Abstract Background Trastuzumab-based therapies are the therapeutic option for HER2 positive (HER2+) breast cancer. HER2 amplification is the only biomarker validated for trastuzumab-based therapies. However, a proportion of tumors become refractory during treatment course. For this reason, the finding of new biomarkers beyond HER2 overexpression to identify patients who would benefit most from trastuzumab regimens is of outstanding importance. Methods Models of trastuzumab-resistant or hypersensitive cells were generated by exposure to trastuzumab. Cell surface, total HER2, and analyses of proteins involved in cell cycle or apoptosis were analyzed by western blotting. Cell proliferation was analyzed by cell counting, cell cycle and apoptosis was evaluated by FACS. Transcriptomic characterization of the cellular models was performed using bioinformatic online tools, and clinico-genomic analyses were performed using the PAMELA clinical trial data. Results Besides differing in sensitivities to trastuzumab, the different cellular models also showed distinct response to other anti-HER2 drugs (lapatinib, neratinib, pertuzumab and T-DM1) used in the clinic. That differential effect was not due to changes in cell surface, total or activated HER2. Trastuzumab caused important induction of cell death in hypersensitive cells but not in parental or resistant cells. Transcriptomic analyses of these cellular models together with querying of online databases allowed the identification of individual genes and gene signatures that predicted prognosis and trastuzumab response in HER2+ breast cancer patients. Conclusion The identification of trastuzumab response biomarkers may be used to select patients particularly sensitive to facilitate the use of trastuzumab-based therapies and refine follow-up guidelines in patients with HER2+ tumors.