Integrin alpha4beta1 regulates migration across basement membranes by lung fibroblasts: a role for phosphatase and tensin homologue deleted on chromosome 10

Abstract

Idiopathic pulmonary fibrosis is a disease that is characterized by fibroblast accumulation and activation in the distal airspaces of the lung. We hypothesized that fibrotic lung fibroblasts migrate/invade across basement membranes by integrin-mediated mechanisms as a means of entering alveoli. We demonstrate that in lung fibroblasts derived from patients with idiopathic pulmonary fibrosis, fibronectin signaling is both necessary and sufficient for basement membrane migration/invasion across basement membranes. This effect is mediated through the alpha5beta1 integrin because blockade of fibronectin-alpha5 integrin ligation attenuated this response. In contrast, ligation of alpha4beta1 integrin inhibits basement membrane invasion by normal lung fibroblasts but not by fibrotic lung fibroblasts. This phenotypic difference is not related to surface expression of the alpha4beta1 integrin, as demonstrated by flow cytometry. In normal lung fibroblasts but not in fibrotic lung fibroblasts, we show that ligation of alpha4beta1 integrin induces a significant increase in phosphatase and tensin homologue deleted on chromosome 10 (PTEN) activity. Fibrotic lung fibroblasts express constitutively less PTEN mRNA and protein as well as phosphatase activity in comparison to normal lung fibroblasts. Together, these data suggest that a loss of alpha4beta1 signaling via PTEN confers a migratory/invasive phenotype to fibrotic lung fibroblasts. Furthermore, this study implicates a loss of PTEN function in the pathophysiology of idiopathic pulmonary fibrosis.

Figure 1

Fibrotic lung fibroblasts (FLFs), but not normal lung fibroblasts (NLFs), migrate across basement membranes in the presence of serum. (A) Matrigel transwell assay. NLFs (open bars) and FLFs (closed bars) were stimulated with serum-free (SF) media or media containing 10% fetal calf serum (FCS). Cell migration is reported as the mean number of cells/high-power field (hpf) ± SEM. (B) Sea urchin extracellular matrix invasion assay. NLFs (open bars) and FLFs (closed bars) were stimulated with SF media or media containing 10% FCS. Cell invasion is reported as the mean percentage of cells invading basement membranes ± SEM. FLFs migrated across basement membranes in the presence of serum, whereas SF media resulted in minimal cell migration. Data are representative results from experiments with fibroblasts from five patients in each group. Each patient’s fibroblasts were assessed at least twice in each assay.

Figure 2

Fibronectin is necessary for full migration across basement membranes by FLFs. (A) Matrigel transwell assay. FLFs were stimulated to cross basement membranes in the presence of 10% FCS or 10% FCS that had been depleted of fibronectin (FCS^Fn−). Cell migration is reported as the mean number of cells/hpf ± SEM. (B) Sea urchin extracellular matrix invasion assay. FLFs were stimulated with FCS or FCS^Fn−. Cell invasion is reported as the mean percentage of cells invading basement membranes ± SEM. FLFs invaded basement membranes significantly in the presence of serum. FCS^Fn− induced no significant basement membrane invasion above that seen for serum-free conditions (data not shown). Data are representative results from experiments with fibroblasts from five patients in each group. Each patient’s fibroblasts were assessed at least twice in each assay.

Figure 3

Fibronectin is sufficient to induce migration across basement membranes by FLFs. Fibroblast migration was assessed by the Matrigel transwell assay using SF media or SF media to which 4-μg/ml plasma fibronectin (Fn) had been added. Cell migration is reported as mean number of cells per hpf (± SEM). The addition of Fn did not induce increased invasion in NLFs. Data are representative results from experiments with fibroblasts from five patients in each group. Each patient’s fibroblasts were assessed at least twice in each assay.

Figure 4

Connecting segment-1 (CS-1) peptide ligation with α₄β₁ integrin attenuates basement membrane migration in NLFs but not FLFs. (A) Matrigel transwell assay of NLFs in response to SF media alone, SF media with the 110-kD fibronectin fragment (110 kD), or a combination of the 110 kD and CS-1 containing fibronectin fragments (110 kD/CS-1). Cell migration is reported as mean number of cells per hpf ± SEM. The CS-1 fibronectin fragment attenuates migration across basement membranes by NLFs but not by FLFs. Migration across basement membranes by FLFs was not different between the 110-kD and the 110-kD/CS-1 groups. Data are representative of experiments with fibroblasts from five patients in each group. Each patient’s fibroblasts were assessed at least twice. (B) Fibronectin-induced migration across basement membranes is inhibited by α₄β₁ ligation. NLFs migrated across basement membranes minimally in response to intact fibronectin when treated with nonspecific IgG (control). In the presence of anti-α₄ integrin antibody, intact fibronectin induced significant fibroblast migration across basement membranes.

Figure 5

Cell-surface expression of the α₄ and α₅ integrin subunits on NLFs and FLFs. Equal numbers of each fibroblast line were incubated with fluorescein isothiocyanate–conjugated anti-α₅ integrin antibody, R-phycoerythrin (PE)-conjugated anti-α₄ integrin antibody, or the appropriate fluorochrome-labeled nonspecific IgG and assessed for fluorescence using flow cytometry. In all cases, each patient’s lung fibroblasts expressed both the α₄ and α₅ integrin subunits on the cell surface. Thin dotted lines represent nonspecific IgG staining, and solid lines represent the integrin-specific fluorescent signal. Data are representative results from experiments with fibroblasts from five patients in each group. To ensure reproducibility, each patient’s fibroblasts were assessed at least twice. UIP = usual interstitial pneumonia.

Figure 6

Phosphatase and tensin homologue deleted on chromosome 10 (PTEN) lipid phosphatase activity is induced by CS-1 fibronectin fragment ligation of the α₄β₁ integrin. Human lung fibroblasts (IMR-90) were incubated for 18 hours with SF media alone, SF media containing 20-nM CS-1 peptide alone, or in the presence of nonspecific IgG or anti-α₄β₁ integrin antibody. Cells were assayed for PTEN activity, and results were expressed as the mean change in phosphate released (± SEM) for each condition, as compared with values obtained at baseline. PTEN lipid phosphatase activity is significantly induced by CS-1 ligation. This effect is specific for ligation of CS-1 with the α₄β₁ integrin. Changes in PTEN activity were not observed in the absence of ligands (data not shown). Results are representative of a duplicate set of experiments. (B) PTEN activity is induced in NLFs but decreased in FLFs in response to CS-1 peptide. NLFs (n = 5; open bar) and FLFs (n = 5; closed bar) were exposed to 20-nM CS-1 peptide, and PTEN immunoprecipitates were assayed for phosphatase activity. NLFs demonstrated increased PTEN activity, whereas FLFs decreased PTEN activity under the same conditions. Results are representative of the mean change in phosphate released by NLFs and FLFs.

Figure 7

PTEN gene and protein expression and phosphatase activity are constitutively diminished in FLFs compared with NLFs. (A) Semiquantitative real-time polymerase chain reaction for PTEN was performed on 100-ng total RNA obtained from NLFs and FLFs. Constitutive PTEN gene expression is significantly diminished in FLFs as compared with NLFs (p = 0.023). (B) Western blot analysis for PTEN was performed on whole cell lysates of fibroblasts from five NLF lines and five FLF lines. NLFs displayed markedly increased levels of PTEN protein in whole-cell lysates than FLFs. To ensure consistent protein loading, the blot was stripped and reprobed with an antibody against β-actin. Relative protein expression was determined as the densitometric ratio of PTEN to β-actin. (C) PTEN phosphatase activity is diminished in FLFs compared with NLFs. Phosphatase activity was assessed in five NLF and five FLF lines. To ensure reproducibility, the experiment was repeated twice. Representative results are shown.