Overexpression of DEK gene is correlated with poor prognosis in hepatocellular carcinoma

  • Authors:
    • Huo‑Chun Yi
    • Ya‑Li Liu
    • Pan You
    • Jin‑Shui Pan
    • Jian‑Yan Zhou
    • Zhen‑Jin Liu
    • Zhong-Ying Zhang
  • View Affiliations

  • Published online on: October 27, 2014     https://doi.org/10.3892/mmr.2014.2781
  • Pages: 1318-1323
Metrics: Total Views: 0 (Spandidos Publications: | PMC Statistics: )
Total PDF Downloads: 0 (Spandidos Publications: | PMC Statistics: )


Abstract

The oncogene DEK was originally identified as one of the parts of the DEK‑CAN fusion gene, arising from the translocation (6;9) in a subtype of acute myeloid leukemia. Since then, DEK has been shown to promote tumorigenesis in a variety of cancer cell types through its roles in inhibiting cell differentiation, senescence and apoptosis. Certain studies have established that DEK is dysregulated in several types of cancer, including hepatocellular carcinoma (HCC). However, its clinical significance in human HCC remains unknown. In this study, the expression of DEK mRNA and protein was examined in 55 surgical HCC specimens and matched non‑tumorous tissues. In addition, the correlation between DEK expression and clinicopathological characteristics and prognosis was analyzed. mRNA and protein levels of DEK were found to be significantly overexpressed in the majority of HCC tumors when compared with matched normal hepatic tissues (P<0.05). In addition, the expression pattern of DEK was closely correlated with differentiation status, portal venous invasion and tumor size (P<0.05). Kaplan‑Meier curves demonstrated that patients with higher DEK expression levels had significantly poorer survival than those with lower DEK expression levels (P=0.003). In addition, Cox regression analysis demonstrated that the level of DEK expression may be a valuable prognostic factor (P<0.05). These results suggested that DEK may play a significant role in hepatocyte differentiation and may serve as a useful prognostic marker and biomarker for the staging of HCC.

Introduction

Hepatocellular carcinoma (HCC) is one of the most common malignancies worldwide. HCC incidence continues to increase, ranking as the second leading cause of cancer-related mortality among males in China (1). Similar to other cancer types, HCC is characterized by an evident multistage process of tumor progression (2). In spite of the developments in surgical treatment strategies and options made available in recent years, the overall prognosis of HCC patients remains extremely poor, and this is largely due to the high frequency of recurrence or metastasis following surgery (3). Thus, the capacity to subjectively predict the risk of recurrence and subsequent prognosis is essential to guide surgery and chemotherapy.

The oncogene DEK was originally identified as one of the parts of the DEK-CAN fusion gene, arising from the translocation (6;9) in a subtype of acute myeloid leukemia patients (4). The overexpression of DEK has been observed in several malignancies including melanoma, bladder cancer, glioblastoma, retinoblastoma and HCC (59). The increasing list of tumor types demonstrating high and easily detectable DEK protein expression indicates the potential of using DEK as a tumor marker (9). A previous study has characterized DEK as a potential urinary and tissue biomarker for transitional cell carcinoma of the bladder of various grades, stages and progression (10). Furthermore, since the amount of DEK expressed in immature cells is greater than that in differentiated cells, it could assist in assessing the differentiation potential of tumor cells (11). Previous studies have identified that DEK overexpression may be linked with the progression of breast cancer, and that DEK may possibly be used not only as a therapeutic target, but also as a breast cancer biomarker for the early diagnosis and prognostic evaluation of breast cancer (1214).

While DEK overexpression has been observed in HCC, little is known about its clinical significance in this context (6,1516). In the present study, we examined the expression level of DEK mRNA and protein in HCC surgical specimens and matched normal hepatic tissues, and then analyzed the correlation between DEK expression and clinicopathological characteristics and patient prognosis.

Materials and methods

Patients and specimens

Fifty-five pairs of HCC tissues and their adjacent non-cancerous tissues were obtained from patients who had undergone surgical resection at Xiamen Zhongshan Hospital affiliated to Xiamen University, China, between July 2007 and November 2010. The histological diagnoses and tumor grades of all samples were confirmed by an experienced pathologist. The tumor grading was determined based on the Edmondson-Steiner classification (17). Patient characteristics collected for analysis included age, gender, tumor size, Edmondson-Steiner grade, hepatitis history, number of tumor nodules and presence of portal venous invasion. None of the patients engaged in this study received chemotherapy or radiotherapy prior to surgery. Follow-up data was obtained following hepatic resection for all 55 patients. The follow-up period was defined as the interval between the date of the surgery and that of the patient’s mortality or the last follow-up. Mortality from other causes was treated as censored cases. This study was authorized by the Research Ethics Committee of Xiamen Zhongshan Hospital affiliated to Xiamen University. Informed consent was obtained from all of the patients. The tissues were collected according to local ethical guidelines and approved beforehand by all participants and the Human Investigation Committee of The Medical College of Xiamen University (Xiamen, China).

All samples were managed anonymously on the basis of ethical and legal standards

For RNA preparation, the resected samples were placed in liquid nitrogen as soon as possible following resection, and stored at −80°C. The remaining tissues were used for routine histopathological examination.

Isolation of total RNA and first strand cDNA synthesis

Total RNA was isolated from frozen tumor and matched normal tissues using TRIzol reagent (Invitrogen Life Technologies, Carlsbad, CA, USA) according to the manufacturer’s instructions. RNA quality was confirmed by agarose gel electrophoresis. Only RNA without DNA contamination and biodegradable 26S rRNA was prepared for succeeding cDNA synthesis (Fig. 1A). Following photometric quantification, 2 μg total RNA was used for a 20 μl reverse transcription (RT) reaction containing 4 μl 5× first strand buffer (Promega Corporation, Madison, WI, USA), 2 μl 0.1 M DTT, 1 μl oligo(dT)15 primer (10 mM), 1 μl dNTPs (10 mM) and diethylpyrocarbonate-treated water. The reaction mixture was incubated for 1 h at 42°C, heated for 5 min at 95°C and placed for 5 min in an ice bath; the first strand of cDNA was either stored at 4°C or used immediately for polymerase chain reaction (PCR).

PCR

Primer sequences were as follows: 5′-TCTGTGAGGTTCTTGATTTGGA-3′ (forward) and 5′-CTGTTCCGTTCCTTTTTACTGC-3′ (reverse) for DEK; 5′-ACCTGACCTGCCGTCTAGAA-3′ (forward) and 5′-TCCACCACCCTGTTGCTGTA-3′ (reverse) for GAPDH. PCR analysis was performed under the following conditions: 4 min at 94°C, 30 cycles of denaturation for 20 sec at 94°C, annealing for 30 sec at 60°C, extension for 40 sec at 72°C, and finally 5 min at 72°C. The amplification products were analyzed on a 1.5% agarose gel. The intensity ratio was the relative expression of DEK normalized to that of GAPDH.

Immunohistochemical staining

A total of 55 HCC tissues and matched normal liver tissues were evaluated by immunohistochemistry. Multiple 5-μm sections were prepared from formalin-fixed, paraffin-embedded tissue blocks. Sections were placed overnight at 4°C with mouse anti-human DEK monoclonal antibody (1:100, (Proteintech Group, Inc., Chicago, IL, USA). The streptavidin-biotin peroxidase complex tertiary system (Boster Bio, Pleasanton, CA, USA) was applied in accordance with the manufacturer’s specifications. Sections were counter-stained using hematoxylin, dehydrated via gradient alcohol and fixed for observation. All sections were examined independently by two pathologists who were unaware of the patients’ clinical status. With regard to the percentage of DEK-positive hepatocytes, immunohistochemical staining was scored as follows: 0, <5% positive; 1+, 5–25% positive; 2+, 26–50% positive; and 3+, >50% positive. Only nuclear expression was considered to be positive staining. Tissues with moderate to strong nuclear staining were designated as the DEK high expression group (scores 2+ and 3+). Tissues designated as the DEK low expression group were either devoid of any nucleus staining or scored as 0 or 1+.

Statistical analysis

SPSS version 16.0 for Windows (SPSS Inc., Chicago, IL, USA) was used for all data analyses. The paired-samples t-test was used to compare the intensity ratio between HCC tissues and matching non-cancerous livers. Pearson’s χ2 test was used to assess differences in the rate of positive immunostaining between hepatic cancer tissues and adjacent tissues. An independent Student’s t-test and Pearson’s χ2 test were used to evaluate the correlations between DEK mRNA and protein expression and clinicopathological variables of HCCs, respectively. The Kaplan-Meier method was employed to analyze patient survival, and the differences in survival were evaluated using the log-rank test. The Cox proportional-hazards model was used to confirm factors independently associated with survival. All P-values were based on two-sided statistical analyses, and P<0.05 was considered to indicate a statistically significant difference.

Results

Significantly increased DEK mRNA and protein expression in HCC tissues

The expression of DEK mRNA was analyzed in 55 HCC samples and matched normal hepatic tissues. Higher expression levels of DEK mRNA were observed in 50 of the HCC tissue samples compared with the matched non-tumor hepatic tissues. DEK mRNA levels in HCC tissues were significantly higher than those in corresponding non-tumorous livers (0.707±0.157 versus 0.391±0.116; t=18.3, P<0.001; Fig. 1B and C). By immunohistochemical analysis, the percentage of positive DEK expression in HCC tissues was significantly higher than that in corresponding normal livers [87.3% (48 of 55) versus 40.0% (22 of 55); P=0.002; Fig. 2C]. Moreover, the DEK protein levels in HCC tissues were significantly higher than those in the corresponding non-tumorous livers (P<0.01, Fig. 2A and B). Furthermore, an increase in nuclear expression of DEK was observed in HCC tumor tissues, both in intensity and in the positive proportion.

On the basis of the immunohistochemistry data, the 55 HCC cases were divided into the low expression group (score 0 or 1+; n=29) and the high expression group (score 2+ or 3+; n=26). The DEK mRNA levels in the DEK high expression group were significantly higher than those in the low expression group (0.841±0.073 versus 0.587±0.106; P<0.001; Fig. 2D).

Taken together, these data confirm the overexpression of DEK at the transcription and translation level in human HCC.

Correlation between DEK expression and clinicopathological characteristics

The immunohistochemistry data were analyzed for the correlation of DEK protein expression with clinicopathological features. The expression of DEK protein showed a positive correlation with tumor size (P=0.001), Edmondson-Steiner grade (P=0.025) and portal venous invasion (P=0.002). However, no significant correlation was observed between DEK protein expression and other clinical characteristics, including age (P=0.143), gender (P=0.385), hepatitis history (P=0.589), number of tumor nodules (P=0.795) and liver cirrhosis (P=0.418) (Table I). The mRNA levels were compared with the clinical data and found to be associated with the same clinical variables as observed in the protein level analysis.

Table I

Correlations between DEK protein expression and clinicopathological variables in 55 cases of hepatocellular carcinoma.

Table I

Correlations between DEK protein expression and clinicopathological variables in 55 cases of hepatocellular carcinoma.

DEK expression

Clinicopathological featuresnLowHighχ2P-value
Gender
 Male4324190.7530.385
 Female1257
Age (years)
 <502611152.1480.143
 ≥50291811
Tumor size (cm)
 ≤52318510.340.001a
 >5321121
Hepatitis history
 Yes4424200.2920.589
 No1156
Edmondson-Steiner grade
 1 and 2141135.0330.025a
 3 and 4411823
Portal venous invasion
 Absent151329.5310.002a
 Present401624
Tumor nodule no.
 Solitary3921180.0670.795
 Multiple (≥2)1688
Liver cirrhosis
 Absent11740.6560.418
 Present442222

{ label (or @symbol) needed for fn[@id='tfn1-mmr-11-02-1318'] } Expression of DEK protein was determined in 55 hepatocellular (HCC) samples by immunohistochemical analysis as described in Materials and methods. Based on the immunohistochemistry data, the 55 cases were divided into a low expression group (score 0 or 1+; n=29) and a high expression group (score 2+ or 3+; n=26). The correlations between the expression of DEK protein and clinicopathological variables of HCC were evaluated by Pearson’s χ2 test.

a P<0.05.

Follow-up and prognostic value of DEK

The Kaplan-Meier method was used to analyze the correlation of DEK expression level with the prognosis of HCC patients. The results indicated that the DEK high expression group had a shorter median survival time than the low expression group (23 versus 39 months), and the overall survival rates for the low and high expression groups were significantly different (P=0.003, log-rank test; Fig. 3).

To elucidate factors that may predict survival following hepatic resection, univariate and multivariate Cox regression analyses were applied. In the univariate analysis, DEK expression level [hazard risk (HR), 2.273; P=0.022), Edmondson-Steiner grade (HR, 1.542; P=0.039) and portal venous invasion (HR, 3.145; P=0.015) were all significantly associated with survival. However, age, gender, tumor size, hepatitis history, number of tumor nodules and liver cirrhosis were not significantly correlated with survival (Table II). In the multivariate analysis, DEK expression level (HR, 2.974; P=0.017), Edmondson-Steiner grade (HR, 2.065; P=0.026) and portal venous invasion (HR, 1.967; P=0.028) were identified to be independent prognostic factors of survival (Table II).

Table II

Multivariate analysis using a Cox proportional-hazards regression model for hepatocellular carcinoma patients.

Table II

Multivariate analysis using a Cox proportional-hazards regression model for hepatocellular carcinoma patients.

Univariate analysisMultivariate analysis


Clinicopathological variablesnHR (95% CI)P-valueHR (95% CI)P-value
Edmondson-Steiner grade0.0390.026
 1 and 21411
 3 and 4411.542 (1.026–2.746)2.065 (1.185–2.934)
Portal venous invasion0.0150.028
 Absent1511
 Present403.145 (1.943–5.627)1.967 (1.173–2.836)
DEK0.0220.017
 Low expression2911
 High expression262.273 (1.474–3.643)2.974 (1.725–4.139)

[i] Insignificant variables with P>0.05 are not listed in the table, including age, gender, tumor size, hepatitis history, number of tumor nodules and liver cirrhosis. HR, hazard risk; CI, confidence interval.

Discussion

DEK was originally identified in the fusion protein DEK-CAN, resulting from the recurrent t(6;9) translocation in a subset of acute myeloid leukemia patients (4). In addition, previous studies have suggested that DEK is ubiquitously expressed in the majority of mammalian cells (1819). Subsequently, DEK was reported to be frequently upregulated in aggressive human tumors including melanoma, glioblastoma, retinoblastoma, bladder cancer and HCC (58,1516). DEK has been proven to boost tumorigenesis in numerous cancer cell types through its function of intervening with cell division; inhibiting cell differentiation, senescence and apoptosis (2022).

In the present study, we measured DEK mRNA and protein expression levels in 55 HCC tissue samples paired with adjacent non-cancerous tissues. The study revealed that levels of DEK mRNA were significantly higher in HCC tissues than in the matching non-cancerous tissues (P<0.001). Although the present study had a higher number of clinical cases, the result was consistent with that from a previous HCC study, which demonstrated that DEK mRNA levels were higher in 4 of 5 primary HCCs compared with matched non-tumorous liver tissues (13). Lü et al also demonstrated that DEK mRNA levels in HCC tissues were higher than those in paraneoplastic tissues (6).

Furthermore, DEK protein expression was analyzed by immunohistochemistry in the same 55 paired specimens. The percentage of cells with positive DEK expression in HCC tissues was significantly higher than that in corresponding non-tumorous livers (87.3 versus 40.0%; P=0.002), comparable with the results from a previous study of DEK protein expression in HCC samples (14). The immunohistochemistry results were in agreement with the mRNA analyses in this study. An increase in nuclear expression of DEK was observed in HCC tumor tissues, both in intensity and in the positive percentage of cells.

In the current study, we divided the HCC cases into low (n=29) and high (n=26) expression groups based on immunohistochemistry scores. DEK mRNA levels in the high expression group were significantly higher than those in the low expression group (P<0.001). Taken together, these results suggest that overexpression of DEK may be a common event in HCC tumorigenesis.

In this study, the DEK expression data obtained from RT-PCR and immunohistochemistry were analyzed for correlation with clinicopathological features. The results revealed that the expression levels of DEK mRNA and protein were correlated with pathological grade, tumor size and portal venous invasion. Significant differences in the expression levels of DEK mRNA and protein were observed when comparing grade 1 and 2 HCC with grade 3 and 4 HCC. These results are in agreement with a previous HCC study which reported that the level of DEK mRNA was correlated with the histological grade (15). However, Lü et al reported that the difference in expression of DEK mRNA between well- and poorly differentiated HCC was not statistically significant (P>0.05) (6).

In the present study, the results demonstrated that increased expression of DEK protein was significantly correlated with poor patient outcomes. However, age, gender, tumor size, hepatitis history, number of tumor nodules and liver cirrhosis had no effect on overall survival (P>0.05), whereas DEK expression, Edmondson-Steiner grade and portal venous invasion were significant predictors of overall survival (P<0.05). The association of DEK expression with overall survival was consistent with a previous study which demonstrated that increased DEK expression was significantly correlated with poor survival in breast cancer patients (12).

A number of tumorigenic functions of DEK may contribute to the inverse correlation between DEK expression and survival of HCC patients. First, DEK is involved in the inhibition of differentiation and facilitation of cellular transformation (20). Second, DEK inhibits cell apoptosis through its interference with p53 functions (24). Lastly, DEK overexpression has been reported to extend cellular life span, supporting the role of DEK as a senescence inhibitor (25).

In summary, overexpression of DEK in human HCC is significantly correlated with the prognosis and differentiation potential of HCC, suggesting that DEK may serve as a useful prognostic marker. Further studies should be carried out to investigate the precise function and molecular mechanism of DEK in the progression of HCC.

Acknowledgements

The authors sincerely thank the patients and their families for their cooperation and participation in the study, as well as the staff involved with the study for data collection and management.

References

1 

Murray CJ and Lopez AD: Mortality by cause for eight regions of the world: Global Burden of Disease Study. Lancet. 349:1269–1276. 1997. View Article : Google Scholar : PubMed/NCBI

2 

Sakamoto M: Pathology of early hepatocellular carcinoma. Hepatol Res. 37(Suppl 2): S135–S138. 2007. View Article : Google Scholar : PubMed/NCBI

3 

Tung-Ping Poon R, Fan ST and Wong J: Risk factors, prevention, and management of postoperative recurrence after resection of hepatocellular carcinoma. Ann Surg. 232:10–24. 2000. View Article : Google Scholar : PubMed/NCBI

4 

von Lindern M, Fornerod M, van Baal S, et al: The translocation (6;9), associated with a specific subtype of acute myeloid leukemia, results in the fusion of two genes, dek and can, and the expression of a chimeric, leukemia-specific dek-can mRNA. Mol Cell Biol. 12:1687–1697. 1992.PubMed/NCBI

5 

Carro MS, Spiga FM, Quarto M, et al: DEK expression is controlled by E2F and deregulated in diverse tumor types. Cell Cycle. 5:1202–1207. 2006. View Article : Google Scholar : PubMed/NCBI

6 

Lü ZL, Luo DZ and Wen JM: Expression and significance of tumor-related genes in HCC. World J Gastroenterol. 11:3850–3854. 2005.PubMed/NCBI

7 

Kappes F, Khodadoust MS, Yu L, et al: DEK expression in melanocytic lesions. Hum Pathol. 42:932–938. 2011. View Article : Google Scholar : PubMed/NCBI

8 

Khodadoust MS, Verhaegen M, Kappes F, et al: Melanoma proliferation and chemoresistance controlled by the DEK oncogene. Cancer Res. 69:6405–6413. 2009. View Article : Google Scholar : PubMed/NCBI

9 

Zheng J, Kohler ME, Chen Q, et al: Serum from mice immunized in the context of Treg inhibition identifies DEK as a neuroblastoma tumor antigen. BMC Immunol. 8:42007. View Article : Google Scholar : PubMed/NCBI

10 

Datta A, Adelson ME, Mogilevkin Y, et al: Oncoprotein DEK as a tissue and urinary biomarker for bladder cancer. BMC Cancer. 11:2342011. View Article : Google Scholar : PubMed/NCBI

11 

Wise-Draper TM, Morreale RJ, Morris TA, et al: DEK proto-oncogene expression interferes with the normal epithelial differentiation program. Am J Pathol. 174:71–81. 2009. View Article : Google Scholar :

12 

Liu S, Wang X, Sun F, et al: DEK overexpression is correlated with the clinical features of breast cancer. Pathol Int. 62:176–181. 2012. View Article : Google Scholar : PubMed/NCBI

13 

Abba MC, Sun H, Hawkins KA, et al: Breast cancer molecular signatures as determined by SAGE: correlation with lymph node status. Mol Cancer Res. 5:881–890. 2007. View Article : Google Scholar : PubMed/NCBI

14 

Privette Vinnedge LM, McClaine R, Wagh PK, et al: The human DEK oncogene stimulates β-catenin signaling, invasion and mammosphere formation in breast cancer. Oncogene. 30:2741–2752. 2011. View Article : Google Scholar : PubMed/NCBI

15 

Kondoh N, Wakatsuki T, Ryo A, et al: Identification and characterization of genes associated with human hepatocellular carcinogenesis. Cancer Res. 59:4990–4996. 1999.PubMed/NCBI

16 

Lin LJ and Chen LT: The role of DEK protein in hepatocellular carcinoma for progression and prognosis. Pak J Med Sci. 29:778–782. 2013. View Article : Google Scholar : PubMed/NCBI

17 

Edmondson H and Steiner P: Primary carcinoma of the liver: a study of 100 cases among 48,900 necropsies. Cancer. 7:462–503. 1954. View Article : Google Scholar : PubMed/NCBI

18 

Sitwala KV, Adams K and Markovitz DM: YY1 and NF-Y binding sites regulate the transcriptional activity of the dek and dek-can promoter. Oncogene. 21:8862–8870. 2002. View Article : Google Scholar : PubMed/NCBI

19 

Hu HG, Scholten I, Gruss C and Knippers R: The distribution of the DEK protein in mammalian chromatin. Biochem Biophys Res Commun. 358:1008–1014. 2007. View Article : Google Scholar : PubMed/NCBI

20 

Grasemann C, Gratias S, Stephan H, et al: Gains and overexpression identify DEK and E2F3 as targets of chromosome 6p gains in retinoblastoma. Oncogene. 24:6441–6449. 2005.PubMed/NCBI

21 

Evans AJ, Gallie BL, Jewett MA, et al: Defining a 0.5-mb region of genomic gain on chromosome 6p22 in bladder cancer by quantitative-multiplex polymerase chain reaction. Am J Pathol. 164:285–293. 2004. View Article : Google Scholar

22 

Castellano G, Torrisi E, Ligresti G, et al: The involvement of the transcription factor Yin Yang 1 in cancer development and progression. Cell Cycle. 8:1367–1372. 2009. View Article : Google Scholar : PubMed/NCBI

23 

Wise-Draper TM, Mintz-Cole RA, Morris TA, et al: Overexpression of the cellular DEK protein promotes epithelial transformation in vitro and in vivo. Cancer Res. 69:1792–1799. 2009. View Article : Google Scholar : PubMed/NCBI

24 

Wise-Draper TM, Allen HV, Jones EE, et al: Apoptosis inhibition by the human DEK oncoprotein involves interference with p53 functions. Mol Cell Biol. 26:7506–7519. 2006. View Article : Google Scholar : PubMed/NCBI

25 

Wise-Draper TM, Allen HV, Thobe MN, et al: The human DEK proto-oncogene is a senescence inhibitor and an upregulated target of high-risk human papillomavirus E7. J Virol. 79:14309–14317. 2005. View Article : Google Scholar : PubMed/NCBI

Related Articles

Journal Cover

February-2015
Volume 11 Issue 2

Print ISSN: 1791-2997
Online ISSN:1791-3004

Sign up for eToc alerts

Recommend to Library

Copy and paste a formatted citation
x
Spandidos Publications style
Yi HC, Liu YL, You P, Pan JS, Zhou JY, Liu ZJ and Zhang Z: Overexpression of DEK gene is correlated with poor prognosis in hepatocellular carcinoma. Mol Med Rep 11: 1318-1323, 2015.
APA
Yi, H., Liu, Y., You, P., Pan, J., Zhou, J., Liu, Z., & Zhang, Z. (2015). Overexpression of DEK gene is correlated with poor prognosis in hepatocellular carcinoma. Molecular Medicine Reports, 11, 1318-1323. https://doi.org/10.3892/mmr.2014.2781
MLA
Yi, H., Liu, Y., You, P., Pan, J., Zhou, J., Liu, Z., Zhang, Z."Overexpression of DEK gene is correlated with poor prognosis in hepatocellular carcinoma". Molecular Medicine Reports 11.2 (2015): 1318-1323.
Chicago
Yi, H., Liu, Y., You, P., Pan, J., Zhou, J., Liu, Z., Zhang, Z."Overexpression of DEK gene is correlated with poor prognosis in hepatocellular carcinoma". Molecular Medicine Reports 11, no. 2 (2015): 1318-1323. https://doi.org/10.3892/mmr.2014.2781