Low SAFB levels are associated with worse outcome in breast cancer patients.

The scaffold attachment factors SAFB1 and SAFB2 have been shown to function as estrogen receptor (ERalpha) co-repressors in breast cancer cells, and to affect many cellular processes such as stress response, RNA processing, and apoptosis. SAFB1 and SAFB2 have also been implicated in breast tumorigenesis: Their shared chromosomal locus at 19p13 is frequently lost in breast cancer, mutations have been identified, and overexpression results in growth inhibition. The purpose of this study was to determine SAFB1/SAFB2 protein expression in human breast tumors, to correlate their expression with either natural progression ("prognostic factor") or with response to Tamoxifen ("predictive factor"), and to analyze potential correlations with tumor characteristics. SAFB1/SAFB2 protein were measured by immunoblotting using a pan-SAFB antibody in tumor extracts from patients with long-term clinical follow-up (n = 289), a subset of whom had received no adjuvant systemic therapy after breast cancer surgery (n = 117) and another subset of whom were treated with adjuvant Tamoxifen (n = 172). SAFB levels were correlated with clinico-pathological variables and patient outcome. SAFB levels varied widely, with 25 tumors not expressing detectable levels of SAFB. SAFB expression was significantly correlated with ERalpha, HER-2, bcl-2 and with expression of other ERalpha coregulators such as SRC-3. There was no association between SAFB expression and disease free survival, however, low SAFB expression was significantly associated with worse overall survival in patients who did not receive adjuvant therapy. This study shows that low SAFB protein levels predict poor prognosis of breast cancer patients, suggesting critical functions of SAFB1 and SAFB2 in breast cancer cells.

SAFB1 mediates repression of immune regulators and apoptotic genes in breast cancer cells.

The scaffold attachment factors SAFB1 and SAFB2 are paralogs, which are involved in cell cycle regulation, apoptosis, differentiation, and stress response. They have been shown to function as estrogen receptor corepressors, and there is evidence for a role in breast tumorigenesis. To identify their endogenous target genes in MCF-7 breast cancer cells, we utilized a combined approach of chromatin immunoprecipitation (ChIP)-on-chip and gene expression array studies. By performing ChIP-on-chip on microarrays containing 24,000 promoters, we identified 541 SAFB1/SAFB2-binding sites in promoters of known genes, with significant enrichment on chromosomes 1 and 6. Gene expression analysis revealed that the majority of target genes were induced in the absence of SAFB1 or SAFB2 and less were repressed. Interestingly, there was no significant overlap between the genes identified by ChIP-on-chip and gene expression array analysis, suggesting regulation through regions outside the proximal promoters. In contrast to SAFB2, which shared most of its target genes with SAFB1, SAFB1 had many unique target genes, most of them involved in the regulation of the immune system. A subsequent analysis of the estrogen treatment group revealed that 12% of estrogen-regulated genes were dependent on SAFB1, with the majority being estrogen-repressed genes. These were primarily genes involved in apoptosis, such as BBC3, NEDD9, and OPG. Thus, this study confirms the primary role of SAFB1/SAFB2 as corepressors and also uncovers a previously unknown role for SAFB1 in the regulation of immune genes and in estrogen-mediated repression of genes.

Estrogen-mediated downregulation of CD24 in breast cancer cells.

We have previously reported on the relevance of the prevalence of CD44(+)/CD24(-/low) cells in primary breast tumors. To study regulation of CD24, we queried a number of publicly available expression array studies in breast cancer cells and found that CD24 was downregulated upon estrogen treatment. We confirmed this estrogen-mediated repression of CD24 mRNA by quantitative real-time PCR in MCF7, T47D and ZR75-1 cells. Repression was also seen at the protein level as measured by flow cytometry. CD24 was not downregulated in the ER alpha negative MDA-MB-231 cells suggesting that ER alpha was necessary. This was further confirmed by ER alpha silencing in MCF7 cells resulting in increased CD24 levels and by reintroduction of ER alpha into C4-12 cells resulting in decreased CD24 levels. Estrogen treatment did not alter half-life of CD24 mRNA and new protein synthesis was not essential for repression, suggesting a primary transcriptional effect. Histone deacetylase inhibition by Trichostatin A completely abolished the repression, but decrease of the ER alpha corepressors NCoR, LCoR, RIP140, silencing mediator of retinoid and thyroid hormone receptors, SAFB1 and SAFB2 by siRNA or overexpression of SAFB2, NCoR and silencing mediator of retinoid and thyroid hormone receptors had no effect. In silico promoter analyses led to the identification of two estrogen responsive elements in the CD24 promoter, one of which was able to bind ER alpha as shown by electrophoretic mobility shift assay and chromatin immunoprecipitation assay. Together, our results show that CD24 is repressed by estrogen and that this repression is a direct transcriptional effect depending on ER alpha and histone deacetylases.