The gene products of two members of the EXT (exostosin) gene family, EXT1 and EXT2, function together as a polymerase in the biosynthesis of heparan sulfate. EXTL2 (EXT-like 2), one of the three EXTL genes in the human genome that are homologous to EXT1 and EXT2, encodes an N-acetylhexosaminyltransferase. We have demonstrated that EXTL2 terminates chain elongation of GAGs (glycosaminoglycans), and thereby regulates GAG biosynthesis. The abnormal GAG biosynthesis caused by loss of EXTL2 had no effect on normal development or normal adult homoeostasis. Therefore we examined the role of EXTL2 in CCl4 (carbon tetrachloride)-induced liver failure, a model of liver disease. On the fifth day after CCl4 administration, the liver/body weight ratio was significantly smaller for EXTL2-knockout mice than for wild-type mice. Consistent with this observation, hepatocyte proliferation following CCl4 treatment was lower in EXTL2-knockout mice than in wild-type mice. EXTL2-knockout mice experienced less HGF (hepatocyte growth factor)-mediated signalling than wild-type mice specifically because GAG synthesis was altered in these mutant mice. In addition, GAG synthesis in hepatic stellate cells was up-regulated during liver repair in EXTL2-knockout mice. Taken together, the results of the present study indicated that EXTL2-mediated regulation of GAG synthesis was important to the tissue regeneration processes that follow liver injury.

INTRODUCTION

GAGs (glycosaminoglycans) are abundant on the surfaces of most cells and in extracellular matrices as components of PGs (proteoglycans) [1]. PGs play important roles in homoeostasis at the cellular and tissue level because they function as cofactors in a variety of biological processes, including cell growth, cell differentiation, embryonic development and postnatal tissue remodelling [2]. These functions of PGs are mediated by interactions between GAG chains and bioactive proteins such as growth factors and cytokines. In addition, diverse protein ligands recognize the specific saccharide sequences of the GAG side chains [3]. Thus abnormal GAG biosynthesis causes disordered cellular function; moreover, it seems to be implicated in various disease conditions.

HS (heparan sulfate) and CS (chondroitin sulfate) are attached to specific serine residues of core proteins through the GAG–protein linkage region GlcUAβ1-3Galβ1-3Galβ1-4Xylβ1-O-Ser. The synthesis of this linkage region is initiated by the addition of Xyl (xylose) to serine residue followed by the addition of two Gal (galactose) residues and is completed by the addition of GlcUA (glucuronic acid), with each reaction being catalysed by the respective specific glycosyltransferases. After a linkage region tetrasaccharide is formed, a GAG chain is built up on this linker by the alternative addition of N-acetylhexosamine and GlcUA residues. GAG chains are modified with sulfotransferases and epimerases to generate structural diversity. Sulfate groups at precise positions along the carbohydrate backbone form specific sulfation motifs, which function as molecular recognition elements for bioactive proteins such as growth factors. That is, GAG polymers encode functional information in a sequence-specific manner as do DNA, RNA and protein polymers. Thus tight regulation of GAG biosynthesis is essential for maintenance of normal cell-level and tissue-level function.

The EXT (exostosin) gene comprises five members: EXT1, EXT2, EXTL1 (EXT-like 1), EXTL2 and EXTL3. EXT1 and EXT2 are associated with HMEs (hereditary multiple exostoses), an autosomal dominant disorder characterized by aberrant bone formation that occurs most commonly in the juxtaepiphyseal regions of the long bones [46]. Because knockout of EXT1, EXT2 or EXTL3 from mice caused a substantial decrease in the amount of HS [79], these genes are thought to play a crucial role in HS biosynthesis in vivo. On the basis of an in vitro study, EXTL2 is known to be an N-acetylhexosaminyltransferase that transfers not only GalNAc (N-acetylgalactosamine), but also GlcNAc (N-acetylglucosamine) to the tetrasaccharide linker via an α1,4-linkage [10]. The function of EXTL2 in GAG biosynthesis has been established only recently [11]. We have demonstrated that EXTL2 can control in vivo GAG biosynthesis via a novel mechanism in which EXTL2 transfers a GlcNAc residue to a phosphorylated tetrasaccharide linker to stop chain elongation [11]. Therefore mice deficient in EXTL2 overproduce GAGs. In the present paper, we report that these excess GAGs in EXTL2−/− mice affect the regeneration processes that follow liver injury.

EXPERIMENTAL

Animals

EXTL2-knockout mice were generated as described previously [11]. Mice were kept under pathogen-free conditions in an environmentally controlled, clean room at the Institute of Laboratory Animals, Kobe Pharmaceutical University; animals were maintained on standard rodent food and on a 12 h light/12 h dark cycle. All experiments were conducted according to the institutional ethical guidelines for animal experiments and safety guidelines for gene manipulation experiments. All animal procedures were approved by the Kobe Pharmaceutical University Committee on Animal Research and Ethics.

qPCR (quantitative real-time PCR) analysis

Total RNA was extracted from cells by the guanidine-phenol method using TRIzol® reagent (Invitrogen) according to the manufacturer's protocols. Aliquots (1 μg) of total RNA were digested with 2 IU of RQ1 RNase-free DNase (Promega) for 30 min at 37°C and then incubated for 10 min at 65°C with stop solution (Promega). For reverse transcription, total RNA (0.75 μg) was treated with M-MLV (Moloney murine leukaemia virus) reverse transcriptase (Invitrogen) using random primers [nonadeoxyribonucleotide mixture; pd(N)9] (Takara Bio). qPCR was conducted using FastStart DNA Master plus SYBR Green I and a LightCycler 1.5 (Roche Applied Science) according to the manufacturer's protocols. The housekeeping gene GAPDH (glyceraldehyde-3-phosphate dehydrogenase) was used as an internal control for quantification. The primers used for qPCR are shown in Supplementary Table S1 (at http://www.biochemj.org/bj/454/bj4540133add.htm).

Disaccharide analysis of GAGs from mouse tissues

GAGs were isolated and purified from mouse tissues and cells as described previously [12]. Briefly, tissues or cells were homogenized, extracted with acetone three times and air-dried thoroughly. The dried materials were digested with 10% (w/v) heat-activated actinase E in 0.1 M sodium borate, pH 8.0, containing 10 mM CaCl2 at 55°C for 48 h. The samples were adjusted to 5% (w/v) trichloroacetic acid and centrifuged (16500 g). The resultant supernatants were extracted with diethyl ether three times to remove trichloroacetic acid, and then neutralized using 20% (w/v) NH4HCO3. The aqueous phase containing 5% (w/v) sodium acetate was adjusted to 80% (v/v) ethanol and left overnight at −30°C. The resultant precipitate was dissolved in 50 mM pyridine acetate (pH 5.0), and subjected to gel filtration on a PD-10 column (GE Healthcare) using 50 mM pyridine acetate (pH 5.0) as an eluent. The flow-through fractions were collected and evaporated to dryness. Purified GAGs were digested with CSase ABC (chondroitinase ABC) from Proteus vulgaris (EC 4.2.2.4) (10 mIU) or with a mixture of Hepase (heparinase) from Flavobacterium heparinum (EC 4.2.2.7) (1 mIU) and HSase (heparitinase) from F. heparinum (EC 4.2.2.8) (1 mIU) at 37°C for 4 h. The digests were derivatized with a fluorophore 2-aminobenzamide and then analysed by HPLC as reported previously [13].

Model of liver disease on the basis of CCl4 (carbon tetrachloride)-induced injury

EXTL2+/+ and EXTL2−/− male mice (8–10 weeks old) were fasted overnight (day 0, 19:00 h, ~day 1, 15:00 h) before being treated with CCl4. On day 1, mice were injected intraperitoneally with a 20% (v/v) solution of CCl4 (Sigma–Aldrich) diluted in olive oil (WAKO chemicals) at a dose of 8 ml/kg body weight. Mice were treated with pentobarbital and killed at 0, 12, 24 or 48 h after administration of CCl4. Blood samples were then collected from abdominal aorta for subsequent biochemical analyses, and liver samples were processed for RNA and protein analyses, histology and immunostaining. Serum ALT (alanine aminotransferase) activity and AST (aspartate aminotransferase) activity were measured using TA-LN kainos (Kainos Laboratories), and serum HGF (hepatocyte growth factor) concentration was measured using Quantikine® ELISA Mouse/Rat HGF (R&D Systems). To monitor cell proliferation in vivo, mice were injected intraperitoneally with 1 ml (1 mg) of BrdU (bromodeoxyuridine) solution (Pharmingen, BD Biosciences) and killed at 3 h after BrdU injection; the In Situ BrdU Detection Kit (Pharmingen, BD Biosciences) and Streptavidin Alexa Fluor® 488 conjugate (S-11223, Invitrogen) were used to monitor BrdU incorporation.

Histological analysis

Samples of liver tissues were fixed with 10% (v/v) Formalin Neutral Buffer Solution (pH 7.4) (062-01661, WAKO Pure Chemical Industries), and embedded in paraffin; paraffin blocks were then cut into 5-μm sections. Some sections were stained with haematoxylin and eosin using the standard procedures.

Localization of HGF in the liver

Paraffin sections (10 μm thickness) of EXTL2+/+ and EXTL2−/− liver at 24 h after administration of CCl4 were deparaffined, incubated in 3% (w/v) H2O2/PBS for 10 min at room temperature (25°C), and washed with PBS. Antigen retrieval was carried out in 0.01 M citrate buffer (pH 6.0) using a microwave oven. After washing with PBS, the sample slides were digested with or without HSase (25 mIU) and Hepase (25 mIU) for 1 h at 37°C. For blocking, the samples were incubated with PBST (PBS/0.1% Tween 20) and 2% (w/v) BSA for 1 h at room temperature, and then reacted with 100 μl/slide of polyclonal antibody specific against mouse/rat HGF conjugated with horseradish peroxidase (894094, R&D Systems). After washing with PBST, detection was carried out using DAB (diaminobenzidine) substrate kit (Pharmingen, BD Biosciences), and then counterstained with haematoxylin.

Isolation of primary mouse hepatocytes

A collagenase perfusion method described previously [14] was used to isolate primary hepatocytes from EXTL2+/+ and EXTL2−/− male mice (8–10 weeks old). Isolated hepatocytes were separated from non-parenchymal cells by low-speed centrifugation technique (hepatocyte purity > 99%). Hepatocyte suspensions were plated at a density of 3.0×105 cells/cm2 on to 35-mm plastic dishes coated with calf-skin collagen (Calbiochem); the final suspension volume was 2 ml; the suspension medium was William's E Medium (Sigma–Aldrich) supplemented with 5 μg/ml of insulin (Roche), 0.1 μM dexamethasone, 10% (v/v) fetal bovine serum, 100 units/ml penicillin and 100 μg/ml streptomycin. Combined supplementation of insulin and dexamethasone can improve the cell attachment efficiency. After cells were allowed to attach to the substrate for 2–4 h, the medium was removed and replaced with fresh supplemented William's E Medium. After another 12 h, the medium was removed and replaced with serum-free William's E Medium that contained 0.1% BSA. After 24 h of incubation, the cells were stimulated with 50 ng/ml recombinant mouse HGF (R&D Systems) for defined periods.

Digestion of hepatocytes with HSase and Hepase or CSase ABC

Isolated hepatocytes were treated with William's E Medium containing both 20 mIU of HSase from F. heparinum (EC 4.2.2.8) and 20 mIU Hepase from F. heparinum (EC 4.2.2.7) (Seikagaku Biobusiness) or with William's E Medium containing 10 mIU of CSase ABC from Proteus vulgaris (EC 4.2.2.4) (Seikagaku Biobusiness) for 2 h in a 5% CO2 incubator at 37°C.

Analysis of the activation of c-Met in mouse liver and hepatocytes

Samples of liver tissue (100 mg wet weight) were taken from animals at 12 or 24 h after administration of CCl4 and were then homogenized in 20 mM Tris/HCl (pH 7.5) containing 1% (v/v) Nonidet P40, 10% (v/v) glycerol, 0.15 M NaCl, 1 mM EDTA, 1 mM sodium orthovanadate (New England BioLabs) and protease inhibitor cocktails [100 mM 4-(2-aminoethyl) benzensulfonyl fluoride hydrochloride, 80 μM aprotinin, 1.5 mM E-64, 2 mM leupeptin hemisulfate monohydrate, 5 mM bestatin and 1 mM pepstatin A; Nacalai Tesque]; each homogenate was then subjected to centrifugation for 15 min at 16500 g and 4°C. The protein concentration in each resulting supernatant was measured using a bicinchoninic acid assay (Thermo Fisher Scientific), and 20 μg of proteins were subjected to SDS/PAGE (4–20% gel).

Unstimulated hepatocytes or those stimulated with 50 ng/ml of recombinant mouse HGF (R&D Systems) for the indicated periods were washed with PBS and lysed in PhosphoSafe Extraction Reagent (Novagen) that contained protease inhibitor cocktails (Nacalai Tesque) for 5 min at room temperature. Protein extracts were dispersed by passing them through a 23-gauge needle attached to a syringe; these suspensions were subjected to centrifugation at 2300 g for 10 min, and the protein concentration in each resulting supernatant was measured using a bicinchoninic acid assay. Protein samples (30 μg each) were mixed with three rounds of SDS loading buffer [187.5 mM Tris/HCl (pH 6.8), containing 6% (w/v) SDS, 30% (v/v) glycerol and 0.03% Phenol Red] and boiled for 5 min. Each sample was subjected to SDS/PAGE (7.5% gel), transferred to Hybond-P PVDF filters (GE Healthcare), and analysed using a procedure described previously [15], involving the ECL Advance™ Western Blotting Detection Kit (GE Healthcare) according to the manufacturer's protocol. Phospho-c-Met (Tyr1234/Tyr1235) was detected with a 1:1000 dilution of rabbit anti-c-Met (Tyr1234/Tyr1235) polyclonal antibody (3126, Cell Signaling Technology); a 1:200 dilution of anti-c-Met (B-2) mouse mAb (monoclonal antibody) (sc-8057, Santa Cruz Biotechnology) was used to detect total c-Met protein.

Cytotoxicity assay

Primary hepatocytes were plated on to a 24-well plate at a density of 8×104 cells/well on day 0. On day 1, cells were washed with PBS and incubated for 24 h in Dulbecco's modified Eagle's medium containing 0.1% BSA, 100 units/ml penicillin and 100 μg/ml streptomycin. On day 2, cells were left unstimulated or were stimulated with 100 ng/ml HGF for 60 min, and then left untreated or treated with 2.5 mM CCl4 for 4 h in the presence of HGF. After cells in each of the four groups were washed with PBS, cellular lactate dehydrogenase was released from cells by adding 1 μl of 1% (v/v) Triton X-100. The enzyme activity of cellular lactate dehydrogenase was measured as an indicator of cell viability using CytoTox-ONE Homogeneous Membrane Integrity Assay (Promega).

Immunohistochemistry for detection of PGs and GAGs

While the mice were under diethyl ether anaesthesia, liver tissues were fixed using the perfusion-fixation method. Briefly, a 27-gauge needle was inserted into the left ventricle; each animal was perfused with 40–50 ml of pre-chilled PBS at steady speed using a perista pump, and then perfused with approximately 50 ml of pre-chilled 4% (w/v) paraformaldehyde/PBS. After perfusion-fixation, liver tissue was extirpated and immersed for fixation in 4% (w/v) paraformaldehyde/PBS overnight at 4°C. Fixed livers were sectioned at a thickness of 30 μm in PBS using a Vibratome (Leica VT1200S). Tissue sections were transferred on to MAS-GP type A slides (Matsunami Glass Ind). After washing with PBS, sections were treated with 0.2% Triton X-100/PBS for 15 min at room temperature. For staining using anti-SDC2 (syndecan 2) and anti-SDC4 antibodies, non-specific binding sites were blocked with 10% (v/v) normal goat serum (Sigma–Aldrich) in PBST for 1 h. For staining using mAbs HepSS-1 (270426), NAH46 (370735), JM403 (370730) (Seikagaku Biobusiness), and CS-56 (C8035, Sigma–Aldrich), 2% (w/v) BSA/PBST was used for blocking. For staining with mAb F69-3G10 (370260, Seikagaku Biobusiness), liver sections were double digested with HSase (5 mIU) and Hepase (5 mIU): liver sections were digested with CSase ABC (10 mIU) for staining with mAbs 2-B-6 (270432) (Seikagaku Biobusiness) respectively; all digestion reactions were carried out in a total volume of 400 μl at 37°C for 40 min. After digestion, sections were blocked with 10% (v/v) goat serum/PBST for 30 min at room temperature, and then treated overnight at 4°C with 0.1 mg/ml AffiniPure Fab Fragment goat anti-(mouse IgG) (H+L) (115-007-003, Jackson ImmunoResearch Laboratories) to block endogenous mouse IgG. After blocking, liver sections were treated with the following primary antibodies: anti-SDC2 antibody (M-140) (sc-15348, Santa Cruz) and anti-SDC4 antibody (H-140) (sc-15350, Santa Cruz), which were used at a dilution of 1:50 with Protein Block Serum-Free Ready-to-Use (X0909, Dako). In addition, liver sections were incubated overnight at 4°C with HepSS-1, NAH46, JM403, CS-56, F69-3G10 or 2-B-6 diluted at 1:200 with Protein Block Serum-Free Ready-to-Use. The following secondary antibodies were used to detect primary antibodies: Alexa Fluor® 594 donkey anti-(rabbit IgG) (H+L) (1:400 dilution) (A-21207, Molecular Probes®), goat anti-(mouse IgM) (μ chain) antibody DyLight™ 488 conjugated (1:400 dilution) (610-141-007, Rockland Immunochemicals), and DyLight™ 488 conjugated AffiniPure Fab Fragment goat anti-(mouse IgG) (H+L) (1:250 dilution) (115-487-003, Jackson ImmunoResearch Laboratories). DNA in nuclei was stained with 5 μg/ml Hoechst 33342. Images were acquired with a Zeiss LSM 700 confocal laser scanning system (Carl Zeiss) that was equipped with an inverted Axio observer Z1 microscope, Objective Plan-Apochromat 10×/0.45, Objective Plan-Apochromat 20×/0.8 and Objective Plan-Apochromat 40×/1.3 Oil; subsequent image analysis was performed with LSM software ZEN 2009 (Carl Zeiss).

Immunohistochemistry for detection of hepatic non-parenchymal cells

Vibratome sections (30 μm/slice) or paraffin sections (10 μm/slice) were stained with rabbit anti-GFAP (glial fibrillary acidic protein) polyclonal antibody (1:200 dilution) (G9269, Sigma–Aldrich) or mouse anti-α-SMA (α-smooth muscle actin) mAb (1A4) (1:50 dilution) (ab7817, Abcam). For formalin-fixed tissue sections embedded in paraffin, an antigen retrieval process was needed. Tissue sections on slides in 0.01 M citrate buffer (pH 6.0) were heated for 5 min in a microwave oven, and slowly cooled down to room temperature for 20 min. For staining with anti-α-SMA antibody, endogenous mouse IgG was blocked overnight at 4°C with 0.1 mg/ml AffiniPure Fab Fragment goat anti-mouse IgG (H+L), and DyLight™ 488 conjugated AffiniPure Fab fragment goat anti-(mouse IgG) (H+L) (1:250 dilution) was used as a secondary antibody.

Statistical analysis

Data are expressed as means±S.D. Statistical significance was determined by Student's t test.

RESULTS

We generated EXTL2-knockout mice to investigate the biological significance of EXTL2-mediated regulation of GAG biosynthesis. EXTL2−/− mice develop normally and are fertile; however, the biosynthesis of GAG in the liver is affected by deficiency of EXTL2 as reported recently [11]. Thus we examined the role of EXTL2 under pathological conditions. GAGs played an important role in the regeneration after liver injury, and liver GAGs were significantly affected by a loss of EXTL2; therefore we investigated whether the abnormal GAG synthesis caused by loss of EXTL2 affected the liver repair following CCl4-induced liver injury. We first checked expression levels of CYP2E1, which is the primary cytochrome P450 that catalyses CCl4 metabolism. Following CCl4 injection, the CYP2E1 mRNA levels in EXTL2+/+ and EXTL2−/− mice were not significantly different (results not shown). We next examined the ratio of liver weight to body weight 5 days after a single administration of CCl4. In wild-type EXTL2+/+ mice, CCl4-treated mice had a higher liver weight/body weight ratio than untreated mice (Figure 1B), as reported previously [16]; this finding indicated that, in wild-type mice, liver regeneration proceeded normally following CCl4-induced liver injury. However, 5 days after a single CCl4 injection, the liver weight/body weight ratio of EXTL2−/− mice was significantly lower than that of EXTL2+/+ mice (Figure 1B). In addition, each lobe from EXTL2−/− mice liver was on average smaller than the respective lobe from EXTL2+/+ mice (Figures 1A and 1C). These observations indicated that liver regeneration processes were inhibited in EXTL2−/− mice. We next assessed liver damage by measuring serum ALT and AST levels. Compared with those in EXTL2+/+ mice, serum ALT and AST activities in EXTL2−/− mice increased significantly within 24 h after administration of CCl4, and the serum ALT level in EXTL2−/− mice at 48 h was slightly, yet significantly, higher than that in EXTL2+/+ mice (Figure 1D). CCl4 is a prototypical inducer of pericentral liver damage and causes hepatocytes close to the central vein to die [17]. As expected, on the basis of the serum ALT and AST activities, central necrosis was evident at 24 and 48 h after administration of CCl4 (Figure 1E, white broken lines). After 120 h, massive inflammatory lymphocytic infiltrates, hallmarks of severe hepatic injury, were observed in EXTL2−/− mice (Figure 1E, white arrows; inset), but not in EXTL2+/+ mice. Furthermore, BrdU incorporation (Figures 2A and 2B) and the expression of CCND1 (cyclin D1) (Figure 2C) were examined as indicators of cell proliferation. Low-magnification images in Figure 2(A) of BrdU incorporation indicate that cell proliferation was remarkably inhibited in EXTL2−/− mice following CCl4-induced liver injury. In Figure 2(B), the numbers of BrdU-positive cells of each cell type [e.g. hepatocytes and non-parenchymal cells (Figure 2A, arrows)] were counted. The results of the present study indicated that the proliferation of hepatocytes, but not of non-parenchymal cells, in EXTL2−/− mice was depressed relative to that in EXTL2+/+ mice. Thus EXTL2−/− hepatocytes may have had lower proliferative activity and higher sensitivity to CCl4 than EXTL2+/+ hepatocytes.

Comparison between EXTL2+/+ and EXTL2−/− mice with regard to CCl4-induced liver injury

Figure 1
Comparison between EXTL2+/+ and EXTL2−/− mice with regard to CCl4-induced liver injury

(A) Appearance of liver taken from EXTL2+/+ and EXTL2−/− mice 5 days after administration of CCl4. (B and C) Body weight and the ratio of liver weight to body weight for EXTL2+/+ mice and for EXTL2−/− mice. Data represent means±S.D. for values from six (B) and three (C) specimens. (D) Serum ALT and AST activities of EXTL2+/+ (open bar) and EXTL2−/− mice (closed bar) were measured 0 (n=4), 24 (n=7) or 48 h (n=6) after administration of CCl4. Data represent means±S.D. *P<0.05, **P<0.01. (E) Haematoxylin and eosin staining in the liver tissues from EXTL2+/+ and EXTL2−/− mice at 0, 24, 48 and 120 h after administration of CCl4. Asterisks indicate a central vein. The necrotic areas are surrounded by white broken lines. White arrows and the inset in EXTL2−/− mice at 120 h show massive inflammatory lymphocytic infiltrates. Scale bar, 100 μm.

Figure 1
Comparison between EXTL2+/+ and EXTL2−/− mice with regard to CCl4-induced liver injury

(A) Appearance of liver taken from EXTL2+/+ and EXTL2−/− mice 5 days after administration of CCl4. (B and C) Body weight and the ratio of liver weight to body weight for EXTL2+/+ mice and for EXTL2−/− mice. Data represent means±S.D. for values from six (B) and three (C) specimens. (D) Serum ALT and AST activities of EXTL2+/+ (open bar) and EXTL2−/− mice (closed bar) were measured 0 (n=4), 24 (n=7) or 48 h (n=6) after administration of CCl4. Data represent means±S.D. *P<0.05, **P<0.01. (E) Haematoxylin and eosin staining in the liver tissues from EXTL2+/+ and EXTL2−/− mice at 0, 24, 48 and 120 h after administration of CCl4. Asterisks indicate a central vein. The necrotic areas are surrounded by white broken lines. White arrows and the inset in EXTL2−/− mice at 120 h show massive inflammatory lymphocytic infiltrates. Scale bar, 100 μm.

Comparison of EXTL2+/+ and EXTL2−/− mice with regard to proliferation of hepatic cells following CCl4-induced liver injury

Figure 2
Comparison of EXTL2+/+ and EXTL2−/− mice with regard to proliferation of hepatic cells following CCl4-induced liver injury

(A) Incorporation of BrdU in EXTL2+/+ or EXTL2−/− mouse liver was measured 48 h after administration of CCl4. White arrows indicate non-parenchymal hepatic cells. Low-magnification images are shown in panels a and c; high-magnification images are shown in panels b and d. (B) BrdU-incorporated cells were counted. Data are expressed as the means±S.D. for two specimens. (C) The expression level of cyclin D1 (CCND1) was assayed using qPCR of samples taken at 0, 12, 24 or 48 h after administration of CCl4. The mean of the relative expression levels of cyclin D1 and the number of specimens is indicated by bar and in parenthesis respectively.

Figure 2
Comparison of EXTL2+/+ and EXTL2−/− mice with regard to proliferation of hepatic cells following CCl4-induced liver injury

(A) Incorporation of BrdU in EXTL2+/+ or EXTL2−/− mouse liver was measured 48 h after administration of CCl4. White arrows indicate non-parenchymal hepatic cells. Low-magnification images are shown in panels a and c; high-magnification images are shown in panels b and d. (B) BrdU-incorporated cells were counted. Data are expressed as the means±S.D. for two specimens. (C) The expression level of cyclin D1 (CCND1) was assayed using qPCR of samples taken at 0, 12, 24 or 48 h after administration of CCl4. The mean of the relative expression levels of cyclin D1 and the number of specimens is indicated by bar and in parenthesis respectively.

We next examined the mechanism by which the proliferation of hepatocytes was depressed in EXTL2−/− liver tissue. Following CCl4-induced injury, the liver immediately initiates repair processes, as evidenced by the release of growth factors such as EGF (epidermal growth factor) and HGF, and the entry of previously resting hepatocytes into the proliferative state [18]. After exposure to CCl4, sinusoidal cells exhibit more than a 5-fold increase in production of HGF [19]. Following toxin-induced injury, HGF functions not only as a proliferation factor, but also as a hepatic survival factor [20,21]. In addition, liver HS can control HGF signalling via interactions between HS chains and HGF [22]. Thus we investigated HGF signalling in EXTL2+/+ and EXTL2−/− mice. The liver HGF mRNA and serum HGF levels of both EXTL2+/+ and EXTL2−/− mice increased to a similar extent in the response to CCl4 (Figure 3A); this finding indicated that there was little difference between EXTL2+/+ and EXTL2−/− mice with regard to HGF production. In addition, HGF focally bound to EXTL2+/+ liver, but not to EXTL2−/− liver, at 24 h after administration of CCl4 (Figure 3B). HGF molecules in EXTL2−/− liver were not accumulated locally, probably because HGF was dispersed throughout EXTL2−/− liver. In addition, the treatment with HSase and Hepase revealed that the HGF binding was dependent on the cell surface HS chains. We next examined phosphorylation of c-Met; this phosphorylation event is an indicator of HGF signalling in the liver. Levels of phosphorylated c-Met increased substantially in EXTL2+/+ mice at 12 and 24 h after administration of CCl4; in contrast, c-Met phosphorylation levels in EXTL2−/− mice were lower following CCl4 administration (Figure 3C). Moreover, we isolated hepatocytes from the livers of EXTL2+/+ and of EXTL2−/− mice and compared these hepatocyte types with regard to their response to HGF. Within 10 min after HGF stimulation, levels of phosphorylated c-Met increased up to 3-fold over baseline levels in EXTL2+/+ hepatocytes; by 30 min after HGF stimulation, these levels were unchanged (Figure 3D). However, the level of phosphorylated c-Met in EXTL2−/− hepatocytes was lower than that in EXTL2+/+ hepatocytes at 10 min after HGF stimulation, and increased gradually during the 30 min following HGF stimulation (Figure 3D). The results of the present study indicated that HGF signalling was attenuated and delayed in EXTL2−/− hepatocytes.

Comparison between EXTL2+/+ and EXTL2−/− mice with regard to HGF signalling in liver and hepatocytes

Figure 3
Comparison between EXTL2+/+ and EXTL2−/− mice with regard to HGF signalling in liver and hepatocytes

(A) The expression level of liver HGF mRNA and the concentration of serum HGF in EXTL2+/+ (○) and EXTL2−/− mice (●) was measured at 0 (n=4), 6 (n=3), 12 (n=4), 24 (n=8) and 48 h (n=6) after administration of CCl4. (B) Liver paraffin sections (10 μm thickness) of EXTL2+/+ and EXTL2−/− mice were digested with or without HSase and Hepase, and then incubated with horseradish peroxidase-conjugated anti-HGF antibody. (C) Phosphorylation of c-Met in EXTL2+/+ (open bar) and EXTL2−/− mouse liver (closed bar) was examined at 12 or 24 h after administration of CCl4. The asterisk shows a non-specific band. Data represent means±S.D. for three (12 h) or four (24 h) specimens. (D) Hepatocytes isolated from EXTL2+/+ or EXTL2−/− mouse liver were treated with recombinant HGF for 0, 10 or 30 min, and then phosphorylation of c-Met was analysed on immunoblots. Similar results were obtained from two independent experiments. Non-specific bands are indicated by an asterisk. (E) Hepatocytes isolated from EXTL2+/+ or EXTL2−/− mouse liver were digested for 4 h with (+) or without (−) CSase ABC and/or double-digested with HSase and Hepase and then incubated in the presence (+) or absence (−) of 50 ng/ml HGF for 10 min. Non-specific bands are indicated by the asterisk. Similar results were obtained in two independent experiments. (F) Hepatocytes isolated from EXTL2+/+ or EXTL2−/− mouse liver were treated with (+) or without (−) 100 ng/ml HGF for 60 min, and treated with (+) or without (−) 2.5 mM CCl4 for 4 h in the presence (+) or absence (−) of HGF. Data represent means±S.D. for three samples, and similar results were obtained in three independent experiments.

Figure 3
Comparison between EXTL2+/+ and EXTL2−/− mice with regard to HGF signalling in liver and hepatocytes

(A) The expression level of liver HGF mRNA and the concentration of serum HGF in EXTL2+/+ (○) and EXTL2−/− mice (●) was measured at 0 (n=4), 6 (n=3), 12 (n=4), 24 (n=8) and 48 h (n=6) after administration of CCl4. (B) Liver paraffin sections (10 μm thickness) of EXTL2+/+ and EXTL2−/− mice were digested with or without HSase and Hepase, and then incubated with horseradish peroxidase-conjugated anti-HGF antibody. (C) Phosphorylation of c-Met in EXTL2+/+ (open bar) and EXTL2−/− mouse liver (closed bar) was examined at 12 or 24 h after administration of CCl4. The asterisk shows a non-specific band. Data represent means±S.D. for three (12 h) or four (24 h) specimens. (D) Hepatocytes isolated from EXTL2+/+ or EXTL2−/− mouse liver were treated with recombinant HGF for 0, 10 or 30 min, and then phosphorylation of c-Met was analysed on immunoblots. Similar results were obtained from two independent experiments. Non-specific bands are indicated by an asterisk. (E) Hepatocytes isolated from EXTL2+/+ or EXTL2−/− mouse liver were digested for 4 h with (+) or without (−) CSase ABC and/or double-digested with HSase and Hepase and then incubated in the presence (+) or absence (−) of 50 ng/ml HGF for 10 min. Non-specific bands are indicated by the asterisk. Similar results were obtained in two independent experiments. (F) Hepatocytes isolated from EXTL2+/+ or EXTL2−/− mouse liver were treated with (+) or without (−) 100 ng/ml HGF for 60 min, and treated with (+) or without (−) 2.5 mM CCl4 for 4 h in the presence (+) or absence (−) of HGF. Data represent means±S.D. for three samples, and similar results were obtained in three independent experiments.

We demonstrated higher amounts of HS and CS in EXTL2−/− hepatocytes than EXTL2+/+ hepatocytes, and some differences in disaccharide composition of HS and CS chains between EXTL2+/+ and EXTL2−/− hepatocytes (Supplementary Table S2 at http://www.biochemj.org/bj/454/bj4540133add.htm). Thus we investigated whether the depressed HGF signalling in EXTL2−/− hepatocytes was associated with an increase in the amount of HS and of CS. Pre-treatment with Hepase and HSase decreases the response of EXTL2+/+ hepatocytes to HGF (Figure 3E); this finding indicated that adequate amounts of HS were required for HGF signalling. The responses of EXTL2+/+ hepatocytes and of EXTL2−/− hepatocytes to HGF were similar after treatment with Hepase and HSase (Figure 3E). In addition, CSase ABC had no effect on the responses of EXTL2+/+ hepatocytes or of EXTL2−/− hepatocytes to HGF. The results of the present study indicated that differences in the response to HGF between EXTL2+/+ and EXTL2−/− hepatocytes were due to the amount of HS. We next compared cell viability in CCl4-treated EXTL2+/+ hepatocytes with that in CCl4-treated EXTL2−/− hepatocytes in the absence or presence of HGF (Figure 3F) because HGF signalling is involved in cell survival, as well as cell proliferation [20,21]. In the absence of HGF, viability of EXTL2+/+ hepatocytes and of EXTL2−/− hepatocytes decreased to the same extent following treatment with CCl4. Pre-treatment with HGF inhibited CCl4-induced cell death of EXTL2+/+ hepatocytes; in contrast, pre-treatment with HGF did not protect EXTL2−/− hepatocytes from the effects of CCl4 (Figure 3F). The results indicated that HGF signalling in the EXTL2−/− hepatocytes was attenuated relative to that in EXTL2+/+ hepatocytes.

We next checked the expression levels of GAGs and some PGs in the liver untreated with CCl4 in order to examine how the expression of GAGs changed and what kinds of PGs were responsible for depressed HGF signalling in EXTL2−/− mice. SDC2 and SDC4, major HSPGs in the liver, were expressed at the surface and in the pericellular matrix of hepatocytes (Figures 4m–4p). We could not examine the expression of other liver PGs such as SDC1 and glypicans, because no antibodies suitable for immunohistochemistry were commercially available. The expression levels of HepSS-1-reactive HS chains (Figures 4a and 4b), NAH46-reactive HS chains (Figures 4c and 4d), JM403-reactive HS chains (Figures 4e and 4f), F58-3G10-reactive HS chains (Figures 4g and 4h), 2-B-6-reactive CS chains (Figures 4i and 4j) and CS-56-reactive CS chains (Figures 4k and 4l) were higher in EXTL2−/− mice than in EXTL2+/+ mice. These GAGs were expressed in the perisinusoidal space (also known as the space of Disse) and on the surface of hepatocytes in both EXTL2+/+ and EXTL2−/− mice (Figure 4). Reportedly, HepSS-1, NAH46 and JM403 recognize any HS domain with a high content of N-sulfated glucosamine, GlcNAc and N-unsubstituted glucosamine residues respectively [23,24], whereas CS-56 binds to 6-O-sulfated GalNAc of CS chains [25]. Monoclonal antibody F58-3G10 and 2-B-6 recognizes HS and CS structures that are newly generated after digestion with HSase and CSase ABC respectively. The results of the present study are summarized in Table 1. In Supplementary Figure S1 (at http://www.biochemj.org/bj/454/bj4540133add.htm), the expression levels of major hepatic HSPGs and CSPGs were investigated at the mRNA and protein levels. The gene expression level of SDCs 1–4, BGN (biglycan) and DCN (decorin) were higher in EXTL2−/− mice than in EXTL2+/+ mice (Supplementary Figure S1A). In addition, the expression of BGN, DCN and VCAN (versican) were elevated in the response to CCl4-induced liver injury (Supplementary Figure S1A). Consistent with these results, the expression levels of SDC2, SDC4 and BGN core proteins increased in EXTL2−/− mice compared with those in EXTL2+/+ mice (Supplementary Figure S1B). Furthermore, we analysed the length of HS and CS chains in EXTL2+/+ and EXTL2−/− liver under normal conditions (Supplementary Figure S2B at http://www.biochemj.org/bj/454/bj4540133add.htm). The length of HS chains in EXTL2+/+ and EXTL2−/− liver was similar, but some CS chains in EXTL2−/− liver were slightly longer than those in EXTL2+/+ liver. The results of the present study suggested that the response of hepatocytes to HGF was affected by the altered synthesis of GAGs and PGs that was caused by the loss of EXTL2.

Table 1
Summary of the results obtained by immunohistochemistry analyses

Intensity of immunohistochemistry staining was defined as negative (−), weak (+), moderate (++) and strong (+++). Supplementary Figure S3 is available at http://www.biochemj.org/bj/454/bj4540133add.htm.

   Immunoreactivity 
Antibody (source) Figure Time after administration of CCl4 (h) EXTL2+/+ EXTL2−/− 
HepSS-1 (Seikagaku Biobusiness) Figure 4 and Supplementary Figure S3 ++ 
 Supplementary Figure S3 24 +++ 
 Figure 5(B) and Supplementary Figure S3 48 ++ +++ 
NAH46 (Seikagaku Biobusiness) Figure 4 and Supplementary Figure S3 ++ +++ 
 Supplementary Figure S3 24 +++ 
 Figure 5(B) and Supplementary Figure S3 48 ++ +++ 
JM403 (Seikagaku Biobusiness) Figure 4 and Supplementary Figure S3 − 
 Supplementary Figure S3 24 ++ 
 Figure 5(B) and Supplementary Figure S3 48 ++ +++ 
F58-3G10 (Seikagaku Biobusiness) Figure 4 ++ +++ 
 Figure 5(B) 48 ++ 
2-B-6 (Seikagaku Biobusiness) Figure 4 ++ +++ 
 Figure 5(B) 48 +++ 
CS-56 (Sigma–Aldrich) Figure 4 ++ 
 Figure 5(B) 48 ++ 
Mouse anti-α-SMA monoclonal antibody (1A4) (Abcam) Figure 7 − − 
  48 
  72 +++ 
  120 ++ +++ 
   Immunoreactivity 
Antibody (source) Figure Time after administration of CCl4 (h) EXTL2+/+ EXTL2−/− 
HepSS-1 (Seikagaku Biobusiness) Figure 4 and Supplementary Figure S3 ++ 
 Supplementary Figure S3 24 +++ 
 Figure 5(B) and Supplementary Figure S3 48 ++ +++ 
NAH46 (Seikagaku Biobusiness) Figure 4 and Supplementary Figure S3 ++ +++ 
 Supplementary Figure S3 24 +++ 
 Figure 5(B) and Supplementary Figure S3 48 ++ +++ 
JM403 (Seikagaku Biobusiness) Figure 4 and Supplementary Figure S3 − 
 Supplementary Figure S3 24 ++ 
 Figure 5(B) and Supplementary Figure S3 48 ++ +++ 
F58-3G10 (Seikagaku Biobusiness) Figure 4 ++ +++ 
 Figure 5(B) 48 ++ 
2-B-6 (Seikagaku Biobusiness) Figure 4 ++ +++ 
 Figure 5(B) 48 +++ 
CS-56 (Sigma–Aldrich) Figure 4 ++ 
 Figure 5(B) 48 ++ 
Mouse anti-α-SMA monoclonal antibody (1A4) (Abcam) Figure 7 − − 
  48 
  72 +++ 
  120 ++ +++ 

Comparison between EXTL2+/+ and EXTL2−/− mice with regard to the expression and localization of GAGs and SDCs

Figure 4
Comparison between EXTL2+/+ and EXTL2−/− mice with regard to the expression and localization of GAGs and SDCs

Liver sections (Vibratome, 30 μm/slice) before administration of CCl4 were stained with various antibodies against GAGs. HepSS-1 (a and b), NAH46 (c and d) and JM403 (e and f) recognize HS structures, -[GlcUAβ1-4GlcNSα1-4]n-, -[GlcUAβ1-4GlcNAcα1-4]n- and -[GlcUAβ1-4GlcNH2α1-4]n- respectively. F58-3G10 (g and h) recognizes HS neo-epitope generated by digesting HS with heparitinese. 2-B-6 (i and j) recognizes 4-O-sulfated CS neo-epitope generated by digesting CS with CSase ABC; CS-56 (k and l) recognizes CS structures, -[GlcUAβ1-3GalNAc(6-O-sulfate)β1-4]n-; SDC2 (m and n), SDC4 (o and p).

Figure 4
Comparison between EXTL2+/+ and EXTL2−/− mice with regard to the expression and localization of GAGs and SDCs

Liver sections (Vibratome, 30 μm/slice) before administration of CCl4 were stained with various antibodies against GAGs. HepSS-1 (a and b), NAH46 (c and d) and JM403 (e and f) recognize HS structures, -[GlcUAβ1-4GlcNSα1-4]n-, -[GlcUAβ1-4GlcNAcα1-4]n- and -[GlcUAβ1-4GlcNH2α1-4]n- respectively. F58-3G10 (g and h) recognizes HS neo-epitope generated by digesting HS with heparitinese. 2-B-6 (i and j) recognizes 4-O-sulfated CS neo-epitope generated by digesting CS with CSase ABC; CS-56 (k and l) recognizes CS structures, -[GlcUAβ1-3GalNAc(6-O-sulfate)β1-4]n-; SDC2 (m and n), SDC4 (o and p).

We next measured the change in the amount of GAGs in EXTL2+/+ and EXTL2−/− mouse liver during liver injury and regeneration processes (Figure 5A); both HS and CS transiently increased within 24 h after CCl4 administration, then HS decreased, and markedly elevated CS levels largely persisted. The results indicated that EXTL2 functions to suppress an excessive increase in GAG biosynthesis during liver regeneration. At 0, 24, 48 and 72 h after CCl4 administration, levels of both HS and CS were significantly higher in the livers of EXTL2−/− mice than in those of EXTL2+/+ mice (Figure 5A). These results indicated that lack of EXTL2 caused the accumulation of excess GAGs in EXTL2−/− mouse liver following liver injury and during the ensuing regeneration. In addition, we analysed the disaccharide composition of HS and CS (Supplementary Figure S2A). At 0 h after administration of CCl4, the disaccharide composition of both HS and CS was similar between the EXTL2+/+ and EXTL2−/− livers. At 24, 48 and 72 h after administration of CCl4, ΔDiHS-NS slightly increased in EXTL2−/− mice compared with EXTL2+/+ mice (Supplementary Figure S2A). In contrast, disaccharide composition of CS was indistinguishable between EXTL2+/+ and EXTL2−/− liver (Supplementary Figure S2A). We next examined what types of cells produced GAGs 48 h after administration of CCl4 (Figures 5B and 5C). F69-3G10- and HepSS-1-reactive HS chains and 2-B-6- and CS-56-reactive CS chains were localized in the perisinusoidal space, and were more highly expressed in EXTL2−/− mice than in EXTL2+/+ mice (Figure 5B). Furthermore, we confirmed the staining patterns of HepSS-1 and CS-56 merged with those of GFAP, known as a marker of stellate cells (Figure 5C). These data indicated that GAG biosynthesis in EXTL2−/− hepatic stellate cells is up-regulated during liver injury and regeneration processes. Furthermore, we examined the expression of specific HS epitopes in the liver using HepSS-1, NAH46 and JM403, which recognize HS with a high content of N-sulfated glucosamines, N-acetylated glucosamines and N-unsubstituted glucosamines respectively (Supplementary Figure S3 at http://www.biochemj.org/bj/454/bj4540133add.htm). All of these HS epitopes were highly expressed in EXTL2−/− liver than in EXTL2+/+ liver, especially following liver injury and during regeneration. The results from the present study are summarized in Table 1.

Comparison between GAGs produced in EXTL2+/+ and EXTL2−/− mice

Figure 5
Comparison between GAGs produced in EXTL2+/+ and EXTL2−/− mice

(A) GAGs were isolated from liver samples taken from EXTL2+/+ (○) or EXTL2−/− (●) mice at 0, 6, 12, 24, 48 or 72 h after administration of CCl4; these GAGs were then subjected to disaccharide composition analysis (see Experimental section). The experiment was performed in triplicate. The error bars represent means±S.D. (B) Liver sections (Vibratome, 30 μm/slice) cut from samples collected at 48 h after CCl4 administration were stained with antibodies against various GAGs. Magnified images are shown in the inset. (C) Liver sections (Vibratome, 30 μm/slice) cut from samples collected at 48 h after CCl4 administration were stained with antibodies against various GAGs and GFAP.

Figure 5
Comparison between GAGs produced in EXTL2+/+ and EXTL2−/− mice

(A) GAGs were isolated from liver samples taken from EXTL2+/+ (○) or EXTL2−/− (●) mice at 0, 6, 12, 24, 48 or 72 h after administration of CCl4; these GAGs were then subjected to disaccharide composition analysis (see Experimental section). The experiment was performed in triplicate. The error bars represent means±S.D. (B) Liver sections (Vibratome, 30 μm/slice) cut from samples collected at 48 h after CCl4 administration were stained with antibodies against various GAGs. Magnified images are shown in the inset. (C) Liver sections (Vibratome, 30 μm/slice) cut from samples collected at 48 h after CCl4 administration were stained with antibodies against various GAGs and GFAP.

In addition, we checked mRNA expression levels of genes in the EXT family, of some genes encoding CS biosynthetic enzymes and of FAM20 following liver injury and during the ensuing regeneration process (Supplementary Figure S4A at http://www.biochemj.org/bj/454/bj4540133add.htm). However, we did not observe a sharp change in the expression of any EXT mRNA, CS biosynthetic enzyme mRNA or FAM20 mRNA. Furthermore, we monitored the change in the expression level of EXTL2 products as described recently [11]; specifically, we measured accumulation of truncated pentasaccharides [GlcNAc-GlcUA-Gal-Gal-Xyl(2-O-phosphate)]. As shown in Supplementary Figure S4(B), steady-state levels of EXTL2 products decreased within 24 h after CCl4 administration, and then increased to basal levels by 48 h after CCl4 administration. Therefore these data indicated that EXTL2 arrests the synthesis of GAGs to prevent excessive production of GAGs in hepatic stellate cells.

We investigated further whether non-parenchymal cells could be affected by loss of EXTL2−/−. The main non-parenchymal cells in the liver are Kupffer cells, sinusoidal endothelial cells and stellate cells. In the present study, we examined Kupffer cells and stellate cells because they are closely associated with hepatocellular function and the liver's response to injury. Liver Kupffer cells can be detected using anti-CD68 antibody because they are macrophage lineage cells. As shown in Supplementary Figure S5(A) (at http://www.biochemj.org/bj/454/bj4540133add.htm), Kupffer cells line the walls of the sinusoids both before and after administration of CCl4. The localization pattern of Kupffer cells in EXTL2−/− mice was similar to that in EXTL2+/+ mice (Supplementary Figure S5A). In addition to non-parenchymal cells, oval cells (known as hepatic progenitor cells in rodents) were also examined. Although hepatocytes are the functional stem cells of the liver [26], hepatic progenitor cells such as oval cells are activated when the proliferative activity of mature hepatocytes is impaired [27]. Oval cells express phenotypical markers of both the biliary epithelium (cytokeratins CK7 and CK19) and the hepatocyte lineages (α-fetoprotein and albumin). We measured the expression level of CK19 to detect oval cells. The number of oval cells had increased 2-fold at 24 h after administration of CCl4 in EXTL2−/− mice, as well as EXTL2+/+ mice (Supplementary Figure S5B).

Hepatic stellate cells undergo a gradual transition from a quiescent cell into an activated, proliferative, contractile and fibrogenic myofibroblast-like cell during liver injury and regeneration process. GFAP is expressed in quiescent stellate cells in vivo, and shows increased expression in the acute response to injury in rats and decreased expression in the chronic phase; thus GFAP is used as an early marker of stellate cell activation [28]. On the other hand, α-SMA is a well-known marker of hepatic stellate cell activation [29]. We used expression of GFAP and of α-SMA as indicators to compare EXTL2+/+ hepatic stellate cells with EXTL2−/− hepatic stellate cells. The number of GFAP-positive cells in EXTL2+/+ mice peaked 24 h after administration of CCl4, but the number of GFAP-positive cells increased more gradually in EXTL2−/− mice than in EXTL2+/+ mice (Figures 6A and 6B). A delay in the increase in the number of GFAP-positive cells in EXTL2−/− mice was not attributed to decreased proliferation because there was little difference in the number of BrdU-incorporated GFAP-positive cells between EXTL2+/+ mice and EXTL2−/− mice (Figure 6C). A previous study indicated that hepatic stellate cells originate from haemopoietic stem cells [30]. Thus differentiation from stem cells into stellate cells might be affected by loss of EXTL2. Activated stellate cells were visualized by an anti-α-SMA antibody (Figure 7 and Table 1). Within 48 h after administration of CCl4, α-SMA-positive cells had appeared in EXTL2+/+ mice. Within 72 h after administration of CCl4, hepatic stellate cells were maximally activated in EXTL2+/+ mice, but most stellate cells were α-SMA-negative in EXTL2−/− mice. At 120 h after administration of CCl4, hepatic stellate cells in EXTL2−/− mice were activated. The results of the present study indicated that the activation of hepatic stellate cells was affected by loss of EXTL2. Changes in the nature of a hepatocyte pericellular matrix caused by the loss of EXTL2 impaired HGF signal-mediated proliferation of hepatocytes, caused severe liver injury and consequently affected the differentiation and activation of hepatic stellate cells. Reportedly, impaired activation of stellate cells in Foxf1+/− (forkhead box f1) mice after CCl4 injury was associated with increased liver cell injury and severe hepatic apoptosis [31]. In contrast, hepatocytes failed to regenerate properly after liver injury in Col-1a1r/r mice, where stellate cells are persistently activated [32]. Activated stellate cells biosynthesize extracellular matrix components containing collagens after liver injury, which are essential for the liver repair process. However, inadequate activation of stellate cells causes overproduction of extracellular matrices, and links to the pathogenesis of fibrosis. Therefore it was investigated whether altered activation of stellate cells after liver injury in EXTL2−/− mice was involved in the development of liver fibrosis. At 5 days after administration of CCl4, collagens were focally, but substantially, accumulated in EXTL2−/− mice (Supplementary Figure S6B at http://www.biochemj.org/bj/454/bj4540133add.htm), although bridging fibrosis had not occurred yet (Supplementary Figure S6A). In addition, an excessive number of ducts were observed in portal tracts of EXTL2−/− mice (Supplementary Figure S6A, white arrowheads). These ducts were identified not as bile ducts, but hepatic arteries, because they were labelled with anti-α-SMA antibody (Supplementary Figure S6C). Reportedly, an excess of arteries was associated with a primary defect in intrahepatic bile duct development [33]. Accordingly, data from the present study indicate that severe liver injury caused by impairment of hepatocyte proliferation also affects biliary epithelial cells in EXTL2−/− mice.

Comparison between EXTL2+/+ and EXTL2−/− mice with regard to GFAP-positive hepatic stellate cells

Figure 6
Comparison between EXTL2+/+ and EXTL2−/− mice with regard to GFAP-positive hepatic stellate cells

(A) Liver sections (paraffin, 5 μm/slice) cut from tissue samples collected at 0, 24, 48, 72 or 96 h after CCl4 administration were stained with anti-GFAP antibody. Magnified images of GFAP-positive cells are shown in the inset. (B) GFAP-positive cells were counted. Data are expressed as the means±S.D. for two specimens. (C) Liver sections (paraffin, 10 μm) cut from tissue samples collected at 48 h after CCl4 administration were stained with anti-BrdU and anti-GFAP antibodies. White arrows and white arrowheads indicate BrdU-incorporated hepatic non-parenchymal cells and hepatocytes respectively. (D) GFAP-positive, BrdU-positive, and GFAP- and BrdU-double positive cells were counted from four to six fields per section, and data are expressed as means±S.D. for two specimens.

Figure 6
Comparison between EXTL2+/+ and EXTL2−/− mice with regard to GFAP-positive hepatic stellate cells

(A) Liver sections (paraffin, 5 μm/slice) cut from tissue samples collected at 0, 24, 48, 72 or 96 h after CCl4 administration were stained with anti-GFAP antibody. Magnified images of GFAP-positive cells are shown in the inset. (B) GFAP-positive cells were counted. Data are expressed as the means±S.D. for two specimens. (C) Liver sections (paraffin, 10 μm) cut from tissue samples collected at 48 h after CCl4 administration were stained with anti-BrdU and anti-GFAP antibodies. White arrows and white arrowheads indicate BrdU-incorporated hepatic non-parenchymal cells and hepatocytes respectively. (D) GFAP-positive, BrdU-positive, and GFAP- and BrdU-double positive cells were counted from four to six fields per section, and data are expressed as means±S.D. for two specimens.

Comparison between EXTL2+/+ and EXTL2−/− mice with regard to α-SMA-positive hepatic stellate cells

Figure 7
Comparison between EXTL2+/+ and EXTL2−/− mice with regard to α-SMA-positive hepatic stellate cells

Liver sections (paraffin, 5 μm/slice) cut from tissue samples taken at 0, 48, 72 or 120 h after CCl4 administration were stained with anti-α-SMA antibody. Green signals show the activated stellate cells.

Figure 7
Comparison between EXTL2+/+ and EXTL2−/− mice with regard to α-SMA-positive hepatic stellate cells

Liver sections (paraffin, 5 μm/slice) cut from tissue samples taken at 0, 48, 72 or 120 h after CCl4 administration were stained with anti-α-SMA antibody. Green signals show the activated stellate cells.

DISCUSSION

In the present study, we showed that abnormal biosynthesis of GAGs which was caused by lack of EXTL2 affected the liver injury and regeneration. As shown in Figure 8, EXTL2+/+ hepatocytes proliferated in response to the HGF produced by Kupffer cells and by endothelial cells after CCl4 administration [34]; the activated stellate cells then synthesized extracellular matrix components containing collagens and PGs to promote matrix remodelling, and, consequently, the liver regenerated normally. In contrast, proliferation of EXTL2−/− hepatocytes in the response to HGF was suppressed, and EXTL2−/− hepatocytes were more vulnerable to CCl4 than EXTL2+/+ hepatocytes. As a result, liver regeneration was impaired in EXTL2−/− mice, and collagen-rich scar matrix was accumulated in foci. Thus we propose that suppressed proliferation of EXTL2−/− hepatocytes is likely to be the primary cause for impaired liver regeneration in EXTL2−/− mice, although we cannot completely exclude the possibility that loss of EXTL2 directly causes changes in the nature of non-parenchymal cells such as Kupffer cells and stellate cells. To elucidate the reason for this, we examined whether HGF signalling was altered in EXTL2−/− liver tissue. HGF is the most potent hepatotrophic factor in liver generation, and its expression increases in response to liver injury. Neutralization of HGF results in impairment and retardation of liver regeneration. Reportedly, HGF binds tightly to HS chains [22], and HS functions as a co-receptor of c-Met [35]. Furthermore, the presence of heparin reportedly increases the potency of HGF, causing elevated downstream signalling effects [35]. However, in spite of the fact that the expression level of HS produced in EXTL2−/− hepatocytes was higher than that produced in EXTL2+/+ hepatocytes, HGF molecules were focally accumulated in EXTL2+/+ liver at 24 h after administration of CCl4, but dispersed throughout EXTL2−/− liver (Figure 3B). HGF molecules might be caught by excess GAGs in EXTL2−/− liver and thus could not be concentrated locally. The response of EXTL2−/− hepatocytes to HGF was attenuated compared with the response of EXTL2+/+ hepatocytes (Figure 3C). The findings of the present study could be explained by inhibition of presentation of HGF molecules to c-Met receptors due to sequestration of HGF molecules by excess HS chains. For example, mice that overexpress agrin, an HSPG, reportedly exhibit disrupted eye development [36]. Overexpression of agrin localized to the extracellular matrix causes a decrease in sonic hedgehog levels in the developing optic nerve [36] because HS chains can bind to sonic hedgehog [37] and retain it outside of optic neurons that express the cognate receptor. Consequently, overexpression of agrin causes down-regulation of the sonic hedgehog pathway in the optic nerve and expansion of PAX6 repressed by the sonic hedgehog pathway. These results suggest that excess HS chains depress the sonic hedgehog signalling pathway. Likewise, we suggest that increased HS chains on the surface of EXTL2−/− hepatocytes reduce the number of HGF molecules bound to c-Met and that this reduction leads to suppression of HGF signalling. Another paper reported that HGF mutants with reduced affinity for HS had enhanced signalling activity in vivo [38]. The authors indicated that one role of HSPGs is sequestration, internalization and degradation of HGF molecules. In addition, functional domain structures with characteristic sulfation patterns may be orderly located throughout an HS chain and function in interacting with bioactive proteins [39]. Reportedly, the minimal HS structures necessary for HGF binding are octasaccharides containing two IdoUA(2-O-sulfate)-GlcN(N-,6-O-sulfate) units [40]. Disaccharide composition analysis of HS chains from mouse embryonic fibroblasts [11], and from mouse cultured hepatocytes (Supplementary Table S2) demonstrated that sulfation patterns of HS were altered by loss of EXTL2. In addition, immunofluorescence analysis showed that EXTL2 deficiency affected the expression levels of specific structures of HS (Figures 4 and 5, and Supplementary Figure S3). Taken together, the HS domain structures in EXTL2−/− hepatocytes might be affected by loss of EXTL2. Thus abnormal sulfation patterns of HS chains might cause an attenuated HGF response in EXTL2−/− hepatocytes. Similarly, EXTL2 deficiency greatly affected sulfation patterns of CS chains in mouse embryonic fibroblasts [11] and in mouse cultured hepatocytes (Supplementary Table S2); the amount of ΔDi-4S markedly increased and that of ΔDi-diSE decreased. In fact, we showed that Wnt3a binds to a specific structure containing ΔDi-diSE units with high affinity [12], although, in the present study, the involvement of CS containing ΔDi-diSE units in liver injury and regeneration was not demonstrated.

Impaired liver regeneration processes in EXTL2−/− mice suffered from liver injury by CCl4

Figure 8
Impaired liver regeneration processes in EXTL2−/− mice suffered from liver injury by CCl4

After administration of CCl4, EXTL2+/+ hepatocytes propagated in response to the HGF produced by Kupffer cells and endothelial cells. The activated stellate cells synthesized extracellular matrix components containing collagens and PGs to promote matrix remodelling, and consequently, the liver regenerated normally. In contrast, proliferation and survival of EXTL2−/− hepatocytes was suppressed because the abnormal GAG biosynthesis that was caused by loss of EXTL2 attenuates HGF signalling. As a result, liver regeneration was impaired, and collagen-rich scar matrix was accumulated in foci.

Figure 8
Impaired liver regeneration processes in EXTL2−/− mice suffered from liver injury by CCl4

After administration of CCl4, EXTL2+/+ hepatocytes propagated in response to the HGF produced by Kupffer cells and endothelial cells. The activated stellate cells synthesized extracellular matrix components containing collagens and PGs to promote matrix remodelling, and consequently, the liver regenerated normally. In contrast, proliferation and survival of EXTL2−/− hepatocytes was suppressed because the abnormal GAG biosynthesis that was caused by loss of EXTL2 attenuates HGF signalling. As a result, liver regeneration was impaired, and collagen-rich scar matrix was accumulated in foci.

In the CCl4-induced injury model of mouse liver damage, HGF/c-Met signalling plays an important role in cell survival and proliferation. Metfl/fl;Alb-Cre+/− conditional knockout mice are unable to maintain an adequate level of hepatocyte proliferation [41], and these mice show increased sensitivity to hepatocyte apoptosis during liver injury [20]; however, Metfl/fl;Alb-Cre+/− conditional knockout mice had no apparent histological or physiological abnormalities under normal conditions. Thus depressed HGF/c-Met signalling impairs recovery from damage when the liver is injured, and accelerates development of liver fibrosis [41]. Like the Metfl/fl;Alb-Cre+/− conditional knockout mice, EXTL2−/− mice grew normally, were fertile and showed no apparent phenotypes under normal physiological conditions. Nonetheless, the adaptive responses of EXTL2−/− mice following liver injury were significantly diminished relative to those of wild-type littermates. The activation level of HGF/c-Met signalling was depressed, hepatocyte proliferation was delayed and hepatocyte cell death was enhanced (Figure 3). Because EXTL2−/− mice experienced more severe liver injury than EXTL2+/+ mice, the ratio of liver weight to body weight was significantly lower in mutants (Figures 1A–1C). In addition, lack of EXTL2 affected hepatic stellate cells; specifically, the ability of these cells to synthesize GAGs was up-regulated during regeneration processes (Figure 5). This up-regulation may have led to the presence of massive inflammatory lymphocytic infiltrates (Figure 1E) and to focal collagen depositions (Supplementary Figure S6B). Hepatocellular injury usually leads to inflammation and activation of the innate immune system, causing release of growth factor and cytokines that can stimulate extracellular matrix synthesis [42,43]. In the absence of EXTL2, up-regulated GAG synthesis cannot be suppressed, and consequently, excess GAGs may be produced, thus causing facilitation of fibrogenesis. As shown in Figure 5(A), although CS remarkably increased during regeneration, CS had no effects on HGF signalling (Figure 3D). It is possible that the prominent increase in CS during regeneration might be implicated in fibrosis. Abnormal CS biosynthesis, owing to the lack of EXTL2, could probably greatly affect the pathological condition of EXTL2−/− mice in chronic phase. Therefore it is likely that reduced EXTL2 expression is one determinant that could convert a normally acute injury into a chronic lesion.

Although we focused on HGF signalling in the present study, other signalling pathways regulated by PGs such as hedgehog and EGFR (EGF receptor) signalling pathways, are also important for liver regeneration. It was reported that healthy livers express low levels of hedgehog ligands and that the ligand production can be stimulated by growth factors/cytokines released during liver injury. Hedgehog ligands act on various types of liver cells and control cell fate decisions. For example, hedgehog ligands stimulate hepatic stellate cells and consequently quiescent hepatic stellate cells trans-differentiate into activated hepatic stellate cells [44]. As shown in Figure 7, the activation of hepatic stellate cells was delayed in EXTL2−/− mice during liver regeneration. Thus we cannot rule out the possibility that the activation of stellate cells was affected by hedgehog pathway impairment due to loss of EXTL2. In addition, the EGFR pathway regulates downstream targets, activating cell cycle genes such as CCND1, and thus controls the entry of the hepatocytes into the cell cycle as well as the HGF pathway [45]. We found that the expression of CCND1 was suppressed in EXTL2−/− mice during liver regeneration compared with EXTL2+/+ mice (Figure 2C). Therefore the EGFR signalling pathway might also be affected by the abnormal GAG biosynthesis caused by loss of EXTL2.

To date, liver HSPGs are reported to regulate lipid metabolism directly and indirectly. SDC1 is the major HSPG in the liver, and works in parallel to, but independently of, the low-density lipoprotein receptor as a receptor for both intestinally derived and hepatic lipoprotein particles [46]. In addition, collagen XVIII, a type of HSPG, presents in the basement membranes underlying endothelial cells, and affects the distribution of lipoprotein lipase and triglyceride metabolism [47]. Furthermore, it has been known that lipid metabolism has an impact on liver regeneration. Although administration of CCl4 cannot induce steatohepatitis [48], mice fed on a high-fat diet and administered CCl4 develop steatohepatitis [49]. The finding implies that aberrant lipid metabolism affects pathological features caused by CCl4. In EXTL2−/− mice, the synthesis of liver HSPGs was affected (Figure 4 and Supplementary Figure S3). Thus there is the possibility that lipid metabolism in EXTL2−/− mice is altered. That is, aberrant lipid metabolism might connect to severe phenotype in EXTL2−/− mice. NASH (non-alcoholic steatohepatitis) is known as a common liver disease. NASH is supposed to develop in a ‘two-hit’ mechanism; first hit, hepatic steatosis (abnormal lipid metabolism); second hit, enhanced lipid peroxidation and increased generation of reactive oxygen species (oxidative stress). If lipid metabolism was affected by loss of EXTL2, progression of NASH might be accelerated in EXTL2−/− mice. Therefore future studies would be needed to clarify the linking of EXTL2 to human diseases.

Abbreviations

     
  • ALT

    alanine aminotranferase

  •  
  • AST

    aspartate aminotransferase

  •  
  • BGN

    biglycan

  •  
  • BrdU

    bromodeoxyuridine

  •  
  • CCl4

    carbon tetrachloride

  •  
  • CCND1

    cyclin D1

  •  
  • CK

    cytokeratin

  •  
  • CS

    chondroitin sulfate

  •  
  • CSase ABC

    chondroitinase ABC

  •  
  • DCN

    decorin

  •  
  • EGF

    epidermal growth factor

  •  
  • EGFR

    EGF receptor

  •  
  • EXT

    exostosin

  •  
  • EXTL

    EXT-like

  •  
  • GAG

    glycosaminoglycan

  •  
  • GalNAc

    N-acetylgalactosamine

  •  
  • GFAP

    glial fibrillary acidic protein

  •  
  • GlcNAc

    N-acetylglucosamine

  •  
  • GlcUA

    glucuronic acid

  •  
  • Hepase

    heparinase

  •  
  • HGF

    hepatocyte growth factor

  •  
  • HS

    heparan sulfate

  •  
  • HSase

    heparitinase

  •  
  • mAb

    monoclonal antibody

  •  
  • NASH

    non-alcoholic steatohepatitis

  •  
  • PBST

    PBS/0.1% Tween 20

  •  
  • PG

    proteoglycan

  •  
  • qPCR

    quantitative real-time PCR

  •  
  • SDC

    syndecan

  •  
  • α-SMA

    α-smooth muscle actin

AUTHOR CONTRIBUTION

Satomi Nadanaka performed most of the experiments; Satomi Nadanaka and Hiroshi Kitagawa designed the research, analysed the data and wrote the paper; Shoji Kagiyama contributed new reagents and assisted in some experiments.

FUNDING

This work was supported by the Ministry of Education, Culture, Sports, Science, and Technology (MEXT)-Supported Program for the Strategic Research Foundation at Private Universities, 2012–2017 (to H.K.), and was supported in part by Grant-in-Aids for Scientific Research on Innovative Areas [grant number 23110003 (to H.K.)] (Deciphering sugar chain-based signals regulating integrative neuronal functions), for Challenging Exploratory Research [grant number 24659039 (to H.K.)], and for Scientific Research (C) [grant number 25460080 (to S.N.)] from the Ministry of Education, Culture, Sports, Science, and Technology (MEXT) of Japan.

References

References
1
Sugahara
K.
Kitagawa
H.
Recent advances in the study of the biosynthesis and functions of sulfated glycosaminoglycans
Curr. Opin. Struct. Biol.
2000
, vol. 
10
 (pg. 
518
-
527
)
2
Bernfield
M.
Gotte
M.
Park
P. W.
Reizes
O.
Fitzgerald
M. L.
Lincecum
J.
Zako
M.
Functions of cell surface heparan sulfate proteoglycans
Annu. Rev. Biochem.
1999
, vol. 
68
 (pg. 
729
-
777
)
3
Salmivirta
M.
Lidholt
K.
Lindahl
U.
Heparan sulfate: a piece of information
FASEB J.
1996
, vol. 
10
 (pg. 
1270
-
1279
)
4
Wise
C. A.
Clines
G. A.
Massa
H.
Trask
B. J.
Lovett
M.
Identification and localization of the gene for EXTL, a third member of the multiple exostoses gene family
Genome Res.
1997
, vol. 
7
 (pg. 
10
-
16
)
5
Wuyts
W.
Van Hul
W.
Hendrickx
J.
Speleman
F.
Wauters
J.
De Boulle
K.
Van Roy
N.
Van Agtmael
T.
Bossuyt
P.
Willems
P. J.
Identification and characterization of a novel member of the EXT gene family, EXTL2
Eur. J. Hum. Genet.
1997
, vol. 
5
 (pg. 
382
-
389
)
6
Saito
T.
Seki
N.
Yamauchi
M.
Tsuji
S.
Hayashi
A.
Kozuma
S.
Hori
T.
Structure, chromosomal location, and expression profile of EXTR1 and EXTR2, new members of the multiple exostoses gene family
Biochem. Biophys. Res. Commun.
1998
, vol. 
243
 (pg. 
61
-
66
)
7
Lin
X.
Wei
G.
Shi
Z.
Dryer
L.
Esko
J. D.
Wells
D. E.
Matzuk
M. M.
Disruption of gastrulation and heparan sulfate biosynthesis in EXT1-deficient mice
Dev. Biol.
2000
, vol. 
224
 (pg. 
299
-
311
)
8
Stickens
D.
Zak
B. M.
Rougier
N.
Esko
J. D.
Werb
Z.
Mice deficient in Ext2 lack heparan sulfate and develop exostoses
Development
2005
, vol. 
132
 (pg. 
5055
-
5068
)
9
Takahashi
I.
Noguchi
N.
Nata
K.
Yamada
S.
Kaneiwa
T.
Mizumoto
S.
Ikeda
T.
Sugihara
K.
Asano
M.
Yoshikawa
T.
, et al. 
Important role of heparan sulfate in postnatal islet growth and insulin secretion
Biochem. Biophys. Res. Commun.
2009
, vol. 
383
 (pg. 
113
-
118
)
10
Kitagawa
H.
Shimakawa
H.
Sugahara
K.
The tumor suppressor EXT-like gene EXTL2 encodes an α1, 4-N-acetylhexosaminyltransferase that transfers N-acetylgalactosamine and N-acetylglucosamine to the common glycosaminoglycan-protein linkage region. The key enzyme for the chain initiation of heparan sulfate
J. Biol. Chem.
1999
, vol. 
274
 (pg. 
13933
-
13937
)
11
Nadanaka
S.
Zhou
S.
Kagiyama
S.
Shoji
N.
Sugahara
K.
Sugihara
K.
Asano
M.
Kitagawa
H.
EXTL2, a member of the EXT family of tumor suppressors, controls glycosaminoglycan biosynthesis in a xylose kinase-dependent manner
J. Biol. Chem.
2013
, vol. 
288
 (pg. 
9321
-
9333
)
12
Nadanaka
S.
Ishida
M.
Ikegami
M.
Kitagawa
H.
Chondroitin 4-O-sulfotransferase-1 modulates Wnt-3a signaling through control of E disaccharide expression of chondroitin sulfate
J. Biol. Chem.
2008
, vol. 
283
 (pg. 
27333
-
27343
)
13
Okada
M.
Nadanaka
S.
Shoji
N.
Tamura
J.
Kitagawa
H.
Biosynthesis of heparan sulfate in EXT1-deficient cells
Biochem. J.
2010
, vol. 
428
 (pg. 
463
-
471
)
14
Seglen
P. O.
Preparation of isolated rat liver cells
Methods Cell Biol.
1976
, vol. 
13
 (pg. 
29
-
83
)
15
Yoshida
H.
Matsui
T.
Yamamoto
A.
Okada
T.
Mori
K.
XBP1 mRNA is induced by ATF6 and spliced by IRE1 in response to ER stress to produce a highly active transcription factor
Cell
2001
, vol. 
107
 (pg. 
881
-
891
)
16
Okazaki
M.
Furuya
E.
Kasahara
T.
Sakamoto
K.
Function of reticuloendothelial system on CCl4 induced liver injury in mice
Jpn. J. Pharmacol.
1985
, vol. 
39
 (pg. 
503
-
514
)
17
Diaz Gomez
M. I.
Fanelli
S. L.
Delgado de Layno
A. M.
Castro
J. A.
Castro
G. D.
Liver nuclear and microsomal CYP2E1-mediated metabolism of xenobiotics in rats chronically drinking an alcohol-containing liquid diet
Toxicol. Ind. Health
2006
, vol. 
22
 (pg. 
367
-
374
)
18
Calabrese
E. J.
Baldwin
L. A.
Mehendale
H. M.
G2 subpopulation in rat liver induced into mitosis by low-level exposure to carbon tetrachloride: an adaptive response
Toxicol. Appl. Pharmacol.
1993
, vol. 
121
 (pg. 
1
-
7
)
19
Maher
J. J.
Cell-specific expression of hepatocyte growth factor in liver. Upregulation in sinusoidal endothelial cells after carbon tetrachloride
J. Clin. Invest.
1993
, vol. 
91
 (pg. 
2244
-
2252
)
20
Huh
C. G.
Factor
V. M.
Sanchez
A.
Uchida
K.
Conner
E. A.
Thorgeirsson
S. S.
Hepatocyte growth factor/c-met signaling pathway is required for efficient liver regeneration and repair
Proc. Natl. Acad. Sci. U.S.A.
2004
, vol. 
101
 (pg. 
4477
-
4482
)
21
Borowiak
M.
Garratt
A. N.
Wustefeld
T.
Strehle
M.
Trautwein
C.
Birchmeier
C.
Met provides essential signals for liver regeneration
Proc. Natl. Acad. Sci. U.S.A.
2004
, vol. 
101
 (pg. 
10608
-
10613
)
22
Lyon
M.
Deakin
J. A.
Mizuno
K.
Nakamura
T.
Gallagher
J. T.
Interaction of hepatocyte growth factor with heparan sulfate. Elucidation of the major heparan sulfate structural determinants
J. Biol. Chem.
1994
, vol. 
269
 (pg. 
11216
-
11223
)
23
van den Born
J.
Salmivirta
K.
Henttinen
T.
Ostman
N.
Ishimaru
T.
Miyaura
S.
Yoshida
K.
Salmivirta
M.
Novel heparan sulfate structures revealed by monoclonal antibodies
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
20516
-
20523
)
24
Suzuki
K.
Yamamoto
K.
Kariya
Y.
Maeda
H.
Ishimaru
T.
Miyaura
S.
Fujii
M.
Yusa
A.
Joo
E. J.
Kimata
K.
, et al. 
Generation and characterization of a series of monoclonal antibodies that specifically recognize [HexA(±2S)-GlcNAc]n epitopes in heparan sulfate
Glycoconjugate J.
2008
, vol. 
25
 (pg. 
703
-
712
)
25
Ito
Y.
Hikino
M.
Yajima
Y.
Mikami
T.
Sirko
S.
von Holst
A.
Faissner
A.
Fukui
S.
Sugahara
K.
Structural characterization of the epitopes of the monoclonal antibodies 473HD, CS-56, and MO-225 specific for chondroitin sulfate D-type using the oligosaccharide library
Glycobiology
2005
, vol. 
15
 (pg. 
593
-
603
)
26
Overturf
K.
Al-Dhalimy
M.
Finegold
M.
Grompe
M.
The repopulation potential of hepatocyte populations differing in size and prior mitotic expansion
Am. J. Pathol.
1999
, vol. 
155
 (pg. 
2135
-
2143
)
27
Dabeva
M. D.
Shafritz
D. A.
Hepatic stem cells and liver repopulation
Semin. Liver Dis.
2003
, vol. 
23
 (pg. 
349
-
362
)
28
Morini
S.
Carotti
S.
Carpino
G.
Franchitto
A.
Corradini
S. G.
Merli
M.
Gaudio
E.
GFAP expression in the liver as an early marker of stellate cells activation
Ital. J. Anat. Embryol.
2005
, vol. 
110
 (pg. 
193
-
207
)
29
Friedman
S. L.
Hepatic stellate cells: protean, multifunctional, and enigmatic cells of the liver
Physiol. Rev.
2008
, vol. 
88
 (pg. 
125
-
172
)
30
Miyata
E.
Masuya
M.
Yoshida
S.
Nakamura
S.
Kato
K.
Sugimoto
Y.
Shibasaki
T.
Yamamura
K.
Ohishi
K.
Nishii
K.
, et al. 
Hematopoietic origin of hepatic stellate cells in the adult liver
Blood
2008
, vol. 
111
 (pg. 
2427
-
2435
)
31
Kalinichenko
V. V.
Bhattacharyya
D.
Zhou
Y.
Gusarova
G. A.
Kim
W.
Shin
B.
Costa
R. H.
Foxf1+/− mice exhibit defective stellate cell activation and abnormal liver regeneration following CCl4 injury
Hepatology
2003
, vol. 
37
 (pg. 
107
-
117
)
32
Issa
R.
Zhou
X.
Trim
N.
Millward-Sadler
H.
Krane
S.
Benyon
C.
Iredale
J.
Mutation in collagen-1 that confers resistance to the action of collagenase results in failure of recovery from CCl4-induced liver fibrosis, persistence of activated hepatic stellate cells, and diminished hepatocyte regeneration
FASEB J.
2003
, vol. 
17
 (pg. 
47
-
49
)
33
Clotman
F.
Libbrecht
L.
Gresh
L.
Yaniv
M.
Roskams
T.
Rousseau
G. G.
Lemaigre
F. P.
Hepatic artery malformations associated with a primary defect in intrahepatic bile duct development
J. Hepatol.
2003
, vol. 
39
 (pg. 
686
-
692
)
34
Noji
S.
Tashiro
K.
Koyama
E.
Nohno
T.
Ohyama
K.
Taniguchi
S.
Nakamura
T.
Expression of hepatocyte growth factor gene in endothelial and Kupffer cells of damaged rat livers, as revealed by in situ hybridization
Biochem. Biophys. Res. Commun.
1990
, vol. 
173
 (pg. 
42
-
47
)
35
Rubin
J. S.
Day
R. M.
Breckenridge
D.
Atabey
N.
Taylor
W. G.
Stahl
S. J.
Wingfield
P. T.
Kaufman
J. D.
Schwall
R.
Bottaro
D. P.
Dissociation of heparan sulfate and receptor binding domains of hepatocyte growth factor reveals that heparan sulfate-c-met interaction facilitates signaling
J. Biol. Chem.
2001
, vol. 
276
 (pg. 
32977
-
32983
)
36
Fuerst
P. G.
Rauch
S. M.
Burgess
R. W.
Defects in eye development in transgenic mice overexpressing the heparan sulfate proteoglycan agrin
Dev. Biol.
2007
, vol. 
303
 (pg. 
165
-
180
)
37
Rubin
J. B.
Choi
Y.
Segal
R. A.
Cerebellar proteoglycans regulate sonic hedgehog responses during development
Development
2002
, vol. 
129
 (pg. 
2223
-
2232
)
38
Hartmann
G.
Prospero
T.
Brinkmann
V.
Ozcelik
C.
Winter
G.
Hepple
J.
Batley
S.
Bladt
F.
Sachs
M.
Birchmeier
C.
, et al. 
Engineered mutants of HGF/SF with reduced binding to heparan sulphate proteoglycans, decreased clearance and enhanced activity in vivo
Curr. Biol.
1998
, vol. 
8
 (pg. 
125
-
134
)
39
Murphy
K. J.
Merry
C. L.
Lyon
M.
Thompson
J. E.
Roberts
I. S.
Gallagher
J. T.
A new model for the domain structure of heparan sulfate based on the novel specificity of K5 lyase
J. Biol. Chem.
2004
, vol. 
279
 (pg. 
27239
-
27245
)
40
Ashikari
S.
Habuchi
H.
Kimata
K.
Characterization of heparan sulfate oligosaccharides that bind to hepatocyte growth factor
J. Biol. Chem.
1995
, vol. 
270
 (pg. 
29586
-
29593
)
41
Marquardt
J. U.
Seo
D.
Gomez-Quiroz
L. E.
Uchida
K.
Gillen
M. C.
Kitade
M.
Kaposi-Novak
P.
Conner
E. A.
Factor
V. M.
Thorgeirsson
S. S.
Loss of c-Met accelerates development of liver fibrosis in response to CCl4 exposure through deregulation of multiple molecular pathways
Biochim. Biophys. Acta
2012
, vol. 
1822
 (pg. 
942
-
951
)
42
Bataller
R.
Brenner
D. A.
Liver fibrosis
J. Clin. Invest.
2005
, vol. 
115
 (pg. 
209
-
218
)
43
Schuppan
D.
Afdhal
N. H.
Liver cirrhosis
Lancet
2008
, vol. 
371
 (pg. 
838
-
851
)
44
Yang
L.
Wang
Y.
Mao
H.
Fleig
S.
Omenetti
A.
Brown
K. D.
Sicklick
J. K.
Li
Y. X.
Diehl
A. M.
Sonic hedgehog is an autocrine viability factor for myofibroblastic hepatic stellate cells
J. Hepatol.
2008
, vol. 
48
 (pg. 
98
-
106
)
45
Collin de L'hortet
A.
Gilgenkrantz
H.
Guidotti
J. E.
EGFR: A master piece in G1/S phase transition of liver regeneration
Int. J. Hepatol.
2012
, vol. 
2012
 pg. 
476910
 
46
Stanford
K. I.
Bishop
J. R.
Foley
E. M.
Gonzales
J. C.
Niesman
I. R.
Witztum
J. L.
Esko
J. D.
Syndecan-1 is the primary heparan sulfate proteoglycan mediating hepatic clearance of triglyceride-rich lipoproteins in mice
J. Clin. Invest.
2009
, vol. 
119
 (pg. 
3236
-
3245
)
47
Bishop
J. R.
Passos-Bueno
M. R.
Fong
L.
Stanford
K. I.
Gonzales
J. C.
Yeh
E.
Young
S. G.
Bensadoun
A.
Witztum
J. L.
Esko
J. D.
Moulton
K. S.
Deletion of the basement membrane heparan sulfate proteoglycan type XVIII collagen causes hypertriglyceridemia in mice and humans
PLoS ONE
2010
, vol. 
5
 pg. 
e13919
 
48
Kanno
K.
Tazuma
S.
Chayama
K.
AT1A-deficient mice show less severe progression of liver fibrosis induced by CCl4
Biochem. Biophys. Res. Commun.
2003
, vol. 
308
 (pg. 
177
-
183
)
49
Kubota
N.
Kado
S.
Kano
M.
Masuoka
N.
Nagata
Y.
Kobayashi
T.
Miyazaki
K.
Ishikawa
F.
A high-fat diet and multiple administration of carbon tetrachloride induces liver injury and pathological features associated with non-alcoholic steatohepatitis in mice
Clin. Exp. Pharmacol. Physiol.
2013
, vol. 
40
 (pg. 
422
-
430
)

Supplementary data