Association of gankyrin protein expression with early clinical stages and insulin-like growth factor-binding protein 5 expression in human hepatocellular carcinoma

Authors

  • Atsushi Umemura,

    1. Department of Clinical Molecular Biology, Graduate School of Medicine, Kyoto University, Kyoto, Japan
    2. Molecular Gastroenterology and Hepatology, Graduate School of Medical Science, Kyoto Prefectural University of Medicine, Kyoto, Japan
    Search for more papers by this author
  • Yoshito Itoh,

    1. Molecular Gastroenterology and Hepatology, Graduate School of Medical Science, Kyoto Prefectural University of Medicine, Kyoto, Japan
    Search for more papers by this author
  • Katsuhiko Itoh,

    1. Department of Clinical Molecular Biology, Graduate School of Medicine, Kyoto University, Kyoto, Japan
    Search for more papers by this author
  • Kanji Yamaguchi,

    1. Molecular Gastroenterology and Hepatology, Graduate School of Medical Science, Kyoto Prefectural University of Medicine, Kyoto, Japan
    Search for more papers by this author
  • Tomoki Nakajima,

    1. Molecular Gastroenterology and Hepatology, Graduate School of Medical Science, Kyoto Prefectural University of Medicine, Kyoto, Japan
    Search for more papers by this author
  • Hiroaki Higashitsuji,

    1. Department of Clinical Molecular Biology, Graduate School of Medicine, Kyoto University, Kyoto, Japan
    Search for more papers by this author
  • Hitoshi Onoue,

    1. Department of Nutritional Science, Faculty of Health and Welfare, Seinan Jo Gakuin University, Kitakyushu, Japan
    Search for more papers by this author
  • Manabu Fukumoto,

    1. Department of Pathology, Institute of Development, Aging, and Cancer, Tohoku University, Sendai, Japan
    Search for more papers by this author
  • Takeshi Okanoue,

    1. Molecular Gastroenterology and Hepatology, Graduate School of Medical Science, Kyoto Prefectural University of Medicine, Kyoto, Japan
    Search for more papers by this author
  • Jun Fujita

    Corresponding author
    1. Department of Clinical Molecular Biology, Graduate School of Medicine, Kyoto University, Kyoto, Japan
    • Department of Clinical Molecular Biology, Graduate School of Medicine, Kyoto University, 54 Shogoin Kawaharacho, Sakyo-ku, Kyoto 606-8507, Japan
    Search for more papers by this author
    • fax: (81) 75-751-4977.


  • Potential conflict of interest: Nothing to report.

Abstract

Gankyrin (also known as PSMD10) is a liver oncoprotein that interacts with multiple proteins including MDM2 and accelerates degradation of the tumor suppressors p53 and Rb. We produced a monoclonal anti-gankyrin antibody and immunohistochemically assessed the clinicopathological significance of gankyrin overexpression in 43 specimens of human hepatocellular carcinoma (HCC). Specific cytoplasmic staining for gankyrin was observed in 62.8% (27/43) of HCCs, which was significantly associated with low TNM stage (P = 0.004), no capsular invasion (P = 0.018), no portal venous invasion (P = 0.008), and no intrahepatic metastasis (P = 0.012). The cumulative survival rate of patients with gankyrin-positive HCC was significantly higher than that with gankyrin-negative HCC (P = 0.037). p53 and MDM2 were positively stained by antibodies in 30.2% and 23.3%, respectively, of HCCs, but neither was inversely associated with gankyrin expression. In the Huh-7 human HCC cell line, overexpression of gankyrin up-regulated expression of insulin-like growth factor binding protein 5 (IGFBP-5), whereas suppression of gankyrin expression by siRNA down-regulated it. Supression of IGFBP-5 expression inhibited proliferation of Huh-7 cells as well as U-2 OS osteosarcoma cells. In HCC specimens, positive staining for IGFBP-5 was observed by immunohistochemistry in 41.9% (18/43), and the level of expression was significantly correlated with that of gankyrin (rho = 0.629, P < 0.001). Conclusion: These results suggest that gankyrin plays an oncogenic role(s) mainly at the early stages of human hepatocarcinogenesis, and that IGFBP-5 inducible by gankyrin overexpression may be involved in it. (HEPATOLOGY 2008.)

Liver cancer is the sixth most common cancer worldwide (626,000 or 5.7% of new cancer cases) and the third most common cause of death from cancer (598,000) in 2002.1 Eighty-two percent of cases are in developing countries, and the areas of high incidence are sub-Saharan Africa, eastern and southeastern Asia, and Melanesia. Histologically, more than 90% of the primary liver cancers are hepatocellular carcinomas (HCCs). Although there are several modalities of treatment for HCC, most patients present with unresectable tumors, and nonsurgical treatments are minimally effective at best.2, 3 Even for those patients who undergo surgical resection, the recurrence rate is very high and the prognosis is poor.2, 4–6 It is therefore important to clarify the mechanisms of human hepatocarcinogenesis and identify molecular targets to develop novel diagnostic, therapeutic, and preventive strategies.

By constructing subtracted complementary DNA (cDNA) libraries, we have previously identified 19 genes overexpressed in HCCs including 2 novel genes.7, 8 One of them was named gankyrin (gann-ankyrin repeat protein; “gann” in Japanese means cancer).9 Gankyrin (also called PSMD10) consists of 7 ankyrin repeats, and its messenger RNA (mRNA) was overexpressed in 34 of 34 HCCs analyzed.9, 10 Independently, gankyrin was isolated as the p28 component or the interactor of the S6b subunit of the 19S regulator of the 26S proteasome.11, 12 The ankyrin repeat is the functional domain involved in protein–protein interactions, and gankyrin has been shown to interact with multiple proteins in addition to S6b. Gankyrin binds to retinoblastoma protein (Rb) and cyclin-dependent kinase (Cdk4), and accelerates phosphorylation and degradation of Rb, which results in release of the E2F transcription factor to activate DNA synthesis genes.9, 13 Gankyrin seems to play a role in cell cycle progression in noncancerous cells as well. Overexpression of gankyrin shortens population doubling time of NIH/3T3 mouse fibroblasts,9 and its up-regulation correlates with cell cycle progression in normal rat primary hepatocytes, oval cells, and human hepatocytes.14, 15

Overexpression of gankyrin confers tumorigenicity to NIH/3T3 cells and inhibits apoptosis in cultured human tumor cells exposed to chemotherapeutic agents.10 The anti-apoptotic activity is attributable, at least partly, to increased degradation of p53, resulting in the reduced transcription of the p53-dependent proapoptotic genes.16 Gankyrin binds to the E3 ubiquitin ligase MDM2 in vitro and in vivo, which increases p53–MDM2 association, thereby facilitating the ubiquitination and subsequent proteasomal degradation of p53 by MDM2. Gankyrin also controls MDM2 auto-ubiquitination and degradation, especially in the absence of p53.16

We produced a mouse monoclonal antibody against human gankyrin and assessed the expression of gankyrin protein in surgically resected HCC specimens by immunohistochemistry. Correlation of gankyrin positivity with clinicopathological findings and expression of p53 and MDM2 in HCC was analyzed. Furthermore, we demonstrated that expression of insulin-like growth factor-binding protein 5 (IGFBP-5) is inducible by overexpression of gankyrin in HCCs.

Abbreviations

3A6C2, mouse monoclonal anti-gankyrin antibody; cDNA, complementary DNA; HCC, hepatocellular carcinoma; IGF, insulinlike growth factor; IGFBP-5, insulin-like growth factor-binding protein 5; MDM2, mouse double minute 2; mRNA, messenger RNA; RT-PCR, reverse transcription polymerase chain reaction; siRNA, short interfering RNA; TNM, tumor-node-metastasis.

Patients and Methods

Patients and Specimens.

HCC tissues and their corresponding noncancerous liver tissues were obtained from 43 and 32 patients, respectively, who had undergone curative hepatectomy at the University Hospital of Kyoto Prefectural University of Medicine between 1992 and 2000. The specimens used were routinely processed, formalin-fixed, and paraffin-embedded. After hematoxylin-eosin staining, all samples were diagnosed as HCC and the tumor-node-metastasis (TNM) classification was made according to the fourth edition of the general rules for the clinical and pathological study of primary liver cancer proposed by the Liver Cancer Study Group of Japan.17 The demographic profiles of the patients are summarized in Table 1. For western blot analysis, HCCs and noncancerous liver tissues were obtained from 3 patients undergoing liver transplantation at the University Hospital of Kyoto Prefectural University of Medicine between 2004 and 2006. No donor organs were obtained from executed prisoners or other institutionalized persons. The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the institutional review boards. Written informed consents were obtained from all patients for subsequent use of their resected tissues.

Table 1. Patient and Tumor Characteristics
CharacteristicNumber (Percentage)
  1. Abbreviations: HCV(+), anti-hepatitis C virus antibody positive; HBV(+), hepatitis B surface antigen positive; (−), negative; AFP, serum alpha-fetoprotein.

Number of patients43
Sex distribution 
 Male27 (62.8%)
 Female16 (37.2%)
Age (years)25–78, median 65
Virus marker 
 HBV(+)/HCV(−)6 (14.0%)
 HBV(−)/HCV(+)28 (65.0%)
 HBV(+)/HCV(+)3 (7.0%)
 HBV(−)/HCV(−)6 (14.0%)
AFP(ng/mL)3.5–39999, median 90
Tumor size (cm)1.6–17.0, median 4.0
Liver cirrhosis 
 Yes29 (67.5%)
 No14 (32.5%)
  Chronic hepatitis13 (30.2%)
  Normal1 (2.3%)
TNM stage 
 I4 (9.3%)
 II22 (51.1%)
 III8 (18.6%)
 IV9 (21.0%)
Histological differentiation 
 Well12 (27.9%)
 Moderate25 (58.1%)
 Poor6 (14.0%)
Capsular formation 
 Yes36 (83.7%)
 No7 (16.3%)
Capsular invasion 
 Yes14 (32.6%)
 No29 (67.4%)
Portal venous invasion 
 Yes9 (20.9%)
 No34 (79.1%)
Intrahepatic metastasis 
 Yes16 (37.2%)
 No27 (62.8%)

Cell Culture and Transfection.

Huh-7 human HCC cells, U-2 OS human osteosarcoma cells, 293T human embryonic kidney cells, mouse lymph node cells, and P3X63Ag8U.1 mouse myeloma cells were cultured in Dulbecco's modified Eagle's medium (Gibco BRL Life Technologies, NY) supplemented with 10% fetal bovine serum as described.16 To assess viable cell numbers, we used the Dojindo Cell Counting Kit-8 (CCK8 kit, Dojindo Laboratories, Kumamoto, Japan) according to the manufacturer's instructions.

The 293T, Huh-7, and U-2 OS cells were transfected with plasmid DNA by using the calcium phosphate method or FuGENE 6 Transfection Reagent (Roche Diagnostics, Mannheim, Germany) as described.16 Short interfering RNA (siRNA) were transfected at a final concentration of 25 nM by using siPORT NeoFX Transfection Agent (Ambion, Austin, TX) following the manufacturer's instructions. Twenty-four hours after transfection, the medium was replaced with fresh medium containing fetal bovine serum, and the culture was continued for another 24 or 48 hours. Then, the cells were harvested for analysis. All transfection assays were repeated at least 3 times.

Plasmids and siRNA.

Human wild-type gankyrin cDNAs, full coding sequence and deletion mutants, were cloned into the mammalian expression vector pMKIT-NEO and expressed as hemagglutinin (HA)-tagged proteins (Fig. 1A). Full-length gankyrin was expressed without a tag as well. To obtain recombinant human gankyrin protein, the full-length cDNA was cloned into an expression vector derived from pET28 (Novagen, EMD Biosciences Inc., San Diego, CA) and expressed as hexahistidine-tagged protein.

Figure 1.

Recognition of gankyrin protein by the monoclonal antibody. (A) Structures of wild-type gankyrin (Full) and its deletion mutants. Numbers on top, N- and C-terminal amino-acid residues. ANK, ankyrin repeat. Black bars, HA tags. (B) Specificity of the antibody. 293T cells were transfected with plasmids expressing the indicated proteins. Cell lysates were analyzed by western blotting (WB) using the anti-gankyrin monoclonal antibody (3A6C2), anti-HA antibody, and anti–β-actin antibody. *Mobility of the endogenous gankyrin. Representative results of 3 repeated experiments are shown. (C) Detection of gankyrin protein in tissues. Lysates were made from hepatocellular carcinoma (HCC, n = 3) and cirrhotic liver tissues (n = 2), and analyzed by WB using antibodies for indicated proteins. HA, hemagglutinin.

To down-regulate gene expression, Silencer Predesigned siRNAs for gankyrin (Ambion) and Stealth Select iRNA: for IGFBP-5 (Invitrogen, Tokyo, Japan), were used together with respective control RNAs.

Antibodies.

To obtain monoclonal antibodies against human gankyrin, recombinant (His)6-gankyrin protein was used as an immunogen. It was dissolved in phosphate-buffered saline (1 mg/mL) and emulsified with an equal volume of Freund's complete adjuvant (Difco, Becton-Dickinson, Franklin Lakes, NJ). Two female BALB/c mice were injected with the emulsion (50 μL/mouse) in the footpad. Two weeks after immunization, the inguinal lymph node cells (4 × 107 cells) were fused with P3X63Ag8U.1 myeloma cells (1 × 107) using polyethylene glycol 1500 (Roche Diagnostics). Fused cells were cultured in 96-well plates at 2 × 105 cell/well. The supernatants were assayed for the anti-gankyrin antibody titer by an enzyme-linked immunosorbent assay using recombinant His-tagged, glutathione-S-transferase (GST)-tagged, and nontagged gankyrin proteins. Selected relevant hybridomas were cloned by the limiting dilution method, and the isotypes of secreted monoclonal antibodies were determined by the IsoStrip kit (Roche Diagnostics) following the manufacturer's instructions. Finally, an IgG2b kappa monoclonal antibody that showed the highest affinity for gankyrin was obtained and named 3A6C2.

For western blot analysis, mouse monoclonal anti-gankyrin antibody (3A6C2), goat polyclonal anti–IGFBP-5 antibody (R&D Systems Inc., Minneapolis, MN), mouse monoclonal anti-HA antibody (12CA5, Roche Diagnostics), and mouse monoclonal anti–β-actin antibody (Chemicon International, Temecula, CA) were used. Horseradish peroxidase–conjugated secondary antibodies against mouse or goat immunoglobulins were obtained from DAKO (Kyoto, Japan).

For immunohistochemistry, mouse monoclonal anti-gankyrin (3A6C2), anti-MDM2 (Ab-4, Oncogene research products, Boston, MA), and anti-p53 (DO-7, DAKO) antibodies, rabbit polyclonal anti-IGFBP-5 antibody (GroPep, Thebarton, Australia), and horseradish peroxidase–conjugated secondary antibodies against mouse or rabbit immunoglobulins (DAKO) were used.

Analysis of Gene Expression.

Extraction of RNA, preparation of cell and tissue lysates, and western blot analysis were performed as described.9 Real-time reverse transcription polymerase chain reaction (RT-PCR) analysis was performed using ABI PRISM 7900 (Applied Biosystems, Foster City, CA) and a 1-step QuantiTect RT-PCR Kit (Qiagen, Cowley, UK) according to the manufacturer's instructions. PCR conditions were 50°C for 30 minutes and 95°C for 15 minutes, followed by 45 cycles of 95°C for 15 seconds, 55°C for 30 seconds, and 72°C for 45 seconds. Specific PCR amplification products were detected by SYBR Green. Transcripts of β-actin were quantified as control. Primer sequences used were as follows: IGFBP-5, AAGAAGCTGACCCAGTCCAA and GAATCCTTTGCGGTCACAAT; gankyrin, GCAACTTGGAGTGCCAGTGAA and TCACTTGAGCACCTTTTCCCA; β-actin, CTACGTCGCCCTGGACTTCGAGC and GATGGAGCCGC- CGATCCACACGG.

The immunohistochemical staining was performed on 4-μm-thick paraffin sections of tissues fixed in buffered formalin. The sections were pretreated with 10 mM citrate buffer (pH 6.1) in a microwave oven for 5 minutes. Endogenous peroxidase activity was blocked with 0.3 % H2O2 for 10 minutes. The sections were incubated with 10% fetal bovine serum for 30 minutes to reduce nonspecific binding, followed by incubation with the primary antibody at 4°C overnight. They were subsequently incubated with horseradish peroxidase–conjugated anti-mouse or rabbit immunoglobulin antibody for 30 minutes. The enzymatic reaction was developed in a freshly prepared solution of 3,3′-diaminobenzidine tetrahydrochloride using DAKO Liquid DAB Substrate-Chromogen Solution for 10 minutes at room temperature. The sections were then counterstained with hematoxylin. The staining pattern, the distribution of the immunostaining in each tissue, and the intensity of the staining were studied in detail. Negative controls were conducted by substituting normal sera of each animal for the primary antibodies. When immunoreactivities were heterogenously observed, cases with moderate or strong staining of nucleus or cytoplasm in more than 5% of the cells were considered positive. To analyze the correlation of the expression levels of gankyrin and IGFBP-5, the staining intensity was expressed as 0 (negative), 1+ (weakly positive), 2+ (moderately positive), or 3+ (strongly positive). In each case the immunoreactivity was determined in 5 random high-powered fields and the count was done independently by 2 observers.

Statistical Analysis.

Categorical variables were compared using Fisher's exact test. Paired comparison of continuous data was performed using the Wilcoxon signed ranks test. To assess whether the 2 variables covary, Spearman's rank correlation coefficient was determined. Cumulative survival curves were calculated by the Kaplan-Meier method and analyzed by the log-rank test. All statistical analyses were performed using the JMP statistical software package (SAS Institute Inc., Cary, NC). A P value less than 0.05 was considered statistically significant.

Results

Clinicopathological Profiles.

Forty-three patients with HCC were recruited in this study, including 27 men and 16 women, with ages ranging from 25 to 78 (median 65) years old. Clinicopathological profiles of the patients and their HCCs are shown in Table 1. Antibody to hepatitis C virus was found in sera of 72% of the patients, and hepatitis B virus surface antigen was positive in 21%.

According to the TNM staging, 60% were stage I to II and 40% were stage III to IV. In noncancerous portions of the resected livers, cirrhosis and chronic hepatitis18 were found in 68% and 30%, respectively, of the specimens, whereas only 1 (2%) was of normal histology. Fibrocapsular formation surrounding HCC was observed in 84% and capsular invasion by HCC cells in 33%. Portal vein involvement and satellite nodules suggesting intrahepatic metastasis were found in 21% and 37%, respectively.

Detection of Gankyrin with the Monoclonal Anti-gankyrin Antibody.

To determine the specificity of the monoclonal anti-gankyrin antibody 3A6C2, we expressed wild-type full-length or truncated gankyrin (Fig. 1A) in 293T cells. The antibody detected all mutants of gankyrin, suggesting that the epitope exists within the third and fifth ankyrin-repeat region (Fig. 1B). The antibody recognized the endogenous gankyrin as well, and no major cross-reacting band was observed.

Because gankyrin mRNA is known to be overexpressed in most HCCs,9 we analyzed the levels of gankyrin protein in HCCs and surrounding noncancerous liver tissues using the 3A6C2 antibody. The protein level of gankyrin was higher in HCC tissues than in noncancerous tissues (Fig. 1C). The mobilities of the gankyrin band were not different among samples.

Immunohistochemical Analysis of Gankyrin Expression.

We next examined the expression of gankyrin protein in HCC and noncancerous liver tissues by immunohistochemistry. The gankyrin signal was observed mainly in the cytoplasm and occasionally in the nucleus of HCC cells (Fig. 2A-C). Although at lower levels compared with those in HCCs, weak but reproducible gankyrin signals were observed in the cytoplasm of the hepatocytes in the noncancerous tissues (Fig. 2D). Expression of gankyrin was not detected in the bile duct cells, blood endothelial cells, or other nonparenchymal cells in the liver tissues. Of 43 HCCs examined, the cytoplasm was stained positively for gankyrin in 27 (63%), and 9 of them (21%) were also positive for nuclear staining. Of 32 noncancerous liver tissues available, gankyrin was positive in 17 (53%).

Figure 2.

Immunohistochemical detection of gankyrin in hepatocellular carcinoma (HCC). HCC sections were stained with mouse monoclonal anti-gankyrin antibody, and counterstained with hematoxylin. Positive immunostaining appears brown. (A) Positive staining for gankyrin in the cytoplasm of most HCC cells. (B) Barely detectable gankyrin signal in some HCC cells. (C) Presence of gankyrin in the nucleus of some HCC cells. (D) Stronger staining for gankyrin in HCC cells (right) than the neighboring cirrhotic hepatocytes (left). Bar, 50 μm.

As shown in Table 2, we analyzed an association between gankyrin protein expression and clinicopathological findings. No significant association between gankyrin expression in HCC cells and sex, age, tumor size, fibrotic change in noncancerous liver tissues, differentiation of the tumor cells, or hepatitis B or C virus infection was observed. Positive cytoplasmic staining for gankyrin of HCC cells was significantly associated with low TNM stage (stage I or II; P = 0.004), no capsular invasion (P = 0.018), no portal venous invasion (P = 0.008), and no intrahepatic metastasis (P = 0.012) of HCC. In noncancerous liver tissues, positive gankyrin staining of hepatocytes was associated with the cytoplasmic gankyrin positivity of HCC cells of the same patient (P = 0.021, Table 3), but not with the parameters examined except for the serum alpha-fetoprotein level (P = 0.015, Table 2).

Table 2. Gankyrin Expression and Clinicopathological Characteristics
 Gankyrin Expression in the Cytoplasm of
HCCP valueNoncancerous LiverP value
Negative (n = 16)Positive (n = 27)Negative (n = 15)Positive (n = 17)
  1. Abbreviations: HCV, anti-hepatitis C virus antibody; HBV, hepatitis B surface antigen; (+), positive or present; (−), negative or absent; AFP, serum alpha-fetoprotein; NS, not significant between any groups or combinations thereof.

Sex distribution      
 Male12150.32810111.000
 Female412 56 
Median age (years)64650.69663620.649
Virus marker  NS  NS
 HBV(+)/HCV(−)33 22 
 HBV(−)/HCV(+)1018 1111 
 HBV(+)/HCV(+)12 20 
 HBV(−)/HCV(−)24 04 
Median AFP (ng/mL)63.095.00.89025.0199.00.015
Median tumor size (cm)4.54.00.0984.54.00.372
Liver cirrhosis (+)9200.3169130.450
TNM stage      
 I and II5210.0048120.467
 III and IV116 75 
Histological differentiation      
 Well570.737630.243
 Moderate and poor1120 914 
Capsular formation (+)15210.22912131.000
Capsular invasion (+)950.018460.712
Portal venous invasion (+)720.008430.678
Intrahepatic metastasis (+)1060.012650.712
Gankyrin nuclear expression      
 Yes090.016250.403
 No1618 1312 
Table 3. Gankyrin Expression and Molecular Histological Markers
 Gankyrin Expression in HCC
NegativePositiveP value
  1. Abbreviations: HCC, hepatocellular carcinoma; non-HCC, noncancerous portion of the resected liver; IGFBP-5, insulin-like growth factor-binding protein 5.

Gankyrin expression in non-HCC   
 Negative (n = 15)870.021
 Positive (n = 17)215 
p53 expression in HCC   
 Negative (n = 30)11191.000
 Positive (n = 13)58 
MDM2 expression in HCC   
 Negative (n = 33)14190.276
 Positive (n = 10)28 
IGFBP-5 expression in HCC   
 Negative (n = 25)13120.026
 Positive (n = 18)315 
IGFBP-5 expression in non-HCC   
 Negative (n = 23)1490.011
 Positive (n = 9)18 

Because expression of gankyrin affects the degradation of p53 and MDM2,16 we examined the expression of p53 and MDM2 as well as gankyrin in HCCs. By immunohistochemistry, nuclear expression of p53 and MDM2 were detected in 30% and 23%, respectively, of 43 HCCs (Fig. 3, Table 3). Positive staining for gankyrin was not associated with the staining for p53 nor MDM2 in HCC cells.

Figure 3.

Immunohistochemical detection of p53 and MDM2 in hepatocellular carcinoma (HCC). HCC sections were stained with antibodies specific to p53 (A and B) or MDM2 (C and D), and counterstained with hematoxylin. Positive immunostaining appears brown. (A) Positive staining for p53 in the nucleus of most HCC cells. (B) Negative p53 in HCC cells. (C) Positive staining for MDM2 in the nucleus of most HCC cells. (D) Negative MDM2 in HCC cells. Bar, 50 μm.

Up-regulation of IGFBP-5 Expression by Gankyrin in HCCs.

Preliminary microarray analysis of the cDNA libraries prepared from U-2 OS cells and Huh-7 cells overexpressing gankyrin suggested that IGFBP-5 mRNA was up-regulated by gankyrin (A. Umemura and J. Fujita, unpublished data). Real-time RT-PCR analysis confirmed that overexpression of gankyrin increased the IGFBP-5 mRNA levels 5.2-fold and 1.7-fold (mean, n = 3 each) in U-2 OS and Huh-7 cells, respectively, and western blot analysis demonstrated that the protein levels were increased as well (Fig. 4A). Conversely, when gankyrin expression was suppressed by siRNA, IGFBP-5 expression was down-regulated (Fig. 4B). In 2 of 3 HCC tissues overexpressing gankyrin, the levels of IGFBP-5 protein were higher compared with those in noncancerous tissues (Fig. 1C). To identify a role that IGFBP-5 might play in HCC cells, we next suppressed IGFBP-5 expression by siRNA. No apoptosis was induced, but viable cell numbers were decreased in Huh-7 as well as U-2 OS cells (Fig. 4C,D, and data not shown), suggesting a growth-promoting effect of IGFBP-5.

Figure 4.

Induction of IGFBP-5 by gankyrin. (A) U-2 OS cells (lanes 1 and 2) and Huh-7 cells (lanes 3 and 4) transiently transfected with plasmids expressing gankyrin or vector alone were analyzed for expression of IGFBP-5 by western blotting using the indicated antibodies. Representative results from more than 3 experiments are shown. (B) Huh-7 cells, mock transfected or transfected with siRNA for gankyrin or control RNA as indicated, were analyzed as in (A). (C) Suppression of IGFBP-5 expression by siRNA. Huh-7 cells were transfected with control RNA or IGFBP-5–specific siRNA. IGFBP-5 transcript levels were determined by real-time RT-PCR and normalized with β-actin levels. Results from 3 repeats were averaged and expressed relative to control. Error bars refer to standard deviation of the average quantitated results. (D) Effect of IGFBP-5 down-regulation on cell growth. U-2 OS and Huh-7 cells were transfected with IGFBP-5 siRNA or control RNA, and 72 hours later viable cell numbers were determined. Values are mean ± standard deviation (n = 3) and expressed relative to controls. ** and *, P < 0.01 and P < 0.05, respectively.

The expression of IGFBP-5 was further examined immunohistochemically in 43 HCC and 32 noncancerous liver tissues (Fig. 5, Table 3). In 42% of HCCs, IGFBP-5 was positively stained in the cytoplasm of HCC cells (Fig. 5A). IGFBP-5 was also detected, although at lower levels, in the cytoplasm of hepatocytes in 28% of the noncancerous tissues (Fig. 5B-D), but not in bile duct cells, blood endothelial cells, or other nonparenchymal cells.

Figure 5.

Immunohistochemical detection of IGFBP-5 in hepatocellular carcinoma (HCC). HCC sections were stained with anti–IGFBP-5 antibody and counterstained with hematoxylin. Positive immunostaining appears brown. (A) Positive staining for IGFBP-5 in the cytoplasm of HCC cells, especially at the invasive boundaries. (B) Presence of IGFBP-5 in noncancerous cirrhotic hepatocytes. (C) Stronger staining for IGFBP-5 in HCC cells (upper) than the neighboring cirrhotic hepatocytes (lower). (D) Positive staining for IGFBP-5 in HCC cells (upper left), but negative in cirrhotic cells (lower right). Bar, 100 μm. (E) Correlation of expression levels of gankyrin and IGFBP-5 in HCCs. The immunostaining levels were expressed as 0 (negative), 1+ (weakly positive), 2+ (moderately positive), or 3+ (strongly positive). Each diamond represents 1 case. The Spearman's rho = 0.629, P < 0.001. (F) Correlation of expression levels of gankyrin and IGFBP-5 in noncancerous hepatocytes determined as in (E). The Spearman's rho = 0.606, P < 0.001.

Specific cytoplasmic staining for IGFBP-5 in HCC cells was associated with low TNM stage (stage I or II; P = 0.013), no portal venous invasion (P = 0.006), low serum alpha-fetoprotein value (P = 0.031), and small tumor size (P = 0.009). No association with capsular invasion or intrahepatic metastasis was observed. There was a significant association between positivities for IGFBP-5 and gankyrin (Table 3), and the levels of expression covaried both in HCCs (ρ = 0.629, P < 0.001) (Fig. 5E) and noncancerous hepatocytes (ρ = 0.606, P < 0.001) (Fig. 5F).

Expression of Gankyrin in HCC and Patient Prognosis.

When we examined the relationship between gankyrin expression in HCC cells and the survival of patients after surgical resection, a significant difference was observed between the patients with gankyrin-positive HCCs and those with gankyrin-negative HCCs (Fig. 6). We found no significant difference in the survival rates between the patients whose HCCs stained positively and negatively for p53, MDM2, or IGFBP-5.

Figure 6.

Survival of patients and expression of molecular markers. The Kaplan-Meier method was used to determine the patient survival and log-rank test to compare survival between patients with HCC grouped according to (A) gankyrin positivity, (B) p53 positivity, (C) MDM2 positivity, and (D) IGFBP-5 positivity. P, positive. N, negative.

Discussion

Gankyrin is as an oncogene, mRNA of which is overexpressed in almost all human HCCs.9, 19 Although less frequent, gankyrin has been found by RNA dot blot analysis to be overexpressed in additional tumors including those of the breast, colon, rectum, stomach, small intestine, pancreas, ovary, lung, and thyroid (A. Umemura and J. Fujita, unpublished data). In the current study, we immunohistochemically examined the gankyrin protein expression in HCCs using the monoclonal anti-gankyrin antibody and found that the protein was highly expressed in the cytoplasm of 63% of HCCs. Tan et al.20 has similarly found overexpression of gankyrin protein in 60% of HCCs using a polyclonal antibody. The reason why the protein is not overexpressed in one-third of HCCs despite overexpression of its mRNA is unknown. The posttranscriptional, translational, and posttranslational regulations of gankyrin expression remain to be elucidated.

According to the 15th follow-up survey by the Liver Cancer Study group of Japan, the cumulative survival rates after surgical removal of HCC are 52.3% and 27.3% at 5 and 10 years, respectively, and better survival rates are associated with fewer numbers of tumors, lack of portal venous invasion, and early clinical stages.4–6 Consistent with these observations, gankyrin positivity of HCC was associated with low TNM stage, lack of capsular invasion, portal venous invasion, and intrahepatic metastasis, and better prognosis of the patients. Patients with hyperdiploid acute lymphoblastic leukemia with more than 50 chromosomes, one of the 6 subtypes of pediatric acute lymphoblastic leukemia, have an excellent prognosis compared with other subtypes, and interestingly, overexpression of gankyrin is 1 of the diagnostic and subclassification markers for it.21 Expression of gankyrin protein may be used as a marker for better prognosis of the patients with HCC as well.

The gankyrin oncoprotein plays a key role in regulation of cell cycle and apoptosis, at least in cultured cells, by inhibiting Rb and p53.10 In a rodent hepatocarcinogenesis model, hypermethylation of the p16INK4A gene and p53 mutation appear at a late stage, whereas gankyrin is overexpressed from early after carcinogen treatment, preceding the loss of Rb protein and adenoma formation.22 Clinically, p53 mutation is not so frequent in HCCs (15%–30%), especially in low-grade or low-stage HCCs.23, 24 Tan et al.20 have immunohistochemically detected gankyrin overexpression in 82%, 63%, and 22% of Edmondson's grade I to II, III, and IV HCCs, respectively. We observed gankyrin positivity in 81% and 35% of low and high TNM stage HCCs, respectively. These results suggest that gankyrin plays an important role(s) at early stages of hepatocarcinogenesis by suppressing Rb, p53 and possibly other tumor suppressors. In advanced HCCs, by contrast, oncogenic mutations probably have accumulated in many genes including p53, and overexpression of gankyrin may not be so crucial as in early stage HCCs. This could explain the present association of gankyrin-negative HCCs with poorer prognosis and the finding that both cases of gankyrin-negative HCCs with gankyrin-positive noncancerous hepatocytes belonged to high TNM stages. This is, however, one of several possible explanations, and further work is necessary to clarify the exact reasons for the observed association.

By immunohistochemical staining, p53 has been detected in 20% to 30% of HCCs.25, 26 Although strong immunohistochemical reactivity for p53 may not be an indicator of the presence of p53 gene mutations as initially suggested,26 it has been associated in some studies with higher proliferative activity, lower differentiation of HCC cells, or poorer survival of patients. Endo et al.27 immunohistochemically detected MDM2 in 28 of 107 (26%) HCCs, and the positive expression correlated with the presence of p53 mutation and poorer prognosis, although it also correlated with smaller HCC size and the absence of vascular invasion. We immunohistochemically detected the expression of p53 and MDM2 in 30% and 23%, respectively, of HCCs, which is in accord with other studies, but no correlation was seen between expression and survival of the patients. Gankyrin accelerates degradation of Rb, p53, and MDM2 in cultured cells.9, 16 Although some correlation between expression of gankyrin and Rb has been suggested in HCC tissues,20 we did not observe significant relationship between the gankyrin positivity and negative staining for p53 nor MDM2. The analysis of individual cells for protein expression, for example by double 2-color immunostaining, may have revealed the presence of some relationship. But most probably, our finding reflects complex interrelated mechanisms regulating the levels of these proteins and also suggests that the relevance of the effects of gankyrin on p53, MDM2, and Rb demonstrated in cultured cells to human hepatocarcinogenic process remains to be firmly established.

The 6 members of IGFBP family (IGFBP-1 through IGFBP-6) are important components of the insulin-like growth factor (IGF) axis, and regulate the activity of both IGF-I and IGF-II polypeptide growth factors.28 IGF-I, IGF-II, and their receptors are expressed in a wide variety of cells, and the liver is the main source of circulating IGF-I. IGFBPs are also secreted by many cell types, and their expression is regulated in a cell-dependent and tissue-type–dependent manner. In the current study, we found up-regulation of IGFBP-5 mRNA and protein levels by overexpression of gankyrin in human osteosarcoma and HCC cell lines and consistently detected a significant association between the protein levels of gankyrin and IGFBP-5 in HCC specimens. In the proximal promoter region of the IGFBP-5 gene, there are several putative transcription-factor–binding sites including those for AP-2, c-Myb, C/EBP, and NF-1, and responsive elements to prostaglandin E2, cyclic adenosine monophosphate, progesterone/retinoic acid, and Akt.28 Whether the effect of gankyrin on IGFBP-5 expression is mediated by these factors is unknown.

The IGFBPs bind IGFs with high affinity, and they are able to enhance or inhibit the activity of IGFs in a cell-specific and tissue-type–specific manner.28 In addition, IGFBPs have IGF-independent effects. There are several reports on the relationship between the IGF axis and HCC.29–31 IGFBP-3 is the most abundant IGFBP present in noncancerous liver tissue and could serve as a negative regulator of cell proliferation in human HCCs.32 Although the presence of IGFBP-5 in numerous tumors and cell lines has been demonstrated, its expression and significance in human HCC have not been documented. We found positive staining for IGFBP-5 in 42% of HCCs, and the positivity correlated with absence of portal venous invasion, low TNM stage, and small tumor size. Although not statistically significant, patients with IGFBP-5–positive HCCs tended to survive longer than those with IGFBP-5–negative HCCs. These findings are essentially similar to those observed for gankyrin. Regarding the effect of IGFBP-5 on cell proliferation, there are contradictory findings.28 In breast cancer cells, many studies have reported inhibition of growth, but there are some indicating a stimulatory effect.33 IGFBP-5 is up-regulated in involuting prostate but is also implicated in growth stimulation of prostate tumor cells.34 We found that down-regulation of IGFBP-5 suppresses growth of Huh-7 HCC cells. Thus, these findings are consistent with a notion that high expression of IGFBP-5 and gankyrin play oncogenic roles in HCCs of early clinical stages. Clarification of the exact roles played by them will shed more light on the molecular mechanisms of human hepatocarcinogenesis and lead to development of new therapeutic and preventive strategies.

Acknowledgements

We thank Dr. R. John Mayer for helpful suggestions.

Ancillary