Clinical significance of the ubiquitin ligase UBE3C in hepatocellular carcinoma revealed by exome sequencing

Authors

  • Jia-Hao Jiang,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion, Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Yan-Feng Liu,

    1. Key Laboratory of Medical Molecular Virology (MOE & MOH), Institute of Biomedical Sciences, Fudan University, Shanghai, P.R. China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Ai-Wu Ke,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion, Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Fang-Ming Gu,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion, Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Yao Yu,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion, Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
  • Zhi Dai,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion, Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
  • Qiang Gao,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion, Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
  • Guo-Ming Shi,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion, Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
  • Bo-Yi Liao,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion, Ministry of Education, Shanghai, P.R. China
    Search for more papers by this author
  • You-Hua Xie,

    1. Key Laboratory of Medical Molecular Virology (MOE & MOH), Institute of Biomedical Sciences, Fudan University, Shanghai, P.R. China
    Search for more papers by this author
  • Jia Fan,

    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion, Ministry of Education, Shanghai, P.R. China
    2. Shanghai Key Laboratory of Organ Transplantation, Shanghai, P. R. China
    Search for more papers by this author
  • Xiao-Wu Huang,

    Corresponding author
    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion, Ministry of Education, Shanghai, P.R. China
    2. Shanghai Key Laboratory of Organ Transplantation, Shanghai, P. R. China
    • Address reprint requests to: Jian Zhou or Xiao-Wu Huang, Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion, Ministry of Education, 180 Fenglin Road, Shanghai 200032, P.R. China. E-mail: zhou.jian@zs-hospital.sh.cn or huang.xiaowu@zs-hospital.sh.cn; fax: +86-21-64037181.

    Search for more papers by this author
  • Jian Zhou

    Corresponding author
    1. Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion, Ministry of Education, Shanghai, P.R. China
    2. Shanghai Key Laboratory of Organ Transplantation, Shanghai, P. R. China
    • Address reprint requests to: Jian Zhou or Xiao-Wu Huang, Liver Cancer Institute, Zhongshan Hospital, Fudan University; Key Laboratory of Carcinogenesis and Cancer Invasion, Ministry of Education, 180 Fenglin Road, Shanghai 200032, P.R. China. E-mail: zhou.jian@zs-hospital.sh.cn or huang.xiaowu@zs-hospital.sh.cn; fax: +86-21-64037181.

    Search for more papers by this author

  • Potential conflict of interest: Nothing to report.

  • Supported by the National Science Foundation for Distinguished Young Scholars of China (Grant No. 81225019), the National Natural Science Foundation of China (Grant No. 30872503, 81172277, 81071992, 81172023), and the National Key Sci-Tech Special Project of China (Grant No. 2012ZX10002-016).

Abstract

Virus-induced hepatocarcinogenesis involves a series of histological developmental processes with the stepwise acquisition of several genetic changes that are necessary for the malignant transformation of hepatocytes. Although genetic alterations are known to be involved in the pathogenesis of hepatocellular carcinoma (HCC), little is known about the contributions of specific genes to this process. To gain insight into the genetic alterations involved in the neoplastic evolution from chronic hepatitis B virus infection to dysplastic nodules (DN) to HCC, we captured and sequenced the exomes of four DNA samples: one DN sample, two HCC samples, and one control peripheral blood sample from a single HCC patient. Mutations in the UBE3C gene (encoding ubiquitin ligase E3C) were observed in both tumor tissues. Then we resequenced the UBE3C gene in a cohort of 105 HCC patients and identified mutations in 17 out of a total of 106 (16.0%) HCC patients. The subsequent experiments showed that UBE3C promoted HCC progression by regulating HCC cells epithelial-mesenchymal transition. Clinically, a tissue microarray study of a cohort containing 323 HCC patients revealed that the overexpression of UBE3C in primary HCC tissues correlated with decreased survival (hazard ratio [HR] = 1.657, 95% confidence interval [CI] = 1.220-2.251, P = 0.001) and early tumor recurrence (HR = 1.653, 95% CI = 1.227-2.228, P = 0.001) in postoperative HCC patients. Conclusion: Our findings indicate that UBE3C is a candidate oncogene involved in tumor development and progression and therefore a potential therapeutic target in applicable HCC patients. (Hepatology 2014;59:2216–2227)

Abbreviations
DN

dysplastic nodules

E3

ubiquitin ligase

EMT

epithelial-mesenchymal transition

HCC

hepatocellular carcinoma

HECT

homologous to E6-AP carboxyl terminus

indels

small insertions and deletions

SNVs

single nucleotide variants

TGF-β1

transforming growth factor beta 1

Hepatocellular carcinoma (HCC) remains the second leading cause of cancer-related deaths worldwide, with ∼700,000 cancer deaths and 750,000 new cases reported each year.[1] While the incidence rates of most cancers have declined in recent decades, the HCC incidence has continued to increase among both men and women.[2] Although recent progress in early diagnosis, surveillance programs in high-risk patients, and surgical interventions (liver resection and transplantation) have significantly improved the survival of patients with HCC, the overall survival (OS) of HCC patients remains poor.[3] The mechanism of hepatocarcinogenesis remains unclear, and our understanding of HCC-associated molecular changes is still lacking.[4] Thus, the discovery of novel HCC susceptibility genes will provide new targets for drug development and gene therapy.

Among the various etiologies related to HCC development, chronic hepatitis B viral (HBV) infection remains the most prominent. Virus-induced hepatocarcinogenesis involves an array of histological developmental processes, and the stepwise acquisition of genetic and genomic changes is necessary for the malignant transformation of hepatocytes, which involves hepatocyte damage, regeneration, and the development of regenerative nodules and dysplastic nodules (DN).[5-7] Hepatocytes with an increasing number of genetic hits gradually develop into precancerous lesions (such as DN) and eventually evolve into neoplasms. A DN is a premalignant lesion that displays a large number of genetic variations and abnormal cytological features, and it is known to invade surrounding tissues and occasionally demonstrate metastatic potential.[8, 9]

Although sequencing studies have made progress in the search for molecular changes in HCC,[10-12] no study has sequenced each histological lesion that is formed during the process of hepatocarcinogenesis. To gain insight into the genetic alterations involved in the neoplastic evolution from chronic HBV infection to DN to HCC, we captured and sequenced the exomes from four DNA samples: one DN sample, two HCC samples, and one control peripheral blood sample from the same HCC patient. The candidate mutated gene(s) identified in these samples were further validated in an additional 105 HCC patients. The functional relevance of these candidates was investigated in HCC cell lines, and clinical relevance was demonstrated in a tissue microarray study of a cohort comprising 323 HCC patients. Our goal was to identify frequent somatic alterations that might describe the biological characteristics of HCC and furnish new approaches for diagnosis and treatment.

Materials and Methods

Patients and Samples

A 58-year-old man who underwent surgical resection for liver neoplasm at Zhongshan Hospital at Fudan University (Shanghai, China) was diagnosed with HCC in June 2011 after suffering from chronic HBV infection for 17 years. The patient had one DN (diameter of 1.8 cm) and two tumors (tumor 1 and tumor 2, with diameters of 2 cm and 3 cm, respectively) with complete encapsulation, microvascular invasion, class II differentiation, and no regional lymph node or distant metastasis (Fig. 1). We obtained peripheral blood samples and three tissue samples, including samples from the DN and the two tumors of this patient.

Figure 1.

Expression and mutations of UBE3C in HCC. (A) The expression and somatic mutations of UBE3C in the patient sample. The upper panel shows hematoxylin-eosin (H&E) and UBE3C staining of the normal liver tissue, DN, and tumor tissue samples. The lower panel shows somatic mutations identified in UBE3C, which were confirmed by Sanger sequencing of the DNA samples from lymphocytes, DN, and tumor tissues. The black arrows indicate the somatic mutation sites. N, normal liver tissue (for H&E and UBE3C staining) or lymphocytes (for Sanger sequencing). (B) UBE3C gene sequencing results in the additional 105 HCC samples. Diverse mutations and amino acid changes were observed between the DNA samples obtained from the peripheral blood specimens (N) and the paired HCC tissues (T), as analyzed by Sanger sequencing. The red arrows indicate the mutation sites.

Genomic DNA samples were extracted from freshly frozen DN tissue, tumor tissues, and mononuclear cells (isolated from the peripheral blood by Ficoll gradient centrifugation) from this HCC patient for exome sequencing. DNA samples were also collected from freshly frozen tumor tissues and matching peripheral blood samples of 105 additional patients for validation. All 105 patients in this study underwent curative hepatectomy for primary liver neoplasms between December 2010 and April 2011 in the Liver Cancer Institute, Zhongshan Hospital, Fudan University (Shanghai, China). UBE3C expression was investigated in a tissue microarray containing samples from 323 HCC patients with complete follow-up data. Patient information for the tissue microarray was reported previously[13] and is detailed in Supporting Text S1. HCC was diagnosed histologically for all patients in this study. Written informed consent was obtained from each patient and ethical approval was obtained from the Zhongshan Hospital Research Ethics Committee.

Whole-Exome Capture Sequencing and Bioinformatics Analysis

Whole-exome capture sequencing methods are described in Supporting Text S1. Candidate somatic mutations were identified according to a bioinformatics pipeline, as described in Supporting Figs. S1 and S2. Briefly, single-end reads were aligned to the human reference genome (hg19, NCBI Build 37.1) using the Burrows-Wheeler Aligner (BWA) (v. 0.5.9-r16). Because duplicated reads were generated during the polymerase chain reaction (PCR) amplification process, single-end reads that aligned to the same genomic positions were removed using SAMtools (v. 0.1.18). Single nucleotide variations (SNVs) were identified using GATK (Genome Analysis Toolkit, v. 1.0.5506) obtained from the Broad Institute (http://www.broadinstitute.org/gatk/).

Validation and Screening of Candidate Mutations

PCR amplification and Sanger sequencing were performed following standard protocols using the primers listed in Supporting Table S1 to verify candidate nonsynonymous somatic variants after exome sequencing and to sequence the coding region of the UBE3C gene for validation in the extended 105-patient dataset.

Phylogenetic Analyses and Multiple Sequence Alignments

The methods used to construct phylogenetic trees and perform multiple sequence alignments are described in Supporting Text S1.

Computational Modeling of Wild-Type and Mutant UBE3C Proteins

The available protein crystal structure of the HECT (homologous to the E6AP C-terminus) domain of E6AP (PDB 1C4Z) was used as the template.[14] A 3D model of the HECT domain was generated using CCP4 Molecular Graphics Superimposition and the molecular docking program AutoDock 4.0. The structural images were drawn using PyMOL software (DeLano Scientific).

Gene Microarrays

Gene microassay analysis was done as previously described.[12]

Cell Lines and Plasmids

The human HCC cell line MHCC-97H was established at our institute and Huh7 cells were obtained from the Chinese Academy of Sciences. The lentiviral vector pLKO.1 TRC was used to construct recombinant lentiviruses. Oligonucleotides encoding hairpin precursors for shUBE3C#1 (5′-CCAGTCATTTATTCGAGGCTA-3′) and shUBE3C#2 (5′-GCCAGACATTACTACTTCCTA-3′) were used to generate short interference RNA (siRNA) constructs. A scrambled sequence (Scr) was used as a control. Recombinant lentiviruses were amplified in HEK293T cells. All transfections were performed as previously described.[13]

Cell Proliferation, Migration, Matrigel Invasion Assays, and In Vivo Metastasis Assays

Cell counting kit-8 (CCK-8) assays, colony formation assays, Matrigel transwell assays, and in vivo metastasis assays were performed as previously described.[13]

Western Blotting, Immunohistochemistry, and Immunofluorescence Assay

Western blotting, immunohistochemistry (IHC), and immunofluorescence were performed as previously described.[13] Specific primary antibodies against UBE3C (HPA039915, Sigma-Aldrich), E-cadherin, vimentin, Snail1, N-cadherin, MMP2, MMP9, and β-actin (Cell Signaling Technology, Beverly, MA) were used. Two experienced pathologists independently assessed all IHC staining. The staining intensity was graded as follows: no staining = 0; weakly positive = 1; moderately positive = 2; and strongly positive = 3. The staining percentage was graded as follows: 0%-25% staining = 1; 26%-50% staining = 2; 51%-75% staining = 3; and 76%-100% staining = 4. The sum of the intensity and percentage scores was used to calculate the final staining score, which was then categorized as low (1-5) or high (6-7).

Statistical Analysis

Statistical analysis was performed with SPSS 19.0 software (SPSS, IBM). The cumulative survival time was calculated using the Kaplan-Meier method and analyzed with the log-rank test. Univariate and multivariate analyses were performed based on the Cox proportional hazards regression model. The Student t test and χ2 test were used as appropriate. P < 0.05 was considered statistically significant.

Results

Identification of Candidate Somatic Mutations Through Exome Sequencing

We performed whole-exome sequencing of genomic DNA obtained from one DN sample, two HCC samples, and the peripheral blood of a single patient. Following the removal of low-quality and duplicate reads, the mean coverage depth of the exome was 79 for normal cells, 77 for DN, and 75 and 78 for tumors 1 and 2, respectively (medians of 59, 57, 56, and 59, respectively). More than 93% of the target exome was covered by at least one read (Table S2). We performed SNVs and insertions and deletions (indels) analysis using a modified bioinformatics pipeline (Figs. S1 and S2) to identify somatic mutations by comparing variants identified in the DN and tumor exome datasets against dbSNP and germline variants present in the peripheral blood control samples. Nonsynonymous somatic SNVs mutations and indels that changed the protein amino acid sequence were filtered and manually confirmed using the Integrative Genomics Viewer (IGV) (Figs. S3-5). We identified 5, 72, and 180 nonsynonymous SNVs that occurred within 255 genes in the DN, tumor 1, and tumor 2, respectively (Table S3, Fig. S1). No nonsynonymous indels were found in the DN. We identified three and nine indels that targeted 12 different genes in tumors 1 and 2, respectively (Table S3, Fig. S2). Five unique nonsynonymous variants were present in the DN DNA within the CADPS, CAPG, FOXD4L1, EXOSC7 and FAM209A gene loci (Table S3). No nonsynonymous variants were found between the DN and tumor DNAs. The UBE3C gene was mutated at p.Leu211Gln and p.Gln531Leu in tumors 1 and 2, respectively (Fig. 1). Both alterations were validated in the original samples using PCR-based Sanger sequencing of the DN, tumor, and normal DNA samples from the same patient. Both mutations were heterozygous and led to amino acid substitutions in the encoded protein (Table 1, Fig. 1).

Table 1. UBE3C Mutations in HCC Samples
GeneExon NumberAllele ChangeAmino Acid Sequence ChangeAmino Acid ChangeMutation TypeCase Numbers With MutationFrequency (%)
  1. a

    One individual demonstrated both p.Leu211Gln and p.Gln531Leu alterations.

UBE3C5A>GAAG>AGGp.Lys130ArgMissense1/10617/106 a (16.0%)
7A>CTAT>TCTp.Tyr208SerMissense3/106
7T>ACTA>CAAp.Leu211GlnMissense1/106
7C>ACCA>CAAp.Pro244GlnMissense1/106
7T>CTTG>TCGp.Leu245SerMissense2/106
13A>TCAA>CTAp.Gln531LeuMissense1/106
20A>TAAC>ATCp.Asn929IleMissense1/106
20G>AGAA>AAAp.Glu959LysMissense7/106
22A>GACT>GCTp.Thr1004AlaMissense1/106

UBE3C Is Frequently Mutated in HCC Samples

We further evaluated all protein-coding exons of the UBE3C gene in an additional 105 HBV-associated HCCs with matched normal controls by way of PCR and direct DNA Sanger sequencing (Table 1, Fig. 1). We identified nine somatic mutations in UBE3C in 17 of the 106 HCC samples (including one sample used in exome sequencing), thus demonstrating a mutation frequency of 16.0%. The mutations were located in exons 5, 7, 13, 20, and 22 of UBE3C (Table 1). Among these mutations, Tyr208Ser, Leu211Gln, Pro244Gln, and Leu245Ser were located in exon 7. Interestingly, a G→A transition at position 3235 of the UBE3C messenger RNA (RefSeq NM_014671.2), which resulted in a Glu959Lys variant, was identified in seven HCC patients, accounting for ∼6.6% of all HCC patients (Table 1).

Locations of UBE3C Mutations and the Structure of the UBE3C Protein

UBE3C, which is located at 7q36.3, is also known as HECTH2, and it encodes the ubiquitin protein ligase E3C belonging to the HECT family of ubiquitin ligases (E3). The mature UBE3C protein contains 1,083 amino acid residues and two characteristic protein domains: an IQ motif and a HECT domain (Fig. 2A). Of all the somatic mutations in UBE3C detected in this study, the Asn929Ile, Glu959Lys, and Thr1004Ala variants were all generated by missense mutations in the UBE3C HECT domain region (Fig. 2A). By aligning the sequences of homologous UBE3C genes among different species, we found that the Asn929, Glu959, and Thr1004 residues have been highly conserved through evolution (Fig. 2C). In particular, Glu959 was found to be highly conserved in UBE3C orthologs and paralogs in different species with very different degrees of genetic relatedness (Fig. S6). To further address the possible structure-function relationship of these variants, the crystal structure of the UBE3C HECT domain was predicted using a computational model (Fig. 2B). According to this structural analysis, the UBE3C HECT domain is a trimer, with Glu959 of UBE3C located on an α-helix of the HECT domain surface that likely plays a role in protein-protein interactions.

Figure 2.

UBE3C mutations, protein structure, and sequence alignments with paralogs in different species. (A) Genomic organization of UBE3C exons and protein domain structure. Locations of the mutations that cause amino acid changes are indicated with red and black arrows in the UBE3C gene (top) and protein (bottom) sequences, respectively. (B) Structural prediction of UBE3C HECT domain mutations. The structure of the UBE3C HECT domain is a trimer (shown in cyan, purple, and green, respectively). The mutated residues are shown in red ribbon format. Glu959 is located on an α-helix of the HECT domain surface near the trimer interface that likely affects protein-protein interactions, suggesting a critical role for this residue in the binding and catalytic activity of the protein. (C) Multiple sequence alignments of UBE3C paralogs from different species. The mutation sites in UBE3C are marked with red arrows. All three of the amino acids mutated in the UBE3C HECT domain are highly conserved among different species, particularly Glu959, which is very highly conserved among species with different degrees of evolutionary relatedness.

UBE3C Is Also Mutated in HCV- or Nonviral-Related HCC Samples

We next investigated all protein-coding exons of the UBE3C gene using Sanger sequencing in an additional 61 samples including 13 hepatitis C virus (HCV)-related HCCs, 19 non-viral HCCs, 12 DNs, and 17 nontumoral liver tissues (including nine cirrhotic and eight noncirrhotic liver tissues). We found that the UBE3C gene was also mutated in HCV- and nonviral-related HCC samples and DNs but not in the nontumoral liver tissues (Table S4, Fig. S7).

UBE3C Is Not Correlated With TP53 and PTEN

To study the correlation of UBE3C status and the presence of other known HCC-related genes, we investigated the genetic alterations of all protein-coding exons and the protein expressions of two famous HCC-related tumor suppressors, TP53 and PTEN, in the previous 17 HCCs with UBE3C mutations. In this study, TP53 was found to be mutated in five HCCs and PTEN was found no mutation in the coding sequences (Table S5, Fig. S8A). Furthermore, the present study found no correlation of UBE3C status and the expression of TP53 and PTEN (Table S6, Fig. S8B).

To understand the impact of UBE3C status on tumor biology, we also investigated differences in gene expression between three paired HCCs with and without UBE3C mutation using complementary DNA (cDNA) microarrays. The signaling pathways and biological processes related to UBE3C are presented in Table S7.

UBE3C Promotes HCC Progression by Inducing Epithelial-Mesenchymal Transition of Tumor Cells

To explore the biological role of UBE3C in human HCC, we evaluated the effects of a loss of UBE3C function on cellular behavior using RNA interference (RNAi) in two human HCC cell lines with high UBE3C protein expression: Huh7 and MHCC-97H. When compared with control cells which were infected with a lentivirus expressing a scrambled siRNA sequence, the knockdown of UBE3C expression in HCC cells significantly inhibited cell proliferation, migration, and invasion in vitro and tumor growth and lung metastasis in a xenograft experiment in nude mice (Fig. 3).

Figure 3.

Functional analysis of UBE3C knockdown in Huh7 and MHCC-97H cells. (A) Knockdown of UBE3C expression in MHCC-97H cells inhibited tumor growth and lung metastasis in a xenograft experiment in nude mice. The black arrow shows the metastatic focus. (B) The CCK-8 assay revealed that UBE3C knockdown significantly reduced the cell growth rate. ***P < 0.001. (C) The knockdown of UBE3C expression significantly inhibited cell proliferation in colony formation assays. The significance of the difference between colony numbers was calculated using a two-tailed t test. ***P < 0.001. (D) Knockdown of UBE3C expression significantly inhibited cell migration and invasion in Matrigel transwell assays. The permeable cells were stained with Giemsa and counted. Cell migration and invasion assays were repeated three times, and the results were statistically analyzed using a two-tailed t test. ***P < 0.001.

Currently, the epithelial-mesenchymal transition (EMT) has been well recognized to represent an important step in the metastatic process. Here we monitored the ability of UBE3C in inducing HCC EMT in the presence of TGF-β1. As shown in Fig. 4, a distinct morphologic difference was observed between Huh7-Scr and MHCC-97H-Scr cells and UBE3C-knockdown cells after 2 days of treatment with TGF-β1. UBE3C-knockdown cells appeared to have more of an epithelial cobblestone-like morphology than the mesenchymal spindle-like morphology characteristic, and to delay the typical EMT phenotype including up-regulation of the epithelial marker E-cadherin and down-regulation of the mesenchymal markers such as vimentin, N-cadherin, and Snail1 (Fig. 4A,B). The EMT phenotype was confirmed further by immunofluorescent staining (Fig. 4B). Moreover, HCC cells overexpressing UBE3C had a mesenchymal phenotype in HCC tissues (Fig. 4C). Thus, we conclude that UBE3C promotes HCC progression by inducing EMT of tumor cells.

Figure 4.

UBE3C induces HCC cells EMT. (A) Different expression of EMT markers, as well as MMP2 and MMP9, were evaluated by western blotting between HCC cells (Huh7 and MHCC97-H) with UBE3C-shRNA and UBE3C-Scr. (B) Phase-contrast microscopy and immunofluorescent staining for EMT markers in HCC cells with UBE3C-shRNA and UBE3C-Scr. (C) Representative HCC cases in tissue microarrays (serial sections) were analyzed by H&E and immunohistochemical staining for UBE3C and EMT markers.

Overexpression of UBE3C Is Correlated With Poor Prognosis in HCC Patients

To investigate the relationship between HCC-associated clinicopathological features (Table S8) and UBE3C mutations, we divided our HCC cases into two groups: cases with and without UBE3C mutations. Statistical analyses revealed that UBE3C mutations were not associated with gender, age, serum alpha-fetoprotein (AFP) levels, liver cirrhosis, tumor size, tumor encapsulation, tumor differentiation, microvascular invasion, tumor TNM stage, or BCLC stage (Table 2).

Table 2. Clinical Features, UBE3C Mutations, and UBE3C Protein Expression in HCC Cases
VariableUBE3C MutationsUBE3C Protein Level
AbsencePresentPLowHighP
  1. a

    Clinical typing of tumors was performed using the tumor node metastasis (TNM) classification system of the American Joint Committee on Cancer (AJCC) and the Union for International Cancer Control (UICC) (7th ed.). HBsAg, hepatitis B surface antigen; AFP, α-fetoprotein.

Gender      
Female1130.55533130.016
Male7814 146131 
Age, years      
≤503440.24893730.822
>505513 8671 
HBsAg      
Negative00NA26190.731
Positive8917 153125 
Serum AFP, ng/mL      
≤ 202760.38651450.590
> 206211 12899 
Liver cirrhosis      
Absence4780.66422140.466
Present429 157130 
Tumor number      
Single8916NA1601160.025
Multiple01 1928 
Tumor size, cm      
≤55290.675100640.041
>5378 7980 
Tumor encapsulation      
Complete70130.84295690.357
Incomplete or none194 8475 
Microvascular invasion      
Absence2960.828104710.115
Present6011 7573 
Edmondson grade      
I-II51140.0521391160.525
III-IV383 4028 
TNM stagea      
I53110.69090510.007
II-III366 8993 
BCLC stage      
0-A5290.67553210.001
B-C378 126123 

Additionally, a UBE3C expression study revealed that UBE3C protein was expressed in the cytoplasm, and UBE3C protein levels were significantly higher in HCC tissues than in paired peritumoral liver tissues. The intratumoral UBE3Chigh group included 44.6% (144/323) of all patients. Moreover, the UBE3Chigh phenotype correlated significantly with gender, tumor number, tumor size, tumor TNM stage, and BCLC stage (Table 2). Univariate analysis also revealed that intratumoral UBE3C expression was significantly associated with recurrence and survival (Table S9), and individuals in the UBE3Chigh group had a significantly worse prognosis than those in the UBE3Clow group (Fig. 5).

Figure 5.

UBE3C expression in HCC tissues and the prognostic significance. (A) Expression of UBE3C in HCC and peritumoral tissues, as evaluated by immunohistochemistry. Representative cases of low and high expression of UBE3C. a1, a3, a5, and a7, 100×; a2, a4, a6, and a8, 200×. (B) Kaplan-Meier analysis revealed that HCC patients with UBE3C overexpression exhibited poorer overall survival and a higher cumulative recurrence rate as compared to patients with low UBE3C expression.

To further confirm the prognostic significance of UBE3C expression, a multivariate Cox proportional hazards regression analysis was performed using all of the variables that were identified as significant by univariate analysis. Multivariate analysis showed that UBE3C expression was an independent prognostic indicator for cumulative recurrence but not OS, and patients in the UBE3Chigh group were 1.407 times more likely to experience recurrence than those in the UBE3Clow group (Table S10). Additional characteristics were also assessed and these results are presented in Table S10.

Discussion

Tumorigenesis is a complex biological process driven by genetic alterations that activate oncogenes or inactivate tumor suppressor genes.[15] Similar to other solid tumors, HCC accumulates a large number of genetic variations during the development of hepatocarcinogenesis. As a normal liver develops an increasing number of genetic variants, it gradually develops precancerous lesions that ultimately evolve into a neoplasm. Hepatocarcinogenesis is known to be a histological developmental process that is induced by multiple factors and steps. Most HCCs evolve from cirrhosis-inducing conditions, with neoplasms induced by chronic HBV or HCV infection evolving from liver cirrhosis to DN and eventually HCC. Taguchi et al.[16] and Kobayashi et al.[17] reported that DNs arising from livers with chronic viral hepatitis infection or cirrhosis were predisposed to neoplasm formation and should be considered precancerous lesions. In our study, we identified five unique nonsynonymous variants in a DN within the CADPS, CAPG, FOXD4L1, EXOSC7 and FAM209A genes; however, these mutations were not identified in either tumor from this patient. In contrast, Miller et al.[18] reported that CADPS was lost in 27.6% of primary central nervous system primitive neuroectodermal tumors and that the loss of this gene was associated with poor prognosis. Moreover, the expression of CAPG, which encodes a member of the gelsolin family of actin regulatory proteins, is elevated in some human cancers and is involved in tumor development and prognosis.[19, 20] The fact that these tumor-related genes were also mutated in the liver DN prompted us to speculate that the DN had already accumulated a number of genetic changes that endowed it with the capacity for malignant transformation prior to the appearance of changes in its cytomorphological features. This finding provides a better understanding of the mechanism of hepatocarcinogenesis.

The recent exome analysis of human HCC have revealed numerous novel mutations of cancer-related genes. According to previous studies, TP53 and CTNNB1 were thought to be the most recurrently mutated tumor suppressor and oncogene for HCC, respectively.[21] Besides, a number of novel somatic alterations for HCC have been identified, such as the ARID2, ARID1A, and LEPR genes.[11, 12, 22, 23] Interestingly, various risk factors for HCC, such as HBV and HCV infections, showed various mutation profiles of the corresponding tumor genome.[23] In this study we identified frequent mutations of the UBE3C gene in human HCCs. Further targeted resequencing of this gene showed that the UBE3C gene was mainly mutated in HBV-related HCCs, rarely in DNs or HCV- and nonviral-related HCCs, and not in the nontumoral tissues. These findings imply that etiology, such as HBV and HCV, may affect the UBE3C mutation profile.

The UBE3C protein contains two characteristic protein domains: an IQ motif and a HECT domain. The IQ motif in the N-terminal region mediates substrate targeting and may serve as a binding and regulatory site,[24] whereas the HECT domain is a highly conserved C-terminal catalytic domain that binds to ubiquitin-conjugating enzymes (E2) and accepts and transfers ubiquitin to the target substrate.[25] Thus, the HECT domain provides the E3 catalytic activity. Furthermore, we identified nine UBE3C mutations in HBV-related HCCs, three of which (Asn929Ile, Glu959Lys, and Thr1004Ala) occurred in the highly conserved HECT domain region of UBE3C. The available crystal structure of the HECT domain of E6AP (also known as UBE3A) reveals a bilobal protein conformation, with a broad catalytic cleft at the junction of the two lobes, which plays a critical role in transferring ubiquitin from the E2 to the E3.[14] In this study, structure prediction analysis revealed that three UBE3C molecules were in close proximity to each other, forming a “center” in the trimer interface of the HECT domain. In light of the catalytic activity of HECT domain, we speculate that this “center” is the enzymatic activity center of UBE3C that binds to substrates and catalyzes reactions. Because Glu959 is situated on an α-helix of the HECT domain surface that may participate in protein-protein interactions, the Glu959Lys mutation in UBE3C may affect its substrate binding capacity, resulting in the excessive degradation of tumor suppressor proteins or a reduced degradation of oncoproteins involved in hepatocarcinogenesis.

To date, no association between UBE3C and tumorigenesis has been established. Previous studies reported that mutations in the HECT domain of E3s often lead to pathophysiological states, including neurological disorders and human cancers.[26-28] Thus, the high frequency of UBE3C mutations in HCCs provides a preliminary connection between UBE3C and human cancer. Furthermore, UBE3C was overexpressed in HCC tissues and promoted HCC progression in vivo and in vitro, extending its role in cancer development. The loss of contact inhibition of cell proliferation, which is a hallmark of cancer, is thought to be associated with cadherin-mediated cell-cell attachments.[29] E-cadherin is expressed in epithelial cells and plays a vital role in cell adhesion and movement. The loss of E-cadherin function or expression is thought to promote tumor progression by increasing proliferation, invasion, and metastasis. Here we showed that UBE3C down-regulated the expression of E-cadherin and induced cancer cells to undergo EMT, which is known to promote carcinoma progression and metastasis in a variety of human cancers, including HCC.[30] Thus, we speculate that UBE3C promotes HCC development and progression by targeting E-cadherin for ubiquitin-mediated proteolysis and inducing EMT in cancer cells.

The ubiquitin-proteasome system regulates crucial cellular functions, such as the cell cycle, DNA repair, cell signaling, and responses to hypoxia, and E3 enzymes play a key role in this process by way of selectively binding to their protein substrates. HECT E3s possess intrinsic catalytic activity and directly catalyze the ubiquitination of substrate proteins, thereby performing pivotal roles in maintaining cellular homeostasis and biological signaling.[31] The dysfunction of HECT E3s frequently contributes to pathological disorders, including various tumors. E6AP is the first identified member of the HECT E3 family.[32] E6AP targets the tumor suppressor protein p53 for ubiquitin-mediated proteolysis, which promotes p53 degradation and contributes to the development of a majority of human cervical cancer cases.[33] Nedd4 was recently reported to promote the proteasomal degradation of PTEN, a tumor suppressor that negatively regulates the phosphatidylinositol 3-kinase (PI3K)/AKT signaling pathway, and Nedd4 expression has also been associated with colorectal cancer, bladder, and prostate carcinomas.[34-36] An increasing number of tumor suppressor molecules have been identified as substrates of HECT E3s, and genetic aberrations in and altered expression of the HECT E3s are often observed in various human tumors.[37, 38] Although the present study found no correlation of UBE3C status and the expression of TP53 and PTEN, we confirmed that UBE3C expression was markedly increased in HCC tissues compared with peritumoral liver tissues. Additionally, we present clinically relevant data showing a correlation between high UBE3C expression and poor prognosis in HCC patients. This finding may be biased by the fact that aggressive clinicopathological features, such as multiple tumors, large tumor size, high TNM and BCLC staging of HCC, were more frequently observed in patients with high UBE3C expression than in those with low expression. Nevertheless, these data indicate that UBE3C protein expression may be a promising prognostic biomarker for HCC.

The ubiquitin-proteasome system yields a plentiful source of molecular targets for anticancer therapy.[39] Proteasome inhibitors such as bortezomib (Velcade, Millennium Pharmaceuticals) are established cancer therapeutics and have been approved by the U.S. Food and Drug Administration for the treatment of multiple myeloma and mantle cell lymphoma.[40] Given the clinical toxicity arising from the lack of specificity of proteasome inhibitors, targeting highly selective E3s will limit the adverse effects and broaden and diversify the scope of anticancer therapies, making them more specific and efficient. Furthermore, our findings indicate that UBE3C may represent a selective target for HCC patients. However, the roles of UBE3C in other types of tumors remain undetermined, and the specific substrate of UBE3C has not yet been identified. Thus, further study is needed to address these issues.

In summary, our findings identify a novel HCC susceptibility gene, UBE3C, as a candidate oncogene with a role in tumor development and progression that may serve as a potential therapeutic target in a subset of HCC patients.

Acknowledgment

We thank Jie Zong and Dai Chen (NovelBioinformatics Ltd., Co.) for technical support in the bioinformatics analysis process.

Ancillary

Advertisement