JOURNAL TRANSCRIPT

---

Cell Tissue Res (2011) 346:79–88 DOI 10.1007/s00441-011-1236-0

REGULAR ARTICLE

TNFR1-mediated signaling is important to induce the improvement of liver fibrosis by bone marrow cell infusion Takuro Hisanaga & Shuji Terai & Takuya Iwamoto & Taro Takami & Naoki Yamamoto & Tomoaki Murata & Toshifumi Matsuyama & Hiroshi Nishina & Isao Sakaida

Received: 25 April 2011 / Accepted: 30 August 2011 / Published online: 11 October 2011 # The Author(s) 2011. This article is published with open access at Springerlink.com

Abstract The importance of TNF-α signals mediated by tumor necrosis factor receptor type 1 (TNFR1) in inflammation and fibrosis induced by carbon tetrachloride (CCl4), and in post-injury liver regeneration including a GFP/CCl4 model developed as a liver repair model by bone marrow cell (BMC) infusion, was investigated. In mice in which TNFR1 was suppressed by antagonist administration or by knockout, liver fibrosis induced by CCl4 was significantly decreased. In these mice, intrahepatic macrophage infiltration and TGF-β1 expression were reduced and stellate cell activity was decreased; however, expression of MMP-9 was also decreased. With GFP-positive BMC (TNFR1 wildtype, WT) infusion in these mice, fibrosis proliferation,

including host endogenous intrahepatic macrophage infiltration, TGF-β1 expression and stellate cell activity, increased significantly. There was no significant increase of MMP-9 expression. In this study, TNFR1 in hosts had a promoting effect on CCl4-induced hepatotoxicity and fibrosis, whereas BMC infusion in TNFR1 knockout mice enhanced host-derived intrahepatic inflammation and fibrosis proliferation. These findings differed from those in WT recipient mice, in which improvement in inflammation and fibrosis with BMC infusion had previously been reported. TNFR1-mediated signaling might be important to induce the improvement of liver fibrosis by bone marrow cell infusion.

T. Hisanaga : S. Terai (*) : T. Iwamoto : T. Takami : N. Yamamoto : I. Sakaida Department of Gastroenterology & Hepatology, Yamaguchi University Graduate School of Medicine, Minami Kogushi 1-1-1, Ube, Yamaguchi 755-8505, Japan e-mail: [email protected]

Keywords Tumor necrosis factor . CCl4 . Bone marrow cell . Liver cirrhosis . Hepatic stellate cell

T. Murata Institute of Laboratory Animals, Yamaguchi University, Minami Kogushi 1-1-1, Ube, Yamaguchi 755-8505, Japan T. Matsuyama Department of Molecular Microbiology and Immunology, Nagasaki University Graduate School of Biomedical Sciences, 1-12-4 Sakamoto, Nagasaki 852-8523, Japan H. Nishina Department of Developmental and Regenerative Biology, Medical Research Institute, Tokyo Medical and Dental University, Yushima 1-5-45, Bunkyo-ku, Tokyo 113-0034, Japan

Introduction The mechanism of hepatotoxicity induced by many factors, including CCl4, has been studied for a long time (Drill 1952) and besides direct cytotoxicity due to these factors, enhanced hepatotoxicity due to inflammatory cell infiltration and stellate cell activation induced by these factors also occurs. TNF-α is an important inflammatory cytokine that induces hepatotoxicity; however, it also initiates liver regeneration after injury. This has been examined in various previous hepatotoxicity, hepatic resection and hepatic failure models. In addition, in recent stem cell studies, TNF-α signals have been found to be essential in the differentiation/proliferation process from stem cells to functional cells. For effective expression of TNF-α signals, of the various receptors, TNFR1 (p55) is important. In

80

studies using TNFR1 knockout (KO) mice, while inhibition of hepatotoxicity has been shown, decreased liver regeneration and hepatocyte proliferation have also been confirmed (Gardner et al. 2003; Simeonova et al. 2001; Sudo et al. 2005; Yamada and Fausto 1998). In our laboratory, using a GFP/CCl4 model, BMC infusion has been shown to improve fibrosis and liver function in cirrhosis (Sakaida et al. 2004, 2008; Terai et al. 2002, 2005). In the liver of mice not treated with CCl4, without hepatic toxicity, repopulation of GFP-positive BMCs into the liver has not been observed, thus highlighting the need for chronic inflammatory signals in cell repopulation and effective expression. In addition, after BMC infusion, alterations in TNFR1 expression and associated increases in various cytokines such as fibroblast growth factor 2 (FGF2) have been recognized (Omori et al. 2004). This also suggested the importance of inflammatory signals, particularly TNF-α signals, in liver function improvement. Therefore, to investigate the importance of TNF-α signals mediated by TNFR1 in liver repair by BMC infusion in recipients, this study using TNFR1 antagonist-administered mice and TNFR1 KO mice as recipients was conducted.

Materials and methods Mice C57BL/6 TNFR1 KO mice were kindly provided by Hiroshi Nishina (Department of Developmental and Regenerative Biology, Medical Research Institute, Tokyo Medical and Dental University). TNFR1 KO mouse typing was performed by PCR analysis using primer typing box [p55R636 and HSV-TK (to confirm induction of silencing gene factor)]. Those primers were p55R636 (5′-GGCTGCAGTCCACGCACTGG-3′), and HSV-TK (5′-ATTCGCCAATGACAAGACGCTGG-3′) (data not shown) (Pfeffer et al. 1993). C57BL/6 female mice were purchased from Chiyoda SLC (Tokyo, Japan). GFP-transgenic mice (TgN(β-act-EGFP)Osb) were kindly provided by Masaru Okabe (Genome Research Center, Osaka University, Osaka, Japan). These mice were maintained in specific pathogen-free housing at the Animal Experiment Facility of Yamaguchi University School of Medicine and cared for in accordance with the animal ethics requirements at Yamaguchi University School of Medicine. Experimental protocol First, to confirm that signals mediated by TNFR1 are important in intrahepatic inflammatory responses and

Cell Tissue Res (2011) 346:79–88

fibrosis induced by CCl4, an experiment using TNFR1 antagonist (anti-mouse TNFR1 antibody, MAB430; R&D Systems, Minneapolis, MN, USA) was conducted in wildtype (WT) mice. In WT mice, CCl4 was administered for 4 weeks to create a liver cirrhosis model; then, by caudal vein injection of 100 μg/body of antagonist, signals mediated by TNFR1 were suppressed in recipients; and thereafter, CCl4 was continued for 1 week. A model was also created with infusion of GFP-positive BMCs after 1 h, when in vivo activity of the antagonist reaches a peak. There were 3 groups: WT (Control) (wild-type, without BMC infusion); WT+A (wild-type, with TNFR1 antagonist only, without BMC infusion); and GFP/WT+A (wildtype, with TNFR1 antagonist and with BMC infusion) (Fig. 1). As will be described later, by blocking TNFR1, suppression of fibrosis and suppression of inflammatory cell infiltration were confirmed. Therefore, as a more highly specific model, a model was created by the following protocol with TNFR1 KO mice as BMC infusion recipients. Six-week-old female C57BL/6 mice and female isogenic TNFR1 KO mice were treated with CCl4 (1.0 ml/kg body diluted 1:3 in corn oil) twice a week for 8 weeks. In the other group, after 4 weeks of CCl4 administration in each group (C57BL/6 wild-type and TNFR1 KO), bone marrow cells (BMC) (1×105 cells) from GFP transgenic mice were injected via the tail vein as previously described (Terai et al. 2003). After 8 weeks, 36 h after the last CCl4 injection, the mice were sacrificed to examine the blood data and liver tissue specimens. The liver was fixed in 4% buffered paraformaldehyde for 24–48 h and paraffin embedded. Blood samples were obtained by cardiac puncture and drawn into a glass tube containing 7.5% EDTA (pH 7.4). After centrifugal separation, the plasma was stored at 4°C. There was a total of 4 groups in this study: WT (Control), wild-type without BMC infusion; KO (Control), TNFR1 KO without BMC infusion; GFP/WT, wild-type with GFP-positive BMC infusion; and GFP/KO, TNFR1 KO with GFP-positive BMC infusion (Fig. 1). Quantitative analysis of liver fibrosis and immunohistochemistry The liver fibrosis area was quantified with Sirius-red staining using an Olympus Provis microscope equipped with a CCD camera (Olympus, Tokyo, Japan). The red area, considered the fibrotic area, was assessed by computer-assisted image analysis with MetaMorph software (Universal Imaging, Downingtown, PA, USA) at a magnification of ×40. The mean value of 10 randomly selected areas per sample was used as the expressed percent area of fibrosis.

Cell Tissue Res (2011) 346:79–88

81

Fig. 1 Experimental protocols. Protocol I In each group of mice, CCl4 (1.0 ml/kg body) was administered twice a week for 5 weeks to create a liver cirrhosis model. In WT+A mice, 100 μg/body of TNFR1 antagonist was injected via tail vein before 1 week of sacrifice. GFP/ WT+A mice were treated with GFP positive BMC (1×105 cells)

infusion via tail vein 1 hour after TNFR1 antagonist injection. Protocol II WT (Control) mice and KO (Control) mice were treated with CCl4 (same as above) for 8 weeks. In GFP/WT mice and GFP/KO mice, after 4 weeks of CCl4 administration, BMCs (1×105 cells) were injected, followed by 4 weeks of CCl4 treatment

Immunohistochemistry of TGF-β1, alpha smooth muscle actin (α-SMA), matrix metalloproteinase (MMP)-9 and F4/80

ab16911; Abcam] staining by the avidin-biotin-peroxidase complex method. Additionally, double immunofluorescent staining was performed to study co-expression of GFP and F4/80 in bone marrow cell-infused mice. The mixture of the first antibodies was GFP and F4/80 noted above. The secondary antibodies, goat anti-rabbit IgG (H+L), Alexa Fluor 488 (A11034,;Invitrogen) (Green) and goat antirat IgG (H+L), Alexa Fluor 568 (A11077; Invitrogen) (Red) were each applied at a concentration of 1:400 in PBS for 60 min at room temperature. Before attaching the coverslip, DAPI (D212; Dojindo Laboratories, Kumamoto, Japan) was applied for counterstaining to visualize all nuclei in the tissue sections. The sections were viewed and photographed with the CCD camera noted above.

Three-μm-thick liver sections were mounted on microscope slides, routinely dewaxed and rehydrated and pretreated with Vector Antigen Unmasking Solutions (Citrate-based, Cat. No. H-3300). For the immunohistochemical analysis, the Vectastain ABC kit (Vector Laboratories, Burlingame, CA, USA) was used for GFP (anti-GFP, rabbit IgG fraction, A11122; Invitrogen, Carlsbad, CA, USA), TGF-β1[TGFβ1(V), SC-146; Santa Cruz Biotechnology], alpha-smooth muscle actin (α-SMA) (alpha smooth muscle actin antibody, ab6594; Abcam, Cambridge, MA, USA), matrix metalloproteinase (MMP)-9 (anti-mouse MMP-9 antibody, AF909; R&D Systems) and F4/80 [F4/80 antibody(BM8),

82

Real-time quantitative PCR analysis Total RNA was isolated from the livers of the mice treated at 4 weeks after the BMC infusion or control CCl4 treatment. The messenger RNA (mRNA) expressions of TGF-β1 and MMP-9 were evaluated using real-time quantitative PCR. Total RNA was extracted using the RNeasy Mini Kit (Qiagen, Hilden, Germany). For cDNA synthesis, AMV reverse transcription reagents were used according to the manufacturer’s instructions (Roche Diagnostic, Pleasanton, CA, USA). Real-time PCR was performed with SYBR Green Master Mix (Roche Diagnostic). The primers used for TGF-β1 were 5′-GAAGCCATCCGTGGCCAGAT-3′ (forward) and 5′-GACGTCAAAAGACAGCACT-3′ (reverse), for MMP-9 were 5′-GGAACTCACACGACATCTTCCA-3′ (forward) and 5′-GAAACTCACACGCCAGAAGAATTT-3′ (reverse) and collagen type 1 alpha were 5′-CGGGCAG GACTTGGGTA-3′ (forward) and 5′-CGGAATCT GAATGGTCTGACT-3′ (reverse). The PCR primers used for mouse glyceraldehyde-3-phospatase dehydrogenase (GAPDH), which was used as an internal control, were: 5′G T C T T C A C C A C C AT G G A G A A G G C - 3 ′ , 5 ′ - AT G CCAGTGAGCTTCCCGTTCAGC-3′. The cycle for PCR was as follows: 1 cycle of 95°C for 20 s; 40 cycles of 3 s at 95°C and 30 s at 60°C; and 1 cycle of 95°C for 15 s, 60°C for 1 min, and 95°C for 15 s. Reactions were performed in a Step One Plus™ real-time PCR system (Applied Biosystems, California, CA, USA) and amounts of all mRNAs were quantified using StepOne™ version 2.1 software (Applied Biosystems). Western blot analysis The liver samples (approximately 40 mg) were homogenized in 1 ml of cell lysis buffer (Cell Signal Technology, Beverly, MA, USA) and 1 mM phenylmethanesulfonyl fluoride (PMSF) and complete mini-centrifuged (Roche Diagnostic). The supernatant represented the whole protein. Then, 40 μg of the protein sample were mixed with the same volume of loading buffer (5% 2-mercaptoethanol and 95% Laemmli Sample Buffer; Bio-Rad Laboratories, Hercules, CA, USA), heated for 4 min at 97°C and separated by 10% SDS-PAGE. The separated bands were transferred to Immobilon-P transfer membrane (Millipore, Billerica, MA, USA), followed by blocking of the membranes for 1 h with blocking buffer (0.1% Tween-20; Wako Pure Chemical Industries, Osaka, Japan), 0.2% IBlock™ reagent (Tropix, Bedford, MA, USA) and 1 mM Tris-HCl buffer (pH 7.5) (Invitrogen). The membranes were then washed with washing buffer (0.1% Tween-20, 1 mM Tris-HCl buffer; pH 7.5) and incubated overnight at 4°C with the primary antibodies against MMP-9 (R&D Sytems), TGF-β1 (Santa Cruz Biotechnology) and β-actin (Abcam)

Cell Tissue Res (2011) 346:79–88

in blocking buffer. Then, after being washed, the membranes were incubated for 1 h at room temperature with the appropriate secondary antibodies. Reactive bands were identified using ECL (Amersham Biosciences, Piscataway, NJ, USA) and autoradiography, according to the manufacturer’s instructions. Statistical analysis Statistical significance was determined using 2-tailed Student'st test. Results are presented as the mean±standard deviation and differences of p