Spreds, inhibitors of the Ras/ERK signal transduction, are dysregulated in human hepatocellular carcinoma and linked to the malignant phenotype of tumors

Article metrics

Abstract

Aberrant activation of the Ras/Raf-1/extracellular-regulated kinase (ERK) pathway has been shown to be involved in the progression of human hepatocellular carcinoma (HCC). However, the mechanism of dysregulation of ERK activation is poorly understood. Recently, we identified Sprouty-related protein with Ena/vasodilator-stimulated phosphoprotein homology-1 domain (Spred) as a physiological inhibitor of the Ras/Raf-1/ERK pathway. In this study, we found that the expression levels of Spred-1 and -2 in human HCC tissue were frequently decreased, comparing with those in adjacent non-tumorous tissue. Moreover, Spred expression levels in HCC tissue were inversely correlated with the incidence of tumor invasion and metastasis. Forced expression of Spred-1 inhibited HCC cell proliferation in vitro and in vivo, which was associated with reduced ERK activation. Spred-1 overexpression also reduced the secretion of matrix metalloproteinase-9 (MMP-9) and MMP-2, which play important roles in tumor invasion and metastasis. In addition, Spred-1 inhibited growth factor-mediated HCC cell motility. These data indicate that the reduction of Spred expression in HCC is one of the causes of the acquisition of malignant features. Thus, Spred could be not only a novel prognostic factor but also a new therapeutic target for human HCC.

Introduction

Cell proliferation is a highly regulated process with multiple levels of control in normal cells, and dysregulated proliferation is one of the most important biological features of cancer. Various growth factors and cytokines activate the Ras/Raf-1/extracellular-regulated kinase (ERK) cascade through different types of membrane receptors, leading to cell proliferation as well as cell differentiation (Johnson and Lapadat, 2002). In human hepatocellular carcinoma (HCC), proliferative activity is higher even in precancerous lesions, such as adenomatous hyperplasia, than in the adjacent non-tumorous liver tissue. Previous studies have shown that various growth factors enhance HCC cell proliferation through the activation of the Ras/Raf-1/ERK pathway and aberrant activation of this pathway is frequently observed in human HCC (Schmidt et al., 1997; Ito et al., 1998; Tsuboi et al., 2004), although most HCC tumors lack oncogenic Ras mutation. In addition, high tumor proliferation induced by ERK activation is closely linked to resistance against any type of therapy, including irradiation therapy and chemotherapy, resulting in shorter survival times for patients.

Furthermore, the prognosis for human HCC depends mainly on the aggressive biological characteristics, including not only rapid cell proliferation but also tumor invasion and metastasis (Marrero and Lok, 2004), which are also correlated with hyperactivation of the Ras/ERK pathway. Tumor invasion and metastasis require tissue disruption and enhanced motility of cancer cells. The expression of matrix metalloproteinase (MMP)-9 and MMP-2, which hydrolyse the extracellular matrix, has been implicated in cancer cell invasion and metastasis (Coussens et al., 2000; Masson et al., 2005). It has been shown that activation of the Ras/Raf-1/ERK pathway upregulates MMP secretion in several cancers (Zaragoza et al., 2002). Ras also regulates cell motility. V12Ras, a constitutively activated form of Ras found in more than 30% of tumors, is known to induce cell migration through actin cytoskeletal rearrangement (Ridley and Hall, 1992; Ridley et al., 1992). Actin-stress fiber formation has been shown frequently in invasive and metastatic lesions of many cancers. These findings suggest that activation of the Ras/Raf-1/ERK pathway is associated with malignant phenotypes of tumors, and the status of this pathway may reflect the prognosis of an HCC patient. Suppression of the ERK pathway also could be an effective therapeutic way for HCC treatment. However, the regulation mechanism of ERK in HCC has been poorly understood.

Recently we isolated Sprouty-related protein with Ena/vasodilator-stimulated phosphoprotein homology-1 (EVH1) domain (Spred) as a negative regulator of the Ras/Raf-1/ERK pathway (Wakioka et al., 2001). The Spred family includes Spred-1, -2 and -3 (Kato et al., 2003). Sprouty-related protein with EVH1 domain-1 and Spred-2 have been found to have an overlapping expression pattern in various organs, including the liver, whereas Spred-3 is expressed exclusively in the brain. Sprouty-related protein with EVH1 domain has a Sprouty-related C-terminal cysteine-rich (SPR) domain and an N-terminal Ena/VASP homology (EVH1) domain. Sprouty-related protein with EVH1 domain, like Sprouty, is localized in the cellular membrane through its SPR domain. Sprouty-related protein with EVH1 domain downregulates the Ras/Raf-1/ERK pathway by interacting with Ras and inhibiting Raf kinase activation. The EVH1 domain is an interaction module found in several proteins implicated in actin-based cell motility (Prehoda et al., 1999). Ena/vasodilator-stimulated phosphoprotein homology-1 domains interact with the consensus proline-rich motif FPPPP and are required for targeting the actin assembly machinery to sites of cytoskeletal remodeling (Volkman et al., 2002). The actin cytoskeleton plays an important role in orchestrating cell migration, axon guidance, phagocytosis and cytokinesis (McAllister et al., 2003).

More recently, we showed that Spred-1 prevents murine osteosarcoma cell proliferation and metastasis through inhibition of the Ras/Raf-1/ERK pathway (Miyoshi et al., 2004). However, the relationship between Spred and human cancer remained to be investigated. Thus, we postulated that abnormal Ras/Raf-1/ERK pathway activation is associated with Spred expression in human HCC tissues. In this study, we found that the expression of Spred-1 and that of Spred-2 were simultaneously decreased in HCC tissue compared with adjacent non-tumorous tissue in about 70% of HCC patients. Interestingly, the Spred expression level was correlated with clinicopathological features of aggressive HCC tumors. The overexpression of Spred-1 inhibited HCC cell proliferation in vitro and in vivo, which was associated with reduced ERK activation. In addition, Spred-1 expression reduced hepatocyte growth factor (HGF)-induced MMP secretion and cell migration. Thus, we present evidence that Spred functions as a tumor suppressor in human HCC.

Results

Expression of Sprouty-related protein with Ena/vasodilator-stimulated phosphoprotein homology-1 domains and extracellular-regulated kinase activation in human human hepatocellular carcinoma tissue

First, we confirmed that ERK activation levels were higher in HCC tissues than in non-tumorous regions, which is consistent with previous reports (Ito et al., 1998). Our investigation revealed that in 17 out of 32 HCCs (53%), phosphorylation of ERK in tumorous regions was higher than in adjacent non-tumorous tissue (Figure 1a and data not shown). We then examined the expression of Spred-1 in human HCC tissues. As shown Figure 1a and b, the expression of Spred-1 were decreased in HCC tissues at both protein and mRNA levels. These results were confirmed by immunohistochemistry for phospho-ERK1, 2 and Spred-1 (Figure 1c). Phosphorylation of ERK was largely detected in the cytoplasm of tumor cells, but not in non-tumorous cells in the tumor regions as well as normal parenchymal cells, whereas the expression of Spred-1 in tumor cells was weaker than that in adjacent non-tumorous cells.

Figure 1
figure1

Extracellular-regulated kinase (ERK) activation and reduction of Sprouty-related protein with Ena/vasodilator-stimulated phosphoprotein homology-1 domain (Spred) expression in human hepatocellular carcinoma (HCC). (a) Immunoblot analysis of phospho-ERK1, 2, Spred-1 and glyceraldehydes-3-phosphate dehydrogenase (GAPDH) in paired noncancerous liver tissues (NT) and tumor tissues (T) from a hepatocellular carcinoma patient. Noncancerous liver and HCC tumor tissues were prepared for Western blot analysis. (b) Representative expressions of Spred-1 and -2 analysed by reverse transcriptase–polymerase chain reaction (RT–PCR). The expression of Spred-1 mRNA, Spred-2 mRNA and GAPDH mRNA (as an internal control) were evaluated by RT–PCR. (c) Photomicrographs of tumor tissue (T) and noncancerous liver tissues (NT) sections immunostained for Phospho-ERK1, 2 and Spred-1. Tumor tissues (T) showed the reduced expression of Spred-1 compared to noncancerous liver tissues (NT).

We further quantified the Spred-1 and -2 mRNA expression of human tissue using real-time reverse transcriptase–polymerase chain reaction (RT–PCR). The expression of Spred-1 and -2 was standardized using G3PDH as the internal control. As shown in Figure 2a, in 27 of 32 samples (84%), the expression of Spred-1 and -2 was reduced in HCC compared with corresponding non-tumorous tissue samples. In 22 of 27 cases (68%), the reduction of Spred-1 and -2 occurred simultaneously.

Figure 2
figure2

(a) Relationship between tumor size and Sprouty-related protein with Ena/vasodilator-stimulated phosphoprotein homology-1 domain (Spred) expression. The copy numbers of Spred-1 and -2 mRNA in tumor tissues (T) and noncancerous liver tissues (NT) determined by real-time reverse transcriptase–polymerase chain reaction (RT–PCR). Expression of Spred-1 and -2 mRNA was standardized using the expression of glyceraldehydes-3-phosphate dehydrogenase (GAPDH) mRNA as an internal control. The tumor tissue (T)/noncancerous tissue (NT) ratio of mRNA was then calculated to determine Spred-1 and -2 mRNA levels in each case. Spred-1 and -2 mRNA ratios (T/NT) were evaluated for correlation with tumor size. Correlation coefficient of Spred-1 and -2 was r=−3.28 and −4.22, respectively. (b) Relationship between capsular invasion and Spred expression. A comparison was made of the distribution of the Spred mRNA ratio (T/NT) between human hepatocellular carcinomas (HCCs) with and without capsular invasion. Spred-1 and -2 mRNA ratios (T/NT) of capsular-invasion-positive cases were significantly lower than those of negative cases. (Spred-1: P=0.075, Spred-2: P=0.0332). (c) Relationship between portal vein invasion and Spred expression. A comparison was made of the distribution of the Spred mRNA ratio (T/NT) between HCCs with and without portal vein invasion. Spred-1 and -2 ratios (T/NT) of portal vein-invasion-positive cases were lower than those of negative cases, although there was no statistical significance. (Spred-1: P=0.2255, Spred-2: P=0.2830). (d) Relationship between intrahepatic metastasis and Spred expression. A comparison was made of the distribution of the Spred mRNA ratio (T/NT) between HCCs with and without intrahepatic metastasis. Spred-1 and -2 mRNA ratios (T/NT) of intrahepatic-metastasis-positive cases were significantly lower than those of negative cases (Spred-1: P=0.0332, Spred-2: P=0.0320).

Sprouty-related protein with Ena/vasodilator-stimulated phosphoprotein homology-1 domain is associated with clinicopathological malignant features of human hepatocellular carcinoma

Human HCC tumors acquire their biological malignant potential, for example, rapid cell growth, invasion and metastasis, during the process of dedifferentiation. We next investigated how Spred is involved in the clinical course of HCC by comparing the expression of Spred and clinicopathological features. The tumor tissue (T)/noncancerous tissue (NT) (T/NT) ratios of Spred-1 and -2 expression decreased with the size of the HCC tumor (Figure 2a). The T/NT ratios of Spred-1 and -2 expression were significantly correlated with the capsular invasion of HCC (Figure 2b). The T/NT ratios of Spred-1 and Spred-2 in HCCs with portal vein invasion were lower than in those without portal vein invasion, although the differences were not statistically significant (Figure 2c). Furthermore, we found that the T/NT ratios of Spred-1 and -2 expression were significantly lower in HCCs with intrahepatic metastasis than in those without metastasis (Figure 2d). Sprouty-related protein with EVH1 domain-1 and -2 expression showed no correlation with age, hepatitis virus infection, histology of adjacent tissue state, degree of HCC differentiation or tumor marker (data not shown). These findings suggest that the reduction of Spred expression is closely linked to the malignant features of HCC.

Sprouty-related protein with Ena/vasodilator-stimulated phosphoprotein homology-1 domain-1 inhibits human hepatocellular carcinoma cell tumorigenesis

We then investigated whether Spreds exhibit inhibitory action on HCC cell growth, invasion and metastasis. First, we studied the effect of Spred-1 on ERK activation and cell proliferation. As shown in Figure 3a, forced expression of Spred-1 suppressed the HGF-induced phosphorylation of ERK activation but not the HGF-induced phosphorylation of Akt in HLF and KYN-2 HCC cells. Then we established HLF and KYN-2 cells stably expressing Spred-1 and C-terminal deletion mutant Spred-1 (ΔC-Spred-1), which functions as a dominant-negative form of Spred-1 and enhances ERK activation (Figure 3b). Overexpression of Spred-1 suppressed HLF cell proliferation and ERK activation, whereas ΔC-Spred-1 augmented cell proliferation (Figure 3c). To elucidate the effect of Spred-1 on tumorigenesis as in vivo HCC cell growth, we subcutaneously injected KYN-2 cells stably expressing Spred-1 and ΔC-Spred-1 into the flank of nude mice. Tumor growth was strikingly suppressed by the overexpression of Spred-1 and was consistent with the suppression of cell proliferation in vitro, whereas stably expressed ΔC-Spred-1 did not inhibit tumorigenesis (Tables 1 and 2). These data suggest that Spred-1 inhibits HCC cell proliferation through suppression of the ERK pathway.

Figure 3
figure3

(a) Sprouty-related protein with Ena/vasodilator-stimulated phosphoprotein homology-1 domain (Spred) inhibited hepatocyte growth factor (HGF)-mediated extracellular-regulated kinase (ERK) phosphorylation in a dose-dependent manner. HLF and KYN-2 cells transfected with myc-tagged wild-type Spred-1 were stimulated in HGF (50 ng/ml) for 3 h. Western blot analysis was performed using an anti-phospho-ERK1,2 antibody, anti-ERK2 antibody, anti-phospho-Akt antibody, anti-Akt antibody and anti-cMyc antibody. (b) Exogenous Spred expression inhibited human hepatocellular carcinoma (HCC) cell proliferation. HLF cells and KYN-2 cells stably expressing wild-type Spred-1 (myc-Spred-1) and a C-terminal Sprouty-related C-terminal cysteine-rich (SPR) domain deletion mutant (myc-ΔC Spred-1) were generated. These cells were plated into 35 mm dishes, and the number of cells was counted daily under G4 18 800 μg/ml addition. Data represent means of three measurements. (c) Stable expression of Spred in HCC attenuated ERK activation. Paretal HLF cell, stable transgene integrated control HLF cell (vector), myc-Spred HLF cell (wild-type Spred-1), and myc-ΔCSpred HLF cell (C-terminal SPR domain deletion mutant of Spred-1) were prepared for Western blot analysis.

Table 1 Clinicopathological features of patients with HCC
Table 2 Spred inhibited the tumorigenesis of HCC cells in vivo

Suppression of expression and secretion of matrix metalloproteinase-2 and matrix metalloproteinase-9 by Sprouty-related protein with Ena/vasodilator-stimulated phosphoprotein homology-1 domain-1

It is well established that the secretion of MMP-2 and -9 from cancer cells leads to hydrolysis of the extracellular matrix, thus enabling cells to break out of their primary site into the circulatory system and from there to metastatic sites. Previous studies have shown the stimulatory effect of growth factors and cytokines on MMP-2 and -9 expression in various cancer cells through activation of the Ras/Raf/ERK pathway. Thus, we investigated the effect of Spred-1 on MMP-2 and -9 expression in HCC cells. Transient Spred-1 overexpression as well as mitogen-induced extracellular kinase (MEK) inhibitor PD98059 (25 μ M), but not PI3K inhibitor LY294002 (10 μ M), inhibited the expression of MMP-2 and -9 mRNA (Figure 4a). We also examined the effect of Spred on the secretion of MMP-2 and -9 from HCC cells by zymography. As shown in the zymogram image in Figure 4b, HLF cells secreted MMP-2 and -9, and the secretion was enhanced by HGF treatment. Strikingly, the stable expression of Spred-1 resulted in the remarkable reduction of MMP-2 and -9 secretion. These data indicate that Spred-1 might prevent the metastasis of cancer cells by the reduction of MMP-2 and -9 expression through the regulation of the Ras/Raf/ERK pathway in HCC cells.

Figure 4
figure4

(a) Sprouty-related protein with Ena/vasodilator-stimulated phosphoprotein homology-1 domain (Spred) reduced human hepatocellular carcinoma (HCC) cell-mediated MMP production through the Ras/Raf-1/extracellular-regulated kinase (ERK) pathway. HLF and KYN-2 cells transfected with myc-tagged wild-type Spred-1 were stimulated in hepatocyte growth factor (HGF) (50 ng/ml) for 3 h. These cells were incubated for 12 h with PD98059 (15 μ M) and LY294002 (30 μ M) as the specific mitogen-induced extracellular kinase (MEK)1 inhibitor and the specific PI3K inhibitor, respectively. Reverse transcriptase–polymerase chain reaction (RT—PCR) analysis was performed using primers for MMP2, MMP9 and glyceraldehydes-3-phosphate dehydrogenase (GAPDH) antibodies. (b) Attenuated secretion of MMP-2, -9 in Spred overexpressed HCC cells. Each HLF stable transfectants were treated for 24 h with HGF (50 ng/ml). Collected media of each cell were performed with zymography assay. Similar results were observed in more than three independent experiments.

Sprouty-related protein with Ena/vasodilator-stimulated phosphoprotein homology-1 domain-1 regulates human hepatocellular carcinoma cell migration

To examine the role of Spred on HCC cell migration, we carried out Boyden chamber assay (Figure 5a). HLF cells stably expressing Spred-1 exhibited decreased cell migration compared to the control and ΔC-Spred-1 stably expressing HLF cells. We also examined cell motility by wound-healing assay. Consistent with Boyden-chamber assays, Spred-1 reduced HGF-induced HCC cell motility in KYN-2 cells (Figure 5b). It has been reported that HGF-induced cell migration requires both the Ras/Raf/ERK pathway and the PI3K/Akt pathway (Potempa and Ridley, 1998). Forced expression of Spred-1 inhibited HGF-mediated migration of HLF cells, although Spred does not regulate the PI3K/Akt pathway (Figure 3a). These findings suggest that Spred regulates HCC cell migration mainly through the Ras/Raf-1/ERK pathway.

Figure 5
figure5

(a) Sprouty-related protein with Ena/vasodilator-stimulated phosphoprotein homology-1 domain (Spred) suppressed human hepatocellular carcinoma (HCC) cell migration. In vitro migration assay: HLF cells stably expressing wild-type Spred-1 (myc-Spred-1) and a C-terminal Sprouty-related C-terminal cysteine-rich (SPR) domain deletion mutant (myc-ΔC Spred-1) were plated on the upper sides of transwells (5 × 104 cells/well). After hepatocyte growth factor (HGF) (50 ng/ml) was added to the lower chamber, cells were allowed to migrate for 24 h. Cells that migrated through the membrane were stained and measured. Data represent the mean s.d. of three measurements. (b) Spred suppressed HCC cell migration. Wound-healing assay: KYN-2 cells stably expressing Spred grown to subconfluence were scraped with a sharp edge to make a cell-free space. Cells were cultured for 24 h with HGF treatment (50 ng/ml), and cells that migrated into the scraped area were observed.

Discussion

Human hepatocellular carcinoma acquires biological malignant characteristics such as prominent growth, invasion, and metastasis with dedifferentiation from precancerous tumors into poorly differentiated tumors. The prognosis of patients with aggressive HCC tumors remains poor, although many advances in clinical therapy have been made. Therefore, prognostic molecular biomarkers of HCC can be invaluable for the clinical evaluation of patients and for tumor control. Still, molecular indicators for the HCC phenotype are not well understood. Although many of human HCC tumors lack the oncogenic Ras (V12Ras) mutation (Stahl et al., 2005), ERK-type mitogen-activated protein (MAP) kinase is highly activated high in HCC as shown in Figure 1. Therefore, in this study, we found that the expression of Spred-1 and -2, which are physiological inhibitors of the Ras/Raf-1/ERK pathway, was frequently decreased in human HCC tissue comparing with that in corresponding non-tumorous tissue. Our study also showed that the expression levels of Spred-1 and -2 were inversely correlated with tumor diameter. In addition, reduced expression of Spred-1 and -2 was associated with tumor invasion and intrahepatic metastasis as clinicopathological features of aggressive HCC.

The activation of the Ras/Raf-1/ERK pathway through stimulated growth-factor receptors or the binding of integrins to extracellular-matrix molecules mediates cell proliferation with the induction of cyclin D1 and the downregulation of endogenous cyclin-dependent kinase inhibitors such as p21/WAF/CIP1 and p27/KIP1(Vaudry et al., 2002). Inhibition of ERK by dominant-negative ERK or antisense nucleotides of ERK exhibits the suppression of cell growth (Bottazzi et al., 1999). Previous studies have shown that overexpression of ERK in HCC is correlated with its proliferative activity (Ito et al., 1998; Tsuboi et al., 2004). The kinase activity of ERK is fully activated through the phosphorylation of threonine (Thr 202) and tyrosine (Tyr 204) of ERK by MAP kinase kinase (MEK), which is activated by Raf. Recently, we reported that Spred-deficient cells showed increased cell proliferation with enhanced phosphorylation of ERK (Nonami et al., 2004). These findings suggest that the expression of Spred may have a profound effect on cell proliferation by regulating ERK activity.

In this study, we found that the forced expression of Spred inhibits HCC cell proliferation in vitro and in vivo, which is associated with attenuated phosphorylation of ERK. Therefore, we speculated that a deficiency of Spred in hepatocytes causes hyperproliferative cell, leading to tumorigenicity. However, neither Spred-1 null mice nor Spred-2 null mice developed any tumors, which was probably due to the redundancy in the physiological function of Spred-1 and -2 because Spred-1 and -2 are coexpressed in various organs and have similar effect on ERK activity (Wakioka et al., 2001; Kato et al., 2003). In our study, expression levels of Spred-1 and -2 were simultaneously decreased in identical HCC tumors at a high rate (22/27:68%). These findings suggest that the reduction of both Spred-1 and -2 is a crucial event in the promotion of human HCC.

The molecular mechanism for the induction of Spred has not yet been fully understood. The loss of Spred expression in HCC could be due to an epigenetic alteration in the gene itself, such as deletion or methylation. However, the coexpression of both Spred-1 and -2 in various tissues as well as the simultaneous downregulation of both Spred-1 and -2 in HCC indicates that Spred-1 and -2 are likely to be similarly regulated. Therefore, the downregulation of Spred-1 and -2 could be due to decreased transcription by the loss of essential transcription factors or the upregulation of negative regulatory factors. Further study is necessary to identify such transcription factors.

The Ras/Raf-1/ERK pathway is also involved in the determination of cell differentiation or proliferation (Marshall, 1995). We previously found that the level of Spred was increased during the differentiation of C2C12 myoblastic cells into myotubes, when ERK activity was decreased. Moreover, constitutive overexpression of Spred as well as dominant-negative Ras (N17-Ras) enhanced the differentiation of C2C12 myoblastic cells with the decreased phosphorylation of ERK. In contrast, dominant-negative Spred (ΔC-Spred and ΔN-Spred) did not differentiate myoblasts and maintained a higher level of ERK activity (Wakioka et al., 2001). In this study, we could not demonstrate whether Spred expression in HCC was associated with the degree of tumor differentiation because our samples mainly consisted of moderately differentiated HCC. However, the level of Spred expression was inversely correlated with clinicopathological features, such as metastasis and invasion, which reflect dedifferentiated tumors. Although the mechanism of dedifferentiation of HCC remains to be elucidated, Raf-1 plays a main role in the differentiation of a hepatoblast to a mature hepatocyte (Mikula et al., 2001; Chen et al., 2004). These findings suggest that the reduction of Spred expression may promote the dedifferentiation of HCC through deregulation of the Ras/Raf-1/ERK pathway. Examining HCC development in Spred-deficient mice using a liver carcinogenesis model might provide new insight into how Spred is involved in the development of human HCC.

MMPs, including MMP-2 and -9, are expressed in a wide variety of human cancers, including HCC, and their roles historically have been associated with invasive and metastatic phenotypes (Nelson et al., 2000). It has been shown that MMP-2 and -9 secretions are upregulated by a variety of growth factors via the Ras/Raf-1/ERK pathway and/or the PI3K/Akt pathway (Zhang et al., 2004). We found that forced expression of Spred-1 strikingly reduced the induction of MMP-2 and -9, suggesting that MMP induction is strongly regulated by the Ras/Raf-1/ERK pathway in addition the PI3K/Akt pathway in HCC cells. HCC tumors exhibiting invasion and metastasis have dysregulated MMP secretion, which might be due to decreased Spred expression. Additionally, Spred might be necessary for the inhibition of angiogenesis in HCC because the activation of MMP-9 coincides with the angiogenic switch in premalignant lesions (Coussens et al., 2000).

Aggressive HCC tumors cause intrahepatic extension and metastasis, resulting in liver failure and death. However, it is not well known how HCC cells adhesion and motility is promoted through the lost differentiated character.

In this study, we found that Spred expression in an HCC tumor was significantly and inversely correlated with the invasion and metastasis of HCC. Furthermore, forced expression of Spred regulated HGF-mediated HCC cell migration. The constitutive active form of Ras (V12Ras) is implicated in cell migration through actin cytoskeletal rearrangement. However, most human HCC tumors lack the oncogenic Ras (V12Ras) mutation (Stahl et al., 2005), and previous reports have suggested that the activation of ERK in HCC is associated mainly with tumor proliferation but not with invasion and metastasis (Ito et al., 1998). Alternatively, the small GTPases Rho and its effector, ROCK pathway is thought to regulate cytoskeletal remodeling and actomyosin contractility (Amano et al., 2000). Previously, we revealed that overexpression of Spred in osteosarcoma cells decreased stress fiber formation mediated by the Rho/ROCK pathway. Furthermore, we confirmed that overexpression of Spred resulted in a marked reduction of actin-stress fiber formation as well as cell migration in heaptoma cells (Figure 5 and data not shown). Thus, reduced expression of Spreds in HCC may contribute to metastasis by upregulating the Rho/ROCK pathway. However, a recent report suggests that Raf-1 physically associates with ROCK and is necessary for Rho/ROCK-regulated cell migration (Ehrenreiter et al., 2005). Therefore, Spred may regulate cell motility through the inhibiting the Raf-1 activation in additon to suppression of the Rho/ROKCK activation. Further study is necessary to clarify the molecular mechanism of Spred-mediated migration suppression by using Raf-1- or ROCK-deficient cells.

In summary, the data presented in this study reveal the importance of Spred in the negative regulation of HCC cell proliferation, invasion and metastasis. We have concluded that the expression level of Spred is a potential prognosis marker and the induction of Spred could be a novel, therapeutic way of dealing with HCC.

Materials and methods

Human tissue samples

Thirty-two paired samples of primary HCCs and their corresponding noncancerous liver tissues were obtained from patients who underwent hepatic resection between 1997 and 1999 at Kurume University Hospital, Fukuoka, Japan (Table 1). These tissue samples were stored in liquid nitrogen until used in experiments. None of the patients had received other treatments previously. Informed consent was obtained from all patients.

Reverse transcriptase–polymerase chain reaction

Total RNA was extracted and purified from frozen tissues using Trizol Reagent (Life Technologies Inc., Carlsbad, CA, USA) according to the manufacturer's instructions. We produced cDNAs from 2 μg of RNA by reverse transcriptase reaction using random hexamer and MULV transcriptase (Roche, Manheim, Germany). For semiquantitative RT–PCR, we used the following primers: Spred-1 (5′-IndexTermTGTGGTATTTAAGACGCAGCCTTCCTCA-3′ (forward) and 5′-IndexTermCATATTGCTTGCCAAAAGCTTCCACAAA-3′ (reverse)), Spred-2 (5′-IndexTermCTCCCTTTCCCCACTCCTTCTTTATTGC-3′ (forward) and 5′-IndexTermGGAGACCCTAGAGAAAGACCCCAAGGAA-3′ (reverse)), MMP-2 (5′-IndexTermACAGGTGATCTTGACCAGAATACCATCG-3′ (forward) and 5′-IndexTermCATTGAACAAGAAGGGGAACTTGCAGTA-3′ (reverse)), MMP-9 (5′-IndexTermACCTCGAACTTTGACAGCGACAAGAAGT-3′ (forward) and 5′-IndexTermCCATCCTTGAACAAATACAGCTGGTTCC-3′ (reverse)), and glyceraldehydes-3-phosphate dehydrogenase (GAPDH) (5′-IndexTermACCACAGTCCATGCCATCAC-3′ (forward) and 5′-IndexTermTCCACCACCCTGTTGCTGTA-3′ (reverse)).

The primers and TaqMan probes for real-time quantitative RT–PCR of Spred-1 and -2 were designed as follows: Spred-1 (5′-IndexTermCTGATCCCTGTTCGTGTGACA-3′ and 5′-IndexTermAGACAAAGCTACCAGGGCTAACC-3′ (reverse)), Spred-2 (5′-IndexTermGACCCCGAGGGAGACTATACAGA-3′ and 5′-IndexTermCCACCGGAGGCAAAACTTC-3′ (reverse)), and TaqMan probe (Spred-1: FAM-IndexTermAGCGACGACAAGTTCTGCTTGCGA, Spred-2: FAM-IndexTermCCTTGCTCGTGCGATACTAGCGACG). Primers and probes for GAPDH (TaqMan GAPDH control reagent kit, Foster City, CA, USA) were purchased from Perkin-Elmer Applied Biosystems. Real-time quantitative PCR was done using the ABI Prism 7000 Sequence Detection System (Perkin-Elmer Applied Biosystems, Foster City, CA, USA), as described previously (Yoshida et al., 2004).

Cell culture and transfection of Sprouty-related protein with Ena/vasodilator-stimulated phosphoprotein homology-1 domain

HLF cells and KYN-2 cells were established from human HCC tissues (Yano et al., 1988; Koga et al., 2001). These cells were cultured at 37°C in 5% CO2 in Dulbecco's modified Eagle's medium (Sigma, St Louis, MO, USA) supplemented with 10% fetal bovine serum, penicillin (100 U/l) and streptomycin (100 μg/ml). pcDNA3-Myc-tagged Spred-1 plasmid (Myc-Spred-1), Myc-tagged ΔC-Spred-1 plasmid (Myc-ΔC-Spred-1), and Myc-tagged N-terminal deletion mutant Spred-1 plasmid (Myc-ΔN-Spred-1) were prepared as described previously (Inoue et al., 2005). Transfection of plasmids was performed by lipofection using a Polyfect transfection reagent (QIAGEN, GmbH, Germany) or Lipofectamin2000 (Invitrogen, Carlsbad, CA, USA). HLF cells and KYN2 cells of stable transformants were cultured in the presence of 800 μg/ml of G418 to maintain transgene expression.

Immunoblot and immunohistochemistry

Immunoblot analysis was carried out by using antibodies for c-Myc, GAPDH (Santa Cruz Biotechnology, Delaware, CA, USA), phospho-ERK1,2 (Thr 202, Tyr 204), ERK-1,2, phospho-Akt (Ser 473), Akt (Cell Signaling, Beverly, MA, USA) and Spred-1 (Wakioka et al., 2001) as described (Sasaki et al., 2003). Immunohistochemistry was performed on 5 μm sections of formalin-fixed, paraffin-embedded tissue. For immunohistochemical staining of phospho-ERK1,2, the sections were put in a microwave for antigen retrieval. The sections were treated with 0.5% hydrogen peroxide in methanol for 10 min and blocked using protein block serum free reagent (DAKO, Glostrup, Denmark) for 15 min. The sections were incubated with antibodies for phosphor-ERK1,2 (Thr 202, Tyr 204) (20G11) and Spred-1 overnight at 4°C, and visualized with peroxidase-diaminobenzidine as a chromogen. After counterstaining with hematoxylin, the sections were dehydrated and permanently mounted.

Cell growth and in vivo tumorigenicity

Cells (1.0 × 105) were plated on six-well plates. At each time point, the cells were trypsinized, and viable cells were counted with trypan-blue stain. Each count was performed in triplicate. To test tumorigenicity in vivo, cells (1.0 × 106) in a total volume of 0.1 ml were injected subcutaneously into the flank of BALB/c nude mice (male, 6 weeks old: Charles River Lab). Tumor growth was monitored weekly for 8 weeks. Tumor size was determined by the product of two perpendicular diameters and the height above the skin surface.

Cell-migration assay and wound-healing assay

Cell migration was monitored using Boyden chambers (Chemicon, Temecula, CA, USA) according to the manufacturer's protocol. Briefly, cells (3.5 × 105) were seeded onto the upper-side transwells in 0.1 ml of a serum-free medium, and 0.5 ml of the same medium was added to the lower chamber. The cells were allowed to adhere for 1 h. Then HGF (50 ng/ml) was added to the lower chamber, and the cells were incubated to migrate for 24 h at 37°C. Cells that migrated on the lower-chamber side of the membrane were stained with Cell Stain and measured the optical density at 560 nm.

To evaluate cell motility, a wound-healing assay was also performed. Cells grown to subconfluence were scraped with a sharp edge to make a cell-free area. Cells that migrated into the scraped area were counted 24 h after scraping.

Zymography

Zymograms of MMP-2 and -9 were performed according to the manufacturer's protocol (Invitrogen). Briefly, cells were seeded to confluence, allowed to attach for 6 h, and then stimulated with HGF (50 ng/ml) for 24 h. The conditioned media were collected 24 h after stimulation, mixed with nonreducing Laemmli's sample buffer, and subjected to electrophoresis in a 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis gel containing 0.1% (W/V) gelatin. The gel was incubated with renaturing buffer for 30 min and developing buffer at 37°C for 12 h. The gel was stained with 0.25% coomassie blue (Liu et al., 2002).

Statistical analysis

For statistical analysis, we used the Student's t-test. A 95% confidence level (P<0.05) was taken to be significant.

References

  1. Amano M, Fukata Y, Kaibuchi K . (2000). Exp Cell Res 261: 44–51.

  2. Bottazzi ME, Zhu X, Bohmer RM, Assoian RK . (1999). J Cell Biol 146: 1255–1264.

  3. Chen L, Zeng Y, Yang H, Lee TD, French SW, Corrales FJ et al. (2004). FASEB J 18: 914–916.

  4. Coussens LM, Tinkle CL, Hanahan D, Werb Z . (2000). Cell 103: 481–490.

  5. Ehrenreiter K, Piazzolla D, Velamoor V, Sobczak I, Small JV, Takeda J et al. (2005). J Cell Biol 168: 955–964.

  6. Inoue H, Kato R, Fukuyama S, Nonami A, Taniguchi K, Matsumoto K et al. (2005). J Exp Med 201: 73–82.

  7. Ito Y, Sasaki Y, Horimoto M, Wada S, Tanaka Y, Kasahara A et al. (1998). Hepatology 27: 951–958.

  8. Johnson GL, Lapadat R . (2002). Science 298: 1911–1912.

  9. Kato R, Nonami A, Taketomi T, Wakioka T, Kuroiwa A, Matsuda Y et al. (2003). Biochem Biophys Res Commun 302: 767–772.

  10. Koga H, Sakisaka S, Harada M, Takagi T, Hanada S, Taniguchi E et al. (2001). Hepatology 33: 1087–1097.

  11. Liu JF, Crepin M, Liu JM, Barritault D, Ledoux D . (2002). Biochem Biophys Res Commun 293: 1174–1182.

  12. Marrero JA, Lok AS . (2004). Gastroenterology 127: S113–S119.

  13. Marshall CJ . (1995). Cell 80: 179–185.

  14. Masson V, de la Ballina LR, Munaut C, Wielockx B, Jost M, Maillard C et al. (2005). FASEB J 19: 234–236.

  15. McAllister SS, Becker-Hapak M, Pintucci G, Pagano M, Dowdy SF . (2003). Mol Cell Biol 23: 216–228.

  16. Mikula M, Schreiber M, Husak Z, Kucerova L, Ruth J, Wieser R et al. (2001). EMBO J 20: 1952–1962.

  17. Miyoshi K, Wakioka T, Nishinakamura H, Kamio M, Yang L, Inoue M et al. (2004). Oncogene 23: 5567–5576.

  18. Nelson AR, Fingleton B, Rothenberg ML, Matrisian LM . (2000). J Clin Oncol 18: 1135–1149.

  19. Nonami A, Kato R, Taniguchi K, Yoshiga D, Taketomi T, Fukuyama S et al. (2004). J Biol Chem 279: 52543–52551.

  20. Potempa S, Ridley AJ . (1998). Mol Biol Cell 9: 2185–2200.

  21. Prehoda KE, Lee DJ, Lim WA . (1999). Cell 97: 471–480.

  22. Ridley AJ, Hall A . (1992). Cell 70: 389–399.

  23. Ridley AJ, Paterson HF, Johnston CL, Diekmann D, Hall A . (1992). Cell 70: 401–410.

  24. Sasaki A, Taketomi T, Kato R, Saeki K, Nonami A, Sasaki M et al. (2003). Nat Cell Biol 5: 427–432.

  25. Schmidt CM, McKillop IH, Cahill PA, Sitzmann JV . (1997). Biochem Biophys Res Commun 236: 54–58.

  26. Stahl S, Ittrich C, Marx-Stoelting P, Kohle C, Altug-Teber O, Riess O et al. (2005). Hepatology 42: 353–361.

  27. Tsuboi Y, Ichida T, Sugitani S, Genda T, Inayoshi J, Takamura M et al. (2004). Liver Int 24: 432–436.

  28. Vaudry D, Stork PJ, Lazarovici P, Eiden LE . (2002). Science 296: 1648–1649.

  29. Volkman BF, Prehoda KE, Scott JA, Peterson FC, Lim WA . (2002). Cell 111: 565–576.

  30. Wakioka T, Sasaki A, Kato R, Shouda T, Matsumoto A, Miyoshi K et al. (2001). Nature 412: 647–651.

  31. Yano H, Maruiwa M, Murakami T, Fukuda K, Ito Y, Sugihara S et al. (1988). Acta Pathol Jpn 38: 953–966.

  32. Yoshida T, Ogata H, Kamio M, Joo A, Shiraishi H, Tokunaga Y et al. (2004). J Exp Med 199: 1701–1707.

  33. Zaragoza C, Soria E, Lopez E, Browning D, Balbin M, Lopez-Otin C et al. (2002). Mol Pharmacol 62: 927–935.

  34. Zhang D, Bar-Eli M, Meloche S, Brodt P . (2004). J Biol Chem 279: 19683–19690.

Download references

Acknowledgements

We thank Ms Masako Sinkawa for her excellent technical assistance and Ms Mieko Hatae and Ms Motoko Gotoh for manuscript preparation. This work was supported by the 21st Century COE Program for Medical Science of Japan.

Author information

Correspondence to T Yoshida.

Rights and permissions

Reprints and Permissions

About this article

Keywords

  • Ras
  • Raf-1
  • ERK
  • spred
  • sprouty
  • HCC

Further reading