We have assessed ®ve signal transduction pathways to determine the role each might play in the malignant transformation of mammary epithelium initiated by neu, heregulin/NDF, TGFa, v-Ha-ras and c-myc in transgenic mice. The study involves a molecular and pharmacologic assessment of Erk/MAP kinase, Jnk/SAP kinase, PI 3-kinase, protein kinase C, and the Src-related kinases Lck and Fyn. Our results indicate that oncogenes capable of transforming mammary gland epithelium activate and require speci®c signal transduction pathways. For example, mammary tumors initiated by neu, vHa-ras and c-myc have high levels of active Erk/MAP kinase and their anchorage independent growth is strongly inhibited by PD098059, an inhibitor of Mek/ MAP kinase kinase. By contrast, Erk/MAP kinase activity is weak in tumors initiated by TFGa and heregulin/NDF and the corresponding cell lines are not growth inhibited by PD098059. Similarly, PI 3-kinase is strongly activated in neu, TGFa and heregulin/NDF initiated tumor cell lines, but not in c-myc or v-Ha-ras initiated tumor cell lines. The anchorage independent growth of all these tumor cell lines are, however, inhibited by the speci®c PI 3-kinase inhibitor LY294001. Further illustrating this oncogene-based speci®city, PP1, a speci®c inhibitor of the Src-like kinases, Lck and Fyn, blocks anchorage-independent cell growth only in the TGFa initiated mammary tumor cell line. Taken together with additional observations, we conclude that certain oncogenes reliably require the recruitment/activation of speci®c signal transduction pathways. Such speci®c relationships between the initiating oncogene and a required pathway may re¯ect a direct activating e ect or the parallel activation of a pathway that is a necessary oncogenic collaborator for transformation in the mammary gland. The work points to a molecular basis for targeting therapy when an initiating oncogene can be implicated; for example, because of ampli®cation, increased expression, genetic alteration, or heritable characteristics.