Multiple forms of TGF-beta 1 in breast tissues: a biologically active form of the small latent complex of TGF-beta 1

The aim of this study was to investigate whether breast cancer growth in vivo could be due to a failure in the activation of TGF-beta 1, a growth factor which has been shown to affect the development of normal breast tissue. Tissue samples of 40 breast carcinomas and the normal adjacent tissue from 37 (henceforth referred to as 'adjacent tissue'), as well as 13 specimens of benign lesions, were included in this study. The specimens were used in vitro to produce conditioned medium (CM), and this was examined for TGF-beta 1 activity by measuring growth inhibition of the mink lung epithelial cell line CCL-64. Immunoblotting and electrophoresis were used to detect the presence of TGF-beta 1 in CM and homogenised tissue samples. We demonstrated that the majority of TGF-beta 1 in breast cancer conditioned medium was biologically active, in direct contrast to CM prepared from benign disease specimens. Furthermore, active TGF-beta 1 was also identified in CM prepared from adjacent tissue, suggesting an important early role for this growth factor in the spread of this disease. Three distinct breast cancer related (BCR) molecular weight species of TGF-beta 1 (12.5/25 kDa, 50 kDa and 95 kDa) were identified. Both the 50 kDa and 95 kDa bands immunoprecipitated by an anti-TGF-beta 1 antibody were also immunoreactive with anti-TGF-beta 1 binding protein antibodies suggesting that the 50 kDa band may comprise at least part of the previously described small latent complex of TGF-beta 1. However, using the CCL-64 cell assay, we were able to demonstrate that the 50 kDa TGF-beta 1 BCR protein was biologically active whereas the large (95 kDa) TGF-beta 1 BCR latent complex protein was not. Adjacent tissue was more likely to contain the 50 kDa form than the tumour tissues (P < 0.05). Similarly, the 50 kDa molecule was also more common in patients who had oestrogen receptor (ER) negative tumours (compared with ER positive patients; P < 0.05) and in those who had received tamoxifen treatment prior to surgery (P < 0.01). In all of these cases, the increase in the incidence of the small active complex form was accompanied by a decrease in the incidence of the high molecular weight complex (95 kDa). We confirmed that, in vitro, the 95 kDa TGF-beta 1 BCR can be proteolytically cleaved to yield a 50 kDa TGF-beta 1 BCR. Finally, we observed a correlation between the presence of the 50 kDa complex protein and reduced levels of plasminogen activator (PA), which was significant in ER negative patients (P < 0.05) and tamoxifen-pretreated patients (P < 0.01). This suggests that the secretion of this active TGF-beta 1 protein may provide breast tumours with a mechanism whereby they can escape oestrogen dependence, and may provide an explanation for the common problem of tamoxifen resistance.