Anti-fibrotic efficacy of nintedanib in pulmonary fibrosis via the inhibition of fibrocyte activity

Human fibrocytes were isolated according to previously described methods [17, 19]. All procedures for consent, sample collection, and privacy protection were approved by the Ethics Committee of Tokushima University Hospital. Human mononuclear cells (HMNC) were isolated from the peripheral blood of healthy volunteers, and cultured on fibronectin-coated dishes. After six to seven days, adherent cells were used as fibrocytes, the phenotype of which was confirmed by a flow cytometric analysis. Monocytes were isolated from HMNC with an automated magnetic cell separation device.

Nintedanib and SB431542 were obtained from Boehringer Ingelheim GmbH & Co. KG (Biberach, Germany). SU5416, a VEGFR-specific inhibitor, was purchased from Abcam (Cambridge, MA). BGJ-398 and imatinib were purchased from Chemietek (Indianapolis, IN). Bleomycin (BLM) was purchased from Nippon Kayaku Co. (Tokyo, Japan).

Mediator concentrations were measured in the cell culture supernatants of fibrocytes, monocytes, and fibroblasts using commercial enzyme-linked immunosorbent assay (ELISA) kits.

Fibrocytes, monocytes, and fibroblasts were lysed and used for immunoblotting as described previously [20].

MRC-5 cells were cultured in the cell culture supernatant of fibrocytes with various concentrations of inhibitors (0–1 μM) or recombinant growth factors (FGF2: 30 ng/ml, PDGF-AA: 100 ng/ml, PDGF-BB: 100 ng/ml, VEGF-A: 100 ng/ml) for 72 h. A [³H] thymidine deoxyribose (³H–TdR) incorporation assay was performed as described previously [3].

HMNC were seeded in fibronectin-coated 6-well plates with growth factor (FGF2: 30 ng/ml, PDGF-BB: 100 ng/ml, VEGF-A: 100 ng/ml), and various concentrations of inhibitors. Each growth factor and inhibitor was added again every 48 h. On day 6, attached cells were stained with Diff-Quick and counted.

Fibrocytes were added to the upper chamber of cell culture inserts with a pore size of 8 μm in the presence or absence of various concentrations of nhibitors (0–100 nM). Growth factors (FGF2: 30 ng/ml, PDGF-BB: 100 ng/ml, VEGF-A: 100 ng/ml) were added to the lower chamber. After 20-h incubation, fibrocytes that had migrated to the bottom surface of the filter were stained with Diff-Quick and counted [17, 19].

Eight-week-old C57BL/6 male mice were purchased from CLEA Japan (Tokyo, Japan). Mice received a single transbronchial instillation of 7.5 mg/kg BLM on day 0. Nintedanib at 60 mg/kg was administered daily by gavage until day 7. Lung tissue was analyzed on day 7 via a fluorescence-activated cell sorter (FACS) analysis and immunohistochemistry [17].

Paraffin-embedded lung sections were stained with primary antibodies and then stained with fluorescence-conjugated secondary antibodies and 4′, 6-diamidino-2-phenylindole. Fluorescence images were captured with a confocal laser scanning microscope and counted [17].

Minced lungs were digested, and the harvested cells were stained with antibodies for CD45, CXCR4 and collagen-1. Stained cells were analyzed using a FACScan flow cytometer (BD Biosciences-Pharmingen, San Diego, CA) [17].

The significance of differences were analyzed using Mann–Whitney U test for unpaired samples, or a one-way ANOVA followed by a Dunnett’s test. Where appropriate, the Kruskal-Wallis H test was applied with Dunn’s test. P values of less than 0.05 were considered to be significant. Statistical analyses were performed using GraphPad Prism programme Ver. 5.01 (GraphPad Software Inc.).

We confirmed the expression of growth factors in fibrocytes as previously reported [18]. In the present study, we compared their expression among monocytes, fibrocytes, and fibroblasts. Based on the targets of nintedanib, FGF2, PDGF-AA, PDGF-BB, VEGF-A, VEGF-B, VEGF-C, and TGFβ-1 were examined in the different culture supernatants using ELISA. Fibrocytes secreted greater amounts of FGF2, PDGF-BB, and VEGF-A than monocytes (Fig. 1a–d). Fibrocytes and fibroblasts both secreted PDGF-AA (Fig. 1b). Only fibroblasts secreted VEGF-C (Fig. 1e). PDGF-AB, TGFβ-1, and VEGF-B were below the detection limit of ELISA. The expression of FGF2 and PDGF-BB from fibrocytes was also confirmed by an immunoblot analysis (Fig. 2). These results suggest that fibrocytes are one of the sources of growth factors in pulmonary fibrosis.

The expression of growth factor receptors on fibrocytes, monocytes, and fibroblasts was examined by an immunoblot analysis. Fibrocytes expressed FGFR2 and VEGFR1. Fibroblasts also expressed FGFR2, and strongly expressed PDGFRα and β (Fig. 2).

In order to examine the effects of culture supernatants of fibrocytes as well as those of nintedanib on the phosphorylation of growth factor receptors, the expression of all receptors and receptor phosphorylation were examined using an immunoblot analysis. The incubation of MRC-5 cells with the culture supernatant of fibrocytes resulted in the phosphorylation of PDGFR, which was inhibited by nintedanib mainly at a concentration of 100 nM or more. However, the inhibitory effects of nintedanib were more potent on the phosphorylation of PDGFR compared to FGFR (Fig. 3a–c). These results indicate that growth factors produced by fibrocytes have a biological activity to stimulate fibroblasts, which can be inhibited by nintedanib.

Next, we investigated the effects of nintedanib on the fibrocyte-induced proliferation of lung fibroblasts. Fibroblast proliferation induced by the culture supernatant of fibrocytes was inhibited by nintedanib mainly at a concentration of 100 nM or more (Fig. 4a). The selective inhibition of FGFR and PDGFR by their specific inhibitors, BGJ-398 and imatinib, respectively, also inhibited fibroblast proliferation (Fig. 4b, c), whereas the VEGFR inhibitor SU5416 did not (Fig. 4d). These results were supported by recombinant FGF2, PDGF-AA, and PDGF-BB also stimulating the proliferation of fibroblasts, whereas VEGF-A did not (Fig. 4e). Taken together, these results show that FGF and PDGF, but not VEGF are important for the fibrocyte-mediated proliferation of fibroblasts.

We examined the effects of nintedanib on the differentiation of fibrocytes. We used human fibrocytes generated from HMNC by culturing them on fibronectin-coated 6-well plates for six days.

Various kinase inhibitors, including nintedanib, BGJ398, imatinib, SU5416, and SB431542, in addition to serum amyloid P were added to the culture medium of human HMNC at the concentrations ranging from 0 to 1 μM (Additional file 2: Figure S1). As reported previously, serum amyloid P inhibited the differentiation of fibrocytes [21]. Specific inhibitors including nintedanib, BGJ398, imatinib, and SU5416 reduced the number of fibrocytes generated by HMNC, whereas SB431542, which is a kinase inhibitor for the receptor of TGFβ, did not.

In order to confirm that these results were due to the activation of growth factor receptors, we investigated whether the addition of the recombinant proteins of growth factors increased the generation of fibrocytes from HMNC. Recombinant FGF2, PDGF-BB, and VEGF-A increased the number of fibrocytes generated by HMNC (Fig. 5a, b, and c). In contrast, BGJ398, imatinib, and SU5416 inhibited growth factor-induced differentiation corresponding to each specific inhibitor (Fig. 5a, b, and c). Nintedanib with the concentration of 100 nM or more also inhibited differentiation promoted by all growth factors (Fig. 5a, b, and c).

In an attempt to rule out the possibility that these results were due to the cytotoxic activity of nintedanib against fibrocytes, we investigated the effects of nintedanib on the viability of human fibrocytes derived from monocytes in vitro. Nintedanib was added to a culture of fibrocytes on day 6. The treatment of fibrocytes with nintedanib at concentrations up to 1 μM for six days did not decrease the number of fibrocytes harvested on day 12 (Additional file 2: Figure S2), indicating that the cellular toxicity of nintedanib on fibrocytes was negligible.

We assessed the effects of kinase inhibitors on the migration of human fibrocytes induced by various growth factors at concentrations up to 100 nM. The number of migrated fibrocytes markedly increased when cells were treated with FGF2, PDGF-BB or VEGF-A. Moreover, BGJ398, imatinib, and SU5416 inhibited migration mediated by growth factors corresponding to each inhibitor (Fig. 6a, b, and c). Nintedanib with the concentration of 30 nM or more also inhibited differentiation promoted by PDGF-BB and VEGF-A, but didn’t inhibited differentiation promoted by FGF2 (Fig. 6a).

We investigated whether nintedanib reduces the number of fibrocytes in a mouse model of BLM-induced pulmonary fibrosis. Mice received a single transbronchial instillation of BLM on day 0. After the administration of nintedanib with the concentration of 60 mg/kg by gavage from days 0 to 7, lung tissues were harvested and analyzed. Paraffin-embedded lung sections were stained for S100A4/fibroblast-specific antigen-1 (FSP-1) and CD45. Because FSP-1 is expressed not only by fibrocytes but also by fibroblasts and endothelial cells, we counted only strongly stained cells as FSP-1 positive cells. Double-positive cells for S100A4/FSP-1 and CD45 in lung sections were defined as fibrocytes as described previously [17] (Fig. 7a). The administration of BLM significantly increased the number of fibrocytes per high-power field in the lungs of BLM-treated mice, while nintedanib at 60 mg/kg significantly reduced these numbers (Fig. 7c). We also counted the number of fibrocytes in the lungs of BLM-treated mice using a three-color FACS analysis (CXCR4, collagen-1, and CD45). The percentage of fibrocytes (CXCR4⁺Col1⁺CD45⁺cells) among all lung cells in BLM-treated mice was elevated on day 7, and nintedanib at 60 mg/kg was found to significantly reduce this percentage (Fig. 7b, d). Consequently, when nintedanib was continuously administrated until day 21, the number of fibrotic lesions in the lungs of BLM-treated mice was reduced (Additional file 2: Figure S3). A quantitative histological analysis showed that the Ashcroft fibrotic score was significantly lower in mice treated with BLM and nintedanib at a dose of 60 mg/kg than in those treated with BLM alone (Additional file 2: Figure S4A). A hydroxyproline colorimetric assay also showed reductions in collagen content in the lungs of mice treated with nintedanib at 60 mg/kg (Additional file 2: Figure S4B). The anti-fibrotic effects of nintedanib were consistent with those described previously [7].