Preclinical models for the evaluation of targeted therapies of metastatic disease.

It has been estimated that approx 60-70% of cancer patients harbor overt or subclinical metastases at diagnosis, and it is the eradication of such systemic disease that largely determines survival. Preclinical tumor model systems employed to evaluate potential new treatment strategies should aim to represent the process and patterns of metastasis of their clinical counterparts as closely as possible. Severe combined immune-deficient (SCID) and nu/nu mice have been extensively used as hosts for the growth of human tumor cell lines and in some cases fresh tumor material. However, in most instances the resulting neoplasms fail to metastasize, and the aberrant immune systems of such animals has limited their use mainly to passive therapies of localized disease. Recently, the development of specially selected tumor variants and the use of appropriate orthotopic sites for implantation has provided several models in which dissemination can be demonstrated. Where the gene coding for a potential target antigen has been cloned, and where its overexpression or mutation is associated with malignancy (e.g., c-erbB-2, H-ras), transgenic mice may yield tumors that will develop in these immunocompetent hosts. In some cases such tumors exhibit metastasis. A third approach is to transfect human genes of interest into appropriate rodent tumors expressing the desired metastatic phenotype. These various approaches are compared with particular reference to mammary carcinoma biology.