Stat3 and CCAAT/enhancer binding protein beta (C/EBP-beta) regulate Jab1/CSN5 expression in mammary carcinoma cells

The human breast cancer cell lines, MCF7, MDA-MB-468, MDA-MB-231, BT-474, ZR-75-1, BT-549, MDA-MB-453, T47D and non-tumorigenic human breast epithelial MCF-10A, MCF-10F, HMEC, and 184A, were purchased from the American Type Culture Collection (Manassas, VA, USA). None were derived directly from tumor tissue for the purposes of this study. MCF7 cells were grown in DMEM. Breast cancer cells were grown in RPMI medium supplemented with 10% FBS. MCF-10A and MCF-10F cells were grown in 50% DMEM, 50% F-12 medium supplemented with 5% horse serum, 100 units/ml penicillin, 100 μg/ml streptomycin, 10 μg/ml insulin, 100 ng/ml cholera toxin, 0.5 μg/ml hydrocortisone, 20 ng/ml recombinant human epidermal growth factor, and 1 mM CaCl₂. The following antibodies were used in the study: C/EBP-α (N-19), C/EBP-β (C-19), GATA-1 (H-200), and Stat3 (C-20) (Santa Cruz Biotechnology, Santa Cruz, CA, USA). The following antibodies were obtained from Cell Signaling (Danvers, MA, USA): Src, Phospho-Stat3(Y705), β-Tubulin, and β-actin. Anti-Flag was obtained from Sigma-Aldrich (St. Louis, MO, USA). IL-6 was obtained from Invitrogen (Carlsbad, CA, USA) and used at 40 ng/mL. The Stattic inhibitor (Sigma, St. Louis, MO, USA) was used at 20 nM.

An antisense primer, P1 was designed 5' of the ATG translation site of the Jab1 gene and end-labeled with T4 polynucleotide kinase and ³²P-γ-ATP, followed by purification using Nu-Clean D25 columns (Shelton Scientific-IBI, Peosta, IA, USA). Total RNA was isolated from MDA-MD-231 cells using TRIzol reagent (Invitrogen, Carlsbad, CA, USA), and 20 μg of RNA was hybridized with 1 × 10⁶ c.p.m. of ³²P-labeled P1 oligonucleotide. The protocol for the Primer Extension System-AMV Reverse Transcriptase (Promega, Madison, WI, USA) was followed. The size of the extended product was determined by electrophoresis on a 6% polyacrylamide gel containing 6 M urea. To determine the size of the primer extended fragment, the labeled Promega marker and a sequencing reaction with Jab1 cDNA as template and primer P1 using the SequiTherm Excel II DNA Sequencing Kit (Epicentre, Madison, WI, USA) were run simultaneously on the gel.

Transcription factor binding sites were predicted using Genomatix software (Munich, Germany); the MatInspector program, which uses TRANSFAC matrices; and the WebGene HCtata program's Hamming clustering method for TATA signal prediction in eukaryotic genes.

To clone the 5'-flanking region of the human Jab1 gene, a bacterial artificial chromosome clone containing the region corresponding to Jab1 was used as a template for PCR. The amplified DNA fragments were subcloned into the luciferase reporter vector pGL3 (Promega, Madison, W, USA). Progressive deletion mutants of the pGL3-Jab1 promoter were created by PCR. The integrity of constructs was confirmed by DNA sequencing. The following primers were used: +83R, -2006F, -2958F, -946F, -658F, -472F, -344F, -127F, and -59F (Table 1).

The PCR reaction contained 100 ng of DNA template, 1.5 mM MgCl₂, 0.2 mM dNTP (Roche Applied Science, Indianapolis, IN, USA), 1 uM of primers, and Taq High Fidelity DNA polymerase (Invitrogen, Carlsbad, CA, USA).

The reverse transcriptase (RT) assay was performed from 2 μg of total RNA using Superscript II RT (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's procedure. A reaction without RT was performed in parallel to ensure the absence of genomic DNA contamination. PCR amplification was carried out as described previously [22]. Conditions for the PCR reaction consisted of an initial denaturation step at 94°C for 5 minutes, followed by 30 cycles of 30 seconds at 94°C, 30 seconds at 60°C, and 30 seconds at 68°C. After a final extension at 72°C for 10 minutes, PCR products were resolved on 1.2% agarose gels and visualized by ethidium bromide transillumination under UV light. Primers used were: Jab1 F296-314, Jab1 R1094-1076, Jab1 R883-864, GAPDH F, and GAPDH R. For these and all following primer sequences please refer to Table 1.

MCF7, MDA-MB-231, and MDA-MB-468 cells were plated into 24-well tissue culture dishes at 4 × 10⁴ cells/well 24 hours before transfection. Transfections were performed in triplicate according to the manufacturer's protocol using Lipofectamine PLUS reagent (Invitrogen, Carlsbad, CA, USA). Briefly, 0.4 μg reporter plasmid Jab1-Luc (Firefly luciferase) together with 10 ng of pRL (Renilla luciferase) were cotransfected. Luciferase assays were performed 36 hours after transfection using a Dual-Luciferase Reporter Assay System (Promega, Madison, W, USA). Firefly and Renilla luciferase activities were read on a Monolight 3010 luminometer (BD Bioscience, Rockville, MD, USA). Firefly luciferase activity was normalized to Renilla luciferase readings in each well. Each experiment was conducted at least twice in triplicate.

Site-directed mutagenesis of CEBP and GATA1 was performed according to the QuickChange II method (Stratagene, La Jolla, CA, USA). The following mutagenic primers were used: CEBP-Mut, GATA1-Mut, and GATA1-Del (see Table 1 for sequences). All mutants were verified by sequencing.

Nuclear extracts were prepared as previously described [1]. Briefly, MCF7 and MDA-MB-468 cells were lysed in 10 mM HEPES-KOH (pH 7.9), 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM DTT, and protease inhibitor cocktail (Roche Applied Science, Indianapolis, IN, USA). After incubation for 10 minute on ice, nuclei were recovered by centrifugation at 3,000 × g at 4°C for one minute and resuspended in 20 mM HEPES-KOH (pH 7.9), 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, and protease inhibitor cocktail. Protein concentrations were determined using the DC Protein Assay (Bio-Rad Laboratories, Hercules, CA, USA). The following double-stranded DNA oligonucleotides were used in the electrophoretic mobility shift assays (EMSAs): -462/-436 CEBP-WT, -462/-436 CEBP-M1, -435/-417 GATA1-WT, and -435/-417 GATA1-M1 (see Table 1 for sequences).

Oligonucleotides were end-labeled with [γ-³²P] ATP (MP Biomedicals, Solon, OH, USA) by T4 polynucleotide kinase (New England Biolabs, Ipswich, MA, USA) and purified with Quick Spin G-50 Columns (Roche Applied Science, Indianapolis, IN, USA). Pre-binding of 5 μg of nuclear extract to 0.05 mg/mL poly(dI:dC) (Roche, Applied Science, Indianapolis, IN, USA) was performed in a buffer containing 20 mM HEPES (pH 7.4), 0.1 mM EDTA (pH 8.0), 75 mM KCl, 2.5 mM MgCl₂, 1 mM DTT, and 5% glycerol for 20 minutes at room temperature (22 to 24°C) before addition of 60,000 c.p.m. of labeled probe. For competition assays, 25- and 100-fold excess cold competitor oligonucleotide duplex was added to the reaction buffer 10 minutes before addition of the labeled probes. For supershift assays, antibodies were added for 20 minutes at 4°C prior to addition of the labeled probe. Reactions were resolved by electrophoresis on a 4.5% nondenaturing polyacrylamide gel run in 0.5 × TBE, vacuum-dried with heat, and exposed to film at -80°C.

The manufacturer's protocol for the chromatin immunoprecipitation (ChIP) Assay Kit (Upstate Biotechnology, Temecula, CA, USA) was followed. Briefly, MCF7 cells or MDA-MB-231 cells transfected with control pcDNA vector or C/EBP-β2 and were incubated with 1% formaldehyde for 20 minutes at 37°C. Cells were collected, lysed, sonicated, and incubated with 4 μg of antibodies to C/EBP-α, C/EBP-β, GATA-1, Stat3, or β-actin overnight. PCR was used to amplify DNA bound to the immunoprecipitated histones after reversing the histone-DNA cross-links. The following primers were used for PCR: -472F and -344R (see Table 1 for sequences).

Small interfering RNA (siRNA) for Stat3, Src, and Control (LUC) were obtained from Dharmacon (Lafayette, CO, USA). Oligonucleotides were transfected using Oligofectamine (Invitrogen, Carlsbad, CA, USA) following the manufacturer's protocol. For luciferase assay experiments, MDA-MB-468 cells were plated at 4 × 10⁴ cells per well in a 24-well tissue culture dish. siRNA (20 nmol and 50 nmol) were transfected in complete medium without antibiotics. The -472 Jab1-Luc construct and pRL were cotransfected 24 hours later using the manufacturer's protocol for Lipofectamine PLUS transfection reagent. Luciferase assays were performed after 48 hours.

To determine the transcription start site of the Jab1 gene, primer extension analysis was performed. An antisense oligonucleotide (P1 primer) was synthesized corresponding to the sequence located at the ATG translational start site according to the published sequence [GenBank accession no. NT008183]. The P1 primer was radio-labeled and extended using avian myeloblastosis virus reverse transcriptase and analyzed on a polyacrylamide-urea gel along with a DNA cycle sequencing reaction using the same primer. The primer extension experiment revealed an extension product of 71 nucleotides using the P1 primer (Figure 1a). The nucleotide 71 bp upstream of the ATG translation start site was designated the +1 transcription initiation site in the numbering of the nucleotide sequence throughout this study.

Based on the determined location of the transcription start site, we analyzed the Jab1 5' flanking region for a functional promoter. Using the MatInspector program, which uses TRANSFAC transcription factor binding site matrices, we identified a number of putative transcription factor binding sites upstream of the Jab1 transcription start site (Figure 1b). A typical TATA box [23] was found 40 bp upstream of the transcription start site, along with a CAAT box at -99 bp. A number of putative transcription factor-binding elements were present in the promoter sequence, including STAT, ELK, p53, E2F, C/EBP, GATA, c-Myb, and AP-1.

To analyze the mechanisms responsible for transcriptional regulation of Jab1, luciferase constructs were generated to evaluate Jab1 promoter activity. A fragment of the human Jab1 promoter region from -2958 to +68 was amplified, sequenced, and subcloned into the luciferase reporter vector pGL3. To identify regulatory sequences that are important for transcriptional control of the Jab1 gene, we created a series of 5'-deletion constructs of the full length Jab1 promoter and analyzed for their ability to drive the expression of luciferase in MCF7 breast cancer cells (Figure 2a). Serial 5'-deletion mutations of the full-length promoter revealed a pattern of functional activity in transfected cells. The results suggest that the core promoter elements that drive maximal promoter activity are within the sequence spanning -472 and -345 upstream of the transcription initiation site, as the -344 Jab1-Luc construct displayed a marked loss of promoter activity (Figure 2a). Similar patterns of activity were seen in other breast cancer cell lines including BT-549, MDA-MB-231, SKBR3, and BT-474 (data not shown). Additionally, we evaluated the promoter strength in mammary tumor cells versus non-tumor cells. Jab1 promoter activity was three times higher in MCF7 breast cancer cells when observing the -472 construct than in MCF-10A normal mammary epithelial cells (Figure 2b). Further analysis of a panel of breast cancer cell lines including BT-474, ZR-75-1, MCF7, BT-549, and MDA-MB-453 revealed higher Jab1 promoter activity compared to normal mammary epithelial cells MCF-10A, HMEC and 184A (Figure 2c). Jab1 protein levels were also found to be higher in the panel of breast cancer cells compared with the normal mammary epithelial cells HMEC and MCF-10A (Figure 2d). Therefore, the transcription factors responsible for promoting Jab1 transcription are either not present in normal cells or not as active as in tumor cells. Another possibility is that normal cells may express transcription suppressors acting on this region. In Figure 2b, the activity of the deletion constructs demonstrated differing ratios of luciferase activity between MCF7 and MCF-10A suggesting that different dominant regulatory factors may exist in each of these cell lines. However, interestingly, the same differential in promoter activity was seen between the -472 and -344 constructs in both cell lines.

Our data suggest that the -472 to -345 (-472/-345) region is key to driving the transcription of Jab1. We therefore sought to identify the transcription factors that regulated Jab1 through this region. Using the TRANSFAC database, we identified a number of putative transcription factor binding sites in the -472/-345 region, including sites for C/EBP (-444/CAAC/-441) and GATA-1 (-435/TATCT/-431) (Figure 1b). Because these transcription factors are activated in some cases of tumorigenesis [24, 25], we speculated that they contributed to the increased transcription of Jab1.

C/EBPs are a family of basic leucine-zipper transcription factors, including C/EBP-α, C/EBP-β, -∂, -ε, -γ, and -ζ [26]. Of these, C/EBP-α, -β, and -∂ are expressed in the mammary gland, with C/EBP-β proposed to play a role in breast cancer [27, 28]. The GATA family of transcription factors is required for erythroid and megakaryocytic differentiation [29]. Unlike other GATA members, GATA-1 has not been associated with any solid tumors, but mutations in GATA-1 are associated with essentially all cases of acute megakaryoblastic leukemia in children with Down syndrome [30].

To pinpoint the functional significance of the C/EBP and GATA-1 binding sites detailed within the Jab1 promoter we performed promoter and EMSA analysis of these regions. Cloning of this region in front of the -105/+83 sequence containing the TATA and CAAT boxes was sufficient to drive Jab1 promoter activity (Figure 3a). To determine whether these putative elements are capable of binding transcription factors, we performed a series of EMSA experiments with nuclear extracts from MCF7 breast cancer cells. The oligonucleotides and mutants for the C/EBP and GATA-1 binding sites are shown in Figure 3b. The -462/436 and -435/-417 probes showed transcription factor binding activity to the DNA containing the C/EBP and GATA-1 binding sites, respectively (Figure 3c). The cold specific oligonucleotides competed efficiently for binding, whereas a cold mutant competitor containing a mutation in the C/EBP or GATA-1 binding site did not. Further, a supershift was observed when the oligonucleotides were incubated with antibodies to C/EBP-β and GATA-1 (Figure 3c).

To assess whether any of these sites is important for expression of Jab1, each was mutated individually in luciferase reporter plasmids. We introduced mutations in the -472/-345 region of the Jab1 promoter to disrupt C/EBP and GATA-1 binding and compared their activity with the -472 Jab1-Luc promoter construct in transient transfection assays. Mutation of either C/EBP or GATA-1 binding sequence reduced Jab1 promoter activity by approximately 40% and 20%, respectively, and mutation of both sites resulted in a reduction of approximately 75% (Figure 3d). Interestingly, C/EBP and GATA binding sites homology from the human and mice promoter regions were found very well conserved [see Figure S1 in Additional file 1].

Next, we examined which of the C/EBP and GATA family members are important for Jab1 promoter activity. We cotransfected different members of the C/EBP and GATA family, C/EBP-α, C/EBP-β, or C/EBP-δ, or GATA1-6, into MCF7 cells, along with the -472 Jab1-Luc plasmid (Figure 4a). Compared with control cells transfected with vector alone, cells transfected with C/EBP-α, C/EBP-β, or GATA-1 showed the greatest increase in -472 Jab1-Luc reporter activity (2.2-, 3.0-, and 2.6-fold, respectively). GATA-2 and GATA-3 also showed a significant increase in activity but for the purpose of this study were not studied in detail.

Interestingly, the transcription factor C/EBP-β has been associated with breast cancer [27, 31]. It is translated into three different isoforms: C/EBP-β1, a 55-kDa liver-enriched activating protein (LAP1); C/EBP-β2, a 42- to 46-kDa protein also called LAP2 and; C/EBP-β3 a 20-kDa liver-enriched inhibitory protein (LIP) [25, 32]. Of the three C/EBP-β isoforms, LAP2 has been specifically observed in breast cancer, and overexpression of this isoform can induce epithelial-mesenchymal transition [28]. LIP, however, is unable to activate gene transcription, but is still able to bind to DNA and dimerize, and therefore acts as a dominant negative [27]. It has been suggested that the LAP/LIP ratio may be an important indicator of C/EBP-β transcription [25]. We found that LAP2 activates the Jab1 promoter (over two-fold), whereas LAP1 had little effect and LIP decreased activity by about 16% (Figure 4b). Further, analysis of the C/EBPβ isoforms in breast cancer cells compared with normal, revealed higher levels of the LAP-1 and -2 isoforms in MCF7, MDA-MB-468 and MDA-MB-231 breast cancer cells compared with normal mammary epithelial cells HMEC and MCF-10A (Figure 4c). Interestingly, the LIP expression was also expressed at higher levels in the breast cancer cells compared with normal (Figure 4c). This in line with the increased endogenous Jab1 expression detected in these breast cancer cells compared with the normal mammary epithelial cells as shown in Figure 2d. We also detected higher GATA-1 expression in the breast cancer cells compared with normal mammary epithelial cells and taken together could be a driving force in leading Jab1 expression in breast cancer.

To further determine whether C/EBP isoforms and GATA-1 bind to the Jab1 promoter, we performed ChIP assays on chromatin obtained from MCF7 cells (Figure 4d). When chromatin was incubated with β-actin, no product was observed. When chromatin was incubated with antibodies to C/EBP-α, C/EBP-β, and GATA-1 and the -472/-344 region was amplified by PCR, a product was observed. These data suggest that C/EBP-α, C/EBP-β, and GATA-1 bind specifically to the -472/-344 region of the Jab1 promoter.

The C/EBP transcription factors form heterodimers with other C/EBP family members and other transcription factors. We observed that Jab1-driven luciferase reporter activity increased upon addition of conditioned medium to MCF7 cells (data not shown). One of the pathways activated by conditioned medium is the Stat3 pathway, which has also been linked with breast tumorigenesis [33]. Moreover, Stat3 has been shown to interact with the C/EBP transcription factors and increase their activity [24]. We therefore investigated whether Stat3 can transactivate the Jab1 promoter, alone or in combination with C/EBP-α or C/EBP-β. Stat3 alone increased Jab1 promoter activity, but Stat3 along with C/EBP-α and C/EBP-β synergistically increased Jab1 promoter activity (Figure 5a), suggesting that Stat3 interacts with C/EBP on the Jab1 promoter. To that end, we observed a STAT general consensus site (TTN5AA) [34] that overlaps the C/EBP site within the Jab1 promoter sequence (-472 to -345) and was not identified by transcription factor database searches (Figure 5b). Stat1, Stat3, Stat4, and Stat92E all bind to this generalized consensus site [34, 35].

We next performed a ChIP assay to determine whether Stat3 could bind with C/EBP-β2 co-operatively to the Jab1 promoter using MDA-MB-231 cells, which have constitutively activated Stat3. Since C/EBP-β2 (or LAP2) seems to be a major activator of the Jab1 promoter (Figure 4), we next focused only on C/EBP-β2. Binding of Stat3 to the Jab1 promoter was increased greater than seven-fold when C/EBP-β2 was transfected into the cells (Figure 5c). Taken together, these data suggest that C/EBP-β2 and Stat3 bind to the Jab1 promoter to increase Jab1 promoter activities.

We further investigated whether inhibition of Stat3 affects Jab1 promoter activity in MDA-MB-468 cells, which have constitutively activated Stat3. Expression of a dominant-negative mutant of Stat3 reduced over 80% Jab1 promoter activity (Figure 5d). The EEE/VV mutation renders the protein incapable of DNA binding [36]. Expression of exogenous wild type Stat3 increased Jab1 expression, whereas the dominant-negative EEE/VV mutation reduced Jab1 protein levels (Figure 5e).

To further demonstrate the regulation of Jab1 by Stat3, we examined the effect of inhibition of Stat3 and also its upstream activator Src using siRNA [37, 38]. Inhibition of Stat3 and Src by siRNA resulted in a dramatic reduction of Jab1 promoter activity (Figure 6a), mRNA levels (Figures 6b and 6c), and Jab1 protein expression (Figure 6d). Similar results were obtained with MDA-MB-231 cells (data not shown). Collectively, these results demonstrate that the Src/Stat3 signaling pathway plays an important role in the regulation of Jab1 transcription.

To determine the biological significance of Stat3-mediated Jab1 expression, the role of Stat3 in normal human breast epithelial cells was investigated. As Jab1 expression in normal mammary epithelial cells is low, we asked whether overexpression of Stat3 could enhance Jab1 transcription in these cells. Ectopic expression of Stat3 in normal mammary epithelial cells (MCF-10A and MCF-10F) resulted in increased Jab1 mRNA and protein levels (Figures 7a and 7b). Therefore, the Stat3 transcription factor is in part responsible for promoting Jab1 transcription in breast cancer cells and is either not present or not as active in the MCF-10A and MCF-10F cells.

We further investigated whether an upstream activator of Stat3, the cytokine IL-6, could be driving increased Jab1 expression. Treatment with IL-6 for 30 minutes increased phosphorylation of Stat3 on tyrosine 705 and resulted in increased Jab1 mRNA (Figure 7c) and protein levels (Figure 7d) within the short time of 30 minutes, 1 hour, and 4 hours that was partially blocked by the addition of the Stat3 inhibitor Stattic. The same trends were observed in the breast cancer cell line T47D and the mammary epithelial cells, MCF-10A. IL-6 also resulted in increased Jab1 promoter activity in MCF7 and T47D cells [See Figure S2 in Additional file 1].

Taken together, it is evident that both IL-6 and Src signaling through Stat3 is contributing to Jab1 transcription and increased expression in breast cancer. Further, it is possible that Stat3 and C/EBPβ could be binding to the Jab1 promoter either separately or together to mediate increased Jab1 transcriptional activity. A proposed model for the activation of Jab1 transcription is shown in Figure 7e.