Ablation of the decorin gene enhances experimental hepatic fibrosis and impairs hepatic healing in mice

Abstract

Accumulation of connective tissue is a typical feature of chronic liver diseases. Decorin, a small leucine-rich proteoglycan, regulates collagen fibrillogenesis during development, and by directly blocking the bioactivity of transforming growth factor-β1 (TGFβ1), it exerts a protective effect against fibrosis. However, no in vivo investigations on the role of decorin in liver have been performed before. In this study we used decorin-null (Dcn-/-) mice to establish the role of decorin in experimental liver fibrosis and repair. Not only the extent of experimentally induced liver fibrosis was more severe in Dcn-/- animals, but also the healing process was significantly delayed vis-à-vis wild-type mice. Collagen I, III, and IV mRNA levels in Dcn-/- livers were higher than those of wild-type livers only in the first 2 months, but no difference was observed after 4 months of fibrosis induction, suggesting that the elevation of these proteins reflects a specific impairment of their degradation. Gelatinase assays confirmed this hypothesis as we found decreased MMP-2 and MMP-9 activity and higher expression of TIMP-1 and PAI-1 mRNA in Dcn-/- livers. In contrast, at the end of the recovery phase increased production rather than impaired degradation was found to be responsible for the excessive connective tissue deposition in livers of Dcn-/- mice. Higher expression of TGFβ1-inducible early responsive gene in decorin-null livers indicated enhanced bioactivity of TGFβ1 known to upregulate TIMP-1 and PAI-1 as well. Moreover, two main axes of TGFβ1-evoked signaling pathways were affected by decorin deficiency, namely the Erk1/2 and Smad3 were activated in Dcn-/- samples, whereas no significant difference in phospho-Smad2 was observed between mice with different genotypes. Collectively, our results indicate that the lack of decorin favors the development of hepatic fibrosis and attenuates its subsequent healing process at least in part by affecting the bioactivity of TGFβ1.

Figure 1

Hepatic fibrosis is accentuated and healing is delayed in Dcn−/− mice. (a) Picrosirius-stained sections from the liver of wild-type and Dcn−/− mice without treatment (first column), after 4 months of thioacetamide treatment (second column), and 4 months after the withdrawal of the thioacetamide (third column). Scale bar = 150 μm. (b) Quantification of hepatic fibrosis by morphometric analysis of livers from wild-type and Dcn−/− animals as indicated. Values represent the mean (%) ± SEM of three animals. All the values from 2 months to 4+4 months have P<0.01. (c) Quantification of hydroxyproline in whole liver extracts of wild type and decorin-null livers. Values represent the mean (%) ± SEM of n=3; *P<0.05.

Figure 2

Changes in collagen type I and decorin in hepatic fibrosis and its healing in Dcn−/− mice. (a) Double immunostaining for collagen I and decorin in representative liver samples from all the various experimentally treated animals as indicated. Collagen is shown as red, decorin as green, nuclei are stained with DAPI. Representative Wt and Dcn−/− controls are shown in Supplementary Figure S2. Scale bar = 200 μm. (b) Quantification of collagen I content in livers from wild-type and decorin-null mice. Results are expressed as mean intensity of dots after densitometry. Wild-type non-treated control was considered as 100%. Values represent the mean ± SD of three animals (*P<0.05). (c) Quantification of decorin levels by immunoblotting of total liver lysates from wild-type untreated mice, and from mice treated for four months with TA or for an additional four months in TA-free conditions (**P<0.01).

Figure 3

Elevation of collagen I, III and IV mRNA levels in the liver of Dcn−/− animals following hepatic injury. (a–c) Quantification of mRNA levels for procollagens type I, III and type IV as determined by quantitative real-time RT PCR analysis. The data represent mean ±SD of n=3 *P<0.05. Levels are expressed in folds, normalized to endogenous β-actin amplified in parallel.

Figure 4

MMP elevation during hepatic fibrosis. (a) Zymogram using liver homogenates of wild-type and Dcn−/− mice without treatment (control), at 4 months of thioacetamide treatment (TA4), and at 4 months after the withdrawal of the chemical (TA4+4). (b,c) Quantification of zymograms as those shown in panel A using scanning densitometry. The values represent the mean intensity (%) ±SD of two experiments run in triplicate; **P<0.01

Figure 5

Elevation of TIMP-1 and PAI-1 mRNA levels in Dcn−/− liver following experimental hepatic fibrosis. (a,b) Real-time RT-PCR analysis of TIMP-1 andPAI-1 mRNA in whole liver of wild-type and decorin-null animals as indicated. The designations are the same as those of Figure 1.The data represent mean ±SD of n=3; **P<0.01 and *P<0.05. The relative expression levels were calculated by using 2^−ΔΔCT method, using β-actin as endogenous control.

Figure 6

Genetic ablation of decorin gene enhances the expression of the TGFβ1-inducible early responsive gene (TIEG) and TGFβ1 in experimental hepatic fibrosis. Real-time RT-PCR analysis of (a) TGFβ1-inducible early response gene (TIEG) and (b) TGFβ1 expression in whole liver extracts. The designations are the same as those of Figure 1. The data represent mean ±SD of three experiments run in triplicate and are presented as expression relative to endogenous β-actin levels; *P<0.05, **P<0.01.

Figure 7

Amounts of Erk1/2, phospho-Erk1/2 (P-Erk1/2), phospho-Smad2 (P-Smad2), phospho-Smad3 (P-Smad3) and α−smooth muscle actin (αSMA) as detected by Western immunoblotting (a). Pictures are representative of immunoblotting analyses of liver homogenates from two independent experiments run in triplicate. The genotype is indicated at the top and the various categories are listed at the bottom, including animals with no treatment (control), following a 4-month exposure to thioacetamide (TA4), and at the end of the regeneration period (TA4+4). β-actin served as loading control. (b) Densitometric quantification of immunoblottings for αSMA P-Smad3, P-p44 and P-p42. The values were normalized on β-actin levels. Data are expressed as mean ±SD of duplicate experiments run in triplicate; **P<0.01.

Figure 8

Analysis of LX2 cell activation by TGFβ. (a) Quantitative real-time RT-PCR analysis of procollagen type 1, and α-smooth muscle actin mRNA levels normalized on GAPDH. The data represent the mean ±S.D. of duplicate experiments run in triplicates. (b) Determination of collagen 1 protein amount by dot blot in medium of LX2 cells with normal (scr) or silenced decorin content (siDcn) with or without TGFβ1 treatment. (c) Diagram of the relative decorin proteoglycan content from media of cells transfected with either scrambled or decorin-specific siRNAs. The data represent the mean ±SD of duplicate experiments run in triplicates; *P<0.05, **P<0.01.

Figure 9

Changes in α-smooth muscle actin (αSMA) in liver sections and in LX2 cells. (a) Immunostaining of αSMA on sections of wild type (Wt) and decorin-null (Dcn−/−) livers after 4 months of TA-exposure and at the end of the regression period (TA4+4). (b) Immuncytochemistry of αSMA performed on cells transfected with scrambled (scr) siRNA or with siRNA specific for decorin (siDcn) in the absence or presence of TGFβ1 for 48 h. αSMA is shown as red and nuclei are counterstained with DAPI. Scale bar = 100 μm for each picture.

Figure 10

Western blot analysis of α-smooth muscle actin (αSMA), phospho-Smad2 (P-Smad2) and phospho-Smad3 (P-Smad3) in LX2 hepatic stellate cells (a). Pictures are representative of immunoblotting analyses of cell lysates from two independent experiments run in triplicate. (b) Quantification of αSMA, P-Smad2 and P-Smad3 content normalized on GAPDH levels. *P<0.05.