Fax: (011) 886 73468056
Expression and prognostic role of tumor suppressor gene PTEN/MMAC1/TEP1 in hepatocellular carcinoma
Article first published online: 1 APR 2003
Copyright © 2003 American Cancer Society
Volume 97, Issue 8, pages 1929–1940, 15 April 2003
How to Cite
Hu, T.-H., Huang, C.-C., Lin, P.-R., Chang, H.-W., Ger, L.-P., Lin, Y.-W., Changchien, C.-S., Lee, C.-M. and Tai, M.-H. (2003), Expression and prognostic role of tumor suppressor gene PTEN/MMAC1/TEP1 in hepatocellular carcinoma. Cancer, 97: 1929–1940. doi: 10.1002/cncr.11266
- Issue published online: 1 APR 2003
- Article first published online: 1 APR 2003
- Manuscript Accepted: 9 DEC 2002
- Manuscript Revised: 25 NOV 2002
- Manuscript Received: 5 JUL 2002
- National Science Council of Taiwan. Grant Numbers: NSC-89-2315-B-182A-005, NSC-89-2320-B-075B-018
- Kaohsiung Veterans General Hospital. Grant Number: VGHKS-89-17
- Chang Gung Memorial Hospital. Grant Number: CMRP-1130
- Academic Excellence Program, Ministry of Education, Taiwan. Grant Number: 89-B-FA08-1-4
- hepatocellular carcinoma (HCC)
Inactivation of the tumor suppressor gene PTEN/MMAC1/TEP1, located on chromosome 10q23, is a common event in advanced stages of diverse human malignancies. However, the prognostic role of PTEN expression in patients with hepatocellular carcinoma (HCC) has not been characterized.
One hundred five resected specimens were collected from patients with HCC. Expression levels of PTEN and p53 in clinical samples were analyzed by immunohistochemistry.
Immunohistochemical analysis of 105 HCC tissue specimens revealed that decreased or absence of PTEN immunostaining was found in 43 specimens (40.9%). Reduced PTEN expression levels were correlated with increased tumor grade (P = 0.017), advanced disease stage (P = 0.016), and elevated serum α-fetoprotein (αFP) levels (P = 0.001). Kaplan–Meier analysis indicated that patients with reduced PTEN levels had shorter overall survival (P = 0.001) and higher recurrence rates (P = 0.0007) compared with patients who had intact PTEN expression. Examining p53 expression unveiled an inverse correlation between p53 overexpression and reduced PTEN expression in patients with HCC (P = 0.004). In addition, patients with p53 overexpression had shorter overall survival compared with patients who were without p53 overexpression (P = 0.0014). Univariate and multivariate analyses revealed that reduced PTEN expression was an independent prognostic factor for survival in patients with HCC.
The current study demonstrated that reduced PTEN expression levels are involved in the pathogenesis of HCC. Moreover, decreased PTEN expression was correlated with tumor progression, high αFP levels, p53 overexpression, and poor prognosis in patients with HCC. Cancer 2003;97:1929–40. © 2003 American Cancer Society.
Hepatocellular carcinoma (HCC) is one of the most prevalent malignancies worldwide, with an annual incidence of 250,000 diagnoses. However, to our knowledge little is known regarding the molecular mechanisms underlying the progression of HCC. Epidemiologic studies have shown that HCC frequently is associated with hepatitis B virus (HBV) infection or hepatitis C virus (HCV) infection, exposure to aflatoxin B1, and alcoholic intake. Hepatocarcinogenesis is considered a multifactorial and multistep process that involves the activation of oncogenes or the inactivation of tumor suppressor genes in different stages of HCC progression. Inactivation of tumor suppressor genes usually occurs as a consequence of mutation of one allele or the loss of one allele (loss of heterozygosity [LOH]). In HCC, LOH has been reported on chromosomes 1p, 4q, 5q, 8p, 10q, 11p, 13q, 16, 17p, and 22q.1 It is known that several tumor suppressor genes participate in the development and progression of HCC, including the p53 gene on chromosome 17q13, the Rb gene on chromosome 13q14, and the APC gene on chromosome 5q21. In addition to these loci, allelic loss or LOH of chromosome 10q frequently has been characterized in HCC, suggesting the presence of tumor suppressor gene(s) on chromosome 10q in HCC.2–4
The tumor suppressor gene PTEN (also known as MMAC1 or TEP1) is located on human chromosome 10q23.5–7 Mutation of PTEN is a common event in advanced stages of diverse human malignancies, occurring in approximately 70% of patients with glioblastoma, 50% of patients with endometrial carcinoma, 50% of patients with prostate carcinoma, and 30% of patients with melanoma.8 In addition, germline mutations in PTEN gene are associated with the dominantly inherited Cowden and Bannayan–Zonana syndromes, which are characterized by the formation of multiple benign tumors and by an increased risk of developing malignant breast and thyroid tumors. The 403-amino acid PTEN protein encodes dual specificity protein phosphatases. The two major substrates for PTEN are phosphatidylinositol triphosphate (PIP3) phospholipids and the oncogene Akt/PKB. Both are important signal transducers for growth, survival, and proliferation of cells.9
The involvement of PTEN inactivation during HCC progression remains to be elucidated. Gene knockout studies revealed that PTEN (+/−) heterozygous mice exhibited neoplasm of the liver, suggesting that loss of PTEN may participate in liver carcinogenesis.10 Furthermore, loss of a PTEN allele was identified in 20–30% of patients with HCC.11, 12 A recent cDNA microarray analysis indicated that levels of PTEN transcripts were decreased in a significant number of patients with HCC.13 However, DNA sequencing analysis indicated that mutations in the PTEN gene rarely occurred in HCC specimens.11, 12, 14, 15 In addition, some authors had a contradictory view that PTEN may not play a role in the pathogenesis of hepatoma.15 To gain further insights into the role of PTEN inactivation during HCC progression, the levels and localization of PTEN protein expression were analyzed in 105 HCC specimens using immunohistochemistry. In addition, the correlations between PTEN expression and clinicopathologic parameters in patients with HCC were investigated using statistical analysis in the current study.
MATERIALS AND METHODS
Raising Antibodies Against PTEN
Human PTEN cDNA was subcloned into pET-15b (Novagene; Madison, WI) for generation of 6xHis-tagged PTEN protein. The pET-15b-PTEN was transformed into E. coli expression host, BL21 cells (Novagene), grown to log phase (outer dimension, 600 nm ≈ 0.5–0.9), and supplemented with 1 mM isopropylthio-β-D-galactoside to induce protein expression. Recombinant PTEN was purified using nitrilotriacetic acid agarose (Qiagen GmbH, Germany) according to procedures recommended by manufacturer. The preimmune serum was collected from rabbits (New Zealand White strain) prior to immunization. For generation of PTEN antibodies, recombinant PTEN protein (100–200 μg in 0.5 mL) was mixed with 0.5 mL of complete adjuvant (Sigma, St. Louis, MO) and was injected subcutaneously into rabbits. Next, antigen was mixed with incomplete adjuvant (Sigma) at a 1:1 ratio and was injected into rabbit every 2 weeks to raise antibodies. The blood was collected periodically by ear bleeding, and the immunoglobulin G (IgG) fraction was purified using protein A-Sepharose 4B columns (Amersham Pharmacia Biotech, Buckinghamshire, UK) when necessary. The titer and specificity of purified antibodies were analyzed by Western blot analysis.
Western Blot Analysis
SAOS-2, HepG2, and PTEN-deficient U87MG cells16 were purchased from the American TypeCulture Collection (Manassas, VA). Cell extracts were prepared using lysis buffer containing 40 mmol/L N-2-hydroxyethyl-piperazine N′-2-ethane sulphonate, 1% Triton X-100, 10% glycerol, and 1 mmol/L phenylmethanesulfonyl fluoride. The cell lysate was centrifuged at 6000 revolutions per minute for 5 minutes at 4 °C with a microcentrifuge and collected to determine protein concentration with a bichloroacetic acid protein assay kit (BCA kit; Pierce, Rockford, IL). An aliquot of 20 μg protein from each sample was separated on 10% sodium dodecyl sulfate-polyacrylamide gels. Proteins were transferred subsequently onto polyvinylidene difluoride membranes (Immobilon-P membrane; Millipore, Bedford, MA). Membranes were incubated with PTEN antibodies or preimmune serum for 1 hour, then washed and incubated with a horseradish peroxidase-conjugated antirabbit immunoglobulin (Ig)G antibody (1:5000 dilution; Vector Laboratories, Burlingame, CA) for 30 minutes. Immunoreactivity was detected by ECL plus chemiluminescence kit (Amersham Pharmacia Biotech).
Resected HCC Specimens
A total of 105 HCC surgical resection specimens were collected from the Department of Pathology at Kaohsiung Chang Gung Memorial Hospital between January 1986 to December 1996. All patients with HCC were diagnosed with resectable tumor(s) after undergoing liver biochemical tests, sonograms, computed tomography scans, and angiography. Patients who died due to complications or postsurgical hepatic failure were excluded from this study. The closing date of follow-up was December 31, 2001. The duration of follow-up was estimated in months. The status of hepatitis markers, αFP levels, and other clinical parameters also were documented for all patients. All HCC specimens were comprised of both tumor and adjacent nontumor parts. Tumor size was recorded as the greatest dimension of each specimen. The background of the nontumor part of each specimen was characterized as cirrhotic or noncirrhotic. Tumors were divided into three groups according to their grade of differentiation: well differentiated tumors (Grade 1 carcinoma; Edmondson–Steiner classification), moderately differentiated tumors (Grade 2 carcinoma; Edmondson–Steiner classification), and poorly differentiated tumors (Grade 3–4 carcinoma; Edmondson–Steiner classification). HCC tumors were classified according to the staging system by the International Union against Cancer with minor modification as follows: Stage I, encapsulated and without evidence of liver or vascular invasion; Stage II, unencapsulated or capsulated and with liver invasion but without vascular invasion; Stage III, invasion of small vessels in the tumor capsule or focal invasion of portal vein branches close to the tumor; and Stage IV, invasion of the portal vein in distal the liver (1 cm away from the tumor capsule), branches of the major portal vein, and the common bile duct or perforation into visceral peritoneum.
The paraffin sections from HCC specimens were deparaffinized, blocked with 3% hydrogen peroxide for 10 minutes, and subjected to antigen retrieval with microwave in 0.01 M citrate buffer for 15 minutes. The slides then were washed twice with phosphate-buffered saline, incubated with PTEN antibodies (1:100 dilution, 2 μg/mL) or with p53 antibodies (1:100 dilution; Dakopatts, Copenhagen, Denmark) for 30 minutes, then detected with peroxidase conjugate using a polymer detection system (catalog no. 87-89431; Zymed, San Francisco, CA) for 30 minutes. The antibody staining was visualized with 3,3′-diaminobenzidine tetrahydrochloride (Sigma) in 0.1 M Tris, pH 7.2, containing 0.01% hydrogen peroxide. The section slides were counterstained with Gill hematoxylin, dehydrated, and mounted. The intensities of PTEN immunostaining in the tumor and nontumor parts of clinical samples were evaluated by two independent pathologists using a scoring system from 0 to 3 to represent null, weak, intermediate, and strong intensities, respectively. The putative normal liver tissue was collected from an individual with traumatic internal bleeding by partial hepatectomy and was used as a positive control for PTEN staining. For a negative control, normal liver tissue was stained with preimmune rabbit serum (1:100 dilution). Reduced PTEN expression was defined as the lower PTEN labeling index in HCC tissues compared with adjacent normal tissue. Because staining variation was found frequently, especially in the nonneoplastic liver tissue between the peritumoral region and the region far away from the tumor, the staining intensity in the major portion of peritumoral liver tissue was scored to represent the labeling index of nontumor parts. The p53 expression in HCC samples was scored as described previously.17, 18 Briefly, expression of p53 in tumor tissues was scored based on the percentage of cells, with positive nuclear staining scored as 0 (< 10%), 1 (11–40%), 2 (41–70%), and 3 (> 71%). p53 positive was defined as nuclear p53 scores ≥ 1, and p53 negative was defined as nuclear p53 scores < 1. Subsequently, the correlation between labeling scores and clinicopathologic features of HCC specimens, including histologic tumor grades (differentiation) and stages, survival outcome, serum levels of αFP, cirrhotic background, hepatitis markers, and tumor sizes, were analyzed by statistical analysis.
Patient ages and tumor sizes were expressed as the mean + the standard deviation. Follow-up after surgery and serum αFP levels were recorded as medians with quartiles. Comparisons between groups of related samples were assessed with the Wilcoxon paired-sample test. Comparisons between groups of independent samples were assessed with a Student t test, one-way analysis of variance (ANOVA), Mann–Whitney U test, or Kruskal–Wallis test. The associations between categorical variables were assessed using the chi-square test or the Fisher exact test. Correlations between continuous variables were determined using the Spearman rank correlation test. Survival rates were calculated using the Kaplan–Meier method, and the difference in survival was compared with the log-rank test. The influence of various clinicopathologic features on overall survival was assessed with a Cox proportional hazards model. A P value < 0.05 was considered statistically significant.
Clinical Parameters of Patients with HCC
The surgically resected specimens were collected from 105 patients with HCC, including 84 males and 21 females who ranged in age from 25 years to 81 years, with a mean age of 55.7 years ± 12.2 years. The overall survival rate was 49.5% (52 of 105 patients) at the end of at least 5 years of follow-up. Periods of follow-up ranged from 2 months to 147 months, with quartiles of 17 months, 54 months, and 68 months. Tumor sizes ranged from 1 to 20 cm, with a mean size of 7.19 cm ± 4.15 cm. The extent of differentiation in HCC samples was graded as well differentiated in 26 samples, moderately differentiated in 52 samples, and poorly differentiated in 27 patients. Tumors were divided into Stage I in 16 samples, Stage II in 33 samples, Stage III in 33 samples, and Stage IV in 23 samples. Due to the few samples that were categorized as Stage I tumors and Stage IV tumors, we simplified the categories as early stage (Stage I and II) and late stage (Stage III and IV) for comparison. In the background of adjacent nontumor sections, cirrhosis was found in 66 sections (62.9%), and noncirrhotic tissue was found in 39 sections (37.1%). Overall, patients who had late-stage disease (Stage III and IV) had larger tumors (P < 0.001; Student t test) and higher α AFP levels (P < 0.001; Mann–Whitney U test) compared with patients who had early-stage disease (Stage I and II). In addition, younger age (P = 0.039; one-way ANOVA) and higher αFP levels (P = 0.001; Kruskal–Wallis test) were apparent in patients with poorly differentiated HCC.
Hepatitis viruses are known to participate in the tumorigenesis of HCC. Thus, the status of HBV and HCV in these patients with HCC was examined. Hepatitis B surface antigen was detected in the sera of 69 patients (65.7%), and anti-HCV was detected in 29 patients (27.6%). Four patients (3.8%) were positive for both HBV and HCV, whereas six patients (5.7%) were negative for both markers. Patients who had HBV positive HCC were younger compared with patients who had HCV positive HCC (P = 0.008). In addition, they had higher serum αFP levels (P = 0.001), a higher male ratio (P = 0.001), and larger tumors (P = 0.041) compared with patients who had HCV positive HCC. However, there was no correlation among hepatitis markers and other parameters that were measured in this study, including tumor grade and disease stage.
Reduced PTEN Expression in HCC Specimens
For immunohistochemical studies, PTEN antibodies were raised against recombinant PTEN protein (with a molecular weight of approximately 45 kilodaltons [kD]) and characterized by Western blot and histologic analyses (Fig. 1). Immunohistochemical studies were performed using PTEN antibodies to analyze the profile of PTEN expression in HCC specimens. PTEN immunostaining was detected in both hepatocytes and tumor cells (Fig. 2) but was not observed in infiltrating mononuclear cells or in sinusoidal lining cells. The intensity of PTEN immunostaining in tumor and nontumor regions was expressed as a labeling score for statistical analysis. In 105 HCC specimens, 43 specimens (40.9%) had lower intensity of PTEN immunostaining in tumors compared with their adjacent, nontumor tissues (Fig. 2). Statistical analysis indicated that PTEN scores in nontumor tissues were significantly higher compared with the scores in HCC tissues (P < 0.001; Wilcoxon paired-sample test).
With regard to subcellular localization, PTEN staining was detected in the cytoplasm and nuclei of HCC samples (Fig. 3). Nuclear PTEN staining was localized primarily in nucleoli and the nuclear membrane, whereas cytoplasmic PTEN staining was diffused in cells. There was a noticeable reduction in cytoplasmic PTEN staining in tumor tissues compared with the staining in their adjacent, nontumor tissues (Fig. 3). In contrast, nuclear PTEN staining decreased in tumor tissues but remained detectable in some specimens (Fig. 3A). Statistical analysis indicated that reduced PTEN expression occurred in the cytoplasm (P < 0.001; Wilcoxon paired-sample test) and the nuclei in HCC tissues (P = 0.002; Wilcoxon paired-sample test).
Correlation of PTEN Expression with Clinicopathologic Parameters
PTEN inactivation is found frequently in patients with advanced-stage HCC.5, 7 In addition, reexpression of PTEN restores the differentiation features of tumor cells.19, 20 Thus, the correlation between PTEN expression and grades of differentiation was investigated in the HCC specimens (Fig. 4, left photomicrographs). There was prominent PTEN staining in well-differentiated and moderately differentiated HCC specimens (Fig. 4A, B). However, PTEN staining was decreased profoundly in poorly differentiated HCC specimens (Fig. 4C, D). The correlation between PTEN expression and clinicopathologic parameters in patients with HCC also was analyzed statistically. Table 1 shows that reduced PTEN expression in patients with HCC was correlated with increased tumor grades (P = 0.017), advanced tumor stages (P = 0.015), and high serum αFP levels (P = 0.001).
|Variable||No. of patients (%)||P value|
|Total||PTEN expression (%)|
|Intact (n = 62 patients)||Reduced (n = 43 patients)|
|Male||84||48 (57)||36 (43)||NSa|
|Female||21||14 (67)||7 (33)||—|
|With||66||36 (55)||30 (45)||NSa|
|Without||39||26 (67)||13 (33)||—|
|Positive||69||38 (55)||31 (45)||NSa|
|Negative||31||22 (71)||9 (29)||—|
|Positive||29||20 (69)||9 (31)||NSa|
|Negative||71||40 (56)||31 (44)||—|
|Well||26||19 (73)||7 (27)||0.017a|
|Moderate||52||32 (62)||20 (38)||—|
|Poor||27||11 (41)||16 (59)||—|
|I and II||49||35 (71)||14 (29)||0.016a|
|III and IV||56||27 (48)||29 (52)||—|
|Negative||66||46 (70)||20 (30)||0.004a|
|Positive||39||16 (41)||23 (59)||—|
|Age (mean yrs ± SE)||—||56.4 ± 1.9||54.8 ± 1.50||NSc|
|Tumor size (mean cm ± SE)||—||6.3 ± 0.51||6.2 ± 0.50||NSc|
|αFP (mean ng/mL ± SE)b||—||1599 ± 774||4127 ± 1832||0.001d|
Correlation between PTEN Expression and p53 Expression in HCC Specimens
The tumor suppressor gene p53 plays an important role in the genesis and progression of HCC.17, 18, 21, 22 Thus, we analyzed p53 expression profiles and compared them with the profiles of PTEN expression in HCC specimens. Immunohistochemical studies demonstrated that p53 staining was localized sporadically in the nuclei in HCC tissues (Fig. 4, right photomicrographs). We used an index system, in which p53 negative indicated the group of patients with p53 scores < 1, and p53 positive indicated the group of patients with p53 scores ≥ 1, to investigate the role of p53 expression in HCC. Statistical analysis demonstrated that p53 positive status was associated with increased grade of differentiation (P = 0.015) and advanced disease stage (P = 0.011; chi-square test) (Table 2) in patients with HCC.
|Variable||No. of patients (%)||P value|
|Total||p53 expression (%)|
|Intact (n = 66 patients)||Reduced (n = 39 patients)|
|Male||84||51 (61)||33 (39)||NSa|
|Female||21||15 (71)||6 (29)||—|
|With||66||37 (56)||29 (44)||NSa|
|Without||39||29 (74)||10 (26)||—|
|Positive||69||42 (61)||27 (39)||NSa|
|Negative||31||20 (65)||11 (35)||—|
|Positive||29||17 (59)||12 (41)||NSa|
|Negative||71||45 (63)||26 (37)||—|
|Well||26||21 (81)||5 (19)||0.015a|
|Moderate||52||32 (62)||20 (38)||—|
|Poor||27||13 (48)||14 (52)||—|
|I and II||49||37 (76)||12 (24)||0.012a|
|III and IV||56||29 (52)||27 (48)||—|
|Negative||62||46 (74)||16 (26)||0.004a|
|Positive||43||20 (47)||23 (53)||—|
|Age (mean yrs ± SE)||—||57.1 ± 1.5||53.5 ± 1.9||NSc|
|Tumor size (mean cm ± SE)||—||6.1 ± 0.4||6.5 ± 0.6||NSc|
|αFP (mean ng/mL ± SE)||—||1599 ± 595||4127 ± 2129||NSd|
Because p53 also is involved in maintaining chromosome stability, cells that lack functional p53 are susceptible to the loss of genetic materials during DNA replication, and they develop allelic loss.19 Because allelic loss is one the mechanisms for PTEN inactivation in tumors, we investigated the correlation between PTEN expression and p53 expression in HCC samples. Table 2 shows that 23 of 43 patients (54%) with reduced PTEN expression had positive p53 status, whereas only 16 of 62 patients (26%) with intact PTEN expression had positive p53 status. Statistical analysis indicated there was an inverse correlation between reduced PTEN expression and p53 overexpression (P = 0.004; chi-square test). Together, these data suggest that reduced PTEN expression frequently occurs in hepatoma cells with defective p53, and vice versa.
Correlation of PTEN Expression and p53 Expression with the Survival of Patients with HCC
It has been shown that functional PTEN expression suppresses the growth and invasive properties of tumor cells, thereby enhancing the survival and outcome in many types of cancer patients.23–27 Thus, we hypothesized that reduced PTEN expression may influence the probability of survival and prognosis in patients with HCC. Indeed, Kaplan–Meier analysis indicated that patients who had reduced PTEN expression levels had a shorter survival compared with patients who had intact PTEN expression (Fig. 5A) (P = 0.001). The median survival rates among patients with HCC in the group with intact PTEN expression and the group with reduced PTEN expression were 76 months and 30 months, respectively. We also examined p53 status in these patients to compare the prognostic value of PTEN with that of p53. Figure 5B shows that patients in the p53 negative group survived significantly longer compared with patients in the p53 positive group (P = 0.0014). The median survival rates for patients with p53 negative HCC and p53 positive HCC were 71 months and 32 months, respectively. These data were comparable to previous reports on p53 expression levels in the prognosis of patients with HCC.17, 18
After undergoing surgical resection, the mortality of patients with HCC results mainly from tumor recurrence. Therefore, the correlation between PTEN expression and tumor recurrence in patients with HCC was investigated. Kaplan–Meier analysis indicated that patients who had reduced PTEN expression had higher recurrence rates compared with patients who had intact PTEN expression (Fig. 6A) (P = 0.0007). For patients with intact PTEN expression, the 1-year, 3-year, and 5-year recurrent rates were 25%, 46%, and 62%, respectively. In contrast, the 1-, 3-, and 5-year recurrent rates for patients with reduced PTEN expression were 54%, 75%, and 84%, respectively. Likewise, patients who had p53 positive status had higher recurrent rates compared with patients who had p53 negative status (Fig. 6B) (P = 0.0014).
Univariate and Multivariate Analyses of Prognostic Factor for Patients with HCC
To evaluate the potential of using PTEN expression for the prognosis of patients with HCC after surgery, univariate analysis using a Cox proportional hazards model showed that both reduced PTEN expression and p53 overexpression were the prime variables for prognosis (Table 3). Furthermore, univariate analysis also identified serum αFP level (P = 0.005), HCC grade (P = 0.004), and HCC stage (P < 0.001) as prognostic factors for the outcome of patients with HCC (Table 3). After adjustment of all clinicopathologic features, multivariate analysis indicated that reduced PTEN expression remained correlated with the prognosis of patients with HCC (P = 0.020; Table 3); however, p53 overexpression was no longer a significant factor (Table 3). In the current study, reduced PTEN expression (P = 0.020), tumor stage (P < 0.001), and serum αFP level (P = 0.033) were independent prognostic factors for patients with HCC after they underwent surgical resection.
|Risk||95% CI||P value||Risk||95% CI||P value|
|PTEN expression (intact or reduced)||2.28||1.37–3.78||0.002a||1.85||1.10–3.09||0.020a|
|p53 expression (positive or negative)||2.23||1.34–3.71||0.002a||1.05||1.0–2.18||0.10|
|Stage (I and II or III and IV)||4.08||2.29–7.26||< 0.001a||3.64||2.03–6.52||< 0.001a|
|αFP (≥ 700 ng/mL or <700 ng/mL)||2.13||1.26–3.59||0.005a||1.81||1.06–3.08||0.033a|
|Age (≥ 65 yrs or < 65 yrs)||0.88||0.52–1.49||0.63||—||—||—|
|Cirrhosis (presence or absence)||1.18||0.69–2.01||0.54||—||—||—|
|HBV (positive or negative)||1.61||0.88–2.94||0.12||—||—||—|
|HCV (positive or negative)||0.64||0.35–1.16||0.14||—||—||—|
|Grade (1 and 2 or 3 and 4)||1.71||1.19–2.44||0.004a||1.02||1.01–1.52||0.45|
|Tumor size (≥ 5 cm or < 5 cm)||1.38||0.83–2.31||0.22||—||—||—|
|No. of tumors (single or multiple)||1.28||0.71–2.29||0.41||—||—||—|
In the current study, we provided evidence for decreased PTEN protein levels in > 40% of HCC samples and demonstrated the diagnostic value of reduced PTEN expression as an independent prognostic factor for patients with HCC after surgery. The immunohistochemical approach allowed examination of PTEN expression levels in the nontumor-tumor boundary as well as in different subcellular compartments in HCC specimens. These finding suggest that PTEN down-regulation frequently occurs in HCC and may constitute a prognostic factor for patients with HCC after surgery.
We observed a unique subcellular localization of PTEN staining in both the cytoplasm and nuclei of HCC specimens. Although PTEN does not possess a motif of nuclear localization signal, several studies have indicated that PTEN protein may be detected in nuclei.28–30 In addition, differential PTEN expression in cytoplasmic and nuclear compartments has been observed in some types of malignancies.28, 29 In endocrine pancreatic tumors, PTEN expression was localized primarily within the nucleus in nonneoplastic islets but shifted to the cytoplasm and plasma membrane in tumor tissues.28 In thyroid tumors, the exclusion of PTEN nuclear staining was associated with increased malignancy.29 Studies in neuronal differentiation have postulated that nuclear PTEN plays a role in cellular differentiation or metabolism of phosphatidylinositol phosphate in the nucleus.30 In the current study, reduced PTEN expression was associated with tumor progression in patients with HCC. Furthermore, both cytoplasmic and nuclear PTEN levels decreased in hepatoma tissues (Fig. 3). This differs from the nuclear exclusion of PTEN staining in malignancies of the pancreas and thyroid.17, 18 The mechanism underlying PTEN distribution in subcellular compartments and its pathologic consequences for disease progression in patients with HCC await clarification.
PTEN can be inactivated through a wide range of mechanisms, including somatic mutations, hemizygous deletions, promoter methylation, decreased transcription and/or translation, increased protein degradation and/or phosphorylation, and differential subcellular compartmentalization.28, 29, 31–33 In the current study, using immunohistochemistry, we found that approximately 40% of HCC samples had reduced PTEN expression levels. However, sequencing analysis of genomic DNA from 20 of those HCC samples did not find any mutations in the PTEN coding region (data not shown). Previous studies also have indicated that somatic mutation of the PTEN gene is a rare event in HCC samples (< 5%).11, 12, 14, 15 Together, these data suggest that gene mutation is not the primary mechanism for PTEN inactivation in HCC. Allelic loss has been proposed as the primary pathway for PTEN inactivation in many types of human malignancies.34, 35 In patients with HCC, the frequency of loss of flanking markers around the PTEN allele was determined at rates of 27% (25 of 89 patients) and 32% (12 of 37 patients).11, 12 The frequency of PTEN loss (≈ 40%) in the current study was closer in range to the rate of allelic loss (27–32%) compared with the rate of gene mutation (< 5%). However, we could not exclude the probable existence of other mechanisms for PTEN inactivation in HCC.
The prognostic role of structural abnormality in the p53 gene for HCC has been deciphered previously.17, 18, 22, 36 In the current study, we demonstrated that p53 overexpression occurs primarily in the nuclei of tumors from patients with advanced-stage HCC (Stage III–IV) and represents a prognostic factor for patients with HCC. Furthermore, an inverse correlation between p53 overexpression and reduced PTEN expression was characterized in HCC samples. The loss of PTEN alleles appears to be related to the lack of functional p53, which normally maintains chromosomal stability. Recently, a p53-binding element was identified in the promoter region of the PTEN gene,37 indicating that p53 may activate PTEN gene expression. There is evidence indicating that PTEN inhibits PI3K/Akt signaling and promotes degradation of Mdm2, thereby increasing the cellular content and transactivation of p53.38 The correlation between p53 overexpression and reduced PTEN expression in the current study supports the finding of an interaction between these two major tumor suppressors.
Recently, there was an immunohistochemical report on PTEN expression in tissues from HCV positive, cirrhotic patients with HCC.39 In that study, the PTEN immunostaining intensity was low in 29 of 46 HCC specimens (63.1%), a rate that is higher compared with the frequency of reduced PTEN expression in HCC specimens in the current study. In agreement with our findings, the authors of that recent report found that reduced PTEN expression is associated with increased grade, advanced tumor stage, high serum αFP level, and shorter overall survival. In the current study, we further revealed the influence of p53 overexpression on PTEN expression levels and showed that PTEN expression is an independent prognostic factor for survival in patients with HCC. However, we did not find a significant correlation between reduced PTEN expression and the status of cirrhosis or infection with HBV or HCV.
The mechanism underlying the poor survival outcome in patients with reduced PTEN expression may be attributed to their advanced stages, high αFP levels, and high recurrence rates (Table 1, Fig. 6A). Recently, we restored PTEN expression in PTEN-deficient hepatoma cells by using adenovirus gene delivery, which led to reductions in Akt/PKB phosphorylation, tumorigenicity, and invasive properties in hepatoma cells. Furthermore, administration of PTEN gene delivery vectors effectively reduced tumor growth and prolonged the survival rates of mice bearing hepatoma (P.-R. Lin et al., unpublished data). These data indicates that PTEN gene delivery may be applicable to the treatment of patients with HCC. Together with the findings from the current study, we propose that the tumor suppressor gene, PTEN, plays an important role during liver carcinogenesis. In addition, PTEN may have diagnostic and therapeutic potential for patients with HCC.
The authors thank Shang-Yun Liu for assistance in immunohistochemistry.
- 32Modulation of hypoxia-inducible factor 1alpha expression by the epidermal growth factor/phosphatidylinositol 3-kinase/PTEN/AKT/FRAP pathway in human prostate cancer cells: implications for tumor angiogenesis and therapeutics. Cancer Res. 2000; 60: 1541–1545., , , et al.
- 39Impact of PTEN expression on the outcome of hepatitis C virus-positive cirrhotic hepatocellular carcinoma patients: possible relationship with COX II and inducible nitric oxide synthase. Int J Cancer. 2002; 100: 152–157., , , et al.