Abstract The development of breast cancer resistance to endocrine therapy results from altered cellular plasticity leading to the emergence of a hormone independent tumour. We have previously shown that the endocrine resistant phenotype is poorly differentiated and has greater metastatic potential when compared with an endocrine sensitive phenotype. The underlying mechanism of how an ER positive, endocrine sensitive tumour can turn on these adaptive mechanisms remains unresolved. We hypothesise that cellular reprogramming can occur as a result of long-term exposure to endocrine treatment which manifests clinically as tumour recurrence. Our aim was to identify the mechanism of breast cancer cellular reprogramming in a clinical context. In this study we adopted a global approach to define transcriptional networks significantly elevated in the breast cancer metastatic setting. Using RNAseq we determined the gene expression profile of three ER positive patients who developed tumour recurrence on tamoxifen treatment. RNAseq was performed on primary tumours, axillary node metastases and subsequent distant metastases. Following bioinformatic analysis, we defined the genes that were significantly elevated in the metastatic tumours. In the metastases, genes clustered away from those in the primary and node, with 209 genes uniquely elevated in the metastatic context, suggesting altered gene expression in response to endocrine therapy. Pathway analysis of genes significantly elevated in the metastatic setting included developmental pathways (WNT signalling and TGFbeta), growth factor signalling pathways (MAPkinase), cell adhesion molecules, junction adherens and pathways in cancer. Further analysis provided a list of clinically relevant transcriptional networks which may facilitate tumour cell de-differentiation. Transcription factors including Myc, SMAD2, ESR, TCF3, CUX1 and CLOCK were identified which potentially drive reprogramming in response to endocrine treatment. We interrogated this data to identify downstream targets of this ‘de-differentiation’ network for new predictive markers of response to endocrine treatment. Bioinformatic analysis identified PAWR, an inhibitor of progenitor cell expansion as a potential SMAD2 target. In a xenograft model of endocrine resistance, PAWR expression was found to be reduced in metastatic tissue from lung, liver and bone compared with the primary tumour. In human breast cancer patients PAWR significantly associated with a good response to endocrine treatment and luminal A status (p = 0.0008), establishing PAWR as a predictor of good response to endocrine therapies. We have identified a transcriptional network which may drive a de-differentiated phenotype following endocrine treatment. Data presented here proposes a mechanism by which apparently less aggressive Luminal A type tumours may fail endocrine treatment and develop subsequent recurrence. This represents a fundamental shift in how we approach the development of resistance to endocrine therapy, establishing a number of biomarkers which will allow tracking of response to endocrine therapy or allow accurate early prediction of those who will respond to treatment. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P6-04-01.