Discussion
In this study we investigated the in vitro effect of nintedanib on the proliferative capacity as well as on the extracellular matrix metabolism of primary human lung fibroblasts obtained from patients with IPF and from non-fibrotic control lungs. We found that the receptor tyrosine kinase inhibitor nintedanib prevented growth-factor induced proliferation of both IPF and control fibroblasts. Furthermore, growth-factor induced collagen synthesis was significantly reduced by nintedanib. Interestingly, the drug up-regulated pro-MMP-2 secretion, but down-regulated its inhibitor TIMP-2.
IPF is a progressive lung disease with a median survival time after diagnosis of 2.5 to 3.5 years, and limited therapeutic options. The pathomechanisms that lead to IPF are not fully understood, however, they ultimately result in enhanced numbers of mesenchymal cells and the accumulation of ECM components in the interstitium. Nintedanib, an intracellular inhibitor of tyrosine kinases, has proven good antitumor/antiproliferative efficacy, and had beneficial effects in patients with IPF. Nintedanib's targets include PDGFR, FGFR, and VEGFR, which have been shown to be involved in lung fibrosis.
Using primary human lung fibroblasts obtained from IPF lungs and from non-fibrotic control lungs, we found that IPF cells expressed higher levels of PDGFR and FGFR compared to controls. This finding is in line with previous data showing an enhanced expression of FGFR1 on myofibroblast-like cells in IPF patients, and up-regulated PDGFRα expression in a rodent model of lung fibrosis. Whereas we found similar levels of VEGFR in fibrotic and non-fibrotic cells, others reported an increased expression of VEGFR1 in IPF patients. However, in the latter study soluble VEGFR was measured in bronchoalveolar lavage fluid, whereas we used protein extracts from pure primary human lung fibroblasts, which makes it difficult to compare the data.
PDGF-BB and bFGF are potent mitogenic and chemotactic factors for mesenchymal cells and both factors were shown to be elevated in IPF patients. Accordingly, we observed a strong mitogenic effect of PDGF-BB and bFGF in primary human lung fibroblasts from IPF and non-fibrotic control lungs. Importantly, the observed mitogenic effect was stronger in the control cells than in IPF cells (2-fold for PDGF-BB, 1.7-fold for bFGF), which is in line with previous data demonstrating a different response to growth factors by normal and IPF fibroblasts. VEGF is an important regulator of neovascularization, and recent data supported the hypothesis that it might be involved in fibrotic diseases. In this regard, marked vascular changes were reported in an animal model of lung fibrosis. Currently, the role of VEGF in IPF is not clear, since VEGF plasma levels were significantly related to radiologic fibrosis scores in patients with idiopathic interstitial pneumonias, but decreased VEGF levels in broncho-alveolar lavage fluid from patients with IPF have been reported. Even though VEGF is considered to be a specific mitogen for endothelial cells, we observed a mitogenic effect of VEGF in IPF, as well as in non-fibrotic control fibroblasts, reaching statistical significance only in the latter ones.
In accordance with previous data, nintedanib inhibited the PDGF-, FGF-, and VEGF-induced fibroblast proliferation in a concentration dependent manner in IPF as well as in non-fibrotic control cells. In our primary human fibroblasts, we observed that the antagonistic capacity of nintedanib was significantly stronger in non-fibrotic cells compared to IPF fibroblasts. This reduced sensitivity of IPF cells towards nintedanib might be explained by the higher PDGFR and FGFR expression in IPF cells. Furthermore, in contrast to our findings on collagen secretion, where only high concentrations of nintedanib prevented growth-factor induced collagen secretion, cell proliferation was potently inhibited by low concentrations of nintedanib.
The excessive accumulation of ECM is a hallmark of IPF, and therefore we studied the effect of nintedanib on ECM metabolism. We found that nintedanib induced the activity and protein levels of MMP-2, and reduced the levels of TIMP-2. MMPs are a family of secreted zinc-containing endopeptidases that degrade proteins of the ECM, and TIMP are their main physiological inhibitors. Both MMP-2 and MMP-9 metabolize various ECM proteins including type IV and type V collagens and gelatin. MMPs are secreted as inactive precursors which have to be activated in the extracellular space. MMP-2 can be activated by type I collagen and thrombin, but also by a membrane-type MMP-dependent pathway involving TIMP-2. An imbalance between MMP and TIMP might be involved in the accumulation of ECM in fibrogenesis. Accordingly, a greater increase in the levels of TIMP than levels of MMP-2 was reported and such an imbalance would favour the enhanced deposition of ECM proteins. Interestingly, TGF-β induced lung fibrosis in mice was primarily due to TIMP up-regulation, and a recent study demonstrated that over-expression of MMP-9 by alveolar macrophages in mice attenuated the fibrotic reaction after bleomycin instillation. Therefore, if TIMP are elevated in IPF, a drug that reduces the secretion of TIMP-2 by fibroblasts is possibly a powerful tool in the therapy of this devastating disease.
To further investigate nintedanib's anti-fibrotic potential, we determined the drug's effect on PDGF-, FGF-, and VEGF-induced collagen secretion. Similar to others we observed no significant modulation of collagen secretion upon stimulation with the three growth-factors in fibroblasts.
TGF-β is a well-known stimulator of collagen production, a pivotal mediator of fibrogenesis, and TGF-β-induced disturbances are critical in IPF. As expected, TGF-β up-regulated total collagens in IPF and control fibroblasts, and this effect was antagonized by the highest concentration of nintedanib (1 μM). To elucidate the underlying mechanism of this effect, several down-stream signalling pathways of TGF-β were studied. It could be demonstrated that nintedanib inhibited TGF-β1-induced phosphorylation of the MAP kinase ERK1/2 and of the protein tyrosine kinase c-Abl. The lack of phosphorylation of Smad2/3 upon TGF-β stimulation in primary human lung fibroblasts is in accordance with previous data. The finding that TGF-β stimulates c-Abl tyrosine kinase is in line with others identifying c-Abl as an important element of TGF-β signalling and crucial for TGF-β-induced synthesis of ECM-proteins. ERK1/2 is a well known component of the non-canonical TGF-β-signalling and inhibition of ERK suppressed collagen expression in vitro. Based on our data we therefore suggest that the antagonistic effect of nintedanib on TGF-β-induced collagen secretion is – at least partly – due to the inhibition of c-Abl and/or ERK1/2. We hypothesize that the kinase inhibitor nintedanib inhibits c-Abl tyrosine kinase, similar to imatinib mesylate, a tyrosine kinase inhibitor specific for PDGFR as well as cAbl. Finally, the ERK1/2 MAP kinase is not TGF-β-specific but is also a downstream target responsive to PDGF receptor activation. In summary, our data indicate that nintedanib is not only able to inhibit the pro-fibrotic effects of PDGF and FGF but is also capable of abrogating the effects of TGF-β – a key mediator in tissue fibrogenesis.
Our study has several limitations: i) we are aware that alveolar epithelial cells are thought to be critical in the initiation and progression of the fibrotic process in IPF. However, the growth factor-induced proliferation of local fibroblasts and the exaggerated accumulation of fibroblast-derived ECM ultimately result in the destruction of the lung parenchyma. Therefore, we feel that studying the effect of nintedanib on fibroblasts is justified; ii) we acknowledge that measuring pro-MMPs is not reflecting their bioactivity which requires the interaction with other cell types.