In FY11, we have continued to focus primarily on elucidating the tumor cell-autonomous components of the switch in activity of TGF-beta from tumor suppressor to pro-progression factor. Our main experimental platform is a xenograft model of breast cancer progression based on the MCF10A human breast epithelial cell line. We have previously demonstrated that TGF-beta switches from tumor suppressor to pro-progression factor in this model, and the high degree of genetic relatedness between the different cell lines of the progression series gives us an exceptionally high signal-to-noise system in which to address mechanisms underlying the TGF-beta switch. Our work in FY11 has focused in three main areas: 1. THE ROLE OF TGF-BETA IN REGULATING CANCER STEM CELL DYNAMICS. During FY11 we have continued to develop and employ our novel functional imaging approach for the identification of the putative cancer stem cell (CSC) population, and we have examined how TGF-beta regulates CSC dynamics. Our CSC reporter responds to the presence of the stem cell master transcription factors Oct4 and Sox2, and marks cells that are enriched for CSC activities, including the ability to divide asymmetrically and to initiate tumorigenesis and metastasis in vivo. We have previously shown that TGF-beta has tumor suppressor activity in the MCF10Ca1h human breast cancer model. Now we have demonstrated that a significant fraction of the CSC population in this model has endogenously activated TGF-beta signaling. By timelapse fluorescence microscopy, we have shown that the majority of these cells are non-proliferative. Furthermore, we find that CSCs are intrinsically more migratory and invasive than the bulk tumor cell population, and that TGF-beta can inhibit these activities in CSCs while having a mildly stimulatory effect on the bulk cell population. Thus TGF-beta appears to play an important role in maintaining quiescence in a subpopulation of CSCs and in keeping this population stationary. In order to extend this approach from transplanted to spontaneously arising tumors, we have generated transgenic mice expressing the same stem cell reporter construct. Multiple founder lines have been obtained and several look promising. Understanding how CSCs are regulated will be critical to development of more effective cancer therapies as these cells are largely resistant to existing therapeutic approaches, leading ultimately to relapse. 2. INSIGHTS INTO TGF-BETA-MEDIATED TUMOR SUPPRESSION FROM THE GENERATION OF A CORE TGF-BETA/SMAD3 GENE EXPRESSION SIGNATURE. TGF-beta antagonists are being developed as cancer therapeutics. However, the complex role of TGF-beta in cancer progression makes it imperative to avoid treating patients whose tumors still have intact tumor suppressive responses to TGF-beta. Published TGF-beta response signatures have not distinguished the tumor suppressive responses from the neutral or tumor promoting responses. Our goal here was to develop a signature that specifically reflects the tumor suppressor activity of TGF-beta with the hope that it provide insights into the underlying biology and be useful for patient stratification in TGF-beta antagonist clinical trials. Smad2 and Smad3 are highly homologous proteins that transduce the TGF-beta signal. Using conditionally immortalized mammary epithelial cells of different Smad genotypes, we have shown that Smad3 is critical for the majority of the biological responses to TGF-beta, with Smad2 playing a lesser or even opposing role. Using shRNA knockdown approaches, we further demonstrated that Smad3 mediates the tumor suppressor activity of TGF-beta in MCF10Ca1h breast tumor xenografts and that this activity is lost in the closely related MCF10Ca1a tumors. We therefore focused on Smad3 to address underlying mechanisms. First we performed genome-wide chromatin immunoprecipitation to identify TGF-beta/Smad3 target genes in the MCF10 progression series. Integration of the promoter occupancy data with global gene expression data both in vitro and in vivo yielded a core Smad3-based gene signature that is associated with TGF-beta-mediated tumor suppression. In a meta-analysis of more than 1300 human breast cancers, high expression of this signature associated with good distant metastasis-free survival, suggesting that the tumor suppressor activity of TGF-beta does indeed play an important role in human breast cancer. Unlike the majority of prognostic signatures, our signature was enriched for differentiation-associated rather than proliferation-associated genes suggesting that it is reflecting a novel aspect of breast cancer biology. Experiments addressing the underlying biological and molecular mechanisms are ongoing. This signature will be tested for ability to predict response to TGF-beta antagonists in pre-clinical models, and will be exploited in gene expression based screens to find novel compounds that might reverse the switch and restore the tumor suppressor activities of TGF-beta. 3. DEVELOPMENT OF TOOLS TO IDENTIFY AND QUANTITATE NON-CANONICAL TGF-BETA SIGNALING IN HUMAN CLINICAL SAMPLES. Recent data from other labs has suggested that TGF-beta signaling via non-canonical mechanisms is associated with tumor promoting effects of TGF-beta, while canonical signaling is critical for tumor suppressor effects. Thus activation of non-canonical TGF-beta signaling may cause TGF-beta to switch from tumor suppressor to tumor promoter. Non-canonical signaling mechanisms include the formation of mixed Smad signaling complexes involving both TGF-beta Smads (Smad2/3) and BMP Smads (Smad1/5/8), and altered signaling due to phosphorylation of TGF-beta Smads in the middle linker region. In situ proximity ligation (Duolink) is a new technique that uses paired antibodies to detect and quantitate protein/protein interactions or complex post-translational modifications in frozen or fixed tissue sections. In FY11 we have made substantial progress towards developing this technique for the assessment of non-canonical TGF-beta signaling. We will apply this approach to look at the onset of non-canonical signaling at different stages in the progression process in pre-clinical and clinical breast cancer material and to assess whether non-canonical TGF-beta signaling might serve as a useful predictive biomarker for TGF-beta antagonist trials (see related project Z1A BC 010881).
Showing the most recent 10 out of 16 publications
