The cancer stem cell hypothesis proposes that cancers arise in stem/progenitor cells through disregulation of self-renewal pathways generating tumors, which are driven by a component of 'tumor-initiating cells' retaining stem cell properties. The HER2 gene is amplified in 20-30% of human breast cancers and has been implicated in mammary tumorigenesis as well as in mediating aggressive tumor growth and metastasis. We demonstrate that HER2 overexpression drives mammary carcinogenesis, tumor growth and invasion through its effects on normal and malignant mammary stem cells. HER2 overexpression in normal mammary epithelial cells (NMEC) increases the proportion of stem/progenitor cells as demonstrated by in vitro mammosphere assays and the expression of stem cell marker aldehyde dehydrogenase (ALDH) as well as by generation of hyperplastic lesions in humanized fat pads of NOD (nucleotide-binding oligomerization domain)/SCID (severe combined immunodeficient) mice. Overexpression of HER2 in a series of breast carcinoma cell lines increases the ALDH-expressing 'cancer stem cell' population which displays increased expression of stem cell regulatory genes, increased invasion in vitro and increased tumorigenesis in NOD/SCID mice. The effects of HER2 overexpression on breast cancer stem cells are blocked by trastuzumab in sensitive, but not resistant, cell lines, an effect mediated by the PI3-kinase/Akt pathway. These studies provide support for the cancer stem cell hypothesis by suggesting that the effects of HER2 amplification on carcinogenesis, tumorigenesis and invasion may be due to its effects on normal and malignant mammary stem/progenitor cells. Furthermore, the clinical efficacy of trastuzumab may relate to its ability to target the cancer stem cell population in HER2-amplified tumors.