NOTCH-1 and NOTCH-4 are novel gene targets of PEA3 in breast cancer: novel therapeutic implications

MDA-MB-231, SKBr3, BT474 and MCF-7 breast cancer cells were purchased from the American Type Culture Collection (Manassas, VA, USA). All cell lines were supplemented with 100 μmol nonessential amino acids and 1% L-glutamine. SKBr3 cells (supplemented with 10% fetal bovine serum (FBS)) and MDA-MB-231 cells (supplemented with 5% FBS) were maintained in Iscove's Minimal Essential Medium (IMEM). MCF-7 cells (supplemented with 10% FBS) were maintained in DMEM/Ham's Nutrient Mixture F-12. BT474 cells (supplemented with 10% FBS) were maintained in DMEM. All cells were cultured in a 37°C incubator with 5% CO₂. MRK-003 GSI was kindly provided by Merck Oncology International, Inc. (33 Avenue Louis Pasteur, Boston, MA, USA). MRK-003 GSI was dissolved in dimethylsulfoxide (DMSO) and stored at -80°C until use. Lactacystin (L6785) was purchased from Sigma-Aldrich (St Louis, MO, USA), dissolved in deionized water and stored at -20°C until use. The pcDNA3.1 expression vector and the PEA3-pcDNA3.1 expression vector were kindly provided by Dr Mein-Chie Hung (The University of Texas MD Anderson Cancer Center, Houston, TX, USA). The pHMB empty and pHMB-TAM-67 expression vectors were kindly provided by Dr Richard Schultz, Department of Immunology and Microbiology, (Loyola University Medical Center, Maywood, IL, USA).

Control scrambled siRNA-a (sc-37007), Notch-1 siRNA (sc-36095), and a smart pool of three distinct PEA3 siRNA (PEA3ia, sc-36205) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). An unrelated control siRNA and PEA3 siRNA (PEA3ib, catalog number 115237) were purchased from Ambion (Austin, TX, USA). A smart pool of four distinct c-Jun siRNA (sc-29223) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Transfection reagents used were Lipofectamine 2000 and Lipofectamine RNAiMAX, which were purchased from Invitrogen (Carlsbad, CA, USA), and FuGENE 6 was purchased from Roche Diagnostics Corporation (Indianapolis, IN, USA). Protocols were performed as described by the manufacturers.

Notch-1 (antibody clone C-20), Notch-4 (antibody clone H-225), PEA3 [16] (product number sc-113), c-JUN (antibody clone G-4) were purchased from Santa Cruz Biotechnology. β-actin (antibody clone AC-15) was purchased from Sigma-Aldrich and used as the loading control

Total RNA was extracted from the cells using the RNeasy Mini Kit (Qiagen, Valencia, CA, USA) according to the manufacturer's protocol. Total cDNA was reverse-transcribed from the total RNA with random hexamers using the MultiScribe™ Reverse Transcriptase Kit (Applied Biosystems, Foster City, CA, USA) according to the manufacturer's protocol. Analysis of transcript relative fold copy number adjusted to hypoxanthine-guanine phosphoribosyl transferase (HPRT), an endogenous control, was carried out by quantitative real-time PCR using iTaq™ SYBR Green Supermix with ROX (Bio-Rad Laboratories, Hercules, CA, USA) and the PlusOne thermal cycler from Applied Biosystems. The following primers were used for detection: Notch-1 (forward primer: 5'-GTCAACGCCGTAGATGACC-3', reverse primer: 5'-TTGTTAGCCCCGTTCTTCAG-3'), Notch-2 (forward primer: 5'-TCCACTTCATACTCACAGTTGA-3', reverse primer: 5'-TGGTTCAGAGAAAACATACA-3'), Notch-3 (forward primer: 5'-GGGAAAAAGGCAATAGGC-3', reverse primer: 5'-GGAGGGAGAAGCCAAGTC-3'), Notch-4 (forward primer: 5'-AACTCCTCCCCAGGAATCTG-3', reverse primer: 5'-CCTCCATCCAGCAGAGGTT-3'), PEA3 (forward primer: 5'-AGGAGACGTGGCTCGCTGA-3'), (reverse primer: 5'-GGGGCTGTGGAAAGCTAGGTT-3'), HEY-1 (forward primer: 5'-TGGATCACCTGAAAATGCTG-3', reverse primer: 5'-TTGTTGAGATGCGAAACCAG-3'), MMP-9 (forward primer: 5'-TCGTGGTTCCAACTCGGTTT-3', reverse primer: 5'-GCGGCCCTCGAAGATGA-3') and IL-8 (forward primer: 5'-CACCGGAAGGAACCATCTCACT-3', reverse primer: 5'-TCAGCCCTCTTCAAAAACTTCTCC-3'), with HPRT (forward primer: 5'-ATGAACCAGGTTATGACCTTGAT-3', reverse primer: 5'-CCTGTTGACTGGTCATTACAATA-3') used as the loading control. Protocols were performed as described by the manufacturers. Transfection conditions for Western blot analysis and RT-PCR were as follows. MDA-MB-231 (4 × 10⁵), MCF-7 (4 × 10⁵), SKBr3 (5 × 10⁵) and BT474 (5 × 10⁵) cells were placed into a six-well culture dish. Twenty-four hours later the cells were transfected with reagents. Forty-eight hours later cells were harvested and protein/RNA was extracted. Notch-4 luciferase constructs were generously provided by Dr Emery Bresnick, Cell & Regenerative Biology, (University of Wisconsin at Madison, Madison, WI, USA). The AP-1 luciferase construct was generously provided by Dr Richard Schultz, Department of Immunology and Microbiology,(Loyola University Medical Center, Maywood, IL, USA). Dual-luciferase assays were performed as described by the manufacturer (Promega, Madison, WI, USA). A pRL-thymidine kinase promoter-driven Renilla luciferase reporter was cotransfected with the Firefly luciferase construct mentioned above as an internal transfection control. Transfection activity was measured using the Veritas Microplate Luminometer (Turner Biosystems, Sunnyvale, CA, USA) and represented as the ratio of Firefly luciferase to Renilla luciferase.

The cells were lysed in radioimmunoprecipitation assay buffer (pH 8.0) containing 50 mmol Tris HCl, 150 mmol NaCl, 1% Nonidet P-40 (NP-40), 0.5% sodium deoxycholate, 0.1% SDS, 25 mmol β-glycerophosphate, 1 mmol sodium orthovanadate, 1 mmol sodium fluoride, 1 mmol phenylmethylsulfonyl fluoride, 1 mg/mL aprotinin and 1 mg/mL leupeptin. Western blot analysis was performed as previously described [21]. NuPAGE Bis-Tris Gels (4% to 12%; Invitrogen) in 3-(N-morpholino)propanesulfonic acid buffer were run at 175 V for 1.5 hours, and proteins were transferred at 38 V for 2 hours using polyvinylidene fluoride membranes. Protein detection was performed using the SuperSignal West Dura Substrate (Thermo Scientific, Rockford, IL, USA) and visualized by using the FUJIFILM™ Las-300 imager (GE-Healthcare Bio-Sciences, Piscataway, NJ, USA).

MDA-MB-231 cells (3 × 10⁶) were plated in 150-cm² Petri dishes. Twenty-four hours later cells were transfected with pcDNA3.1, PEA3-pcDNA3.1 and PEA3-pcDNA3.1, together with PEA3 siRNA, for 48 hours. The cells were cross-linked with 1% formaldehyde and lysed in SDS lysis buffer (1% SDS, 10 mmol ethylenediaminetetraacetic acid (EDTA), 50 mmol Tris HCl, pH 8.1). The lysates were sonicated using the Branson Sonifier model 250 (Branson Ultrasound Corp., Danbury, CT, USA) at output 4.5, duty cycle 50, and pulsed 10 times. The lysate was then diluted 1:10 in immunoprecipitation dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mmol EDTA, 16.7 mmol Tris HCl, 167 mmol NaCl, pH 8.1). Approximately 300 to 700 μg of total precleared protein in lysates were incubated with 4 μg of PEA3 antibody (sc-113X; Santa Cruz Biotechnology) or mouse immunoglobulin G (IgG) overnight. Fifty microliters of protein G-plus agarose beads (sc-2002; Santa Cruz Biotechnology) were added to the immune complexes for 2 hours while gently rocking. Immune complexes/beads were washed in low-salt buffer (0.1% SDS, 1% Triton X-100, 2 mmol EDTA, 20 mmol Tris HCl, 150 mmol NaCl, pH 8.1), high-salt buffer (0.1% SDS, 1% Triton X-100, 2 mmol EDTA, 20 mmol Tris HCl, 500 mmol NaCl, pH 8.1), LiCl buffer (1% sodium deoxycholate, 1% NP-40, 0.25 mol LiCl, 1 mmol EDTA, 10 mmol Tris HCl, pH 8.1) and Tris-EDTA buffer (1 mmol EDTA and 10 mmol Tris HCl) at 4°C with agitation. The protein/antibody complexes from beads were eluted in freshly prepared elution buffer (1% SDS, 0.1 mol sodium bicarbonate, pH 8.0). Cross-linking of proteins and DNA was reversed by heating at 65°C overnight while gently rocking. The protein was degraded using a proteinase solution (0.5 mol EDTA, 1 M Tris HCl, pH 8.0, 10 mg/mL proteinase K) and incubated at 52°C for 1 hour. DNA was isolated using the QIAquick PCR purification kit (Qiagen). Enrichment at promoter sites was detected by quantitative PCR using iTaq™ SYBR Green Supermix with ROX (Bio-Rad Laboratories). C_t threshold values were normalized to the IgG control. The following promoter region primers were used in detection: Notch-1 promoter regions not containing ETS or AP-1 sequences served as negative controls (forward primer: 5'-GTGCACACGGCTGTCCG-3', reverse primer: 5'-GCGACAACTGGCGACTGAA-3'), 2× PEA3 (forward primer: 5'-GCTGCAAGAGCCAAGATGAA-3', reverse primer: 5'-GGTGCCTGTGTTGAAAGCTCT-3'), c-ETS (forward primer: 5'-CTCCTGGCGCTTAACCAGG-3', reverse primer: 5'-CCAGAAAGCACAAACGGGTC-3'), AP-1(A) (forward primer: 5'-GCCTCCTTAGCTCACCCTGA-3'), reverse primer (5'-TCTTCAGAGGCCCCCTGC-3'), AP-1(B) (forward primer: 5'-TCCGCAAACCAGGCTCTG-3', reverse primer: 5'-ATTGGGGTGCAGTGCCG-3'), Sp-1/PEA3 (forward primer: 5'-CACCTCGACTCTGAGCCTCAC-3', reverse primer: 5'-CTCTTCCCCGGCTGGCT-3'). Notch-4 promoter regions not containing ETS or AP-1 sequences served as negative controls (forward primer: 5'-GGCGAGGATTCTAATGTGGA-3', reverse primer: 5'-CCCTGA GTGAAAGGGTGAAG-3'), CBF-1 (forward primer: 5'-TGGTACCACCCTGGTCAGTAT-3', reverse primer: 5'-TGCTCAGGCATTATGAGCTATG-3'), AP-1(A) (forward primer: 5'-AGCTGCCACTGACACCTTCT-3', reverse primer: 5'-CAGGGAATGCCAGTCAGAAT-3'), AP-1(B) (forward primer: 5'-ACTTCCCCAGGGGTTGTC-3', reverse primer: 5'-CTTCCTCCTCGGCCTGCT-3') and PEA3/ETS (forward primer: 5'-GGGTTCCTCTTCCCCATACC-3', reverse primer: 5'-TCATTTTCCCATCACCTTCCTT-3').

MDA-MB-231 cells (3 × 10⁶) were plated in 150-cm² Petri dishes for 24 hours. The cells were transfected with pcDNA3.1, PEA3-pcDNA3.1 and PEA3-pcDNA3.1, together with PEA3 siRNA, for 48 hours. The cells were cross-linked with 1% formaldehyde and lysed in SDS lysis buffer (1% SDS, 10 mmol EDTA, 50 mmol Tris HCl, pH 8.1). The lysates were sonicated using the Branson Sonifier 250 at output 4.5, duty cycle 50, and pulsed 10 times. The lysate concentration was ascertained and was equally diluted in immunoprecipitation dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mmol EDTA, 16.7 mmol Tris HCl, 167 mmol NaCl, pH 8.1). Approximately 300 to 700 μg of total precleared lysates were incubated with 3 μg of PEA3 antibody (sc-113X; Santa Cruz Biotechnology) or mouse IgG overnight. Thirty microliters of protein G-plus agarose beads (sc-2002, Santa Cruz Biotechnology) were added to the immune complexes for 2 hours while gently rocking. Immune complexes and beads were washed three times in PBS. The pellet was resuspended in Laemmli sample buffer with 5% β-mercaptoethanol (Bio-Rad Laboratories) and heated for 5 minutes at 95°C while vigorously shaking. Western blot analysis was used to detect coimmunoprecipitation proteins as described above. Antibodies used for detection were PEA3 (16), c-FOS (H-125), c-JUN (G-4) and Fra-1 (R-20).

MDA-MB-231 cells (1 × 10⁵) were seeded into a six-well plate. After 24 hours, the cells were treated/transfected with DMSO/scrambled, control siRNA, DMSO/PEA3 siRNA, MRK-003 (10 μmol)/scrambled siRNA or MRK-003 (10 μmol)/PEA3 siRNA. Briefly, 24 and 48 hours posttreatment/posttransfection, the cells and media were isolated and permeabilized with 70% ethanol. The cells were then pelleted, washed in 5% bovine calf serum and resuspended in 10 μg/mL RNase/PBS solution. The cells were stained with propidium iodide (100 μg/mL) and analyzed by flow cytometry (FACSCanto; BD Biosciences, San Jose, CA, USA).

MDA-MB-231 cells (1.5 × 10⁵) were seeded into a six-well plate for 24 hours. Cells were treated/transfected with DMSO/scrambled siRNA, DMSO/PEA3 siRNA, MRK-003 (10 μmol)/scrambled siRNA and MRK-003 (10 μmol)/PEA3 siRNA. Briefly, after 24 and 48 hours of treatment/transfection, 1 × 10⁶ cells were isolated in 1× annexin binding buffer (0.1 mol 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid/NaOH, 140 mol NaCl, 25 mmol CaCl₂, pH 7.4) and treated with both fluorescein isothiocyanate-annexin V stain and propidium iodide. After 1-hour incubation, the samples were subjected to analysis by flow cytometry (FACSCanto; BD Biosciences).

MDA-MB-231 cells (1.5 × 10⁵) were seeded into a six-well plate for 24 hours. Cells were treated/transfected with DMSO/scrambled siRNA, DMSO/PEA3 siRNA, MRK-003 (10 μmol)/scrambled siRNA and MRK-003 (10 μmol)/PEA3 siRNA. Briefly, after 24 and 48 hours of treatment/transfection, cells and media were harvested and washed in PBS. The cells were stained with trypan blue and counted under a standard slide microscope at ×40 magnification.

In a six-well plate, a bottom layer of 1 mL of 0.8% agar dissolved in IMEM was added. Two milliliters of a 2% methylcellulose/IMEM solution (supplemented with 5% FBS, 1% nonessential amino acids and 1% L-glutamine) were added on top of the agar. MDA-MB-231 cells (1 × 10³ cells/mL) were treated/transfected with DMSO/scrambled siRNA, DMSO/PEA3 siRNA, MRK-003 (5 μmol)/scrambled siRNA or MRK-003 (5 μmol)/PEA3 siRNA for 24 hours. Treated cells were added directly to the methylcellulose solution. The assay was left untouched for 14 days at 37°C and 5% CO₂. Colonies were stained with crystal violet for 1 hour, photographed and counted under a standard light microscope at ×40 magnification. Nine fields per well were counted.

To investigate the role of PEA3 in Notch expression, we examined the effect of two independent PEA3 RNA interference sequences on the expression of Notch-1 through Notch-4 receptor mRNA as measured by real-time PCR. Notch-1 and Notch-4 transcripts were decreased by nearly 50% and 70%, respectively, upon PEA3 knockdown by either PEA3ia or PEA3ib siRNA. Notch-3 transcript levels remained unchanged. Interestingly, Notch-2 levels showed a moderate but significant increase upon PEA3 knockdown, which, upon further investigation, may prove to be advantageous, since it has been correlated with proapoptotic and breast tumor suppressive function [30] (Figure 1A, lower left graphs). PEA3 transcripts were measured as a control for the efficacy of PEA3 knockdown using the two types of PEA3 siRNA (Figure 1A, upper left graph).

To determine whether the effect on Notch-1 and Notch-4 transcripts correlated with protein expression, protein lysates from cells transfected with either a scrambled, control siRNA (SCRBia) or PEA3 siRNA (PEA3ia) for 48 hours were subjected to Western blot analysis. The results showed a reduction of Notch-1 full-length and transmembrane receptor protein levels by 50% (Figure 1B, upper right panel). To visualize the effects of PEA3 siRNAa on the rapidly turned over Notch-4 receptor, we treated MDA-MB-231 cells with or without lactacystin, a specific proteasome inhibitor, for 24 hours before lysing cells for total proteins. Consequently, N4IC protein levels were decreased by 90% upon PEA3 knockdown (Figure 1B, upper right panel).

To confirm the effect of PEA3 siRNAa on the levels and activity of PEA3, PEA3 and its classical downstream targets were measured by real-time PCR after transfection with control or PEA3 siRNA for 48 hours. PEA3 transcripts were significantly reduced (Figure 1A, left graph) as well as its known target gene MMP-9 (Figure 1C). Moreover, HEY-1, a classic downstream target of the Notch signaling pathway, was also significantly reduced upon PEA3 knockdown. Taken together, these results suggest that PEA3 positively regulates Notch-1 and Notch-4 receptor expression and affects Notch activity in MDA-MB-231 cells.

Our investigations are the first to show that PEA3 is a transcriptional activator of Notch-1 and Notch-4 in MDA-MB-231 breast cancer cells, an example of triple-negative breast cancer. We next assessed whether this is a common mechanism among the other subtypes of breast cancer cells. We used the single pool of PEA3 siRNAa because similar results were obtained for the two independent siRNA. MDA-MB-231 (triple-negative), MCF-7 (luminal A), BT474 (luminal B) and SKBr3 (HER2⁺) cells were transfected with either scrambled or PEA3 siRNAa. Normalized to MDA-MB-231 cells, MCF-7 cells had slightly less PEA3 transcripts, followed by BT474 cells and then SKBr3 cells (Figure 2A). Notch-1 (Figure 2B) and Notch-4 (Figure 2C) transcript levels were compared with either scrambled or PEA3 siRNA among the cell lines. PEA3 regulation of Notch-1 was seen in MDA-MB-231, SKBr3 and MCF-7 cells, but not BT474 cells (Figure 2B). However, PEA3-mediated effects on Notch-4 transcripts were observed only in MDA-MB-231 cells (Figure 2C). This could be due in part to the fact that PEA3 basal transcript levels were very low in other cell types (Figure 2A). Motivated by this notion, we asked whether we could drive the expression of Notch-4 by increasing levels of PEA3 in breast cancer cells that express low endogenous PEA3 levels. To address this question, BT474 and SKBr3 cells were transfected with either pcDNA3.1 vector alone or PEA3-pcDNA3.1 expression plasmid. Notch-1 transcripts remained unchanged when PEA3 was overexpressed (Figures 2D and 2E). Notch-4 transcript levels were significantly increased ninefold in BT474 (Figure 2D) and sixteenfold in SKBr3 breast cancer cells (Figure 2E).

Taken together, PEA3 is required for Notch-1 transcription in at least three subtypes of breast cancer cells: MDA-MB-231, SKBr3 and MCF-7. However, PEA3 is an activator of Notch-4 transcription in MDA-MB-231 cells where endogenous PEA3 levels are high. Exogenous expression of PEA3 in BT474 and SKBr3 cells provides evidence that it is also a potential activator of Notch-4 transcription in other breast cancer subtypes.

To determine the mechanism by which PEA3 regulates Notch-1 and Notch-4 transcription, we scanned the promoter regions of Notch-1 and Notch-4 using the National Center for Biotechnology Information Entrez GenBank [59, 60] and identified several ETS binding sites (Figures 3B and 3C, respectively). Interestingly, both promoters contained two AP-1 sites adjacent to ETS sites. A classic CBF-1 site was also previously identified in the Notch-4 promoter [23] and was used as a negative control for PEA3 recruitment. Primers were subsequently designed to include these identified regions. Primers flanking promoter regions containing no known consensus sites were used as negative controls. We then performed chromatin immunoprecipitation (ChIP) experiments on lysates from MDA-MB-231 cells that were transfected with either pcDNA3.1 vector alone or pcDNA3.1-PEA3 expression plasmid.

To confirm that PEA3 was overexpressed in MDA-MB-231 cells, immunoprecipitation (IP) followed by Western blot analysis was performed to detect PEA3 protein (Figure 3A). The IP followed by Western blot analysis showed that endogenous PEA3 could not be detected in pcDNA3.1-transfected MDA-MB-231 cells. However, cells transfected with the PEA3 expression plasmid expressed PEA3 protein (Figure 3A). The ChIP analysis was performed on cells transfected with vector alone or with PEA3 expression plasmid (Figures 3B and 3C). Endogenous PEA3 was not enriched in either the Notch-1 or Notch-4 promoter regions in cells transfected with pcDNA3.1. This could be due to weak antibody avidity or, more likely, could be a result of PEA3 protein instability [61]. PEA3 enrichment was found in both Notch-1 and Notch-4 promoter regions within 1 kb of the start site when PEA3 was overexpressed compared to vector alone-transfected cells (Figures 3B and 3C). A PEA3 siRNA smart pool was used to confirm the specificity of PEA3 recruitment and the PEA3 antibody. Figure 4A shows that PEA3 siRNA significantly decreased the enrichment of PEA3 in the identified promoter regions of Notch-1 and Notch-4 compared to a control siRNA. The Western blot shown in Figure 4B confirms that the overexpressed PEA3 in MDA-MB-231 cells was almost completely knocked down using the PEA3 siRNAa smart pool.

To address whether PEA3 acts with AP-1 in Notch-1 and Notch-4 promoter regions, since it is known to work in accordance with c-Jun on several other genes [34, 35, 50], transfected lysates with either phMB-vector or phMB-TAM-67, a dominant-negative form of c-Jun missing the transactivation domain [62], were subjected to ChIP analysis. PEA3 enrichment on the Notch-1 promoter was not affected by TAM-67 (Figure 5A). However, PEA3 recruitment to an ETS site adjacent to an AP-1 site approximately 70 nt upstream from the start site of the Notch-4 promoter was significantly decreased by expression of TAM-67 (Figure 5B), indicating that the transactivation domain of c-Jun is required for PEA3 recruitment to the Notch-4 promoter. To confirm that TAM-67 reduces AP-1 transcriptional activity, an AP-1 luciferase reporter assay was performed in MDA-MB-231 cells transfected with either phMB vector alone or phMB-TAM-67. The results showed that the TAM-67-transfected cells contained more than 50% less AP-1 reporter activity than vector alone-transfected cells (Figure 5). To determine whether c-Jun specifically is required for PEA3 enrichment in the identified Notch-4 promoter region, transfected lysates with either control or c-Jun siRNA smart pool were subjected to ChIP analysis. In agreement with our previous findings, PEA3 enrichment in the Notch-4 promoter region was significantly decreased upon knockdown of c-Jun (Figure 5B), suggesting that c-Jun is required for PEA3 enrichment in the identified Notch-4 promoter region. We confirmed that c-Jun protein was knocked down by siRNA on the basis of Western blot analysis (Figure 5B). We also confirmed that c-Jun protein was in a complex with PEA3 in MDA-MB-231 nuclear extract by performing a co-IP assay (Figure 5C).

To assess whether the results of the ChIP studies were not due to artificial interactions as a result of PEA3 overexpression, we performed a Notch-4 luciferase reporter assay using a minimal promoter region (-650 to +1) which contains the same AP-1/ETS region just -70 nt upstream from the start site. The wild-type construct contained an intact AP-1 and ETS consensus site at -70 nt. The mutant AP-1 construct contained an ablated AP-1 consensus site, but the adjacent ETS binding site was preserved [32]. Knockdown of endogenous PEA3 using siRNA significantly decreased wild-type AP-1-containing reporter activity (Figure 6, right graph) or mutant AP-1-containing reporter activity (Figure 6, inset) compared to the scrambled control. These results suggest that both PEA3 andAP-1 are critical for the regulation of the -70 nt region within the Notch-4 promoter in MBA-MD-231 cells. As a control study, we observed no significant difference in AP-1 luciferase reporter activity upon PEA3 knockdown, indicating that PEA3 was not mediating effects on the Notch-4 promoter indirectly through AP-1 regulation (Figure 6, left graph). Herein we provide the first evidence that both PEA3 and c-Jun are required to activate the Notch-4 promoter.

To determine which members of the Fos family of transcription factors are required for Notch-4 transcription, we performed real-time PCR to detect Notch-4 transcripts in response to knockdown of Fra-1 and c-FOS, which are the major forms of Fos in MDA-MB-231 cells. The results showed that Notch-4 transcripts were decreased 2.5-fold when Fra-1 was knocked down similarly to PEA3 or c-JUN knockdown (Figure 7A, upper graph). Conversely, c-FOS siRNA increased Notch-4 transcripts twofold compared to a scrambled control siRNA (Figure 7A, lower graph). Figure 7B demonstrates the efficacy of knockdown by PEA3, c-Jun, Fra-1 and c-FOS siRNA as measured by real-time PCR and Western blot analysis. These results, together with those of the previous ChIP studies, suggest that PEA3 regulates Notch-4 transcription by potentially interacting with c-Jun and Fra-1. The results also indicate that c-FOS is a potential transcriptional repressor of Notch-4.

Notch is vital for proliferation and cell fate determination [63], whereas PEA3 is critical for cell migration and invasion [64, 65]. Both aberrant and unregulated activities can lead to overall malignancy and poor survival [25, 40]. To understand their biological significance and their effect in MDA-MB-231 triple-negative cells, we explored dual inhibition of PEA3 and Notch using the specific PEA3 siRNA smart pool and the preclinical MRK-003 GSI, respectively. MDA-MB-231 cells were transfected with scrambled control or PEA3 siRNA alone, treated with vehicle or MRK-003 GSI, or a combination of PEA3 siRNA plus MRK-003 GSI thereof for 48 hours. Independently, PEA3 knockdown or Notch inhibition had little effect on the cell cycle compared to control (Figure 8A). PEA3 knockdown or MRK-003 GSI treatment of MDA-MB-231 cells alone showed a modest but not significant increase in G₁ phase arrest, a modest decrease in the S phase and little effect on the G₂/M phase (Figure 8A). However, the combination of PEA3 knockdown and MRK-003 GSI treatment resulted in a significant increase in the G₁ phase compared to control (P < 0.05) (Figure 8A). In the presence of both PEA3 knockdown and MRK-003 GSI treatment, there was a significant reduction in the S phase of the cell cycle compared to control (P < 0.05). Also, both PEA3 knockdown and MRK-003 GSI treatment resulted in a significant reduction in the G₂/M phase of the cell cycle compared to control (P < 0.05) (Figure 8A). These results indicate that both PEA3 and Notch are critical for the proliferation of MDA-MB-231 cells.

We then asked whether dual inhibition using both PEA3 knockdown and Notch inhibition affect cell viability, potentially through increased apoptosis. Independently, PEA3 knockdown showed no change in apoptosis as measured by (1) annexin V staining (Figure 8B) or (2) cell viability using trypan blue exclusion (Figure 8C). MRK-003 GSI treatment alone increased apoptosis to 30% (Figure 8B) and decreased cell viability to 60% (Figure 8C). Importantly, the combination of PEA3 knockdown and MRK-003 GSI treatment significantly increased apoptosis to almost 40% as measured by positive staining of annexin V cells (Figure 8B) and diminished cell viability to 50% (Figure 8C). These results, taken together with the cell cycle data, indicate that both PEA3 and Notch activities are critical for cell proliferation and survival in MDA-MB-231 breast cancer cells.

To address whether dual inhibition of PEA3 and Notch activities affect anchorage-independent growth as an in vitro measure of tumorigenicity, we performed a colony formation assay with MDA-MB-231 cells that were transfected with either scrambled control or PEA3 siRNA and treated with vehicle or MRK-003 GSI independently or in combination. We observed a reduction of almost 55% in the number of colonies when PEA3 was knocked down or 61% when Notch was inhibited using a GSI in cells compared to vehicle + control siRNA control (Figures 9A and 9B). Furthermore, the number of colonies was reduced further to 20% upon PEA3 knockdown and GSI treatment compared to vehicle + control siRNA (Figures 9A and 9B).

The results so far have indicated that dual inhibition of PEA3 and Notch signaling inhibits both anchorage-dependent and anchorage-independent cell growth in vitro more effectively than either treatment alone. The results from the anchorage-independent assay suggest that transient knockdown of PEA3 using siRNA is sufficient to inhibit the formation of colonies for up to two weeks. On the basis of these results, we asked whether transient knockdown of PEA3 could inhibit tumor formation in vivo and whether treatment with MRK-003 GSI in combination could prevent tumor formation. We performed an orthotopic MDA-MB-231 tumor xenograft study in athymic mice. MDA-MB-231 cells were transfected with control or PEA3 siRNA in vitro (Figure 9D) and then injected into the mammary fat pads of female athymic mice. The mice were randomized to injection with vehicle control or MRK-003 GSI. The results showed that either PEA3 knockdown or GSI treatment significantly reduced tumor growth by almost similar rates compared to vehicle + scrambled control siRNA (Figure 9C).

We have shown for the first time, to our knowledge, that PEA3 is a novel activator of Notch-1 and Notch-4 transcription in different subtypes of breast cancer cells and could prove to be an important therapeutic target, possibly upstream of Notch-1 and/or Notch-4 signaling.