Reticulon 4B (Nogo-B) is a novel regulator of hepatic fibrosis

Authors


  • Potential conflict of interest: Nothing to report.

Abstract

Nogo-B, also known as Reticulon 4B, plays important roles in vascular injuries. Its function in the liver is not understood. The aim of this study was to characterize Nogo-B in liver fibrosis and cirrhosis. Nogo-B distribution was assessed in normal and cirrhotic human liver sections. We also determined the levels of liver fibrosis in wild-type (WT) and Nogo-A/B knockout (NGB KO) mice after sham operation or bile duct ligation (BDL). To investigate the mechanisms of Nogo-B's involvement in fibrosis, hepatic stellate cells were isolated from WT and NGB KO mice and transformed into myofibroblasts. Portal pressure was measured to test whether Nogo-B gene deletion could ameliorate portal hypertension. In normal livers, Nogo-B expression was found in nonparenchymal cells, whereas its expression in hepatocytes was minimal. Nogo-B staining was significantly elevated in cirrhotic livers. Fibrosis was significantly increased in WT mice 4 weeks after BDL compared with NGB KO mice. The absence of Nogo-B significantly reduced phosphorylation of Smad2 levels upon transforming growth factor β (TGF-β) stimulation. Reconstitution of the Nogo-B gene into NGB KO fibroblasts restored Smad2 phosphorylation. Four weeks after BDL, portal pressure was significantly increased in WT mice by 47%, compared with sham-operated controls (P = 0.03), whereas such an increase in portal pressure was not observed in NGB KO mice (P = NS). Conclusion: Nogo-B regulates liver fibrosis, at least in part, by facilitating the TGFβ/Smad2 signaling pathway in myofibroblasts. Because absence of Nogo-B ameliorates liver fibrosis and portal hypertension, Nogo-B blockade may be a potential therapeutic target in fibrosis/cirrhosis. (HEPATOLOGY 2011;)

Nogo-B, also known as Reticulon 4B, is a member of the reticulon (Rtn) family of proteins that are primarily localized to the endoplasmic reticulum. In mammalian cells, there are four Rtn genes, Rtn-1, -2, -3, and -4, and each gene encodes multiple isoforms. For Rtn 4, there are three gene products: Nogo-A, -B, and -C. Nogo-A and -C are highly expressed in the central nervous system, with Nogo-C also present in the skeletal muscle.1, 2 Nogo-A is known to inhibit axonal growth and repair, whereas the function of Nogo-C is not well known.1, 2 Nogo-B is found in many tissues and regulates vascular remodeling by enhancing migration of endothelial cells, while inhibiting migration and proliferation of smooth muscle cells1, 3 in pathological vascular conditions caused by ischemia, atherosclerosis, and other insults. Its role in wound healing indicates that it may have relevance to hepatic injury.4-8

Liver fibrosis and cirrhosis are the consequence of a sustained wound healing response to chronic hepatic injury from a variety of causes, including cholestasis and drug-induced and metabolic diseases.9 Because Nogo-B plays an important role in wound-healing,4 we hypothesized that Nogo-B may be involved in liver fibrosis and cirrhosis. Currently, no studies have examined the role of Nogo-B in the liver. Thus, we first examined Nogo-B expression in normal and cirrhotic human livers. Second, we determined whether Nogo-B is involved in liver fibrosis/cirrhosis using Nogo-B knockout mice. Third, we investigated the mechanisms by which Nogo-B contributes to liver fibrosis/cirrhosis.

We found that Nogo-B is highly expressed in nonparenchymal cells and is significantly elevated in cirrhosis, whereas its absence ameliorates hepatic fibrosis. The absence of Nogo-B impaired phosphorylation of Smad2 in response to transforming growth factor β (TGF-β), the most potent fibrogenic cytokine,10, 11 in hepatic stellate cells (HSCs) that were transformed into myofibroblasts (MF-HSCs). Furthermore, the absence of Nogo-B blocked the development of portal hypertension. These data suggest that Nogo-B promotes liver fibrosis through activation of the TGF-β signaling pathway in MF-HSCs.

Abbreviations

α-SMA, α-smooth muscle actin; BDL, bile duct ligation; eNOS, endothelial nitric oxide synthase; ER, endoplasmic reticulum; HSC, hepatic stellate cell; Hsp90, heat shock protein 90; LSEC, liver sinusoidal endothelial cell; MEF, mouse embryonic fibroblast; MF-HSC, myofibroblast; mRNA, messenger RNA; NGB KO, Nogo-A/B knockout; PCR, polymerase chain reaction; Rtn, reticulon; TGF-β, transforming growth factor β; TGF-βRI, transforming growth factor β receptor I; WT, wild-type.

Materials and Methods

Human Liver Specimens

Archival specimens of human livers, used for immunohistochemistry of Nogo-B, were obtained from patients with liver cirrhosis who underwent surgery in the Department of Surgery I, School of Medicine, University of Occupational and Environmental Health, Kitakyushu, Japan. The Clinical Ethics Committee of the University of Occupational and Environmental Health approved the study.

Animals

All animal studies were approved by the Institutional Animal Care and Use Committees of Yale University and Veterans Affairs Connecticut Healthcare System, and performed in adherence with the National Institutes of Health Guidelines for the Use of Laboratory Animals. Nogo-A/B knockout (NGB KO) mice were a gift from Stephen Strittmatter12 or Mark Tessier-Lavigne.13

Bile Duct Ligation Surgery

Male NGB KO mice at 2 months of age and their age-matched littermate wild-type (WT) mice underwent either sham operation or bile duct ligation (BDL).14 In sham operation, the bile duct was exposed, but not ligated.15 Liver and spleen samples from these mice were collected 1 and 4 weeks after surgery. The samples were snap-frozen in liquid nitrogen and stored at −80°C until used.

CCl4 Intraperitoneal Injection

Mice were injected with CCl4 using a method described by Constandinou et al.16 with minor modifications. In brief, CCl4 was mixed with olive oil in a fume hood. For intraperitoneal injections, one part of olive oil was added to three parts of CCl4. The CCl4 mixture (≈7.5 M CCl4) was administered to mice every 3 days (7.5 μmol/μL/g body weight/injection) for 8 weeks. To insure constant dosing, each injection was normalized to the current body weight. The same normalized dose of olive oil was injected to control mice.

Histological Analyses

Five-micrometer-thick sections of paraffin-embedded liver samples were stained with hematoxylin and eosin for structural evaluation and with Sirius red for the evaluation of fibrosis. Fibrosis was determined by calculating the percent Sirius red–positive area (i.e., collagen-positive area) over the total area analyzed. Because there was no difference in Sirius red–positive areas in sham-operated livers between 1-week and 4-week time points, results from sham-operated groups were pooled. Image J1.41o (Wayne Rasband, National Institutes of Health, Bethesda, MD) was used for image analysis of the entire liver sections. At least 20 images per liver section were randomly taken and used for the analyses.

Rat Liver Cells

Liver sinusoidal endothelial cells (LSECs), hepatocytes, and HSCs were isolated from male Sprague-Dawley rats at the Yale Liver Center Core Facility. To confirm the purity of LSEC and HSC isolation, we assessed endothelial nitric oxide synthase (eNOS) and α-smooth muscle actin (α-SMA) expression, respectively.

Immunohistochemistry

Formalin-fixed human liver samples, embedded in paraffin, were cut into 6-μm-thick sections and used for immunolabeling as described.17 A primary antibody used was goat anti-human Nogo-B (N-18, 1:100; Santa Cruz Biotechnology, Santa Cruz, CA; recognizing epitopes 1-18). These sections were then incubated with a secondary antibody conjugated with horseradish peroxidase (1:200) at room temperature for 1 hour. Sections incubated with rabbit immunoglobulin G in place of a primary antibody, or incubated with a secondary antibody only, served as negative controls.

Western Blot Analysis

Liver samples or cells were homogenized in a lysis buffer containing 50 mM Tris HCl, 0.1 mM ethylene glycol tetraacetic acid, 0.1 mM ethylene diamine tetraacetic acid, 0.1% sodium dodecyl sulfate, 0.1% deoxycholic acid, 1% (vol/vol) Nonidet P-40, 5 mM sodium fluoride, 1 mM sodium pyrophosphate, 1 mM activated sodium vanadate, 0.32% protease inhibitor cocktail (Roche Diagnostics, Mannheim, Germany), and 0.027% Pefabloc. The lysates were centrifuged at 14,000 rpm for 10 minutes at 4°C. The protein concentrations in the supernatants were quantified using the Lowry assay. An equal amount of protein (50-80 μg) from each sample was loaded onto sodium dodecyl sulfate–polyacrylamide gel electrophoresis gels and transferred to polyvinylidene fluoride or nitrocellulose membranes. Membranes were probed with antibodies that recognize rabbit anti-Nogo serum (1761A; 1:10,000), mouse anti–α-SMA (1:1,000; Sigma, St. Louis, MO), heat shock protein 90 (Hsp90), and eNOS (1:1,000; BD Biosciences, San Jose, CA) and mouse anti–β-actin (1:1,000; Sigma). After washing, membranes were incubated with fluorophore-conjugated secondary antibodies (either 680 nm or 800 nm emission). Detection and quantification of bands was performed using the Odyssey Infrared Imaging System (Li-Cor Biotechnology, Lincoln, NE). Hsp90 or β-actin was used as a loading control.

Quantitative Real-Time Polymerase Chain Reaction Analysis

The total RNA from approximately 50 mg of frozen mouse livers was isolated using TRIZOL reagent (Sigma) and further purified using an RNA cleanup kit (Qiagen Sciences, Germantown, MD). RNA concentrations were measured using a DU530 spectrophotometer (Beckman Coulter, Brea, CA). Four micrograms of the total RNA was used as a template to synthesize complementary DNA using SupertScript II Reverse Transcriptase (Invitrogen, Carlsbad, CA) and 10 mM dNTP Mix (Invitrogen). Then, complementary DNA was diluted 5 times to be used as a real-time polymerase chain reaction (PCR) template. Real-time PCR was performed using ABI 7500 SDS software (Applied Biosystems, Foster City, CA). TaqMan gene expression assays (Applied Biosystems) were used for measurement of glyceraldehyde 3-phosphate dehydrogenase (Mm99999915_g1), collagen type Iα, transforming growth factor-β1 (Mm01178820_m1), and transforming growth factor β receptor 1 (Mm00 436971_m1) gene expression.

Hydroxyproline Assay

Frozen liver tissues (50 mg) were homogenized in 1 mL 6N HCl and heated at 110°C in a heating block for 20 hours. After cooling, the samples were filtered through a 1 ml syringe

filter (Pall Corporation, Ann Arbor, MI). Fifty microliters of filtered homogenates was neutralized with 450 μL of 2.2% NaOH in citrate acetate buffer (50 g/L citric acid monohydrate, 12 mL/L acetic acid, 120 g/L sodium acetate (H2O)3, and 34 g/L NaOH) in a 1.5-μL tube. Two-hundred fifty microliters of chloramine-T solution (0.141 g chloramine-T, 2 mL H2O, 3 mL methoxyethanol, and 5 mL citrate acetate buffer) was added to 500 μL neutralized homogenate, standard hydroxyproline solution, or citrate acetate buffer and incubated for 20 minutes at room temperature. Two-hundred fifty microliters of perchloric acid was added into the reaction mixture and incubated for 20 minutes. Finally, 250 μL of dimethylbenzaldehyde solution (2 g dimethylbenzaldehyde in 10 mL methoxyethanol) was added and incubated at 60°C for 20 minutes. After cooling, the absorbance was measured at 550 nm using a Synergy 2 Multi-Mode Microplate reader (BioTek Instruments, Winooski, VT).

Mouse Embryonic Fibroblasts and Reconstitution of NGB KO Mouse Embryonic Fibroblasts with Nogo-B Gene

Isolation.

Mouse embryonic fibroblasts (MEFs) were isolated from embryonic day 13.5 embryos and immortalized by serial passaging. Briefly, pregnant mothers were sacrificed on postcoital day 13.5 by cervical dislocation. After spraying with 70% ethanol, the abdomen of the mother was opened, and the uterus containing the embryos was transferred in sterile phosphate-buffered saline. Using fine forceps and scissors, embryos that contained fetal membranes were extracted from the uterus and transferred to fresh sterile phosphate-buffered saline. The rest of the procedure was performed in a tissue culture hood. Embryos were separated from placenta and membranes, eviscerated, and decapitated, and embryonic bodies were transferred to trypsin–ethylene diamine tetraacetic acid. Using razor blades, embryonic bodies were minced into fine tissue pieces and incubated for 10 minutes in 5% CO2 at 37°C. Complete media was added, and tissues were disaggregated by repeated pipetting and plated for culture. Media was refreshed the next day. On day 3, cells were either frozen down, passaged for further experiments, or immortalized.

Immortalization.

NGB KO MEFs and corresponding WT MEFs were immortalized by serial replating of cells. Briefly, MEFs were split to one-third twice per week during the rapid growth phase. Splitting was reduced to one-half or cells were merely replated as cells entered into senescence. Immortalization occurred following 10-20 passages. Cells were used as pooled immortalized clones.

Reconstitution.

Immortalized NGB KO MEFs were transfected with HA-tagged Nogo-B pIRES vector to generate reconstituted MEFs. Knockout cells transfected with empty pIRES vector alone served as negative controls. Cells expressing HA-tagged Nogo-B or empty vector were selected according to their G418 (neomycin) antibiotic resistance.

Isolation of HSCs and Transformation to Myofibroblasts

Mouse primary HSCs were isolated from WT and NGB KO mice by in situ pronase-collagenase perfusion followed by density gradient centrifugation, as described.18, 19 Those HSCs were maintained in Dulbecco's modified Eagle's medium/F12 medium supplemented with 10% heat-inactivated fetal bovine serum, penicillin (100 U/mL), streptomycin (100 μg/mL), 1%-amphotericin B, and gentamycin (20 μg/mL), and used between 1 and 7 days as HSCs or after 14 days of cell culture as activated and transformed myofibroblasts (MF-HSCs).20

TGF-β Signaling Study

Either MEFs, HSCs, or MF-HSCs were seeded at a density of 1 × 105/well in 6-well tissue culture plates and allowed for overnight attachment in the medium containing 10% heat-inactivated fetal bovine serum. Cells were then replaced with serum-free medium, incubated for 16 hours and stimulated with 0, 0.5, 1.0, or 5.0 ng/mL of recombinant human TGF-β1 (R&D Systems, Minneapolis, MN) for 24 hours. Cell lysates were collected and used for western blot analysis.

Portal Pressure Measurement

Portal pressures were measured in mice with sham operation or 4-week BDL as described.21 In brief, the tip of a 30-gauge needle was inserted into the ileocolic vein. The needle length was joined to a short length of PE-10 tubing, which in turn was joined to PE-50 tubing and connected to a Hewlett-Packard pressure transducer. The readings were monitored and saved on a computer using the analog-to-digital PowerLab system (AD Instruments, Colorado Springs, CO).

Statistical Analysis

Results are expressed as the mean ± SE. Statistical analysis was performed with SPSS 14.0 statistical software (SPSS, Chicago, IL). Results were assessed using one-way analysis of variance followed by a Student t test.P < 0.05 was considered significant.

Results

Nogo-B Is Highly Expressed in Nonparenchymal Cells and Minimally in Hepatocytes in Human Liver.

The tissue distribution of Nogo-B was first assessed in human archival livers. In normal livers, Nogo-B was highly expressed in nonparenchymal cells, but was absent in hepatocytes (Fig. 1A). Similarly, cirrhotic livers showed high expression of Nogo-B in fibrotic areas and minimally or none in hepatocytes (Fig. 1B).

Figure 1.

Nogo-B is expressed in human livers and localizes primarily to nonparenchymal cells in regions of fibrotic liver injury. (A) Nogo-B expression in normal liver shows prominent staining of bile ducts (arrow) and hepatic sinusoids (arrowheads) with absent staining over hepatocytes (asterisks). (B) Nogo-B expression is especially prominent in fibrotic portal tracts (arrow), whereas immunolabeling of hepatocytes remains negligible. Respective low-power (left) and high-power (right) images are shown.

Nogo-B Is Also Expressed in Nonparenchymal Cells and Minimally in Hepatocytes in Rats.

We then determined Nogo-B protein expression in rat primary cells (LSECs, hepatocytes, and HSCs) (Fig. 2). As observed in human livers, nonparenchymal cells, such as LSECs and HSCs, expressed Nogo-B, whereas hepatocyte lysates only revealed a band slightly below 40 kDa. This may represent a nonspecific binding or a splice variant, but was consistently less than the predicted molecular weight of Nogo-B. Nogo-A protein (molecular weight ≈250 kDa) was not detected in all the cell types.

Figure 2.

Differential expression patterns of Nogo-B demonstrate that Nogo-B is expressed by primarily by nonparenchymal cells of the liver. Nogo-B expression is shown in rat liver cells, including LSECs, hepatocytes, and HSCs. eNOS was used as an LSEC marker; α-SMA was used as an HSC marker. β-Actin was used as a loading control. The asterisk indicates a band seen only in hepatocyte lysates that is smaller than the molecular weight of Nogo-B.

Nogo-B Protein Levels Are Significantly Elevated in Fibrotic/Cirrhotic Livers.

We performed western blot analysis in experimental models of fibrosis/cirrhosis in mice. Nogo-B has two isoforms, Nogo-B1 (lower band with ≈45 kDa) and B2 (upper band with ≈48 kDa).4 Hsp90 was used as a loading control, because expression of this protein is not affected by cirrhosis (data not shown). We analyzed Nogo-B levels in mouse cirrhotic livers that were generated by way of intraperitoneal CCl4 injection (Fig. 3A). The bar graphs depict the total expression of both Nogo-B1 and -B2 isoforms. There was a significant increase in Nogo-B expression: a three-fold increase with CCl4 injection (P = 0.001). We also analyzed Nogo-B levels in mouse cirrhotic livers that were generated by way of BDL (Fig. 3B). Similar to CCl4-induced fibrosis/cirrhosis, mouse livers isolated 4 weeks after BDL showed a significant five-fold increase in Nogo-B expression (Fig. 3B) compared with sham-operated controls. Collectively, these data indicate that Nogo-B levels are elevated in fibrotic/cirrhotic livers.

Figure 3.

Nogo-B levels are significantly increased in mice with cirrhosis. Both Nogo-B2 (upper band with ≈48 kDa) and Nogo-B1 (lower band with ≈45 kDa) were up-regulated. (A) Fibrotic livers isolated from mice given intraperitoneal injection of CCl4 for 8 weeks. Control mice were given olive oil for 8 weeks (n = 4 per group). (B) Cholestatic livers isolated from mice 4 weeks after sham operation or BDL (n = 3 per group).

Nogo-B Gene Deletion Reduces Hepatic Fibrosis.

Fibrosis

We used Sirius red staining for the assessment of liver fibrosis in WT and NGB KO mice that underwent sham operation or BDL (Fig. 4A). The livers were isolated either after 1 week (early stage of cirrhosis) or 4 weeks (late stage of cirrhosis) after surgery. Fig. 4B shows the lack of Nogo-B protein expression in NGB KO livers and an increase in Nogo-B expression in WT BDL livers. One week after BDL, both WT and NGB KO livers showed a significant three-fold increase in fibrosis as indicated by Sirius red staining (Fig. 4A), but there was no difference in fibrosis between WT and NGB KO livers. However, 4 weeks after BDL, NGB KO livers showed significantly less fibrosis than WT livers, indicating that Nogo-B gene deletion blocked the progression of fibrosis (Fig. 4A).

Figure 4.

Nogo-B gene deletion reduces the progression of liver fibrosis. (A) Representative images of Sirius red staining for the detection of collagens as an indicator of fibrosis. Percent of the Sirius red–positive areas over the total areas of livers analyzed is shown (n = 5-7 per group). (B) Representative western blot analysis confirms absence of Nogo-B proteins in NGB KO livers. (C) α-SMA protein levels are significantly increased in livers isolated from WT mice 4 weeks after BDL, compared with those of NGB KO mice.

α-SMA.

Levels of α-SMA protein (Fig. 4C), a marker of activated HSCs,22 were determined in livers isolated from sham-operated and BDL mice. We observed that WT mice showed a significant increase in hepatic α-SMA levels 4 weeks after BDL, compared with NGB KO mice. This result indicates that Nogo-B gene deletion decreased the number of α-SMA positive cells (e.g., HSCs) in cirrhotic livers.

Collagen Type Iα Expression and Hydroxyproline Levels

Quantitative real-time PCR analysis confirmed significantly lower expression of collagen type Iα messenger RNA (mRNA) in NGB KO BDL livers compared with WT BDL livers 4 weeks after surgery (Fig. 5A). Hydroxyproline levels were also significantly lower in NGB KO BDL livers than in WT BDL livers (P < 0.04) (Fig. 5B). Furthermore, in WT mice, hydroxyroline levels were increased 4.5-fold in BDL livers compared with sham-operated livers, whereas the increase was just two-fold in NGB KO mice. These results suggest that Nogo-B gene deletion reduces fibrosis in the later stages of cirrhosis.

Figure 5.

Nogo-B gene deletion significantly reduces collagen type Iα and TGF-β expression levels, and hydroxyproline content in cirrhotic mouse livers. Quantitative real-time PCR analysis was performed in livers collected 4 weeks after sham operation or BDL. Data were normalized against glyceraldehyde 3-phosphate dehydrogenase expression (n = 5-6 per group). (A) Collagen type Iα mRNA levels. (B) Hydroxyproline levels of livers. (C) TGF-β mRNA levels. (D) TGF-βRI mRNA levels (n = 5 per group).

TGF-β and TGF-β Receptor I

TGF-β expression was significantly increased in WT BDL livers 4 weeks after surgery compared with their sham-operated counterparts (P = 0.04) (Fig. 5C). In contrast, there was no increase in TGF-β expression in NGB KO BDL livers compared with sham-operated livers. Moreover, TGF-β expression was significantly higher in WT than in NGB KO livers (P = 0.04) after BDL. Unlike TGF-β, its receptor (TGF-βRI) expression was unchanged in response to BDL in both WT and NGB KO mice (Fig. 5D).

Nogo-B Gene Deletion Impairs Smad2 Phosphorylation in Response to TGF-β Stimulation in MEFs and MF-HSCs.

Using MEFs isolated from WT and NGB KO mice, we determined Smad2 phosphorylation, a downstream effector of TGF-β signaling known as the most potent fibrogenic pathway, in response to TGF-β stimulation (Fig. 6A). We found that phosphorylated Smad2 levels were significantly increased in a TGF-β dose-dependent manner in WT MEFs, but not in NGB KO MEFs (Fig. 6A). Furthermore, when NGB KO MEFs were reconstituted with Nogo-B, Smad2 phosphorylation was restored upon TGF-β stimulation (Fig. 6B). These data strongly suggest that Nogo-B enhances the TGF-β/Smad2 signaling pathway in fibroblasts.

Figure 6.

Absence of Nogo-B blunts TGF-β–mediated Smad2 phosphorylation in MEFs and MF-HSCs. (A) Phosphorylated Smad2 levels in MEFs isolated from WT and NGB KO mice. These cells were treated with 0, 0.5, and 1.0 ng/mL TGF-β for 24 hours. (B) Phosphorylated Smad2 levels in NGB KO MEFs reconstituted with HA-tagged Nogo-B (RC) or empty lentiviral vector (KO). These cells were also treated with 0, 0.5, and 1.0 ng/mL TGF-β for 24 hours. (C) Phosphorylated Smad2 levels in MF-HSCs (21 days after HSC culture) generated from WT and NGB KO mice. These cells were also treated with 0, 1, and 5.0 ng/mL TGF-β for 24 hours. (D) Nogo-B immunofluorescent image of MF-HSCs (21 days after hepatic stellate cell culture) generated from WT and NGB KO mice. Nogo-B is shown in red and the nucleus in blue. Representative data from 3-5 independent experiments.

We then tested whether this finding in MEFs can be recapitulated in MF-HSCs. We generated MF-HSCs by transforming quiescent HSCs isolated from WT and NGB KO mice into myofibroblasts through at least 2 weeks of cell culture (Fig. 6D) and used them to perform the same experiments as above in MEFs. These MF-HSCs expressed α-SMA (data not shown), indicating the activation and transformation of quiescent HSCs into myofibroblasts. Similar to MEF cells, Smad2 phosphorylation in WT MF-HSCs was increased in a TGF-β dose-dependent manner, but was not in NGB KO cells (Fig. 6C). Taken together with the results in MEFs (Figs. 6A,B), this finding strongly suggests that Nogo-B facilitates Smad2 phosphorylation in response to TGF-β in MF-HSCs and that Nogo-B deletion impairs TGF-β/Smad2 signaling in MF-HSCs.

Nogo-B Gene Deletion Blocks the Development of Portal Hypertension.

At 4 weeks post-BDL, WT mice displayed a significantly elevated portal pressure compared with sham-operated control mice (P < 0.05). In contrast, NGB KO mice did not show such a significant increase in portal pressure 4 weeks after BDL (Fig. 7A). This finding demonstrates that the amelioration of hepatic fibrosis seen in NGB KO mice reduces experimental portal hypertension. WT mice with BDL also showed a significant increase in spleen weights compared with sham-operated control mice (P < 0.0001), whereas NGB KO with BDL did not. Because the development of portal hypertension increases spleen weights,23 the significant increase in spleen weights observed only in WT mice 4 weeks after BDL is consistent with the significantly elevated portal pressure observed in WT mice.

Figure 7.

Nogo-B gene deletion blocks the development of portal hypertension in cirrhotic mice. (A) Portal pressure was measured 4 weeks after sham operation or BDL in mice (n = 7-11 per group). Portal pressures in BDL mice were compared with their sham-operated control mice. (B) Spleen weights of WT and KO mice 4 weeks after sham operation or BDL (n = 7-11 per group).

Discussion

This study demonstrates that Nogo-B is expressed by nonparenchymal liver cells and plays a critical role in hepatic fibrosis. Hepatic Nogo-B levels are significantly elevated in cirrhosis, whereas the absence of Nogo-B ameliorates liver fibrosis and portal hypertension. Loss of Nogo-B decreases TGF-β expression, the most potent fibrogenic stimulus, in cirrhotic livers. Furthermore, we found that Nogo-B facilitates TGF-β signaling by enhancing phospho-Smad2 levels in MEFs as well as MF-HSCs. Taken together, our results suggest that Nogo-B promotes liver fibrosis by: (1) enhancing TGF-β expression or TGF-β–expressing cells in fibrotic/cirrhotic livers and (2) facilitating the TGF-β signaling pathway in MF-HSCs. Because Nogo-B deletion ameliorates these processes, Nogo-B is a new regulator for liver fibrosis.

Little is known about Nogo-B's function in the liver. Although Nogo-B is known to play important roles in pathological vascular conditions in response to vascular injuries, such as ischemia and atherosclerosis,4-8 the molecular mechanisms by which Nogo-B mediates these conditions are largely unknown. Our study, for the first time, demonstrates a new role of Nogo-B for liver fibrosis and potential mechanisms through which Nogo-B facilitates fibrosis. Because the TGF-β signaling pathway also mediates fibrosis in other organs,10, 11 our findings might be applicable to fibrosis in other organ systems.

We observed a significant elevation in Nogo-B levels in two different mouse models of fibrosis/cirrhosis, including intraperitoneal CCl4 injection for 8 weeks and 4-week post-BDL. Elevated Nogo-B levels in liver fibrosis/cirrhosis may be due to an increase in Nogo-B expression in nonparenchymal cells per se and/or an increase in the number of Nogo-B–expressing nonparenchymal cells in liver cirrhosis. Because Nogo-B plays a role in cell proliferation of vascular cells,5 Nogo-B may contribute to the regulation of cell proliferation in those nonparenchymal cells in response to liver injury. In hepatic fibrosis, namely, an increase in Nogo-B levels may reflect enhanced cell proliferation of collagen-producing cells, such as MF-HSCs. In fact, α-SMA levels, a marker of MF-HSCs, were significantly higher in WT mice than in NGB KO mice 4 weeks after BDL, when fibrosis was significantly more prominent in WT livers than in NGB KO livers. Because MF-HSCs are the primary source of α-SMA22 in liver fibrosis, a decreased MF-HSC population may be attributable to decreased MF-HSC proliferation or possibly an increase in MF-HSC apoptosis, which in turn contributes to blocking progression of fibrosis in NGB KO livers.

Transforming growth factor β is the most important cytokine that promotes fibrosis22, 24 and is markedly increased during fibrogenesis.24 Because almost all cell types express TGF-β receptors, TGF-β influences all steps of fibrosis. Furthermore, multiple cell types secrete TGF-β, including apoptotic parenchymal cells and activated myofibroblasts.24 Therefore, we investigated whether Nogo-B alters TGF-β and TGF-β receptor expression in cholestatic livers (4 weeks post-BDL). We found that NGB KO significantly reduced TGF-β expression without altering the levels of TGF-β receptor expression. We thus speculate that a decrease in TGF-β expression observed in NGB KO cholestatic livers is due to a decrease in TGF-β expression in TGF-β–producing cells and/or a decrease in proliferation of those TGF-β–producing cells during fibrogenesis.

A complementary finding to the diminished TGF-β levels is that Nogo-B gene deletion impairs the TGF-β/Smad2 signaling pathway in fibrogenic MF-HSCs and MEFs. We demonstrated that Nogo-B gene deletion decreases the level of Smad2 phosphorylation in response to TGF-β stimulation, whereas Nogo-B gene restoration to knockout MEFs restores TGF-β signaling. Precisely how Nogo-B enhances Smad2 phosphorylation upon TGF-β stimulation is unclear. Because reticulon proteins like Nogo-B play important roles in endoplasmic reticulum (ER) tubular structure and ER-Golgi trafficking,3, 25 it is possible that the antifibrotic effects of Nogo-B gene deletion may be mediated by impaired ER-Golgi trafficking. The presence of Nogo-B may be critical for Smad2 phosphorylation by enhancing translocation of (1) TGF-βRI/II from the cytosol to the plasma membrane and/or (2) Smad2 from the cytosol to the TGF-β RI/II complex at the plasma membrane. Identifying the detailed mechanisms through which Nogo-B mediates Smad2 signaling will help to identify a novel profibrotic pathway.

In addition, Nogo-B may regulate the total levels of Smad2 protein within cells. Despite the fact that Smad2 phosphorylation is impaired in NGB KO MEFs, total Smad2 protein levels were 15-fold higher in NGB KO MEFs than in WT MEFs. This increase, however, was not observed in NGB KO MF-HSCs. This may indicate species-specific or tissue-specific regulation of Smad2 protein levels by Nogo-B.

The transformation of HSCs into MF-HSCs is a key step in hepatic fibrogenesis.9, 22 TGF-β signaling, an inducer of HSC activation, is impaired by Nogo-B gene deletion. Thus, the presence of Nogo-B is critical for maintaining normal TGF-β signaling in HSCs. This is reflected in the attenuated hepatic fibrogenesis in NGB KO mice.

Although we focused on the TGF-β signaling pathway in MF-HSCs, our studies also identified that Nogo-B was expressed by other nonparenchymal cells, including LSECs, cholangiocytes, and Kupffer cells. These other cell types may also be involved in liver fibrogenesis.24, 27 As shown in Supporting Fig. 2, decreased cholangiocyte proliferation in NGB KO cholestatic livers could contribute to decreased fibrosis. It is also possible that the knockout of Nogo-B in LSECs contributes to the reduction of liver fibrosis by decreasing hepatic angiogenesis. Because Nogo-B is instrumental in vascular remodeling and angiogenesis,5 the reduction in hepatic fibrosis may be related to a blunted angiogenic effect.27 Furthermore, Nogo-B expressed in Kupffer cells could play a role in liver fibrosis by generating TGF-β at the site of liver injury. However, we did not observe a noticeable difference in F4/80-positive cells between WT and NGB KO mice 4 weeks after BDL (late stage of fibrosis/cirrhosis) (Supporting Fig. 2). A study by Yu et al.4 showed impaired macrophage infiltration in adductor and gastrocnemius muscles of NGB KO mice 3 days after ischemia. Thus, although the number of Kupffer cells is not much different between these mice at 4 weeks post-BDL, it is possible that infiltration of Kupffer cells may be influenced in the earlier stages of fibrosis/cirrhosis. Although the current study identified Nogo-B's role in TGF-β signaling in fibroblasts, future studies on its effects on nonparenchymal cells may provide additional information on its role in hepatic fibrogenesis.

Nogo-B does not appear to be essential for liver development, because NGB KO mice are viable,5 and we did not observe any structural differences in normal livers between WT and NGB KO mice. However, as the results from this study show, Nogo-B plays an important role in liver fibrosis/cirrhosis. Interestingly, loss of Nogo-B does not inhibit early fibrotic responses, as no difference was observed in fibrosis between WT and NGB KO livers 1 week after BDL. Loss of Nogo-B, however, blocks the progression of fibrosis/cirrhosis 4 weeks after BDL in sharp contrast to WT mice, in which liver fibrosis was progressive at this time point. Thus, Nogo-B appears to be critical for perpetuating liver injury but not for initiating early fibrotic responses.

Finally, a critical role for Nogo-B in hepatic fibrosis was also observed in animals with experimental portal hypertension. Absence of Nogo-B blocked the development of portal hypertension. This likely reflects the blunted fibrotic response in Nogo-B KO mice 4 weeks after BDL. Blocking Nogo-B expression may thus have clinical implications.

In conclusion, this study identifies Nogo-B as a novel regulator of liver fibrosis through activation of the TGF-β/Smad2 signaling pathway. These results indicate that Nogo-B warrants future investigation as a potential therapeutic target in chronic liver diseases.

Acknowledgements

We thank Ms. Nina Sheung and Dr. Jonathan Dranoff for hepatic stellate cell isolation, and Ms. Yuko Ueda for technical support.

Ancillary