Hypoxia Promotes Invadosome Formation by Lung Fibroblasts

Abstract

Lung parenchymal hypoxia has emerged as a cardinal feature of idiopathic pulmonary fibrosis (IPF). Hypoxia promotes cancer cell invasion and metastasis through signaling that is dependent upon the lysophosphatidic acid (LPA) receptor, LPA₁ (LPAR1). Abundant data indicate that LPA₁-dependent signaling also enhances lung fibrogenesis in IPF. We recently reported that fibroblasts isolated from the lungs of individuals with IPF have an increased capacity to form subcellular matrix-degradative structures known as invadosomes, an event that correlates with the degree of lung fibrosis. We therefore hypothesized that hypoxia promotes invadosome formation in lung fibroblasts through LPA₁-dependent signaling. Here, it is demonstrated that invadosome formation by fibroblasts from the lungs of individuals with advanced IPF is inhibited by both the tyrosine receptor kinase inhibitor nintedanib and inhibition of LPA₁. In addition, exposure of normal human lung fibroblasts to either hypoxia or LPA increased their ability to form invadosomes. Mechanistically, the hypoxia-induced invadosome formation by lung fibroblasts was found to involve LPA₁ and PDGFR-Akt signaling. We concluded that hypoxia increases the formation of invadosomes in lung fibroblasts through the LPA₁ and PDGFR-Akt signaling axis, which represents a potential target for suppressing lung fibrosis.

Keywords: LPA1; fibroblasts; hypoxia 2; idiopathic pulmonary fibrosis; invadosomes LPA; nintedanib.

Figure 1

Hypoxic areas are found in the lungs of mice with pulmonary fibrosis. Images of lung tissue from (A) healthy and (B) bleomycin-exposed mice. Images are representative of three healthy and five bleomycin-exposed mice. Tissues are labeled for collagen (blue) via Masson’s trichrome staining and for hypoxia (brown) via immunohistochemistry using an anti-pimonidazole antibody. Higher magnification images show the architecture of the lung tissues. For the bleomycin-exposed mouse, two regions of interest were magnified, namely, a normal region and a fibrotic lesion, to visualize the presence of hypoxia and collagen. The scale of the highest magnification images is 50 µm. (C) Quantification of pimonidazole positive area staining in non-fibrotic (NF) areas of healthy lung tissues or in non-fibrotic (NF; n = 3) and fibrotic (FIB; n = 5) areas of bleomycin-exposed mice lung tissues. Bars represent the mean +/− SEM. ** p < 0.01, Mann–Whitney test.

Figure 2

Hypoxia increased invadosome formation in lung fibroblasts. The percentage of invadosome-forming cells were assessed in lung fibroblasts cultured on fluorescent gelatin coverslips under normoxic or hypoxic conditions. Two coverslips were made for each experimental condition, and 3 × 300 cells per coverslip were counted (total 6 × 300 cells). (A) Primary mouse lung fibroblasts (p-MLF) isolated from healthy (Ctrl) (n = 6) or bleomycin (BLM)-exposed (n = 5) mice and cultured for 40 h in hypoxia; (B) primary human lung fibroblasts (p-HLF), where each dot represents a cell culture from one healthy individual. (n = 6) after 20 h of culture under hypoxia; and (C) HFL-1 cell line (n = 9). (D) The number of invadosomal structures identified through the colocalization of F-actin and cortactin was counted per cell (HFL-1 cell line). (E) Confocal micrographs of human fibroblasts stained for F-actin (red), cortactin (green), and nucleus (blue). Arrowheads show some representative colocalizations of F-actin and cortactin. Scale bar = 10 μm. For all experiments, data were obtained from at least three different mice or individuals in at least three independent experiments. Data are expressed as mean ± SEM. * = p < 0.05, ** = p < 0.01, *** = p < 0.001, **** = p < 0.001. ns = not significant.

Figure 3

Hypoxia and invadosome-related gene expression in healthy controls and IPF patients of increasing disease severity: (A) heatmap illustrating the relative expression of each gene for each healthy control (N = 84) and IPF patient presenting mild (N = 67) or severe (N = 35) % predicted DLCO; (B,C) bar graphs illustrating the relative expression of invadosome (B) or hypoxia-associated (C) genes for healthy controls (N = 84) and IPF patients presenting mild (N = 67) or severe (N = 35) % predicted DLCO. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

Figure 4

Invadosome, hypoxia, fibrosis, and metabolism-related gene expression correlation in lung tissues from IPF patient cohort GSE47460: (A) heatmap illustrating Spearman correlation coefficient of selected gene pairs; (B) heatmap illustrating associated p-values for the correlations presented in (A).

Figure 5

Invadosome markers are present in hypoxic collagen-rich areas of IPF lungs. Representative images of lung tissue from healthy and IPF patients. Tissues are labeled for collagen (blue) via Masson’s trichrome staining and for cortactin, phospho-cortactin, CAIX, or TKS5 (brown) via immunohistochemistry. N = 4; scale bar = 50 μm.

Figure 6

LPA increases invadosome formation in healthy lung fibroblasts through LPA₁. The percentage of invadosome-forming cells were assessed in lung fibroblasts cultured on fluorescent gelatin-coated coverslips in the presence or absence of LPA and the LPA₁ antagonist AM095 used at the indicated concentrations. Two coverslips were made for each experimental condition, and 3 × 300 cells per coverslip were counted (total 6 × 300 cells). (A) Primary mouse lung fibroblasts (p-MLF) isolated from healthy mice after 40 h in hypoxia; (B) HFL-1 cell line and (C) primary human lung fibroblasts (p-HLF) after 20 h in hypoxia. Each dot represents a cell culture from one healthy individual. n = 5. For all experiments, data were obtained from at least three different mice or individuals in at least three independent experiments. Data are expressed as mean ± SEM. * = p < 0.05, ** = p < 0.01, *** = p < 0.001.

Figure 7

Hypoxia increased invadosome formation in healthy lung fibroblasts through LPA₁ and RTK. The percentage of invadosome-forming cells were assessed in lung fibroblasts cultured on fluorescent gelatin-coated coverslips in normoxia or hypoxia and in the presence or absence of LPA₁ antagonists (AM095 or BMS-986020) or nintedanib. Two coverslips were made for each experimental condition, and 3 × 300 cells per coverslip were counted (total 6 × 300 cells). (A) Primary mouse lung fibroblasts (p-MLF) isolated from healthy mice after 40 h under hypoxia; (B) HFL-1 cell line; and (C–E) primary human lung fibroblasts (p-HLF). Each dot represents a cell culture from one healthy individual (n = 4–5) after 20 h under hypoxia. For all experiments, data were obtained from at least three different mice or individuals in at least three independent experiments. Data are expressed as mean ± SEM. * = p < 0.05 and ** = p < 0.0.

Figure 8

LPA₁ was involved in the activation of PDGFRβ and Akt under hypoxia in healthy lung fibroblasts. Primary human lung fibroblasts were incubated under normoxia or hypoxia in the presence or absence of nintedanib 0.2 µM, AM095 0.2 µM, or BMS-986020 0.2 µM. Immunoblots of total cell lysates. (A) Cells were incubated under hypoxia for 30 and 120 min. Densitometric analysis is relative to normoxia. (B) Representative experiment of 4–5 independent experiments involving PDGFRβ immunoprecipitation (IP:PDGFRβ) followed by immunoblotting for phosphorylated PDGFR (Y751) and total PDGFRβ with corresponding ratios relative to hypoxia. Below, immunoblots of total cell lysates showing phosphorylated PDGFR (Y857), total PDGFRβ, phosphorylated Akt (S473), and total Akt with corresponding ratio relative to normoxia. (C) Immunoblots of total cell lysates showing phosphorylated PDGFR (Y857), total PDGFRβ, phosphorylated Akt (S473), and total Akt with corresponding ratio relative to normoxia. Corresponding densitometric analysis relative to hypoxia of four to five independent experiments. Each dot represents a cell culture from one healthy individual. Data are expressed as mean ± SEM. * = p < 0.05 and ** = p < 0.01.