This study shows that galectin-3 is an essential mediator of TGF-β–induced lung fibrosis. This was manifest by reduced myofibroblast activation and collagen production and reduced TGF-β1–induced EMT of galectin-3−/− AECs. Galectin-3 deletion reduced phosphorylation and nuclear translocation of β-catenin but had no effect on Smad2/3 phosphorylation. A novel inhibitor of galectin-3, TD139, blocked TGF-β–induced β-catenin activation in vitro and in vivo and attenuated the late-stage progression of lung fibrosis after bleomycin. Moreover, patients with stable IPF had elevated levels of galectin-3 in the BAL fluid and serum compared with patients with NSIP and control subjects, and this rose sharply during an acute exacerbation suggesting that galectin-3 may be a marker of active fibrosis in IPF.
There is increasing evidence that EMT may be a major source of pathogenic myofibroblasts during pulmonary fibrogenesis and contributes to the formation of fibroblastic foci in mice and humans (25
). Mice expressing β-galactosidase exclusively in lung epithelial cells express mesenchymal markers after TGF-β1 expression in vivo
). We show that TGF-β1−induced EMT in primary AECs is also dependent on galectin-3. It is important to distinguish between factors that induce EMT rather than those that stimulate the growth of contaminating mesenchymal cells or promote the death of epithelial cells. Our results indicate that there is no significant difference in the proliferation of WT or galectin-3−/−
fibroblasts and no evidence of increased cell death between WT and galectin-3−/−
Fibrocytes express mesenchymal and hematopoietic markers (30
) and are elevated in the blood of patients during an acute fibrotic exacerbation of IPF (32
) and have also been found in IPF lung tissue (33
). We found that bleomycin-induced lung injury resulted in a marked increase in fibrocyte recruitment to the damaged lung; however, we found no difference in fibrocyte recruitment between WT and galectin-3−/−
mice (Figure E5). Taken together our results suggest that galectin-3 regulates TGF-β1–mediated EMT and myofibroblast activation rather than affecting fibroblast numbers or fibrocyte recruitment.
Our results suggest that reducing galectin-3 at the cell surface reduces the cell surface expression of TGFβR without affecting the total expression of TGFβR or receptor affinity for TGF-β1. This is most likely caused by reduced cell surface receptor retention as a result of loss of galectin-3 binding to polylactosamine residues on TGF-β receptors (34
). TGF-β1 signals predominantly by Smad-dependent pathways and Smad3-deficient mice are protected from TGF-β1–induced fibrosis (35
). However, there is evidence for TGF-1–induced activation of non–Smad-dependent signaling in EMT (36
), in particular, interactions between TGF-β1 and the β-catenin pathway (38
). In response to TGF-β1, β-catenin is liberated from the E-cadherin adherens junctions and translocates to the nucleus where it mediates activation of transcription factors promoting collagen transcription (39
). However, the predominant pathway involved in TGF-β1–mediated EMT seems to be highly cell type and context dependent (36
). Mice lacking the α3β1 integrin show full phosphorylation of Smad2 but a reduced interaction of Smad2 with phosphorylated β-catenin resulting in reduced TGF-β1–mediated EMT and fibrosis (41
). In MDCKII cells loss of tight junctions was Smad independent, whereas complete loss of E-cadherin and transformation to a mesenchymal phenotype were dependent on Smad signaling (43
). The role of these non-Smad pathways during EMT in the lung is largely unknown. However, the Wnt/β-catenin signaling pathway is aberrantly activated in IPF and nuclear β-catenin localization is observed in cells forming fibroblast foci (44
). β-catenin and TGF-β1 can independently or cooperatively regulate target gene transcription, which play an important role in EMT (39
). Our results demonstrate that galectin-3 does not affect Smad3 or Smad2 activation or Smad2 association with pTyr654–β-catenin but regulates TGF-β1–induced EMT by cooperative regulation of the Wnt signaling pathway, resulting in activation and nuclear translocation of β-catenin by an inhibition of GSK-3β phosphorylation and activity. The increased basal β-catenin activation in WT AECs compared with galectin-3−/−
cells is most likely a result of spontaneous EMT observed in WT cells in culture probably caused by activation of AECs plated on the collagen-fibronetric matrix (42
). However crucially, we saw no difference in basal β-catenin activation in cells treated with exogenous recombinant galectin-3 and no difference in control-treated WT and galectin-3−/−
mice in vivo
suggesting that there is no real difference in basal activation in vivo
We suggest that although the Smad pathway is necessary it is not sufficient to induce EMT in lung AECs. A recent study by Li and coworkers (46
) highlights the importance of lung epithelial cell TGFR-II expression in driving EMT and fibrosis after bleomycin. Interestingly, in this study deletion of TGFR-II did not fully block TGF-β1–induced Smad signaling, which could suggest additional non-Smad pathways are necessary for EMT and fibrosis to occur. This has parallels with the present study, which shows that reduced surface expression of TGFR-II (in the absence of galectin-3) allows Smad signaling but prevents EMT and fibrosis. We propose that TGF-β1 increases galectin-3 expression in the fibrotic lung, which stimulates EMT and myofibroblast differentiation. By anchoring TGF receptors at the cell surface, galectin-3 may provide an optimal framework that allows the receptors to signal by the accessory pathways necessary for full EMT to occur (Figure E6). Although the mechanisms of this effect have yet to be defined, differential internalization of TGF-β receptors is thought to be important for regulating the duration and directionality of signaling (47
), and that undefined regulatory mechanisms exist that direct sequestration into different endocytic compartments, which can either promote Smad signaling or induce receptor degradation (48
). The Snail family of transcription factors (SNAI1 and SNAI2) is induced by TGF-β by Smad and non-Smad pathways and function to inhibit E-cadherin transcription leading to the development of EMT (13
). The effect of galectin-3 on the expression and function of these transcription factors requires further study.
Galectin-3 is markedly up-regulated in fibroproliferative areas in the lung of patients with UIP. Serum galectin-3 concentration is stable over time, showing little variation during the stable phase of UIP but during an acute exacerbation, serum galectin-3 levels rise significantly. Thus, our observations in patients mirror those seen in mice where galectin-3 expression correlates with the level of active fibrosis. Our results suggest that serum galectin-3 levels may help distinguish UIP from NSIP clinically and identify patients undergoing an acute exacerbation. This requires further study in a larger patient cohort.
The bleomycin model of fibrosis is widely used as a model of human IPF (49
) and as a screen to evaluate novel antifibrotic drugs for clinical use. As with Ad–TGF-β, galectin-3−/−
mice were protected from the profibrotic effects of bleomycin. In screening for antifibrotic drugs it is critical to distinguish between potential antiinflammatory and antifibrotic effects because preventing progression of fibrosis has more clinical relevance (50
). We administered the galectin-3 inhibitor TD139 during the fibrotic phase of bleomycin-induced lung injury, which fully blocked the progression of fibrosis. TD139 is a novel synthetic inhibitor of galectin-3. TD139 has high affinity for galectin-3 with a Kd
14 nM and galectin-1 Kd
10 nM, but low affinity for galectins 2, 4N, 4C, 7, 8N, or 9N. In contrast to galectin-3, which is associated with chronic inflammation, the in vivo
administration of galectin-1 prevents the development of chronic inflammation and impairs the ongoing disease in a number of experimental models of autoimmune diseases (53
). Galectin-1 has been shown to suppress collagen expression and renal fibrosis (57
). Thus, the antifibrotic effects of TD139 are most likely caused by its blocking galectin-3 function.
Our results show that blocking galectin-3 function is both preventative and therapeutic in reducing lung fibrosis, suggesting that galectin-3 inhibition is an exciting novel therapeutic target to treat patients with IPF. In addition, TD139 may be a compound for further drug development for treatment of lung fibrosis.