Metastatic tumor antigen 1 is closely associated with frequent postoperative recurrence and poor survival in patients with hepatocellular carcinoma†
Article first published online: 27 FEB 2008
Copyright © 2008 American Association for the Study of Liver Diseases
Volume 47, Issue 3, pages 929–936, March 2008
How to Cite
Ryu, S. H., Chung, Y.-H., Lee, H., Kim, J. A., Shin, H. D., Min, H. J., Seo, D. D., Jang, M. K., Yu, E. and Kim, K.-W. (2008), Metastatic tumor antigen 1 is closely associated with frequent postoperative recurrence and poor survival in patients with hepatocellular carcinoma. Hepatology, 47: 929–936. doi: 10.1002/hep.22124
Potential conflict of interest: Nothing to report.
- Issue published online: 27 FEB 2008
- Article first published online: 27 FEB 2008
- Manuscript Accepted: 28 OCT 2007
- Manuscript Received: 27 AUG 2007
Metastatic tumor antigen 1 (MTA1) is known to play a role in angiogenic processes as a stabilizer of hypoxia-inducible factor 1-α (HIF1-α). In this study, we examined whether overexpression of MTA1 affects the recurrence of hepatocellular carcinoma (HCC) after surgical resection and the survival of the patients. A total of 506 HCC patients who underwent hepatic resection were included in the study. They were followed up for a median of 43 months (range, 1-96 months) after hepatectomy. MTA1 expression levels were determined by the proportion of immunopositive cells (none, all negative; +, <50%; ++, >50%). The relationships between MTA1 expression and the HCC histological features, the appearance of recurrent HCC after surgical resection, and the survival of the patients were examined. Eighty-eight cases (17%) of the HCCs were positive for MTA1, although the surrounding liver tissues were all negative for MTA1; 62 cases were + and 26 cases were ++. Increased MTA1 expression levels in HCC were correlated with larger tumors (P = 0.04), perinodal extension (P = 0.03), and microvascular invasion (P = 0.008). Histological differentiation had marginal significance (P = 0.056). However, there was no association between MTA1 expression and age, sex, Child-Pugh class, and capsule invasion of HCC. Interestingly, MTA1 expression levels were significantly greater in hepatitis B virus (HBV)-associated HCC compared with hepatitis C virus (HCV)-associated HCC (P = 0.017). The cumulative recurrence rates of MTA1-positive HCCs were markedly greater than those of MTA1-negative HCCs (P < 0.0001). The cumulative survival rates of patients with MTA1-positive HCCs were significantly shorter than those of patients with MTA1-negative HCCs (P < 0.0001). In conclusion, our data indicate that MTA1 is closely associated with microvascular invasion, frequent postoperative recurrence, and poor survival of HCC patients, especially in those with HBV-associated HCC. (HEPATOLOGY 2008;47:929–936.)
Hepatocellular carcinoma (HCC) is a common malignancy that is the 5th most common cancer in the world.1–3 Although surgery is 1 of the best treatment modalities for HCC, fewer than 10%–20% of HCC patients are candidates for surgery because of unresectable size and number of tumors, poor liver function reserve, multiple intrahepatic or distant metastasis, and frequent vessel invasion at diagnosis.4–6 Even in HCC patients who are good candidates for surgery, frequent recurrence after surgery is a major limitation to long-term survival.7–10
Various metastasis-associated proteins of cancers have been studied and isolated. Among them, the recently identified metastatic tumor antigen 1 (MTA1) is known to increase the migration and invasion of various tumor cells in vitro.11–13 MTA1 has also been suggested to play a role in angiogenic processes as a stabilizer of hypoxia-inducible factor 1α (HIF1α).14–16 Thus, it has been reported that MTA1 overexpression is closely correlated with an aggressive course in several human cancers, such as breast, prostate, colorectal, gastric, and esophageal cancers.17–24 However, few data are available regarding the role of MTA1 in invasion, recurrence, and survival in HCC patients.25, 26
In this study, we examined whether the high expression levels of MTA1 in HCC tissues can affect the recurrence of HCC after surgical resection and consequently affect patient survival.
Patients and Methods
A total of 506 patients who underwent hepatic resection for HCC from 1998–2003 at the Asan Medical Center, University of Ulsan, College of Medicine, Seoul, Korea, were included in the study. The clinical characteristics of the 506 patients are listed in Table 1. The patients were followed-up for a median period of 43 months (range, 1–96 months) after hepatectomy. The recurrence and survival were determined using medical records at the last follow-up date. In cases in which a patient was lost to follow-up for more than 3 months, we evaluated the recurrence and survival based on information obtained by visiting the patient's home or a village office near the patient's residence.
|Age (years, mean ± SD)||56 ± 10|
|Disease severity (CH:LC)||138:368|
|Child-Pugh class (A/B/C)||383/73/50|
|Causes of HCC (HBV/HCV/Both/NBNC)||397/29/8/72|
|Size of HCC (cm)*||4 (0.7–21)|
|Follow-up after hepatectomy (months)*||43 (1–96)|
Tissue Microarray Construction.
The construction of tissue microarrays has been described previously14. Formalin-fixed, paraffin-embedded tissue samples were arrayed using a tissue-arraying instrument (Beecher Instruments, Silver Spring, MD). Briefly, representative areas of each tumor were selected and marked on the hematoxylin-eosin (H&E)-stained slide, and its corresponding tissue block was sampled. The designated zone of each donor block was punched with a tissue cylinder of 1 mm diameter, and the sample was transferred to a recipient block. Each sample was arrayed in duplicate to minimize tissue loss.
Immunohistochemical Staining of MTA1 in Human HCC Tissue Microarray.
Immunohistochemical staining for MTA1 was performed using the avidin-biotin-peroxidase complex method with 3, 3′-diaminobenzidine as a chromogen and an LSAB kit (DAKO, Carpinteria, CA). Paraffin-embedded tissue-microarray blocks, which included HCC and surrounding non-neoplastic liver tissues, were sectioned at 5-μm intervals. Slides were deparaffinized with xylene and rehydrated in a series of graded alcohol. The slides were then incubated in 3% hydrogen peroxide for 10 minutes to block endogenous peroxidase activity. To increase the immunoreactivity, antigen retrieval was performed in citrate buffer (pH 6.0) for 10 minutes in a steam oven. The primary antibody against MTA1 (Santa Cruz Biochemistry, Santa Cruz, CA) was used at a dilution of 1:200. Subsequently, secondary biotinylated antibody and avidin–biotin complex reagent were applied and the sections were counterstained with Harris hematoxylin. For a negative control, sections were incubated with Tris-buffered saline containing 2% goat serum and 1% bovine serum albumin instead of primary antibody.
Evaluation of Immunostaining.
We examined a proportion of tumor cells that were positive for MTA1 in the tumor cell nuclei. For each spot, areas of the most intense or predominant staining patterns were scored. Their staining intensity was usually positively correlated with the proportion of positive tumor cells. That is, the staining values appeared to cluster (1) in a group of cases with less than half of weekly positive tumor cells or (2) in a group of cases with more than 90% of strongly positive tumor cells. Based on the findings, we used the criteria to simplify the classification and analyze the clinical data effectively; (1) 0% (none, −); (2) MTA1 low group (less than 50%, +); and (3) MTA1 high group (more than 50%, ++). The nuclear staining was diffuse and there were no other staining patterns such as membranous, nucleolar, or speckled patterns that can be observed in cases of another nuclear proteins. Two independent observers determined MTA1 expression levels using the arrays, and both observers reexamined specimens with discrepant scores to determine a consensus score.
To avoid confusion of analysis and focus on the role of MTA1 as a prognostic factor in HCC recurrence and overall survival, we performed this study using the following standards: First, we considered “only objective evidences on imaging studies” as “HCC recurrence” for cumulative recurrence rate and did not consider “probability of recurrence” as “HCC recurrence,” even in patients who died of presumed progressed liver dysfunction. Second, all deaths of hepatic origin combined with or without progressed liver dysfunction were considered for cumulative survival rate (overall survival). We did not include nonhepatic origin deaths in determining cumulative survival rates. Based on the same criteria, multivariate analyses for recurrence and survival were performed using Cox regression hazard model.
Data were expressed as mean ± standard deviation or median (range). For statistical significance, Student t test and chi-squared test were used for comparisons of variables between groups. The cumulative relapse and survival rates were evaluated by the Kaplan-Meier method, and differences were determined by the log-rank test. A multivariate analysis was carried out to identify the independent predictor for recurrence and survival using the Cox regression hazard model. A value of P < 0.05 was regarded as significant. All analyses were performed using SPSS 13.0 software (SPSS Inc., Chicago, IL).
MTA1 Expression Frequencies in HCC and Surrounding Liver Tissues
Of the 506 HCC samples, MTA1 was stained in 88 (17%), but none of the surrounding liver tissues were stained. Of positive 88 samples, 62 were + and 26 showed ++ expression (P < 0.001) (Fig. 1).
Tumor Size and MTA1 Expression Levels
The level of MTA1 expression was higher in HCC patients with larger tumors. A total of 150 patients had HCC less than 3 cm diameter, and 333 had HCC of more than 3 cm diameter. Of the HCCs of less than 3 cm diameter, MTA1 expression was negative in 87%, + in 11%, and ++ in 2%. Of the HCCs greater than 3 cm in diameter, MTA1 expression was negative in 79%, + in 14%, and ++ in 7%, respectively (P = 0.04) (Fig. 2).
Tumor Type and MTA1 Expression Levels
The tumor types were analyzed in 434 cases. The nodular type was noted in 263 patients, nodular with perinodal extension in 74 patients, multinodular confluent in 70 patients, pedunculated in 6 patients, and diffuse infiltrative in 21 patients. The level of MTA1 expression was lower in the nodular type of HCC than in other types of HCC (perinodal, multinodular confluent, pedunculated, and diffuse infiltrative types) (P = 0.03) (Fig. 3).
Histological Differentiation and MTA1 Expression Levels
A total of 469 HCC samples were available for histological differentiation analysis. Edmondson-Steiner (E-S) grades 1, 2, 3, and 4 were noted in 38, 154, 213, and 64 patients, respectively. MTA1 expression levels in E-S grade 1 were negative in 92%, + in 8%, and ++ in 0%. MTA1 expression levels in E-S grade 2 were negative in 85%, + in 11%, and ++ in 4%. MTA1 expression levels in E-S grade 3 were negative in 80%, + in 13%, and ++ in 7%. MTA1 expression levels in E-S grade 4 were negative in 72%, + in 20%, and ++ in 7%. Increased MTA1 expression levels tended to be associated with worse histological differentiation of HCC (P = 0.056) (Fig. 4).
Microvascular Emboli and MTA1 Expression Levels
We examined microvascular emboli in 452 HCC frozen tissues. Among them, 102 (22.6%) showed microvascular emboli. MTA1 expression levels in patients with microvascular emboli on frozen tissues were much greater than those of patients without microvascular emboli (negative in 84%, + in 12%, and ++ in 4%, versus negative in 72%, + in 18%, and ++ in 10%, respectively; P = 0.008) (Fig. 5).
The Causes of HCC and MTA1 Expression Levels
Different causes of liver disease were associated with differences in MTA1 expression levels. Of the 484 patients, 380 had hepatitis B virus (HBV), 27 had hepatitis C virus (HCV), and 8 had both HBV and HCV. A total of 69 patients had a non-viral hepatitis cause of HCC. Interestingly, MTA1 was expressed in 80 of 380 patients with HBV-associated HCC (21.1%), but in only 1 patient (3.4%) with HCV-associated HCC (P = 0.017). The levels of MTA1 expression in HBV-related HCC were negative in 79%, + in 15%, and ++ in 6%. MTA1 expression levels in HCV-associated HCC were negative in 96%, + in 0%, and ++ in 4% (Fig. 6). None of HCCs from the patients with HBV and HCV coinfection had MTA1 expression (0%, 0/8). Among 72 HCC patients with nonviral causes, 69 patients were available for evaluation of MTA1 staining. Of 69 patients, 7% (5/69) had level 1 (+) expression and 3% (2/69) level 2 (++) expression of MTA1.
Correlation of Other Clinicopathological Factors and MTA1 Expression Levels
There was no association between MTA1 expression levels and age, sex, Child-Pugh class of liver disease, decompensation of liver function, or capsule invasion of HCC (Table 2).
|Age (years)*||57 (4–88)||58 (34–76)||57 (29–71)||NS|
|Capsule invasion (%)||21||29||22||NS|
Recurrence and Survival Rates According to MTA1 Expression Levels
The 1-year, 3-year, and 5-year cumulative recurrence rates of MTA1-positive HCC were much higher than those of MTA1-negative HCC (45%, 61%, and 73% versus 23%, 40%, and 51%, respectively; P < 0.001). The cumulative recurrence rates in patients with high MTA1 expression levels (++) at 1, 3, and 5 years were 41%, 72%, and 93%, respectively, which were much higher than those in patients with + (39%, 54%, and 65%, respectively) and negative MTA1 expression (25%, 39%, and 51%, respectively) levels (P < 0.001) (Fig. 7).
The 1-year, 3-year, and 5-year cumulative survival rates of patients with MTA1-positive HCC were significantly shorter than those of patients with MTA1-negative HCC (71%, 54%, and 44% versus 89%, 72%, and 61%, respectively; P < 0.001). The cumulative survival rates in patients not showing MTA1 expression were 89%, 72%, and 61% at 1, 3, and 5 years, respectively. However, in patients with + MTA1 expression, the cumulative survival rates at 1, 3, and 5 years were 77%, 61%, and 51%, respectively, and in patients with ++ MTA1 expression, the cumulative survival rates were 54%, 39%, and 24%, respectively (P < 0.001) (Fig. 8). On subgroup analysis for the patients with HBV infection, MTA1 expression levels were still significantly correlated with the cumulative recurrence rates (Fig. 9) and the cumulative survival rates (Fig. 10).
MTA1, tumor size (>3.0 cm), histological differentiation (E-S III/IV), non-nodular tumor type, capsule invasion, portal vein thrombosis, and microvascular invasion were potential candidates for multivariate analysis of recurrence and survival (P < 0.1 on univariate analysis). On multivariate analysis, positive MTA1 staining (especially ++), a larger tumor size (>3 cm in diameter), portal vein thrombosis, and microvascular invasion were independent prognostic factors for postoperative recurrence and survival (Table 3).
|OR, 95% CI||P value||OR, 95% CI||P value|
|MTA1 (+)||1.510, 1.008–2.262||0.045||1.668, 1.092–2.549||0.018|
|MTA1 (++)||3.248, 1.871–5.638||<0.001||2.532, 1.383–4.634||0.003|
|Tumor size (>3 cm)||1.807, 1.230–2.655||0.003||2.336, 1.432–3.813||0.001|
|PVT||2.220, 1.454–3.390||<0.001||1.752, 1.110–2.763||0.016|
|Capsular invasion||–||NS||1.421, 0.970–2.082||0.072|
|E-S grade (III/IV)||–||NS||–||NS|
|MVI||1.719, 1.212–2.438||0.002||1.550, 1.061–2.266||0.023|
Extrahepatic Metastasis According to MTA1 Expression Levels
Of 446 patients who undertook an evaluation about HCC at the last follow-up, 48% (213/446) had no recurrence. Single and multiple intrahepatic metastases were in 21% (95/446), and 17% (77/446), respectively. Extrahepatic metastasis occurred in 14% (61/446). Interestingly, incidence of positive MTA1 staining (+ or ++) in each group was 12% (26/213), 19% (18/95), 23% (18/77), and 31% (19/61), respectively (P = 0.004). Regarding extrahepatic metastasis, it occurred more frequently in the MTA1-positive group (+ or ++) than in the negative group [23% (19/81) versus 12% (42/365), P = 0.005].
HCC is one of typical hypervascular tumors, and the treatment of it is difficult because of frequent distant metastasis at diagnosis and recurrence after primary treatment. MTA1, which is an 80-kDa protein component of the nuclear remodeling and deacetylation complex histone deacetylase, was recently identified as a metastasis-associated protein.27–30 MTA1 plays an important role in the adverse course of various human cancers, including breast, prostate, lung, colorectal, gastric, and esophageal cancers, through migration and invasion of cancer cells toward surrounding and distant tissues.17–24 In HCC, only a small number of studies have examined the role of MTA1 in vascular invasion and survival after treatment. Moon et al.25 reported that high MTA1 expression levels were correlated with large tumors and vascular invasion, based on analysis of paraffin sections from 45 HCC specimens. Hamatsu et al.26 showed that high MTA1 messenger RNA expression levels were more often seen in poorly differentiated HCC (83%) than in those of well or moderately differentiated HCC (33%); however, these authors found no association between MTA1 gene expression levels and cancer invasion to the portal vein or intrahepatic metastasis. Disease-free survival rates after hepatectomy were significantly lower in patients with high MTA1 gene expression levels, and the MTA1 gene might be more closely related to the poorer differentiation and intrahepatic metastasis of HCC.26 Therefore, the authors suggested that high MTA1 gene expression levels might be a prognostic indicator after curative hepatectomy for HCC. In the current study, the MTA1 expression levels in HCC tissues were greater in patients with large tumors, worse histological differentiation, and microvascular emboli. However, there was no association between MTA1 expression level and age, sex, Child-Pugh class of underlying liver cirrhosis, decompensation of liver function, or capsule invasion of HCC. These results are in agreement with those of previous reports.25, 26
In our study, patients with higher MTA1 expression levels showed high recurrence rates and shorter survival times after curative treatment. The 1-year, 3-year, and 5-year cumulative recurrence rates after curative hepatectomy were 45%, 61%, and 73%, respectively, in MTA1-positive HCC, and 23%, 40%, and 51%, respectively, in MTA1-negative HCC (P < 0.001). The 1-year, 3-year, and 5-year cumulative survival rates were 71%, 54%, and 44%, respectively, in MTA1-positive HCC, and 89%, 72%, and 61%, respectively, in MTA1-negative HCC (P < 0.001). Larger tumors and tumors with worse histological differentiation have the potential to invade microvessels and metastasize to distant sites from primer tumors, thereby leading to shorter disease-free survival despite curative surgery. MTA1 is thought to be an important factor for tumor cell migration, invasion, and metastasis, and the MTA1 expression levels were much higher in HCC patients with large tumors and poor differentiation.
MTA1 expression is related to tumor invasiveness, but the mechanism has not yet been determined. In this study, MTA1 was expressed in 80 of 380 patients with HBV-associated HCC (21.1%), but in only 1 patient (3.4%) with HCV-associated HCC (P = 0.017). This result indicates that MTA1 is an important factor for invasion and metastasis in HBV-associated HCCs. Moon et al.25 reported that hepatitis B virus X protein (HBx) induces angiogenesis by stabilizing HIF-1α. They showed that HBx increases the transcriptional activity and protein level of HIF-1α under both normoxic and hypoxic conditions by reducing the binding of von Hippel-Lindau protein to HIF-1α and preventing ubiquitin-dependent degradation of HIF-1α. These authors also reported that HBx, HIF-1α. and vascular endothelial growth factor were more strongly detected in dysplastic lesions than in non-neoplastic lesions in HBx-transgenic mice.31 MTA1 also increases angiogenesis by stabilizing the HIF-1α protein. MTA1 overexpression increased the transcriptional activity and stability of HIF-1α protein. MTA1 induces the deacetylation of HIF-1α by increasing the expression level of histone deacetylase 1.14–16 These findings indicate a close correlation between MTA1-associated metastasis and HIF-1α–induced tumor angiogenesis. Although MTA1 has been reported to be associated with tumor cell migration, metastasis, and invasiveness in vitro, it is still inconclusive in human.11–13 In this study, we demonstrated that MTA1 overexpression in HCC may contribute to the extrahepatic metastasis as well as vascular invasion. Taken together, it suggests that MTA1 may be in the middle of mechanisms associated with distant metastasis through vascular invasion.
According to other reports, MTA1 is an essential downstream effector of the c-MYC oncoprotein.32 HBx not only potentiates c-MYC–induced liver oncogenesis in transgenic mice but also inhibits the ubiquitination and proteasomal degradation of c-MYC.33, 34 HBx also promotes hepatocellular carcinoma metastasis by facilitating tumor-cell invasion and subsequent destruction of the extracellular matrix by upregulation of matrix metalloproteinases.35 Collectively, both MTA1 and HBx induce angiogenesis by stabilizing HIF-1α and increasing c-MYC oncogene function. Our data, which show almost no MTA1 expression in HCV-associated HCCs compared with HBV-associated HCCs, indicate that there might be a relationship between MTA1 and HBx. However, we did not investigate this possible correlation between MTA1 and HBx experimentally in this study.
In conclusion, as in some other human cancers, MTA1 expression is closely associated with larger tumor size, worse histological differentiation, microvascular invasion, frequent postoperative recurrence, and poor patient survival, especially in HBV-associated HCCs. Therefore, the expression levels of MTA1 in HCC tissues might be an important prognostic marker after curative surgery.