Susceptibility to liver fibrosis in mice expressing a connective tissue growth factor transgene in hepatocytes

Authors

  • ZhenYue Tong,

    1. Center for Cell and Developmental Biology, The Research Institute at Nationwide Children's Hospital, Columbus, OH
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Ruju Chen,

    1. Center for Cell and Developmental Biology, The Research Institute at Nationwide Children's Hospital, Columbus, OH
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Daniel S. Alt,

    1. Center for Cell and Developmental Biology, The Research Institute at Nationwide Children's Hospital, Columbus, OH
    Search for more papers by this author
  • Sherri Kemper,

    1. Center for Cell and Developmental Biology, The Research Institute at Nationwide Children's Hospital, Columbus, OH
    Search for more papers by this author
  • Bernard Perbal,

    1. Laboratory of Molecular Oncology and Virology, UFR de Biochimie (Department of Biochemistry), University of Paris, Paris, France
    Search for more papers by this author
  • David R. Brigstock

    Corresponding author
    1. Center for Cell and Developmental Biology, The Research Institute at Nationwide Children's Hospital, Columbus, OH
    2. Department of Surgery, The Ohio State University, Columbus, OH
    3. Department of Molecular and Cellular Biochemistry, The Ohio State University, Columbus, OH
    • Center for Cell and Developmental Biology, Room WA2022, The Research Institute, Nationwide Children's Hospital, 700 Children's Drive, Columbus OH 43205
    Search for more papers by this author
    • fax: 614-722-5892.


  • This work is dedicated to the memory of Dr. Donald R. Cooney (1943–2008), who was a champion of surgical research and established the Pediatric Surgery Research Laboratory at Nationwide Children's Hospital in 1991.

  • Potential conflict of interest: Nothing to report.

Abstract

Connective tissue growth factor (CCN2) is a matricellular protein that is up-regulated in many fibrotic disorders and coexpressed with transforming growth factor β. CCN2 promotes fibrogenesis and survival in activated hepatic stellate cells, and injured or fibrotic liver contains up-regulated levels of CCN2 that are produced by a variety of different cell types, including hepatocytes. To investigate CCN2 action in vivo, transgenic FVB mice were created in which the human CCN2 gene was placed under the control of the albumin enhancer promoter to elevate hepatocyte CCN2 levels. Production of human CCN2 (hCCN2) messenger RNA and elevated CCN2 protein levels was demonstrated in transgenic livers, whereas levels of endogenous mouse CCN2 were comparable between transgenic and wild-type mice. Liver histology and liver function tests were unaffected in transgenic animals. However, after chronic administration of CCl4, α-smooth muscle actin (α-SMA)–expressing cells and collagen deposition were increased as a function of the dosage of the hCCN2 transgene (hccn2+/+ > hccn2+/− > hccn2−/−). Moreover, CCl4-induced serum hyaluronic acid, hepatic tissue levels of α-SMA or acid-soluble collagen, and messenger RNA expression of α-SMA, collagen α1 (I), matrix metalloprotease-2, or tissue inhibitor of metalloprotease-1 were greater in transgenic mice than in wild-type mice. Transgenic mice also exhibited enhanced hepatic deposition of collagen 2 weeks after bile duct ligation. Conclusion: Production of elevated CCN2 levels in hepatocytes of transgenic mice in vivo does not cause hepatic injury or fibrosis per se but renders the livers more susceptible to the injurious actions of other fibrotic stimuli. These studies support a central role of CCN2 in hepatic fibrosis and demonstrate a role of the microenvironment in regulating the profibrotic action of CCN2. (HEPATOLOGY 2009.)

Connective tissue growth factor (CCN2) is a 349-residue mosaic protein that participates in critical processes such as differentiation, development, tumor growth, angiogenesis, placentation, and wound healing.1 The actions of CCN2 are complex and reflect its function as a matricellular protein whereby it resides as a nonstructural matrix- or cell-associated molecule that regulates cellular responsiveness to a myriad of extracellular signals and environmental cues.2 These actions are likely achieved, at least partly, through its engagement of specific integrins and heparan sulfate proteoglycans on the cell surface.2

Fibrosis is the most common pathophysiological condition in which CCN2 has been implicated and fibrotic conditions in the liver, lung, kidney, heart, and skin are all strongly associated with CCN2 overexpression.1 In the liver, CCN2 messenger RNA (mRNA) expression is elevated in patients suffering from chronic viral hepatitis, primary biliary cirrhosis, primary sclerosing cholangitis, cryptogenic, alcoholic liver disease, congenital hepatic fibrosis, or nonalcoholic steatohepatitis.3 CCN2 expression is also elevated in fibrotic livers from rodents subjected to bile duct ligation or exposure to CCl4 or N-nitrosodimethylamine.4–6 High hepatic levels of CCN2 are associated with its entry into the circulation, as demonstrated by higher serum CCN2 concentrations in patients with biliary atresia or chronic viral hepatitis compared with control patients, with increasing CCN2 concentrations correlating with the progression of liver fibrosis.7, 8

Hepatocyte-derived CCN2 likely contributes to fibrotic pathways in the liver, at least in part, through its paracrine stimulation of activated hepatic stellate cells (HSCs) in which fibrogenic and survival pathways are promoted by CCN2.3, 9 Hepatocyte expression of CCN2 is usually low but is enhanced in response to liver injury or insult such as partial hepatectomy, drug-induced toxicity, ethanol metabolism, infection with hepatitis C virus, or cancer.10–13 This effect may be due in part to the inductive action of TGF-β, which stimulates CCN2 expression in hepatocytes via activation of PiI3K/p38MAPK and/or ALK5-dependent Smad 3.14, 15 In these studies, we have investigated the profibrotic aspects of elevated hepatocyte CCN2 levels in transgenic mice in which the human CCN2 gene was expressed under the control of the albumin enhancer promoter.

Abbreviations

α-SMA, α-smooth muscle actin; BDL, bile duct ligation; CCN2, connective tissue growth factor; cDNA, complementary DNA; FACS, fluorescence-activated cell sorting; HSC, hepatic stellate cell; hCCN2, human CCN2; mCCN2, mouse CCN2; mRNA, messenger RNA; PCR, polymerase chain reaction; qRT-PCR, quantitative real-time polymerase chain reaction; RT-PCR, reverse-transcriptase polymerase chain reaction; TGF-β, transforming growth factor β.

Materials and Methods

Cloning of Human CCN2 Expression Vector.

A 6,600-bp SacI fragment of human CCN2 (hCCN2) genomic DNA16 was amplified by polymerase chain reaction (PCR) using forward primer 5′-CGTGCCAACCATGACCGA-3′ and reverse primer 5′-CGGAATTCTCTAATGAGTTAATGTCTCTCACTCTCTG-3′. The resulting product was digested with SacII and EcoRI and ligated into the SacII/EcoR1 sites of pTRE (Clontech, Mountain View, CA) to produce pTRE/hCCN2. To achieve expression by hepatocytes, the cytomegalovirus promoter in pTRE/hCCN2 was replaced with the mouse albumin enhancer/promoter sequence in pBluescript (KS-) plasmid p2335A-1.17 Briefly, a 2.3-kb SacI/Kpn fragment from p2335A-1 containing the mouse albumin enhancer/promoter was blunted by T4 DNA polymerase and ligated into the blunted Xhol/SacII sites of pTRE/hCCN2 to generate “pALB/hCCN2” (Fig. 1A) which was confirmed by nucleotide sequencing.

Figure 1.

Cloning and testing of human CCN2 transgene construct and production of native CCN2 by hepatocytes in vivo. (A) A 1.9-kb fragment containing the human CCN2 gene was placed downstream of the albumin enhancer to produce pALB/hCCN2 as described in Materials and Methods. Primers P1 and P2 were used for genotyping. (B) Wild-type FVB/n mice were administered (a) 6 g/kg ethanol or (b) saline by way of intraperitoneal injection at 0, 8, and 16 hours. Mice were sacrificed 8 hours later and either (a,b) fixed tissue was processed for in situ hybridization to detect CCN2 mRNA or (c) hepatocytes were isolated from fresh tissue and tested for the presence of CCN2 protein by FACS analysis using a CCN2 polyclonal antibody. (C) FL83B hepatocytes were either mock-transfected (lane 1) or transiently transfected with pALB/hCCN2 (lane 1) or mock-transfected (lane 2) for 48 hours, after which RNA or protein were subsequently extracted from parallel cultures of cells. The results of reverse-transcriptase PCR using primers designed to detect (a) mCCN2 mRNA, (b) hCCN2 mRNA, and (c) mouse β-actin are shown. Immunoprecipitation using anti-CCN2 showed that (d) higher levels of CCN2 protein were present in lysates from hCCN2-transfected cells as compared with the mock-transfected counterparts, whereas (e) levels of β-actin determined by way of western blotting were comparable.

pALB/CCN2-Mediated CCN2 Overexpression in Hepatocytes In Vitro.

FL83B mouse hepatocytes (American Type Culture Collection, Manassas, VA) were grown in 6-well plates in Hams F-12 medium (Mediatech Inc., Manassas, VA) containing 10% bovine serum. At ≈80% confluency, the medium was replaced with 1 mL fresh Hams F-12 medium containing 2% bovine serum and 15 μL FuGene 6 transfection reagent (Roche Diagnostics Corp., Indianapolis, IN). Cells were then transfected with 2 μg pALB/hCCN2 for 48 hours. RNA was isolated using RNA STAT-60 Total RNA Isolation Reagent (Tel Test Inc., Friendswood, TX). Reverse-transcriptase PCR (RT-PCR) was performed on 2 μg total RNA using SuperScript II Reverse Transcriptase (Invitrogen) to generate the complementary DNA (cDNA). The conditions for the reverse-transcriptase reaction were: 25°C for 10 minutes, 42°C for 50 minutes, and 70°C for 15 minutes followed by PCR using the following primers: hCCN2 (632 bp), 5′-TCCCTGCATCTTCGGTGGTA-3′ and 5′-CCTCGCCGTCAGGGCAC-3′; mouse CCN2 (mCCN2) (734 bp), 5′-CCGCACTGCCCCGCC-3′ and 5′-CCCGCAGAACTTAGCCC-3′; and β-actin (324 bp), 5′-AGCTTGCTGTATTCCCCTCCATCGTG-3′ and 5′-AATTCGGATGGCTACGTACATGGCTG-3′. PCR conditions were 2 minutes at 94°C, 35 cycles at 94°C for 30 seconds, annealing at 56°C for 40 seconds, and 72°C for 1.5 minutes, followed by final extension for 10 minutes at 72°C. For the last 6 hours of the 48-hour transfection period, the cells were metabolically labeled with 100 μCi/mL [35S]cysteine/methionine (Trans35S-label, MP Biomedical, Irvine, CA) prior to lysis in immunoprecipitation buffer (50 mM TrisHCl [pH 7.4] containing 150 mM NaCl, 1% Nonidet P40, 0.25% DOC, 1 mM ethylene diamine tetraacetic acid, and 0.1% sodium dodecyl sulfate) and subsequent pull-down of 100 μg total cellular protein with anti-CCN218 or western blot of 50 μg protein using monoclonal anti–β-actin (clone AC-15; Sigma-Aldrich, St. Louis, MO).

Generation of pALB/CCN2 Transgenic Mice.

Animal procedures were approved by the Institutional Animal Care and Use Committee of The Research Institute at Nationwide Children's Hospital (Columbus, OH).

A 7.1-kb linearized fragment of Not1-digested pALB/hCCN2 (Fig. 1A) was microinjected into single-cell FVB/n mouse eggs that were then implanted into pseudo-pregnant female FVB/n mice (The Jackson Laboratory, Bar Harbor, ME), yielding 30 offspring (3 litters) of which two (one male, one female) were positive for ALB/hCCN2 DNA by PCR screening (data not shown). Using these as founders, two essentially identical transgenic lines, ZY1 and ZY2, were established. Data reported here were obtained from the ZY2 line, which has been maintained continuously for 60 months. All of the experiments described were performed on male mice between 4 and 28 weeks of age. The three genotypes (hccn2+/+, hccn2+/−, hccn2−/−) required for this study were initially determined by screening using semiquantitative PCR and Southern blotting, then confirmed by quantitative real-time PCR (qRT-PCR).

CCl4 Administration.

Wild-type or transgenic male mice that were 4-6 weeks of age were injected intramuscularly once per day on 4 successive days each week for 5 weeks with either 30 μL of vegetable oil or a mixture of 0.5 μL CCl4 (Sigma-Aldrich) in 29.5 μL of vegetable oil. This dose of CCl4 induced submaximal hepatic fibrosis in wild-type FVB mice as established from dose–response studies (data not shown). Upon sacrifice, blood was collected and individual liver lobes were tied and harvested either immediately and snap-frozen in liquid nitrogen for subsequent RNA or protein extraction, or after in situ perfusion using phosphate-buffered saline followed by 4% paraformaldehyde (Sigma-Aldrich) for histological analysis of fixed tissue.

Ethanol Administration.

Wild-type FVB mice received ethanol (6 g/kg) or saline intraperitoneally at 0, 8, and 16 hours prior to sacrifice at 24 hours. A portion of the liver was fixed and processed for in situ hybridization to detect mCCN2 mRNA, while hepatocytes were isolated from the remainder of the fresh unfixed liver tissue by in situ collagenase digestion19 and subjected to fluorescence-activated cell sorting (FACS) using anti-CCN2 antibody essentially as described.20

Bile Duct Ligation.

Bile duct ligation (BDL) or sham operations were performed on 4- to 6-week-old hccn2−/− or hccn2+/+ mice as described,21 after which the animals were maintained under normal housing conditions for 14 days. During this period, survival was >90% in all groups of mice. On day 14 after the operation, mice were sacrificed, and their livers were processed for histological evaluation.

Semiquantitative PCR.

Three-week-old mice were genotyped by way of PCR of DNA obtained from tail clippings. DNA was isolated by sequential protease K digestion and phenol-chloroform extraction. PCR was performed using transgene-specific primers, one of which spanned the junction between the albumin enhancer/promoter and hCCN2 (P1: 5′- AGAGCGAGTCTTTCTGCACACA-3′; P2: 5′-GAGAGAATCACGACCCTGACTT-3′) (Fig. 1A). PCR conditions were 2 minutes at 94°C, 35 cycles at 94°C for 30 seconds, annealing at 55°C for 40 seconds, and extension at 72°C for 1.5 minutes, followed by 10 minutes at 72°C. The 1,044-bp hCCN2 PCR product was detected by way of ethidium bromide staining after agarose gel electrophoresis.

Southern Blotting.

Genomic DNA (5 μg) from mice tails was digested in a 40 μL final volume with 20 U StuI in Reaction buffer II (New England Biolabs, Ipswich, MA). DNA fragments were separated in a 0.7% agarose gel and transferred to Hybond N+ membrane (Amersham-Pharmacia, Piscataway, NJ). The membrane was probed with a previously described 1.1 kb human CCN2 cDNA18 that was labeled using a Radprime labeling kit (Invitrogen).

Quantitative PCR.

Quantitative PCR was performed with TaqMan-minor groove binder, primers, probes, and TaqMan Universal Master Mix using an ABI Prism 7500 (all from Applied Biosystems Inc., Foster City, CA). Primers were designed using Primer Express Software (version 3.0; Applied Biosystems). Genotypes were established by determination of the amount of hCCN2 product relative to that of mouse glucagon after 40 cycles. Specific conditions were 50°C for 2 minutes, 95°C for 10 minutes, 40 cycles of 95°C for 15 seconds, and then 58°C for 1 minutes. Each amplification reaction was performed in triplicate on 20 ng of genomic DNA. The number of transgene copies was established from hCCN2 plasmid DNA standard curves. Primer sequences were: hCCN2, CATCCCCCACCCCTCTCT (forward), CCGTACCACCGAAGATGCA (reverse), TCTCCAGCCAAAGAT (probe); glucagon, CACAACATCTCGTGCCAGTCA (forward), ATCTGCATGCAAAGCAATATAGCT (reverse), TGGGATGTACAATTTCAA (probe).

Histology.

Perfused livers were fixed with 4% paraformaldehyde for 24 hours and then embedded in paraffin. Sections of 5 μm thickness were cut and stained with hematoxylin-eosin. For immunohistochemical detection of α-smooth muscle actin (α-SMA), slides were incubated with monoclonal mouse anti–α-SMA immunoglobulin G (Dako, Glostrup, Denmark) followed by development with UltraTek reagents and AEC chromogenic substrates (all from ScyTek Laboratories, Logan, UT) and hemotoxylin counterstain. CCN2 immunohistochemistry was performed using affinity-purified rabbit anti-CCN2 immunoglobulin G as described.22 CCN2 in situ hybridization was performed with digoxygenin-labeled RNA sense or antisense probes generated from a pCRII vector containing mCCN2 cDNA.23 Collagen was detected by staining sections with 0.1% Sirius Red (Sigma-Aldrich). Protein staining for α-SMA or collagen was quantified over 10 ×20 fields per mouse using ImagePro software (Media Cybernetics Inc., Bethesda, MD).

RT-PCR and qRT-PCR of Liver RNA.

RT-PCR of CCN2 mRNA isolated from frozen liver tissue was performed as described above for FL83B mouse hepatocytes. For qRT-PCR of mouse liver RNA, the reverse-transcription step was performed as described above for mouse FL83B cells. Real-time amplification was then performed with SYBRGreen PCR Master Mix reagent (Applied Biosystems) using an ABI Prism 7500. The 20 μL reaction volume contained 30 ng cDNA, 10 μL 2× PCR Master mix, and 0.4μL of each primer. After initial denaturation at 95°C for 10 minutes, 40 PCR cycles were run at 95°C for 15 seconds, followed by 60°C for 1 minute. The primers were: mouse β-actin, 5′-TGTTACCAACTGGGACGACA-3′ (forward), 5′-CTTTTCACGGTTGGCCTTAG-3′ (reverse); mouse α-SMA, 5′-GGCTCTGGGCTCTGTAAGG-3′ (forward), 5′-CTCTTGCTCTGGGCTTCATC-3 (reverse); mouse collagen α1 (I), 5′-CCAAGGGTAACAGCGGTGAA-3′ (forward), 5′-CCTCGTTTTCCTTCTTCTCCG-3′ (reverse); mouse tissue inhibitor of metalloprotease-1, 5′-GCATCTCTGGCATCTGGCATC-3′ (forward), 5′-GCGGTTCTGGGACTTGTGGGC-3′ (reverse); mouse matrix metalloprotease-2, 5′-TTCCCCCGCAAGCCCAAGTG-3′ (forward), 5′-GAGAAAAGCGCAGCGGAGTGACG-3′ (reverse); and mouse TGF-β1, 5′-GGTTCATGTCATGGATGGTGC-3′ (forward), 5′-TGACGTCACTGGAGTTGTACGG-3′ (reverse).

Western Blotting and Marker Assays.

Liver extracts were prepared by homogenization of 100 mg tissue in 1 mL immunoprecipitation buffer (see above). Clarified tissue supernatant containing 50 μg of total protein was subjected to SDS-PAGE and western blotting using anti-CCN2,18 anti–β-actin (see above), or monoclonal anti–α-SMA (clone 1A4; Sigma-Aldrich). Intensities of immunoreactive bands were quantified using image analysis software (Scion Corp, Frederick, MD). Serum levels of aspartate aminotransferase and alanine aminotransferase were determined by the Mouse Phenotyping Shared Resource (Ohio State University Comprehensive Cancer Center, Columbus, OH). Serum hyaluronic acid levels were determined using a commercial kit (Corgenix Inc., Broomfield, CO). Acid-soluble hepatic collagen was determined by way of Sircol assay (Biocolor Ltd., Carrickfergus, UK).

Statistical Analysis.

Each experiment was performed between three and eight times. Data were typically obtained from three to five animals in each experimental group using duplicate or triplicate determinations for each animal. For comparison of individual data points, the Student t test was applied. Statistical significance was set at P < 0.05.

Results

Production of Native CCN2 by Mouse Hepatocytes In Vivo.

We first verified that mouse hepatocytes naturally produce CCN2 in vivo in response to hepatic insult. Ethanol administration caused a robust increase in CCN2 mRNA expression in the liver parenchyma as assessed by in situ hybridization and an increase in the frequency of CCN2-positive hepatocytes from 13% to 85% as assessed with FACS (Fig. 1B), thus supporting studies of CCN2 expression in mouse hepatocytes using a transgenic approach.

Cloning and Expression of pALB/CCN2 in Mouse Hepatocytes.

The functionality of pALB/hCCN2 was verified by way of transient transfection of mouse FL83B hepatocytes in vitro. Using species-specific CCN2 primers, mock-transfected cells were shown by RT-PCR to contain only the 734 bp mCCN2 transcript, whereas this transcript plus the 632-bp hCCN2 transcript was present in pALB/hCCN2-transfected cells (Fig. 1C, panels a,b). As assessed by immunoprecipitation, transfected cells also contained greater quantities of the 38-kDa CCN2 protein when compared with the mock-transfected cells (Fig. 1C, panel d). β-Actin mRNA expression and protein production were comparable in mock- and CCN2-transfected cells (Fig. 1C, panels c,e). It was thus concluded that the pALB/hCCN2 plasmid was functional in hepatocytes.

Production and Phenotypic Characterization of hCCN2 Transgenic Mice.

Limited PCR performed on tail DNA showed that the intensity of the hCCN2 PCR product was greater in hccn2+/+ (homozygous) than in hccn2+/− (heterozygous) mice, and was absent from wild-type (hccn2−/−) mice (Fig. 2A, upper panel). These data were supported by Southern blotting using a hCCN2 probe (Fig. 2A, lower panel) and were subsequently verified by quantitative PCR (Fig. 2B). The hCCN2 transgene copy number was 2.89 ± 0.15 for hccn2+/− (n = 35) and 4.77 ± 0.16 for hccn2+/+ (n = 11).

Figure 2.

Genotypic characterization and CCN2 production in livers of hCCN2 transgenic mice. (A) Semiquantitative PCR (upper panel) and Southern blotting (lower panel) of DNA isolated from mouse tails. (B) Quantitative PCR in which each DNA sample was analyzed in three separate reactions (mean ± standard deviation). Representative data are shown for three individual mice, each of which has a distinct hccn2 genotype. (C) RT-PCR of hepatic RNA from wild-type or transgenic mice using mCCN2-specific or hCCN2-specific primers confirming the presence of hCCN2 transcripts in transgenic but not wild-type mice. (D) CCN2 immunohistochemistry of liver sections from (a) wild-type (hccn2−/−) or (b) transgenic (hccn2+/+) mice. (E) FACS analysis of CCN2-positive hepatocytes isolated from the livers of wild-type or transgenic (hccn2+/+) mice. The inset shows a CCN2 western blot of whole liver lysate normalized for total protein loading from wild-type (lane 1) or transgenic (hccn2+/+) (lane 2) mice.

Hepatic hCCN2 mRNA was verified by way of RT-PCR, which further demonstrated that hCCN2 expression was greater in hccn2+/+ than hccn2+/− mice, and that no hCCN2 expression was evident in hccn2−/− littermates (Fig. 2C). Hepatic expression of endogenous mCCN2 mRNA was identical between all three strains of mice (Fig. 2C). Immunohistochemistry revealed stronger CCN2 staining in the parenchyma of hccn2+/+ livers as compared with hccn2−/− livers (Fig. 2C). Additionally, increased amounts of immunoreactive CCN2 in hccn2+/+ mice compared with hccn2−/− were demonstrated by way of western blotting of liver lysates (Fig. 2E, inset) and FACS analysis of isolated hepatocytes (Fig. 2E).

CCN2 transgenic mice were alert, healthy, active, and exhibited normal growth curves, weights, appetite, fertility, pregnancy rate, gestational period, litter size, and birth weight (data not shown). Individual ZY2 transgenic mice were maintained up to 29 months of age with no clinical symptoms in either males or females. Neither hccn2+/+ nor hccn2+/− mice exhibited any histological abnormalities. Hematoxylin-eosin staining of liver sections was unremarkable and showed no differences compared with wild-type littermates (Fig. 3A). The staining pattern of α-SMA or collagen in transgenic mice was also unremarkable and was identical to that of wild-type mice, with α-SMA present in the smooth muscle cells of the blood vessels and collagen fibers localized to the vasculature as expected (Fig. 3A). Serum liver function tests (aspartate aminotransferase, alanine aminotransferase) were not significantly altered in transgenic mice (Fig. 3B).

Figure 3.

CCN2 transgenic mice exhibit normal hepatic histology and LFTs. (A) Histological features of livers from wild-type or transgenic mice. Liver sections of mice were stained with hematoxylin-eosin (top row), anti–α-SMA (middle row), or Sirius Red (bottom row) (magnification ×20; insets show detail at ×40). Data are representative of 24 mice observed from eight individual experiments (n = 3 per group). (B) Serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) levels in individual wild-type or transgenic mice over the first 28 weeks of life.

Enhanced Susceptibility of CCN2 Transgenic Mice to Liver Fibrosis.

When animals were chronically treated with a dose of CCl4 that was selected to induce submaximal stimulation of hepatic fibrosis in wild-type mice, the fibrotic response to this dose was noticeably enhanced in transgenic mice (Figs. 4, 5). This difference was evident when liver sections were stained either for α-SMA (Fig. 4A) or collagen (Fig. 5A) and was more pronounced in homozygous mice than heterozygous mice. As assessed by image analysis, the amount of staining for α-SMA or collagen was positively correlated with the hccn2 genotype and a statistically significant increase was apparent in hccn2+/+ mice as compared with hccn2−/− mice (P < 0.05) (Figs. 4B, 5B). A statistically significant exacerbated fibrotic response to CCl4 in hccn2+/+ mice was also evident when hepatic extracts were assessed for α-SMA levels by way of western blotting (Fig. 4C) or for acid-soluble collagen content by Sircol assay (Fig. 5C) (P < 0.05). As assessed by qRT-PCR, treatment of either hccn2−/− mice or hccn2+/+ mice with CCl4 caused an increase in the mRNA expression of α-SMA, collagen α1 (1), matrix metalloprotease-2, and tissue inhibitor of metalloprotease-1 (Fig. 6A). However, CCl4-induced expression of all transcripts was greater in hccn2+/+ mice compared and was statistically higher (P < 0.05) for α-SMA, collagen α1 (1), and tissue inhibitor of metalloprotease-1. Expression of TGF-β1 mRNA in wild-type or transgenic mice was moderately enhanced after chronic CCl4 treatment, but this response was not exacerbated by the presence of the CCN2 transgene (Fig. 6A). Finally, in CCl4-treated animals, circulating hyaluronic acid levels were significantly greater in hccn2+/+ mice as compared with hccn2−/− mice (P < 0.05) (Fig. 6B).

Figure 4.

Enhanced CCl4-induced α-SMA production in CCN2 transgenic mice as compared with wild-type mice. (A) Mice were treated intramuscularly once per day on 4 successive days each week for 5 weeks with either 30 μL of vegetable oil or a mixture of 0.5 μL CCl4 in 29.5 μL of vegetable oil. Livers were resected, fixed, and processed for α-SMA immunohistochemistry. Staining is representative of 24 mice observed from eight individual experiments (n = 3 per group). (B) α-SMA staining on tissue sections was quantified by image analysis of 10 ×20 fields (mean ± standard deviation). *P < 0.05 versus hccn2−/−. (C) Presence of α-SMA in hepatic tissue extracts relative to that of β-actin as assessed by densitometric scanning of western blots from four animals per group (mean ± standard deviation). *P < 0.05 versus hccn2−/−. The inset shows a representative western blot.

Figure 5.

Enhanced CCl4-induced collagen deposition CCN2 transgenic mice as compared with wild-type mice. (A) Mice were treated as described in Fig. 4, and liver sections were stained with Sirius Red to detect collagen. Staining is shown at ×20 and is representative of 24 mice observed from eight individual experiments (n = 3 per group). (B) Sirius Red staining on tissue sections was quantified by image analysis of 10 ×20 fields (mean ± standard deviation). *P < 0.05 versus hccn2−/−. (C) Sircol assay of acid-soluble collagen in hepatic tissue extracts (mean ± standard deviation; n = 5 per group). *P < 0.05 versus hccn2−/−.

Figure 6.

Enhanced CCl4-induced production of fibrosis-related gene products in CCN2 transgenic mice as compared with wild-type mice. (A) qRT-PCR of mRNA from hccn2−/− or hccn2+/+ livers after chronic exposure to CCl4 for 5 weeks. Data are triplicate determinations (mean ± standard deviation) of relative mRNA expression from three mice in each group and are representative of three experiments. *P < 0.05 versus hccn2−/−, oil; **P < 0.05 versus hccn2+/+, oil; #P < 0.05 versus hccn2−/−, CCl4. (B) Serum hyaluronic acid levels in hccn2−/− or hccn2+/+ mice after chronic CCl4 treatment for 5 weeks (mean ± standard deviation; n = 5 per group). *P < 0.05 versus hccn2−/−.

To verify that the profibrotic effect of the CCN2 transgene after liver injury was not restricted to the effect of hepatotoxins such as CCl4, mice alternatively underwent BDL, and their livers were examined for collagen deposition 2 weeks later. As detected by Sirius Red staining, there was a marked increase in BDL-induced collagen deposition in hccn2+/+ mice as compared with hccn2−/− mice (P < 0.05) (Fig. 7), whereas no fibrotic response was present in sham-operated mice (data not shown). These data thus show that the effects of distinct fibrotic stimuli were exacerbated by expression of the hepatic hCCN2 transgene.

Figure 7.

Enhanced BDL-induced collagen deposition in CCN2 transgenic mice as compared with wild-type mice. (A) Mice were subjected to BDL and, 2 weeks later, livers were removed, sectioned, and stained with Sirius Red (magnification ×20). (B) Quantification of Sirius Red staining of tissue sections by image analysis of 10 ×20 fields from four mice in each group (mean ± standard deviation). *P < 0.05 versus hccn2−/−.

Discussion

Research over the last 15–20 years has firmly established that CCN2 is involved in fibrosis. Although many studies have focused on aspects of CCN2 gene regulation at the molecular level, the precise role of CCN2 in fibrogenic cascades has been difficult to ascertain because of the lack of availability of high-quality purified CCN2 with which to perform critical studies and because its actions as a matricellular protein are difficult, if not impossible, to fully model using in vitro approaches. Moreover, only a limited number of transgenic systems have been reported for CCN2,24–26 and none of these have addressed either its action in the liver or, somewhat surprisingly, its fibrotic properties. To address these deficiencies, we created transgenic mice that overproduced hepatic CCN2 under the control of the albumin enhancer/promoter in hepatocytes, a legitimate cell type given its ability produce CCN2 in response to various insults, often downstream of TGF-β10–15 (Fig. 1).

Our major finding was that hepatocyte overexpression of CCN2 per se was nonfibrotic but that it resulted in a potentiation of CCl4- or BDL-stimulated hepatic fibrosis. It is important to note that, under normal circumstances, there was no histopathology or evidence of HSC activation in the livers of CCN2 transgenic mice. In vitro studies have shown that CCN2 does not interact with quiescent HSCs but that it potentiates fibrogenic and survival pathways in activated HSCs, which express up-regulated levels of CCN2 receptors such as integrin αvβ3.27 This fact, coupled with the absence of other injury cascades in transgenic livers that might otherwise cause HSC activation, likely explains the inability of the CCN2 transgene alone to drive fibrosis.

Fibrosing stimuli ordinarily cause the elaboration of multiple signaling pathways as opposed to changes in CCN2 expression alone. Thus, the pathophysiological role of CCN2 is best evaluated in the context of the fibrogenic cascades, of which it is a normal component. We therefore conducted additional studies in transgenic mice subjected to CCl4 treatment or BDL and found that the presence of the hepatocyte CCN2 transgene constitutes a first hit that sensitized the liver to a different second hit, resulting in more advanced fibrosis.

The increase in fibrosis susceptibility in CCN2 transgenic mice may be related to the additive effects of the hCCN2 protein on intrinsic CCN2 target pathways that were activated by CCl4 or BDL, or their cooperative binding interactions with other molecules in the fibrogenic cascade such as fibronectin or TGF-β1. For example, CCN2 binds to fibronectin28 and increases integrin-dependent adhesion to fibronectin by fibroblasts,29 chondrocytes,30 hepatic oval cells,31 or activated mouse HSCs (unpublished data). Also, TGF-β1 binds directly to CCN2, which increases TGF-β bioactivity due to its enhanced affinity for the type II TGF-β receptor,32 a phenomenon that may account for the reported synergism between CCN2 and TGF-β in stimulating dermal fibrosis in vivo.33 A synergistic action of this type seems plausible in our model, because many fibrotic readouts downstream of TGF-β were amplified in CCl4-treated transgenic mice versus wild-type mice, even though expression of CCl4-induced TGF-β itself was comparable between the two groups of animals. Although there are no other transgenic CCN2 fibrosis models with which to draw comparisons, expression of a CCN2 transgene in mouse podocytes in vivo was insufficient to cause glomerular damage, but when challenged in a setting of a streptozotocin-induced type 1 diabetes the transgenic mice exhibited increased glomerular pathology as compared with wild-type mice.25 This exacerbated response included enhanced matrix expansion that was proposed to reflect the combined matrigenic actions of CCN2 and other factors such as TGF-β.25 Lastly, cooperative interactions between CCN2 and TGF-β in vivo were shown in Balb/c mice which, though genetically resistant to Smad-3–mediated bleomycin-induced lung fibrosis, became susceptible to fibrosis when treated concomitantly with bleomycin and a CCN2-expressing adenovirus.34 With respect to genetic resistance, the FVB strain used in our study is relatively resistant to hepatic fibrosis,35 raising the question of whether this phenomenon might account for the lack of hepatic phenotype in FVB hCCN2 transgenic mice. However, we have back-crossed the hCCN2 transgene onto the more susceptible Balb/c strain35 and have found that liver histology remains normal over all generations examined (up to and including F11; data not shown). Thus, the various in vivo studies to date show that CCN2 overexpression alone does not cause tissue damage and/or fibrosis and that additional changes in the microenvironment are required for the deleterious functions of CCN2 to be manifested, a mode of action that is entirely consistent with its function as a matricellular protein.

In conclusion, our studies demonstrate that expression of an albumin promoter-driven CCN2 transgene in mouse hepatocytes in vivo does not cause hepatic injury or fibrosis but renders the livers more susceptible to the injurious actions of another fibrotic stimulus. These studies provide further evidence in support for a central role of CCN2 in hepatic fibrosis and emphasize the role of the microenvironment in regulating the profibrotic action of CCN2.

Acknowledgements

We thank Dr. Richard Palmiter (University of Washington, Seattle, WA) for providing p2335A-1. We acknowledge the help of Dr. Donna Kusewitt with liver function tests and Dr. Andrei Radulescu with animal surgery.

Ancillary