FoxO3 an important player in fibrogenesis and therapeutic target for idiopathic pulmonary fibrosis

Abstract

Idiopathic pulmonary fibrosis (IPF) is a progressive and fatal parenchymal lung disease with limited therapeutic options, with fibroblast-to-myofibroblast transdifferentiation and hyperproliferation playing a major role. Investigating ex vivo-cultured (myo)fibroblasts from human IPF lungs as well as fibroblasts isolated from bleomycin-challenged mice, Forkhead box O3 (FoxO3) transcription factor was found to be less expressed, hyperphosphorylated, and nuclear-excluded relative to non-diseased controls. Downregulation and/or hyperphosphorylation of FoxO3 was reproduced by exposure of normal human lung fibroblasts to various pro-fibrotic growth factors and cytokines (FCS, PDGF, IGF1, TGF-β1). Moreover, selective knockdown of FoxO3 in the normal human lung fibroblasts reproduced the transdifferentiation and hyperproliferation phenotype. Importantly, mice with global- (Foxo3^-/-) or fibroblast-specific (Foxo3_f.b^-/-) FoxO3 knockout displayed enhanced susceptibility to bleomycin challenge, with augmented fibrosis, loss of lung function, and increased mortality. Activation of FoxO3 with UCN-01, a staurosporine derivative currently investigated in clinical cancer trials, reverted the IPF myofibroblast phenotype in vitro and blocked the bleomycin-induced lung fibrosis in vivo These studies implicate FoxO3 as a critical integrator of pro-fibrotic signaling in lung fibrosis and pharmacological reconstitution of FoxO3 as a novel treatment strategy.

Keywords: fibroblast; forkhead box O transcription factors; idiopathic pulmonary fibrosis; myofibroblast; transdifferentiation.

Figure 1. Expression and regulation of FoxO3 in lung fibroblasts of IPF patients and in human donor lung fibroblasts in response to pro‐fibrotic stimuli

1. A

   Proliferation rate of N‐HLF (isolated from healthy donor; n = 3) and IPF‐HLF (isolated from IPF patients; n = 3) in the absence or presence of 5% FCS was measured by BrdU incorporation.

2. B

   mRNA expression of myofibroblasts markers (ACTA2, COL1A1, and COL3A1) in N‐HLF (n = 9) and IPF‐HLF (n = 7) by qPCR.

3. C–F

   Expression and activity of FoxO3 in N‐HLF and IPF‐HLF (passages 4–5). (C) mRNA expression analysis of FOXO3 by qPCR (n = 7/group). (D, left panel) Representative Western blot of FoxO3. (D, right panel) Densitometrically quantified data of FoxO3/GAPDH expression ratio (n = 6–7/group). (E) Representative Western blots of p‐FoxO3 (Thr32), p‐FoxO3 (Ser253). (F) Densitometrically quantified data of p‐FoxO3 (Thr32) or p‐FoxO3 (Ser253) to GAPDH expression ratio (n = 6–7/group).

4. G–J

   Representative Western blots of p‐FoxO3 (Thr32), p‐FoxO3 (Ser253), and FoxO3 in serum‐starved (48 h) N‐HLF (n = 3) that were stimulated with 5% FCS (G), PDGF‐BB (H), IGF‐1 (I), or TGF‐β1 (J) as indicated.

5. K

   Representative immunocytochemistry (ICC) images of FoxO3 localization in N‐HLF (n = 4) stimulated without/with TGF‐β1 (10 ng/ml) as indicated. Scale bar = 50 μm.

6. L

   mRNA expression of FOXO3 by qPCR (n = 7; left panel) and (middle panel) representative Western blot of FoxO3 in N‐HLF stimulated with TGF‐β1 for 24 h (right panel). Densitometrically quantified data, represented as fold change of FoxO3/GAPDH expression ratio (n = 4).

Data information: Data are expressed as mean ± SEM. In (A–D and F), data were analyzed using Student's t‐test, *P < 0.05, **P < 0.01, ***P < 0.001 versus donor. In (L), data were analyzed using paired t‐test, *P < 0.05, ***P < 0.001 versus control. In all blots, GAPDH was used as a loading control.

Figure EV1. PDGF induces FoxO3 phosphorylation and nuclear exclusion of FoxO3

1. Densitometry quantitation ratio of p‐FoxO3 (Thr32) (left panel) and p‐FoxO3 (Ser253) (right panel) in serum‐starved (48 h) N‐HLF (n = 3) that were stimulated without/with PDGF‐BB as indicated. Quantification is represented as a fold change to control (time corresponded non‐stimulated cells). Data were analyzed using repeated‐measures ANOVA, **P < 0.01 versus control 15 min, ^§ P < 0.05, ^§§ P < 0.01, ^§§§ P < 0.001 versus control 30 min.

2. ICC of FoxO3 in N‐HLF that was serum‐starved for 48 h, and stimulated with PDGF‐BB (60 ng/ml) as indicated. Control image panel represents cells that were left non‐stimulated for 6 h. TO‐PRO3 (blue) was used to label nuclei. FoxO3 and TO‐PRO3 images were overlaid to visualize nuclear and cytoplasmic localization of FoxO3. Images are representative of n = 3. Scale bar = 50 μm.

Figure EV2. Different N‐HLFs respond in similar manner to various growth factors

1. A–C

   Western blots of p‐FoxO3 (Thr32), FoxO3, and GAPDH in serum‐starved (48 h) N‐HLF (n = 3) that was stimulated with 5% FCS (A), PDGF‐BB (B), or IGF‐1 (C) as indicated. Densitometry quantified data of p‐FoxO3 (Thr32) to FoxO3 expression ratios, represented as a fold change to non‐stimulated cells.

Figure 2. FoxO3 knockdown in control human lung fibroblasts mimics the pro‐proliferative and myofibroblast phenotypes

1. A

   Representative Western blots of FoxO3 in control human lung fibroblasts untransfected or transfected with scramble siRNA or FoxO3 siRNA. Densitometry quantified data of FoxO3/GAPDH expression ratio (n = 3).

2. B, C

   Human lung fibroblasts from donors (n = 2–3) were transfected with scramble siRNA or FoxO3 siRNA. (B) Transfected cells were stimulated with 5% FCS, IGF‐I (200 ng/ml), and PDGF‐BB (60 ng/ml) or left non‐stimulated, and cell proliferation was measured by BrdU incorporation. (C) mRNA expression of fibroblasts to myofibroblasts markers (ACTA2, COL1A1, and COL3A1) was analyzed by qPCR.

Data information: Data are expressed as mean ± SEM. In (A and B), data are represented as a percentage of control, non‐stimulated scramble siRNA‐transfected cells. In (A), data were analyzed using repeated‐measures one‐way ANOVA. In (B), data were analyzed using one‐way ANOVA. In (C), data were analyzed using Student's t‐test. *P < 0.05, ***P < 0.001 versus non‐stimulated scramble siRNA; ^§§§ P < 0.001 versus stimulated scramble siRNA.

Figure 3. Aggravation of bleomycin‐induced lung fibrosis in mice with global‐ and fibroblast‐specific Foxo3‐knockout mice

1. A, B

   Fibroblasts were isolated from saline‐treated mice lungs at day 21 post‐instillation and mice lungs after 14 or 21 days of bleomycin instillation. (A) mRNA expression analysis of Foxo3 by qPCR (n = 4–5/group). (B, left panel) Representative Western blots of p‐FoxO3 (Thr32), FoxO3, and α‐Sma protein levels. Actb was used as a loading control. (B, right panel) Densitometry quantified data of FoxO3, p‐FoxO3 Thr32, and α‐Sma‐to‐Actb expression ratio (n = 4/group).

2. C

   Percentage of survival of mice challenged with bleomycin. WT (littermates) bleomycin (n = 6 mice), Foxo3 _f.b ^−/− (n = 6 mice) and Foxo3 ^−/− bleomycin (n = 9 mice).

3. D–F

   Lung function measurements of mice at day 14 after instillation, (D) total lung capacity, (E) lung compliance, and (F) lung tissue resistance.

4. G

   Hydroxyproline levels in mouse lungs.

5. H

   Representative H&E staining of whole left lung (upper panel) and higher magnification (lower panel). Scale bar = 100 μm.

6. I

   Fibrotic score.

Data information: Data are expressed as mean ± SEM. In (D–I) WT saline (n = 7 mice), Foxo3 ^−/− (n = 6 mice), Foxo3 _f.b ^−/− (n = 5 mice), WT bleomycin (n = 6 mice), Foxo3 _f.b ^−/− bleomycin (n = 4 mice), and Foxo3 ^−/− bleomycin (n = 7 mice). Data were analyzed using one‐way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001 versus WT saline. ^§ P < 0.05, ^§§ P < 0.01, ^§§§ P < 0.001 versus WT bleomycin.

Figure EV3. Foxo3 knockout (global‐ and fibroblast‐specific) influences immune cell composition in bleomycin‐instilled mice lungs

1. A–D

   Immunofluorescence staining was performed on WT, Foxo3 ^−/− and Foxo3 _f.b ^−/− mice lung sections (saline‐ and bleomycin‐instilled) using CD68, CD45, and CD3 antibodies. Representative pictographs depicting CD68 (A), CD45 (B), and CD3 (C) staining in green from mice (n = 3) in each group. DAPI was used as a nuclear stain. Scale bar = 50 μm. Fluorescence intensities of CD3‐stained sections (n = 5/6 per group) were quantified using ImageJ software and normalized to DAPI intensity. Data are expressed as mean ± SEM and were analyzed using repeated‐measures one‐way ANOVA, ***P < 0.001 versus WT saline group and ^§§ P < 0.01 versus WT bleomycin group. m1, m2, and m3 represent three different mice evaluated in each group.

Figure 4. UCN‐O1 inhibits human lung fibroblasts proliferation and fibroblast‐to‐myofibroblasts transdifferentiation in response to pro‐fibrotic stimuli

1. A–C

   Serum‐starved (48 h) N‐HLF (n = 3) were stimulated with 5% FCS (A), PDGF‐BB (60 ng/ml) (B), or IGF‐1 (200 ng/ml) (C) in the presence of UCN‐01 or vehicle (DMSO) as indicated and cell proliferation was measured by BrdU incorporation after 24 h. Data represent percentage of control; vehicle‐treated cells (A) or serum‐starved cells (B and C).

2. D–F

   Serum‐starved (48 h) N‐HLF (n = 3) were left non‐stimulated or stimulated with TGF‐β1 (10 ng/ml) in medium containing UCN‐01 (50 nM) or vehicle (DMSO) for an additional 24 h and the following analyses were performed: (D) mRNA expression analysis of fibroblast‐to‐myofibroblasts markers (COL1A1, COL3A1, and ACTA2) expression by qPCR. (E, left panel) Representative Western blots of collagen 1a (Col1a) and α‐SMA. GAPDH was used as a loading control. (E, right panel) Densitometry quantified data of Col1a and α‐SMA‐to‐GAPDH expression ratio. (F) ICC of α‐SMA (red, upper) and Col1a (green, lower). TO‐PRO3 (blue) was used to label nuclei. Scale bar = 50 μm.

Data information: Data are expressed as mean ± SEM. In (A–D), data were analyzed using one‐way ANOVA, n.s. = not significant, *P < 0.05, **P < 0.01, ***P < 0.001 versus vehicle‐treated cells. Data in (E) were analyzed using repeated‐measures one‐way ANOVA, *P < 0.05, ***P < 0.01 versus vehicle‐treated cells.

Figure 5. UCN‐01 inhibits FoxO3 inactivation in human lung fibroblasts stimulated with growth factors

1. A–I

   Serum‐starved (48 h) N‐HLF were stimulated with 5% FCS (A–C), IGF‐1 (200 ng/ml) (D–F), or PDGF‐BB (60 ng/ml) (G–I) in medium containing UCN‐01 (50 nM) or wortmannin (Wort.) (500 nM) or vehicle (DMSO) for 30 min and Western blot analysis was performed with the antibodies as described. (A, D, and G) Representative Western blots of p‐FoxO3 (Thr32), p‐FoxO3 (Ser253), FoxO3, p‐AKT (Thr308), AKT, and GAPDH. (B, C, E, F, H, and I) Densitometry quantified data of p‐FoxO3 (Thr32) or p‐FoxO3 (Ser253) to FoxO3 and p‐AKT (Thr308) to AKT expression ratios, represented as a fold change to non‐stimulated cells.

2. J

   ICC of FoxO3 in N‐HLF that were serum‐starved for 48 h and were stimulated with 5% FCS or PDGF‐BB (60 ng/ml) or IGF‐1 (200 ng/ml) and treated with 50 nM UCN‐01 or 500 nM wortmannin or vehicle control (DMSO) for 6 h. TO‐PRO3 (blue) was used to label nuclei. FoxO3 and TO‐PRO3 images were overlaid to visualize nuclear and cytoplasmic localization of FoxO3. Images are representative of n = 3. Scale bar = 50 μm.

Data information: Data are expressed as mean ± SEM and were analyzed using repeated‐measures one‐way ANOVA, *P < 0.05, **P < 0.01, ***P < 0.001 versus vehicle‐treated cells.Source data are available online for this figure.

Figure 6. UCN‐01 inhibits FoxO3 inactivation in human lung fibroblasts stimulated with TGF‐β1

1. A–D

   Forty‐eight hours serum‐starved N‐HLF were left non‐stimulated or stimulated with TGF‐β1 (10 ng/ml) in medium containing UCN‐01 (50 nM) or vehicle (DMSO) for 24 h. (A) mRNA expression of FOXO3 by qPCR. Data are expressed as mean ± SEM and were analyzed using one‐way ANOVA, **P < 0.01, ***P < 0.001 versus vehicle‐treated cells. (B) Representative Western blots of p‐FoxO3 (Thr32), p‐FoxO3 (Ser253), FoxO3, p‐AKT (Thr308), AKT and GAPDH. (C, D) Densitometry quantified data, represented as fold change (n = 3). In all panels, cells were serum‐starved for 48 h before stimulations. Data are expressed as mean ± SEM and were analyzed using repeated‐measures one‐way ANOVA, n.s. = not significant, *P < 0.05, **P < 0.01, ***P < 0.001 versus vehicle‐treated cells.

Source data are available online for this figure.

Figure 7. Preceding knockdown of FoxO3 reduces inhibitory effect of UCN‐01 on human lung fibroblast proliferation and myofibroblast transdifferentiation in response to pro‐fibrotic stimuli

1. N‐HLF were transfected with scramble siRNA or FoxO3 siRNA. 6 h after transfection, cells were serum‐starved for 36 h and then stimulated with 5% FCS in the presence or absence of UCN‐01 (50 nM), and cell proliferation was measured by BrdU incorporation after 24 h. Scramble siRNA‐transfected cells were left non‐stimulated for 24 h as an additional control. Data represent percentage of control, scramble siRNA non‐stimulated cells (n = 2–3). Data are expressed as mean ± SEM and were analyzed using one‐way ANOVA, **P < 0.01 versus 5% FCS‐scramble siRNA.

2. N‐HLF was transfected with scramble siRNA or FoxO3 siRNA. 6 h after transfection, cells were serum‐starved for 36 h and then stimulated with TGF‐β1 (10 ng/ml) in the presence or absence of UCN‐01 (50 nM) for 24 h and mRNA expression of COL1A1 and COL3A1 were measured with qPCR (n = 3). Data are expressed as mean ± SEM and were analyzed using one‐way ANOVA, n.s. = not significant, **P < 0.01 versus scramble siRNA, ^§ P < 0.05 versus TGF‐β1‐scramble siRNA, ^# P < 0.05 versus UCN‐01‐scramble siRNA.

Figure EV4. UCN‐01 inhibits FoxO3 nuclear exclusion of FCS‐, IGF‐1‐ or PDGF‐BB‐ stimulated IPF‐HLF

Serum‐starved (48 h) IPF‐HLF were stimulated with 5% FCS or PDGF‐BB (60 ng/ml) or IGF‐1 (200 ng/ml) and treated with 50 nM UCN‐01 or vehicle control (DMSO). After 6 h of treatment, ICC assessed FoxO3 cellular localization. Control panel represents cells that were treated for 6 h with 50 nM UCN‐01 or vehicle control (DMSO). TO‐PRO3 (blue) was used to label nuclei. FoxO3 and TO‐PRO3 images were overlaid to visualize nuclear and cytoplasmic localization of FoxO3. Scale bar = 50 μm. Images are representative of n = 3.

Figure 8. UCN‐01 inhibits human IPF lung fibroblast pathological phenotypes and attenuates lung fibrosis induced by bleomycin injury

1. A

   Serum‐starved (48 h) IPF‐HLF (n = 5) were stimulated with 5% FCS and treated with UCN‐01 (10, 50 100, and 200 nM) or vehicle (DMSO) or left untreated, and cell proliferation was measured by BrdU incorporation after 24 h.

2. B

   IPF‐HLF cells that were serum‐starved for 48 h and stimulated with TGF‐β1 (10 ng/ml) and treated with UCN‐01 (50 nM) or vehicle (DMSO) for 24 h. (B, left panel) Western blots of p‐FoxO3 (Thr32), p‐FoxO3 (Ser253), FoxO3, p‐AKT (Thr308), AKT, and GAPDH. (B, right panel) ICC of α‐SMA (red) and Col1a (green). TO‐PRO3 (blue) was used to label nuclei. Scale bar = 50 μm. Images are representative of n = 3.

3. C

   Scheme shows experimental setup.

4. D–F

   Lung function measurements of mice, (D) total lung capacity, (E) lung compliance, and (F) tissue resistance.

5. G

   Representative H&E staining of whole left lung.

6. H

   Fibrotic score.

Data information: Data are expressed as mean ± SEM. Data in (A) were analyzed using one‐way ANOVA, n.s. = not significant, ***P < 0.001 versus vehicle‐treated cells, ^§§§ P < 0.001 versus 50 nM UCN‐01‐treated cells, ^## P < 0.01 versus 100 nM UCN‐01‐treated cells. In (C–H): saline (n = 8 mice), bleomycin non‐treated (n = 7 mice), bleomycin vehicle‐treated (n = 8 mice), UCN‐01 5 mg/kg.bw (n = 10 mice), and UCN‐01 7.5 mg/kg.bw (n = 7 mice). In (D–F and H), data were analyzed using one‐way ANOVA, n.s. = not significant, *P < 0.05, **P < 0.01,***P < 0.001 versus vehicle‐treated mice.Source data are available online for this figure.

Figure 9. UCN‐01 inhibits lung fibrosis via regulation of FoxO3

1. A–C

   Lung function measurements of mice, (A) total lung capacity, (B) lung compliance, and (C) tissue resistance.

2. D

   Hydroxyproline levels in mice lungs.

3. E, F

   Stimulation with TGF‐β1 (10 ng/ml) in the presence or absence of UCN‐01 (50 nM) for 24 h, followed by (E) Western blotting for p‐FoxO3 (Thr32), FoxO3, AKT, GAPDH, and (F) qPCRs for COL1A1 and COL3A1 (n = 3).

4. G

   Model of signaling events involving FoxO3 during IPF pathogenesis. In lung fibroblasts, FoxO3 with nuclear localization leads to transcription of genes involved in regulation of apoptosis, migration, differentiation, and cell‐cycle inhibition. Pro‐fibrotic factors (TGF‐β1, IGF‐1, PDGF‐BB) bind to their receptors and activate PI3K, followed by phosphorylation of AKT. AKT phosphorylates and thereby inactivates FoxO3, followed by nuclear exclusion and degradation of FoxO3, resulting in progression of lung fibrosis. UCN‐01 inhibits AKT phosphorylation resulting in FoxO3 reactivation and attenuation of disease progression. Transforming growth factor β receptor (TGF‐βR), transforming growth factor β1 (TGF‐β1), platelet‐derived growth factor receptor (PDGFR), platelet‐derived growth factor‐BB (PDGF‐BB), insulin‐like growth factor receptor (IGFR), and insulin‐like growth factor‐1 (IGF‐1).

Data information: Data are expressed as mean ± SEM. In (A–C), WT saline (n = 13 mice), Foxo3 ^−/− saline mice (n = 6 mice), WT bleomycin vehicle‐treated (n = 14 mice), Foxo3 ^−/− bleomycin vehicle‐treated mice (n = 7 mice), WT bleomycin‐UCN‐01 7.5 mg/kg.bw (n = 11 mice), Foxo3 ^−/− bleomycin‐UCN‐01 7.5 mg/kg.bw mice (n = 5 mice). (D) Collagen deposition: WT saline (n = 4 mice), Foxo3 ^−/− saline mice (n = 6 mice), WT bleomycin vehicle‐treated (n = 5 mice), Foxo3 ^−/− bleomycin vehicle‐treated mice (n = 6 mice), WT bleomycin‐UCN‐01 7.5 mg/kg.bw (n = 5 mice), Foxo3 ^−/− bleomycin‐UCN‐01 7.5 mg/kg.bw mice (n = 5 mice). Data in (A–D) were analyzed using one‐way ANOVA, *P < 0.05, **P < 0.01, and ***P < 0.001 versus WT saline mice; ^# P < 0.05 versus WT bleomycin‐UCN‐01 mice; ^§§§ P < 0.001 versus WT bleomycin‐UCN‐01 mice. Data in (E and F) were analyzed using one‐way ANOVA, *P < 0.05, ***P < 0.001 versus TGF‐β1‐EV and ^§ P < 0.05 versus UCN‐01‐EV.

Figure EV5. Anti‐proliferative effects of UCN‐01 are mediated via FoxO3

1. A, B

   N‐HLF was transfected with empty vector (EV) or AKT mutant (AKT mut) plasmid. 6 h after transfection, cells were serum‐starved for 36 h and then stimulated with 5% FCS in the presence or absence of UCN‐01. From the above‐treated samples, after 24 h, Western blots [p‐FoxO3 (Thr32), FoxO3, AKT, AKT mut, GAPDH] and cell proliferation measurements (BrdU incorporation) were performed. Data represent percentage of control, EV non‐stimulated cells (n = 3). Data are expressed as mean ± SEM and were analyzed using one‐way ANOVA, *P < 0.05, **P < 0.01, and ***P < 0.001 versus 5% FCS‐EV, and ^§§§ P < 0.001 versus UCN‐01‐EV.