Background
The familiy of Activator Proteins-2 (AP-2, Tcfap2) comprises 5 highly conserved DNA-binding transcription factors referred to as AP-2α, AP-2β, AP-2γ, AP-2δ and AP-2ε (or Tcfap2a-e) [
1]. They preferentially bind GC-rich consensus-sequences in their target genes, which results in transcriptional regulation either as a stimulatory or repressive event [
1,
2]. All AP-2 proteins share a modular protein structure consisting of a proline/glutamine-rich transactivation domain at the amino terminus, followed by a highly conserved central basic region and a helix-span-helix motif at the carboxyl terminus. DNA-binding is mediated by the basic region and requires dimer formation of two AP-2 proteins via the helix-span-helix motifs. AP-2 proteins are involved in manifold cellular functions such as proliferation, differentiation and apoptosis and play pivotal roles during embryonic development and carcinogenesis [
1]. Especially the family members AP-2α and AP-2γ have been implicated in breast cancer [
3]. Under physiological conditions expression of AP-2α and AP-2γ is restricted to either the luminal or the myoepithelial compartment in the breast, respectively [
4,
5]. In contrast simultaneous overexpression of AP-2α and AP-2γ has been observed in breast carcinoma [
4], but their impact on mammary tumorigenesis is still discussed controversially [
6].
Several studies propose a tumor-suppressive function for AP-2α in breast tissue [
7‐
9] and data from clinical trials indicate that tumor progression is accompanied by loss of AP-2α expression [
10]. Moreover, loss of heterozygosity on chromosome position 6p22, where AP-2α is mapped to, is frequently observed in breast cancer specimens [
11]. Accordingly, high expression of AP-2α in invasive breast cancer is correlated with favorable overall survival rates in patients [
4].
Furthermore, in transgenic mouse studies, overexpression of AP-2α in the mammary epithelium resulted in impaired mammary gland growth caused by a reduction in proliferation and a simultaneous increase in apoptosis [
12]. This fits
in vitro data where loss of AP-2α lead to a decrease of apoptosis and an increased resistance towards chemotherapeutic drugs [
8]. These effects might be partly mediated through control of expression of the CdK-Inhibitor p21
WAF and of p53 by AP-2α, which supports the idea that AP-2α acts as a tumor suppressor [
9,
13,
14].
In contrast, little is known about the actual role of AP-2γ in breast carcinogenesis. Overexpression of AP-2γ is frequently seen in breast tumors and breast cancer cell lines [
15,
16] and has recently been shown to correlate with poorer response to hormone therapy and reduced patient survival in invasive breast cancer [
17,
18]. In addition, increased expression of AP-2γ has been associated with poorly differentiated breast tumor samples [
4]. Gene amplification has been proposed as potential mechanism leading to overexpression of AP-2γ because the gene maps to the genomic locus 20q13.2, which is frequently amplified in breast cancer [
19]. Transgenic mouse studies provide further evidence for a role of AP-2γ in breast tumor formation: MMTV-driven overexpression of AP-2γ in the mammary epithelium leads to increase in proliferation and an impaired differentiation [
20]. Moreover, bi-transgenic mice expressing the oncogenic HER-2/neu receptor tyrosine kinase and AP-2γ showed an accelerated tumor progression compared to single-transgenic HER-2/neu mice [
21]. Together, these findings point towards a causal involvement of AP-2γ in the etiology of breast cancer.
AP-2 proteins form either homo- or heterodimers [
1]. There is evidence that under physiological conditions the net effect whether a cell proliferates, undergoes apoptosis or differentiates, may depend on the balance between different AP-2 isoforms [
22]. So far, experiments addressing AP-2 functions in breast cancer cells were using either overexpression of, or interference with particular AP-2 isoforms. Overexpression of transcription factors may lead to occupancy of promoters not used under physiological conditions, which masks specific AP-2 functions. In the latter case, interference with a particular isoform may cause a shift in AP-2 dimer composition and hence obscure the physiological function.
In a complementary approach, we now set out to investigate the role of AP-2 proteins in a global manner by interfering with them functionally. To this end, we conditionally overexpressed a dominant-negative mutant (ΔAP-2γ) in the N202.1A murine breast cancer cell line expressing AP-2α and AP-2γ endogenously. Interference with AP-2 proteins influenced the expression of genes involved in chemo- and radiation-sensitivity and resulted in the induction of apoptosis. These results establish an important role for AP-2 transcription factors in breast cancer cells determining their sensitivity towards chemotherapeutic agents and ionizing radiation. Expression of AP-2 proteins in breast tumors might therefore be of prognostic value. The results also suggest that AP-2 proteins or their target genes may be promising targets for therapeutic intervention in breast cancer.
Methods
Plasmids
pUHG172-1neo (rtTA) was kindly provided by H. Bujard (Heidelberg, Germany). In order to yield pBIEGFP, EGFP was cloned from pEGFP-C3 into pBI-4 (kindly provided by H. Bujard (Heidelberg, Germany) as an NheI/XbaI fragment. ΔAP-2γ was created by deleting 474 bp (158 aa of the NH-terminus) of the 5'-portion of the murine AP-2γ cDNA. Upon introduction of a start codon using a NotI/NaeI-flanked oligo, ΔAP-2γ was cloned into pBIEGFP using NotI/Sal I restriction sites to generate pBIEGFPΔAP-2γ. The BMP-4 Firefly luciferase construct was kindly provided by M. Moser, (Martinsried, Germany) and the CMV-Renilla luciferase plasmid was obtained from Promega (Mannheim, Germany).
Cell Lines and Cell Culture
The murine breast cancer cell line N202.1A is derived from MMTV-HER-2/neu transgenic mice and was kindly provided by P.-L. Lollini (Bologna, Italy) [
22]. N202.1A cells were grown in DMEM-Glutamax supplemented with 20% fetal calf serum and penicillin/streptomycin (Invitrogen, Karlsruhe, Germany). Cells were kept under standard conditions using a cell culture type incubator at 37°C under 7.5% CO
2. Functional assays were performed 96 h after addition of doxycycline to the cell culture media (2 μg/ml, BD Biosciences, Heidelberg, Germany).
Transient and stable Transfection
For stable transfection N202.1A cells were transfected using 39 μg plasmid DNA (ratio plasmid - resistance gene 10: 1) using 107.5 μl Lipofectamine 2000 (Invitrogen, Karlsruhe, Germany) per 10 cm culture dish according to the manufacturer's protocol. N202.1A cells were selected at 850 μg/ml G418 (Calbiochem, Darmstadt, Germany) for the pUHG172-1neo and 200 μg/ml Hygromycin B (BD Biosciences, Heidelberg, Germany) for pBIEGFPΔAP-2γ and pBIEGFP. For BMP-4 luciferase assays N202.1A cells were transfected with 900 ng plasmid (870 ng BMP-4 Firefly luc +30 ng CMV-Renilla luc) and 1.75 μl Lipofectamine 2000 (Invitrogen, Mannheim, Germany) per 24 well culture dish according to the manufacturer's protocol.
Microscopy and Image Processing
Cells were visualized using a Leica-DM-IRB microscope (Bensheim, Germany), fitted to a Microfire digital camera (Optronics, Goleta, CA, USA) and image processing was performed applying Adobe Photoshop and Illustrator software.
Co-Immunoprecipitation
HCT116 cells were transfected with equal amounts of AP-2α- and ΔAP-2γ - expression plasmids. Cells were lysed after 48 hours with non-denaturing lysis buffer. Co-IP was performed with 20 μl DYNABEADS®(Invitrogen Cat.no. 199.03D, Invitrogen, Karlsruhe, Germany) und 1,5 μg anti-AP-2α antibody (H79, Santa Cruz, Heidelberg, Germany). 150 μg protein lysate per sample was loaded. Western blot with anti-AP-2γ antibody (6E4/4, Santa Cruz, Heidelberg, Germany) followed.
Western Blot Analysis, Luciferase Assays and Giemsa Staining
Western Blot was performed using the following primary antibodies: AP-2γ (6E4/4, 1:200 Upstate, New York, USA), AP-2α (H79, 1:200 Santa Cruz Heidelberg, Germany), AP-2β (1:1000), AP-2δ (1:1500) and AP-2ε (1:1500), all kindly provided by M. Moser (Martinsried, Germany). The following secondary antibodies were used: goat anti-rabbit-HRP (1:2000 DAKO, Hamburg, Germany) and rabbit anti-mouse HRP (1:1000, DAKO, Hamburg, Germany). Protein lysates from HeLa (cervical carcinoma) N2A (neuroblastoma) cell lines and human keratinocytes or in vitro synthesized proteins, respectively, served as antibody positive controls for the different AP-2 isoforms in Western Blot analysis. BMP-4 promoter luciferase assays were performed using the "Dual Luciferase Reporter Assay Systems Kit" (Promega, Mannheim, Germany) according to the manufacturer's protocol. 48 h upon induction (2 μg/ml doxycycline), N202.1A cells were transfected with BMP-4-Firefly luc, lysed, and luciferase activity was measured at 562 nm (Berthold Technologies, Bad Wildbad, Germany). Transfection efficiencies were normalized using a CMV-driven Renilla Luciferase (CMV-Renilla luc). To quantifiy cell numbers after treatment with chemotherapeutic compounds and irradiation cells were stained with Giemsa. For this purpose cells were washed twice in ice-cold PBS and fixed in methanol for 5 min at room temperature, stained for 5 min in Giemsa solution (Merck, Darmstadt, Germany) and washed gently in H2O. Cells were quantified using ImageJ software (Adriamycin treatment) or three independent fields of visions were counted (irradiation) in each experiment.
Proliferation
Determination of S-phase index was carried out using the "Click-iT™ EdU Imaging Kit" (Invitrogen, Karlsruhe, Germany) according to the manufacturer's instructions. Briefly, 96 h after addition of doxycycline (2 μg/ml) N202.1A cells were incubated with 100 μM thymidine analogon EdU for 30 min at 37°C, fixed in 3.7% formaldehyde for 15 min and washed twice in 3% BSA/PBS. Permeabilization was achieved using 0.5% Triton X-100/PBS for 20 min and cells were washed twice with 3% BSA/PBS. Subsequently, the fixed cells were incubated with the reaction cocktail containing Alexa Fluor®594-azide for 30 min and washed with 3% BSA/PBS. Imaging using a Leica DM-IRB (Leica, Bensheim, Germany) microscope was followed by quantification of Edu Alexa Fluor®594-azide staining in three fields of vision (magnification ×100) in three independent experiments. On average, a field of vision contained 218 cells.
Transmission (TEM) - and scanning electron microscopy (SEM)
For TEM-analysis, cells were fixed in 1.25% glutaralaldehyde in 0.1 M cacodylate buffer (pH 7.4) for 1 h and post-fixed in 2% OsO4 in cacodylate. Following dehydration in a graded series of ethanol, the specimens were embedded in a propylene-epon-mixture, mounted as 50 nm thin sections and post-contrasted with 3.5% uranylacetate. TEM analysis was performed using a CM10 microscope (Philips, Eindhoven, Netherland).
For SEM cells were fixed in 2% glutaralaldehyde in 0.1 M sodium cacodylate (pH 7.3) for 20 min, transferred to 0.1% aqueous tannic acid and rinsed with distilled water. All specimens were dehydrated through a graded series of ethanol and critical point-dried from CO2 in 10 cycles using a Balzers CPD 030 (BAL-TEC, Schalksmuehlen, Germany). Dried specimens were mounted on aluminium sample holders and coated with a 2 nm layer of platinum/palladium in a HE 208 sputter coating device (Cressington, Watford, UK). SEM analyis was performed with an XL 30 SFEG (Philips, Eindhoven, Netherlands).
Caspase 3/7 Acitivity and AnnexinVC3.18 staining
Caspase 3/7 activity was quantified using the "Caspase-Glo® 3/7 Assay Kit" (Promega, Mannheim, Germany) according to manufacturer's instructions. Non-induced or cells induced with doxycycline (2 μg/ml) for 96 h were incubated with "Caspase-Glo® 3/7 Reagent" for 3 h followed by measurement of luminescence at 562 nm (Berthold Technologies, Bad Wildbad, Germany). AnnexinVC3.18 staining was carried out using the "AnnexinVC.18 Kit" (Sigma-Aldrich, Munich, Germany). Doxycycline-induced cells were washed 3 times in PBS followed by 3 cycles of binding buffer for 1 min. In the next step cells were incubated with the AnnexinVC3.18 conjugate (AnnexinVC3.18:binding buffer, 1:100) for 10 min at room temperature, washed 3 times in binding buffer and then subjected to fluorescence microscopy. Cells treated for 2.5 h with Staurosporine (2 μM) served as positive controls (Sigma-Aldrich, Munich, Germany).
Irradiation
The irradiation of the N202.1 cells 96 h after addition of doxycyline (2 μg/ml) at 105 Gy was performed using a linear accelerator (Siemens Mevaton MD2, Siemens Medizintechnik, Munich, Germany). The photon energy of 6 MeV at a dose rate of 2 Gy/min was chosen. The field size was set to 20 cm × 20 cm at a SSD (skin to surface distance) of 100 cm. The beam divergence was approx. 11°. The cells were irradiated in a 6-well plate using a RW3-Phantom (PTW) at dose maximum. The photon beam was calibrated according the DIN 6800-2 protocol for a water equivalent energy dose. The RBW-factor for 6 MeV irradiation is almost 1, so the energy dose was equivalent to the biological dose.
RNA isolation and qRT-PCR
Total RNA was isolated from the various clones using Trizol Reagent (Invitrogen, Karlsruhe, Germany) according to the manufacturer's instructions. 1 μg of DNase-treated (DNA-free™ kit, Ambion, Austin, TX) total RNA was reverse transcribed using RETROscript™ reagents (Ambion, Austin, TX). RNA was heat-denatured for 3' at 85°C and the reaction was incubated at 42°C for 1 hour and 10' at 92°C. Quantitative Real Time PCR (qRT-PCR) reactions were carried out in 96 well plates using SYBRGreen
®Master Mix (Applied Biosystems, Foster City, CA), specific primers and 10 ng total RNA converted into cDNA in 10 μl final volume. Fluorescence was measured using an ABI Prism
® 7300 (Applied Biosystems, Foster City, CA) detection system according to the manufacturer's instructions. Primers were purchased from QIAGEN (QuantiTect
® Primer Assays). Relative quantitations to control cells were performed: first, each Ct value was corrected for the Ct
r of the reference gene, GAPDH, and then the Ct of each sample was subtracted from the Ct of control cells (Ct
0). The relative amount of template (Q) was therefore calculated as:
. All samples were run in triplicates and mean and standard deviation calculated as described in Bookout et al. 2003 [
23]. QuantiTect
® Primer Assay catalogue numbers are as follows: QT00101297 Mm_Tcfap2c_1_SG QuantiTect Primer Assay (200) (NM_009335, NM_001159696); QT00265524 Mm_Egr3_1_SG QuantiTect Primer Assay (200) (NM_018781); QT00096131 Mm_Ctgf_1_SG QuantiTect Primer Assay (200) (NM_010217); QT01044295 Mm_Sema3b_1_SG QuantiTect Primer Assay (200) (NM_009153); QT00157381 Mm_Nrp1_1_SG QuantiTect Primer Assay (200) (NM_008737); QT01061599 Mm_Gsta3_1_SG QuantiTect Primer Assay (200) (NM_010356); QT00100653 Mm_Gzme_1_SG QuantiTect Primer Assay (200) (NM_010373); QT00134064 Mm_Tnfaip3_1_SG QuantiTect Primer Assay (200) (NM_009397). QT00105483 Mm_Fst_1_SG QuantiTect Primer Assay (200) (NM_008046).
Northern Blot
Northern Blot analysis was carried out as described in Jäger et al. [
20] using the murine AP-2γ cDNA as a probe.
Whole genome expression analysis
Microarray analysis of gene expression in response to expression of ΔAP-2γ/EGFP (Δ#7 + Δ#15) or EGFP (Δ#5 + Δ#11) in N202.1A breast cancer clones (2 μg/ml doxycyline for 96 h) was performed using the Illumina BeadChip system (Illumina, Inc, San Diego, CA). We used 500 ng of total RNA to obtain labeled, amplified cRNA for each sample to hybridize the Illumina Ref-8 BeadChips according to manufacturer's instructions (Illumina, Inc, San Diego, CA). Arrays were scanned with an IlluminaBeadArray Reader confocal scanner and data processed and analyzed using IlluminaBeadStudio software (Illumina, Inc, San Diego, CA). Raw Illumina data were rank invariant normalized with BeadStudio software (Illumina, Inc, San Diego, CA), which was also used to assess differential expression between the ΔAP-2γ and control clones, based on three RNA preparations from each clone after subtraction of the background obtained with control clones following doxycyclin treatment. After normalization, genes were filtered by their 'detection' value, which had to be 0.99 (significantly detected), in the three samples. Subsequently, we identified differentially expressed genes using the Illumina custom error model implemented in BeadStudio, which provides an expression difference score ('DiffScore') taking into account background noise and sample variability [
24]. We chose a DiffScore threshold of 30, corresponding to a p value of 0.001, with a False Discovery Rate (FDR) lower than 5%. To restrict the analysis to the most regulated genes, an additional filtering criterion was that the average expression fold-change between ΔAP-2γ/EGFP- (Δ#7 or Δ#15) and EGFP- (Δ#5 or Δ#11) expressing clones had to be at least 1.5-fold, which lead to the identification of 139 modulated transcripts (49 decreased and 90 increased). Sample permutation analysis confirmed that under these conditions the FDR was well below 5%.
Gene Set Enrichment Analysis
Gene Set Enrichment Analysis (GSEA) is a computational method used to look for overlaps between the AP-2-driven gene set obtained from the microarray analysis and modulated genes present in the Molecular Signature Database (MSigDB) following chemical and genetic perturbations [
25].
Chromatin immunoprecipitation (ChIP) assays
ChIP was performed using the ChIP-IT™ Kit (Active Motif, Carlsbad, CA) reagents and protocols. Primer pairs were designed on the TFAP2 binding site containing regions (identified by TRANSFAC) using the Primer-BLAST software
http://www.ncbi.nlm.nih.gov/tools/primer-blast/index.cgi?LINK_LOC=NcbiHomeAd. The following primers were used: Tnfaip3FW: 5'-CCCCTAACGGAGGCACTCTTCCAC-3'; Tnfaip3RV: 5'-CCGCCTCCTCCAGGTCTTCCTAGCCC-3'; CtgfFW: 5'-AGGAAGTCTC GGGCCTCTTCTCTTTGA-3';CtgfRV: 5'-TCAAGTGGCTGACCACATCATCTGCAC-3'.
PCR was performed using Platinum® Taq DNA Polymerase according to the manufacturer's instructions.
Discussion
Although single AP-2 proteins have been implicated in tumorigenesis in various types of cancers, their concerted action in initiation and progression of breast cancer is far from being understood. Previous work put its focus on the analysis of the role of single AP-2 isoforms in the etiology of breast cancer. Using a dominant-negative AP-2 mutant, we demonstrated that simultaneous interference with both, AP-2α and AP-2γ leads to a decrease in proliferation and induction of apoptosis in N202.1A breast cancer cells. Furthermore the cells became increasingly sensitive to chemotherapeutic drugs and irradiation.
AP-2α and AP-2γ were suggested to act as decisive pacemakers for cellular fates such as proliferation, apoptosis and differentiation in the mammary epithelium under physiological conditions [
5,
20]. Therefore, their inappropriate spatio-temporal expression might contribute to malignant transformation. While in the normal breast tissue AP-2α and AP-2γ show non-overlapping expression patterns [
5], coexpression of AP-2α and AP-2γ was detected in undifferentiated breast carcinomas [
4]. There is increasing evidence that the net outcome whether a cell proliferates, undergoes apoptosis or differentiates depends on the balance of various AP-2 proteins [
22]. Here, interference with AP-2 proteins using a dominant-negative AP-2 mutant in N202.1A breast cancer cells resulted in a decreased proliferation rate. Remarkably, overexpression of AP-2γ in the mammary gland using a transgenic mouse approach caused an increase in proliferation as well [
20], highlighting the notion that certain AP-2 proteins might trigger proliferation.
Wajapeyee et al. correlated increased AP-2α expression with an enhanced rate of apoptosis after treatment with cytostatic compounds [
7,
8]. Our results, however, suggest that global interference with AP-2 proteins rather sensitizes cells to apoptosis and anticancer treatment regimen. We obtained conflicting results to those reported for the function of AP-2α alone because we could detect apoptosis and an increased sensitivity towards chemotherapeutic drugs and irradiation after global interference with AP-2 proteins. Our approach benefits from the fact that the dominant-negative AP-2 mutant presumably interferes with all AP-2 homo- and heterodimers present in N202.1A breast cancer cells, which possibly occupy different target gene promoters than AP-2α alone. While this approach can not distinguish between the functions of individual AP-2 family members, it suggests that interfering with AP-2γ may sensitize to apoptosis even in the absence of functional AP-2α.
The whole genome expression analysis revealed 139 differentially expressed genes upon interference with AP-2 proteins. Their function might explain the decrease in proliferation rate and at the same time the increase in chemo- and radiation-sensitivity upon repression of AP-2 function also based on modulations of some of these genes previously found following chemical and genetic perturbations [
25]. On mRNA level the following genes were found upregulated upon interference with AP-2 function suggesting that AP-2 proteins act as repressors of these genes:
Connective Tissue Growth Factor (CTGF) is described to suppress proliferation in breast cancer cells and other tumor entities and induces apoptosis at least in part by activation of caspase-3 [
31‐
33]. Moreover, low expression levels of
CTGF in breast cancer patients have been correlated with a poorer clinical outcome [
34].
Neuropilin 1 (NRP-1) receptor has been described to be involved in induction of apoptosis in breast cancer cell lines by inhibiting the Akt-signaling pathway [
35,
36]. The tumor suppressor gene
Tumor necrosis factor, alpha-induced protein 3 (Tnfaip3) is an inhibititor of the pro-survival activity of NFκB [
37].
Remarkably, the
ATP-binding-cassette, subfamilyB (MDR/TAP) member 9/GSTA3 was downregulated upon interference with AP-2 function. It is member of a family, which has been shown to contribute to the Multidrug Resistance (MDR) phenotype [
30]. Additionally this efflux transporter family has been reported to detoxify reactive oxygen species (ROS) generated by irradiation, [
29] which is a potential mechanism how tumor cells acquire resistance towards irradiation-induced damage. This might represent a mechanism of sensitization of N202.1A breast cancer cells towards chemotherapeutic drugs and irradiation caused by the abrogation of AP-2 function.
According to the analysis using ingenuity software, the AP-2 target genes predominantly cluster to NFκB, Tp-53, Ras signalling and Calcium signalling. In summary the concerted deregulation of the candidate genes might render cells more susceptible to apoptosis thus resulting in an increased sensitivity towards chemotherapeutic drugs and irradiation. It remains to be elucidated, whether the candidate genes are direct targets of AP-2 transcription factors or are regulated by an indirect secondary mechanism.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
VT constructed the vectors, generated the cells, performed the induction, radiation and chemotheraphy experiments and writing. FO performed array and ChIP-analyses bioinformatics and interpretation, writing and revision. RJ contributed to conception, acquisition of data, interpretation and writing and performed RNA isolation and northern blots. DE contributed to conception and design, bioinformatics, interpretation, writing and revision of data. GK helped with TEM analyses. SS performed co-immunoprecepitation and western-blot. SG helped with the radiation experiments, DT contributed to interpretation, writing and revision, HS contibuted to conception and design, interpretation, writing and revision. All authors read and approved the final manuscript.