The etiology of human breast cancer is thought to involve a complex interplay of genetic, hormonal, and dietary factors that are superimposed on the physiological status of the host. Attempts to derive a cohesive picture of how these factors participate in the etiology of breast cancer have been confounded by a lack of information on specific mutations associated with the initiation and progression of the disease. We have undertaken a strategy that is aimed at determining, on a molecular level, those genetic alterations in primary breast tumor DNA that have a statistically significant association of patient history, tumor characteristics, and patient postsurgical prognosis. The int-2 gene is frequently activated in mouse mammary tumors by the mouse mammary tumor virus (MMTV) via insertional mutagenesis. We previously found that int-2 is amplified 2- to 15-fold in 14% of primary human breast tumor DNAs (n = 118) and has a significant association (p < 2 x 10[-6]) with a poor prognosis for the patient. We have now defined the amplification unit on chromosome 11q13 and found that 17 of 18 tumors amplified for int-2 were also amplified for hst and bcl-1. The hst and int-2 genes are both related to basic fibroblast growth factor. The bcl-1 locus defines the translocation t(11;14) breakpoint in B-cell chronic lymphocytic leukemias. In a lymph node metastasis of a breast cancer patient, int-2 but not hst was expressed. In 10% of the primary breast tumor DNAs (n = 122) the c-erbB-2 gene was amplified. The c-erbB-2 gene is related to the gene encoding the epidermal growth factor receptor. However, in contrast to one published report, we found no significant association with patient postsurgical prognosis or other aspects of patient history and tumor characteristics. In other studies, we detected the expression of MMTV-related human RNA sequences in breast tumor-derived cell lines and have obtained recombinant cDNA clones of these sequences for further analysis.