High intensity ERK signal mediates hepatocyte growth factor-induced proliferation inhibition of the human hepatocellular carcinoma cell line HepG2.

Hepatocyte growth factor (HGF) induces growth stimulation of a variety of cell types, but it also induces growth inhibition of several types of tumor cell lines. The molecular mechanism of the HGF-induced growth inhibition of tumor cells remains obscure. We have investigated the intracellular signaling pathway involved in the antiproliferative effect of HGF on the human hepatocellular carcinoma cell line HepG2. HGF induced strong activation of ERK in HepG2 cells. Although the serum-dependent proliferation of HepG2 cells was inhibited by the MEK inhibitor PD98059 in a dose-dependent manner, 10 microM PD98059 reduced the HGF-induced strong activation of ERK to a weak activation; and as a result, the proliferation inhibited by HGF was completely restored. Above or below this specific concentration, the restoration was incomplete. Expression of constitutively activated Ha-Ras, which induces strong activation of ERK, led to the proliferation inhibition of HepG2 cells, as was observed in HGF-treated HepG2 cells. This inhibition was suppressed by the MEK inhibitor. Furthermore, HGF treatment and expression of constitutively activated Ha-Ras changed the hyperphosphorylated form of the retinoblastoma tumor suppressor gene product pRb to the hypophosphorylated form. This change was inhibited by the same concentration of MEK inhibitor needed to suppress the proliferation inhibition. These results suggest that ERK activity is required for both the stimulation and inhibition of proliferation of HepG2 cells; that the level of ERK activity determines the opposing proliferation responses; and that HGF-induced proliferation inhibition is caused by cell cycle arrest, which results from pRb being maintained in its active hypophosphorylated form via a high-intensity ERK signal in HepG2 cells.

Hepatocyte growth factor (HGF) induces growth stimulation of a variety of cell types, but it also induces growth inhibition of several types of tumor cell lines. The molecular mechanism of the HGF-induced growth inhibition of tumor cells remains obscure. We have investigated the intracellular signaling pathway involved in the antiproliferative effect of HGF on the human hepatocellular carcinoma cell line HepG2. HGF induced strong activation of ERK in HepG2 cells. Although the serum-dependent proliferation of HepG2 cells was inhibited by the MEK inhibitor PD98059 in a dose-dependent manner, 10 M PD98059 reduced the HGF-induced strong activation of ERK to a weak activation; and as a result, the proliferation inhibited by HGF was completely restored. Above or below this specific concentration, the restoration was incomplete. Expression of constitutively activated Ha-Ras, which induces strong activation of ERK, led to the proliferation inhibition of HepG2 cells, as was observed in HGF-treated HepG2 cells. This inhibition was suppressed by the MEK inhibitor. Furthermore, HGF treatment and expression of constitutively activated Ha-Ras changed the hyperphosphorylated form of the retinoblastoma tumor suppressor gene product pRb to the hypophosphorylated form. This change was inhibited by the same concentration of MEK inhibitor needed to suppress the proliferation inhibition. These results suggest that ERK activity is required for both the stimulation and inhibition of proliferation of HepG2 cells; that the level of ERK activity determines the opposing proliferation responses; and that HGF-induced proliferation inhibition is caused by cell cycle arrest, which results from pRb being maintained in its active hypophosphorylated form via a high-intensity ERK signal in HepG2 cells.
Hepatocyte growth factor (HGF) 1 is a mesenchymal cell-derived protein that is mitogenic for primary hepatocytes as well as other cell types (1)(2)(3)(4). HGF consists of a 62-kDa heavy chain and a 32-34-kDa light chain linked together by a disulfide bond (5,6). HGF is identical to scatter factor, which dissociates epithelial cell colonies into individual cells and induces a scattered fibroblastic morphology (7)(8)(9)(10). HGF also stimulates the migration of epithelial cells (11) as well as the invasion of carcinoma cells (12) and causes epithelial cells grown on three-dimensional collagen gels to form branching tubules (13). In addition, HGF protects liver progenitor cells from apoptosis (14). Thus, HGF is widely recognized as a multifunctional cytokine. In vivo, HGF is a potent angiogenic factor (15,16) and is involved in organ regeneration (17) and tumorigenesis (18). Analysis of mice lacking HGF revealed that HGF is essential for organogenesis during normal embryonic development (19 -21).
HGF is also identical to the fibroblast-derived tumor cytotoxic factor (22); and although HGF stimulates the growth of some tumor cell lines (23)(24)(25)(26)(27)(28)(29), the growth of a number of other tumor cell lines is inhibited (22-24, 26 -31). This growth inhibitory effect of HGF was also observed in vivo. When HGF transfectants of Fao hepatoma cells were transplanted into nude mice, the tumor formation of the transfectants was suppressed compared with the parental cells (31). Exogenous administration of HGF decreases the DNA synthesis of diethylnitrosamine-induced rat liver tumors (32). Furthermore, c-mycinduced hepatocarcinogenesis is inhibited by HGF in a transgenic mouse model coexpressing c-myc and HGF (33).
The diverse biological effects of HGF are transduced through the activation of its high-affinity receptor, the c-met protooncogene product (the c-Met receptor) (34 -37). The mature form of the c-Met receptor is a heterodimeric protein consisting of a 50-kDa extracellular ␣-subunit and a transmembrane 145-kDa ␤-subunit containing a tyrosine kinase domain in the cytoplasmic region (38,39). Binding of HGF to the c-Met receptor induces activation of the tyrosine kinase. The tyrosine kinase domain is highly phosphorylated at two tyrosine residues (Tyr 1234 and Tyr 1235 ) that are essential for the catalytic activity of the enzyme (40,41). In addition, two tyrosine residues located in the carboxyl-terminal region of the ␤-subunit (Tyr 1349 and Tyr 1356 ) are phosphorylated (42). These phosphorylated tyrosine residues provide binding sites for the Src homology 2 domain of the p85 regulatory subunit of phosphoinositide 3-kinase (PI3K), pp60 c-src , phospholipase C␥, and Shc (42). The Src homology 2 domain of Grb2 binds to Tyr 1356 (43). As a result of these interactions, HGF stimulates the activation of PI3K, pp60 c-src , phospholipase C␥, and Ras-ERK signaling pathways.
The molecular mechanism of the HGF-induced growth inhibition of tumor cells remains obscure. In this study, we ana-lyzed the HGF signaling pathways downstream of the c-Met receptor in the human hepatocellular carcinoma cell line HepG2, the proliferation of which is markedly inhibited by HGF. We found that both serum-dependent proliferation and HGF-induced proliferation inhibition of HepG2 cells were inhibited by blocking the MEK-ERK pathway, that HGF induced a strong activation of ERK, and that reduction of this activation to a weak activation restored the proliferation of HepG2 cells. We also found that HGF induced the hypophosphorylated form of pRb and that inhibition of ERK activation suppressed this induction. These findings suggest that stimulation and inhibition of proliferation of HepG2 cells are determined by the level of ERK activity and that the proliferation inhibition of HepG2 cells triggered by HGF results from cell cycle arrest caused by pRb being maintained in an active hypophosphorylated form through high-intensity ERK signaling.

EXPERIMENTAL PROCEDURES
Reagents-Reagents were obtained as follows: anti-Ha-Ras antibody from Santa Cruz Biotechnology; anti-ERK2 antibody from Upstate Biotechnology, Inc.; anti-pRb antibody from Pharmingen; horseradish peroxidase-conjugated anti-mouse and anti-rabbit immunoglobulins from Amersham Pharmacia Biotech; recombinant human HGF from the Research Center of Mitsubishi Chemical Corp.; PD98059 and LY294002 from Calbiochem; and U0126 from Promega.
Plasmids-Plasmid constructs pOPRSVI-Ha-Ras(G12V) and p3Јss were provided by Dr. Y. Kaziro (Tokyo Institute of Technology. The plasmid pTK-SEAP, which contains the secreted alkaline phosphatase (SEAP) gene under the control of the thymidine kinase basal promoter of the herpes simplex virus, was obtained from CLONTECH. pRRE (collagenase-1)-SEAP and pSRE(c-fos)-SEAP were constructed by inserting the Ras-responsive element (RRE) in the collagenase-1 promoter or the serum response element (SRE) in the c-fos promoter, respectively, into pTK-SEAP. pRRE(collagenase-1)-SEAP contains the SEAP reporter gene under the control of RREs, (AGAGGATGTTATA-AAGCATGAGTCAG) 4 , with essential Ets-binding sites. pSRE(c-fos)-SEAP contains the SEAP reporter gene under the control of SREs, (CCCTTACACAGGATGTCCATATTAGGACATCTGCGTCAGC) 4 , with Ets-binding sites to which the Ets subfamily ternary complex factors bind.
Cell Culture and Transfection-Parenchymal hepatocytes were isolated from an adult mouse by the modified in situ perfusion method as described (44). HepG2 cells, MKN74 cells, and hepatocytes were cultured in DMEM, RPMI 1640 medium, and William's medium E, respectively, containing 10% fetal bovine serum (FBS), 100 units/ml penicillin, and 100 g/ml streptomycin in a humidified atmosphere of 5% CO 2 at 37°C. The plasmid (2.2 g) was transfected into the cells using the FuGENE-6 reagent (Roche Molecular Biochemicals), and the transfected cells were cultured for 2 days. The cells were then cultured in selective medium containing hygromycin B (400 g/ml) or G418 (1 mg/ml).
Cell Proliferation Assay-Cells were seeded at a proper density of cells/well in 12-well plates and cultured with DMEM or RPMI 1640 medium containing 10% FBS for 1 day. Cells were subsequently treated with reagents and cultured further. The cells were harvested after trypsinization, and the number of cells were counted using a hemocytometer.
Measuring DNA Synthesis-DNA synthesis was analyzed by measuring the incorporation of the pyrimidine analog 5-bromo-2Ј-deoxyuridine (BrdUrd) during DNA synthesis. Cells were seeded at a proper density of cells/well in 96-well plates and cultured in DMEM (HepG2 cells), RPMI 1640 medium (MKN74 cells), or William's medium E (hepatocytes) containing 10% FBS. After the proper time, the medium was replaced with fresh medium, and cells were cultured in the presence or absence of HGF. BrdUrd was added to the cells 24 h before measuring DNA synthesis. After 24 h, the BrdUrd incorporation during DNA synthesis was measured using a cell proliferation enzyme-linked immunosorbent assay system (Amersham Pharmacia Biotech).
In-gel Kinase Assay-80 g of protein from the cell lysates were separated on SDS-12.5% polyacrylamide gel containing myelin basic protein (0.5 mg/ml; Sigma) as an ERK substrate. Gels were washed with 20% (v/v) 2-propanol and 50 mM Tris-HCl, pH 8.0, to remove SDS and then with 50 mM Tris-HCl, pH 8. After extensive washing with 5% (w/v) trichloroacetic acid and 1% (w/v) tetrasodium pyrophosphate, the gels were dried, and phosphorylation of myelin basic protein was detected by autoradiography.
SEAP Reporter Assay-Cells were seeded at a density of 1.25 ϫ 10 5 cells/well in six-well plates containing DMEM. After 24 h, the medium was replaced with fresh medium, and plasmid pTK-SEAP, pRRE(collagenase-1)-SEAP, or pSRE(c-fos)-SEAP (2.2 g) was transfected into the cells using the FuGENE-6 reagent. At 24 h after transfection, the cells were pretreated with PD98059 (0 -50 M) for 1 h. The cells were subsequently treated with HGF (50 ng/ml) or isopropyl-␤-D-thiogalactopyranoside (IPTG; 5 mM) for 2 days. The culture medium was then harvested, and detached cells were removed by centrifugation. The SEAP assay was performed using the SEAP reporter assay kit (Toyobo), and chemiluminescence was quantitated on a Wallac 1420 ARVOsx multi-label counter.

Effect of HGF on the Proliferation of HepG2 and MKN74
Cells-It has been shown that the proliferation of the human hepatocellular carcinoma cell line HepG2 is markedly inhibited by HGF (26,27,31). Thus, we used this cell line in this study to examine the mechanism of growth inhibition of tumor cells by HGF. For comparison, we also used the human gastric carcinoma cell line MKN74, the proliferation of which is enhanced by HGF (45), because nontransformed primary cultured hepatocytes are not suitable for counting cell numbers, which is thought to be a proper method for evaluating cell proliferation. To confirm the opposing proliferation responses of HepG2 and MKN74 cells to HGF, these cells were cultured in medium containing 10% FBS in the presence or absence of HGF, and the cell numbers were counted. Fig. 1A shows a time course of the effect of HGF (50 ng/ml) on the proliferation of HepG2 and MKN74 cells. The proliferation of HepG2 cells was inhibited by HGF. The inhibitory effect first appeared at 3 days and became marked at 4 days. The number of cells was ϳ0.45fold less compared with HGF-untreated control cells at 4 days. On the other hand, the proliferation of MKN74 cells was stimulated by HGF. The stimulatory effect first appeared at 2 days and became marked at 3 and 4 days. The number of cells was ϳ2-fold greater compared with HGF-untreated control cells at 4 days. These results confirmed the opposing proliferation responses of HepG2 and MKN74 cells to HGF. In addition to affecting proliferation, HGF induced scattering of cell colonies in HepG2 and MKN74 cells and hepatocytes (data not shown).
Next, we examined the effect of HGF on DNA synthesis of HepG2 and MKN74 cells by measuring BrdUrd incorporation during DNA synthesis. The DNA synthesis of HepG2 cells was inhibited by HGF (Fig. 1B). On the other hand, similar to primary cultured hepatocytes, the DNA synthesis of MKN74 cells was stimulated by HGF (Fig. 1B). These results indicate that the effect of HGF on the proliferation (counting cell numbers) of HepG2 and MKN74 cells correlates well with the effect on DNA synthesis (measuring BrdUrd incorporation). Thus, HepG2 and MKN74 cells were used to investigate the intracellular signaling pathway involved in the opposing effects of HGF on cell proliferation.
Effect of MEK-specific Inhibitors on the Proliferation of HepG2 and MKN74 Cells Treated with HGF-Since the MEK-ERK pathway plays a central role in signaling cell growth (46,47), it may be involved in the proliferation inhibition of HepG2 cells triggered by HGF. To examine whether the MEK protein is involved in the HGF-induced proliferation inhibition of HepG2 cells, the effects of specific inhibitors (PD98059 and U0126) (48 -50) of MEK, an upstream kinase of ERK, were tested. Cells were pretreated with the MEK-specific inhibitors and then cultured in medium containing 10% FBS in the presence or absence of HGF. The proliferation of MKN74 cells observed in both the presence and absence of HGF was inhibited by PD98059 in a dose-dependent manner (Fig. 2B). The proliferation of HepG2 cells cultured in the absence of HGF was inhibited by PD98059, similar to MKN74 cells, whereas the proliferation of HepG2 cells inhibited by HGF was restored by PD98059 ( Fig. 2A). It should be noted that PD98059 completely restored the proliferation inhibited by HGF at a con-centration of 10 M, whereas above or below 10 M, the restoration was incomplete. The growth curve for cells cultured in the presence of PD98059 (10 M) and HGF was similar to that for untreated cells (data not shown). A similar restoration of HepG2 cell proliferation was observed with U0126, and a concentration of 2 M completely restored the proliferation inhibited by HGF (Fig. 2C). These results suggest that MEK activity is essential for the serum-induced proliferation stimulation of HepG2 cells and the serum-and HGF-induced proliferation stimulation of MKN74 cells and, in addition, that it is also required for the HGF-induced proliferation inhibition of HepG2 cells.
It has been shown that the PI3K pathway is involved in cell proliferation (51,52). Thus, to examine whether the PI3K pathway participates in the HGF-induced proliferation inhibition of HepG2 cells, the effect of a PI3K-specific inhibitor (LY294002) was tested. The proliferation of HepG2 cells cultured in both the presence and absence of HGF was inhibited by LY294002 in a dose-dependent manner. The proliferation inhibition induced by HGF was not affected by LY294002 (Fig. 2D). These results suggest that PI3K is not necessary for the HGF-induced proliferation inhibition of HepG2 cells, although it is required for the serum-induced proliferation stimulation. In addition, we also examined whether the stress MAPK pathway is involved in the HGF-induced proliferation inhibition, using a p38-specific inhibitor (SB203580) (53). The proliferation inhibition induced by HGF was not affected by SB203580 (data not shown). Activity of ERK2 in HepG2 and MKN74 Cells Treated with HGF-ERK is the main downstream target of MEK (46,47), which phosphorylates and activates ERK. Thus, to determine whether ERK participates in the HGF-induced proliferation inhibition of HepG2 cells, we examined the phosphorylation and activity of ERK2 in HepG2 and MKN74 cells treated with HGF by immunoblot analysis and in-gel kinase assay, respectively. ERK2 was phosphorylated by treatment with HGF, and this phosphorylation lasted for 12 h (Fig. 3, panels a, ϪPD98059) and reached almost the basal level after 48 h (data not shown) in both HepG2 and MKN74 cells. ERK2 activity was more strongly stimulated in HepG2 cells treated with HGF than in MKN74 cells (Fig. 3, panels b, ϪPD98059). PD98059 at a concentration of 10 M, which completely restored the proliferation of HepG2 cells inhibited by HGF, also reduced the phosphorylation and activity of ERK2 in HepG2 cells stimulated with HGF, but low levels of ERK2 phosphorylation and activity were still detected (Fig. 3). In addition, 10 M PD98059 reduced the phosphorylation and activity of ERK2 in MKN74 cells treated with HGF, although the reduction in activity was not as prominent (Fig. 3). The phosphorylation and activity of ERK2 were reduced by PD98059 in a dose-dependent manner, and they were slightly lower in the presence of 30 M PD98059 than in the presence of 10 M PD98059 and were not detected in the presence of 50 M PD98059 (data not shown). These results suggest that the strong activation of ERK induces the proliferation inhibition, whereas the weak activation of ERK induces the proliferation stimulation of HepG2 cells.
Correlation between HGF Dose-dependent Inhibition of Proliferation and Activation Level of ERK2 in HepG2 Cells-The correlation between the proliferation response of cells to HGF and the level of ERK activity induced by HGF was suggested by the experiments using specific inhibitors of MEK. To obtain further support for this correlation, we examined the relationship between HGF dose-dependent proliferation inhibition and the activation level of ERK2 in HepG2 cells. This is a more direct assessment that can avoid side effects of long-term treatment of inhibitors. Cells were cultured in medium containing 10% FBS in the presence of various concentrations of HGF, and the cell numbers were counted. The proliferation of HepG2 cells was inhibited by HGF in a dose-dependent manner. The significant inhibitory effect first appeared at 1 ng/ml, and maximal inhibition was at 30 ng/ml and was maintained up to at least 100 ng/ml (Fig. 4A). The phosphorylation and activity of ERK2 in HepG2 cells treated with various concentrations of HGF were examined by immunoblot analysis and in-gel kinase assay, respectively. ERK2 phosphorylation and activation higher than the serum-induced level or the PD98059 (10 M)inhibited level in the presence of HGF were first detected at 1 ng/ml, and they were stimulated by HGF in a dose-dependent manner (Fig. 4B). It should be noted that the concentration of HGF (1 ng/ml) at which the proliferation inhibitory effect first appeared was equivalent to the concentration of HGF (1 ng/ml) that first stimulated the phosphorylation and activation of ERK2 to a higher level compared with the serum-induced level or the PD98059 (10 M)-inhibited level in the presence of HGF. These results indicate that the level of ERK activity induced by serum leads to the proliferation stimulation of HepG2 cells, whereas cell proliferation inhibition begins when ERK activity exceeds this level.
Constitutively Activated Ha-Ras Inhibits the Proliferation of HepG2 Cells-To obtain further support for the suggestion that the strong activation of ERK is required for the proliferation inhibition of HepG2 cells, we stably expressed constitutively activated Ha-Ras, which induces strong activation of ERK (54), in HepG2 cells and analyzed its effect on the proliferation of the cells. To obtain HepG2 cell clones expressing constitutively activated Ha-Ras, an IPTG-inducible system was used because the expression of constitutively activated Ha-Ras may affect the proliferation of the cells. An expression plasmid encoding a Lac repressor protein (p3Јss) (55) was transfected into HepG2 cells. Cells resistant to hygromycin B were selected and expanded. The expression of the repressor protein was assessed by immunoblot analysis (data not shown). A clone that expressed a high level of Lac repressor protein was selected and used for further analysis. Inhibition of the proliferation of this clone by HGF was more prominent than that of uncloned HepG2 cells, and IPTG had no effect on the proliferation (Fig.  5B, Parental cells). An expression plasmid (pOPRSVI-Ha-Ras(G12V)) encoding a Ha-Ras mutant in which glycine 12 is substituted with valine (56) was transfected into the HepG2 cells expressing a high level of Lac repressor protein. Clones resistant to G418 were selected and expanded. Induced expression of the Ha-Ras(G12V) protein in these clones was assessed by immunoblot analysis. Two clones (V1 and V2) in which the level of Ha-Ras(G12V) protein was increased by treatment with IPTG were selected (Fig. 5A).
The proliferation of V1 cells was inhibited by the induced expression of Ha-Ras(G12V), as was observed with HGF treatment (Fig. 5B). The level of proliferation stimulated by serum in V2 cells was lower than that in V1 cells. However, the proliferation of V2 cells was inhibited by the induced expression of Ha-Ras(G12V), although the inhibitory level was lower than that with HGF treatment (Fig. 5B). This lower level of inhibition may be due to a lower expression of Ha-Ras(G12V). These results indicate that sustained activation of Ha-Ras is able to induce the proliferation inhibition of HepG2 cells, similar to HGF treatment. The induced expression of Ha-Ras(G12V) also induced scattering of cell colonies in both V1 and V2 cells (data not shown).
Effect of PD98059 on the Proliferation and ERK-mediated Transcriptional Activation of V1 Cells Treated with HGF and/or IPTG-To examine whether MEK activity is involved in the proliferation inhibition of HepG2 cells induced by constitutively activated Ha-Ras, the effect of PD98059 on V1 cells treated with HGF and/or IPTG was tested. Cells were pretreated with PD98059 and cultured in medium containing 10% FBS in the presence or absence of HGF and/or IPTG. The proliferation of V1 cells cultured in the absence of HGF and IPTG was inhibited by PD98059 in a dose-dependent manner. PD98059 restored the proliferation inhibited by HGF or Ha-Ras(G12V) expression alone or both treatments together. The most effective concentrations of PD98059 were 10 and 5 M for HGF treatment and induced expression of Ha-Ras(G12V), respectively. The concentration of PD98059 shifted to 30 M when both treatments were performed. At these specific concentrations, PD98059 stimulated the proliferation slightly more than the proliferation induced by serum. Above or below these concentrations, the restoration was incomplete (Fig. 6A). These results suggest that MEK activity is involved in the proliferation inhibition of HepG2 cells induced by constitutively activated Ha-Ras.
To determine whether ERK is involved in the Ha-Ras(G12V)induced proliferation inhibition and to examine the mechanism of the PD98059 concentration dependence on the complete restoration of proliferation inhibited by the various treatments, we analyzed the effect of PD98059 on transcriptional activation controlled by the RRE of the collagenase-1 promoter and the SRE of the c-fos promoter, which are mediated by ERK (57-59). To analyze transcriptional activation, we made constructs that contained the SEAP reporter gene under the control of RRE (collagenase-1) or SRE(c-fos). Transcriptional activity controlled by RRE(collagenase-1) and SRE(c-fos) was enhanced by HGF and the induced expression of Ha-Ras(G12V), and this activity was reduced by PD98059 in a dose-dependent manner. The same concentrations of PD98059 most effective in restoring the proliferation of HepG2 cells inhibited by HGF (10 M), the induced expression of Ha-Ras(G12V) (5 M), or both (30 M) were also required to reduce the transcriptional activation induced by the respective treatments to near the serum-induced level (Fig. 6B, panels b and c). Transcriptional activation by the thymidine kinase promoter as a negative control was largely unaffected by treatment with HGF, IPTG, or PD98059 (Fig. 6B, panel a). These results suggest that the proliferation inhibition of HepG2 cells induced by constitutively activated Ha-Ras is mediated through the ERK pathway. In addition, the differences in the most effective concentrations of PD98059 required to restore the proliferation of V1 cells following HGF treatment, the induced expression of Ha-Ras(G12V), or both may result from the different amounts of PD98059 required to suppress the varying degrees of ERK activity induced by the treatments to a level that leads to proliferation stimulation.
Phosphorylation State of pRb-Since it is conceivable that the HGF-induced proliferation inhibition results from cell cycle arrest, we analyzed the phosphorylation state of the retinoblastoma tumor suppressor gene product pRb. pRb is an active transcriptional repressor that binds to transcription factors such as members of the E2F family (60 -65). There are three phosphorylated forms of pRb: unphosphorylated, hypophosphorylated, and hyperphosphorylated (66,67). Cyclin D-Cdk4/6 complexes activate pRb in early G 1 by hypophosphorylation (68 -70). The active hypophosphorylated pRb binds to transcription factors such as E2F proteins. Cyclin E-Cdk2 complexes perform the initial inactivating hyperphosphorylation of pRb at the late G 1 restriction point transition (68 -73). Inactivation of pRb by hyperphosphorylation in the late G 1 phase causes the release of E2F proteins and activation of transcription of E2F-regulated genes important for DNA synthesis, resulting in entry into the S phase (74). Thus, in cycling cells, pRb alternates between a hypophosphorylated form present in the early G 1 phase and a hyperphosphorylated form after passage through the restriction point in late G 1 and continuing through FIG. 3. Effect of PD98059 on ERK2 phosphorylation and activity in HepG2 and MKN74 cells stimulated with HGF. Cells were seeded at a density of 5 ϫ 10 5 cells/100-mm dish in DMEM/FBS for HepG2 cells and 1 ϫ 10 6 cells/100-mm dish in RPMI 1640 medium/FBS for MKN74 cells. After 24 h, the medium was replaced, and cells were cultured for 2 days. The medium was replaced with fresh medium; and following pretreatment with or without PD98059 (10 M) for 1 h, the cells were stimulated with HGF (50 ng/ml), and their lysates were prepared at the indicated times. Proteins (30 g) from the cell lysates were separated by SDS-10% PAGE and transferred to nitrocellulose membrane. Immunoblot analysis was performed with an anti-ERK2 antibody. Phosphorylation of ERK2 proteins is indicated by a shift to a slower electrophoretic mobility (panels a). The phosphotransferase activity of ERK2 was determined by an in-gel kinase assay and is indicated by phosphorylation of myelin basic protein as an ERK2 substrate (panels b). Similar results were obtained in at least three (immunoblot analysis) or two (in-gel kinase assay) independent experiments. the S, G 2 , and M phases. Thus, the hyperphosphorylation of pRb is a hallmark of G 1 -S transition. The hyperphosphorylated pRb was predominant in HepG2 cells cultured in the presence of serum. Hyperphosphorylated pRb was decreased and hypophosphorylated pRb was increased in HepG2 cells treated with HGF. Hyperphosphorylated pRb was restored by pretreatment with 10 M PD98059, which suppressed the effect of HGF on HepG2 cell proliferation (Fig. 7A). Similar results were obtained with V1 cells. Hyperphosphorylated pRb was decreased and hypophosphorylated pRb was increased in V1 cells treated with HGF or IPTG. Hyperphosphorylated pRb was restored by pretreatment with 5 M PD98059, which suppressed the effect of the induced expression of Ha-Ras(G12V) on V1 cell proliferation (Fig. 7B). These results suggest that HGF maintains pRb in the hypophosphorylated form by inhibition of hyperphosphorylation or stimulation of dephosphorylation of the hyperphosphorylated form via a strong activation of ERK and that the HGF-induced proliferation inhibition is caused by cell cycle arrest at late G 1 phase. DISCUSSION The ERK pathway is generally thought to be the central signaling cascade of growth stimulation by growth factors. In this study, we showed that HGF induced ERK activity in MKN74 cells, the proliferation of which is stimulated by HGF. In addition, we found that a MEK inhibitor blocked the proliferation stimulation, suggesting that the ERK pathway plays a crucial role in the HGF-induced proliferation stimulation of the human gastric carcinoma cell line. Similar to proliferation stimulation, we also showed that HGF induced ERK activity in HepG2 cells, the proliferation of which is inhibited by HGF, and that a reduction of this induction by MEK inhibitors suppressed the HGF-induced proliferation inhibition. Furthermore, expression of constitutively activated Ha-Ras, which induces activation of the ERK pathway, led to the proliferation inhibition of HepG2 cells, and a MEK inhibitor suppressed this proliferation inhibition. Thus, the ERK pathway also appears to be essential for the proliferation inhibition of the human hepatocellular carcinoma cell line induced by HGF.
ERK activity, as assessed by an in-gel kinase assay, was much higher in HGF-treated HepG2 cells than that in HGFtreated MKN74 cells. Moreover, when a MEK inhibitor reduced the HGF-induced ERK activity in HepG2 cells to a level similar to that found in HGF-treated MKN74 cells, the proliferation of HepG2 cells was restored to a level similar to that of HGFuntreated cells. In addition, HGF induced proliferation inhibition and ERK activation in a dose-dependent manner, and the proliferation inhibitory effect appeared when ERK activity ex-

FIG. 4. Correlation between HGF dose-dependent proliferation inhibition and the activation level of ERK2 in HepG2 cells.
A, HGF dose-dependent proliferation inhibition of HepG2 cells. Cells were seeded at a density of 5 ϫ 10 4 cells/well (12-well plates) in DMEM/ FBS. After 24 h, the medium was replaced with fresh medium, and cells were cultured in the presence of the indicated concentrations (0, 1, 5, 10, 15, 30, 50, and 100 ng/ml in the left panel and 0, 0.01, 0.1, and 1 ng/ml in the right panel) of HGF for 4 days. Cell numbers were counted in the presence of trypan blue. The viable cell number is indicated. Each value represents the mean Ϯ S.D. of triplicate determinations from a representative experiment. (The significant inhibitory effect first appeared at 1 ng/ml (p Ͻ 0.02).) Similar results were obtained in at least three independent experiments. B, HGF dose-dependent activation of ERK2 in HepG2 cells. Cells were seeded at a density of 5 ϫ 10 5 cells/100-mm dish in DMEM/FBS. After 24 h, the medium was replaced, and cells were cultured for 2 days. The medium was replaced then with fresh medium; cells were stimulated with the indicated concentrations of HGF for 5 min; and their lysates were prepared. Proteins (30 g) from the cell lysates were separated by SDS-10% PAGE and transferred to nitrocellulose membrane. Immunoblot analysis was performed with an anti-ERK2 antibody. Phosphorylation of ERK2 proteins is indicated by a shift to a slower electrophoretic mobility (panel a). The phosphotransferase activity of ERK2 was determined by an in-gel kinase assay and is indicated by phosphorylation of myelin basic protein as an ERK2 substrate (panel b). Similar results were obtained in at least three (immunoblot analysis) or two (in-gel kinase assay) independent experiments.

FIG. 5. Constitutively activated Ha-Ras inhibits the proliferation of HepG2 cells.
A, immunoblot analysis of Ha-Ras(G12V) protein expression in HepG2 cells. HepG2 cells possessing Ha-Ras(G12V) under the control of the lac operator (V1 and V2 cells) were seeded at a density of 3 ϫ 10 5 cells/100-mm dish in DMEM/FBS. After 24 h, the medium was replaced with fresh medium, and cells were cultured in the presence (ϩ) or absence (Ϫ) of IPTG (5 mM) before their lysates were prepared at the indicated times. Proteins (30 g) from the cell lysates were separated by SDS-12.5% PAGE and transferred to nitrocellulose membrane. Immunoblot analysis was performed with an anti-Ha-Ras antibody. B, effect of the induced expression of Ha-Ras(G12V) on the proliferation of HepG2 cells. Cells were seeded at a density of 3 ϫ 10 4 cells/well (12-well plates) in DMEM/FBS. After 24 h, the medium was replaced with fresh medium, and cells were cultured in the presence (ϩ) or absence (Ϫ) of IPTG (5 mM) for 2 days. The medium was then replaced with fresh medium, and cells were cultured in the presence or absence of IPTG (5 mM) and HGF (50 ng/ml) for 4 days. Cell numbers were counted in the presence of trypan blue. The viable cell number is indicated. Each value represents the mean Ϯ S.D. of triplicate determinations from a representative experiment. Similar results were obtained in at least three independent experiments. ceeded the level reduced by MEK inhibitors that led to complete restoration of proliferation inhibited by HGF. The most effective concentrations of the MEK inhibitor PD98059 required to suppress the proliferation inhibition of HGF-treated HepG2 cells or HepG2 cells expressing constitutively activated Ha-Ras (Ha-Ras(G12V)) were 10 and 5 M, respectively, whereas the concentration of PD98059 needed to suppress the proliferation inhibition of Ha-Ras(G12V)-expressing HepG2 cells treated with HGF shifted to 30 M. Similar concentrations of PD98059 were required to suppress the activity of ERK signaling to the serum-induced level found in HGF-untreated HepG2 cells. These results suggest that the level of ERK signaling activity has the range suitable for leading to proliferation stimulation and that when ERK signaling activity is below this range, proliferation stimulation is not induced; on the other hand, when ERK activity exceeds this range, proliferation inhibition is induced. Thus, the level of ERK activity may determine the proliferation response of cells to HGF.
The diverse biological effects of HGF are transduced through activation of the c-Met receptor. The expression and phosphorylation level of c-Met were not significantly different between HepG2 and MKN74 cells treated with HGF (data not shown). It has been reported that there is no significant correlation between the multifunctional activity of HGF and the number of cell-surface c-Met receptors (30). Thus, it is conceivable that differences in post-receptor signal transduction are responsible for the opposing proliferation responses of HepG2 and MKN74 cells to HGF. Expression of constitutively activated Ha-Ras (Ha-Ras(G12V)) induced the proliferation inhibition of HepG2 cells, and the level of the inhibition was similar to that of HepG2 cells treated with HGF. Furthermore, the proliferation inhibition was suppressed by treatment with a MEK inhibitor. These results suggest that the proliferation inhibition of HepG2 cells induced by HGF is mediated through the Ras protein. It has been shown that activated Raf, a downstream target of Ras, induces the proliferation inhibition of fibroblasts (75,76) and small lung cancer cells (77), suggesting that high-FIG. 6. Effects of PD98059 on the proliferation and ERK-mediated transcriptional activation of V1 cells treated with HGF and/or IPTG. A, effects of PD98059 on proliferation. Cells were seeded at a density of 3 ϫ 10 4 cells/well (12-well plates) in DMEM/FBS. After 24 h, the medium was replaced with medium containing the indicated concentrations of PD98059. After 1 h, HGF (50 ng/ml) and IPTG (5 mM) were added to the medium, and cells were cultured in the absence (E) or presence of HGF (q), IPTG (OE), or HGF and IPTG (f) for 6 days. (The medium was replaced with fresh medium containing these reagents at 3 days.) Cell numbers were counted in the presence of trypan blue. The viable cell number is indicated. Each value represents the mean Ϯ S.D. of triplicate determinations from a representative experiment. Similar results were obtained in at least three independent experiments. B, effects of PD98059 on ERK-mediated transcriptional activation. V1 cells were transiently transfected with plasmid pTK-SEAP (panel a), pRRE(collagenase-1)-SEAP (panel b), or pSRE(c-fos)-SEAP (panel c). At 24 h after transfection, the cells were pretreated with PD98059 (0 -50 M) for 1 h; HGF (50 ng/ml) or IPTG (5 mM) was subsequently added to the medium; and cells were cultured for 2 days in the absence (white bars) or presence of HGF (50 ng/ml; black bars), IPTG (5 mM; striped bars), or HGF and IPTG (hatched bars) for 2 days. SEAP assay was performed as described under "Experimental Procedures." The average -fold increase compared with untreated cells (indicted by the lines) is indicated. Each value represents the mean Ϯ S.D. of triplicate determinations from a representative experiment. Similar results were obtained in at least three independent experiments. FIG. 7. Immunoblot analysis of the phosphorylation state of pRb treated with HGF or IPTG in HepG2 and V1 cells. Cells were seeded at a density of 3 ϫ 10 5 cells/100-mm dish in DMEM/FBS. After 24 h, the medium was replaced with fresh medium, and cells were cultured for 2 days. The medium was replaced with fresh medium; and following preincubation with (ϩ) or without (Ϫ) PD98059 (5 or 10 M) for 1 h, HGF (50 ng/ml) or IPTG (5 mM) was added. Cells were then cultured (the medium was replaced with the fresh medium at 2 days) with or without HGF or IPTG, and their lysates were prepared at the indicated times. Proteins (30 g) from the cell lysates were separated by SDS-8% PAGE and transferred to a polyvinylidene difluoride membrane. Immunoblot analysis was performed with an anti-pRb antibody. The pRb phosphorylation state in HepG2 (A) and V1 (B) cells was determined. Hyperphosphorylated pRb is indicated by a shift to a slower electrophoretic mobility. Similar results were obtained in at least three independent experiments. intensity Ras-ERK signaling induces proliferation inhibition in several types of cells. Thus, HGF treatment of HepG2 cells is likely to induce high-intensity signaling of the Ras-ERK pathway, which leads to the proliferation inhibition of the cells. Alternatively, it is possible that a pathway other than the Ras-Raf pathway may induce the strong activation of ERK in HGF-treated HepG2 cells.
Two mechanisms could cause the proliferation inhibition of HepG2 cells induced by HGF: one is the result of apoptosis, and the other is cell cycle arrest. Although it has been shown that HGF induces apoptosis of several cells (78 -82), we did not detect any DNA fragmentation or chromatin condensation, which are indexes of apoptosis, in HGF-treated HepG2 cells (data not shown). Thus, the HGF-induced proliferation inhibition may not result from apoptosis. On the other hand, we found that the hyperphosphorylated form of pRb was decreased and the active hypophosphorylated form of pRb was increased in HepG2 cells treated with HGF or expressing constitutively activated Ha-Ras and that the situation was reversed by a MEK inhibitor. These results suggest that the HGF-induced proliferation inhibition is caused by cell cycle arrest at the late G 1 phase and that this cell cycle arrest is mediated through high-intensity ERK signaling. Active Raf induces expression of Cdk inhibitors in fibroblasts and small lung cancer cells (75)(76)(77), resulting in the proliferation inhibition of these cells. Moreover, it has been shown that HGF induces expression of the Cdk inhibitors p21 and p27 in HepG2 cells (83,84). Thus, high-intensity ERK activity in HepG2 cells treated with HGF may induce expression of Cdk inhibitors, which in turn leads to increases in the hypophosphorylated form of pRb. However, our preliminary results showed that no expression of the Cdk inhibitors p21 and p27 was induced in HepG2 cells treated with HGF. Transforming growth factor-␤ induces the proliferation inhibition of HepG2 cells that is led by cell cycle arrest. Transforming growth factor-␤ treatment leaves pRb in a hypophosphorylated form without expression of Cdk inhibitors in HepG2 cells. Down-regulation of a 45-kDa human protein with Cdkactivating kinase activity is responsible for this cell cycle arrest (85). Thus, it is possible that the proliferation inhibition of HepG2 cells by HGF is mediated by a similar mechanism, although the Ras-ERK signaling may not be involved in the transforming growth factor-␤-induced proliferation inhibition of HepG2 cells. Further examination is required to identify molecules that link the strongly activated ERK to the hypophosphorylated form of pRb.
HGF is a potent mitogen for hepatocytes, whereas it inhibits the growth of a number of hepatoma-derived cells. In this study, we have shown that the proliferation inhibition of HepG2 cells is led by a strong activation of the ERK pathway. The proliferation of the human hepatocellular carcinoma cell line HuH7 is inhibited by HGF (31). We found that PD98059 restored the proliferation of HuH7 cells inhibited by HGF (data not shown), suggesting that strong activation of the ERK pathway mediates the proliferation inhibition of HuH7 cells. Thus, the antiproliferative effect of HGF on hepatoma-derived cells appears to be mediated through a strong activation of the ERK pathway. Further studies are needed to understand whether the same signaling pathway leads to the HGF-induced proliferation inhibition of other cancer-derived cells.