Flavones hydroxylated at 5, 7, 3' and 4' ameliorate skin fibrosis via inhibiting activin receptor-like kinase 5 kinase activity

Abstract

Skin fibrosis is mainly characterized by excessive collagen deposition. Studies have recently identified a number of flavonoids with variable structures that have the potency of inhibiting collagen synthesis and thus attenuating organ fibrosis. In this study, we found that flavones with 5, 7, 3', 4' hydroxy substitution reduced collagen expression most efficiently. Among those flavones, luteolin, quercetin, and myricetin were selected for follow-up. In vivo, the three compounds ameliorated skin fibrosis and reduced collagen deposition. Further analysis showed the compounds had significant inhibition on the proliferation, activation and contractile ability of dermal fibroblasts in vitro and in vivo. More importantly, we revealed that luteolin, quercetin, and myricetin selectively downregulated the phosphorylation of Smad2/3 in TGF-β/Smads signaling via binding to activin receptor-like kinase 5 (ALK5) and impairing its catalytic activity. We also found flavones with 5, 7, 3', 4' hydroxy substitution showed stronger affinity with ALK5 compared with other flavonoids. Herein, we identified at least in part the underlying molecular basis as well as the critical structures that contribute to the antifibrotic bioactivity of flavones, which might benefit drug design and modification.

Fig. 1. Correlation between hydroxy substitution in flavones and their effectiveness on Col1a2 and Col3a1 inhibition.

Flavones with no less than four hydroxyls induce significant inhibition on Col1a2 and Col3a1 expression compared with vehicle in a 109 candidate flavones and b 39 positive compounds. c The frequency of hydroxy substitutions in 39 positive compounds. d The effectiveness of flavones with different hydroxy substitutions in 39 positive compounds. Flavones containing 5, 7, 3′, 4′ hydroxy substitution exhibits significantly stronger inhibition on collagen expression compared with flavones with other structures in e 109 candidate flavones and f 39 positive compounds. Data are the mean ± SD (three independent experiments); **P < 0.01, ***P < 0.001

Fig. 2. Luteolin, quercetin or myricetin ameliorates skin fibrosis and excessive collagen deposition in vivo.

a Representative images of fibrosing skin tissue in bleomycin-induced skin fibrosis model from DMSO-treated mice and drug-treated mice after hematoxylin and eosin (H&E), Masson’s trichrome and Picrosirius red staining. b Quantitative analysis of dermal thickness and collagen density of fibrosing skin from vehicle-treated mice and drug-treated mice in bleomycin-induced skin fibrosis model. c Representative images of macroscopy, H&E and Picrosirius red staining of hypertrophic scar (HS) tissue in mechanical load-induced HS model from DMSO-treated mice and drug-treated mice. d Quantitative study of gross scar, cross-section area, and collagen density of HS tissue from DMSO-treated mice and drug-treated mice in HS model. Data are the mean ± SD (three independent experiments); *P < 0.05, **P < 0.01, ***P < 0.001

Fig. 3. Luteolin, quercetin or myricetin inhibits fibroblast proliferation in vivo and in vitro.

Representative images of hypertrophic scar (HS) in a fibrosing skin in bleomycin-induced model and b mechanical load-induced HS model from DMSO-treated mice and drug-treated mice stained for proliferating cell nuclear antigen (PCNA) and quantitative analysis of PCNA-positive cells. c Representative immunofluorescence images and quantitative results of EdU incorporation assay in human dermal fibroblasts (HDFs). EdU is shown by red fluorescence and nucleus by blue fluorescence. d Representative images and quantitative results of colony formation assay exhibit an evidently reduced number of HDF colony (purple) after drug treatment. Data are the mean ± SD (three independent experiments); *P < 0.05, ***P < 0.001

Fig. 4. Luteolin, quercetin or myricetin suppresses the activation and contractile behavior of dermal fibroblast.

Representative images and quantitative analysis of fibrosing tissues from in vivo models (a bleomycin-induced model; b mechanical load-induced model) stained for α-SMA show significant suppression induced by luteolin, quercetin or myricetin on fibroblast activation. c Immunofluorescence staining for α-SMA and F-actin in cultured human dermal fibroblasts (HDFs) after incubation with TGF-β1 (5 ng/ml) and DMSO or drugs for 48 or 72 h. α-SMA is shown by red fluorescence and F-actin by green fluorescence. d Representative images and quantitative results of collagen gel contraction assay show that luteolin, quercetin or myricetin significantly inhibited HDF-mediated gel contraction with exogenous TGF-β1 (5 ng/ml) treatment. Data are the mean ± SD (three independent experiments); **P < 0.01, ***P < 0.001

Fig. 5. Luteolin, quercetin or myricetin downregulates the phosphorylation of Smad2/3.

a−c Western blotting and densitometric analysis demonstrate that luteolin, quercetin or myricetin dose-dependently inhibits the expression of phosphorylated Smad2/3 without affecting the other factors in TGF-β/Smads signaling pathway or Smad1/5/9 in BMP signaling. d Representative images of immunofluorescence staining of human dermal fibroblasts exhibit a significant reduction in TGF-β1-induced Smad3 (green) translocation into the nuclei. e Luteolin, quercetin or myricetin has no significant impact on the expression of TGF-β1 based on ELISA assay. f Luteolin, quercetin or myricetin does not induce notable effect on the expression of TGFβRI or TGFβRII according to western blotting. Data are the mean ± SD (three independent experiments); *P < 0.05, **P < 0.01, ***P < 0.001

Fig. 6. 5, 7, 3′, 4′ hydroxyl-containing flavones selectively bind to activin receptor-like kinase 5 (ALK5).

a Schematic of computational docking simulation of 5, 7, 3′, 4′-hydroxylated flavones interaction with the ATP-binding site of ALK5. Amino acid residues interacting with the compounds are shown (yellow) with the hydrogen bonds indicated by dashed lines. b Mean docking score of 5, 7, 3′, 4′ hydroxyl-containing flavones vs. flavones with other structures. c Docking simulation of luteolin, quercetin, and myricetin binding to the catalytic site of ALK5. Amino acid residues interacting with the compounds are indicated and hydrogen bonds are represented with straight dashed lines. **P < 0.01

Fig. 7. Luteolin, quercetin or myricetin binds directly to activin receptor-like kinase 5 (ALK5) and hinders ALK5-mediated phosphorylation of Smad3.

a Dose−response curve of LanthaScreen^TM Eu kinase binding assay demonstrated a dose-dependent decrease in emission ratios induced by luteolin, quercetin or myricetin. Luteolin, quercetin or myricetin inhibited ALK5 activity with IC₅₀ values of 1.062, 1.480, and 4.083 μM, respectively. b In vitro phosphorylation of GST-Smad3 by ALK5 was significantly reduced by luteolin, quercetin or myricetin. Data are the mean ± SD (three independent experiments)