A virus-like particle-based connective tissue growth factor vaccine suppresses carbon tetrachloride-induced hepatic fibrosis in mice

Connective tissue growth factor (CTGF) has been recognized as a central mediator and promising therapeutic target in hepatic fibrosis. In this study, we generated a novel virus-like particle (VLP) CTGF vaccine by inserting the 138–159 amino acid (aa) fragment of CTGF into the central c/e1 epitope of C-terminus truncated hepatitis B virus core antigen (HBc, aa 1–149) using a prokaryotic expression system. Immunization of BALB/c mice with the VLP vaccine efficiently elicited the production of anti-CTGF neutralizing antibodies. Vaccination with this CTGF vaccine significantly protected BALB/c mice from carbon tetrachloride (CCl4)-induced hepatic fibrosis, as indicated by decreased hepatic hydroxyproline content and lower fibrotic score. CCl4 intoxication-induced hepatic stellate cell activation was inhibited by the vaccination, as indicated by decreased α-smooth muscle actin expression and Smad2 phosphorylation. Vaccination against CTGF also attenuated the over-expression of some profibrogenic factors, such as CTGF, transforming growth factor-β1, platelet-derived growth factor-B and tissue inhibitor of metalloproteinase-1 in the fibrotic mouse livers, decreased hepatocyte apoptosis and accelerated hepatocyte proliferation in the fibrotic mouse livers. Our results clearly indicate that vaccination against CTGF inhibits fibrogenesis, alleviates hepatocyte apoptosis and facilitate hepatic regeneration. We suggest that the vaccine should be developed into an effective therapeutic measure for hepatic fibrosis.


Results
The recombinant protein HBcΔCTGF 138-159 assembles into VLPs and elicits high titres of anti-CTGF neutralizing antibodies. Based on the in silico prediction and previous reports on the structure-function relationship of CTGF, a 22-amino acid peptide (CTGF 138-159 , 138 SMDVRLPSPDCPFPRRVKLPGK 159 ) within the VWC/CR domain of CTGF was selected as the antigen epitope. Using genetic engineering techniques, this polypeptide segment was inserted into the major immunodominant region (c/e1 epitope) of C-terminus (aa150-183)truncated HBc (HBcΔ ) to generate the recombinant protein HBcΔ CTGF  . HBcΔ protein was prepared as a control with the same method. The recombinant proteins were prokaryotically expressed and purified with Ni-NTA chromatography followed by size exclusion chromatography (Fig. 1a,b). Transmission electron microscopy confirmed that both HBcΔ and HBcΔ CTGF 138-159 assembled into VLPs within E. coli (Fig. 1c).
To test the neutralization ability of the anti-HBcΔ CTGF 138-159 serum, its effects on the CTGF-induced expressions of α -smooth muscle actin (α -SMA), the alpha 2 chains of collagen I (COL1A2) and tissue inhibitor of metalloproteinase-1 (TIMP-1) mRNAs in HSC-T6 rat hepatic stellate cells were examined. First, we determined the proper concentrations of rhCTGF on HSC-T6 cells. The results showed that rhCTGF at 10, 20, and 50 ng/mL did not significantly affect the proliferation of HSC-T6 (Fig. 2a) whereas rhCTGF at all these three concentrations significantly induced the upregulation of α -SMA, COL1A2 and TIMP-1 in HSC-T6. rhCTGF at 10 ng/mL showed weaker effects than it at 20 and 50 ng/mL, while no siginificant difference was discerned between the latter two concentrations on α -SMA, COL1A2 or TIMP-1 expression (Fig. 2b-d and data not shown). Based on the above results, we performed the neutralization test with the addition of rhCTGF at 20 ng/mL. As shown in Fig. 2b-d, the addition of the serum from HBcΔ CTGF 138-159 -immunized mice, but not that from either HBcΔ -immunized or normal control (NC) mice, at 1:20 and 1:100 dilution significantly abrogated rhCTGF-induced upregulation of α -SMA, COL1A2 and TIMP-1 whereas 1:500 dilution of the serum showed negligble effect ( * P > 0.05, # P < 0.05, § P < 0.005). Moreover, the anti-HBcΔ CTGF 138-159 serum at 1:20 dilution exhibited significantly stronger effect than that at 1:100 dilution. Subsequent experiment with the purified mouse IgGs showed similar results ( Supplementary Fig. S1). The purified IgG from HBcΔ CTGF 138-159 -immunized mice dose-dependently inhibited rhCTGF-induced upregulation of α -SMA, COL1A2 and TIMP-1 in HSC-T6 cells, while the IgGs from HBcΔ -immunized and NC mice had no effect. These results verified that HBcΔ CTGF 138-159 immunization elicited the production of CTGF-neutralizing antibodies.
We did not find any behavioural abnormalities or obvious pathological alterations in the vital organs of the immunized mice for five months after the first immunization (data not shown).
Immunization with HBcΔCTGF 138-159 protects mice from CCl 4 -induced hepatic fibrosis by suppressing HSC activation and decreasing the expression of profibrogenic factors. In the fibrosis experiment, forty BALB/c mice were equally divided into four groups: HBcΔ /CCl 4 , HBcΔ CTGF 138-159 /CCl 4 , CCl 4 and NC groups. The mice in the HBcΔ /CCl 4 and HBcΔ CTGF 138-159 /CCl 4 groups were immunized with the recombinant proteins as described above, and those in the NC and CCl 4 groups received an equal volume of PBS instead. ELISAs demonstrated that the antibody-producing profiles were similar to those in the above experiment. One week after the fifth immunization, the mice in the three CCl 4 groups received i.p. CCl 4 twice per week for six weeks (the protocol is illustrated in Fig. 3a). Sirius red staining and hydroxyproline determination revealed that six weeks of CCl 4 injection resulted in obvious collagen deposition and hepatic structural alterations (Fig. 3b) and significantly elevated hepatic hydroxyproline content (Fig. 3c) in the CCl 4 and HBcΔ /CCl 4 groups. However, the hydroxyproline content (Fig. 3c) and Ishak fibrosis score (Table 1) in the HBcΔ CTGF 138-159 /CCl 4 group were significantly lower than those in either the CCl 4 or HBcΔ /CCl 4 group, whereas those in the latter two groups were similar. These results clearly demonstrate that vaccination with HBcΔ CTGF 138-159 significantly protects mice from CCl 4 -induced hepatic fibrosis.
The expression of α -SMA, an indicator of myofibroblasts (MFBs), was detected via immunohistochemical staining and Western blotting to evaluate the effect of the vaccination on HSC activation during hepatic fibrosis. Immunohistochemistry showed that α -SMA expression in normal controls was confined to the smooth muscle cells in the walls of portal and central veins. After repeated CCl 4 injections, many α -SMA-positive cells were Negative-staining electron microscopy revealed that both recombinant HBcΔ and HBcΔ CTGF 138-159 assembled into VLPs (b). Scale bars = 100 nm. After five immunizations with HBcΔ CTGF 138-159 or HBcΔ VLPs, the production of anti-CTGF and anti-HBc antibodies in the BALB/c mice was determined via ELISAs with plates coated with HBc (c), OVA-CTGF 138-159 (d) and rhCTGF (e). Western blotting showed that the antiserum from the mice immunized with HBcΔ CTGF 138-159 recognized a 37-kD band in fibrotic liver tissue corresponding to rhCTGF (f). VLP, virus-like particle; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; ELISA, enzyme-linked immunosorbent assay; OVA, ovalbumin; rhCTGF, recombinant human connective growth factor. present in and around the fibrous septa as well as the perisinusoidal spaces of residual hepatic parenchyma. Compared with the CCl 4 and HBcΔ /CCl 4 groups, in which α -SMA-positive cells were strongly and diffusely immunostained in the liver tissue, the HBcΔ CTGF 138-159 /CCl 4 group showed fewer hepatic α -SMA-positive cells (Fig. 4a). Computer-assisted semi-quantitative analysis revealed that the average α -SMA-positive area in the HBcΔ CTGF 138-159 /CCl 4 group was significantly smaller than that in either the CCl 4 or HBcΔ /CCl 4 group (P < 0.001) (Fig. 4b). To support the above results, Western blotting confirmed significantly lower α -SMA expression in the HBcΔ CTGF 138-159 /CCl 4 group compared with the CCl 4 and HBcΔ /CCl 4 groups (Fig. 4c).
During α -SMA immunohistochemical staining, we observed that some α -SMA-positive cells around the fibrotic septa had some morphological features of hepatocytes, such as large size and big, round nuclei. To identify these α -SMA-positive cells, we performed double-labelling immunofluorescence using antibodies against albumin (a marker of hepatocytes) combined with α -SMA (a marker of MFBs) or FSP-1 (fibroblast-specific protein-1). There were a few double-labelled α -SMA and albumin as well as FSP-1 and albumin double-labelled cells distributed near the fibrous septa, suggesting that hepatocyte EMT may contribute to the population of MFBs in CCl 4 -induced liver fibrosis (Fig. 5).
Smad2 is a key mediator of TGF-β 1 signalling and a major regulator of CTGF expression 29 . Western blotting showed that the ratio of phosphorylated Smad2 (pSmad2) to total Smad2 was greatly increased in CCl 4 -induced fibrotic livers, whereas immunization with HBcΔ CTGF 138-159 significantly abolished this increase along with inhibiting hepatic fibrosis (Fig. 6e).

Vaccination alleviates hepatocyte apoptosis and promotes hepatocyte proliferation in fibrotic livers.
To determine the effect of vaccination against CTGF on hepatocyte injury and death, we assessed hepatocyte apoptosis via terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labelling assay (TUNEL). Few TUNEL-positive cells were present in the NC livers. Six weeks of CCl 4 exposure led to greatly increased numbers of TUNEL-positive hepatocytes in the three CCl 4 groups. HBcΔ CTGF 138-159 /CCl 4 group showed a significantly lower apoptosis index (AI) compared with either HBcΔ /CCl 4 or CCl 4 group (P < 0.05), whereas there was no significant difference in AIs between the latter two groups (P > 0.05) (Fig. 7a,b).
The proliferation of hepatocytes was evaluated using proliferating cell nuclear antigen (PCNA) immunostaining. The increased PCNA LIs in all the three CCl 4 groups were significantly higher than those in the NC group  (P < 0.001), whereas the HBcΔ CTGF 138-159 /CCl 4 group exhibited a significantly higher PCNA labelling index (PCNA LI) than the CCl 4 and HBcΔ /CCl 4 group (P < 0.001) (Fig. 7c,d).
To determine whether the increased proliferation of hepatocytes in the HBcΔ CTGF 138-159 /CCl 4 group is attributable to the direct neutralization of CTGF, the effects of rhCTGF and the anti-sera on the proliferation of human embryo liver cells L02 were evaluated. rhCTGF at a dose of 50 ng/mL significantly inhibited the proliferation of L02 cells (P < 0.05), whereas the addition of 1:100 diluted anti-HBcΔ CTGF 138-159 serum significantly abolished the proliferation-inhibitory effect of 50 ng/mL rhCTGF on L02 cells (P < 0.05). Because of the nutritional effect of the mouse serum, the addition of the other two sera at the same dilution resulted in slightly higher proliferation rates compared with that in the rhCTGF + PBS group (Fig. 7e).

Discussion
The present study demonstrated that a HBc VLP-based CTGF vaccine efficiently elicited the production of anti-CTGF neutralizing antibodies in mice. Immunization with this vaccine significantly attenuated hepatic fibrogenesis, alleviated hepatocyte apoptosis and improved hepatocyte proliferation in CCl 4 -intoxicated mouse livers.
Unlike other growth factors and cytokines, CTGF does not have a unique receptor to which it binds with high affinity to induce signal transduction. Alternatively, CTGF functions via interactions with other growth factors or cytokines to modify their signalling through multiple mechanisms depending on the cell type and context. A full-length CTGF is composed of four domains: an insulin-like growth factor-binding protein domain (IGFBP, aa 27-97), a von Willebrand factor type C or cysteine-rich domain (VWC/CR, aa 101-167), a thrombospondin type 1 homology domain (TSP-1, aa 199-243), and a carboxy-terminal cysteine knot domain (CT, aa 256-330) 30 . Each of these domains may exert individual biological functions. Interestingly, CTGF is present not only in its full-length form but also as fragments in the serum, urine and extracellular space 31 . The N-terminal half (IGFBP and VWC/CR domains) of CTGF has been proposed to modulate MFB differentiation and collagen synthesis 32 , whereas the C-terminal domains contribute to fibroblast proliferation 33,34 . The VWC/CR domain of CTGF can bind to TGF-β 1 to exert a synergistic effect for TGF-β 1-mediated fibrogenesis and can bind to BMPs to inhibit their antifibrotic effect 35 . Therefore, the VWC/CR domain is considered to be the primary profibrogenic domain within CTGF 32,35 . In the present study, we selected a twelve-amino acid (138-159) fragment derived from the VWC/CR domain of CTGF as the antigen epitope. This vaccine significantly attenuated CCl 4 -induced hepatic fibrosis in mice, and the anti-serum elicited by this vaccine largely abolished CTGF-mediated extracellular matrix production by HSC-T6 cells. This suggests that the VWC/CR module is essential for the profibrogenic effect of CTGF. The profibrogenic function of CTGF is primarily dependent on its binding with TGF-β 1 through the VWC/CR domain, which further enhances the binding of TGF-β 1 to its receptors and amplifies the signalling 35 . The vaccination targeting the VWC/CR domain of CTGF decreased the levels of p-Samd2, a representative downstream signalling mediator of TGF-β 1 36 , indicating that the vaccination abated the TGF-β 1 signalling. We also prepared another VLP-based vaccine by inserting a twenty-amino acid fragment from the hinge region between the TSP-1 and CT domains into HBcΔ (HBcΔ CTGF 240-259 ). The antibodies against this epitope inhibited bone nodule formation and mineralization in rats 37 and suppressed neovascularization induced by HT1080 cells or MDA231 cells on chicken chorioallantoic membrane 38 . However, although HBcΔ CTGF 240-259 efficiently elicited the production of specific antibodies, vaccination with HBcΔ CTGF 240-259 had no effect on CCl 4 -induced fibrosis in mice (data not shown). This result supports the assertion that the profibrogenic effect of CTGF is exerted predominantly by the N-terminus modules.
TGF-β 1 is known to be a key profibrogenic cytokine and a major modulator of CTGF expression. TGF-β induces the expression of CTGF in hepatocytes through a Smad2 signalling pathway 29,39 and in HSCs via a STAT3-dependent pathway 40,41 . PDGF-B promotes CTGF expression in HSCs through TGF-β 1 production 40 . CTGF also up-regulates its own expression through a positive feedback loop 12 . Whether CTGF modulates the hepatic expression of TGF-β 1 and PDGF-B remains to be elucidated. In this study, we observed that the hepatic expressions of TGF-β 1, PDGF-B and CTGF as well as Smad2 phosphorylation were decreased by the CTGF vaccination, indicating that CTGF is not only a downstream effector but also an important positive modulator of hepatic TGF-β 1 and PDGF-B expression and, in fibrotic liver tissues, there are complex mutual promotion and positive feedback effects between these factors. The neutralization of CTGF can lead to the diminishment of the profibrogenic cytokine cascade and results in hepatic fibrosis attenuation.
TIMP-1, which is primarily produced by activated HSCs and Kupffer cells, can inhibit the degradation of accumulated ECM by blocking MMPs and suppressing HSC apoptosis 42,43 . Thus, TIMP-1 plays a critical role in liver matrix remodelling. In this study, TIMP-1 was significantly up-regulated in fibrotic livers. Vaccination with HBcΔ CTGF 138-159 abolished this increase. The addition of CTGF up-regulated the expression of TIMP-1 in HSC-T6 cells, and the neutralization of CTGF by the anti-CTGF serum abated this up-regulation, suggesting that TIMP-1 is one of the downstream mediators of CTGF in profibrogenesis.  controversial 44,45 . In a previous study, we found that some hepatocyte-like liver cells were desmin-positive in fibrotic mouse livers. This result suggests that the EMT of hepatocytes may represent a source of ECM-producing MFBs 46 . In the present study, double-labelling immunofluorescence showed that some albumin-positive cells in the mouse fibrotic liver sections were positively stained with α -SMA or FSP-1, suggesting that some hepatocytes undergo EMT and may contribute to MFBs in CCl 4 -induced hepatic fibrosis.
In addition to an anti-fibrotic effect, the CTGF vaccination resulted in significantly decreased apoptosis and increased regeneration of hepatocytes. The restoration of the hepatic structure is considered to provide favourable conditions for both the survival and regeneration of hepatocytes by improving the micro-circulation and composition of the extracellular matrix. For example, TIMP-1 is reported to exhibit an inhibitory effect on hepatocyte proliferation 47,48 . Moreover, TGF-β 1 has been reported to impair hepatocyte proliferation and induce hepatocyte apoptosis [49][50][51] . Our in vitro experiment demonstrated that rhCTGF significantly inhibited the proliferation of L02 human hepatocyte cells and that the neutralization of rhCTGF with the anti-CTGF serum reversed this inhibitory effect. This indicates that the neutralization of CTGF and the subsequent down-regulation of other CTGF-modulated growth factors and cytokines also contribute to decreased apoptosis and increased regeneration of hepatocytes in the vaccinated mice.
The effectiveness and safety of passive immunization (utilizing Abs) against deleterious pathological factors has been verified in clinical practice. The success of TNFα mAb agents in the clinical treatment of rheumatoid arthritis and ankylosing spondylitis are excellent examples of this. FG-3019 13 , an anti-CTGF mAb developed by FibroGen, Inc., is undergoing phase II clinical trials to evaluate its safety and efficacy in liver fibrosis due to chronic Hepatitis B infection and idiopathic pulmonary fibrosis. However, vaccines have notable advantages compared with therapeutic Abs. Because the half-life of IgG Abs is approximately 2-3 weeks, frequent injections at weekly rates are necessary to maintain effective threshold levels of Abs. Abs have immunogenicity that is inherent to the heterologous epitopes of the Ig molecule; thus, repeated use of Abs can induce an anti-antibody response. This may represent a major limitation for the prolonged use of passive immunization in chronic disease. Moreover, the high costs due to the difficulties in manufacturing and purification limit the wide use of therapeutic antibodies. By contrast, active immunization with vaccines may largely overcome these defects of therapeutic Abs.
In conclusion, our study verified that vaccination against CTGF with HBcΔ CTGF 138-159 markedly inhibited fibrogenesis, alleviated hepatocyte apoptosis and promoted hepatocyte proliferation in CCl 4 -intoxicated mouse livers. This approach could potentially be developed into an efficient, safe, and convenient therapeutic strategy for managing chronic fibrotic liver diseases. Furthermore, because CTGF plays a universal role in tissue fibrosis, this vaccine may also be used to inhibit fibrosis in other organs and tissues.

Methods
Selection of antigenic peptides. Antigenic peptide prediction was performed based on the hydrophilicity/hydrophobicity analyses of amino acids (http://tools.immuneepitope.org) and the epitopes of amino acid residues (http://www.cbs.dtu.dk/services/BepiPred/ and http://www.imtech.res.in/raghava/bcepred). Using information from the high-resolution solution structure of CTGF, priority was given to the peptide sequences located in the receptor binding sites and/or at the terminal regions.
The plasmids pET28-HBcΔ and pET28-HBcΔ CTGF 138-159 were transformed into E. coli BL21(DE3), respectively. Expressions of HBcΔ and HBcΔ CTGF 138-159 fusion proteins were induced by the addition of isopropyl-1-thio-β -D-galactopyranoside. The recombinant proteins were preliminarily purified using Ni-NTA affinity chromatography under the native conditions. The VLPs were further purified with sepharose CL-4B size exclusion chromatography (AKTAPrime Plus, GE, Fairfield, CT). Expression and purification were monitored via SDS-PAGE followed by Coomassie brilliant blue staining. To verify the formation of VLPs, the purified proteins were negatively stained with 3% phosphotungstic acid and examined using transmission electron microscopy (H-600, Hitachi, Japan). For immunization, the purified proteins were dialysed against PBS and concentrated to 0.5 mg/mL. Animal and experimental protocols. Specific pathogen-free, 6-week-old male BALB/c mice were purchased from the Experimental Animal Center, School of Medicine, Xi'an Jiaotong University (Permit Number: 2011-54). All animals received humane care and were housed for one week prior to the experiments. They were allowed free access to water and a laboratory chow diet. All animal procedures were carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The maintenance of mice and all experiment protocols were approved by the Institutional Animal Ethics Committee of Xi'an Jiaotong University.
To verify the antigenicities of the recombinant proteins, eighteen mice were assigned to three equal groups (HBcΔ CTGF 138-159 , HBcΔ and NC) and received an intraperitoneal (i.p.) immunization of HBcΔ CTGF 138-159 (50 μ g in 0.1 mL), HBcΔ (50 μ g in 0.1 mL) or an equal volume of PBS biweekly five times. The production of specific antibodies was monitored using ELISAs, and the plates were coated with 100 ng/well of either OVA-conjugated CTGF 138-159 polypeptide (the polypeptide was synthesized using an automated peptide synthesizer and cross-linked to OVA by 1-ethyl-3-(3-dimethylaminopropy) carbodiimide (EDC)) or recombinant HBc (purchased from ProSpec-Tany TechnoGene Ltd, Ness Ziona, Israel). The production of anti-CTGF antibodies was further confirmed via ELISAs in 20 ng/well rhCTGF (ProSpec-Tany TechnoGene Ltd, Ness Ziona, Israel) and Western blotting. One week after the fifth immunization, three mice in each group were euthanized. The blood was collected to isolate the serum for the subsequent experiments. Five months after the first immunization, the remaining mice were sacrificed. The liver, lungs, heart, and kidneys were harvested, fixed with 10% formalin, sectioned, and stained with haematoxylin and eosin to evaluate any adverse effects.
In the fibrosis experiment, forty mice were assigned to four equal groups: NC, CCl 4 , HBcΔ /CCl 4 , and HBcΔ CTGF 138-159 /CCl 4 . The mice were immunized as described above with the exception that the first two groups were given an equal volume of PBS instead of the recombinant proteins. One week after the fifth immunization, the mice in the three CCl 4 groups received i.p. CCl 4 (1 mL/kg dissolved in olive oil to reach a final concentration of 20%) twice per week for six weeks. The mice in the NC group were given an equal volume of olive oil. Three days after the final CCl 4 injection, the mice were sacrificed. The blood samples and the left lobe of the liver were collected.

Cell culture and neutralization tests. Rat hepatic stellate cells HSC-T6 and human embryo liver cells L02
(both from the Type Culture Collection of the Chinese Academy of Sciences, Shanghai, China) were maintained in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% foetal bovine serum (FBS) at 37 °C under 5% CO 2 . All the experiment was performed in triplicate wells and repeated three times.
For the cell proliferation assay, HSC-T6 and L02 cells were inoculated into 96-well plates at 300 cells per well and allowed to adhere for 24 hours. Then, the complete medium was replaced with DMEM supplemented with 1% FBS, rhCTGF (10, 20 and 50 ng/mL), and the serum (1:100 dilution as indicated) from the mice immunized with HBcΔ CTGF 138-159 , HBcΔ or NC, respectively. Cell proliferation was measured via CCK-8 (Dojindo, Kyushu, Japan) assay following the manufacturer's instructions.
In the neutralization test, HSC-T6 cells were seeded into 6-well plates at a density of 5 × 10 5 cells/well and maintained in complete medium to approximately 50% confluence. Then, the cells were cultured in DMEM supplemented with 10% FBS and rhCTGF (20 ng/mL) and the anti-serum (at 1:500, 1:100 or 1:20 dilution) for 48 hours. Finally, the cells were harvested and the total RNA was extracted for further experiments.
The IgGs were purified by caprylic acid─ ammonium sulfate precipitation followed by desalting through sephadex G-25 chromatography. About 10 mg IgG was obtained from 2 mL pooled serum for each group. The neutralization effect of the purified mouse IgG on rhCTGF was tested as described above except that the anti-sera were replaced with the purified IgG at concentrations of 10, 50 and 250 μ g/mL.

Western blotting.
To determine the specificity of the antibodies elicited by HBcΔ CTGF 138-159 , 100 ng of rhCTGF and 100 μ g of fibrotic liver tissue lysate were loaded on a SDS-PAGE and transferred onto a nitrocellulose membrane. After being blocked for one hour in 10% skim bovine milk, the membrane was incubated in the anti-CTGF mouse serum (1:500 dilution) at 4 °C overnight. Horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG (Thermo, Pierce, Rockford, IL) served as the second antibody. Finally, the blots were developed with enhanced chemiluminescence.
To detect the expression of α -SMA, phospho-Smad2/ Smad2, TIMP-1, TGF-β 1 and PDGF in the mouse liver tissues, the tissues were lysed in radio immunoprecipitation assay buffer supplemented with protease inhibitors. The extracts were resolved on SDS-PAGE and transferred onto NC membranes. Mouse anti-α -SMA mAb (Thermo, Lab Vision, Fremont, CA), rabbit anti-phospho-Smad2 (Ser465/467) mAb, rabbit anti-Smad2 mAb, mouse anti-TIMP-1 mAb, mouse anti-TGF-β 1 mAb and mouse anti-PDGF-B mAb were used as the primary antibodies, and HRP-conjugated goat anti-mouse IgG or HRP-conjugated goat anti rabbit IgG were used as secondary antibodies. The expression of β -actin (mouse anti-β -actin mAb) served as the internal control. All of the antibodies except those specifically indicated were from Cell Signaling Technology (Danvers, MA). Bands were quantified with Image-J software (National Institutes of Health, Bethesda, MD). The relative protein abundance in each sample was normalized to that of β -actin.

Real-time quantitative reverse transcriptase polymerase chain reaction (RQ-PCR). Total
RNA was extracted from HSC-T6 cells and the liver tissues with TRIzol reagent (Thermo, Life Technologies, Carlsbad, CA) and quantified via NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Carlsbad, CA). Reverse transcription was performed with a PrimeScript TM RT reagent Kit (TaKaRa, Dalian, China). The relative abundance of each mRNA in the sample was determined using RQ-PCR with the corresponding primers (Supplementary Table S1 Bio-Rad iQ5 2.0 Standard Edition Optical System Software (Bio-Rad, Hercules, CA). Data were analysed using the Δ Δ CT method, and β -actin served as an internal control. The results are presented as the mean ± SEM of triplicate reactions from three separate experiments.
Hepatic hydroxyproline content. The hepatic hydroxyproline contents were determined using an alkaline lysis method with a commercially available kit (Nanjing Jiancheng Biotech, LTD). Hepatic hydroxyproline content was expressed as μ g/mg wet liver tissue.
Histology, immunohistochemistry and immunofluorescence. The formalin-fixed paraffin-embedded liver tissues were 5-μ m sectioned, deparaffinized, rehydrated, and subsequently stained with haematoxylin & eosin and Sirius red to assess the hepatic architectural alterations and collagen deposition. The degree of fibrosis was evaluated semi-quantitatively according to the Ishak system 52 .
Immunohistochemistry was performed using the Histostain TM -Plus SP kit (Thermo Fisher Scientific, Carlsbad, CA). The rehydrated sections were submerged in 0.3% (v/v) H 2 O 2 /methanol for 20 min to quench the endogenous peroxidase activity. Next, the sections were blocked with 10% (v/v) normal goat serum in PBS for 1 h and incubated with the primary antibodies (mouse anti-α -SMA mAb, Thermo, Lab Vision, Fremont, CA, 1:800 diluted; rabbit anti-FSP-1 polyclonal antibody, Abcam, Cambridge, MA, 1:100 diluted or mouse anti-PCNA mAb, Thermo, Lab Vision, Fremont, CA, 1:500 diluted) at 4 °C overnight. After three washes with PBS-T, the sections were incubated with Streptavidin-HRP for 20 min at room temperature, washed, and processed using diaminobenzidine (DAB) at room temperature for 5 min. Finally, the sections were counterstained with haematoxylin and mounted.
Immunohistochemical images of α -SMA in the tissue sections were obtained using a light microscope (Olympus BX51, Olympus, Tokyo, Japan) equipped with a DP70 digital camera. Ten random fields (100× ) were quantified using Image-Pro Plus 5.0, a commercially available software package from Media Cybernetics (Silver Spring, MD). The expression of α -SMA was expressed as the mean percentage of the positively stained areas out of the field areas. PCNA LI was evaluated by counting 1,000 hepatocytes from multiple random areas and was expressed as the percentage of positive cells.
Tunel. Apoptotic hepatocytes were labelled in situ using a TUNEL peroxidase detection kit (DeadEnd TM Colorimetric TUNEL System, Promega, Madison, WI) in accordance with the manufacturer's protocol. The nuclei were counterstained with haematoxylin. The AI is expressed as the percentage of TUNEL-positive hepatocytes using the same method for evaluating PCNA LI.

Statistical analysis.
All quantitative data were expressed as the mean ± standard error (SEM). To assess the statistical significance of the inter-group differences of the quantitative data, Bonferroni's multiple comparison tests were performed after one-way analysis of variance (ANOVA), followed by Bartlett's tests to determine the homology of variance. The Mann-Whitney U-test and the Pearson's correlation coefficient (R) calculation were used to compare qualitative variables. The differences were considered significant at a p value of < 0.05.