Background
The ERBB2 (Her2/Neu) gene encodes a tyrosine kinase receptor whose abnormal activity is linked to oncogenesis in breast cancer. In fact, ERBB2 gene amplification is found in 20−30% of primary breast tumors, and it is usually associated with poor clinical prognosis. In these tumors, ErbB2 receptor overexpression activates several intracellular signalling pathways, such as the Ras/Erk and PI3K/AKT pathways [
1], whose effects on c-MYC oncogene transcription and Myc protein stability have been demonstrated [
2]. The treatment of ERBB2-amplified breast tumor cells with the ErbB2-specific antibody trastuzumab causes cell cycle arrest accompanied by a decrease in PI3K/Akt activity and the downregulation of c-MYC and D-type cyclins; on the other hand, ectopic expression of c-MYC in ERBB2-overexpressing SKBr3 cells partially rescues the cells from functional ERBB2 inactivation [
3,
4]. Several studies have reinforced the significance of c-MYC as an ERBB2 effector and the functional role that the two genes play in breast cancer progression (for a review, see [
5]).
The c-MYC gene is regulated at multiple levels. One of the regulators, the Myc promoter-binding protein-1 (MBP-1), was originally identified in HeLa cells as a transcriptional repressor which binds to the human c-MYC P2 promoter, negatively affecting transcription. This factor competes for the TATA-binding protein (TBP) and prevents the formation of the transcription initiation complex [
6,
7]. MBP-1 is a short form of the 48 kDa alpha-enolase protein, lacking the first 96 amino acid. Several studies support the existence of a single ENO1 gene transcript from which both alpha-enolase and MBP-1 arise through the use of alternative translation initiation sites [
8,
9]. More recently, it has been reported that a shorter variant transcript, originating from intron III of the ENO1 gene, may contribute to MBP-1 expression in a variety of normal tissues and cancer cells [
10]. Exogenous MBP-1 expression inhibits the growth of breast tumors in nude mice [
11], induces cell death in neuroblastoma cells [
12], suppresses proliferation in non-small-cell lung cancer cells [
13], and induces G0–G1 growth arrest in chronic myeloid leukemia cells [
14]. Moreover, a role for MBP-1 in tumor invasion and metastasis has been proposed for follicular thyroid carcinoma and gastric cancer [
15,
16]. MBP-1 may exert its function as a single factor, in concert with other factors, or through physical interaction with its identified cellular partners: MIP-2/sedlin [
17], histone deacetylase 1 (HDAC1) [
18], the kelch protein NS1-BP [
19], and the Notch 1 receptor intracellular domain [
20]. Besides c-MYC, only one other direct target of MBP-1, the COX2 gene, has been identified so far [
16].
Consistent with its negative regulatory role on cell growth, the endogenous level of MBP-1 in tumor cells is low; in MCF-7 breast cancer cells, glucose concentration and hypoxia have been reported to modulate MBP-1 expression and its binding to the c-MYC promoter, consequently affecting cell proliferation [
21,
22]. Thus, MBP-1 appears to be one of the factors controlling cell growth and proliferation, and alterations in its expression level induced by the tumor microenvironment may contribute to cancer development.
Our previous studies have indicated that MBP-1 is expressed and easily detectable in normal breast epithelial cells, but a loss of expression occurs in most primary invasive ductal carcinomas (IDC) of the breast. Furthermore, MBP-1 expression inversely correlates with expression levels of the ErbB2 and Ki67 proteins [
23]. On the basis of these observations, we hypothesized a direct functional link between MBP-1 and the ERBB2 gene in human breast carcinomas.
In the present study, we provide evidence that MBP-1 inhibits the expression of the ERBB2 gene in SKBr3 breast cancer cells by interacting with the promoter region. In addition, we show that HDAC1 is recruited to the same region of the ERBB2 promoter which is bound by MBP-1. Finally, we report a significant correlation between MBP-1, HDAC1 and ERBB2 protein expression in primary breast carcinomas. Taken together, our findings indicate that the ERBB2 gene is a target of MBP-1 and suggest that the concomitant high expression of MBP-1 and HDAC1 may be considered a diagnostic marker for IDC.
Methods
Cell culture and tumor tissues
The ERBB2-amplified human breast cancer cell line SKBr3, was purchased from American Type Culture Collection (ATCC, Rockville, MD). Cells were cultured in DMEM medium supplemented with 10% fetal bovine serum, 2 mM glutamine and 100 μg/ml penicillin/streptomycin (Invitrogen, Carlsbad, CA).
Tumor tissue samples were from 45 patients submitted to routine histopathological examination at the Anatomic Pathology Unit of La Maddalena Hospital in Palermo. All experiments using human tissues were performed with the written patients’ informed consent and with the approval of Institutional Review Boards from La Maddalena Hospital.
Reporter and effector plasmid constructs
The construction of the effector plasmid pFlag-MBP-1 has been described previously [
19]. For the reporter constructs, the relevant regions of the ERBB2 promoter, including 44 base pairs (bp) of the first exon, were obtained by PCR amplification of genomic DNA from a human-mouse hybrid cell line containing only chromosome 17 [
24]. Three DNA fragments, spanning 306-, 558- and 787-bp, were amplified with primers containing restriction sites and cloned into the luciferace vector pGL3-basic (Promega, Madison, WI). In order to confirm the nucleotide sequence and the correct orientation of the cloned fragments, the three reporter plasmids, pG-E300, pG-E500 and pG-E700 were subjected to cycle-sequencing on an ABI 3130 genomic analyzer, according to the manufacturer’s instructions (Applied Biosystems, Foster City, CA),
Cell transfection and luciferase reporter assay
SKBr3 cells were transfected with Lipofectamine LTX reagent in OptiMem medium as instructed by the manufacturer (Invitrogen). For RT-PCR, western blot and ChIP analyses 1.5x106 cells in 10 mm culture dishes were transfected with either the pFlag-MBP1 (3.5 or 7.5 μg) or pFlag-CMV plasmid (7.5 μg) and cell extracts were prepared 48 hrs after transfection. An aliquot of the transfected cells was routinely monitored for transfection efficiency by immunofluorescence assay and Western blot analysis with anti-Flag antibodies. Only samples yielding more than 70% transfected cells and lysates with no detectable Flag-MBP-1 breakdown products were used for further analysis.
For immunofluorescence assays, 1.5x10
5 SKBr3 cells were grown onto glass coverslips in 12-well culture plates for 24 hrs, then transfected with either 750 ng of pFLAG-MBP1 or pEGFPN1 plasmid (Clontech, Mountain View, CA), as described previously [
19].
For reporter assays cells (6×105) were transfected with 750 ng of the pGL-cmp luciferase reporter construct and 250 ng of the β-galactosidase expressing vector pSVβ-gal (Promega, Madison, WI), the latter used as an internal control plasmid to monitor transfection efficiency. In cotransfection experiments with the pFLAG-MBP1 effector vector (1.25 μg), the total amount of DNA was kept constant by addition of the empty expression plasmid. Luciferase and beta-galactosidase activities were measured independently in duplicate using the Bright-Glo Luciferase Assay and Beta-Glo Assay Systems (Promega, Madison, WI) and a Turner 20/20 luminometer (Turner Designs, Inc., Sunnyvale, CA). Luciferase activity was normalized with respect to beta-galactosidase activity. All transfections were performed in triplicate and results from three independent experiments are expressed as mean ± SD.
Total RNA isolation and quantitative real-time PCR
Total RNA was extracted using Trizol reagent (Invitrogen, Carlsbad, CA) according to the manufacture’s instructions. RNA was reverse-transcribed with the Superscript II reverse transcriptase (Invitrogen, Carlsbad, CA) and cDNA amplified as described previously [
23] using either c-MYC or ERBB2 specific primers (Qiagen, Hilden, Germany) and Power SYBER Green PCR ready-mix in a 7300 thermal cycler (Applied Biosystems, Foster City, CA), primer sequences are listed in (Additional file
1: Table S1). PCR conditions were: denaturation at 95C° for 3 minutes, followed by 35 cycles at 95C° for 20 seconds, 60C° for 15 seconds, and 72C° for 15 seconds, and a final extension at 72°C for 7 minutes. Reaction specificity was controlled by post-amplification melting curve analysis and agarose gel electrophoresis of the amplified products. To correct for the experimental variations between samples, Ct value of TBP mRNA was determined in each PCR reaction using specific primers (Qiagen, Hilden, Germany). Data shown were generated from three independent experiments performed in triplicates and are expressed as mean ± SD. Comparison and statistical analysis were performed using Student
t test.
Immunofluorescence and microscopy
SKBr3 breast cancer cells were seeded onto glass coverslips in a 12-well plate culture vessel, 48–72 hrs post-transfection cells were fixed with 3.7% paraformaldehyde in phosphate buffered saline (PBS) and then permeabilized with 0.3% Triton X-100 in PBS. To detect endogenous ErbB2 and ectopically expressed Flag-MBP-1 proteins cells were incubated with 1 ug/ml of mouse anti-ErbB2 (sc-80898, Santa Cruz Biotechnology, Santa Cruz, CA) and rabbit anti-Flag (F7425, Sigma Chemical Company, St Louis, MO) primary antibodies in PBS containing 0.2% Tween 20. AlexaFluor 488-conjugated goat anti-rabbit IgG and AlexaFluor 594-conjugated goat anti-mouse IgG (Invitrogen, Carlsbad, CA) at a dilution of 1:600 were used as secondary antibodies. DNA was counterstained with 4′6-diamidino-2-phenylindole (DAPI) and the coverslips were mounted onto glass slides with Slowfade reagent (Invitrogen, Carlsbad, CA). Primary-antibody-omission demonstrated the specificity of the immunostaining. Immunofluorescence microscopy was performed with either a Leica DM-RA2 microscope, or a Leica TCS SP5 confocal laser-scanning microscope and confocal optical sections were created using Leica confocal software.
Immunoblotting and immunohistochemistry
Total cell lysates from transfected cells were prepared in RIPA buffer (50 mM TrispH 7.4, 150 mM NaCl, 1% Triton X-100, 0.1% SDS, 1% sodiumdeoxycholate, 1 mM EDTA, 0.5 mM DTT) supplemented with protease and phosphatase inhibitors (Sigma Chemical Company, St Louis, MO). Frozen normal and tumor tissues were homogenized and lysates prepared as described previously [
23]. Protein concentrations of tissue and cell lysates were determined by the Bradford protein assay (BioRad, Hercules, CA). Samples (30–40 ug) were separated on 4-12% polyacrylamide gradient gels (Invitrogen, Carlsbad, CA), and transferred to PVDF membrane, according to the manufacturer’s instructions (Amersham Biosciences, Sweden). Membranes were probed with primary antibodies: rabbit anti-Flag (F7425, Sigma Chemical Company, St Louis, MO, dilution 1:200), rabbit anti-ErbB2, (18299-1-AP, Proteintech, dilution 1:100), mouse anti-Myc (sc-40, Santa Cruz Biotechnology, Santa Cruz, CA, dilution 1:200) rabbit anti-HDAC1 (ab7028, Abcam, Cambridge, UK, dilution 1:500) and horseradish peroxidase-conjugated secondary antibodies (Amersham Bioscience, Sweden). Membranes were additionally probed with mouse beta-actin antibody (AC-15, Sigma Chemical Company, St Louis, MO) as a loading control. Detection was performed with a chemiluminescent substrate (Pierce Biotechnology, Rockford, IL) and signals were quantified by densitometric analysis employing the AlphaEasyFc software (Alpha Innotech Corporation, Johannesburg, South Africa).
Immunohistochemistry was performed on tissue serial sections of archived formalin-fixed, paraffin-embedded tissue blocks from patients as described previously [
23], using primary antibodies against ErbB2 (4B5, Ventana Medical System, dilution 1:500), MBP-1/alpha-enolase (monoclonal antibodies ENO-19/8 and ENO-276/3, 1.0 ug/ml, [
23]) and HDAC1 (ab7028, Abcam, dilution 1:1000). To confirm the specificity of immunoreactions, the primary antibody was either omitted or replaced by non-immune IgG. Tissue slides were evaluated blindly by two authors (ER and CL). The imunohistochemical grading scale used to evaluate the intensity and percentage of MBP-1-positive cells has been described previously [
23]. Tumors were graded as ErbB2-positive with a score of 3+ and negative with a score of 0 or 1+, according to common pathological guidelines. Tumors ErbB2-positive 2+ were further evaluated by in situ hybridization (FISH) with a dual-color probe (PathVysion ErbB2/CEP17; Vysis, Downers Grove, IL, USA), according to manufacturer’s instructions, and scored positive when ErbB2 gene amplification was found. Immunohistochemical score for HDAC1 expression in each tissue section was calculated as the percentage of positively stained cells on total cells.
Chromatin immunoprecipitation (ChIP) assay
In vivo MBP-1 and HDAC1 occupancy at the ERBB2 and c-MYC promoter was investigated using a ChIP assay kit (Upstate Biotech, Billerica, MA). Sheared chromatin samples from either pFlag-MBP1- or pFlag-CMV-transfected SKBr3 cells were separately immunoprecipitated with rabbit anti-Flag, anti-HDAC1 or anti-acetylated Histone H4 polyclonal antibodies (Upstate Biotech, Billerica, MA). The recovered DNA was analyzed by quantitative real-time PCR as described previously [
25], using primers specific to either ERBB2 or c-MYC promoter, and to unrelated sequences as a negative control (Additional file
1: Table S1). A DNA sample representing 10% of the total input chromatin was also included as a positive control. The data shown are means ± standard deviations (SD) from three independent experiments performed in triplicates and are expressed as percentage of total input DNA.
Statistical analysis
Group comparison and statistical analyses were performed using the software tools in GraphPad Prism version 4.02 for Windows (GraphPad Software, Inc. La Jolla, CA, USA). All tests of statistical significance were two-tailed and p-values less than 0.05 were considered statistically significant.
Discussion
In this study, we provide novel observations regarding the transcriptional control of the ERBB2 gene in SKBr3 breast cancer cells. The human ERBB2 gene is frequently amplified in breast tumors, and its high expression is associated with poor prognosis. However, substantial evidence suggests that the increased level of ERBB2 mRNA depends on active gene transcription in addition to gene amplification [
30]. Several positive and negative regulatory elements have been characterized in the ERBB2 gene proximal promoter as well as in the 5
′-flanking sequence up to 6 kb and in the first intron [
31‐
34]. Altogether, these studies indicate the involvement of several factors regulating ERBB2 gene transcription in breast cancer cells. Among positive regulators, members of the AP-2 and Ets families of transcription factors are required for maximal ERBB2 promoter activity and have been associated with the overexpression of the gene in breast cancer (for a review, see [
35]); in addition, the multifunctional transcription factor YY1 has been shown to cooperate with AP-2 to stimulate ERBB2 promoter activity through the AP-2 binding sites [
36]. Other transcription factors have been identified as negative regulators of ERBB2 expression in breast cancer (for a review, see [
37]): e.g., PEA3, an Ets DNA-binding protein that targets a DNA motif in the ERBB2 gene promoter [
38]; FOXP3, an X-linked breast cancer tumor suppressor which represses the transcription of the ERBB2 gene by interacting with forkhead DNA-binding motifs in the promoter [
39]; the zinc-finger transcription factor GATA4, part of a negative feedback regulatory loop with the ERBB2 gene [
40]. Although the functional relationships between positive and negative transacting factors still remain largely unexplored, overall, these data illustrate the complexity of ERBB2 gene transcriptional control.
In this context, we report that the c-MYC gene repressor MBP-1 negatively regulates ERBB2 gene transcription in SKBr3 breast cancer cells by targeting regulatory sequences in the promoter region. Through chromatin immunoprecipitation, we have located the MBP-1 binding region between nucleotide −514 and −262, relative to the major transcriptional start site of the ERBB2 gene, and demonstrated the concomitant recruitment of HDAC1 to the same region. Furthermore, our ChIP assays have indicated a decreased AcH4 occupancy at the same ERBB2 promoter region targeted by MBP-1 and HDAC1, suggesting a regulatory role for HDAC1 in MBP-1 repression activity, although we cannot exclude the involvement of other HDAC family members.
HDAC1 has positive and negative effects on gene transcription [
41] and, like all the HDACs, lacks a DNA-binding domain; thus, it must be associated with a DNA-binding protein in order to target a specific chromatin region (reviewed in [
42]). For example, to repress transcription, HDAC1 interacts with the transcription factor E2F in a complex containing BRM, BRG1, and SUV39H1 [
43]. Ghosh et al. previously demonstrated that MBP-1 physically associates with HDAC1 in vitro and in vivo, although the MBP-1-mediated repression of the c-MYC P2 promoter seems to be independent of HDAC1 [
18]. Our results support this previous observation concerning the c-MYC promoter and, conversely, suggest that MBP-1 represses ERBB2 gene transcription by recruiting the HDAC1 protein to its promoter. Therefore, MBP-1-mediated transcriptional repression may occur through different mechanisms, likely depending on the chromatin structure and the nucleotide sequence of the promoter. MBP-1 can block the assembly of the basal transcription complex by competing with TBP, as reported for the c-MYC P2 promoter [
7], or it may bind the promoter regulatory sequences and recruit HDAC1, as we suggest here, for the ERBB2 gene. The differences we observed in the recruitment of HDAC1 to ERBB2 and the c-MYC P2 promoter strongly support this last hypothesis.
Overall, our data suggest the existence of a novel transcriptional regulatory network that modulates ERBB2 expression, though detailed investigations using different cellular models are needed to dissect this network and define the molecular mechanisms underlying MBP-1/HDAC1-mediated transcriptional repression of the ERBB2 gene in breast cancer.
We also report a significant inverse correlation between ERBB2 expression and both MBP-1 (r= −0.278, p= 0.031) and HDAC1 (r= −0.267, p= 0.037) protein levels in primary breast tumors, and, accordingly, we propose MBP-1/HDAC1/ERBB2 relative expression as a diagnostic marker in breast IDC. Our results are in agreement with previous observations that have associated the reduction of HDAC1 transcript and protein levels with progression from normal mammary epithelium to ductal carcinoma in situ (DCIS) and to IDC [
27‐
29].
Furthermore, it has been independently reported that the expression of either MBP-1 or HDAC1 is a predictor of good disease-free survival, and both proteins are independent prognostic factors in breast cancer patients [
23,
29]. Despite the limited number of patients examined in this study, the significant positive correlation we observed between MBP-1 and HDAC1 expression in ErbB2-negative IDC suggests that their concomitant high expression may have a stronger diagnostic and prognostic significance in this tumor subtype.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
Conceived and designed the experiments: FC CM ER AG SF. Performed the experiments: FC CM AF MLP CL GP. Analyzed the data: FC CM ER AG SF. Contributed reagents/materials/analysis tools: FC CM CL GP AG SF. Wrote the paper: FC AG SF. All authors read and approved the final manuscript.