596.
.
OpenUrl Abstract/FREE Full Text

[246] Morgenstern JP,

[247] Land H

[248] ↵
Zuber J, et al.
(2011) Toolkit for evaluating genes required for proliferation and survival using tetracycline-regulated RNAi. Nat Biotechnol 29(1):79–83.
.
OpenUrl CrossRef PubMed

[249] Zuber J, et al.

[250] ↵
Sobie EA
(2011) An introduction to MATLAB. Sci Signal 4(191):tr7.
.
OpenUrl Abstract/FREE Full Text

[251] Sobie EA

[252] ↵
Li B,
Dewey CN
(2011) RSEM: Accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics 12:323.
.
OpenUrl CrossRef PubMed

[253] Li B,

[254] Dewey CN

[255] ↵
Jeanmougin M, et al.
(2010) Should we abandon the t-test in the analysis of gene expression microarray data: A comparison of variance modeling strategies. PLoS One 5(9):e12336.
.
OpenUrl CrossRef PubMed

[256] Jeanmougin M, et al.

[257] ↵
Hochberg Y,
Benjamini Y
(1990) More powerful procedures for multiple significance testing. Stat Med 9(7):811–818.
.
OpenUrl CrossRef PubMed

[258] Hochberg Y,

[259] Benjamini Y

[260] ↵
Benjamini Y,
Hochberg Y
(1995) Controlling the false discovery rate: A practical and powerful approach to multiple testing. J R Stat Soc Series B Stat Methodol 57(1):289–300.
.
OpenUrl

[261] Benjamini Y,

[262] Hochberg Y

[263] ↵
Krzywinski M,
Altman N
(2014) Visualizing samples with box plots. Nat Methods 11(2):119–120.
.
OpenUrl CrossRef PubMed

[264] Krzywinski M,

[265] Altman N

[266] ↵
Streit M,
Gehlenborg N
(2014) Bar charts and box plots. Nat Methods 11(2):117.
.
OpenUrl CrossRef PubMed

[267] Streit M,

[268] Gehlenborg N

[269] ↵
Gao J, et al.
(2013) Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Sci Signal 6(269):pl1.
.
OpenUrl Abstract/FREE Full Text

[270] Gao J, et al.

[271] ↵
Bastien RRL, et al.
(2012) PAM50 breast cancer subtyping by RT-qPCR and concordance with standard clinical molecular markers. BMC Med Genomics 5:44.
.
OpenUrl CrossRef PubMed

[272] Bastien RRL, et al.

[273] ↵
Zhu J, et al.
(2009) The UCSC Cancer Genomics Browser. Nat Methods 6(4):239–240.
.
OpenUrl CrossRef PubMed

[274] Zhu J, et al.

[275] ↵
Goldman M, et al.
(2015) The UCSC Cancer Genomics Browser: Update 2015. Nucleic Acids Res 43(Database issue):D812–D817.
.
OpenUrl Abstract/FREE Full Text

[276] Goldman M, et al.

[277] ↵
Towns J, et al.
(2014) XSEDE: Accelerating scientific discovery. Comput Sci Eng 16(5):62–74.
.
OpenUrl CrossRef

[278] Towns J, et al.

[279] ↵
Andrews S
(2010) FastQC: A quality control tool for high throughput sequence data. Available at www.bioinformatics.babraham.ac.uk/projects/fastqc/. Accessed February 15, 2015.
.

[280] Andrews S

[281] ↵
Bolger AM,
Lohse M,
Usadel B
(2014) Trimmomatic: A flexible trimmer for Illumina sequence data. Bioinformatics 30(15):2114–2120.
.
OpenUrl Abstract/FREE Full Text

[282] Bolger AM,

[283] Lohse M,

[284] Usadel B

[285] ↵
Trapnell C,
Pachter L,
Salzberg SL
(2009) TopHat: Discovering splice junctions with RNA-Seq. Bioinformatics 25(9):1105–1111.
.
OpenUrl Abstract/FREE Full Text

[286] Trapnell C,

[287] Pachter L,

[288] Salzberg SL

[289] ↵
Li H, et al., 1000 Genome Project Data Processing Subgroup
(2009) The Sequence Alignment/Map format and SAMtools. Bioinformatics 25(16):2078–2079.
.
OpenUrl Abstract/FREE Full Text

[290] Li H, et al., 1000 Genome Project Data Processing Subgroup

[291] ↵
Anders S,
Pyl PT,
Huber W
(2015) HTSeq--a Python framework to work with high-throughput sequencing data. Bioinformatics 31(2):166–169.
.
OpenUrl Abstract/FREE Full Text

[292] Anders S,

[293] Pyl PT,

[294] Huber W

[295] ↵
Huang W,
Sherman BT,
Lempicki RA
(2009) Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 4(1):44–57.
.
OpenUrl CrossRef PubMed

[296] Huang W,

[297] Sherman BT,

[298] Lempicki RA

[299] ↵
Huang W,
Sherman BT,
Lempicki RA
(2009) Bioinformatics enrichment tools: Paths toward the comprehensive functional analysis of large gene lists. Nucleic Acids Res 37(1):1–13.
.
OpenUrl Abstract/FREE Full Text

[300] Huang W,

[301] Sherman BT,

[302] Lempicki RA

[303] ↵
Hulsen T,
de Vlieg J,
Alkema W
(2008) BioVenn - a web application for the comparison and visualization of biological lists using area-proportional Venn diagrams. BMC Genomics 9:488.
.
OpenUrl CrossRef PubMed

[304] Hulsen T,

[305] de Vlieg J,

[306] Alkema W

[307] ↵
Brown RW,
Chirala R
(1995) Utility of microwave-citrate antigen retrieval in diagnostic immunohistochemistry. Mod Pathol 8(5):515–520.
.
OpenUrl PubMed

[308] Brown RW,

[309] Chirala R

Main menu

User menu

Search

Repression of p63 and induction of EMT by mutant Ras in mammary epithelial cells

Significance

Abstract

Results

Discussion

Materials and Methods

Plasmids, Cell Culture, and Generation of Stable Cell Lines

Protein Extraction and Western Blotting

qRT-PCR

Luciferase Reporter Assays

Invasion and Soft Agar Assays

Bioinformatics/RNA-Seq

Chromatin Immunoprecipitation

Tissue Microarrays

Acknowledgments

Footnotes

References

Citation Manager Formats

Article Classifications

You May Also be Interested in

Similar Articles

Articles

PNAS Portals

Information

Main menu

User menu

Search

Significance

Abstract

Results

Discussion

Materials and Methods

Plasmids, Cell Culture, and Generation of Stable Cell Lines

Protein Extraction and Western Blotting

qRT-PCR

Luciferase Reporter Assays

Invasion and Soft Agar Assays

Bioinformatics/RNA-Seq

Chromatin Immunoprecipitation

Tissue Microarrays

Acknowledgments

Footnotes

References

Citation Manager Formats

Article Classifications

Sign up for Article Alerts

Jump to section

You May Also be Interested in

Similar Articles

Articles

PNAS Portals

Information