The BCL9-2 proto-oncogene governs estrogen receptor alpha expression in breast tumorigenesis.

The majority of human breast cancers express estrogen receptor alpha (ER), which is important for therapy with anti-estrogens. Here we describe the role of BCL9-2, a proto-oncogene previously characterized as co-activator of Wnt/ß-catenin signaling, for mammary tumorigenesis in mice and human. ER positive human breast cancers showed overexpression of BCL9-2 and tamoxifen treated patients with high BCL9-2 demonstrated a better survival. BCL9-2 was upregulated during puberty and pregnancy in normal mammary epithelia, but downregulated in the involuted gland. BCL9-2 overexpression in vivo delayed the mammary involution and induced alveolar hyperplasia. Moreover, aged BCL9-2 transgenic mice developed ductal-like mammary tumors with high nuclear ER expression. We found, that primary cell cultures of BCL9-2 breast tumors responded to tamoxifen treatment. Moreover, BCL9-2 regulated the expression of ER and the proliferation of human breast cancer cells independently of ß-catenin. Finally, we describe a novel mechanism, how BCL9-2 regulates ER transcription by interaction with Sp1 through the proximal ESR1 gene promoter. In summary, BCL9-2 induces ER positive breast cancers in vivo, regulates ER expression by a novel ß-catenin independent mechanism in breast cancer cells, and might predict the therapy response to tamoxifen treatment.

100 ng TOP or FOP reporter and 25 ng tk-Renilla control for 48 hours followed by treatment with Wnt3a-or control-conditioned media produced in L cells (ATCC) for additional 24 hours. Firefly luciferase values were normalized to Renilla controls.

Chromatin immunoprecipitation (ChIP), re-ChIP and Immunoprecipitation
ChIP and re-ChIP assays on MCF7 were performed after chromatin crosslink with 1% formaldehyde and sonification on a Diagenode BioRuptor (3-times 30 cycles of 10 sec on/off at high power) using the EZ-ChIP Kit (Merck Millipore) according to manufacturer`s protocol. For ChIP assays after mithramycin A treatment, MCF7 were pretreated 24 hours either with 0.1% EtOH as vehicle or 100 nM mithramycin A (Sigma). The elution of DNA and reverse protein/DNA crosslink was performed using Chelex-resin solution (Bio-Rad Laboratories) or using IPure kit (Diagenode) according to manufacturer`s protocol. Mouse and rabbit IgG (m IgG, rb IgG) were used as negative controls and for the normalization of the amplified signals. The fold enrichment was calculated as percent of input DNA. For co-IP's, cells were incubated on ice in RIPA buffer (50 mM Tris-HCl, pH 7.4; 150 mM NaCl; 1 mM EDTA, pH 8; 1% NP40; 0.25% Na-desoxycholate) containing 1x protease inhibitor cocktail (Roche). After centrifugation, the supernatants were incubated with the respective antibodies and Protein G sepharose (GE Healthcare) according to standard protocols. The immunoprecipitated proteins complexes were analyzed by Western Blot. ChIP experiments and co-immunoprecipitation were repeated at least three times. Primary antibodies are in Suppl. Table 2. Primers for PCR and qRT-PCR analysis of the precipitated chromatin complexes are provided in Suppl. Table 3.

Immunoreactive scores
BCL9 and BCL9-2 scores were calculated for the intensity and percentage of positive epithelial cells as previously described [1; 2]. Staining for nuclear ER and PR were scored according to the Remmele score [3] (IRS 0-12). Membranous Her2 scores were from 0 to 3+.
Tissues with scores of 0 and 1 for ER, PR, or Her2 respectively were evaluated as negative, the cases with higher scores as positive, according to the European guidelines for quality assurance in breast cancer screening and diagnosis (4 th edition, 2006).

GEO Data analysis
The Gene Expression Omnibus (GEO) GSE6532 dataset [4] with normalized gene expression values of human Affymetrix arrays was retrieved from the NCBI data repository. The dataset was created from early stage patients with ER+ breast cancers and provides information of ER positivity, tamoxifen treatment, and patient's survival. Array data from 263 tamoxifen treated patients were grouped as high or low relative to the median gene expression of BCL9-2 (228065_at).

Statistical methods
Box Plot analyses were performed for the scores of preneoplastic tissue changes and the immunoreactive scores and the significances were calculated using the Mann-Whitney U-test provided by the SPSS package 19. Cumulative hazard for the risk of breast tumors of mice were determined by Kaplan-Meier analysis using SPSS 19. Frequency of mammary tumor development in the mouse models were analyzed by the Chi-Square Test (Suppl. Table 1) and graphs for the dose-response of primary control and BCL9-2 tumor cells were plotted after nonlinear regression analysis using Graph Pad Prism 5. Survival analysis for relapse free patient survival was analyzed by Kaplan-Meier plots and significance was calculated using the Cox Proportional Hazards Model with the free statistical software R (version 2.14.1). For all other experiments, graphs were calculated as means ± SEM and statistics were evaluated by the Student's t-test using Microsoft Excel. P-values of <0.05 were considered to be statistically significant.

Image acquisition
Immunostains were visualized on a Leica DM5000 B upright microscope with 20x and 40x objectives, colonies from 2D-collagen assays on a Leica DM IRB inverted microscope with a 10x objective and Carmine whole mount stains on a Leica MZ FLIII stereo microscope at