We have developed a cell model of airway smooth muscle hypertrophy in which primary human bronchial smooth cells are treated with TGF-β. We have found that, within 48 h of treatment, TGF-β increases cell size and protein synthesis, protein abundance of α-smooth muscle actin and smMHC, formation of actomyosin filaments, and shortening in response to ACh. This model holds advantages over two previous models of airway smooth muscle hypertrophy, prolonged serum deprivation of canine tracheal myocytes (4
) and human bronchial smooth muscle cell lines transduced with a temperature-sensitive large T antigen (7
). In both models, cell cycle arrest, induced either by serum withdrawal or degradation of large T with consequent release of p53 and induction of the cyclin-dependent kinase inhibitor p21Cip1
, induces the formation of large myocytes with increased abundance of contractile proteins. However, since airway smooth muscle cells are apparently quiescent in vivo
, as evidenced by the absence of Ki67 staining (2
), the physiologic relevance of these models is unclear. Also, in both models, the precise stimulus for airway smooth muscle growth is not known. In contrast, TGF-β is not only known to be increased in the airways of patients with asthma, but expression is relatively higher in those with severe asthma compared with those with less severe disease (8
), consistent with recent reports demonstrating airway smooth muscle hypertrophy in individuals with severe (2
) but not mild asthma (3
The hypertrophic response to TGF-β is not unexpected, as similar results have been shown in vascular smooth muscle cells and cardiac myocytes, and the effects of TGF-β on the transcription of contractile protein genes is well established (13
). In the present manuscript, we confirmed that TGF-β increases steady state levels of α-smooth muscle actin mRNA. However, we reasoned that, to induce hypertrophy, TGF-β must also enhance the efficiency of mRNA translation. We based this hypothesis on the following information. First, selective alterations in the transcription rates of individual genes or groups of genes are unlikely to be sufficient to account for the global increases in protein synthesis observed during hypertrophy. Second, in the intact heart, the availability of mRNA is not generally limiting for increases in overall cell protein synthesis during growth (32
). Finally, translational control mechanisms have been shown to contribute to both cardiac (33
) and skeletal muscle hypertrophy in vivo
), not to mention our previous cell culture models of airway smooth muscle hypertrophy (4
). We found that TGF-β had no effect on α-actin mRNA stability, and pilot pulse-chase studies have shown no degradation of radiolabeled α-actin protein 48 h after withdrawal of hot probe (not shown), suggesting that protein degradation does not play a role. On the other hand, TGF-β increased α-smooth muscle synthesis in the presence of actinomycin D, an inhibitor of gene transcription, demonstrating that TGF-β increases the translation of α-actin mRNA into protein.
As noted above, there are three general mechanisms by which TGF-β might increase translation. First, translation of the majority of eukaryotic mRNAs is initiated through a 7-methylguanosine cap structure at the 5′ end of mRNA. The cap is recognized and “clamped” by eIF4E, which in turn associates with inhibitory 4E-BPs. 4E-BP1 undergoes phosphorylation at multiple sites, which results in its release from eIF4E, thereby increasing the availability of eIF4E for binding to eIF4G, eIF4F complex formation, and cap-dependent translation (21
). Second, concurrent with the preparation of mRNA, the preinitiation complex must be formed. eIF2, a multimer consisting of α, β, and γ subunits, functions to recruit methionyl tRNA and conduct it as a tRNA-eIF2-GTP ternary complex to the 40S ribosomal subunit, to form the 43S preinitiation complex. Finally, the translation of mRNAs with 5′ TOP tracts, many of which encode elongation factors and ribosomal proteins involved in mRNA translation, is upregulated by successive phosphorylation of mTOR, S6K-1, and the S6 ribosomal protein.
In the present study, we found that treatment with TGF-β induced phosphorylation of 4E-BP. Further, inhibition of 4E-BP phosphorylation by a chemical inhibitor of PI 3-kinase, LY294002, attenuated TGF-β–induced airway smooth muscle cell enlargement and α-smooth muscle actin expression. While these data suggest that 4E-BP phosphorylation is required for hypertrophy in this context, these experiments must be viewed with caution, not only because of the potential nonspecific effects of a chemical inhibitor, but also because correlations between phenotypic change and dephosphorylation of 4E-BP do not prove causality. Indeed, it is conceivable that LY294002 blocks TGF-β–induced phenotypic change by blocking transcription as well as translation. To directly test the hypothesis that 4E-BP phosphorylation is required for TGF-β–induced hypertrophy, we expressed AA-4E-BP1, a nonphosphorylatable mutant of 4E-BP1 that dominantly binds to eIF4E, in primary human bronchial smooth muscle cells by retroviral infection. AA-4E-BP1 also inhibited TGF-β–induced airway smooth muscle hypertrophy, demonstrating that 4E-BP phosphorylation, a prerequisite for eIF4E release and cap-dependent translation, is required for this process. The requirement of 4E-BP phosphorylation and cap-dependent translation initiation in the regulation of cell size has been demonstrated only once previously, in human U2OS osteosarcoma cells expressing the same phosphorylation site-defective mutant of 4EBP1 (37
). In the latter study, the 4E-BP mutant blocked the ability of eIF4E overexpression to increase cell growth.
On the other hand, we found that rapamycin, an inhibitor of mTOR, had only modest effects on TGF-β–induced 4E-BP phosphorylation and phenotypic change, despite inhibiting phosphorylation of p70 S6 kinase. These data suggest that, in TGF-β– treated cells, 4E-BP phosphorylation may occur in a PI 3-kinase–dependent, mTOR-independent manner. As noted above, it has recently been shown that PI 3-kinase may directly phosphorylate 4E-BP in vitro
). Further, since TGF-β–induced hypertrophy occurred in the presence of p70 S6 kinase dephosphorylation, these data suggest that the upregulation of 5′ TOP mRNAs is not required for the observed hypertrophic response. These data vary from those obtained in serum-deprived canine tracheal myocytes, in which rapamycin blocks S6K1 phosphorylation and phenotypic change (6
). On the other hand, these data are consistent with a recent report in which deletion of S6 kinase failed to attenuate insulin-like growth factor-1 or PI 3-kinase–mediated cardiac hypertrophy (38
In conclusion, we have developed a new model of airway smooth muscle hypertrophy in which primary cells are treated with TGF-β. TGF-β increased cell size and total protein synthesis, expression of α-smooth muscle actin and smMHC, formation of actomyosin filaments, and cell shortening to ACh. Further, in addition to transcriptional effects, TGF-β increased the translation of α-actin mRNA into protein. Finally, phosphorylation of 4E-BP is required for the TGF-β–induced hypertrophy, suggesting that eIF4E-, cap-dependent translation is necessary for this process. Further studies examining the translational control pathways regulating airway smooth muscle hypertrophy may provide new insight into the pathogenesis of severe asthma, and lead to new therapeutic interventions.