Based largely on studies of cell lines in vitro and of transgenic mouse models, disruptions of transforming growth factor (TGF) beta signaling are thought to contribute to the development and progression of human breast cancer. However, whether and how TGF-beta signaling becomes disrupted during human breast cancer development in vivo remains largely unknown. To address this question, we have compared the patterns of expression and activation of the postreceptor components of the TGF-beta signaling pathway, the so-called Smads, in human breast cancer cell lines with those in breast carcinoma specimens. None of the breast carcinoma cell lines were growth arrested by TGF-beta in vitro. Each of the tumor cell lines expressed normal levels of Smad2 and -3. Moreover, TGF-beta treatment induced phosphorylation of Smad2 (Smad2P) in each of these lines, except those that lacked TGF-beta type II receptors. Moreover, only one of the cell lines failed to express Smad4. Among 456 cases of human breast carcinoma assembled in tissue microarrays, the majority (92%) expressed Smad2, Smad2P, as well as Smad4, indicating their ability to proliferate within a microenvironment that contains bioactive TGF-beta. Thirty cases (6.6%) failed to express Smad2P, suggesting the loss of TGF-beta receptor signaling. Nine cases (2%) failed to express Smad4, and 3 of these also failed to express Smad2P. Thus, the phenotypes of breast tumors in vivo paralleled that of human breast cancer cell lines in terms of Smad2P and Smad4 expression. Loss of Smad signaling was not associated with any particular histological subtype, histological or nuclear grade, estrogen- or progesterone receptor expression, or HER2/neu expression. Loss of Smad4 was inversely correlated with the presence of axillary lymph node metastases. Most importantly, among patients with stage II breast cancer, lack of Smad2P expression in the tumor was strongly associated with shorter overall survival. Finally, analysis of a small cohort of hereditary breast cancers failed to reveal any association between BRCA1 or BRCA2 genotype and alterations in Smad signaling.