TGF-beta1-induced Smad 3 binding to the Smad 7 gene: knockout of Smad 7 gene transcription by sense phosphorothioate oligos, autoregulation, and effect on TGF-beta1 secretion: bleomycin acts through TGF-beta1.

Bleomycin produces its fibrogenic effect, at least in part, by TGF-beta1 secretion. Treatment of IMR-90 human embryonic lung fibroblasts with bleomycin at 0.5 microg/ml results in a 1.6-fold increase of TGF-beta1 as determined by a specific ELISA assay for TGF-beta1 after acidification of the conditioned media. This elevation of TGF-beta1 secretion is furthermore enhanced in vivo by TGF-beta1 autoinduction of the TGF-beta1 gene. To demonstrate TGF-beta1 autoinduction, the fibroblasts were pretreated with 12.5 ng/ml TGF-beta1, washed extensively to remove any residual TGF-beta1, and then allowed to incubate for 24 h in AIM V synthetic serum-free media. The media when assayed using the ELISA assay contained a 1.6-fold increase of TGF-beta1. The distal promoter of the human TGF-beta1 gene contains a Smad 3 element (CAGGACA), which is homologous to the Smad 3 binding element motif (CAGA). The nuclear extracts of human embryonic lung fibroblasts treated for either 15 min or 24 h with TGF-beta1 did not demonstrate specificity of binding of a protein(s) to the homologous Smad 3 element as determined by cold wild-type oligodeoxynucleotide competition experiments. However, specific Smad 3 binding to the Smad 3 element (GTCTAGAC) found in proximal promoter of the Smad 7 gene was observed by cold oligo competition and supershift assays using a goat polyclonal Smad 3 antibody in the presence and absence of an N-terminal Smad 3 peptide. To determine the functionality of this Smad 3 binding to the Smad 3 element in the proximal promoter of the Smad 7 inhibitory gene to TGF-beta1 secretion, fibroblasts were transiently pretransfected with double-stranded phosphorothioate oligo "decoys" containing the Smad 7/Smad 3 element in the presence of plasmin to convert latent TGF-beta1 to active TGF-beta1. Under these conditions, which simulate the in vivo situation of 2.2-fold increase of total active TGF-beta1 was observed. Fibroblasts were also pretransfected with these double-stranded oligo "decoys," washed, then treated with TGF-beta1, washed and incubated in AIM V for an additional 24 h. In this latter experiment, a superinduction of TGF-beta1 secretion was observed. We propose that these oligo "decoys" bind Smad 3 preventing this initiation factor from binding to the Smad 7/Smad 3 element thereby decreasing the transcription of the Smad 7 gene. The decrease of the inhibitory Smad 7 would result in less binding of this Smad inhibitor to the Type I TGF-beta receptor and less antagonism of active TGF-beta1, more autoinduction of the TGF-beta1 gene, and more of the fibrogenic effects of TGF-beta1.