Abstract. An ideal mouse model should closely mimic a clinical situation. However, for most models available, this is not the case since clinical trials frequently fail to reproduce the highly encouraging therapeutic results obtained from preclinical studies performed using mouse models. In this study, in the process of extending the application of our previously established breast tissue-derived orthotopic and metastatic (BOM) model, the human breast cancer cell line MDA-MB-231 failed to exhibit any osteotropic features that differed from previous studies. Our observations suggest that a human tissue-specific microenvironment could be an essential requirement for a successful mouse model of breast cancer. Here, multiple in vivo breast cancer models were used to confirm this hypothesis. Human breast tissue and cancellated bone were transplanted subcutaneously to female severe combined immunodeficiency disease (SCID) mice by different assemblies, to build several mouse models termed 'breast-breast', 'breastbone', 'bone-bone', 'MFP (mouse mammary fat pad)-bone', and 'MFP-breast' models. Two human breast cancer cell lines, MDA-MB-231 and MDA-MB-231BO, and the mouse breast cancer cell line TM40D were used. All cancer cells were labeled with GFP for gross observation. In addition, transplanted human tissues and various mouse tissues including bone, lung, liver, mesentery were examined microscopically. Based on morphological, immunohistochemical, and enzymohistochemical evidence obtained from several comparative experiments in 'breast-breast', 'breast-bone' and 'bone-bone' models, the BOM model was proved to be feasible and reliable. The organ tropism of the breast cancer cell line, which was derived from a mouse model by intracardiac inoculation in a pure mouse microenvironment, was reconsidered. The behavior of breast cancer cells in the mouse model was altered in response to the varying microenvironment. The results in this study suggest the human tissue-specific microenvironment is most likely an essential requirement in mouse models of breast cancer.