2075 Background: Clinical efficacy of ErbB1/ErbB2 kinase inhibitors is limited by the development of acquired autoresistance. Developing clinically relevant models to study the mechanism(s) of acquired autoresistance will identify therapeutic strategies to prevent its onset. Methods: BT474 cells, an ErbB2-overexpressing/ER+ breast cancer line sensitive to lapatinib-induced apoptosis were chronically exposed to lapatinib, simulating chronic administration of lapatinib in the clinic. Lapatinib-resistant cells (rBT474) and single cell clones were established. Baseline gene expression in parental BT474 and rBT474 cells was compared using Human Affymetrix oligonucleotide arrays. Sequential tumor biopsies were analyzed by quantitative IHC as part of a clinical trial using lapatinib monotherapy (1500mg/d) in patients with metastatic breast cancer. Results: Gene expression and protein analysis indicates that acquired resistance to lapatinib is mediated by increased estrogen receptor (ER) signaling rather than loss of ErbB2 expression or insensitivity of the ErbB2-MAPK-PI3K pathways to lapatinib. Activation of ER signaling results from de-repression of FOXO3a and transient up-regulation of caveolin-1, both facilitating ER transcriptional activity. Gene interference studies confirm that acquired resistance involves a switch in survival dependence and regulation of survivin from ErbB2 to ER. Since ErbB2 signaling is operative in rBT474 cells, treating parental BT474 cells with lapatinib plus (i) fulvestrant, or (ii) estrogen deprivation prevents the onset of acquired lapatinib resistance compared with either therapy alone. Importantly, increased ER signaling occurs in patients with ErbB2-overexpressing/ER+ breast cancers treated with lapatinib monotherapy, consistent with changes in cell lines. Conclusions: Chronic lapatinib monotherapy induces a shift in the dependence of cell survival from ErbB2 to ER. Combining lapatinib with selected anti-estrogens prevents its onset in cell lines and warrants clinical validation. Establishing models of acquired resistance simulating the clinic provides a tool to identify strategies to improve treatment options for patients with ErbB2-overexpressing breast cancers. [Table: see text]