FoxO3a (Forkhead Box O3a) deficiency protects Idiopathic Pulmonary Fibrosis (IPF) fibroblasts from type I polymerized collagen matrix-induced apoptosis via caveolin-1 (cav-1) and Fas

Abstract

Idiopathic Pulmonary Fibrosis is a lethal fibrotic disease characterized by the unrelenting proliferation and persistence of fibroblasts in a type I collagen-rich matrix that result in an expanding reticular network of fibrotic tissue. However, the underlying mechanism responsible for the persistence of myofibroblasts in IPF remains unclear. During normal tissue repair, unwanted fibroblasts are eliminated during collagen-matrix contraction by a mechanism whereby high PTEN activity suppresses Akt. We have previously found that FoxO3a, a transcriptional activator of apoptosis-inducing proteins, is inactivated in IPF fibroblasts resulting from aberrantly high PI3K/Akt activity due to inappropriately low PTEN activity. Here we demonstrate that this low FoxO3a activity confers IPF fibroblasts with resistance to collagen-mediated apoptosis. We show that the mechanism by which low FoxO3a activity confers IPF fibroblasts with an apoptotic resistant phenotype involves suppression of Fas expression as a result of down regulation of cav-1 expression via a PTEN/Akt-dependent pathway. We demonstrate that PTEN over-expression or Akt inhibition increases FoxO3a expression in IPF fibroblasts, resulting in up-regulation of caveolin-1. We show that FoxO3a binds to the cav-1 promoter region and ectopic expression of FoxO3a transcriptionally increases cav-1 mRNA and protein expression. In turn, we show that overexpression of caveolin-1 increases Fas levels and caspase-3/7 activity and promotes IPF fibroblast apoptosis on polymerized type I collagen. We have found that the expression of caveolin-1, Fas and cleaved caspase-3 proteins in fibroblasts within the fibroblastic foci of IPF patient specimens is low. Our data indicate that the pathologically altered PTEN/Akt axis inactivates FoxO3a down-regulating cav-1 and Fas expression. This confers IPF fibroblasts with an apoptosis-resistant phenotype and may be responsible for IPF progression.

Figure 1. Low FoxO3a confers IPF fibroblasts with resistance to polymerized collagen-mediated apoptosis.

A & B. Serum starved IPF and control fibroblasts (n = 5, each) were permitted to attach to polymerized collagen for 48 h in the absence of serum and cell morphology (A) and cell viability (B) were examined. Shown are 50X magnification images of cell morphology. C & D. IPF and control fibroblasts infected with adenovirus expressing FoxO3a (F3) or empty vector GFP were cultured on collagen for 48 h in serum free medium and cell morphology (C) and cell viability (D) were examined. Shown is 50X magnification of cell morphology. Note that forced expression of FoxO3a significantly increased apoptosis in IPF fibroblasts on collagen matrix. *p = 0.003 versus GFP control. The assay was obtained from triplicate.

Figure 2. FoxO3a regulates cav-1 mRNA and protein expression.

A. Upper, shown is a representative Western blot analysis for cav-1 protein expression in control (n = 4) and IPF fibroblasts (n = 4) cultured on type I polymerized collagen matrix in serum free medium for 24 h. Lower, serum starved IPF (n = 8) and control fibroblasts (n = 6) were cultured under the same condition. Shown is cav-1/actin expression ratio of control and IPF fibroblasts on polymerized collagen. *p = 0.02 versus control fibroblasts. B. ChIP assay was carried out with primers as described in the Materials and Methods. Putative FoxO3a consensus binding sequences were underlined within C1 region. 1/25 volume (2 µl) of recovered DNA with primers flanking C1 region was used for PCR and 139 bp of ChIP-PCR product was amplified. ChIP-PCR without recovered DNA was carried out using C1 forward and reverse primers as a negative control (ND). ChIP-PCR without FoxO3a antibody using C1 forward and reverse primers was also carried out as an antibody control (NAB). ChIP-PCR from immunoprecipitate with normal rabbit IgG using C1 forward and reverse primers was performed as an IgG control (IgG). Image was obtained from a different gel. M: 1kb DNA ladder. C. Quantitative RT-PCR was carried out using primers as described in Materials and Methods. Shown is quantitative RT-PCR normalized to 18S rRNA in FoxO3a−/− and FoxO3a+/+ cells. *p = 0.002. The assay was obtained from triplicate. D. Upper panel, FoxO3a −/− cells were infected with adenovirus expressing wild type FoxO3a (F3) or empty vector GFP, and cultured on polymerized collagen for 24 hrs. Cav-1 protein expression was measured by Western analysis. GAPDH was used as a loading control. Lower panel. FoxO3a −/− cells transfected with constructs expressing wild type FoxO3a (WT), or S253A mutant FoxO3a (S253A) or a FoxO3a construct with all Akt phosphorylation residues mutated with alanine (TM) were cultured on polymerized collagen for 24 hrs. Cav-1 protein levels were measured by Western blot analysis. Vec: empty vector. Images were obtained from same Western blot.

Figure 3. FoxO3a deficiency suppresses cav-1 expression in IPF fibroblast on collagen matrix.

A. IPF and control fibroblasts infected with adenovirus expressing wild type FoxO3a (F3), dominant negative FoxO3a (DF) or empty vector (GFP) were cultured on polymerized collagen for 24 h in serum free medium. After total RNA was isolated, quantitative RT-PCR was carried out with cav-1 primers as described in the Materials and Methods. Shown is quantitative RT-PCR for cav-1 mRNA levels normalized to 18S rRNA in IPF and control fibroblasts. *p = 0.03 versus GFP in control fibroblasts, **p = 0.02 versus GFP in IPF fibroblasts. ***p = 0.016 versus GFP in IPF fibroblast. The assay was obtained from triplicate from 2 control and 1 IPF fibroblasts. B. IPF and control fibroblasts expressing wild type FoxO3a (F3), mutant FoxO3a (DF) or empty vector (GFP) were transfected with a luciferase construct containing consensus Forkhead binding sites (FHRE-Luc). Cells were attached to polymerized collagen in serum free medium for 24 h and luciferase activity was measured as described in the Materials and Methods. Non: cells without transfection of FHRE-Luc construct. The assay was obtained from triplicate. C. Left upper panel, shown is representative Western blot analysis (WB) of cav-1 protein expression in IPF fibroblasts expressing wild type (F3), dominant negative FoxO3a (DF) or GFP control and cultured on polymerized collagen for 24 h. Left middle panel, nuclear fraction was isolated from IPF fibroblasts cultured on collagen as described in the Materials and Methods and unphosphorylated FoxO3a protein level was measured. NC : nuclear fraction. Lamin A was used as a loading control. Left lower panel : cytoplasmic fraction was isolated from IPF fibroblasts cultured on collagen and phosphorylated FoxO3a (p-FoxO3a) was measured. CP : cytoplasmic fraction. GAPDH is shown as a loading control. Right panel, densitometric analysis of cav-1/GAPDH expression ratio. D. Left panel, control fibroblasts expressing wild type (F3), dominant negative FoxO3a (DF) or GFP control and cultured on polymerized collagen for 24 h and cav-1 protein expression was measured. GAPDH is shown as loading control. Right panel, densitometric analysis of cav-1/GAPDH expression ratio. All blots represent 3 independent assays.

Figure 4. Aberrant function of the PTEN/Akt axis inactivates FoxO3a and suppresses cav-1 expression.

A. Left panel. IPF and control fibroblasts infected with adenovirus expressing hyperactive Akt (HA), dominant negative Akt (DA) or empty vector GFP were cultured on polymerized collagen and cav-1 protein expression was examined by Western blot analysis. GAPDH is shown as a loading control. Right panel, cav-1/GAPDH expression ratio was quantified by densitometry. *p = 0.017, **p = 0.008 versus GFP control. B. Control fibroblasts infected with adenovirus co-expressing either wild type FoxO3a & hyperactive Akt or empty GFP & hyperactive Akt were cultured on polymerized collagen in serum free medium for 24 h, and Western analysis was carried out to measure cav-1, p-Akt, FoxO3a and GAPDH levels. Note that cav-1 level is increased when FoxO3a and hyperactive Akt were co-expressed. C. Left, IPF fibroblasts infected with adenovirus expressing wild type PTEN (WP) or empty vector GFP were cultured on polymerized collagen for 24 h and cav-1 protein expression was examined by Western blot analysis. GAPDH is shown as loading control. Right, cav-1/GAPDH levels were quantified by densitometry. D. Left, control fibroblasts infected with adenovirus expressing mutant PTEN (MP) or empty vector GFP were cultured on polymerized collagen for 24 h and cav-1 protein levels were examined by Western blot analysis. GAPDH is shown as loading control. Right, cav-1/GAPDH levels were quantified by densitometry. All blots represent at least 3 independent assays.

Figure 5. Low FoxO3a and cav-1 function confers IPF fibroblasts with an apoptotic-resistant phenotype via Fas suppression.

A & B. IPF and control fibroblasts over-expressing cav-1 or empty vector GFP were cultured on collagen for 48 h in serum free medium and cell morphology (A) and cell viability (B) were examined. Shown is 50X magnification of cell morphology. C. Serum starved IPF and control fibroblasts were cultured on polymerized collagen for 24 h in the absence of serum, and Fas/GAPDH protein expression level and caspase-3/7 activity were measured as described in the Materials and Methods (n = 4 lines each). *p = 0.02, **p = 0.015 versus control fibroblasts. D. Upper panel. IPF and control fibroblasts were cultured on collagen as a function of time, and Fas and GAPDH protein expression was measured. Lower panel. Shown is the Fas/GAPDH protein expression ratio in IPF and control fibroblasts cultured on polymerized collagen as a function of time (n = 2 lines each). E. 2×10⁵ control and IPF fibroblasts cultured on polymerized collagen in the absence of serum were ligated with 0–2 µg of human Fas activating antibody for 18 h. Viable cells were measured as described in the Materials and Methods. Shown are the % viable cells exposed to the various concentrations of Fas activating antibody compared with that of untreated cells. *p = 0.016 versus non treated cells. Results were obtained from triplicates.

Figure 6. Forced expression of FoxO3a increases Fas expression via cav-1 in IPF fibroblasts and promotes apoptosis on polymerized collagen matrix.

A. IPF fibroblasts infected with adenovirus over-expressing FoxO3a (F3), cav-1 (C-1) or empty vector GFP were cultured on polymerized collagen for 48 h in serum free medium. Left panel, Fas, FoxO3a, cav-1 and GAPDH expression were measured by Western blot analysis. Three independent assays are represented. Right panel. Fas/GAPDH protein expression ratio was quantified. *p = 0.02 versus GFP. **p = 0.05 versus GFP. B. IPF fibroblasts infected with adenovirus over-expressing FoxO3a (F3), cav-1 (C-1) or empty vector GFP were cultured on polymerized collagen for 48 h and caspase-3/7 activity was measured. Results were obtained from triplicates. C. IPF fibroblasts infected with adenovirus over-expressing FoxO3a (F3), cav-1 (C1), FoxO3a in the presence of 10 nM of Cav-1 siRNA (F3+C-1S), or empty vector GFP with 10 nM control siRNA (GFP/Con) were cultured on polymerized collagen for 48 hrs. Fas, FoxO3a, cav-1 and GAPDH expression were examined by Western blot analysis. Blot represents 3 independent assays. D. IPF fibroblasts infected with adenovirus expressing FoxO3a alone (F3) or IPF fibroblasts over-expressing FoxO3a in the presence of 10 nM of Cav-1 siRNA (F3+C-1S) or GFP/Control siRNA (GFP/Con) were cultured on polymerized collagen for 48 hrs. Left panel, Fas expression was quantified. *p = 0.03 versus GFP/Con. Results were obtained from 3 independent assays. Middle panel, Caspase-3/7 activity was measured. *p = 0.01 versus GFP/Con. Results were obtained from triplicates. Right panel, viable cells were quantified. *p = 0.01, **p = 0.05 versus GFP/Con, respectively. Results were obtained from quadruplicates. E–F. 2×10⁵ IPF fibroblasts infected with adenovirus over-expressing FoxO3a (F3), cav-1 (C1) or empty vector GFP were cultured on polymerized collagen in the presence of 10 nM Fas or control siRNA for 48 h in serum free medium. Fas, FoxO3a, cav-1 and GAPDH expression were measured by Western blot analysis (E). Cell viability (F, upper panel) and caspase-3/7 activity (F, lower panel) were measured as described in Materials and Methods. Results were obtained from triplicates.

Figure 7. Fibroblasts within the fibroblastic foci of IPF patient lung specimens express inactive FoxO3a and α-smooth muscle actin but not cav-1 and cleaved caspase-3.

Upper upper panel, frozen serial sections of IPF patient lung specimens (F/F: fibroblastic foci) were immunostained with anti-α smooth muscle actin antibody (upper left), phosphorylated FoxO3a (inactive FoxO3a, upper middle), or anti-cav-1 (upper right), and IHC was carried out as described in the Materials and Methods (n = 3). Upper lower panel, shown are IHC images obtained from control tissue without primary antibodies as negative controls (N/C). Lower upper panel. IHC analysis of IPF (left) and control (middle) lung tissue was performed using anti-cleaved caspase-3 antibodies. N/C : IHC images obtained from control tissue without cleaved caspase-3 antibody as a negative control. Lower lower panel. Fas expression was measured from frozen sections of IPF (left) and control (middle) lung specimens. N/C : IHC images obtained from control tissue without Fas antibody as a negative control. Note that cav-1, cleaved caspase-3 and Fas expression were absent or very low in cells within the fibroblastic foci.

Figure 8. Schematic illustrating the mechanism by which FoxO3a deficiency confers IPF fibroblasts with resistance to polymerized collagen-mediated apoptosis.

When control fibroblasts are cultured on polymerized collagen, Akt activity is suppressed by elevated PTEN activity, resulting in an increase in FoxO3a activity. Active FoxO3a then transcriptionally increases cav-1 level and cells undergo apoptosis via high Fas expression and caspase-3/7 activity. In contrast, PTEN activity is low in response to IPF fibroblast attachment to collagen matrix, which activates Akt thereby suppressing FoxO3a function. Cav-1 transcription is low due to inactive FoxO3a, thereby suppressing Fas expression & caspase-3/7 activity, which subsequently protects IPF fibroblasts from polymerized collagen-induced apoptosis.