Osteopontin: Versatile modulator of liver diseases
Department of Gastroenterology and Hepatology, Saitama Medical Center, Saitama Medical University, Saitama, Japan
Correspondence: Professor Sumiko Nagoshi, Department of Gastroenterology and Hepatology, Saitama Medical Center, Saitama Medical University, 1981 Kamoda, Kawagoe-shi, Saitama, 350-8550, Japan. Email: email@example.com
Osteopontin (OPN) is a multifunctional protein, involved in pathological conditions including inflammation, immunity, angiogenesis, fibrosis and cancer progression in various tissues. Hepatic inflammation and fibrosis induced by feeding with a diet deficient in methionine and choline (MCD diet) were markedly attenuated in OPN knockout mice when compared with wild-type mice in the model of non-alcoholic steatohepatitis (NASH). Hepatic cholangiocytes, myofibroblastic stellate cells and natural killer T cells were suggested to secret OPN in mice fed an MCD diet. Plasma and hepatic OPN levels were significantly higher in patients with NASH with advanced fibrosis than in those with early fibrosis. Hepatic OPN mRNA level was correlated with hepatic neutrophil infiltration and fibrosis in patients with alcoholic liver diseases. In those with hepatocellular carcinoma (HCC), OPN levels in plasma and HCC were prognostic factors after liver resection or transplantation. Downregulation of OPN inhibited tumor growth and lung metastasis in nude mice implanted with HCC cells. The single nucleotide polymorphism in the promoter region of the OPN gene was shown to be associated with activity of hepatitis in chronic hepatitis C patients, prognosis in patients with HCC, and growth and lung metastasis of HCC xenografts in nude mice. OPN was reported to be a downstream effecter of Hedgehog pathway, which modulates hepatic fibrosis and carcinogenesis. This review focuses on the roles of OPN in hepatic inflammation, fibrosis and cancer progression. Further elucidation of cellular interactions and molecular mechanisms associated with OPN actions may contribute to development of novel strategies for treatment of the liver diseases.
OSTEOPONTIN (OPN) WAS first described as a phosphoprotein secreted by a transformed cell line in 1979. Several years later, the bone-specific sialoprotein was cloned as a matrix protein and termed “osteopontin”. OPN was independently identified as early T-lymphocyte activation 1 (Eta-1), which was abundantly expressed in activated T cells. Thereafter, it has been reported that OPN has important roles in the development of various inflammatory conditions. OPN was also shown to act as a cytokine essential for the initiation of T-helper 1 immune responses in mice.
Osteopontin is physiologically expressed in the kidney and bone, while OPN expressions are found in various organs under pathological conditions. Hepatic expression of OPN was first confirmed in the activated Kupffer cells, macrophages and stellate cells in the inflammatory and necrotic areas in rats after carbon tetrachloride intoxication. OPN was shown to contribute to the migration of macrophages into the lesions.[5, 6] Recent reports suggested that plasma OPN levels were predictive of liver fibrosis in patients with chronic hepatitis B, chronic hepatitis C, alcoholic liver disease and non-alcoholic steatohepatitis (NASH). Furthermore, it has been reported that OPN was expressed in various cancers, including hepatocellular carcinoma (HCC),[11-13] and played important roles in growth, invasion and metastasis of cancer, angiogenesis and inhibition of apoptosis.[14, 15]
Structure, Receptors and Modifications of OPN
OSTEOPONTIN IS COMPOSED of approximately 300 amino acids with a molecular mass ranging 40–80 kD due to varied post-translational modifications such as glycosylation, phosphorylation, sulfation and enzymatic cleavage. OPN contains a classical cell-binding motif, an arginine-glycine-asparate (RGD) domain, which binds to the cell surface RGD-recognizing integrins such as αvβ1, αvβ3 and α5β1. Next to the RGD domain, OPN is cleaved by proteases including thrombin and plasmin. The serine-valine-valine-tyrosine-glycine-leucine-arginine (SVVYGLR) domain in humans and serine-leucine-alanine-tyrosine-glycine-leucine-arginine (SLAYGLR) domain in mice and rats, require cleavage by thrombin to be recognized by non-RGD-recognizing integrins such as α4β1 and α9β1.[17-19] OPN is further cleaved at a position within the SVVYGLR domain, by matrix metalloproteinases (MMP), such as MMP-3 and MMP-7. OPN also binds to the spliced variant form of CD44 (CD44v), but a precise binding site has not been elucidated. Different forms of OPN protein can exert distinct biological functions.
There are two isoforms of OPN, a secreted form of OPN (sOPN) and an intracellular form of OPN (iOPN) (Fig. 1). sOPN staining had perinuclear distribution which appeared in Golgi, and iOPN staining had perimembrane distribution. sOPN is secreted through the endoplasmic reticulum and Golgi, and exerts its effects by binding to the cell surface receptors. On the other hand, iOPN is co-localized with the CD44-ezrin-radixin-moesin (CD44-ERM) complex that played a role in cell motility.[22-24] iOPN is also involved in signal transduction pathways of innate immune receptors, such as Toll-like receptors, and is translocated into the nucleus during mitosis.
Kupffer Cells and Macrophages in Injured Rat Liver
IN CARBON TETRACHLORIDE-INTOXICATED rats, activated Kupffer cells and macrophages increased in number in the necrotic areas. OPN mRNA expression was prominently higher in Kupffer cells, hepatic macrophages and stellate cells immediately after isolation from these intoxicated rats than in the cells from normal rats. In heat-killed Propionibacterium acnes-treated rats, marked accumulation of macrophages developed in the liver later than 3 days after the intoxication and its extent became maximal between 5 and 7 days. In these rats, increased OPN mRNA expression in the liver occurred at 1 day with its peak at 3 days, and OPN mRNA expression in Kupffer cells and hepatic macrophages isolated at 7 days was extremely increased. Chemotaxis assay, using a cell culture chamber precoated with OPN, showed that OPN promoted migration of Kupffer cells isolated from normal rats in a dose-related manner. These results suggested that sOPN secreted by Kupffer cells and hepatic macrophages at inflammatory sites, contributed to macrophage recruitment into the sites of liver injury (Fig. 1). Furthermore, iOPN co-localized with the CD44-ERM complex may promote migration of macrophages.
OPN in Non-Alcoholic Fatty Liver Disease (Nafld)
OPN in patients with NAFLD
NON-ALCOHOLIC FATTY LIVER disease is characterized by accumulation of excess hepatocellular triglyceride, which is associated with obesity and insulin resistance. Hepatic expressions of OPN and its receptor, CD44 mRNA, were significantly correlated with the grade of hepatic steatosis, plasma alanine aminotransferase level and hepatic insulin resistance in morbidly obese patients. NASH, a subset of NAFLD, shows hepatic necroinflammatory changes and often progresses to liver fibrosis and cirrhosis. Hepatic OPN mRNA expression was significantly increased in patients with NAFLD than in controls with normal liver, while the expression was significantly decreased in those with NASH compared to those without NASH. Among patients with NASH, immunohistochemical expression of OPN was significantly greater in the liver with advanced fibrosis than in that with early fibrosis. Plasma OPN level was also significantly higher in patients with NASH with advanced fibrosis than in those with early fibrosis.
OPN roles in hepatic steatosis and inflammation in mouse models of NAFLD
In obese mice fed a high-fat diet, OPN deficiency protected against hepatic steatosis and inflammation.[29, 30] Decreased steatosis in OPN–/– mice fed a high-fat diet was paralleled by improved whole-body insulin sensitivity, mainly due to reduction in hepatic insulin resistance and gluconeogenesis. In addition, hepatocyte ballooning, portal leukocyte infiltration and hepatic macrophage accumulation were attenuated by genetic OPN deficiency. Antibody-mediated OPN neutralization also inhibited accumulation of hepatic macrophages and insulin resistance, induced by a high-fat diet in mice (Fig. 2).
In mice fed with a diet deficient in methionine and choline (MCD diet) for up to 12 weeks, hepatic steatosis started by 1 week and inflammation appeared at 1 week and progressed to significant hepatitis by 4 weeks, and fibrosis appeared at 8 weeks. In these mice, hepatic expression of OPN protein was markedly increased at 1 day after the beginning of MCD diet and persisted up to 8 weeks, whereas OPN mRNA expression was increased at 4 weeks. OPN protein expression was predominantly localized to hepatocytes, not inflammatory cells, assessed by immunohistochemistry at 3 days and at 8 weeks. Moreover, hepatic inflammation induced by MCD diet was markedly reduced in OPN–/– mice compared with OPN+/+ mice, while histological steatosis and liver triglyceride levels were similar among these mice. This may be because mice fed MCD diet lose weight and do not show insulin resistance, unlike human or other diet-induced rodent models of NAFLD.
OPN roles in fibrosis progression in mouse models of NAFLD
Osteopontin mRNA was increased during culture-related activation of rat quiescent hepatic stellate cells to myofibroblastic stellate cells. Furthermore, incubation of hepatic stellate cells with OPN induced their collagen production, transforming growth factor-β receptor upregulation, proliferation and migration. These results suggested a potential role for OPN in the progression of hepatic fibrosis.
Hepatic fibrosis induced by MCD diet was attenuated in OPN–/– mice compared with OPN+/+ mice.[31, 33] Moreover, hepatic OPN mRNA and protein levels were not affected by feeding an MCD diet in genetically leptin-deficient, obese and diabetic ob/ob mice, which developed steatohepatitis but not liver fibrosis after the feeding. In contrast, the diet had a stimulatory effect on OPN mRNA and protein levels in hyperleptinemic, obese and diabetic db/db mice, which exhibited a lesser degree of steatosis, but greater histological inflammation and marked pericellular fibrosis by the diet.
Recently, it was reported that OPN was induced by Hedgehog signaling and directly promoted profibrogenic responses in steatohepatitis (Fig. 2). Hedgehog pathway can promote activation of quiescent hepatic stellate cells to myofibroblastic stellate cells. In patients with NAFLD, accumulation of Hedgehog ligands and expression of Hedgehog-target genes were significantly correlated with hepatic fibrosis stage. Furthermore, Hedgehog-mediated accumulation of natural killer T (NKT) cells contributed to fibrosis progression of NASH in mice and humans. As shown in Figure 3, after the binding of the Hedgehog ligand to the Patched (Ptc) receptor, glioma-associated oncogene (Gli) is activated by release from a large protein complex and translocated to the nucleus to function as a transcriptional activator. The consensus DNA-binding sequences of Gli-1 were indentified in the 5′ regions of OPN, and gel shift analysis confirmed Gli-1 protein could bind to the oligonucleotides of OPN promoter region. Syn et al. analyzed hepatic OPN expression and fibrosis grade, using Ptc-deficient (Ptc+/−) mice with haploinsufficiency of Ptc, which exhibit overly active Hedgehog signaling. Ptc+/− mice expressed more OPN mRNA and protein, and developed more severe fibrosis in the liver compared with wild-type mice after feeding an MCD diet. OPN-immunoreactive cells were mostly bile duct cells in both Ptc+/− and wild-type mice. Hepatic stellate cells isolated from Ptc+/− mice expressed higher mRNA levels of Gli-2, OPN, collagen and α-smooth muscle actin (α-SMA) compared with the cells from wild-type mice. Neutralizing OPN with RNA aptamers significantly reduced collagen and α-SMA expressions, but had little effect on Gli-2 expression in stellate cells from Ptc+/− mice. Furthermore, in patients with NASH, ballooned hepatocytes produced Hedgehog ligands and were surrounded by Gli-2 positive stromal cells expressing myofibroblastic markers. These findings suggested that OPN induced by Hedgehog pathway activation, promoted fibrogenic responses in NASH.
It was reported that NKT cells could promote liver fibrogenesis by producing profibrotic cytokines such as Hedgehog ligands, OPN, interleukin (IL)-4 and IL-13. Mice genetically deficient in NKT cells developed significantly less hepatic fibrosis and liver injury, with significantly reduced hepatic and plasma OPN levels compared to wild-type mice after feeding with MCD diet. Activated NKT cells generated OPN and Hedgehog ligands, and neutralizing OPN with aptamers or inhibition of Hedgehog signal transduction attenuated the fibrogenic effect of NKT cells on hepatic stellate cells. These findings suggested that OPN can function as a paracrine factor, secreted by cholangiocytes or NKT cells, and also as an autocrine factor to promote fibrogenesis in hepatic stellate cells (Fig. 2).
It was suggested that Sex-determining region Y-box 9 (Sox9) was downstream of Gli-2 and responsible for OPN expression in hepatic stellate cells. Co-localized staining for OPN and Sox9 was found in spindle-shaped hepatic stellate cells in the area of fibrosis in mice fed an MCD diet. In adult human hepatic stellate cell lines, LX2 cells, a Hedgehog agonist, increased SOX9 and OPN proteins and siRNA abrogation of Sox9 attenuated the effect of the Hedgehog agonist on OPN expression. Similarly, overexpression of Sox9 rescued the inhibitory effect of a Hedgehog antagonist on OPN expression in the cells.
OPN in Neutrophil Infiltration and Fibrosis in Alcoholic Liver Diseases
HEPATIC OPN MRNA level was correlated with hepatic neutrophil infiltration and fibrosis in patients with alcoholic liver disease. Hepatic expressions of uncleaved and thrombin-cleaved forms of OPN protein, and OPN mRNA were significantly increased in rat alcoholic steatohepatitis models.[42, 43] It was also shown that the extent of hepatic neutrophil infiltration was significantly correlated with the level of cleaved form of OPN in the model. OPN protein was localized predominantly to the hepatocytes surrounding the inflammatory foci,[42, 43] and OPN mRNA expression was found within biliary epithelium, suggesting that OPN was secreted from biliary epithelium. Interestingly, hepatic expressions of uncleaved and cleaved forms of OPN protein and OPN mRNA, and hepatic neutrophil infiltration extent were significantly higher in females compared with males in the experiment.
OPN in HCC
OPN expressions in patients with HCC
OSTEOPONTIN PROTEIN WAS immunohistochemically labeled in 40–55% of HCC,[11, 44-46] and OPN mRNA was overexpressed in 55% of HCC. The immunohistochemical staining and RNA in situ analysis were observed in the cytoplasm of HCC cells, but not in nuclei.[11, 12, 44] OPN positive HCC cells were scattered in the periphery of cancer nodules adjacent to stromal cells, or dispersed in the cancer nodules.[11, 46, 47] OPN protein and/or mRNA overexpression was significantly associated with large size,[12, 45] late tumor stage,[12, 48] poor differentiation,[12, 45, 46, 48] capsular infiltration,[11, 44, 45] vascular invasion,[44, 46] lymph node invasion and intrahepatic metastasis[12, 13, 46] of HCC.
Plasma OPN levels were significantly higher in patients with HCC than in patients with chronic liver diseases without HCC, and healthy controls.[47, 49] In patients after curative resection of hepatitis B virus (HBV)-related HCC, plasma OPN levels significantly decreased after a transient fluctuation, and increased again at the time of tumor recurrence.
As a marker for the diagnosis of HCC in patients with cirrhosis, plasma OPN level had a greater area under curve (AUC) value than α-fetoprotein (AFP)[47, 49, 51] and protein induced by vitamin K absence/antagonist-II by receiver–operator curve (ROC) analysis. Furthermore, the combination of OPN and AFP levels enhanced sensitivity and specificity in detecting HCC.
Moreover, plasma OPN levels were reported to be useful as a prognostic factor after liver resection or transplantation in patients with HCC of tumor–node–metastasis (TNM) stage I, II or III.[48, 52] In a prospective study, TNM stage and plasma OPN level measured prior to tumor resection were highly significant predictors of overall survival (OS) and disease-free survival (DFS) in patients with HCC in China. Preoperative plasma OPN level and Edmondson's grade were also identified as independent predictors for prognostic factor for OS and DFS in patients with TNM stage I of HBV-related HCC.
Increased expression of OPN protein in HCC was also shown to be an independent predictor of poor OS and/or poor DFS in patients undergoing liver transplantation and resection of HCC.[44, 54, 55] Finally, a meta-analysis revealed high OPN expression in HCC predicted poor OS (hazard ratio, 1.37; 95% CI, 1.21–1.55) and DFS (hazard ratio, 1.62; 95% CI, 1.24–2.11) of HCC after liver resection, liver transplantation or transarterial chemoembolization.
OPN roles in growth, invasion and metastasis of HCC in vitro and in vivo
It has been reported that OPN plays significant roles in the metastasis of HCC in vivo and in vitro. However, the effects of OPN on the growth of HCC cells were controversial. In nude mice, s.c. injected with human HCC derived cell line HCCLM3 which showed high metastatic potential and expressed high level of OPN, simultaneous injection of anti-OPN antibody significantly inhibited lung metastasis of the cells, but the antibody did not affect tumor growth. In contrast, i.v. injection of anti-OPN antibody twice per week significantly inhibited tumor growth and angiogenesis as well as lung metastasis in nude mice implanted with HCCLM3 cells. Downregulation of OPN by shRNA inhibited tumor growth and lung metastasis of HCCLM3 cells in the implanted nude mice. OPN antisense oligonucleotides significantly inhibited lung metastasis in mice bearing orthotopic xenografts with the human metastatic HCC cell line HCCLM6, although tumor weight was not reduced. Transfection of OPN significantly enhanced migration and invasion, but not proliferation of the human HCC SMMC-7721 cell line, which was weakly tumorigenic and non-metastatic, and expressed a low level of OPN. On the other hand, overexpression of OPN via transfection significantly stimulated proliferation of Huh-7 cells in vitro and growth of tumors in nude mice injected with the cells.
These discrepancies may be explained by the following observation, suggesting the necessary level of OPN for tumor growth was much lower than that for metastasis of HCC cells. Each of the Lentiviral-mediated miRNA against OPN, Lenti.OPNi-2 and Lenti.OPNi-3, significantly suppressed the migration of HCCLM3 cells in vitro and lung metastasis of HCCLM3 xenografts in nude mice. On the other hand, Lenti.OPNi-3, but not Lenti.OPNi-2, significantly inhibited proliferation of cultured HCCLM3 cells and tumor growth in nude mice. The downregulation degrees of OPN expression were 78% and 95% by Lenti.OPNi-2 and Lenti.OPNi-3, respectively. Suppression of tumor growth may be more difficult compared with prevention of metastasis during the OPN-targeting HCC therapy.
The molecular mechanisms of tumor progression and metastasis, influenced by tumor cell-derived OPN, have been investigated especially in breast cancer, but they are still poorly explored in HCC. OPN silencing by shRNA resulted in an increase of Bax expression, inhibition of Bcl-2/Bcl-xL and XIAP expressions and nuclear factor-κB activation, and induction of mitochondria-mediated apoptosis in HCCLM3 cells. Specific suppression of OPN inhibited MMP-2[58, 61, 63] and urokinase-type plasminogen activator (uPA) expressions[58, 61] in HCC cells. Addition of OPN to the medium or transfection of OPN enhanced expressions of MMP-2[59, 61] and uPA in HCC cells.
OPN Polymorphism in Liver Diseases
MOCHIDA ET AL. PREVIOUSLY detected four SNP in the promoter region of the OPN gene; single nucleotide polymorphisms (SNP) at nt −155, nt −616 and nt −1748, which showed linkage disequilibrium to each other, and an independent SNP at nt −443. It was also demonstrated that SNP at nt −443 was a marker of activity of hepatitis in patients with hepatitis C virus (HCV) infection.[64, 65] Moreover, the efficacy of interferon-based therapies was more effective in patient with T/T at nt −443 than those with C/C or C/T at nt −443.[66, 67] On the other hand, there were no differences in the frequencies of alleles at nt −155, nt −443 and nt −616 between the hepatitis C patients with HCC and those without HCC.
Recently, Hamaoka et al. showed peripheral platelet counts at the time of HCC detection were greater in females with homozygous deletion at nt −155 and C/C or C/T at nt −443 than in those showing other alleic combinations among the hepatitis patients with HCV infection, while no such difference was observed in males. It is well known that the platelet counts decrease with the progression of liver fibrosis in patients with persistent HCV infection. Thus, HCC may develop in the early stage of liver fibrosis after HCV infection in females with such a genetic background.
Dong et al. demonstrated that OPN SNP at nt −443 was significantly associated with OS and time to recurrence (TTR) in the patients with HCC. Multivariate analysis identified allele C/C at nt −443 as a significant independent predictor of increased OS and long TTR. Tumor growth and lung metastasis were enhanced in nude mice implanted with HepG2 cells transfected with OPN promoter-reporter constructs containing allele T at nt −443 compared with allele C. They showed oligonucleotides with allele T at nt −443 increased transcriptional activity and OPN protein level compared with allele C. However, Hamaoka et al. presented that the transcriptional activity was greater in oligonucleotides with allele C at nt −443 than in those with allele T. The reason for the discrepancy remains unclear.
OSTEOPONTIN IS INVOLVED in hepatic inflammation and fibrogenesis in alcoholic and non-alcoholic steatohepatitis. OPN is also linked to progression and metastasis of HCC. OPN expressions were observed in a variety of liver cells, including Kupffer cells, hepatic macrophages, stellate cells, bile duct cells, NKT cells, hepatocytes and HCC cells. OPN is altered through cleavage, splicing or post-translational modifications and has two isoforms, sOPN and iOPN. Recently, OPN was shown to be a downstream effecter of Hedgehog pathway. Therefore, elucidation of a multiplicity of functions of OPN depending on the structure and cellular interactions, could develop novel therapeutics and biomarkers for the liver diseases.