Abstract 1090: Development of predictive biomarkers for anti-TGF-beta therapy

TGF-betas are pleiotropic growth factors with complex roles in tumorigenesis. The TGF-betas have many activities that could promote metastasis, including the enhancement of invasion and migration, the promotion of angiogenesis, and the suppression of immune surveillance. Overexpression of TGF-beta in many advanced human tumors correlates with metastasis and poor prognosis, and there is now considerable interest in developing TGF-beta antagonists for cancer therapy. Indeed, a Phase I oncology trial using a fully humanized anti-TGF-beta monoclonal antibody was recently successfully completed. There is now an urgent need for good predictive biomarkers for patient stratification in future clinical trials. Using different syngeneic mouse transplant models of metastatic breast cancer, we have found that anti-TGF-beta antibodies suppress metastasis in several models, including the widely-used 4T1 model. However, TGF-beta antibody treatment has no effect or even stimulates metastasis in other models. This panel of mouse models with differing responses to anti-TGF-beta therapy represents a powerful experimental platform for the identification of predictive biomarkers. Based on existing knowledge, we have assessed a number of plausible candidates. Global gene expression analyses have defined several distinct molecular subtypes of breast cancer with different clinical characteristics. The “Claudin-low” subtype is a particularly poor prognosis subset of the triple negative ER-PR-HER2- breast cancers. By definition these tumors have low claudin and E-cadherin expression and a mesenchymal phenotype. Since TGF-betas are strong inducers of the mesenchymal phenotype, the claudin-low subtype may represent a particularly good target for anti-TGF-beta therapy. Similarly, overexpression of the homeobox protein Six-1 has recently been shown to cause TGF-beta to switch from tumor suppressor to pro-metastatic factor in breast cancer and might also be a predictor of a positive therapeutic response to TGF-beta antagonism. Here we have assessed the claudin-low status of a number of mouse mammary tumor models, by RTQ-PCR for key markers, using cells cultured in vitro. In addition, we determined nuclear Six1 levels by Western blot. Neither claudin-low status nor Six1 expression were predictive of response to anti-TGF-beta therapy. We are assessing additional candidate biomarkers and moving forward with discovery approaches. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 1090. doi:10.1158/1538-7445.AM2011-1090