Apoptosis

, Volume 19, Issue 1, pp 30–41

MicroRNA-29a protects against acute liver injury in a mouse model of obstructive jaundice via inhibition of the extrinsic apoptosis pathway

Authors

  • Mao-Meng Tiao
    • Department of PediatricsKaohsiung Chang Gung Memorial Hospital and the Graduate Institute of Clinical Medical Sciences, Chang Gung University College of Medicine
  • Feng-Sheng Wang
    • Department of Medical ResearchKaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine
  • Li-Tung Huang
    • Department of PediatricsKaohsiung Chang Gung Memorial Hospital and the Graduate Institute of Clinical Medical Sciences, Chang Gung University College of Medicine
  • Jiin-Haur Chuang
    • Department of SurgeryKaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine
  • Ho-Chang Kuo
    • Department of PediatricsKaohsiung Chang Gung Memorial Hospital and the Graduate Institute of Clinical Medical Sciences, Chang Gung University College of Medicine
  • Ya-Ling Yang
    • Department of AnesthesiologyKaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine
    • Department of PediatricsKaohsiung Chang Gung Memorial Hospital and the Graduate Institute of Clinical Medical Sciences, Chang Gung University College of Medicine
Original Paper

DOI: 10.1007/s10495-013-0909-4

Cite this article as:
Tiao, M., Wang, F., Huang, L. et al. Apoptosis (2014) 19: 30. doi:10.1007/s10495-013-0909-4

Abstract

Recent studies have shown that microRNA-29 (miR-29) is significantly decreased in liver fibrosis, as demonstrated in human liver cirrhosis, and that its downregulation influences the activation of hepatic stellate cells. In addition, both cleaved caspase-3 production and apoptosis play a role in cholestatic liver injury. However, it is unknown if miR-29 is effective in modulating the extent of injury. We employed miR-29a transgenic mice (miR-29aTg mice) and wild-type (WT) littermates to clarify the role of miR-29 in hepatic injury and fibrogenesis, using the bile duct-ligation (BDL) mouse model. After BDL, all three members of the miR-29 family were significantly downregulated in the livers of WT mice, and miR-29b and miR-29c were significantly downregulated in the livers of the miR-29aTg mice. Liver function, as measured by alanine transaminase and aspartate transaminase activity, was significantly improved in the miR-29aTg mice than in the WT littermates, following 1 week of obstructive jaundice. In addition, overexpression of miR-29a was associated with a significant downregulation of the expression of collagen-1α1, collagen-4α1, phospho-FADD, cleaved caspase-8, cleaved caspase-3, Bax, Bcl-2, PARP, and nuclear factor-κB, as well as an upregulation of phospho-AKT expression. In addition, there were significantly fewer TUNEL-positive liver cells in the miR-29aTg group than in the WT littermates after BDL. Our results indicate that miR-29a decreases cholestatic liver injury and fibrosis after BDL, at least partially, by modulating the extrinsic rather than intrinsic pathway of apoptosis.

Keywords

CaspaseMir-29CholestasisTUNEL

Abbreviations

BDL

Bile duct ligation

ECM

Extracellular matrix

HSC

Hepatic stellate cells

LDH

Lactate dehydrogenase

TLR

Toll like receptor

TUNEL

TdT-mediated dUTP biotin nick-end labeling

WT

Wide type

XIAP

X-linked inhibitor of the apoptotic protein

Introduction

Persistent liver injury because of cholestasis and hepatitis may result in liver fibrosis involving a range of cell types [1, 2]. Hepatic stellate cells (HSC) become activated and undergo morphologic and functional trans-differentiation into contractile myofibroblastic cells responsible for the production of extracellular matrix (ECM) in the injured liver [13]. During liver fibrogenesis, collagen type I is the predominant isoform deposited into the perisinusoidal space. However, collagen type IV, which constitutes less than 10 % of total collagen in the normal liver, is most dramatically upregulated in fibrosis [4]. The mechanisms underlying the regulation of ECM gene expression in activated HSCs are of interest as potential therapeutic targets.

Recent studies have shown that the expression of microRNA (miR)-132 and miR-29, which consists of miR-29a, miR-29b, and miR-29c, is significantly decreased in fibrotic livers, as demonstrated in human liver cirrhosis as well as in two different models of liver injury induced by bile duct ligation (BDL) and carbon tetrachloride (CCl4). This downregulation influences HSC activation [46]. In addition, serum levels of miR-29a are significantly lower in patients with advanced liver cirrhosis than in healthy controls or patients with early fibrosis [5]. Moreover, overexpression of miR-29 in murine HSCs resulted in downregulation of collagen expression, including collagen-1α1 and collagen-1α2 [5, 7], through a mechanism directly targeting the mRNA expression of ECM genes. These findings suggest that miR-29 may be an important therapeutic target in chronic liver disease.

Obstructive jaundice, which is associated with bacterial translocation, cytokine activation, inflammatory cell infiltration, and endotoxemia, results in significant morbidity and mortality [8, 9]. Previous investigations have strongly suggested that Fas-mediated hepatocyte apoptosis [10] and cleaved caspase-3 production play central roles in cholestatic liver injury [1113]. We have also observed that expression of the pro-apoptotic regulator Bax was increased after BDL but could be down-regulated by steroid treatment [14]. In contrast, miR-29 may promote apoptosis of hepatocellular carcinoma via the inhibition of two anti-apoptotic molecules, Bcl-2 and Mcl-1[15]. However, the biological role of miR-29 in BDL-induced cholestatic liver injury has not been addressed. Thus, we hypothesized that miR-29 might protect from cholestatic liver injury via modulation of the apoptosis pathway and liver fibrosis. Thus, we employed miR-29a transgenic mice (miR-29aTg mice) to clarify the roles of miR-29a in hepatic injury and fibrogenesis, as well as to elucidate the association of miR-29a with the apoptotic pathway in a mouse model of obstructive jaundice [16].

Materials and methods

Construction and breeding of miR-29a transgenic mouse colony

First, we first successfully developed the miR-29a transgenic mice (FVB/miR-29aTg). Briefly, human phosphoglycerate kinase (PGK) promoter and human miR-29a precursor full-length cDNA were cloned from the cDNA tissue library using PCR protocols. A bioinformatics search indicates that mouse miR-29a (GGAUGACUGAUUUCUUUUGGUGUUCAGAGUCAAUAGAAUUUUCUAGCACCAUCUGAAAUCGGUUAUAAUGAUUGGGGA) is high homology with human miR-29a (AUGACUGAUUUCUUUUGGUGUUCAGAGUCAAUAUAAUUUUCUAGCACCAUCUGAAAUCGGUUAU). Bioinformatics also reveal that putative mRNAs targeted by mouse miR-29a (putative 6228 mRNA targets) are highly similar to human miR-29a (putative 6777 mRNA targets). The cDNA was inserted into the pUSE expression vector, and the linear hPGK-miR-29a-BGH poly-A cDNAs were cloned and purified. All PCR products and cDNA clones were validated by gene sequencing and matched via the National Center for Biotechnology Information (NCBI) bio-information station. The designed constructs were transferred into fertilized eggs from the FVB/N mice. The eggs were next transferred into ICR foster mothers. We increased the colony size for further studies. Animal housing was in a temperature-controlled, pathogen-free environment with a 12-hour light/dark cycle; all mice had free access to water and standard food. The treatment of the mice followed a protocol that was approved by the Animal Ethics Committee of Chang Gung Memorial Hospital. The body weight of each mouse was checked before the following procedures were conducted.

Animal model and experimental protocol

FVB male mice (National Animal Center of Academia Sinica, Taipei, Taiwan) weighing 25–35 g, were used for all of the experiments and were maintained on standard laboratory mice food in a 12-h light/dark cycle. The mice were randomly divided into either the OJ group or the sham group, depending on whether the mice had received ligation or sham ligation of the common bile duct, as described in a previous study [17]. The mice were killed 1 week after the procedure. Blood was drawn for liver function tests, and the liver was harvested. Liver tissues were snap-frozen so that mRNA and protein expression could be determined later. The samples were kept at −80 °C until the biochemical analyses were conducted.

Blood biochemistry

Serum samples were analyzed using a Hitachi 7600 modular chemistry analyzer (Hitachi High-Technologies Corporation, Tokyo, Japan) in order to determine the levels of alanine aminotransferase (ALT) [18], aspartate aminotransferase (AST) [5], lactate dehydrogenase (LDH), and total bilirubin, which were used as indices of hepatocellular injury.

RNA isolation and real-time quantitative reverse transcription polymerase chain reactions

In order to quantify the amount of miR-29 in the tissue samples, we performed real-time, quantitative reverse-transcription PCR with the ABI 7700 Sequence Detection System (TaqMan; Applied Biosystems, Inc., Foster City, CA). Total microRNA was isolated using MicroRNA Isolation kits (BioChain Institute, Inc, Hayward, CA), according to the manufacturer’s instructions. U6 gene expression was used to normalize gene and microRNA expression. Templates were preamplified using 2× TaqMan® PreAmp Master Mix and 10× MegaplexTM PreAmp Primers and then PCR-amplified using 2× TaqMan® Universal PCR Master Mix. The relative quantification of the gene expression was based on the comparative threshold cycle (CT) method in which the amount of the target was determined to be 2−(ΔCT target−ΔCT calibrator) or 2−ΔΔCT. The PCR products were then electrophoresed on a 2 % agarose gel in order to confirm the amount of the products. Validation experiments were performed in duplicate, and amplification efficiencies were validated.

Western blot analysis

Liver-homogenized crude proteins (30 μg) were treated with sample buffer, boiled for 10 min, separated using 12 % sodium dodecyl sulfate–polyacrylamide gels, and transferred to a nitrocellulose membrane. Blots were incubated with the primary antibodies against collagen-1α1 (#sc-8784-R, Santa Cruz Biotechnology, Dallas, TX), collagen-4α1 (#AP7369a, ABGENT, CA), phospho-Fas-associated death domain (FADD; #2785, Cell Signaling, MA), FADD (#3523-S, Epitomics, CA), cleaved caspase-3 (#9661, Cell Signaling), cleaved caspase-8 (#4790, Cell Signaling), cleaved caspase-9 (#9504, Cell Signaling), Bax (#2772, Cell Signaling), Bcl-2 (#2876, Cell Signaling), cleaved-PARP (#9542, Cell Signaling), AKT (#9272, Cell Signaling), p-AKT (#9275, Cell Signaling), p65 (#13752 Cayman, MI), phospho-p65 (#3033, Cell Signaling), XIAP (#2042, Cell Signaling) and GAPDH (#2251-1, Epitomics). After washing with TBST and incubating with horseradish peroxidase-coupled anti-rabbit immunoglobulin-G antibodies (dilution, 1:10,000) at room temperature for 2 h, the blots were developed with enhanced chemiluminescence detection (GE Healthcare Bio-sciences AB, Uppsala, Sweden) and exposed to film. The signals were quantified with densitometry.

Immunohistochemistry

Immunohistochemical staining of the caspases was performed on the paraffin-embedded, formalin fixed archival liver tissue. There were six liver specimens from each group. The 2-μm sections were deparaffinized, treated with 3 % hydrogen peroxide to inactivate the endogenous peroxidase activity, and microwaved for 10 min in 10-mM citrate buffer to retrieve the antigen. The sections were then incubated with rabbit anti-caspase 3(ab4051, abcam, UK, 1:100 dilution), and eight antibody (sc-5263, Santa Cruz Biotechnology, Dallas, 1:100 dilution) at 4 °C overnight and detected by the second antibody with the SuperPicTure™Polymer detection kit (#87-8963, Zymed Laboratories, South San Francisco, CA) and DAB chromogen (# K3467, DAKO, CA), according to the manufacturer’s instructions. The control slides were incubated without treatment with a primary antibody.

Immunofluorescence staining

Liver tissues were embedded in TissueTek® OCTTM (optimal cutting temperature) compound (Sakura Finetek) and frozen at −80 °C for storage. Frozen sections (4-μm thick) were prepared using a cryostat (CM3050 S, Leica). Cryosections were fixed with isotonic PBS and 4 % paraformaldehyde solution for 1 h. To block non-specific background staining, the samples were incubated in a solution containing 1 % BSA for 30 min. After washing with PBS, the slides were incubated with the primary antibodies. An anti-collagen-1α1 primary antibody (1:2,000 dilution; Sigma–Aldrich) was used. Alexa Fluor® 488- and Alexa Fluor® 594-conjugated secondary antibodies (Molecular Probes) were used. Samples were co-stained with DAPI (4′,6-diamidino-2-phenylindole; Molecular Probes) to facilitate visualization of the nuclei. The stained cells were mounted with fluorescent mounting medium (Dako Cytomation) and visualized using Olympus microscopy. All of the exposure gains and rates were consistent among samples. The fluorescence intensities were quantified on independent color channels.

TdT-mediated dUTP biotin nick-end labeling assay

Immediately after sacrifice, representative sections were cut from the mid-right lobe of the liver in the mice. Sections were obtained from six mice in each group. After formalin fixation, these paraffin-embedded liver sections were stained and evaluated. Double-strand DNA breaks were detected by TdT-mediated dUTP biotin nick-end labeling (TUNEL), according to the method used in our previous study, with some modifications [19]. We used the ApopTag Plus Peroxidase In Situ Apoptosis Detection Kit (CHEMICON International, Inc.) for the TUNEL assay.

Deparaffinized sections were washed with distilled water and treated with Protein Digestion Enzyme for 15 min at 37 °C. After washing with three changes of PBS, the sections were treated with a TdT solution, incubated with 3 % hydrogen peroxide for 5 min to block endogenous peroxidase activity, and then treated with peroxidase-conjugated antibody for 10 min at room temperature. After washing in PBS, the reacted sections were immersed in 30, 30-diaminobenzidine solution with 3 % hydrogen peroxide and counterstaining with methyl green to facilitate visualization of the nick-end labeling.

Statistical analysis

All values in the figures and tables are expressed as mean ± standard error. Quantitative data were analyzed using the 1-way analysis of variance [20] when appropriate. The least significant difference (LSD) test was used for post hoc testing when appropriate. Correlations between quantitative variables were assessed using Pearson’s coefficient. Two-sided p values less than 0.05 were considered statistically significant.

Results

miR-29 family members were downregulated in the mouse model of obstructive jaundice

Mice with miR-29a overexpression were viable and reproduced normally. The behavior of the transgenic mice was not noticeably different from that of controls, and there were no overt lesion noted in the liver of any of the mice. Our observation also consistence with the findings that mice with reduced hepatic miR-29ab1 expression were reproduced normally and no overt lesion noted in the liver [21]. These observations indicate that liver development is not dependent on hepatic expression of miR-29a. miR-29 family members showed a striking relationship to numerous genes encoding collagen and other ECM proteins and were downregulated in liver fibrosis [5]. One week after BDL, miR-29a, b, and c in WT mice (p = 0.006, p < 0.001, and p < 0.001, respectively), as well as miR-29b and miR-29c in miR-29aTg littermates (p = 0.003, and p < 0.001, respectively), were significantly downregulated (Fig. 1). Furthermore, we found no reduction in miR-29a expression between sham-operated and BDL miR-29aTg mice.
https://static-content.springer.com/image/art%3A10.1007%2Fs10495-013-0909-4/MediaObjects/10495_2013_909_Fig1_HTML.gif
Fig. 1

Downregulation of miR-29 during murine liver fibrosis. One week after bile duct ligation, expressions of miR-29a, miR-29b, and miR-29c in wild-type (WT) mice (p = 0.006, p < 0.001, and p < 0.001, respectively), as well as miR-29b and miR-29c in miR-29aTg littermates (p = 0.003 and p < 0.001, respectively), were significantly downregulated. Data are expressed as mean ± standard error in 8 samples from miR-29a transgenic mice and 6 samples from WT littermates. *p < 0.05 between the groups

Over-expression of miR-29a significantly reduced hepatocellular damage in cholestatic mice

Ligation of the common bile duct significantly increased total bilirubin levels, AST, ALT, and LDH in miR-29aTg mice and WT littermates (all p < 0.001, Table 1). No statistical significance was obtained in total bilirubin levels, AST, and ALT between the miR-29aTg mice and the WT littermates in the group that underwent the sham operation. With overexpression of miR-29a, levels AST and ALT were reduced significantly (by 39 and 38 %) in the miR-29aTg mice compared with the WT littermates in the BDL group (p = 0.002 and p < 0.001, respectively). However, there was no statistically significant difference in LDH levels between miR-29aTg mice and WT littermates in the BDL group (1644.2 ± 260.9 vs. 1354.3 ± 158.5, p = 0.244). This indicated that overexpression of miR-29a protected against cholestatic liver damage.
Table 1

Comparison of the changes of liver functions in the four groups

Group

Number

Bilirubin (mg/dl)

AST (IU/l)

ALT (IU/l)

LDH (IU/l)

Sham/WT

n = 6

0.60 ± 0.02

82 ± 11

36 ± 5

515 ± 46

BDL/WT

n = 6

9.26 ± 0.57a

966 ± 126a

686 ± 71a

1354 ± 159a

Sham/miR-29aTg

n = 8

0.61 ± 0.01

84 ± 7

36 ± 8

612 ± 62

BDL/miR-29aTg

n = 8

9.40 ± 0.91a,c

623 ± 59a,b,c

442 ± 40a,b,c

1644 ± 260a,c

Data expressed as mean ± standard error

WT wide type, BDL bile duct ligation, Tg transgenic, AST aspartate aminotransferase, ALT alanine aminotransferase, LDH lactate dehydrogenase

ap < 0.05 versus Sham/WT

bp < 0.05 versus BDL/WT

cp < 0.05 versus Sham/miR-29aTg

Fibrosis in WT and miR-29a transgenic mice after BDL

To investigate the effect of miR-29a overexpression on the progression of fibrosis, we studied the expression of ECM proteins during hepatic fibrogenesis. Using western blot analysis, we found significantly higher expression of collagen-1α1 and collagen-4α1 protein in tissues from the BDL group than in tissues from the sham group (both p < 0.001; Fig. 2) in WT mice. However, there was a weaker induction of collagen-1α1 and collagen-4α1 in miR-29aTg mice with cholestasis (p = 0.236 and p = 0.841, respectively). Moreover, miR-29a overexpression significantly downregulated collagen-1α1 and collagen-4α1 protein expression in miR-29aTg mice with cholestasis compared with the WT littermates (both p < 0.001). To further characterize collagen-1α1 protein expression, in vivo immunofluorescence staining was performed. As illustrated in Fig. 2c and Supplement Fig. S1, compared with the sham-operation group (a), the BDL group (b) of WT mice exhibited stronger collagen-1α1 immunoreactivity (red). Moreover, miR-29a overexpression (c, d) significantly downregulated collagen-1α1 immunoreactivity in the miR-29aTg mice with cholestasis (d) compared with the WT littermates, which indicated that miR-29a might have an impact on fibrogenesis in early cholestasis.
https://static-content.springer.com/image/art%3A10.1007%2Fs10495-013-0909-4/MediaObjects/10495_2013_909_Fig2_HTML.gif
Fig. 2

Overexpression of miR-29a in the murine model resulted in downregulation of collagen-1α1 (A) and collagen-4α1 (B) in the liver of mice after bile duct ligation. There were significantly higher expressions of collagen-1α1 and collagen-4α1 protein in tissues from the BDL group than in tissues from the sham-operated group (both p < 0.001) in WT mice. Moreover, miR-29a overexpression significantly downregulated collagen-1α1 and collagen-4α1 protein expression in miR-29aTg mice with cholestasis compared with the WT littermates (both p < 0.001). C For immunofluorescence analysis of collagen-1α1 expression in BDL and sham-operated mice, sections of livers from both WT and miR-29aTg mice were stained with collagen-1α1 (red) and processed with DAPI (blue) for nuclear localization. Compared with the sham-operated group (a), the BDL group (b) of WT mice exhibited stronger collagen-1α1 immunoreactivity. Moreover, miR-29a overexpression (c, d) significantly downregulated collagen-1α1 immunoreactivity in miR-29aTg mice with cholestasis (d) compared with the WT littermates. Data are expressed as mean ± SE standard error in 8 samples from miR-29a transgenic mice and 6 samples from WT littermates. *p < 0.05 between the groups (Color figure online)

Overexpression of miR-29a significantly decreased apoptotic proteins and increased anti-apoptotic proteins after BDL

Next, we next investigated the expression of the pro-apoptotic members of FADD, Bax, and PARP, as well as the anti-apoptotic members of Bcl-2, Mcl-1, Bcl-xL and Akt, between the groups. In the early stage of BDL, we found significantly higher expression of the pro-apoptotic proteins phospho-FADD, Bax, and cleaved-PARP, as well as anti-apoptotic proteins of Bcl-2 and phospho-Akt, in tissue from the BDL group than from the sham group in WT mice (p < 0.001, p < 0.001, p = 0.018, p < 0.001, and p < 0.001, respectively; Figs. 3, 4). Notably, compared with WT mice, the miR-29aTg littermates showed significantly downregulated phospho-FADD, Bax, Bcl-2, and cleaved-PARP, as well as upregulated phospho-Akt expressions in cholestasis (p < 0.001, p < 0.001, p < 0.001, p = 0.002, and p < 0.001, respectively). At the same time, the overexpression of miR-29a significantly inhibited Bcl-2 protein expression in BDL mice (p < 0.001), as well as marginally inhibited the expression in sham-operated mice (p = 0.15), which was consistent with the finding that Bcl-2 is a direct target of miR-29 [15]. However, over-expression of miR-29a did not change the Mcl-1 and Bcl-xL expressions after BDL (data not shown).
https://static-content.springer.com/image/art%3A10.1007%2Fs10495-013-0909-4/MediaObjects/10495_2013_909_Fig3_HTML.gif
Fig. 3

Comparison of FADD (a), Bax (b), and PARP (c) expression by western blot in liver tissue of miR-29a transgenic mice and WT littermates that underwent sham operation and bile duct ligation (BDL). There were significantly higher expressions of pro-apoptotic proteins phospho-FADD, Bax, and cleaved-PARP in tissues from the BDL group than from the sham-operated group in WT mice (p < 0.001, p < 0.001, and p = 0.018, respectively). Notably, miR-29aTg mice with cholestasis showed significantly downregulated expressions of phospho-FADD, Bax, and cleaved-PARP compared to the expressions in the WT littermates (p < 0.001, p < 0.001 and p = 0.002, respectively). Data are expressed as mean ± standard error in 8 samples from miR-29a transgenic mice and 6 samples from WT littermates. *p < 0.05 between the groups

https://static-content.springer.com/image/art%3A10.1007%2Fs10495-013-0909-4/MediaObjects/10495_2013_909_Fig4_HTML.gif
Fig. 4

Comparison of a Bcl-2 and b Akt expression by western blot in liver tissue of miR-29a transgenic mice and WT littermates underwent sham operation and bile duct ligation (BDL). There were higher expressions of the anti-apoptotic proteins Bcl-2 and phospho-Akt, in tissues from the BDL group than from the sham-operated group in WT mice (p < 0.001 and p < 0.001, respectively). Notably, miR-29aTg mice with cholestasis showed significantly downregulated Bcl-2 and up-regulated phospho-Akt expression compared with WT littermates (both p < 0.001). Data are expressed as mean ± standard error in 8 samples from miR-29a transgenic mice and 6 samples from WT littermates. *p < 0.05 between the groups

Overexpression of miR-29a significantly decreased apoptotic cell death after BDL

To further assess whether apoptosis is involved in this liver damage, the activation of the apoptotic machinery was measured by using the expression of caspases and the extent of TUNEL staining. The expression of caspase-8, caspase-9, and caspase-3 increased significantly after BDL than after the sham operation (p = 0.003, p < 0.001, and p < 0.001, respectively; Fig. 5). Moreover, after BDL, the overexpression of miR-29a significantly downregulated caspase-8 and caspase-3 expression (p < 0.001, and p < 0.001, respectively) but had no impact on caspase-9 expression (p = 0.998). To investigate the correlation of caspase-8 and caspase-9 with caspase-3, we performed a set of general linear models between the groups. Univariate analysis showed that the caspase-8 and caspase-9 levels were positively and significantly correlated with the caspase-3 levels (R = 0.721, p < 0.001 and R = 0.490, p = 0.021, respectively). To show the effect of overexpression of miR-29a, we correlated the levels of caspase-8 and caspase-9 with those of caspase-3 in miR-29aTg mice. Interestingly, there was a significant correlation only between caspase-8 and caspase-3 expression; there was no correlation between caspase-9 and caspase-3 expression in miR-29aTg mice (R = 0.780, p = 0.005 and R = 0.286, p = 0.423, respectively). Both showed significant correlation in WT littermates (R = 0.829, p = 0.002 and R = 0.673, p = 0.023, respectively). To further characterize caspase-8 and caspase-3 protein expression in the liver in response to cholestasis, immunohistochemical staining was carried out. There was significantly higher caspase-8 (Supplement Fig. S2) and caspase-3 (Fig. 5d) immunoreactivity in the liver tissue of the sham-operated and BDL groups than in the control group, which was expressed constantly in the cytoplasm of hepatocytes (arrowhead), bile ductular epithelial cells (small arrow), and some nonparenchymal cells morphologically identical to Kupffer cells (large arrow). Moreover, overexpression of miR-29a decreased the expression of caspase-8 and caspase-3 after BDL. The trend of caspase-3 and caspase-8 protein expression was the same as that of the western blot results. Cytochrome c is often released from mitochondria during the early stages of apoptosis of intrinsic pathway [22]. However, there was no significant difference between miR-29aTg mice and WT littermates in cholestasis (data not shown). Furthermore, we confirmed by TUNEL staining that apoptosis was significantly elevated in the WT mice after BDL and was decreased by the overexpression of miR-29a in the miR-29aTg mice (Fig. 6). Finally, our results indicated that miR-29a decreases cholestatic liver injury and fibrosis after BDL, at least in part, via modulation of the extrinsic rather than the intrinsic pathway of apoptosis.
https://static-content.springer.com/image/art%3A10.1007%2Fs10495-013-0909-4/MediaObjects/10495_2013_909_Fig5_HTML.gif
Fig. 5

Comparison of cleaved-caspase 8 (A), caspase-9 (B), and caspase-3 (C) expression by western blot in liver tissue of miR-29a transgenic mice and WT littermates that underwent sham operation and bile duct ligation (BDL). The expressions of caspase-8, caspace-9, and caspase-3 increased significantly after BDL than after sham operation (p = 0.003, p < 0.001, and p < 0.001, respectively). Moreover, the overexpression of miR-29a significantly downregulated the expression of caspase-8 and caspase-3 (p < 0.001 and p < 0.001, respectively), but had no impact on that of caspase-9, after BDL (p = 0.998). D There was significantly higher caspase-3 immunoreactivity in the liver tissue of the sham-operated and BDL groups than in the control group, which was expressed constantly in the cytoplasm of hepatocytes (arrowhead), bile ductular epithelial cells (small arrow), and some nonparenchymal cells morphologically identical to Kupffer cells (large arrow). Moreover, overexpression of miR-29a decreased the expression of caspase-3 after BDL. Data are expressed as mean ± standard error in 8 samples from miR-29a transgenic mice and 6 samples from WT littermates. *p < 0.05 between the groups

https://static-content.springer.com/image/art%3A10.1007%2Fs10495-013-0909-4/MediaObjects/10495_2013_909_Fig6_HTML.gif
Fig. 6

Apoptosis by TdT-mediated dUTP biotin nick-end labeling (TUNEL). A (a), The TUNEL stain in wild-type littermates that underwent sham operation; (b) the increased liver cell apoptosis (arrow) in wild-type littermates that underwent bile duct ligation; (c), miR-29a transgenic mice that underwent sham operation; (d), the TUNEL-positive liver cells in the miR-29a Tg group after BDL were significantly less than those in the WT littermates. B Data are expressed as mean ± standard error in 8 samples from miR-29a transgenic mice and 6 samples from WT littermates. *p < 0.05 between the groups

Overexpression of miR-29a significantly decreased NF-κB after BDL

Obstructive jaundice, which has been shown to be associated with endotoxemia, results in enhanced NF-κB activation and transcriptional activation of pro-inflammatory cytokines [23, 24]. In order to assess whether overexpression of miR-29a affects NF-κB activation, we investigated the expression of phospho-p65 between groups. As shown in Fig. 7, we found significantly higher expression of phospho-p65 in tissue from the BDL group than in tissue from the sham-operated group of WT mice (p = 0.002). Notably, miR-29aTg mice showed significantly downregulated expression of phospho-p65 compared to the expression in the WT littermates (Fig. 7b; p = 0.002).
https://static-content.springer.com/image/art%3A10.1007%2Fs10495-013-0909-4/MediaObjects/10495_2013_909_Fig7_HTML.gif
Fig. 7

Overexpression of miR-29a significantly decreased NF-κB after BDL. There were significantly higher expressions of phospho-p65 in tissues from the BDL group than from the sham-operated group of WT mice (p = 0.002). Notably, compared with WT mice, miR-29aTg littermates showed significantly downregulated phospho-p65 in B (p = 0.002). Data are expressed as mean ± standard error in 8 samples from miR-29a transgenic mice and 6 samples from WT littermates. *p < 0.05 between the groups

Discussion

In the present study, we first demonstrated that overexpression of miR-29a in cholestatic mice significantly inhibited hepatocellular damage and liver fibrosis; it also induced a significant decrease in the levels of the pro-apoptotic proteins phospho-FADD, Bax, PARP, cleaved caspase-8, and caspase 3, and caused a significant increase in the level of the anti-apoptotic protein phospho-AKT, as well as a significant decrease in NF-κB, thereby leading to a significant decrease in cellular injury and apoptosis. In order to identify that miR-29a plays a role in the functional regulation of an apoptotic pathway, we showed that the overexpression of miR-29a significantly downregulated caspase-8 and caspase-3 expression, but had no impact on caspase-9 expression. In the cholestatic mice model, as there was a good correlation of caspase-8 expression, but not caspase 9, with caspase-3 expression in miR-29aTg mice. In addition, overexpression of miR29a significantly downregulated phospho-FADD protein expression in the extrinsic apoptotic pathway, but there was no change in the cytochrome c and X-linked inhibitor of the apoptotic protein, which binds to and inhibits caspase-9 expression in the intrinsic apoptotic pathway [25] (Supplement Fig. S3). Finally, as expected, caspase-8, but not caspase-9, was predicted to be an miR-29a target by miRBase (http://microna.sanger.ac.uk) and was more likely to be repressed by miR-29a overexpression. Together, these findings indicate that overexpression of miR-29a might decrease apoptosis after BDL, at least in part, via modulation of the extrinsic rather than intrinsic pathway of apoptosis.

Liver fibrosis is a complex process modulated by a set of signaling pathways. It is well known that the stimulation of HSC by transforming growth factor-β (TGF-β) is a crucial event in liver fibrogenesis because of its impact on myofibroblast transition and ECM induction. TGF-β secreted by hepatocytes, Kupffer cells, and sinusoidal endothelial cells causes HSCs to activate, transdifferentiate, and secrete ECM [26]. Moreover, Roderburg et al. reported that in vitro activation of HSCs led to a downregulation of all miR-29-members during 8 days of culturing and that TGF-β mediated the downregulation of miR-29 in HSCs [5]; this effect appeared to be specific to HSCs, because TGF-β treatment of Huh-7.5 cells did not downregulate miR-29 expression [7]. Moreover, overexpression of miR-29 in murine HSCs resulted in downregulation of collagen expression [5, 7] through directly targeting the mRNA expression of ECM genes. In the contrast, another study has shown increased fibrosis and mortality in miR29ab1-knockout mice following the administration of CCl4 [21]. Our results confirmed the antifibrogenic role of miR-29a. In addition, miR-29 plays an important role of sex differences in fibrotic responses. In a study by Zhang et al., miR-29a and miR-29b were significantly downregulated in the livers of male but not female mice after 4 weeks of CCl4 treatment, and estrogen was shown to inhibit CCl4-induced hepatic injury through the induction of hepatic miR-29 [27]. Estrogen enhanced the expression of miR-29a/b through suppression of the NF-κB signal pathway, which regulates a decrease in miR-29 expression [27]. In a similar fashion, we demonstrated that overexpression of miR-29a in male mice inhibited acute liver injury and reduced fibrosis in response to cholestatic liver injury.

To date, little is known about how miR-29 functions in blocking HSC activation. In recent studies of renal fibrosis, it was demonstrated that Smad3 mediated TGF-β1-induced downregulation of miR-29 by binding to the promoter of miR-29 [28, 29]. Furthermore, miR-29 acted as a downstream inhibitor and therapeutic miRNA for TGF-β1/Smad3-mediated renal fibrosis. However, a study by Suh et al. showed that the TGF-ß signaling pathway is unlikely to be responsible for the reduction in miR-29 expression between quiescent and proliferative dermal fibroblasts [28]. Thus, it is still unknown whether liver fibrosis follows TGF-β1/Smad3-mediated liver fibrosis. Recently, we were the first to demonstrate that obstructive jaundice is associated with the downregulation of toll-like receptor 7 (TLR7) expression and upregulation of profibrogenic gene expression in HSCs [29]. Selective activation of TLR7 may modulate the profibrogenic phenotype in activated HSCs with cholestatic liver injury [29]. Interesting, miR-29a also can function by binding as a ligand to TLR7 and can trigger a TLR-mediated prometastatic inflammatory response that may ultimately lead to tumor growth and metastasis [30]. These findings raise the question of whether miR-29a can regulate the profibrogenic phenotype in activated HSCs via the TLR7 signal pathway. If so, miR-29a may be an important therapeutic target in chronic liver disease. Further study will be necessary to investigate this area.

FADD is a central element in the Fas-mediated cell death pathway [31]. FADD binds specifically to Fas and caspase-8 through its death domain to form a death-inducing signaling complex during apoptosis. On the contrary, activation of the phosphatidylinositol-3 kinase-Akt pathway has been demonstrated as an important cell-surviving signal in a variety of cell types [32]. Our data suggest that miR-29a decreases the levels of the pro-apoptotic proteins phospho-FADD, Bax, cleaved PARP, caspase-8, and caspase-3 and significantly increases the level of the anti-apoptotic protein phospho-AKT, which leads to a decrease in apoptosis as measured by TUNEL staining. However, Xiong Y et al. showed that miR-29 sensitizes hepatocellular carcinoma cells to apoptosis by targeting two antiapoptotic molecules, Bcl-2 and Mcl-1 [15]. In keeping with the results of Xiong Y et al., we also found that over-expression of miR-29a resulted in lower levels of Bcl-2 in miR-29aTg mice than in WT littermates. Similarly, Hatano et al. reported that dominant negative FADD, not overexpression of Bcl-xL, significantly reduced TUNEL-positive cells after BDL [33]. These results indicate that hepatic injury shortly after BDL involves a Fas-mediated and Bcl-xL-insensitive apoptotic pathway [33]. Thus, the prosurvival effect of miR-29a in cholestatic liver injury reflects how miR-29a is differently regulated in cancer cells and normal hepatocytes in response to external stimulation.

Although some studies have strongly suggested that hepatocellular cholestatic injury is due to Fas-mediated hepatocyte apoptosis [10], others have concluded that necrosis, rather than apoptosis, represents the main mechanism of hepatocyte death in chronic cholestasis [34]. In a previous study [13] and in this study, we found that the levels of LDH, a marker of cellular necrosis [35], were increased in both miR-29aTg and WT mice after BDL. However, there was no statistical difference between miR29aTg mice and WT littermates after BDL, which indicated that overexpression of miR-29a decreases cholestatic liver injury through modulation of the apoptotic pathway rather than by necrosis.

Obstructive jaundice has been shown to be associated with the activation of NF-κB, with transcriptional activation of pro-inflammatory cytokines [23, 24]. Intraperitoneal injection of LPS into mice resulted in significant downregulation of all miR-29 members, as well as stimulation of primary HSCs and primary hepatocytes, with downregulation of miR-29 via NF-κB [5]. Here, we showed that overexpression of miR-29a was associated with significant downregulation of NF-κB in obstructive jaundice. In view of the role of NF-κB in inflammation, the inhibition of NF-κB activation could contribute to the anti-inflammatory effects of miR-29a, thereby decreasing cellular apoptosis in early cholestasis after BDL. However, recently, the overexpression of miR-29 has been shown to alter genes not predicted to be miR-29 targets or to leave genes unchanged that were predicted to be miR-29 targets by TargetScan in fibroblasts [28]. This is because microRNAs normally do not act in linear signaling cascades but instead are able to integrate signals from distinct upstream signaling pathways [36].

This study had some potential limitations. First, we did not study the effects of miR-29a on the expression of pro- and anti-apoptotic proteins in each different hepatic cell type after BDL. Second, we did not study which type of cells in liver are responsible for the majority of the change in pro- and anti-apoptotic proteins, all of which are important and require further investigation. Third, further studies will be required to evaluate the mechanisms by which miR-29 blocks HSC activation and a direct target gene which has responsible for resistance to apoptosis in BDL animal model.

In summary, overexpression of miR-29a significantly reduced the expression of pro-apoptotic proteins and enhanced phospho-AKT protein expression, thereby leading to a decrease in cellular apoptosis, liver injury, and fibrosis in cholestasis.

Acknowledgments

The authors thank Yuan-Ting Chuang for her assistance in this study. This study was supported by a grant from the Chang Gung Memorial Grant CMRPG 8B0991 and 8B1461, Taiwan.

Conflict of interest

The authors have no conflict of interest to declare.

Supplementary material

10495_2013_909_MOESM1_ESM.tif (77 kb)
Fig. S1Quantitative measurement of collagen-1α1 expression by immunofluorescence staining in liver tissue of miR-29a transgenic mice and WT littermates underwent sham operation and bile duct ligation (BDL). Data are expressed as mean ± standard error in 4 samples from miR-29a transgenic mice and 4 samples from WT littermates (TIFF 76 kb)
10495_2013_909_MOESM2_ESM.tif (1.9 mb)
Fig. S2Immunoreactive caspase 8 staining between miR-29aTg mice and wild type littermates. There was significantly higher caspase-8 immunoreactivity in the liver tissue of the sham-operated and BDL groups than in the control group, which was expressed constantly in the cytoplasm of hepatocytes (arrowhead). Moreover, overexpression of miR-29a decreased the expression of caspase-8 and caspase-3 after BDL (TIFF 1900 kb)
10495_2013_909_MOESM3_ESM.tif (595 kb)
Fig. S3Comparison of the X-linked inhibitor of the apoptotic protein (XIAP) expression by western blot in liver tissue of miR-29a transgenic mice and WT littermates that underwent sham operation and bile duct ligation (BDL). There were significantly downregulated XIAP protein expression in both miR-29aTg mice and WT littermates after BDL. However, overexpression did not change of XIAP protein expression after BDL. Data are expressed as mean ± standard error in 8 samples from miR-29a transgenic mice and 6 samples from WT littermates (TIFF 595 kb)

Copyright information

© Springer Science+Business Media New York 2013