Oncostatin M (OSM) is a member of the IL-6 family of cytokines. Mice deficient in the OSM receptor (OSMR-/-) showed impaired liver regeneration with persistent parenchymal necrosis after carbon tetrachloride (CCl4) exposure. The recovery of liver mass from partial hepatectomy was also significantly delayed in OSMR-/- mice. In contrast to wildtype mice, CCl4 administration only marginally induced expression of tissue inhibitor of metalloproteinase (TIMP)-1 and TIMP-2 genes in OSMR-/- mice, correlating with the increased gelatinase activity of matrix metalloproteinase (MMP)-9 and matrix degradation in injured livers. The activation of STAT3 and expression of immediate early genes and cyclins were decreased in OSMR-/- liver, indicating that OSM signaling is required for hepatocyte proliferation and tissue remodeling during liver regeneration. We also found that CCl4 administration in IL-6-/- mice failed to induce OSM expression and that OSM administration in IL-6-/- mice after CCl4 injection induced the expression of cyclin D1 and proliferating cell nuclear antigen, suggesting that OSM is a key mediator of IL-6 in liver regeneration. Consistent with these results, administration of OSM ameliorated liver injury in wildtype mice by preventing hepatocyte apoptosis as well as tissue destruction. In conclusion, OSM and its signaling pathway may provide a useful therapeutic target for liver regeneration. (HEPATOLOGY 2004;39:635–644.)
The liver has a remarkable ability to regenerate in response to liver injury due to various causes such as partial hepatectomy, toxic exposure, or virus infection.1, 2 Liver parenchymal cells, hepatocytes, are normally in the quiescent G0 phase and reenter the cell cycle following injury to restore its mass, architecture, and function quickly. Most of the previous studies have focused on the mechanisms of hepatocyte proliferation after liver injury and a number of growth factors and cytokines have been implicated in regulating these complex processes.3–12 Among them, an important role for IL-6 in liver regeneration has been shown in IL-6 knockout (IL-6-/-) mice.11, 12 In addition, recent studies by using a liver-specific conditional knockout mouse showed that activation of signal transducer and activator of transcription protein 3 (STAT3), a key molecule of IL-6 signaling, was required for the initial step of liver regeneration and for the hepatoprotection in many types of liver injury.13–15
Oncostatin M (OSM) is a member of IL-6 family cytokines that include IL-6, IL-11, leukemia inhibitory factor, ciliary neurotrophic factor, cardiotropin-1, and novel neurotrophin-1/B-cell stimulating factor-3.16–18 These cytokines share the gp130 receptor subunit as a common signal transducer.15 Mouse OSMR is composed of the OSM-specific β subunit and gp130.19 Ligand binding to the receptor complex activates the Janus tyrosine kinases (Jak1, Jak2, and Tyk2) and the activated Jaks in turn activate downstream pathways such as SHP-2 tyrosine phosphatase and STAT3. Although IL-6 family cytokines share the common signal transducer (gp130), we previously reported various OSM-specific biological activities during mouse development, e.g., stimulation of production of definitive hematopoietic progenitors in a primary culture of the aorta-gonad-mesonephros region of embryonic day 11.5 (E11.5) mouse embryo20 and stimulation of proliferation of Sertoli cells in neonatal testes.21 We also demonstrated that OSM specifically promoted differentiation of fetal hepatocytes, as evidenced by expression of several metabolic enzymes, accumulation of glycogen and lipid, ammonia clearance, and morphological maturation in primary culture of fetal hepatocytes.18, 22–24 These biological activities are specific to OSM among IL-6 family members. In adult liver, OSM stimulates the expression of acute phase proteins25 and tissue inhibitor of metalloproteinase-1 (TIMP-1) in vitro26–29 and in vivo.30 It is well known that TIMP-1 plays an important role in remodeling of extracellular matrix during liver regeneration31, 32 and that the expression of TIMP-1 is most strongly induced by OSM among the IL-6 family cytokines in several types of cells.27–29
We have recently generated OSMR-deficient mice and found that the mutant mice are viable and fertile, while hematopoiesis is compromised.33 Thus, despite the fact that OSM strongly stimulates differentiation of fetal hepatocytes in vitro, OSM is dispensable for normal liver development. However, the role of the OSM/OSMR system in liver injury and regeneration remains unclear. In this article, we demonstrate that OSMR-/- mice exhibit impaired liver regeneration, i.e., delayed hepatocyte proliferation, persistent liver necrosis, and increased tissue destruction. We also show that administration of OSM ameliorates CCl4-induced acute liver failure. Furthermore, we found that the injection of CCl4 in IL-6-/- mice failed to induce OSM and that OSM administration in IL-6-/- mice recovered the regenerative liver response, suggesting that OSM is a key downstream mediator of IL-6 in liver regeneration.
OSM, oncostatin M; CCl4, carbon tetrachloride; NPCs, nonparenchymal cells; PCs, parenchymal cells; OSMR-/-, OSMR null; PCNA, proliferating cell nuclear antigen; STAT3, signal transducer and activator of transcription protein 3; TIMP, tissue inhibitor of metalloproteinase; MMP, matrix metalloproteinase.
Materials and Methods
All experiments were performed using males of wildtype (WT) C57/BL6 mice (Nihon SLC, Hamamatsu, Japan), IL-6-/- mice (Jackson Laboratory, Bar Harbor, ME; strain name B6. 129S6-Il6[tm1kopf]) and OSMR-/- mice (OSMR-/- mice was originally generated in the C57BL6/SV129 mix background and were bred with C57/BL6 for 8 generations) at the age of 8–12 weeks. OSMR-/- mice develop normally and display some alteration in the hematopoietic system.33 OSMR-/- mice were maintained by mating homozygous siblings. All animal studies were performed according to the institutional guideline, which is in agreement with the National Institutes of Health guidelines.
CCl4 Injury and Partial Hepatectomy.
CCl4 liver injury was induced by i.p. injection of 20% (v/v) solution of CCl4 (Wako Pure Chemicals, Osaka, Japan) in olive oil at a dose of 7 μl/g body weight. Three to six mice in each cohort were sacrificed at indicated times after injury. For partial hepatectomy, mice were subjected to conventional 70% partial hepatectomy under pentobarbital anesthesia.34 At the time of hepatectomy, the weight of the resected liver was measured and this value was used for the estimation of the initial liver weight (the resected liver = 0.7 × total liver). The activities of serum aminotransferase were measured with a commercially available kit (GPT-UV test wako and GOT-UV test wako; Wako Pure Chemicals) according to the manufacturer's instructions.
Administration of Recombinant Mouse OSM.
In the studies of the effects of administration of OSM on CCl4-induced acute liver injury, WT mice were subcutaneously (s.c.) injected with recombinant mouse OSM (R&D Systems, Minneapolis, MN) at a dose of 1 mg/kg of body weight at 20 minutes before and at 8, 16, and 24 hours after CCl4 injection. In rescue experiments, OSM (300 μg/kg of body weight) was s.c. injected at 1 hour after CCl4 injection in IL-6-/- mice. PBS was injected in the same way as a control.
Liver Cell Fractionation and Flow Cytometry.
Liver cells were isolated by two-step collagenase perfusion method.35 Cell suspension was centrifuged at 500 rpm for 1 minute. Nonparenchymal cells (NPCs) were prepared from the supernatant and further fractionated by magnetic cell sorting (MACS). Briefly, NPCs were incubated with FITC-conjugated anti-CD45 antibody (BD Pharmingen, San Diego, CA) and then incubated with anti-FITC MicroBeads (Miltenyi Biotec, Bergisch Gladbach, Germany). Cell separation was performed by autoMACS (Miltenyi Biotec). Flow cytometric analysis of OSMR expression in NPCs was performed by FACScalibur (Becton Dickinson, San Jose, CA). Monoclonal antibody against mouse OSMRβ was purchased from MBL (Code No. D059-3; Nagoya, Japan).
RNA Extraction, Northern Blot Analysis, and RT-PCR.
Total RNA was prepared using TRIZOL (Invitrogen, Carlsbad, CA). For Northern blot, 10 μg of total RNA was denatured with 18% formaldehyde, separated by electrophoresis on 1% agarose, and transferred to a nylon membrane. Membranes were hybridized with digoxigenin (DIG)-labeled cDNA probes and further incubated with alkaline phosphatase-labeled anti-DIG Ab (Roche Diagnostics, Indianapolis, IL). Blots were developed with CDP-star (New England Biolabs, Beverly, MA). For RT-PCR analysis, first-strand cDNA was synthesized by using the First-Strand cDNA Synthesis Kit (Amersham Biosciences, Piscataway, NJ). Synthesized cDNA samples were used as templates for PCR amplification for OSM, OSMR, and GAPDH. Quantitative PCR analysis was also performed to measure the mRNA levels for OSM and GAPDH by using real-time PCR (LightCycler; Roche Diagnostics). PCR primers used in this study were as follows, OSM (5′-TCCGCCTCCAAAACCTGAACAC-3′ and 5′-ATGGTATCCCCAGAGAAAGC-3′), OSMR (5′-ATCCAAAGGCTCCGCAGGAC-3′ and 5′-GTAAGGTTGCAGGTCAAGGC-3′), GAPDH (5′-ACCACAGTCCATGCCATCAC-3′ and 5′-TCCACC-ACCCTGTTGCTGTA-3′).
Cryostat sections (7 μm) were stained with hematoxylin and eosin with the standard protocol. For proliferating cell nuclear antigen (PCNA) staining, sections were incubated with anti-PCNA mAb (DakoCytomation, Copenhagen, Denmark) followed by biotin-conjugated anti-mouse IgG. Immunoreactive nuclei for PCNA were then visualized by using the Vectastain ABC Elite kit (Vector Laboratories, Burlingame, CA). Apoptotic hepatocytes were visualized by TUNEL staining (Roche Diagnostics).
Liver extracts were separated by SDS-PAGE under the reducing condition and transferred to a polyvinylidene difluoride membrane (Millipore, Bedford, MA). The membranes were incubated with first antibodies followed by HRP-conjugated anti-rabbit (or mouse) IgG and detected with enhanced chemiluminescence (Amersham Biosciences). Antibodies used in this study were: STAT3 (cat. 9132; Cell Signaling), phospho-STAT3 (cat. 9131; Cell Signaling), cyclin D1 (M-20; Santa Cruz), PCNA (PC-10; DakoCytomation), α-tubulin (H-300; Santa Cruz).
NIH image 1.62 (National Institutes of Health) was used to measure densities of blots. Densities of each blot were normalized by the intensity of STAT3 or α-tubulin (immunoblot) or by the intensity of GAPDH (Northern blot) in the same sample.
Liver extracts (80 μg) were separated by 10% SDS-PAGE containing 1 mg/ml of gelatin (Zymogram gels; Invitrogen). Zymographic analysis was performed according to the manufacturer's instruction. In situ gelatin zymography was performed using a gelatin-coated polyester film (MMP in situ zymofilm; Wako Pure Chemical) according to the manufacturer's instruction.
Expression of OSM and OSMR During Liver Regeneration.
It is well known that carbon tetrachloride (CCl4) induces free radical-mediated lipid peroxidation in liver, which results in release of liver enzymes prior to centrilobular necrosis in the liver.36 While expression of both OSM and OSMR mRNA was barely detectable in normal liver, they were rapidly induced within 1 hour following CCl4 administration (Fig. 1A). Next, we fractionated liver cells into parenchymal cells (hepatocytes) and nonparenchymal cells (NPCs) by conventional two-step collagenase perfusion in combination with low-speed centrifugation at various times after injury and found that the expression of OSM was specific to NPCs (Fig. 1B). We further fractionated NPCs at 48 hours after CCl4 administration based on the expression of CD45, a pan-specific hematopoietic marker, by magnetic cell sorting (MACS) using anti-CD45 mAb. Fluorescein-activated cell sorting (FACS) analysis indicated that the purity of each cell fraction was more than 90% and that about 80% of CD45+ cells expressed Mac-1, a macrophage marker (unpubl. data). RT-PCR analysis showed that expression of OSM in NPCs was specific to CD45+ cells (Fig. 1C), suggesting that macrophages (Kupffer cells) are a major source of OSM in the injured liver. In contrast, the expression of OSMR was detected in all types of liver cells, i.e., CD45+ NPCs, CD45- NPCs and hepatocytes (Fig. 1B–D). FACS analysis showed that most of CD45- cells and one-third of CD45+ cells expressed OSMR (Fig. 1D). We reproducibly observed the induction of OSMR mRNA in hepatocytes after CCl4 exposure, while expression of OSMR detected by FACS was unchanged in NPCs during liver injury (data not shown).
Enhanced Liver Injury in OSMR-/- Mice Following CCl4 Treatment and Anti-hepatitis Activity of OSM.
To test the possibility that OSMR-mediated signaling is involved in liver regeneration, we analyzed regenerative responses following CCl4-induced acute liver injury in OSMR-/- mice. Mortality following a single injection of 50 μl of CCl4 was 3.7% in WT mice (n = 27) and 23.1% in OSMR-/- mice (n = 26). WT livers showed a centrilobular necrosis at 48 hours, which was almost completely recovered at 72 hours after CCl4 treatment (Fig. 2A–D). In contrast, OSMR-/- liver at 72 hours after CCl4 treatment exhibited extensive parenchymal necrosis throughout the liver and the injury was persistent up to 5 days after CCl4 administration, when the WT liver had completely recovered from the injury (Fig. 2E–H, and data not shown). Significant fat deposition was also observed throughout OSMR-/- liver parenchyma at 48 hours after the injury (Fig. 2C,G), suggesting that the liver damage caused by CCl4 was much more severe in the OSMR-/- mice. Consistent with these results, although both aspartate aminotransferase (AST) and alanine aminotransferase (ALT) activities in the serum, hallmarks of liver injury, were increased until 48 hours after CCl4 treatment in both cohorts, the recovery form the injury was significantly delayed in OSMR-/- mice (Fig. 3A,B). Enhanced liver injury in OSMR-/- mouse suggested that administration of OSM might reduce CCl4-induced injury. As expected, administration of OSM in WT mice significantly reduced centrilobular necrosis and serum injury markers 48 hours after CCl4 injection (Figs. 2I,J and 3C,D). As it is known that CCl4 induces hepatocyte apoptosis as well as necrosis,12, 35 we examined whether administration of OSM prevents hepatocyte apoptosis by the terminal deoxynucleotidyl transferase-mediated dUTP biotin nick end labeling (TUNEL) reaction. The number of TUNEL-positive hepatocytes in OSM-treated livers was significantly low, i.e., ∼20% of control (PBS-treated) livers (Fig. 3E).
Impaired Hepatocyte Proliferation Induced by CCl4 Treatment or Hepatectomy in OSMR-/- Livers.
Immunohistochemical staining of proliferating cell nuclear antigen (PCNA), a marker for G1/S phase of cell cycle, was used to determine whether the absence of OSMR caused abnormal hepatocyte proliferation after injury.38, 39 No PCNA-positive hepatocytes were observed in WT and OSMR-/- livers before CCl4 treatment, indicating that hepatocytes were in the G0 stage in the steady state. Administration of CCl4 resulted in the appearance of many PCNA-positive hepatocytes at 48 hours after injury in WT mice and PCNA disappeared thereafter (Fig. 4A,B). In contrast, the number of PCNA-positive hepatocytes in OSMR-/- livers was ∼20% of that in WT mice at 48 hours and reached the peak at 60 hours after injury, indicating that proliferation of hepatocytes was significantly delayed in OSMR-/- mice.
We next examined the recovery of liver mass after 70% hepatectomy (Fig. 4C). In WT mice, remnant liver restored the original mass almost completely at 7 days after hepatectomy. In contrast, the restoration of liver mass was significantly delayed in OSMR-/- livers. Impaired liver regeneration in OSMR-/- mice was further demonstrated by expression of cyclin genes and immediate early genes (IEGs) (Fig. 5A–C). In WT livers, expression of cyclin D1, which is induced in the G1 phase and promotes the G1/S transition, was strongly induced at 48 hours and remained high until 72 hours after injury, whereas expression of cyclin D1 in OSMR-/- livers at 48 hours was barely detectable. Although expression of cyclin D1 was observed at 60 and 72 hours in OSMR-/- liver after CCl4 administration, the expression level was lower than that in WT liver. We also found similar expression patterns of cyclin D2 and cyclin D3 after injury in OSMR-/- livers. Consistent with the abnormality of G1 phase, expression of other cyclin genes, cyclin E (G1/S), cyclin A2 (S), and cyclin B1 (G2/M), was also delayed and decreased in OSMR-/- livers. IEGs including c-fos, c-jun, and junB, and STAT3 are normally expressed during the pre-replicative phase of regeneration.31, 40 However, expression of those genes was significantly reduced in OSMR-/- liver (Fig. 5B,C). As it is well established that AP-1 plays an important role in triggering proliferation of hepatocytes after hepatectomy or CCl4 injury,31 decreased expression of these IEGs may lead to alteration of the G1 phase in OSMR-/- livers.
It is believed that rapid activation of STAT3 after injury is important for G0/G1 phase transition of hepatocytes.13, 31, 40 Recent works also demonstrated that STAT3 activation is important for hepatoprotection against a variety of hepatic injuries.14, 15 Western blotting with anti-phosphorylated STAT3-specific Ab revealed that the level of phosphorylated STAT3 in WT liver was significantly increased at 2 hours after the injury and remained high until 4 hours (Fig. 5D), consistent with the previous report that STAT3 DNA binding activity measured by the electrophoretic mobility shift assay was increased in injured liver.11, 12 Compared to WT livers, phosphorylation of STAT3 was significantly reduced in OSMR-/- livers after the injury. Densitometric analysis showed that the relative level of activated STAT3 (P-STAT3/STAT3) was 34.3% at 2 hours and 36.7% at 4 hours in OSMR-/- livers compared to WT (Fig. 5E).
Altered Expression and Activities of TIMPs and MMP-9 in OSMR-/- Livers.
Remodeling of extracellular matrix (ECM) occurs following liver injury and is strictly regulated by matrix metalloproteinases (MMPs) and tissue inhibitor of matrix metalloproteinases (TIMPs).31, 32, 41 While expression of TIMP-1 and TIMP-2 genes in liver was undetectable under normal conditions, it was strongly induced at 48 hours after injury and decreased thereafter in WT liver (Fig. 6A). In contrast, the expression of TIMP-1 by the injury was almost undetectable in OSMR-/- livers. Upregulation of TIMP-2 by the injury was also suppressed in OSMR-/- livers as well. Consistent with this result, administration of OSM in WT mice strongly enhanced TIMP-1 expression compared to PBS-treated livers (Fig. 6B). On the other hand, expression of TIMP-2 was not induced by the administration of OSM (Fig. 6B), suggesting that the low level expression of TIMP-2 in CCl4-treated OSMR-/- liver was a secondary effect by the lack of OSM function.
We then analyzed the expression and activities of MMP-2 and MMP-9 by gelatin zymography analysis (Fig. 6C). Gelatinase activity in liver without injury was associated with the proteins corresponding to pro-MMP-2 and pro-MMP-9, whereas no gelatinase activity was found to be associated with the active forms of MMP-2 and MMP-9 in both WT and OSMR-/- livers. In WT livers, gelatinase activity corresponding to both pro-MMPs was increased in response to the injury and active forms of both MMPs were detected from 24–96 hours after the injury. Induction of MMP-9 activity by injury was much higher than that of MMP-2. Significantly, gelatinase activity corresponding to both pro-MMP-9 and active MMP-9 in OSMR-/- liver was much stronger than in WT livers, while MMP-2 activity in OSMR-/- liver was similar to that in WT liver. In addition, in situ zymography of liver sections of OSMR-/- mice showed elevated gelatinase activity throughout the liver compared to WT mice (Fig. 6D,E). Elevation of MMP-9 and suppression of TIMP-1 and TIMP-2 after the injury were likely to enhance matrix degradation in OSMR-/- livers. In contrast, in situ gelatinase activity in OSM-treated livers was significantly lower than control livers (Fig. 6F,G).
OSM Is a Downstream Mediator of IL-6 in Liver Regeneration.
The phenotypes of OSMR-/- mice described above were quite similar to those of IL-6-/- mice.11, 12 To investigate the relationship between IL-6 and OSM, we analyzed the expression of OSM after CCl4 exposure by real-time PCR. Expression of OSM by CCl4 was biphasic and was reduced in IL-6-/- mice. Especially, the first peak of OSM expression at 2 hours after CCl4 exposure was almost completely diminished in IL-6-/- mice (Fig. 7A). On the other hand, there was no difference in the levels of mRNA and serum concentration of IL-6 between WT and OSMR-/- mice (data not shown). These results suggest that OSM is downstream of IL-6 following CCl4 exposure. We then administered OSM in IL-6-/- mice to test if it recovers the regenerative response. Administration of OSM (300 μg/kg body weight) at 1 hour after CCl4 injection induced the phosphorylation of STAT3 at 2 and 4 hours in IL-6-/- liver (Fig. 7B). The kinetics of STAT3 activation by OSM in IL-6-/- mice was the same as WT (Fig. 7C). We then examined protein levels of cyclin D1 and PCNA by Western blot. In IL-6-/- livers, the levels of cyclin D1 and PCNA at 48 and 72 hours after CCl4 injection were significantly decreased compared with WT livers. The administration of OSM at 1 hour after CCl4 exposure reverted the protein levels of cyclin D1 and PCNA completely (Fig. 7D,E).
Mice deficient in OSMR showed extensive liver injury by CCl4 and parenchymal degeneration was persistent for 5 days after CCl4 administration. TNF-α has been known to enhance CCl4-induced liver injury, since treatment of mice with soluble TNF-α receptor ameliorated liver injury significantly.42 TNF-α induces not only necrosis but also apoptosis in hepatocytes. In addition, TNF-α also stimulates production and secretion of MMP-9 and enhances tissue destruction.43, 44 It was reported previously that administration of OSM suppressed LPS-induced TNF-α production and lethality in mice.45 Consistently, the serum TNF-α level in OSMR-/- after the injury was higher than that in WT, and administration of OSM decreased the CCl4-induced serum TNF-α level (data not shown). These results suggest that the enhanced liver injury in OSMR-/- mice is caused by the increased level of serum TNF-α and that OSM may be a negative regulator of TNF-α during liver injury. In fact, a significant level of OSMR expression was detected in CD45+ NPCs, including Kupffer cells (Fig. 1D), a major source of TNF-α. In addition, a number of studies have shown that STAT3 activated by IL-6 exhibited antiapoptotic effects through the induction of Bcl-2, Bcl-xL, and FLICE inhibitor protein (FLIP), which in turn inhibit the activities of FLICE and caspase-3 in hepatocytes.12, 46, 47 Adenovirus-mediated expression of an active form of STAT3 also induced the expression of redox-associated protein, redox factor-1, and protected Fas-mediated apoptosis in the liver.15 Administration of OSM also reduced significantly TUNEL-positive apoptotic hepatocytes after CCl4 administration (Fig. 3E). Therefore, increased apoptosis in OSMR-/- as well as IL-6-/- livers could be due to reduced activation of STAT3 in their livers. However, Wuestefeld et al.48 reported that hepatectomy induced no STAT3 activation and delayed expression of cyclin A and cyclin E in gp130 conditional knockout mice, but DNA synthesis in the mice was normal after hepatectomy. The reason for this apparent discrepancy between these results is not clear. As the gp130 gene was deleted conditionally by pIpC-induced expression of Cre, pIpC that induces interferon might have altered gene expression in addition to Cre, which could result in the phenotype described. Alternatively, genetic background could make the difference: the conditional gp130-/- mice were generated in NMRI background,48 while our OSMR-/- as well as IL-6-/-11, 12 mice were C57/BL6.
CCl4-treated OSMR-/- mice showed high mortality and decreased PCNA-positive hepatocytes, indicative of impaired G1/S transition after the injury (Fig. 4A,B). The number of PCNA-positive cells reached a peak at 48 hours after injury in WT livers, while the peak shifted to 60 hours in OSMR-/- liver (Fig. 4B). Suppression of STAT3 activation and reduced expression of IEGs and cyclin D1 after the injury in OSMR-/- mice likely induce the alteration of early G1 phase in mutant livers. Consistent with these observations, restoration of liver mass after 70% hepatectomy was delayed in OSMR-/- livers (Fig. 4C). Especially, the recovery rate during the initial 4 days was markedly reduced in OSMR-/- livers, while it was comparable to WT livers after 4 days. Although we did not examine the direct effects of OSM on the cell cycle progression, these results suggest that OSMR-mediated signaling is required for the early proliferative response of hepatocytes to hepatectomy as well as toxic injury.
The defects in liver regeneration in OSMR-/- mice are similar to the phenotypes that have been shown in IL-6-/- mice.11, 12 Because both OSM and IL-6 are cytokines that share the gp130 receptor subunit as the common signal transducer and are produced in Kupffer cells in response to injury, there may be functional redundancy between OSM and IL-6 in liver. However, as deletion of either one of these cytokine functions by gene targeting resulted in a similar phenotype, one possibility is that OSM and IL-6 are on the same pathway, i.e., one induces the other. In fact, it was reported that expression and/or release of IL-6 was induced by OSM in some kinds of cells in vitro49, 50 and in vivo.30 However, the expression of IL-6 following CCl4 exposure was not altered in OSMR-/- livers and no difference was found in the serum IL-6 levels after CCl4 exposure between WT and OSMR-/- mice (data not shown), suggesting that expression of IL-6 is independent of OSM in regenerating livers. In addition, expression of IL-6R and gp130 was also unaffected in OSMR-/- livers (data not shown), indicating that IL-6 is not downstream of OSM in liver regeneration. Conversely, evidence indicates that OSM is downstream of IL-6. Real-time PCR analysis showed that expression of OSM after CCl4 exposure was greatly decreased in IL-6-/- mice (Fig. 7A), suggesting that OSM is a downstream mediator of IL-6 in acute liver injury. Administration of OSM in IL-6-/- mice at 1 hour after CCl4 exposure resulted in phosphorylation of STAT3 with the normal kinetics and degree and in the induction of cyclin D1 and PCNA (Fig. 7). These results strongly support our hypothesis. The level of the STAT3 phosphorylation in OSMR-/- livers was 34% of WT at 2 hours after the injury (Fig. 5E), while that in IL-6-/- mice was 16% (Fig. 7C). Strong suppression of STAT3 phosphorylation in IL-6-/- mice may be due, in part, to the decreased expression of OSM in IL-6-/- mice by CCl4 exposure. The present results suggest that IL-6 induces the expression of OSM, and then both IL-6 and OSM induce full phosphorylation of STAT3 cooperatively. In addition, as OSMR was expressed in NPCs as well as hepatocytes (Fig. 1), the suppression of STAT3 phosphorylation in OSMR-/- livers may reflect the decreased responsiveness of NPCs to OSM.
MMPs and TIMPs are thought to play an essential role for tissue remodeling and repair during liver injury. It has also been reported that ECM degradation by MMPs during the early phase of liver regeneration was essential for the proliferation of hepatocytes.31, 32, 41 Several reports demonstrated that OSM induced the expression of TIMP-1 more effectively than any other members of the IL-6 family cytokines in several cells27–29 and that adenovirus-mediated OSM expression induced the expression of TIMP-1 in the liver.30 Consistent with these reports, we showed that TIMP-1 expression in the liver was strongly induced by OSM administration (Fig. 6B). Although it is well known that TGF-β is responsible for the induction of TIMP-1 during acute liver injury,31, 32 the expression of TIMP-1 following CCl4 injury was barely detectable in OSMR-/- livers (Fig. 6A). These results suggest that OSM is a major regulator of TIMP-1 expression and that target cell types are different between IL-6 and OSM during liver injury. In fact, OSMR was highly expressed in NPCs and the expression of TIMP-1 during injury was mostly restricted to NPCs.41 Although MMP-9 activity after the CCl4 administration was enhanced in OSMR-/- livers, OSM failed to affect the expression of MMP-9 in primary cultures of both PCs and NPCs (data not shown). Therefore, MMP-9 activity appears to be indirectly upregulated in OSMR-/- mice. Because TNF-α was shown to induce MMP-9 during liver injury and the serum TNF-α level in OSMR-/- mice was higher than in WT mice, the enhanced MMP-9 activity in OSMR-/- liver may be caused by the increased TNF-α level. Our results strongly suggested that enhanced MMP-9 activity and suppression of TIMP-1 and TIMP-2 expression accelerated ECM degradation during acute liver injury in OSMR-/- mice. In a fulminant hepatitis model in mouse, TNF-α in combination with GalN induced remarkable MMP-9 production, and pretreatment of mice with BB-94, a wide-range inhibitor of MMPs, suppressed lethal hepatitis.43 Adenovirus-mediated expression of TIMP-1, an endogenous inhibitor of MMPs, decreased tissue damage in the joint of TNF-α transgenic mice.44 As administration of OSM induced the expression of TIMP-1 and blocked ECM degradation in the liver (Fig. 6B,F,G), anti-acute hepatitis activity of OSM may be due to the induction of TIMP-1 in the liver in addition to inhibition of TNF-α production and STAT3-mediated antiapoptosis.
Taken together, the present results provide evidence for an essential role of OSMR-mediated signaling in survival and proliferation of hepatocytes, as well as tissue remodeling via the expression of TIMP-1 in NPCs during liver regeneration. Furthermore, the present results also indicate that OSM acts as a key downstream mediator of IL-6 in acute liver injury.
We thank the members of the Stem Cell Regulation project of the Kanagawa Academy of Science and Technology for helpful discussion.