|Home | About | Journals | Submit | Contact Us | Français|
Liver cirrhosis, which is caused by the accumulation of extracellular matrix materials, is a serious clinical problem that can progress to hepatic failure. Transforming growth factor‐β (TGFβ) plays a pivotal role in extracellular matrix production, but bone morphogenetic protein (BMP)‐7, a member of the TGFβ superfamily, can antagonise the fibrogenic activity of TGFβ.
In this study, we examined whether adenovirus‐mediated overexpression of BMP‐7 (Ad‐BMP‐7) antagonised the effect of TGFβ in vitro and in vivo.
In primary cultured rat stellate cells and the LX‐2 human stellate cell line, induction of BMP‐7 by Ad‐BMP‐7 infection decreased the expression of collagen 1A2 mRNA and smooth muscle α‐actin in the presence or absence of TGFβ, via Smad 1/5/8 phosphorylation. BMP‐7 triggered the mRNA expression of inhibitors of differentiation 2 (Id2) in LX‐2. Although endogenous expression of BMP‐7 was hardly detectable, Smad1 and Id2 overexpression increased BMP‐7 expression in LX‐2. A liver fibrosis model was induced by the repetitive intraperitoneal injection of thioacetamide (200 mg/kg body weight) twice per week for up to 7 weeks. In rats administered Ad‐BMP‐7 via the tail vein, hydroxyproline content and the areas stained by Sirius red dye in the liver were significantly reduced compared to controls. Ad‐Id2 also reduced fibrosis.
These data demonstrate that BMP‐7, Smad 1/5/8 and Ids interact to antagonise hepatic fibrogenesis.
Hepatic stellate cells (HSC) are included in the group of sinusoidal constituent cells that play multiple roles in the liver. In the normal liver, HSC in the space of Disse surround the sinusoids as pericytes and store vitamin A‐associated lipid droplets.1 When the liver parenchyma suffers from chronic injury due to disease, elimination of damaged hepatocytes leads HSC to become activated by multiple peptide, lipid and gaseous mediators that are released from hepatocytes, Kupffer cells, endothelial cells and infiltrating inflammatory cells.2
HSC activation accompanies phenotypic transformation into myofibroblast‐like cells, as represented by the expression of smooth muscle α‐actin (αSMA) and by the production of extracellular matrix materials (ie, collagens, fibronectin, laminin and proteoglycans), thereby contributing to the progression of fibrosis in the chronically damaged liver.3,4 HSC activation is driven particularly by transforming growth factor‐β (TGFβ).5 Activated HSC produce TGFβ1 which stimulates their own collagen gene expression in an autocrine loop. TGFβ1 also upregulates tissue inhibitors of metalloproteinases6 that exhibit anti‐apoptotic effects towards myofibroblast‐like cells.7 In addition, a contribution of TGFβ1 to organ fibrogenesis has recently been suggested through induction of epithelial‐mesenchymal transition (EMT) in the injured kidney, which process was suppressed by bone morphogenic protein (BMP)‐7.8
BMPs, which belong to the TGFβ superfamily, are 30–35 kDa hetero‐ or homodimeric proteins originally identified as signals capable of inducing ectopic cartilage and bone upon non‐systemic injection in mammals.9 Several studies have since demonstrated that these proteins play essential roles during embryonic development10,11 and some BMP KO mice exhibit embryonic lethality.12,13 However, the functions of BMPs in adult organ tissues have not been fully elucidated. Recently, BMP‐7 was reported to counteract some of the profibrogenic action of TGFβ in renal mesangial cells14 and in the injured kidney.8 BMPs usually do not affect the growth of epithelial cells,15 although they exhibit weak antiproliferative effects towards thyroid carcinoma cells.16 The activation of BMP receptors by their ligands leads to the phosphorylation and activation of Smad1, Smad5 and Smad8 and to their heterodimerisation with Smad4, followed by their translocation into the nucleus, resulting in suppression of the nuclear accumulation of Smad3 by TGFβ1 stimulation.17
We have hypothesised here that introduction of BMP‐7 by adenoviral infection might suppress the profibrogenic response of the liver. We show that adenoviral gene transfer of BMP‐7 suppresses the expression of type I collagen in cultured HSC and in an in vivo thioacetamide (TAA)‐induced liver fibrosis model via upregulation of an inhibitor of differentiation 2 (Id2), which is a negative regulator for basic helix‐loop‐helix (bHLH) transcription factor. Furthermore, we show that adenoviral gene transfer of Id2 attenuates liver fibrosis in vivo.
Pathogen‐free male Wistar rats were obtained from SLC (Shizuoka, Japan). Animals were housed at a constant temperature and had free access to laboratory chow and water. Procedures were performed under the control of the animal care committee of Osaka City University in accordance with the Guideline on Animal Experiments in Osaka City University and the Japanese Government Animal Protection and Management Law.
We used the Adenovirus Cre/loxP Kit (Takara Biomedicals, Tokyo, Japan) and the Adenovirus Expression Vector Kit (Takara Biomedicals) to generate two types of recombinant adenovirus vector. One adenovirus vector expresses the Cre recombinase, AxAw‐CAG‐Cre, under the control of the CAG (a fusion of the cytomegalovirus enhancer and the chicken β‐actin promoter plus part of the 3′ untranslated region of rabbit β‐globin) promoter. The other vector contains a potent expression unit that is activated by Cre recombinase‐mediated removal of an upstream loxP‐flanked ”stuffer” sequence. As a consequence, co‐transfection of these two adenovirus vectors to either cells or animals induces transcription of the down‐stream transgenes of interest (BMP‐7, Id2, Smad6 and Smad1). Cosmid pAxCALNL‐GFP, ‐LacZ, ‐BMP‐7, ‐Id2, ‐Smad6 and ‐Smad1 were constructed by the insertion of green fluorescent protein (GFP) cDNA, LacZ cDNA, mouse BMP‐7 (mBMP‐7) cDNA, mouse Id2 cDNA, mouse Smad1 cDNA and mouse Smad6 cDNA, respectively, into the Swa I cloning site of cosmid pAxCALNLw.18 GFP and LacZ were used to validate this Cre/loxP system in vitro and in vivo. Using the COS‐TPC method,19 recombinant adenoviruses of CAG‐Cre, LNL‐GFP, LNL‐LacZ, LNL‐BMP‐7 and LNL‐Id2 were generated by transfecting the cosmids of individual adenoviruses into HEK‐293 cells as described in the manufacturer's protocol. The titre of the recombinant adenoviruses was measured by the 50% tissue culture infectious dose (TCID50) method.20 Each adenovirus vector was used at the concentration of 1×109 PFU in vivo and the concentration of 50 multiplicity of infection (MOI) PFU/cell number) in vitro.
HSC were isolated from male Wistar rats and transgenic mice, which expressed a construct that includes the sequences from −17 kb to +54 bp of the mouse pro α2(I) collagen gene which was cloned upstream of the firefly luciferase reporter gene, as previously described.21 Isolated HSC were suspended in Dulbecco's modified Eagle's medium (DMEM; Sigma, St. Louis, MO) supplemented with heat‐inactivated 10% fetal calf serum (FCS; Sigma), 100 IU/ml penicillin and 100 μg/ml streptomycin, and were plated at 5×105 cells/cm3 on non‐coated 35 mm dishes or six‐well plastic plates (Falcon 3001 or Falcon 3046, respectively, Becton Dickinson, Flanklin Lakes, NJ). The human hepatic stellate cell line LX‐222 was cultured in DMEM containing 100 IU/ml penicillin and 100 μg/ml streptomycin and supplemented or not with 10% FCS.
Primary cultured HSC were co‐infected with 50 MOI of LNL‐GFP, LNL‐LacZ, LNL‐BMP‐7 or LNL‐Id2 together with 50 MOI of CAG‐Cre for 1 h at day 4 of primary culture and incubated in serum‐free DMEM for 48 h. They were then incubated in the presence or absence of TGFβ1 (5 ng/ml) for another 24 or 48 h. Subconfluent LX‐2 cells were infected with 50 MOI of CAG‐Cre and 50 MOI of LNL‐mBMP‐7, LNL‐Smad6, LNL‐Smad1, LNL‐Id2 or LNL‐GFP.
Hepatic fibrosis was induced in rats by intraperitoneal injection of TAA (200 mg/kg body weight; Wako, Osaka, Japan) twice per week for 4 weeks. Rats then each received 2×109 PFU of adenovirus vectors via the tail vein and were repetitively injected with TAA for 3 days, 7 days or 3 weeks as shown in fig 11.. Liver tissues were either snap‐frozen in liquid nitrogen or fixed in 4% buffered paraformaldehyde for immunostaining.
Fixed liver tissues were dehydrated and then embedded in Polybed. Sections were cut at 4 μm thickness. Following deparaffinisation and hydration, the sections were stained for 1 h in 0.1% (w/v) Sirius red (Direct Red 80, Aldrich, Milwaukee, WI) in a saturated aqueous solution (about 1.2% w/v) of picric acid (Wako, Osaka, Japan). After staining, the slides were rinsed for 30 min in 0.01 N HCl to remove unbound dye. Following dehydration by an alcohol series, the slides were mounted and observed under a light microscope (HC 2500, Fuji Photo Film, Tokyo, Japan). For semiquantitative analysis of liver fibrosis, ten fields from each slide were randomly selected, recorded and the red stained area per total area (mm2/mm2) was measured using Macscope Analyzer (Mitani, Fukui, Japan).
A microwave‐based antigen retrieval method was used to enhance antigenic capture. Fixed liver samples from rats infected with both LNL‐mBMP‐7 and CAG‐Cre were embedded in paraffin. Antigen unmasking of deparaffinised sections (4 μm thickness) was accomplished by heating in phosphate‐buffered saline buffer in a microwave oven for 15 min. The sections were then allowed to cool for 20 min. Immunohistochemistry for anti‐BMP‐7 was performed according to the method previously described.23 The sections were blocked against endogenous peroxidase using 3% H2O2. The sections were treated first with a 1:1000‐diluted goat polyclonal anti‐BMP‐7 (Sigma) and then with horseradish peroxidase‐conjugated anti‐immunoglobulins. Immunocomplexes were visualised by 3,3′‐diaminobenzidine tetrahydrochloride (DAB; Wako, Osaka, Japan). The tissue sections were counterstained with haematoxylin for 10 s. BMP‐7 expression was documented using an HC 2500 camera (Fuji Photo Film).
Wet liver samples (100 mg) were subjected to acid hydrolysis to determine the amount of hydroxyproline as previously described.24 The data were expressed as hydroxyproline (mg)/wet liver weight (g).
Samples were homogenised in a lysis buffer (50 mM HEPES, 150 mM NaCl, 1.5 mM MgCl2, 5 mM EDTA, 10% glycerol, 1% Triton X‐100, 0.5% Nonidet P‐40, 10 mM NaF, 1 mM Na3VO4, pH 7.5) supplemented with a protease inhibitor cocktail from Roche (Summerville, NJ). To determine nuclear protein, cells were fractionated into cytoplasm and nuclei by using a ProteoExtract Subcellular Proteome Extraction kit (Merck Biosciences) according to the manufacturer's protocol. After heat denaturation at 95°C for 5 min, the samples (10 or 20 μg protein) were subjected to sodium dodecyl sulfate‐polyacrylamide gel electrophoresis and then transferred onto a nitrocellulose membrane (Bio‐Rad, Richmond, CA). The membranes were subsequently treated with 5% skim milk and then with primary antibodies overnight at 4°C. The following antibodies were used: anti‐phosphorylated Smad 1/5/8 antibody (1:1000 dilution; Cell Signaling Technology, Beverly, MA), anti‐collagen α2 (I) antibody (1:1000 dilution; Santa Cruz Biotechnology, Santa Cruz, CA), anti‐αSMA antibody (1:1000 dilution; Sigma), anti‐Smad3 antibody (1:1000 dilution; Zymed, San Francisco, CA), anti‐histone H1 antibody (1:1000 dilution; MBL, Nagoya, Japan), anti‐GAPDH antibody (1:25000 dilution; Chemicon, Temecula, CA). After vigorous washing, the membranes were incubated with secondary horseradish peroxidase‐conjugated anti‐immunoglobulins for 2 h at room temperature. After washing, immunoreactive bands were visualised by using an enzyme‐linked chemiluminescence detection reagent (Amersham, Buckinghamshire, UK) and documented using an LAS 1000 analyser (Fuji Photo Film).
Cell extracts were prepared using Cell Culture Lysis Reagent (Promega, Madison, WI). The luciferase activity was measured according to the manufacturer's protocol (Promega) using a luminometer (Wallac 1420 Victor 2 multilabel counter system; PerkinElmer, Waltham, MA) and normalised for protein concentrations measured by DC Protein Assay (Bio‐Rad, Hercules, CA). Values are the means±SD of five independent experiments.
Expression of collagen 1A2 (COL1A2), BMP‐7 and Id2 mRNAs was measured by a reverse transcription polymerase chain reaction (RT‐PCR) method using TaqMan One‐Step RT‐PCR Master Mix Reagents and the Applied Biosystems Prism 7700 system (PE Applied Biosystems, Foster City, CA) according to the previously reported procedure.25 Total RNA was isolated from HSC or whole liver tissues using ISOGEN (Nippon Gene, Tokyo, Japan). Primers and the oligonucleotide probe used are shown in table 11.. Individual gene expression was normalised by GAPDH mRNA expression. As a standard reaction, cDNA corresponding to 200, 40, 8, and 1.6 ng of total RNA from one sample was examined and used as a reference value. We analysed each sample by using 100 ng of total RNA. The conditions of RT‐PCR were as follows: 20 min at 50°C (stage 1, reverse transcription), 10 min at 95°C (stage 2, reverse transcription inactivation and AmpliTaq Gold activation), and then 40 cycles of amplification for 15 s at 94°C and 1 min at 60°C (stage 3, PCR).
Values reported in the figures are the means±SD of three or more independent samples. The results were analysed by unpaired Student's t test. Statistical significance was set at p<0.05.
We first observed that endogenous BMP‐7 gene expression was hardly detectable in primary cultured HSC. As shown in fig 2A2A,, adenoviral gene transfer of BMP‐7 (Ad‐BMP‐7) augmented BMP‐7 mRNA expression, which was much greater than in Ad‐LacZ treated HSC. As shown in fig 2B2B,, BMP‐7 overexpression decreased COL1A2 mRNA in the presence or absence of 5 ng/ml of TGFβ.
To investigate whether BMP‐7 activates a Smad 1/5/8 signal, Western blot analysis for phospho‐Smad 1/5/8 was performed. The phosphorylation of Smad 1/5/8 was upregulated by exogenously added BMP‐7 in the presence or absence of TGFβ1 stimulation (fig 2C2C).). BMP‐7 decreased the protein level of type I collagen and αSMA in primary cultured HSC (fig 2C2C).). To investigate the effect of BMP‐7 on collagen promoter activity, HSC were isolated from transgenic mice harbouring the COL1A2 upstream sequence fused to luciferase, and luciferase activity was assayed. As shown in fig 2D2D,, COL1A2 promoter activity was inhibited by BMP‐7.
We further examined the effect of BMP‐7 using LX‐2, a human HSC cell line. As shown in fig 3A3A,, BMP‐7 decreased the expression of COL1A2 mRNA in LX‐2, which is consistent with the data obtained from primary cultured rat HSC (fig 2B2B).). The inhibitory effect of BMP‐7 on COL1A2 expression was blocked by Smad6 overexpression (fig 3A3A).). These data clearly demonstrate that BMP‐7 decreased the expression of COL1A2 via Smad 1/5/8 phosphorylation. BMP‐7 also inhibited nuclear localisation of Smad3 in LX‐2 cells (fig 3B3B).
Id2 is known to be upregulated by BMP‐7 and to antagonise Smad 2/3 signalling. We utilised LX‐2 cells to confirm this effect in HSC. As shown in fig 3C3C,, overexpression of BMP‐7 increased Id2 expression in LX‐2 cells. Smad1 overexpression also increased Id2 expression. Expression of BMP‐7 was barely detectable in untreated LX‐2 cells. However, both Smad1 and Id2 overexpression increased BMP‐7 expression in LX‐2 cells (fig 3C3C).). Id2 overexpression decreased the expression of COL1A2 mRNA in LX‐2 cells (data not shown).
Based on these in vitro data, we hypothesised that BMP‐7 inhibits the development of fibrosis in TAA‐treated rat liver. Liver fibrosis was induced by intraperitoneal injection of TAA (200 mg/kg body weight) twice per week for 4 weeks. Rats then each received 2×109 PFU of CAG‐Cre and LNL‐BMP‐7 via the tail vein and were repetitively injected with TAA until sacrifice. Expression of BMP‐7 was evaluated by immunohistochemistry of TAA‐treated rat livers that were co‐infected with Ad‐CAG‐Cre and Ad‐LNL‐BMP‐7. As shown in fig 4A4A,, BMP‐7 was expressed mainly in hepatocytes, sinusoidal cells and fibroblasts of periportal regions 3 days after adenovirus injection (fig 4A4A).). In these livers, expression of αSMA was already suppressed as early as 3 days after Ad‐BMP‐7 infection (fig 4B4B).). We confirmed that expression of BMP‐7 mRNA peaked at 3 days after injection, as shown in fig 4C4C,, and returned to basal levels by 3 weeks after injection. Ad‐BMP‐7 induced Id2 mRNA expression that peaked at 3 days and returned to basal levels 3 weeks after injection (fig 4D4D).
Repetitive injection of TAA for 7 weeks induced prominent hepatic fibrosis. A single administration of Ad‐CAG‐NCre and Ad‐LNL‐BMP‐7 via the tail vein 4 weeks after the initiation of TAA injection considerably reduced the extent of fibrosis, as confirmed by morphometric analysis of Sirius red‐stained liver sections (fig 5A,B5A,B)) and quantitative estimation of hydroxyproline content in the liver (fig 5C5C).). At this time point, αSMA and collagen expression in the liver was persistently decreased (fig 5D,E5D,E).
Finally, we examined whether Id2 suppressed fibrosis in TAA‐treated rat liver. Liver fibrosis was induced by repetitive injection of TAA for 7 weeks. A single administration of Ad‐CAG‐NCre and Ad‐LNL‐Id2 via the tail vein 4 weeks after the initiation of TAA injection reduced the extent of fibrosis, as confirmed by morphometric analysis of Sirius red‐stained liver sections (fig 6A,B6A,B)) and quantitative estimation of hydroxyproline content in the liver (fig 6C6C)) similarly to the effects of Ad‐BMP‐7. Expression of COL1A2 mRNA was also reduced by Ad‐Id2 (fig 6D6D).
Liver fibrosis is largely dependent on local TGFβ activity, since TAA and carbon tetrachloride‐induced hepatic fibrosis is attenuated by the administration of Smad7‐expressing adenovirus26 and soluble decoy proteins for TGFβ receptor, respectively.27 We had therefore predicted that BMP‐7, which reportedly antagonises the effects of TGFβ in renal fibrosis,8 might have a similar effect against hepatic fibrosis. Indeed, the present study clearly demonstrated that BMP‐7 overexpression by Ad‐BMP‐7 attenuated the expression of COL1A2 mRNA, collagen protein and αSMA in primary cultured HSC and suppressed liver fibrosis induced in rat liver by TAA treatment.
Activation of BMP receptor phosphorylates Smad1, Smad5 and Smad8, as shown in fig 2C2C,, whereas Smad2 and Smad3 are phosphorylated by TGFβ receptor‐dependent signalling. One of the mechanisms by which BMP‐7 antagonises TGFβ signalling is by interference of phospho‐Smad 1/5/8 by Smad3 for specific DNA binding and/or recruitment of co‐factors.17 Another mechanism is by enhancing Smad7 expression via Smad1 signalling.28 In addition, independently of Smad, BMP‐7 can initiate intracellular pathways which can activate Jun N‐terminal kinase, p38 kinase and PI3 kinase/Akt pathways, which can protect hepatocytes.29,30
Id proteins, which were upregulated in HSC by BMP‐7 stimulation ((figsfigs 3 and 44),), play important roles in the biological function of BMP‐7.31 Id proteins were identified as negative regulators of bHLH transcription factors. Subsequently, retinoblastoma (Rb) and Ets family members were also shown to interact with Id proteins.32,33 They act as positive regulators for cell proliferation and negative regulators for cell differentiation.34 They lack a basic DNA binding region but possess an HLH dimerisation motif, allowing them to interact with the ubiquitous bHLH transcription factor.32 As a result, these heterodimers containing Id proteins are unable to bind DNA. COL1A2 promoter activity is regulated by Smad3, Sp1 and Ets transcription factors.35 As mentioned above, Id proteins associate with Ets transcription factors and shut down their activity.34 Therefore, Id proteins might regulate COL1A2 promoter activity by inhibiting Ets. Alternatively, although Ids are not oncogenic proteins, the potential of Ids for epithelial cell (hepatocyte) proliferation may also contribute to ameliorating the fibrotic change in a manner similar to the effect of hepatocyte growth factor (HGF).36
Collagen synthesis in HSC and fibroblasts in the liver is regulated at the transcriptional and post‐transcriptional levels.37 In the present study, we demonstrated that BMP‐7 directly affected the COL1A2 promoter activity of HSC (fig 2D2D).). Our recent study using pulmonary myofibroblasts also demonstrated that BMP‐7 decreased COL1A2 promoter activity to ~70% of normal, whereas BMP‐7 lowered COL1A2 mRNA expression to 25–50% of normal.38 BMP‐7 is thus presumed to influence the stability and degradation of COL1A2 mRNA. This issue may require further analysis in future studies.
BMP‐7 is also reported to antagonise EMT in tissue repair, a hallmark of which is the reduction of E‐cadherin expression.39 E‐cadherin is the central component of cell‐cell adhesion junctions and is required for the formation of epithelium. On the other hand, the expression of αSMA is also a marker of mesenchymal myofibroblasts. TGFβ can induce EMT, which can be suppressed by BMP‐7 in renal tubular epithelial cells.8 In the liver, fetal hepatocytes are reported to undergo EMT,40,41 but there is no report describing EMT in adult hepatocytes. As shown in fig 2C2C,, expression of αSMA was actually suppressed by the overexpression of BMP‐7 in primary cultured rat HSC both in the presence and absence of TGFβ, suggesting that BMP‐7 might be involved in the reversion of myofibroblast‐like cells to HSC. To further examine whether adult hepatocytes have a potential for undergoing EMT, we stimulated primary cultured rat hepatocytes with BMP‐7. However, there was no significant impact on the expression of E‐cadherin by BMP‐7 stimulation (data not shown).
Adenovirus vector systems are useful to transfer and express genes in hepatocytes, even in liver fibrosis models.42 A single injection of adenoviral vectors in the tail vein attenuated fibrogenesis as shown in infigsfigs 5 and 66.. One disadvantage of this system is the inability to repeat administration due to acquisition of the immune response against the virus. This problem may be overcome by several new methodologies. First, new adenoviral vectors that express the transgene of interest together with either the dimeric human monoclonal antibodies against CD40 and CD8043 or CTLA4Ig,44 which mediate immunogenic responses to the virus, have been developed. Second, repeated retrograde administration of recombinant adenoviral vector into the common bile duct may be useful to successfully re‐express the transgene in the liver despite the existence of neutralising antibodies in the serum.45 Third, adenoviral vectors are being developed that do not express any viral protein.46,47 These experiments could eventually contribute to the development of gene therapy for hepatic fibrosis using either adenoviral vectors or other viral transduction systems.
The authors thank Ms M Yamauchi and Ms M Nakazawa for excellent technical assistance. This work was supported in part by a grant from the Ministry of Education, Science, Sports and Culture of Japan (C16590150 to KI) and a grant from the NIH (DK56621 to SLF) as well as the Feld Fibrosis Center.
αSMA - smooth muscle α‐actin
bHLH - basic helix‐loop‐helix
BMP‐7 - bone morphogenetic protein‐7
COL1A2 - type I α2 collagen
DMEM - Dulbecco's modified Eagle's medium
FCS - fetal calf serum
GFP - green fluorescent protein
HSC - hepatic stellate cells
Id2 - inhibitors of differentiation 2
MOI - multiplicity of infection
PFU - plaque forming units
TAA - thioacetamide
TGFβ - transforming growth factor‐β
Competing interests: None declared.