Oncogene (2003) 22, 4656–4663 & 2003 Nature Publishing Group All rights reserved 0950-9232/03 $25.00 www.nature.com/onc

Downregulation of -associated protein tyrosine phosphatase-1 (SAP-1) in advanced human hepatocellular

Hidenobu Nagano1,2, Tetsuya Noguchi*,1, Kenjiro Inagaki1, Seitetsu Yoon1, Takashi Matozaki3, Hiroshi Itoh2, Masato Kasuga1 and Yoshitake Hayashi2

1Division of , Digestive and Kidney Diseases, Department of Clinical Molecular Medicine, Kobe University Graduate School of Medicine, 7-5-1 Kusunoki-cho, Chuo-ku, Kobe 650-0017, Japan; 2Division of Surgical , Department of Biomedical Informatics, Kobe University Graduate School of Medicine, 7-5-1 Kusunoki-cho, Chuo-ku, Kobe 650-0017, Japan; 3Biosignal Research Center, Institute for Molecular and Cellular Regulation, Gunma University, Maebashi 371-8512, Japan

SAP-1 (-associated protein tyrosine phos- phosphorylation of cellular targets of these enzymes phatase-1) is a transmembrane-type protein tyrosine (Cantley et al., 1991; Fantl et al., 1993). Protein tyrosine phosphatase that has been implicated as a negative phosphatases (PTPs) likely also affect , regulator of integrin-mediated signaling. The potential given that they counteract the action of protein tyrosine role of this enzyme in hepatocarcinogenesis has now been kinases (Tonks, 1993; Walton and Dixon, 1993; Hunter, investigated by examining its expression in 32 surgically 1995). Indeed, certain PTPs have been associated with excised human (HCC) speci- the development of some types of cancer. For example, mens. Both immunohistochemical and immunoblot ana- the expression of HPTPb is decreased in many lyses revealed that normal tissue, as well as tissue (De Vries et al., 1991), and the encoding affected by chronic hepatitis or cirrhosis, contained PTPg is frequently deleted in both renal and lung substantial amounts of SAP-1. The expression level of cancers (LaForgia et al., 1991; Tsukamoto et al., SAP-1 in 75% of well-differentiated HCCs was similar to 1992). or higher than that observed in the surrounding non- Human hepatocellular carcinoma (HCC) has been cancerous tissue. In contrast, the abundance of SAP-1 in thought to develop as a consequence of combinations of 85.7% of moderately differentiated HCCs and in all abnormalities in various such as those encoding poorly differentiated HCCs was greatly reduced com- Rb, p53, , and telomerase (Rogler and Chisari, 1992; pared with that in the adjacent tissue. Indeed, SAP-1 was Nishida et al., 1993; Tahara et al., 1995). As in other cell almost undetectable in 83.3% of poorly differentiated types, the tyrosine phosphorylation of cellular proteins HCCs. Furthermore, expression of recombinant SAP-1 in is indispensable for the growth and differentiation of two highly motile human HCC cell lines resulted in a hepatocytes (Palmer et al., 1999). In addition, the change in morphology and a marked reduction in both expression of several PTPs is upregulated during migratory activity and growth rate. In conclusion, these regeneration of the liver after partial hepatectomy in results indicate that SAP-1 expression is downregulated mice and rats (Higashitsuji et al., 1995; Kitamura et al., during the dedifferentiation of human HCC, and that this 1995; Saito et al., 1996). These observations thus downregulation may play a causal role in disease implicate tyrosine phosphorylation events and, hence, progression. PTPs in hepatocellular function under both physiologi- Oncogene (2003) 22, 4656–4663. doi:10.1038/sj.onc.1206588 cal and pathological conditions. Dysfunction of PTPs may thus contribute to the development of HCC, Keywords: tyrosine phosphatase; cell migration; de- although few studies to date have implicated specific differentiation; hepatoma PTPs in the pathogenesis of this disease (Ikuta et al., 1994; Kitamura et al., 1995). SAP-1 (stomach cancer-associated protein tyrosine phosphatase-1) was first identified as a transmembrane- Introduction type (receptor-like) PTP that is abundant in a subset of pancreatic and cell lines (Matozaki Many genetic alterations in both oncogenes and tumor et al., 1994). This enzyme belongs to the class 2 suppressor genes contribute to multistep carcinogenesis subfamily of receptor-like PTPs and is structurally in humans (Sugimura, 1992). The fact that several similar to HPTPb, DEP-1, and DPTP10D (Krueger oncogene products, including Bcr-Abl and v-Src, et al., 1990; Tian et al., 1991; Ostman et al., 1994). We possess high protein tyrosine kinase activity suggests have recently shown that SAP-1 mediates the depho- that carcinogenesis is promoted by the aberrant sphorylation of several focal adhesion-associated sub- strates and, consequently, inhibits cell proliferation by *Correspondence: T Noguchi; E-mail: [email protected] suppressing growth factor-elicited mitogenic signaling as Received 5 August 2002; revised 17 March 2003; accepted 17 March 2003 well as by inducing apoptotic cell death (Noguchi et al., Tyrosine phosphatase SAP-1 in human hepatocellular carcinoma H Nagano et al 4657 2001; Takada et al., 2002). The enzymatic activity of Table 1 Clinical characteristics of the study subjects SAP-1 is also increased in cells that exhibit density- WD-HCC MD-HCC PD-HCC dependent growth arrest (Noguchi et al., 2001). Characteristic (n ¼ 12) (n ¼ 14) (n ¼ 6) Furthermore, the abundance of SAP-1 mRNA increases during differentiation of breast cancer cells in response Sex (male/female) 12/0 12/2 4/2 Mean (range) age (years) 58.8 (52-67) 56.3 (34-72) 65.0 (63-66) to anticancer drugs (Keane et al., 1996). These observa- HBsAg/HCV Ab/alcohol 2/10/0 6/8/0 0/5/1 tions suggest that SAP-1 might act as a tumor Serum AFP (ng/ml) 4-303 21-19,840 5117-313,930 suppressor by generating a signal that results in growth Histological background arrest and differentiation of cancer cells. However, the Chronic hepatitis (n)2 0 0 potential roles of this PTP in carcinogenesis have Liver cirrhosis (n)10146 remained unclear. The differentiation status of each tumor was classified into three SAP-1 is abundant in the liver (Matozaki et al., 1994) categories according to the criteria of Edmondson and Steiner: well- and might therefore contribute to the initiation or differentiated (WD) HCC, corresponding to grades I to II; moderately progression of HCC. To investigate this hypothesis, differentiated (MD) HCC, corresponding to grades II or III; and we have now examined the expression of SAP-1 in poorly differentiated (PD) HCC, corresponding to grades III or IV. Disease etiology was classified according to serum positivity for human HCCs at various stages of differentiation. hepatitis B virus surface antigen (HBsAg) or for antibodies to hepatitis Our results show that the abundance of SAP-1 is C virus (HCVAb), or to alcohol intake. AFP, a-fetoprotein inversely related to the aggressiveness of HCC. Further- more, we show that forced expression of SAP-1 in highly motile HCC cells resulted in a morphological change that was accompanied by a marked reduction in (Figure 2a). Consistent with the immunohistochemical cell migration. data, the amount of SAP-1 in WD-HCC tissue was greater than that in MD-HCC or PD-HCC (Figure 2b). In contrast, the expression of SAP-1 in the surrounding Results cirrhotic tissue of these different tumors appeared similar (Figure 2b). The relation between the level of SAP-1 expression Expression of SAP-1 in human HCCs and the differentiation status of HCC was then analysed HCC specimens from 32 individuals were examined for statistically. Among the three stages of HCC differentia- SAP-1 expression by immunohistochemical analysis tion, the abundance of SAP-1 decreased in the rank with two different antibodies specific for this protein. order WD-HCC4MD-HCC4PD-HCC (Po0.01 for The clinical characteristics of the patients are provided WD-HCC versus MD-HCC and for MD-HCC versus in Table 1, and representative immunohistochemical PD-HCC) (Figure 3). Furthermore, the amount of SAP- data are shown in Figure 1. In normal hepatocytes, 1 in either MD-HCC or PD-HCC tissue was signifi- SAP-1 was detected predominantly at the cell membrane cantly smaller than that in the respective noncancerous and to a lesser extent in the cytoplasm (Figure 1a, e). tissue, with this reduction being more pronounced for Immunoreactivity was not apparent in sinusoidal PD-HCC (Figure 3). The abundance of SAP-1 in endothelial cells, whereas epithelial cells of the adjacent noncancerous tissue did not differ significantly yielded a moderate signal (Figure 1a, e). The endothe- among individuals with WD-HCC, with MD-HCC, or lium of the portal tract vessels was weakly positive for with PD-HCC. Given that human HCC progresses SAP-1, but mesenchymal cells were negative (Figure 1a, through dedifferentiation of the tumor, these results e). Nine of 12 (75%) well-differentiated (WD) HCCs indicate that the level of SAP-1 expression is inversely expressed SAP-1 at a level similar to or higher than that related to tumor aggressiveness. apparent in the surrounding noncancerous tissue, although the remaining three WD-HCCs also expressed Effects of ectopic expression of SAP-1 on HCC cell a substantial amount of this protein (Figure 1b, f and morphology and migration Table 2). In 12 of 14 (85.7%) specimens of moderately differentiated (MD) HCC, the abundance of SAP-1 in A reduced abundance of SAP-1 might contribute to the the cancerous tissue was substantially lower than that in acquired characteristics of advanced HCC such as the the adjacent noncancerous tissue (Figure 1c, g and abilities to metastasize and to invade tissue. Alterna- Table 2). All six poorly differentiated (PD) HCCs tively, it might be a consequence of disease progression. contained much smaller amounts of SAP-1 than did To distinguish between these possibilities, we transfected the surrounding noncancerous tissue (Figure 1d, h HLF and HLE cells, both of which are highly motile and Table 2); indeed, SAP-1 was almost undetectable and were derived from human PD-HCC (Doi et al., in five of six (83.3%) PD-HCCs (Table 2). Thus, the 1975), with an expression vector encoding SAP-1. level of expression of SAP-1 in human HCC appeared Endogenous SAP-1 was virtually undetectable in both positively related to the differentiation status of the of these cell lines (Figure 4). Several independent tumor. transfectants that stably expressed exogenous SAP-1 Immunoblot analysis revealed that the abundance of were generated from each cell line (Figure 4). Two SAP-1 in liver tissue affected by chronic hepatitis or clones established from HLF cells (F-SAP-A and cirrhosis was similar to that in normal liver tissue F-SAP-C) and one clone from HLE cells (E-SAP-d)

Oncogene Tyrosine phosphatase SAP-1 in human hepatocellular carcinoma H Nagano et al 4658

Oncogene Tyrosine phosphatase SAP-1 in human hepatocellular carcinoma H Nagano et al 4659 Table 2 Quantification of SAP-1 expression in HCC and surrounding noncancerous tissue Level of SAP-1 expression

Specimen 1 2 3 4

Cancerous tissue WD-HCC (n)0264 MD-HCC (n)1841 PD-HCC (n)5100 Surrounding tissue Chronic hepatitis (n)0 0 2 0 Liver cirrhosis (n)02208 Figure 3 Scoring of the intensity of SAP-1 staining according to The immunohistochemical signal intensity for SAP-1 expression was the differentiation status of 32 human HCC specimens. The level of classified as either 1 (faint), 2 (weak), 3 (moderate), or 4 (strong) SAP-1 expression in cancerous lesions (T) and the adjacent noncancerous tissue (N) was quantified by immunohistochemical analysis. Data are means7s.d. N.S., not significant; *Po0.01

Figure 4 Generation of HCC cell transfectants overexpressing SAP-1. Lysates were prepared either from nontransfected parental HLF (F-Parent) or HLE (E-Parent) cells, from their corresponding Figure 2 Immunoblot analysis of SAP-1 expression in human control lines transfected with the empty vector (F-Cont and E- HCCs at various stages of differentiation. (a) Expression of SAP-1 Cont, respectively), or from independent clones expressing in normal liver tissue or liver tissue affected by either chronic exogenous SAP-1 (F-SAP-A, F-SAP-C, E-SAP-a, E-SAP-b, hepatitis or cirrhosis. (b) Expression of SAP-1 in cancerous lesions E-SAP-g, and E-SAP-d). Samples were subjected to immunoblot (T) and the adjacent noncancerous tissue (N) of representative analysis with polyclonal antibodies to SAP-1 individuals with WD-, MD-, or PD-HCC. Duplicate samples were probed with polyclonal antibodies to actin to verify the presence of equal amounts of total protein loaded in each lane. Data are representative of three independent experiments performed with flattened, less polarized morphology and formed com- each specimen pact colonies in which cells were tightly connected with each other. Fluorescence analysis with rhodamine- were studied most extensively, although the other clones labeled phalloidin revealed that parental HLF cells showed similar phenotypes (data not shown). and cells transfected with the empty vector possess Under normal culture conditions, parental cells of substantial number of both actin stress fibers that cross each cell line, as well as cells transfected with the empty the cell body and filopodia-like microspikes at the cell vector, exhibited an elongated, spindle-like morphology, periphery (Figure 5c). In marked contrast, F-SAP-C appeared to be loosely associated, and freely migrated cells exhibited few actin stress fibers, with disorganized out of colonies (Figure 5a, b). In contrast, both F-SAP- actin filaments apparent in the subcortical region and C (Figure 5a) and E-SAP-d (Figure 5b) cells exhibited a cytoplasm (Figure 5c). Thus, forced expression of SAP-1 b————————————————————————————————————————————————— Figure 1 Immunohistochemical analysis of SAP-1 expression in human HCCs at various stages of differentiation. Surgically excised normal liver (a, e), WD-HCC (b, f), MD-HCC (c, g), and PD-HCC (d, h) tissues were fixed, embedded in paraffin, sectioned, and examined immunohistochemically with rabbit polyclonal antibodies to SAP-1 (a–d) or with a mouse monoclonal antibody (3G5) to SAP-1 (e–h). Normal hepatocytes (a, e) and the WD-HCC (b, f) exhibit a moderate level of SAP-1 expression. The abundance of SAP-1 in the MD-HCC is lower than that in the adjacent noncancerous tissue. The PD-HCC yielded only faint staining for SAP-1 (right halves of (d) and (h)). Neither normal rabbit immunoglobulin nor normal mouse immunoglobulin used as a negative control yielded detectable signals (data not shown). Data are representative of three independent experiments performed with each specimen. T, cancerous lesion; N, adjacent noncancerous tissue. Original magnification (a–h), Â 100. Higher magnification (insets in a and e), Â 400

Oncogene Tyrosine phosphatase SAP-1 in human hepatocellular carcinoma H Nagano et al 4660

Figure 5 Effect of overexpression of SAP-1 on the morphology of HCC cells. The indicated cell lines derived from HLF (a) or HLE (b) cells were photographed during exponential growth under normal culture conditions. Random fields were examined with the use of a light microscope equipped with phase-contrast optics. Original magnification, Â 100. (c) The indicated lines derived from HLF cells were seeded on glass coverslips and subsequently fixed and stained with rhodamine-labeled phalloidin. Cells were examined with a confocal microscope. Original magnification, Â 630

in HCC cells induced morphological changes and disruption of actin-based cytoskeleton. We previously observed similar changes induced by SAP-1 overexpres- sion in Chinese hamster ovary cells (Noguchi et al., 2001). We also examined the migratory capacity of the established HCC cell lines with the use of a Boyden chamber assay, in which cells move through a porous membrane coated with collagen type I. A substantial number of both parental HLF cells and cells transfected with the empty vector migrated in this assay (Figure 6). However, the number of cells that migrated through the membrane was markedly reduced for two independent cell lines (F-SAP-A and F-SAP-C) expressing exogenous SAP-1 (Figure 6).

Effect of ectopic expression of SAP-1 on HCC cell Figure 6 Effect of overexpression of SAP-1 on the migration of growth HCC cells. The indicated lines derived from HLF cells were seeded onto porous membranes that had been coated with collagen type I To further address the potential role of SAP-1 in the and placed in Boyden multiwell chambers. After incubation for pathogenesis of HCC, we monitored the growth rate of 48 h at 371C, the number of cells that had migrated through the membrane was determined as described in the Materials and the established cell lines. As shown in Figure 7, the methods section. Data are expressed as absorbance at 570 nm and growth rate of HLF cells expressing SAP-1 was are means7s.d. of triplicate determinations from three indepen- substantially lower than that of control cells. dent experiments. N.S., not significant; *Po0.01

Discussion expression of SAP-1; the protein was thus most abundant in WD-HCC, which is characterized by low We have shown that the dedifferentiation of human rates of growth and (Kim et al., 1995). HCC is accompanied by a decrease in the level of Furthermore, expression of recombinant SAP-1 in

Oncogene Tyrosine phosphatase SAP-1 in human hepatocellular carcinoma H Nagano et al 4661 structurally similar to SAP-1, during the development of human breast cancer (Ardini et al., 2000) is similar to that of SAP-1 during HCC progression. Tissue invasion and metastasis are characteristic features of advanced HCC and are critically regulated by cell motility (Genda et al., 1999). If the observed downregulation of SAP-1 expression was simply a consequence of dedifferentiation of HCC, forced ex- pression of the recombinant protein would not be expected to have a substantial effect on the behavior of cells derived from PD-HCC. Indeed, we have shown that expression of SAP-1 in highly motile HLF and HLE cells resulted in a change in morphology as well as a marked decrease in migratory activity. These results suggest that SAP-1 normally inhibits the motility of HCC cells and that this PTP plays a role in keeping the tumor cells well differentiated. Together with the observation that the abundance of SAP-1 decreases at the later stages of HCC, our results support the notion that downregulation of SAP-1 contributes, at least in part, to the enhanced abilities of advanced HCC to Figure 7 Effect of overexpression of SAP-1 on the growth of HCC metastasize and invade tissue. cells. The indicated lines derived from HLF cells were seeded in 96- With regard to the mechanism by which SAP-1 well plate and cultured in the presence of 10% fetal bovine serum. inhibits the migration of HCC cells, we have previously At the indicated times thereafter, the rate of cell growth was monitored as described in the Materials and methods section. Data shown that this PTP negatively regulates integrin- are expressed as absorbance at 450 nm and are means7s.d. of mediated signaling by catalyzing the dephosphorylation triplicate determinations from three independent experiments. of p130Cas and focal adhesion kinase (Noguchi et al., *Po0.05 2001), both of which play important roles in cell migration (Hanks and Polte, 1997; Schlaepfer et al., 1999). Furthermore, increased expression and tyrosine highly motile HCC cells resulted in a change in phosphorylation of these proteins promote the migra- morphology that was accompanied by a marked tory and invasive activities of various cancer cells, reduction in migratory activity and growth rate. These including HCC cells (Klemke et al., 1998; Tamura et al., results, together with our previous observations that 1999; Imamura et al., 2000). SAP-1-mediated depho- SAP-1 suppresses mitogenic and survival signaling sphorylation of these focal adhesion-associated proteins, (Noguchi et al., 2001; Takada et al., 2002), suggest that therefore, likely results in inhibition of the migration of this PTP negatively regulates the progression of human HCC cells. Integrin-mediated signaling has been shown HCC. to induce disassembly of intercellular adherens junctions Most PD-HCCs contained almost undetectable mediated by E-cadherin, a key step of cancer invasion amounts of SAP-1, indicating that the expression of and metastasis (Genda et al., 2000; Ojakian et al., 2001). this protein is downregulated at the later stages, but not It is thus possible that expression of SAP-1 prevents at the early stage, of human HCC. This observation HCC metastasis not only by inhibiting cell motility but appears inconsistent with our previous data showing also by stabilizing cell–cell adhesion in the primary that SAP-1 is abundant in colorectal cancer tissue but cancer nest. not in the corresponding normal tissue (Matozaki et al., In conclusion, our results demonstrate that the level 1994). However, we also showed previously that the of SAP-1 expression correlates with the differentiation abundance of SAP-1 decreases with the progression of status of human HCC. Furthermore, overexpression of colorectal tumors (Seo et al., 1997), as was the case with this PTP both affected the morphology of HCC cells as HCC in the present study. Furthermore, for a few well as markedly reduced their migratory activity and individuals with WD-HCC, the level of expression of growth rate. Together, our results suggest that the SAP-1 was greater in the cancerous tissue than in the downregulation of SAP-1 may play a causal role in the adjacent noncancerous tissue. As far as we are aware, progression of this fatal disease. the expression of PTPs in human HCC has not previously been examined in relation to differentiation status, although extracellular signal-regulated kinases (Ito et al., 1998) and the proto-oncogene c-MYC Materials and methods (Kawate et al., 1999) have been shown to be upregulated in PD-HCC. Our results thus identify SAP-1 as a PTP Patients and tissue samples whose expression is induced at the early stage of human HCC specimens including adjacent noncancerous tissue were HCC and suppressed at later stages. Interestingly, the obtained with written informed consent from 32 individuals pattern of expression of RPTPa, a receptor-like PTP admitted to Kobe University Hospital; 27 of the specimens

Oncogene Tyrosine phosphatase SAP-1 in human hepatocellular carcinoma H Nagano et al 4662 were obtained by surgical excision and five by autopsy. The either of pSRa containing human SAP-1 cDNA (Matozaki clinical features of the patients are shown in Table 1. A total of et al., 1994) or of the empty vector with the use of an Effectene 30 of the study subjects exhibited underlying cirrhosis and the Transfection Kit (QIAGEN, Hilden, Germany). The cells were remaining two had chronic hepatitis. The etiologic factors cultured for 48 h in normal growth medium and then exposed included hepatitis B virus (eight patients), hepatitis C virus (23 to G418 (600 mg/ml) (GIBCO BRL). The resulting G418- patients), and alcohol (one patient). The differentiation status resistant colonies were isolated 3 weeks after transfection, and of each tumor was classified into three categories as described stable transfectants were identified by immunoblot analysis by Edmondson and Steiner (1954). Two normal liver tissue with polyclonal antibodies to SAP-1 as described above. specimens were obtained by surgical excision from individuals with liver metastasis of colon cancer. All specimens were split Assay of cell morphology into two portions, one of which was immediately frozen in liquid nitrogen and stored at À801C until protein extraction, Cells were seeded at a density of 1 Â 105 cells/ml on 60-mm and the other was fixed in 10% formalin and embedded in a dishes and cultured in a humidified incubator containing 5% paraffin block. The university ethics committee approved this CO2. Exponentially growing cells were examined under a light study. microscope equipped with phase-contrast optics (TM-PSF, Nikon) and random fields were photographed. For staining of Immunohistochemistry filamentous actin, cells seeded on glass coverslips were washed with PBS, fixed with 3.7% formaldehyde in PBS for 20 min, Immunohistochemical staining was performed with the use of permeabilized with 0.5% Triton X-100 in PBS for 5 min, and a large volume DAKO LSAB Kit (DAKO, Copenhagen, incubated for 2 h in TBS-T (20 mm Tris-HCl (pH 7.6), 150 mm Denmark). In brief, fixed sections were depleted of paraffin by NaCl, 0.05% Tween-20) containing 5% nonfat dry milk, 10% three treatments (each for 5 min) with xylene, dehydrated in fetal bovine serum, and 1% bovine serum albumin. The cells ethanol, and incubated with 3% hydrogen peroxide for 15 min were then incubated for 1 h at room temperature with in order to quench endogenous peroxidase activity. After rhodamine-labeled phalloidin (0.1 mg/ml) (Sigma). After three digestion with 0.1% trypsin for 5 min at room temperature, the washes with PBS, the cells were examined with a laser-scanning sections were incubated first with a blocking solution for confocal microscope (Bio-Rad model MRC-1024), and built- 30 min and then overnight at 41C with either rabbit polyclonal up images were constructed. antibodies (1 : 50 dilution) (Seo et al., 1997) or mouse monoclonal antibody 3G5 (1 : 200 dilution) (Noguchi et al., 2001) to SAP-1 diluted in phosphate-buffered saline (PBS). Assay of cell migration The sections were washed with 50 mm Tris-HCl (pH 7.6) and Cell migration was assessed with the use of a QCM-Collagen I then incubated for 1 h at room temperature first with Quantitative Cell Migration Assay Kit (Chemicon, Temecula, biotinylated secondary antibodies to rabbit or mouse immu- CA, USA). In brief, cells deprived of serum for 24 h were noglobulin G (DAKO) and then with horseradish peroxidase- detached from culture dishes by treatment with Hanks’ conjugated streptavidin (Amersham Pharmacia Biotech, Little balanced salt solution containing 0.53 mm EDTA and 0.05% Chalfornt, UK). The sections were then counterstained with trypsin. The cells (2.5 Â 105 in 0.5 ml of serum-free medium hematoxylin, mounted on cover slides, and examined with a containing 5% bovine serum albumin) were added to each of light microscope (OPTIPHOT, Nikon, Tokyo). The signal the collagen type I-coated membranes (pore size, 8 mm) that intensity was classified into one of the four levels by two had been placed at the bottom of the upper wells. After independent pathologists who were not aware of the specimen incubation for 48 h at 371C in a humidified incubator identity. containing 5% CO2, cells that had migrated through the membrane were exposed to a staining solution for 30 min at Immunoblot analysis room temperature. The stain was then eluted from the cells with 300 ml of an extraction buffer, the absorbance of which at Liver tissue or cultured cells were homogenized on ice in RIPA 570 nm was then measured with a microtiter plate reader. buffer (20 mm Tris-HCl (pH 7.4), 150 mm NaCl, 2 mm EDTA, 1% sodium deoxycholate, 0.1% SDS, 1% Nonidet P-40) containing 5 mm NaF, 1 mm phenylmethylsulfonyl fluoride, Assay of cell growth aprotinin (10 mg/ml), and leupeptin (1 mg/ml). The homogenate Cells were seeded in 96-well culture plates at a density of 1 was clarified by centrifugation at 10 000 g for 15 min at 4 C, 1 Â 103 per well and cultured in the presence of 10% fetal after which the protein concentration of the supernatant was bovine serum at 371C in a humidified incubator containing 5% normalized. Samples containing either 400 mg of tissue protein CO2. The culture medium was changed every 3 days, and or 100 mg of cellular protein were suspended in Laemmli the rate of cell growth was determined every 24 h with the use sample buffer, boiled, and then fractionated by SDS– of a Cell Proliferation Assay Kit (Chemicon, Temecula, CA, polyacrylamide gel electrophoresis on a 7.5% gel. Immunoblot USA). analysis with polyclonal antibodies to SAP-1 or to actin (Sigma) was performed with the use of the ECL detection system (Amersham Pharmacia Biotech) as described pre- Statistical analysis viously (Kameda et al., 1990; Morita et al., 1997). Data are expressed as means7s.d. Continuous variables were evaluated by the Mann–Whitney U test, and discrete variables Cells and transfection were assessed by the Kruskal–Wallis test. A P-value of o0.05 HLF and HLE cells were obtained from the Japanese Cancer was considered statistically significant. Research Resources Bank (Tokyo, Japan) and were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% Acknowledgements fetal bovine serum, penicillin (100 U/ml), and streptomycin This study was supported by a grant-in-aid for scientific (100 mg/ml) (GIBCO BRL, Grand Island, NY, USA). Cells research (B)(2) 10470049 from the Ministry of Education, (70% confluent in 60-mm dishes) were transfected with 1 mg Science, Sports, and Culture of Japan.

Oncogene Tyrosine phosphatase SAP-1 in human hepatocellular carcinoma H Nagano et al 4663 References

Ardini E, Agresti R, Tagliabue E, Greco M, Aiello P, Yang Huebner K. (1991). Proc. Natl. Acad. Sci. USA, 88, LT, Menard S and Sap J. (2000). Oncogene, 19, 4979–4987. 5036–5040. Cantley LC, Auger KR, Carpenter C, Duckworth B, Graziani Matozaki T, Suzuki T, Uchida T, Inazawa J, Ariyama T, A, Kapeller R and Soltoff S. (1991). Cell, 64, 281–302. Matsuda K, Horita K, Noguchi H, Mizuno H, Sakamoto C De Vries L, Li RY, Ragab A, Ragab-Thomas JM and Chap H. and Kasuga M. (1994). J. Biol. Chem., 269, 2075–2081. (1991). FEBS Lett., 282, 285–288. Morita Y, Hayashi Y, Wang Y, Kanamaru T, Suzuki S , Ohta Doi I, Namba M and Sato J. (1975). Gann, 66, 385–392. K, Yamamoto M, Saitoh Y, Itoh H and Doe WF. (1997). Edmondson HA and Steiner PE. (1954). Cancer, 7, 462–503. Hepatology, 25, 856–861. Fantl WJ, Johnson DE and Williams LT. (1993). Annu. Rev. Nishida N, Fukuda Y, Kokuryu H, Toguchida J, Yandell Biochem., 62, 453–481. DW, Ikenaga M, Imura H and Ishizaki K. (1993). Cancer Genda T, Sakamoto M, Ichida T, Asakura H and Hirohashi S. Res., 53, 368–372. (2000). Lab. Invest., 80, 387–394. Noguchi T, Tsuda M, Takeda H, Takada T, Inagaki K, Genda T, Sakamoto M, Ichida T, Asakura H, Kojiro M, Yamao T, Fukunaga K, Matozaki T and Kasuga M. (2001). Narumiya S and Hirohashi S. (1999). Hepatology, 30, 1027– J. Biol. Chem., 276, 15216–15224. 1036. Ojakian GK, Ratcliffe DR and Schwimmer R. (2001). J. Cell Hanks SK and Polte TR. (1997). BioEssays, 19, 137–145. Sci., 114, 941–952. Higashitsuji H, Arii S, Furutani M, Imamura M, Kaneko Y, Ostman A, Yang Q and Tonks NK. (1994). Proc. Natl. Acad. Takenawa J, Nakayama H and Fujita J. (1995). Oncogene, Sci. USA, 91, 9680–9684. 10, 407–414. Palmer HJ, Tuzon CT and Paulson KE. (1999). J. Biol. Chem., Hunter T. (1995). Cell, 80, 225–236. 274, 11424–11430. Ikuta S, Itoh F, Hinoda Y, Toyota M, Makiguchi Y and Imai Rogler CE and Chisari FV. (1992). Semin. Liver Dis., 12, K. (1994). J. Gastroenterol., 29, 727–732. 265–278. Imamura F, Mukai M, Ayaki M and Akedo H. (2000). Jpn. J. Saito S, Hashimoto N, Tokita K, Toyoda M, Suzuki Y, Cancer Res., 91, 811–816. Goldstein BJ, Suzuki H, Makino H and Saito Y. (1996). Ito Y, Sasaki Y, Horimoto M, Wada S, Tanaka Y, Kasahara Biochem. Biophys. Res. Commun., 227, 406–412. A, Ueki T, Hirano T, Yamamoto H, Fujimoto J, Okamoto Schlaepfer DD, Hauck CR and Sieg DJ. (1999). Prog. Biophys. E, Hayashi N and Hori M. (1998). Hepatology, 27, 951–958. Mol. Biol., 71, 435–478. Kameda T, Yasui W, Yoshida K, Tsujino T, Nakayama H, Ito Seo Y, Matozaki T, Tsuda M, Hayashi Y, Itoh H and Kasuga M, Ito H and Tahara E. (1990). Cancer Res., 50, 8002–8009. M. (1997). Biochem. Biophys. Res. Commun., 231, 705–711. Kawate S, Fukusato T, Ohwada S, Watanuki A and Morishita Sugimura T. (1992). Science, 258, 603–607. Y. (1999). , 57, 157–163. Tahara H, Nakanishi T, Kitamoto M, Nakashio R, Shay JW, Keane MM, Lowrey GA, Ettenberg SA, Dayton MA and Tahara E, Kajiyama G and Ide T. (1995). Cancer Res., 55, Lipkowitz S. (1996). Cancer Res., 56, 4236–4243. 2734–2736. Kim SR, Kang KB, Soh CG, Kim JH, Hayashi Y, Hanioka K Takada T, Noguchi T, Inagaki K, Hosooka T, Fukunaga K, and Itoh H. (1995). J. Gastroenterol. Hepatol., 10, 498–508. Yamao T, Ogawa W, Matozaki T and Kasuga M. (2002). Kitamura T, Nakamura K, Mizuno Y and Kikuchi K. (1995). J. Biol. Chem., 277, 34359–34366. Jpn. J. Cancer Res., 86, 811–818. Tamura M, Gu J, Takino T and Yamada KM. (1999). Cancer Klemke RL, Leng J, Molander R, Brooks PC, Vuori K and Res., 59, 442–449. Cheresh DA. (1998). J. Cell Biol., 140, 961–972. Tian SS, Tsoulfas P and Zinn K. (1991). Cell, 67, 675–685. Krueger NX, Streuli M and Saito H. (1990). EMBO J., 9, Tonks NK. (1993). Semin. Cell Biol., 4, 373–377. 3241–3252. Tsukamoto T, Takahashi T, Ueda R, Hibi K and Saito H. LaForgia S, Morse B, Levy J, Barnea G, Cannizzaro LA, Li F, (1992). Cancer Res., 52, 3506–3509. Nowell PC, Boghosian-Sell L, Glick J, Weston A, Harris Walton KM and Dixon JE. (1993). Annu. Rev. Biochem., 62, CC, Drabkin H, Patterson D, Croce CM, Schlessinger J and 101–120.

Oncogene