Discoidin domain receptor 1 contributes to the survival of lung fibroblast in idiopathic pulmonary fibrosis

Abstract

Idiopathic pulmonary fibrosis (IPF), characterized by fibroblast proliferation and accumulation of extracellular matrix, including collagen, is a chronic progressive disorder that results in lung remodeling and fibrosis. However, the cellular mechanisms that may make fibroblasts resistant to apoptosis have not been completely elucidated. Discoidin domain receptor 1 (DDR1), a receptor tyrosine kinase whose ligand is collagen, is expressed in vivo and contributes in vitro to leukocyte differentiation and nuclear factor (NF)-kappaB activation, which may play an important role in fibroblast survival. In this study, we examined in vivo and in vitro DDR1 expression and its role in cell survival using fibroblasts obtained from IPF and non-IPF patients. Immunohistochemically, fibroblasts present in fibroblastic foci expressed endogenous DDR1. The DDR1 expression level was significantly higher in fibroblasts from IPF patients, and the predominant isoform was DDR1b. In IPF patients, DDR1 activation in fibroblasts inhibited Fas ligand-induced apoptosis and resulted in NF-kappaB nuclear translocation. Suppression of DDR1 expression in fibroblasts by siRNA abolished these effects, and an NF-kappaB inhibitor abrogated the anti-apoptotic effect of DDR1 activation. We propose that DDR1 contributes to fibroblast survival in the tissue microenvironment of IPF and that DDR1 up-regulation may occur in other fibroproliferative lung diseases as well.

Figure 1

Immunohistochemistry of the biopsied lung of an IPF patient. C: Fibroblasts in the fibroblastic foci show strong positive staining for DDR1. C and E: Inflammatory cells in the IPF lesion are also stained positive for DDR1. E: The bronchoepithelial cells are negative for DDR1. In the non-IPF lung, only alveolar macrophages are stained weakly positive for DDR1 (arrowheads). A and B: H&E staining; D and G: nonspecific rabbit IgG; E: negative control for second antibody. Original magnifications: ×300 (A); ×500 (B–H).

Figure 2

Flow cytometry analysis of cultured fibroblasts. The percentage of α-SMA-positive fibroblasts was significantly higher in the IPF patients than in the non-IPF patients. The percentage of α-SMA-positive fibroblasts was significantly higher in the ALI patients than in the non-IPF patients. The percentage of DDR1-positive fibroblasts was significantly higher in the IPF patients than in the ALI and non-IPF patients. The percentage of DDR1-positive/α-SMA-positive fibroblasts was significantly higher in the IPF patients than in the ALI and non-IPF patients. A: Representative data; B: comparison between seven different fibroblasts from seven different IPF patients, four different fibroblasts from four different ALI patients, and six different fibroblasts from six different non-IPF patients. *P < 0.01, Bonferroni-Dunn test with one-way factorial analysis of variance.

Figure 3

Western blot analysis for DDR1 expression in 1 × 10⁷ fibroblasts. The amount of total DDR1 was significantly higher in fibroblasts from the IPF patients than in those from the ALI and non-IPF patients. In IPF patients, DDR1b was the predominant isoform, whereas no significant difference in DDR1 isoforms was observed between ALI and non-IPF patients. The amount of DDR1b was significantly higher in fibroblasts from the IPF patients than in those from the ALI and non-IPF patients. A: Representative data; B: comparison of DDR1/actin ratio between seven different fibroblasts from seven different IPF patients, four different fibroblasts from four different ALI patients, and six different fibroblasts from six different non-IPF patients. *P < 0.01, Bonferroni-Dunn test with one-way factorial analysis of variance. With regard to DDR1b expression levels, no significant difference was observed among the passage numbers. C: Representative data; D: comparison between seven different fibroblasts from seven different IPF patients and six different fibroblasts from six different non-IPF patients. The amount of total DDR1 from 1-mg samples of total lung tissue was significantly higher in fibroblasts from the IPF patients than in those from the non-IPF patients. In the IPF patients, DDR1b was the predominant isoform, whereas no significant difference was observed between the DDR1 isoforms in the non-IPF patients. The amount of DDR1b in the total lung tissue was significantly higher in the IPF patients than in the non-IPF patients. E: Representative data; F: comparison of DDR1/actin amount ratio between five different fibroblasts from five different IPF patients and four different fibroblasts from four different non-IPF patients. *P < 0.01, Bonferroni-Dunn test with one-way factorial analysis of variance.

Figure 4

Effect of DDR1 agonistic antibodies and collagen on the FasL-induced apoptosis of fibroblasts. FasL-induced apoptosis of fibroblasts was observed in both groups. Collagen inhibited the FasL-induced apoptosis of fibroblasts from IPF patients; however, it showed no effect in the case of non-IPF patients. Neutralizing antibodies specific for β1-integrin, another collagen receptor, did not affect the anti-apoptotic effect of collagen. The DDR1 agonistic antibodies inhibited the FasL-induced apoptosis of fibroblasts from the IPF patients in a dose-dependent manner; however, they showed no effect in the case of the non-IPF patients. In IPF patients, the percentage of Annexin V-positive fibroblasts treated with FasL and 50 μg/ml of collagen was significantly lower than those treated with FasL alone. In IPF patients, the percentage of Annexin V-positive fibroblasts treated with FasL and 1 mg/ml DDR1 agonistic antibodies was significantly lower than those treated with FasL alone. Thus, when fibroblasts were treated with collagen or DDR1 agonistic antibodies, the percentage of Annexin V-positive fibroblasts was significantly lower in the IPF patients than in the non-IPF patients. Control IgM did not show any effect on the FasL-induced apoptosis. A: Representative data; B: comparison of the percentages of Annexin V-positive cells between seven different fibroblasts from seven different IPF patients and six different fibroblasts from six different non-IPF patients. **P < 0.05, *P < 0.01, Bonferroni-Dunn test with one-way factorial analysis of variance.

Figure 5

A: Western blot analysis of DDR1 autophosphorylation in 1 × 10⁷ fibroblasts from IPF patients. The membrane was probed with anti-phosphotyrosine antibodies, anti-DDR1 antibodies (C-20), or anti-Shc antibodies. Arrows indicate phosphorylated DDR1 or phosphorylated Shc. DDR1 autophosphorylation (top arrow) and the Shc recruitment (bottom arrow) were observed after stimulation with monomeric collagen, fibrillar collagen, and DDR1 agonistic antibodies but not with control IgM. Representative data of five individual experiments using samples from five different donors are shown. B and C: Electrophoretic mobility shift assay analysis of NF-κB translocation. Monomeric collagen, fibrillar collagen, and DDR1 agonistic antibodies induced NF-κB nuclear translocation in fibroblasts from the IPF patients; however, this was not observed in the case of the non-IPF patients. B: Representative data; C: data from five different patients in each group; *P < 0.01, Bonferroni-Dunn test with one-way factorial analysis of variance.

Figure 6

Effect of siRNA specific for DDR1. The siRNA specific for DDR1 purchased from Santa Cruz Biotechnology apparently inhibited the endogenous DDR1 expression in fibroblasts from IPF patients. A: Representative data of five individual experiments using samples from five different donors are shown. The siRNA specific for DDR1 abolished the anti-apoptotic effect of collagen and DDR1 agonistic antibodies on fibroblasts from IPF patients. The percentage of Annexin V-positive fibroblasts was significantly higher in fibroblasts treated with FasL and siRNA than those treated with FasL alone. Negative control siRNA did not show any effect on the FasL-induced apoptosis. B: Representative data; C: comparison of the percentages of Annexin V-positive cells from seven different fibroblasts from seven different IPF patients in each group. *P < 0.01, Bonferroni-Dunn test with one-way factorial analysis of variance.

Figure 7

Effect of DDR1 siRNA on the NF-κB nuclear translocation. DDR1 siRNA significantly attenuated the NF-κB nuclear translocation that was induced by collagen or DDR1 agonistic antibodies in fibroblasts from IPF patients. Negative control siRNA did not show any affect on the NF-κB nuclear translocation. A: Representative data; B: data from five different patients in each group; *P < 0.01, Bonferroni-Dunn test with one-way factorial analysis of variance.

Figure 8

Effect of the NF-κB inhibitor CAPE on the anti-apoptotic effect of collagen or DDR1 agonistic antibodies. CAPE abolished the NF-κB nuclear translocation in fibroblasts from IPF patients. A: Representative data of five individual experiments using samples from five different donors are shown. CAPE significantly attenuated the NF-κB nuclear translocation that was induced by collagen or DDR1 agonistic antibodies in fibroblasts from IPF patients. Dimethyl sulfoxide, the buffer of CAPE, did not show any effect on the NF-κB nuclear translocation. B: Representative data; C: data from seven different patients in each group. *P < 0.01, Bonferroni-Dunn test with one-way factorial analysis of variance.