Reactive oxygen species are required for maintenance and differentiation of primary lung fibroblasts in idiopathic pulmonary fibrosis

Abstract

Background: Idiopathic pulmonary fibrosis (IPF) is a progressive and fatal illness whose pathogenesis remains poorly understood. Recent evidence suggests oxidative stress as a key player in the establishment/progression of lung fibrosis in animal models and possibly in human IPF. The aim of the present study was to characterize the cellular phenotype of fibroblasts derived from IPF patients and identify underlying molecular mechanisms.

Methodology/principal findings: We first analyzed the baseline differentiation features and growth ability of primary lung fibroblasts derived from 7 histology proven IPF patients and 4 control subjects at different culture passages. Then, we focused on the redox state and related molecular pathways of IPF fibroblasts and investigated the impact of oxidative stress in the establishment of the IPF phenotype. IPF fibroblasts were differentiated into alpha-smooth muscle actin (SMA)-positive myofibroblasts, displayed a pro-fibrotic phenotype as expressing type-I collagen, and proliferated lower than controls cells. The IPF phenotype was inducible upon oxidative stress in control cells and was sensitive to ROS scavenging. IPF fibroblasts also contained large excess of reactive oxygen species (ROS) due to the activation of an NADPH oxidase-like system, displayed higher levels of tyrosine phosphorylated proteins and were more resistant to oxidative-stress induced cell death. Interestingly, the IPF traits disappeared with time in culture, indicating a transient effect of the initial trigger.

Conclusions/significance: Robust expression of α-SMA and type-I collagen, high and uniformly-distributed ROS levels, resistance to oxidative-stress induced cell death and constitutive activation of tyrosine kinase(s) signalling are distinctive features of the IPF phenotype. We suggest that this phenotype can be used as a model to identify the initial trigger of IPF.

Figure 1. IPF Fibroblasts Display a Pro-Fibrotic Phenotype.

The baseline phenotype of control (n = 4) and IPF cell lines (n = 7) grown in complete medium was analyzed at different culture passages. Cell lysates from sub-confluent cultures were immune-blotted with anti-α-SMA antibodies. Panel A shows a representative western blot of α-SMA (upper) and tubulin (lower) proteins in control and IPF fibroblasts at early and late passages. Relative statistics for early (full bars) and late (dashed bars) passage cells are also reported. *p<0.05 versus early passage control cells; ^**p<0.05 versus early passage IPF cells. Panel B shows real time PCR expression levels of type-I collagen mRNA in control and IPF fibroblasts at early (full bars) and late (dashed bars) culture passages. *p<0.001 versus early passage control cells; **p<0.001 versus early passage IPF cells. Panel C shows the proliferation rate of early (full bars) and late (dashed bars) passage control and IPF fibroblasts. *p<0.001 versus early passage control cells; **p<0.001 versus early passage IPF cells. Data are expressed as mean fold change ± SD. Late passage control and IPF fibroblasts stained for senescence associated β-galactosidase activity are shown in panel D (magnification 10X). Arrows show β-galactosidase positive blue cells. The mean percentage of β-galactosidase positive cells ± SD in late passage cells (dashed bars) is reported.

Figure 2. The IPF Phenotype is not Dependent upon the Autocrine Influence of TGF-beta.

Early passage control fibroblasts (n = 3) were cultured in the presence of medium obtained from control or IPF cell cultures for 48 h. Panel A shows a representative western blot of α-SMA (upper) and tubulin (lower) proteins and relative statistics of three independent experiments. Representative immune-blots for TGF-β1 in cell extracts and culture media of early passage control and IPF fibroblasts are shown in panel B. Culture media were concentrated by means of centrifugal filter devices YM-10 (Centricon, Millipore). In the middle, 5 ng of rhTGF-β1 has been loaded as positive control. Panel C shows real time PCR expression levels of TGF-β mRNA in control (n = 4) and IPF fibroblasts (n = 7) at early (full bars) and late (dashed bars) culture passages. Early passage IPF fibroblasts (n = 3) were treated for 30 min with a specific TGF-β receptor ALK5 inhibitor (10 µM) and changes of α-SMA expression were assessed at the protein level. Representative immune-blots for α-SMA/tubulin expression and relative statistics are reported in panel D. All data are expressed as mean fold change ± SD.

Figure 3. Chronic Oxidative Stress Induces a Pro-Fibrotic Phenotype.

Changes of α-SMA expression were investigated by means of western blot in early control fibroblasts (n = 3) upon exogenous administration of hydrogen peroxide. Panel A shows changes of α-SMA expression upon cell treatment with increasing concentrations of H₂O₂ for 48 h. Conversely, panel B shows changes of α-SMA by treating cells with a fixed concentration of H₂O₂ (200 µM) at different time points. A representative immune-blot of α-SMA/tubulin is also shown. Panel C shows type-I collagen expression, assessed by real time PCR, in cells incubated for 48 h in the presence of H₂O₂ (200 µM). Expression of α-SMA in H₂O₂ (200 µM)-treated early passage IPF fibroblasts (n = 3) over time is shown in panel D. All data are expressed as mean fold change ± SD and are representative of three independent experiments. *p<0.001 versus basal; **p<0.05 versus basal.

Figure 4. IPF Fibroblasts are Resistant to Acute Oxidative Stress-Induced Cell Death.

Induction of cell death upon exogenous oxidative stress was assessed by means of flow cytometry in early and late passage control (n = 3) and IPF (n = 3) cells. Panels A and C show the distribution of the mean percentages of PI-positive cells treated for 15 min with increasing concentration of H₂O₂, in early and late passage cells, respectively. Panels B and D show the distribution of the mean percentages of PI-positive cells treated with a fixed H₂O₂ concentration (1 mM) at different time points in early and late passage cells, respectively. All data are expressed as mean percentage ± SD of fluorescent cells and are representative of three independent experiments. *p<0.05 versus control cells.

Figure 5. Negative Modulation of the Pro-Fibrotic Phenotype by ROS Scavenging.

Control and IPF fibroblasts were treated for 48 h with the ROS scavenger NAC (5 mM). Panels A and B show representative immune-blots for α-SMA/tubulin and relative statistics in early passage control (n = 3) and IPF (n = 3) cells, respectively. Panel C illustrates changes of type-I collagen expression in early passage IPF fibroblasts. Changes of α-SMA and type-I collagen expression in NAC-treated late passage IPF fibroblasts (n = 3) are shown in panels D and E, respectively. Data are expressed as mean fold change ± SD and are representative of three independent experiments. *p<0.05 versus basal.

Figure 6. IPF Cells Produce High Levels of ROS.

The baseline ability of control (n = 4) and IPF (n = 7) cells to generate ROS was measured at different culture passages with different methods. The intracellular content of ROS was first measured fluorimetrically in DCHF-DA loaded cells at early passages. Panel A shows levels of DCF fluorescence in early (full bars) and late (dashed bars) passage control and IPF cells. *p<0.05 versus control cells. Representative flow cytometry dot plots of the distribution of DCF fluorescence in control and IPF fibroblasts at different culture passages are shown in panel B. Panel C shows O₂ ^- production, estimated by means of superoxide dismutase-inhibitable cytocrome c reduction, respectively in early (full bars) and late (dashed bars) passage control and IPF fibroblasts. All data are reported as mean value ± SD and are representative of three independent experiments. *p<0.01 versus early passage control cells; **p<0.05 versus early passage control cells; ***p<0.001 versus early passage IPF cells.

Figure 7. ROS Generation in IPF Fibroblasts Occurs Through the Activation of a Membrane NADPH Oxidase-Like System.

Basal levels of ERK phosphorylation were assessed in control (n = 4) and IPF (n = 7) fibroblasts at different culture passages. Further, to determine whether a functional NADPH oxidase complex was involved in the generation of ROS, changes of p-ERK expression were analysed in response to treatment with DPI (20 µM), a flovoprotein inhibitor. Modulation of p-ERK levels was analysed upon 30 min of cell treatment. Representative immune-blots of p-ERK/tubulin and relative statistics are respectively shown in panels A and B. *p<0.001 versus basal control cells; **p<0.001 versus basal IPF cells. To investigate whether ERK signalling was inducible upon oxidative stress, early passage control (n = 3) and IPF fibroblasts (n = 3) were treated with increasing concentrations of H₂O₂ for 30 min. Pre-treatment of cells with the ROS scavenger NAC (10 mM for 1 h) was also analysed. Representative immune-blots and relative statistics of three independent experiments are reported in panels C (*p<0.01 versus basal; **p<0.01 versus 1 mM H₂O₂ treated cells) and D (*p<0.05 versus basal; **p<0.01 versus basal). To address whether ERK activation is necessary for α-SMA expression, modulation of peroxide-induced α-SMA (H₂O₂ 200 µM for 2 h) was analysed in early passage control cells (n = 3) pre-treated with UO126 (10 mM for 30 min). Similarly, α-SMA expression was evaluated in early passage IPF fibroblasts (n = 3) in response to UO126. Panel E shows representative immune-blots of α-SMA/tubulin and relative statistics. All data are expressed as mean fold change ± SD. *p<0.05 versus H₂O₂ treated cells; **p<0.05 versus basal.

Figure 8. Constitutive Activation of Receptor Tyrosine Kinases in IPF Cells is Necessary for Differentiation into Myofibroblasts.

Baseline levels of p-tyrosine proteins were assessed by western blotting in early and late passage control (n = 4) and IPF (n = 7) cells. Representative immune-blots are shown in panel A. Then, to address whether p-tyrosine signalling was inducible upon exogenous oxidative stress early passage control fibroblasts (n = 3) were treated for 15 min with increasing concentration of H₂O₂. Pre-treatment of cells with the ROS scavenger NAC (10 mM for 1 h) was also analysed. Modulation of p-tyrosine levels are shown in a representative immune-blot in panel B (left). Conversely, representative changes of p-tyrosine levels in early passage IPF cells (n = 3) treated with NAC (10 mM for 1 h) are shown on the right. To investigate whether myofibroblast differentiation is dependent upon tyrosine kinase signalling, changes of α-SMA expression were assessed by means of western blotting in early passage control (n = 3) and IPF (n = 5) fibroblasts treated for 30 min with the RTKs inhibitors AG1296 (anti-PDGF-R, 2 µM) and AG1478 (anti-EGF-R, 2 µM). Representative immune-blots of α-SMA/tubulin and relative statistics are shown in panel C. *p<0.001 versus untreated IPF; **p<0.01 versus untreated IPF. Finally, to investigate whether RTKs inhibition abrogates peroxide-induced α-SMA expression early passage control cells (n = 3) were pre-treated for 30 min with the RTKs inhibitors AG1296 (2 µM) and AG1478 (2 µM) and then exposed for 2 h to H₂O₂ (200 µM). Panel D shows representative immune-blots of α-SMA/tubulin and relative statistics. Results are expressed as mean fold change ± SD and are representative of three independent experiments. *p<0.05 versus basal; **p<0.001 versus H₂O₂ treated cells.