Osteopontin (OPN) is a highly modified integrin-binding extracellular matrix glycophosphoprotein produced by cells of the immune system, epithelial tissue, smooth muscle cells, osteoblasts, and tumor cells. Extensive research has elucidated the pivotal role of OPN in cell signaling that controls inflammation, tumor progression, and metastasis. OPN interaction with the integrin receptors expressed on inflammatory cells through its arginine–glycine–aspartate (RGD) and non-RGD motifs promote migration and adhesion of cells. In the liver, it has been reported that hepatic Kupffer cells secrete OPN facilitating macrophage infiltration into necrotic areas following carbon tetrachloride liver toxicity. Recent work has highlighted the role of OPN in inflammatory liver diseases such as alcoholic and nonalcoholic liver disease and T-cell–mediated hepatitis. The role of OPN in hepatocellular carcinoma (HCC) has also generated significant interest, especially with regards to its role as a prognostic factor. OPN therefore appears to play an important role during liver inflammation and cancer. In this review we will present data to demonstrate the key role played by OPN in mediating hepatic inflammation (neutrophils, monocytes/macrophages, and lymphocytes) and its role in HCC. Greater understanding of the pathophysiologic role of OPN in hepatic inflammation and cancer may enable development of novel inflammation and cancer treatment strategies.
Osteopontin (OPN) is an arginine–glycine–aspartate (RGD)–containing acidic member of the small integrin–binding ligand N-linked glycoprotein (SIBLING) family of proteins. Other members of the SIBLING proteins include bone sialoprotein, dentin matrix protein 1, dentin sialophosphoprotein, and matrix extracellular phosphoglycoprotein (Fisher et al., 2001). More than 25 years ago, OPN was first identified as a major sialoprotein of the bone and has been described as a matrix protein with potential to bridge between cells and hydroxyapatite (Oldberg et al., 1986). OPN is also synonymously mentioned in literature as a secreted phosphoprotein 1 (SPP1)—this name is based on its early identification as the major phosphoprotein secreted by cultured cells (Craig et al., 1990). The protein is produced in a variety of tissues: brain, liver, gastrointestinal tract, lung, bone, cardiac tissues, joints, and kidney, and appears in a variety of biologic fluids including blood, urine, milk, and seminal fluid (Denhardt and Noda, 1998; Sodek et al., 2000). In addition, because of its role as a cytokine produced by activated lymphocytes and macrophages, it has been referred to as early T-cell activation factor (Eta-1; Patarca et al., 1989). Although, the role of OPN in several pathophysiological events in a variety of tissues has been extensively reviewed, there are few comprehensive reviews that have highlighted the potential role of OPN in the liver. Here, we attempt to provide an insight into the implicated roles of OPN in hepatic inflammation, toxicity, and hepatic tumors.
OPN GENE STRUCTURE
The human OPN gene occurs at the long arm of chromosome 4 (4q21–4q25). Of the seven exons present, six of them (exons 2–7) contain coding sequences which are about 954 bp and the first exon is untranslated. The 5′ upstream sequence of OPN gene contains a number of potential regulatory sequences. Regulatory sequences include a TATA-like sequence found at position −27 to −22 (Sodek et al., 2000), CCAAAT-like sequence scattered throughout the 5′ upstream region from −73 to −2190, and the vitamin D responsive element at positions −698 to −684 and −1892 to −1878. An interferon regulatory factor-1 binding sequence, AACTGA, is also identified at positions −1270 to −1264. In addition, potential binding sites for transcription of OPN gene include AP-2, Ets-1, E2A, TCF, and Myb which are characterized near the 5′ cap site of rat and OPN human promoter (Sodek et al., 2000)
REGULATION OF OPN EXPRESSION AND ITS TRANSCRIPTIONAL REGULATORS
The mechanisms by which OPN are increased during inflammation and cancer are not clear. OPN expression is affected by a number of substances including hormones (e.g., vitamin D3 and estrogen) cytokines, and growth factors. Several inflammatory mediators and growth factors such as interleukin-1 (IL-1), tumor necrosis factor alpha, and platelet-derived growth factor are known to stimulate OPN transcription via activation of protein kinase C (Denhardt and Noda, 1998). In addition several cis and trans-regulatory elements within OPN have been investigated to establish the mechanisms by which OPN gene transcription occur. The human, pig, rat, mouse, and chicken OPN promoters have been sequenced. As mentioned previously in the OPN gene structure section, the promoters of OPN show a number of consensus regulatory sequences including a TATA box, an inverted CCAAT box, a GC box, AP1, PEA3, and Ets-1 binding sequences. Recently, a unique transcriptional regulatory mechanism involving members of heterogeneous nuclear ribonucleoproteins (chromatin associated RNA binding proteins) is reported to be involved in the macrophage expression of OPN (Gao et al., 2005). Also, the positive role of β-catenin/Lef-1, Ets, and AP-1 transcription factors to stimulate OPN gene transcription in rat mammary cell lines is recently reported. In this study, it is shown that the presence of these transcription factors in human breast cancer is responsible in part for the OPN over expression and metastatic process (El-Tanani et al., 2004).
OPN METABOLISM, RECEPTORS, STRUCTURE, AND FUNCTION
OPN is a secreted phosphoprotein consisting of 260–301 amino acids with apparent molecular mass on sodium dodecyl sulfate–polyacrylamide electrophoresis ranging from 44 to 80 kDa due to extensive posttranslational modifications such as phosphorylation, glycosylation, sulfation, and enzymatic cleavage. The range of OPN molecular weights is also determined by the electrophoretic conditions. Thrombin and matrix metalloprotease (MMP) cleavage of OPN results in the generation of N- and C-terminal fragments of OPN, which typically generates a band around 40 kDa. The intact OPN generates a band around 66 kDa and the glycosylation and phosphorylation of OPN also alter the molecular weights of OPN by approximately 5 kDa (Denhardt et al., 2001a,b). Polymeric form of OPN has been recently reported which produces a band around 200 kDa that is mediated by transglutaminase (Higashikawa et al., 2007). OPN contains a hydrophobic leader sequence typical of secreted protein that can bind to a variety of cell surface integrins (Denhardt et al., 2001a,b). OPN has a protease-hypersensitive site that separates the integrin and the non–integrin-binding CD44 binding domains. The major integrin-binding sites observed on OPN include the RGD-cell binding domain known to bind αv/β3, αv/β1, α5/β1, α8/β1, and αv/β5 integrins and the SVVYGLR domain in human (Fig. 1 and supplementary data; SLAYGLR in rat and mouse OPN) which binds to α9/β1 and α4/β1 integrins (Diao et al., 2004; Smith and Giachelli, 1998; Yokosaki et al., 1999, 2005). It should be noted that most of these integrins bind to the N-terminal thrombin cleaved fragment of OPN containing both the RGD and the SVVYGLR domain. The SVVYGLR domain is cryptic and the integrins bind to this domain only when the OPN is cleaved by thrombin at the thrombin cleavage site (Green et al., 2001). The thrombin cleavage motif immediately adjacent to the SVVYGLR domain has a conserved sequence RSK, present in most species which suggests the requirement of OPN cleavage by thrombin for some of its physiologic functions (Denhardt et al., 2001a,b). Thus, the thrombin cleavage of OPN releases the SVVYGLR receptor-binding domain responsible to carry out distinct signaling functions. The SVVYGLR site is unusual as it lacks a critical acidic residue present in other binding motifs for integrins α9/β1 and α4/β1. The negatively charged aspartic acid and glutamic acid are considered to be the critical acidic residues in OPN important in integrin-binding ligands. It is thought that the thrombin cleavage of Arginine 168 in OPN creates a free carboxylic acid group, enabling the SVVYGLR motif to engage these integrins. Thus, the free C-terminus of SVVYGLR provides an acidic group required for its interaction with α9/β1 and α4/β1 integrins (Green et al., 2001). In addition to thrombin cleavage of OPN, there are reports in literature suggesting that full-length OPN is also a substrate for cleavage by MMP-3 and -7. In contrast to the enhanced OPN function by thrombin cleavage, MMP-3 and -7 cleavage of full-length OPN within the SVVYGLR site is reported to interfere with the interaction of OPN with its receptors such as α9/β1 and α4/β1 integrins (Yokosaki et al., 2005). However, it is not known if thrombin cleaved N-terminal OPN can be further cleaved by MMP-3.
In addition to proteolytic cleavage, OPN–receptor interactions may also be determined at the transcriptional levels, because three OPN splice variants have been identified (Denhardt et al., 2001a). Although the N-terminal of OPN consists of integrin-binding sites, the C-terminal fragment (non-RGD) of OPN is demonstrated to bind to the CD44 receptor also known as hyaluronic acid receptor (Fig. 1). It is also reported that CD44 binds to the N-terminal region of OPN independent of the RGD sequence suggesting multiple CD44 domains (O'Regan and Berman, 2000), and the interaction of OPN and CD44 remains somewhat controversial. Other domains such as transglutamination, heparin binding, hydroxyapatitite, and calcium binding regions are also identified within the OPN molecule. Recently, the enhanced functional role of polymeric OPN, resulting from cross-linking of OPN by transglutaminase is reported (Higashikawa et al., 2007).
With respect to OPN function, the protein has been reported to play a multifaceted role in several physiologic and pathophysiological functions. The role of OPN in inflammation, immunity, bone remodeling, kidney stone formation, oncogenesis, angiogenesis and cancer progression, vascular calcification, and apoptosis has been investigated. With regards to inflammation, OPN is involved in several inflammatory and immune responses leading to glomerular nephritis (Denhardt and Guo, 1993; Denhardt et al., 2001a,b; Giachelli and Steitz, 2000; O'Regan and Berman, 2000; Rittling and Denhardt, 1999), CCl4-induced hepatotoxicity (Kawashima et al., 1999), and puromycin-induced nephrotoxicity (Denhardt et al., 2001a,b). A recent study demonstrates the involvement of OPN in neutrophil chemotaxis (Koh et al., 2007). OPN is known to be secreted by activated T-lymphocytes inducing macrophage infiltration to the sites of inflammation. The precise mechanisms by which OPN mediates inflammation are unclear, although the roles of OPN interaction with CD44 and integrins (such as αv/β3) to affect Th1 versus Th2 cytokines in inflammation have been tested.
THE CONTRIBUTION OF OPN TO HEPATIC INFLAMMATION AND TOXICITY
The innate immune response provides the first line of defense against microbes and toxins crossing the intestinal barrier and reaching the liver (Janeway and Medhitov, 2002). Microbes and toxins reaching the systemic circulation are frequently encountered by Kupffer cells (resident macrophages in the liver) before they are rapidly cleared (Gregory and Wing, 1998; Nagy, 2003). In addition, Kupffer cells generate the inflammatory response leading to recruitment of inflammatory cells such as neutrophils, monocytes, T and B lymphocytes, as well as natural killer (NK) and natural killer T (NKT) cells to the liver ultimately facilitating liver damage.
OPN as a Mediator of Hepatic NKT Cell and Neutrophil Infiltration
T-cell–mediated liver diseases such as hepatitis induced by concavalin A (Con A), autoimmune hepatitis and viral hepatitis are classic examples of inflammatory liver disease where there is increased neutrophil and lymphocyte (NKT and T-cell) infiltration followed by hepatocyte necrosis (Fujii et al., 2005; Zhu et al., 2007). A variety of immunoregulatory cytokines associated with Th1 and Th2 immune responses are implicated during T-cell–mediated hepatitis. The mechanistic link between OPN and NKT cells was recently tested in a Con A–induced hepatitis mouse model (Diao et al., 2004). In this study, OPN-deficient mice (OPN−/− mice backcrossed 11 times to B6 mice) developed significantly lower liver injury as evidenced by lower serum transaminase levels and increased survival rates when compared with the wild type (WT) controls. NKT cells were determined to be the major T-cell phenotype in the controls during the development of Con A–induced hepatic injury: this injury was significantly decreased in the OPN-deficient mice. The authors reported that in addition to full-length OPN, the thrombin cleaved OPN (cOPN) product was also present in the liver of WT mice following Con A injection. In verification of the in vivo data, in vitro migration assays showed that infiltrating leukocytes purified from the liver after Con A injection migrated toward the thrombin cleaved form of OPN more efficiently than to the full-length form of OPN. Neutrophils were the predominant leukocyte population that responded to the cleaved OPN. Furthermore, antibodies directed against β1 and α4 integrins inhibited the migration of liver infiltrating cells toward the cleaved OPN. Also, antibodies that specifically recognize and block the SLAYGLR sequence in mouse OPN inhibited cell migration induced by cleaved OPN and decreased liver injury (Diao et al., 2004). Together, these results suggest the role of cleaved OPN and a strong interaction between cleaved OPN and integrin receptor in neutrophil infiltration and liver injury was noted in this model.
The ability of OPN to attract neutrophils during the course of Con A–induced hepatitis is proposed to be due to the following. Intrahepatic resident NKT cells express α9 and α4 integrins, which are receptors for the thrombin cleaved from of OPN. Then, after Con A–induced activation, NKT cells secrete OPN which is cleaved by thrombin in the liver. The interaction of NKT cells and the thrombin cleaved from of OPN through its receptors further activates NKT cells which likely produces macrophage inflammatory protein-2 (MIP-2), a known chemotactic factor for neutrophils. Thrombin cleaved form of OPN also interacts with its receptors α9 and α4 integrins on neutrophils so that neutrophils become activated and contributes to additional liver damage (Diao et al., 2004). Interestingly, the specific contribution of OPN versus MIP-2 to hepatic neutrophilic inflammation was not comprehensively investigated in this model. Nevertheless, this study clearly demonstrated that OPN produced by NKT cells can contribute to hepatic neutrophil infiltration and liver injury in Con A–induced hepatitis model, and focused attention on the interaction of OPN and neutrophils in regulating the innate immune response.
OPN as a Mediator of Hepatic Macrophage Infiltration
One of the first descriptions of OPN was a protein induced early in T cell activation, known as ETA-1. The purified protein showed saturable binding to macrophages, and injection of the protein into mice resulted in an inflammatory infiltrate rich in macrophages, indicating that OPN is chemotactic for these cells in vivo (Singh et al., 1990). Interestingly, this infiltrate also contained abundant polymorphonuclear leukocytes (neutrophils). The chemotactic peptide fMLP, when injected into rats similarly attracted a rich cellular infiltrate containing numerous macrophages. Infusion of injected animals with a neutralizing antibody to OPN reduced this macrophage accumulation by 60% (Giachelli et al., 1998), providing strong evidence that OPN regulates the chemotactic response of macrophages. Together, these experiments suggest that OPN binds to macrophages in vitro, and is both necessary and sufficient for macrophage accumulation.
In addition, it is clear that activated macrophages themselves produce abundant OPN. In vivo, OPN expression in infiltrating macrophages has been demonstrated in a variety of pathological situations such as myocardial necrosis (Murry et al., 1994), pulmonary fibrosis (Takahashi et al., 2001), and interstitial monocyte infiltration in the kidney (Okada et al., 2000), among others. In granulomatous diseases, such as sarcoidosis, and tuberculosis, total OPN expression is elevated, and the protein is found in macrophages in the granulomas (Carlson et al., 1997; O'Regan and Berman, 2000). OPN is also associated with macrophage differentiation. It was identified as one of genes most highly upregulated during differentiation of monocytes to macrophages (Krause et al., 1996) and is upregulated during PMA-induced differentiation of HL-60 cells along the monocyte pathway (Atkins et al., 1998). In RAW264.7 transformed macrophage cells, down regulation of OPN expression with small inhibitory RNA (siRNA) resulted in reduced expression of markers of macrophage differentiation, suggesting that endogenous OPN expression supports the differentiated phenotype in these cells (Nystrom et al., 2007). Together, these results support the idea that OPN is expressed in differentiated macrophages, and is important for their function.
OPN Expression and Function in Hepatic Macrophages and Kupffer Cells
OPN expression has been associated with various kinds of injury, including wound healing in the skin (Liaw et al., 1998), a variety of renal pathologies (reviewed in Xie et al., 2001), in myocardial injury (Zhao et al., 2007b), and in the brain following damage caused by ischemia (Denhardt et al., 2001a,b; Wang et al., 1998). In the liver, carbon tetrachloride (CCl4) treatment causes significant liver damage, accompanied by increased numbers of macrophages in the tissue. OPN expression is strongly upregulated in the livers of CCl4 treated rats, with peak expression 2 days after CCl4 administration (Kawashima et al., 1999). Expression of OPN was highest in Kupffer cells, macrophages, and hepatic stellate cells, consistent with a role of OPN in the function of these cell types. In another model of liver injury, treatment of rats with heat-killed propionibacterium Acnes (pA) caused substantial macrophage infiltration accompanied by macrophage-rich granuloma formation, but not necrosis (Wang et al., 2000). Again, OPN expression was upregulated in the livers of pA treated rats, and expression of OPN protein was localized to hepatic macrophages and Kupffer cells. OPN expression preceded granuloma formation, suggesting that OPN may be chemotactic for macrophages in this injury model. Together, these results suggest that OPN is expressed in macrophages and related cells at sites of hepatic injury and that the protein contributes to the host response to infection or injury. Because OPN stimulates cell migration, it may act as a chemotactic factor in the recruitment of macrophages to sites of liver injury.
The Contribution of OPN to Neutrophil Infiltration in Alcoholic Liver Disease
Alcoholic liver disease (ALD) is a classic example by which liver injury is mediated by inflammatory cells including lymphocytes, macrophages, and neutrophils. Steatohepatitis is one of the major complications of heavy alcohol (EtOH) drinking in humans (Bautista, 2002; Diehl, 2002; French, 2002; Jaeschke, 2002; Ramaiah and Jaeschke 2007a,b; Ramaiah et al., 2004). Chronic ethanol ingestion initially leads to alcoholic steatosis (AS) resulting in hepatic fat accumulation, and this pathologic stage is considered to be mostly benign although there are reports which suggest the contrary. Alcoholic steatohepatitis (ASH) is the next step in the progression of ALD which is characterized by hepatic fat accumulation, inflammatory cell infiltration, and parenchymal injury. Hepatic neutrophil infiltration during ASH is reported to a consistent finding in chronic human alcoholics and efforts are directed toward investigating the precise mechanisms for hepatic neutrophil infiltration.
In order to understand the reason for hepatic neutrophil infiltration during ASH, we hypothesized that OPN produced from liver (hepatocytes macrophages and biliary epithelium) during early stage of ALD mediates hepatic neutrophil transmigration. We tested the induction of OPN in a previously established rodent Lieber–DeCarli model of ASH wherein male Sprague–Dawley rats were fed EtOH-containing Lieber–DeCarli liquid diet for 6 weeks followed by Lipopolysaccharide (LPS) injection (Apte et al., 2005; Enomoto et al., 1999; Ramaiah et al., 2004). Interestingly, higher neutrophilic inflammation, necrosis, and liver injury in this model correlated well with the levels of both intrahepatic and circulating levels of OPN (Apte et al., 2005). Because of the known increased biologic activity of the cOPN due to exposure of the integrin-binding sites on OPN, we also tested the levels of hepatic cOPN in this model and noted significantly higher levels of the cleaved form of OPN. Again, higher cOPN induction correlated with higher neutrophil infiltration and liver injury. The high degree of correlation noted between OPN induction and hepatic neutrophil infiltration was further confirmed by cause and effect studies, wherein enhanced peritoneal fluid neutrophil infiltration was noted in rats injected with OPN intraperitoneally in a rat peritonitis model (Banerjee et al., 2006). Although the precise mechanism by which hepatic OPN induction mediates hepatic neutrophil infiltration was not tested, this study points out that hepatic OPN induction may be contributing to the neutrophilic inflammation during the histogenesis of ASH.
Because of the evidence that OPN mediates hepatic neutrophil infiltration during ALD, the primary author's laboratory tested the role of OPN in mediating higher neutrophil infiltration and liver injury noted in females following alcohol ingestion. Using the rodent model of EtOH-containing Lieber–DeCarli diet model, we noted higher hepatic neutrophil infiltration in the females correlating with the increased liver injury in females using Sprague–Dawley rats (Banerjee et al., 2006). Higher expression of the secreted form of cOPN and full-length OPN was noted in the females in the ASH rats compared with the males and the OPN expression was directly correlated with the degree of hepatic neutrophil infiltration. Because of the random multifocal nature of neutrophilic infiltrates within the liver, we tested the precise localization of OPN. Interestingly, both the OPN messenger RNA (mRNA) and protein was predominantly localized within the biliary epithelium and few necrotic hepatocytes as measured by in situ hybridization, immunofluorescence, and immunohistochemistry. The OPN mRNA data correlated well with the protein expression. Neutralizing OPN antibody experiments designed to confirm the ability of OPN to attract neutrophils to the liver resulted in significant inhibition by almost 50% of neutrophil accumulation in the liver (Banerjee et al., 2006).
To determine the mechanism by which OPN mediated higher neutrophil infiltration in females, we investigated the role of neutrophil-associated integrins such as α9, α4, and β1 integrins which are receptors for OPN within the SVVYGLR and RGD sites of OPN. We noted higher expression of α9, α4, and β1 integrins in the female ASH model suggesting that OPN was mediating its neutrophil infiltration effects through these integrins (unpublished studies). Also, ongoing studies carried out using a living cell array wherein naïve neutrophils were exposed to varying concentration of OPN showed neutrophil activation and chemotaxis. IL-8 was employed as a positive control to compare OPN-mediated neutrophil activation. These studies further confirm the neutrophil chemotactic role of OPN causing higher neutrophil infiltration and injury in the female rodent ALD model.
A Link between OPN and Nonalcoholic Fatty Liver Disease
Nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH) form a spectrum of disease from simple accumulation of fat to cirrhosis and end stage liver disease (Angulo, 2002; Diehl, 2002; Sahai et al., 2004). The progression of hepatic pathology during NAFLD and NASH is somewhat similar to the pathogenesis of ASH, although the mechanisms are unclear. Recently, OPN has been identified as an important cytokine whose expression is increased early in the course of the disease in an experimental dietary rodent model of NASH (Sahai et al., 2004). The dietary model for causing NASH in mice in this study is based on the administration of diet deficient in methionine and choline for up to 12 weeks. Upregulation of OPN mRNA levels was noted following the dietary treatment at 4 and 8 weeks with little effect during early periods of treatment. The control mice had minimal amounts of OPN mRNA expression. In this study, the role of OPN was confirmed in experiments involving OPN knockout mice where significantly decreased liver injury and fibrosis was noted compared with their WT counterparts (Sahai et al., 2004). It should be pointed out that this study did not investigate the role of OPN-mediated inflammation as a contributing factor toward fibrosis. Also, the role of cleaved OPN in the progression of fibrosis or in the occurrence of inflammation during NASH was not investigated in this study.
OPN and Acetaminophen-Mediated Liver Injury
In an effort to better understand the basis for drug-induced liver injury comparative gene and protein expression were identified between resistant (SJL) and susceptible (C57BL6) strains of mice in acetaminophen-(APAP)–induced liver disease (Welch et al., 2006). The goal was to uncover potential risk factors that contribute to differential sensitivity in APAP-mediated liver damage between C57BL6 and SJL mice. Based on the micro array data, higher OPN mRNA was identified in the C57BL6 mice as compared with the SJL mice suggesting that OPN may play a role in higher liver injury following APAP treatment. In support of this hypothesis, OPN knockout mice were found to be more resistant to APAP-mediated liver injury than their WT counterparts (Welch et al., 2006). The proteomics data further confirmed the role of OPN in mediating higher APAP-induced liver injury in the susceptible strain of mice. Although the precise mechanisms by which decreased OPN in SJL mice results in lower liver injury are not clear, the authors noted lower numbers of NK and NKT cells in the SJL mice compared with the C57BL6 mice. Based on this, the authors believe that NK and NKT cells are the main OPN-producing cells and thus there are lower hepatic levels of OPN in the SJL mice. The authors speculate that lower OPN levels eventually fail to activate sufficient numbers of NKT cells in the resistant SJL mice compared with the B6 susceptible strain, although not tested in this study.
The Role of OPN in CCl4 Toxicity and Granuloma Formation in the Liver
To directly test the idea that OPN is chemotactic for inflammatory cells in the liver, transgenic mice were created that express OPN in hepatocytes under the control of the serum amyloid P-component promoter (Mochida et al., 2004). These mice developed mononuclear cell infiltration starting at 12 weeks of age, and becoming more extensive with aging. This infiltrate was primarily cytotoxic T cells, identified by CD8 and HLA-DR expression. When these mice were treated with CCl4, the extent of necrosis and cellular infiltration was reduced compared with control mice, suggesting that OPN overexpression in hepatocytes may play a protective role (Koh et al., 2005). Studies using OPN-deficient mice have further clarified the role of OPN in liver injury. Granuloma formation in the liver was induced by β-glucan administration to both WT and OPN-deficient mice (Morimoto et al., 2004). In WT mice, β-glucan administration induced OPN expression in the liver at both the RNA and protein levels; the kinetics of the changes in OPN expression paralleled that of granuloma formation, and OPN expression was detected within the granulomas. OPN−/− mice had significantly reduced granuloma numbers and size 2 weeks after β-glucan administration. Although there was no difference in the histological appearance of the granulomas between the two genotypes, there was a twofold reduction in macrophage, CD4+ T cells, and dendritic cell accumulation in the livers of OPN−/− mice as compared with WT animals. Furthermore, these authors went on to show OPN overexpression (under the control of the Eμ promoter) increases granuloma size, persistence, and the resulting fibrosis (Morimoto et al., 2004).
Confirming the results of Kawashima, Lorena et al. (2006) found that OPN expression was increased in mice after treatment with CCl4, and that the protein was localized to hepatic macrophages. Liver damage, as assessed by ALT and AST in serum, was twofold worse in OPN-deficient mice as in WT controls, measured 1 day after CCL4 treatment, and the area of necrosis was increased 2 and 3 days after treatment. This was accompanied by reduced numbers of macrophages in the livers of OPN−/− mice, increased fibrosis, and a reduction in inducible nitric oxide synthase expression. These results are consistent with a role for OPN in mediating macrophage infiltration.
Taken together, these studies clearly demonstrate that OPN can regulate macrophage infiltration and potentially differentiation and function in the liver. It appears that the overall effect of OPN, however, can be protective or deleterious, perhaps depending on the cellular mileu in each individual pathologic condition. Given the complex and poorly understood functions of macrophages in both innate and acquired immunity, it is not surprising that the role of OPN in regulating these cells is equally complex, and it is likely that OPN is important in several different aspects of the response of the liver to injury.
THE ROLE OF OPN IN HEPATOCELLULAR CARCINOMA
Liver cancer, or hepatocellular carcinoma (HCC), ranks third among causes of cancer death worldwide, and is one of the most frequently seen cancers in Asia. The 5-year survival rate for this malignancy is depressingly low, ranging from 4% to 6% in different countries, although there are indications that this is slowly improving (Tang et al., 2004). Risk factors for this disease include infection with either hepatitis B or C, and alcoholic cirrhosis; exposure to environmental carcinogens such as alflatoxin is an important etiologic factor in developing countries (El Serag and Mason, 1999). OPN is highly expressed in many malignancies, and the expression level of OPN in tumor tissues or in blood of cancer patients has been positively correlated with worse prognosis in many cancer types (see Rittling and Chambers, 2004 for review). HCC is no exception.
OPN as a Prognostic Factor
Given the poor survival rate for patients with metastatic HCC, there is much interest in the identification of prognostic factors. In studies of surgically resected HCC, differential display was used to identify genes overexpressed in HCC as compared with matched normal liver tissues (Gotoh et al., 2002; Pan et al., 2003). OPN was identified as highly over expressed, and overexpression correlated with tumor grade, tumor stage, and early recurrence. Using gene expression profiling, Ye et al. (2003) developed a gene expression signature with the goal of discovering genes whose expression could distinguish primary from metastatic HCC. Although gene expression in primary and metastatic tumors was overall extremely similar, there were significant changes in gene expression between tumors in patients with and without metastatic spread (Ye et al., 2003). Again, OPN was identified as one of the genes whose expression was most highly up regulated in metastatic as compared with nonmetastatic tumors.
Plasma OPN and HCC
Subsequent studies have refined our knowledge of how OPN expression is associated with HCC. When gene expression was measured as absolute values, rather than as a ratio between tumor and normal tissue, OPN again was identified as a gene highly up regulated in tumor tissue (Luo et al., 2006). In hepatitis B virus (HBV)–associated HCC, amplification of chromosome 4q21, close to the spp (OPN) locus, suggested that part, but unlikely all, of the upregulation of OPN expression may result from gene amplification (Huang et al., 2006). OPN protein as well as mRNA levels are correlated with poor prognosis. OPN identification by immunohistochemistry in HBV-positive HCC was strongly positively correlated with portal vein and lymph node invasion (p < 0.01) and negatively correlated with worse disease-free and overall survival (p < 0.001) (Xie et al., 2007). Similarly, plasma OPN was significantly higher than in patients with chronic liver disease or in healthy individuals. Within the HCC patients, higher plasma OPN positively correlated with reduced liver function, as defined by increasing Child-Pugh class, and with tumor stage (Kim et al., 2006), suggesting the use of plasma OPN as a prognostic factor. In a prospective study, plasma OPN was measured in the plasma of HCC patients prior to tumor resection. These patients were divided into two groups based on the level of expression of OPN (< 200 ng/ml and ≥ 200 ng/ml)—those with the higher level of plasma OPN were found to have significantly higher 1- to 2-year postoperative rates of recurrence. In addition, plasma OPN level was a highly significant predictor of overall survival and disease-free survival (p < 0.001) (Zhang et al., 2006). Together these studies showing that OPN is highly expressed in HCC, and is correlated with worse prognosis underline the importance of OPN in HCC, and suggest that the protein enhances tumor development and metastases. Patients with hepatitis B infection frequently develop liver cirrhosis: recently it was shown that elevated plasma OPN is an excellent predictor of cirrhosis in these patients (Zhao et al., 2007a). However, the role of the protein in these diseases is still poorly understood.
Function of OPN in Cancer and Metastasis
The mechanism by which OPN enhances tumor development and particularly metastasis is still poorly understood. OPN enhances migration of many different cell types through its interaction with integrins, and the ability of cancers to metastasize is strongly associated with migration, so this aspect of OPN function is likely important in metastasis. Other effects of OPN on tumor cells are also likely important, however, and these have been recently reviewed (Chakraborty et al., 2006; Rangaswami et al., 2006; Tuck et al., 2007). Intriguingly, OPN was recently shown to upregulate hyaluronic acid synthase, which may contribute to survival of cells in the absence of adhesion, another key feature of metastatic cells (Cook et al., 2006). This observation may underlie OPN's ability to enhance growth of cancer cells in soft agar, which is the in vitro characteristic most associated with tumorigenesis (Wu et al., 2000). While many of these studies were performed in breast cancer cells, it is likely that the function of OPN is similar is all kinds of cancer, and that these functions are relevant to liver cancer and its metastasis as well.
CONCLUSIONS, LIMITATIONS, AND FUTURE PERSPECTIVES
This review focuses on the functional roles of OPN in hepatic inflammation, toxicity, and cancer. OPN expression is significantly increased in response to hepatic inflammation and carcinoma and alteration of OPN function can occur due to its cleavage by proteases and thrombin. Although there is evidence for induction of this protein during hepatic inflammation and cancer, the precise mechanism by which OPN functions during these pathologic events is not well studied. Therefore, much remains to be learned about these mechanisms and the functional contributions of OPN in order to establish appropriate anti OPN therapeutic strategies within the liver.
It is important to point out the potential therapeutic usefulness of OPN, because hepatic inflammation appears to be important for the pathogenesis and progression of liver diseases. Although, the role of proinflammatory cytokines, chemokines, and adhesion molecules has been implicated in higher hepatic inflammation, other classes of molecules have also been suggested. With the identification of OPN upregulation in hepatic inflammatory diseases, the pursuit for the role of OPN in attracting inflammatory cells to the liver has garnered increased attention. However, several of the following challenging questions still need to be addressed by investigators to fill the knowledge-gaps as it relates to liver. (1) What is the precise mechanism by which OPN mediates hepatic inflammatory cell infiltration? Although it is clear that there are receptors for OPN (such as β1 and β3 integrins) on neutrophils, macrophages, and lymphocytes, it is not exactly known if OPN directly mediates integrin–receptor signaling to contribute to inflammation or if this is an indirect effect mediated by changes in hepatic macrophage function such as through macrophage-derived cytokines. Furthermore, the type of inflammatory cell integrins and the precise OPN motif responsible for hepatic inflammation is also not well studied. For example, the detailed contribution of α9β1 and α4β1 neutrophil integrins in OPN-mediated hepatic neutrophil transmigration during liver toxicity needs to be tested in detail. Also, the role of cleaved OPN (such as from thrombin), dimeric and polymerized OPN (transglutaminase-mediated), and splice variants of OPN has not been studied in inflammation associated with liver toxicity. (2) Is there a difference in the function of OPN induced within hepatocytes, mononuclear cells and biliary epithelium? Although some OPN appears to be constitutive in the liver, there is clearly a secretory form of OPN which appear to have different function. For example, although low amounts of constitutive OPN are expressed within the bile ducts, there is higher expression of the secretory form within the biliary epithelium and hepatocytes following ethanol administration in rats. However, the inflammatory cells are not always directed toward the bile ducts, instead they migrate toward the hepatocytes undergoing necrosis. (4) What is the molecular basis for OPN induction during hepatic inflammation? Although the basis for OPN induction has somewhat been addressed at the level of the OPN gene promoter binding sites, the role of the parent toxic compound, or its metabolite or even systemic endotoxemia as noted during ALD has not been evaluated in regulating OPN expression. Finally, (5) although extracellular OPN in the liver has been implicated for regulation of inflammation, the intracellular function of OPN has not been thoroughly investigated.
Major limitations for understanding the role of OPN in the liver include the lack of reliable antibodies to consistently detect various forms of OPN, presence of several posttranslational modifications within OPN (cleavage by thrombin and MMPs, phosphorylation, glycosylation and polymerization) and different motifs on OPN whose functions are yet to be determined. With respect to the role of OPN in liver cancer, one of the best strategies to combat liver cancer is by early diagnosis and effective treatment. In spite of the number of cancer biomarkers currently used routinely, their usefulness still remains limited due to lack of sensitivity, specificity, and predictive values for population screening. Because some studies have established an association between elevated OPN levels in patients’ tumors or blood with poor prognosis, it could represent a tumor marker for use in liver cancer which is yet to be fully exploited. Although OPN expression may be elevated in pathologies other than cancer, in cases where cancer is already established as the primary pathology, plasma OPN level may be a useful independent prognostic factor for disease stage. Hopefully, this review will generate additional interest on this molecule in the area of liver and facilitate cross interactions providing new ideas and strategic focus.
Supplementary data are available online at http://toxsci.oxfordjournals.org/.
National Institutes of Health grants (AA016316) to S.K.R. and (R01-DK067685) to S.R.
- inflammatory cell
- signal transduction
- liver diseases
- carcinoma, hepatocellular
- kupffer cells
- neoplasm metastasis
- carbon tetrachloride
- tumor cells
- toxic effect
- prognostic factors
- tumor progression