Cancer Letters 357 (2015) 510–519
Contents lists available at ScienceDirect
Cancer Letters j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / c a n l e t
Original Articles
RNA over-editing of BLCAP contributes to hepatocarcinogenesis identified by whole-genome and transcriptome sequencing Xueda Hu a,b,1, Shengqing Wan b,1, Ying Ou c,1, Boping Zhou a,d,1, Jialou Zhu b, Xin Yi b, Yanfang Guan b, Wenlong Jia b, Xing Liu c, Qiudao Wang c, Yao Qi c, Qing Yuan c, Wanqiu Huang e, Weijia Liao f, Yun Wang c, Qinghua Zhang c, Huasheng Xiao c, Xinchun Chen a,d, Jian Huang a,c,d,* a
Shenzhen Key Laboratory of Infection and Immunity, Shenzhen Third People’s Hospital, Guangdong Medical College, Shenzhen 518112, China BGI-Shenzhen, Shenzhen 518083, China c Shanghai-MOST Key Laboratory for Disease and Health Genomics, Chinese National Human Genome Center and National Engineering Center for Biochip at Shanghai, Shanghai, China d Guangdong Key Laboratory of Diagnosis & Treatment for Emerging Infectious Disease, Shenzhen Third People’s Hospital, Guangdong Medical college, Shenzhen 518112, China e Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China f Hepatology Institute of Guilin Medical University, Guilin, Guangxi Zhuang Autonomous Region, China b
A R T I C L E
I N F O
Article history: Received 2 October 2014 Received in revised form 26 November 2014 Accepted 2 December 2014 Keywords: RNA over-editing BLCAP gene Hepatocarcinogenesis Whole-genome and transcriptome sequencing
A B S T R A C T
Hepatocellular carcinoma (HCC) is one of the most common cancers worldwide, although the treatment of this disease has changed little in recent decades because most of the genetic events that initiate this disease remain unknown. To better understand HCC pathogenesis at the molecular level and to uncover novel tumor-initiating events, we integrated RNA-seq and DNA-seq data derived from two pairs of HCC tissues. We found that BLCAP is novel editing gene in HCC and has over-editing expression in 40.1% HCCs compared to adjacent liver tissues. We then used RNA interference and gene transfection to assess the roles of BLCAP RNA editing in tumor proliferation. Our results showed that compared to the wild-type BLCAP gene, the RNA-edited BLCAP gene may stably promote cell proliferation (including cell growth, colony formation in vitro, and tumorigenicity in vivo) by enhancing the phosphorylation of AKT, mTOR, and MDM2 and inhibiting the phosphorylation of TP53. Our current results suggest that the RNA over-editing of BLCAP gene may serve as a novel potential driver in advanced HCC through activating AKT/mTOR signal pathway. © 2014 Elsevier Ireland Ltd. All rights reserved.
Introduction Hepatocellular carcinoma (HCC) is a highly malignant tumor with a poor clinical outcome and represents the third most common cause of cancer-related deaths worldwide [1]. Many predisposing environmental factors can contribute to liver cancer, such as infection with hepatitis B virus or hepatitis C virus, chronic exposure to Aflatoxin B1, and alcoholic cirrhosis [2]. In addition, previous studies have indicated that the development of HCC is a multistep process characterized by the accumulation of genetic and epigenetic alterations, and many of the genetic abnormalities contributing to HCC, such as potential oncogenes and tumor suppressor genes, are well known. Furthermore, with the rapid development of genomic technology, several studies have demonstrated the possibility of
* Corresponding author. Tel.: +86 21 51320142; fax: +86 21 51320142. E-mail address: [email protected] (J. Huang). 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.canlet.2014.12.006 0304-3835/© 2014 Elsevier Ireland Ltd. All rights reserved.
characterizing the genomic profile of HCC [3–7]; these efforts have mainly focused on identifying the somatic DNA variations in tumors and investigating the relationships between these mutations and various clinical features. The transfer of genetic information, as described by the central dogma, has always been considered to be faithful and deterministic. Therefore, cancer genomic studies have paid the most attention to DNA sequence variations, the ultimate source of genetic information, as the basis of individual differences in oncogenesis. Moreover, most studies that have examined mRNA and protein levels have only analyzed their expression differences rather than the sequence differences between individuals. In addition, there are known exceptions to the one-to-one relationship between DNA and mRNA sequences. RNA editing is the post- or co-transcriptional modification of RNA nucleotides (nt) from their complementary DNA sequences [8]. In humans, the most frequent form of RNA editing is the conversion of adenosine to inosine; the transcription machinery subsequently recognizes these inosines as guanosines [9]. With DNA-seq and RNA-seq as well as next-generation sequencing platforms becoming routine tools in molecular research, several
X. Hu et al./Cancer Letters 357 (2015) 510–519
bioinformatic methods have been established to identify RNA editing sites on a genome-wide basis via high-throughput sequencing data [8,10–12]. Additionally, some genome-wide studies of human RNA editing have been reported [13,14], although these studies were mostly performed at the population genetics level and examined healthy individuals, and very few studies have examined cancers. To analyze the RNA editome in HCC and further understand its potential role in hepatocarcinogenesis, we integrated DNA-Seq and RNA-Seq analysis of two pairs of HCC tissues, their matched nontumor tissues, and 2 HCC-related cell lines. Our analysis focused on the RNA editing of BLCAP (bladder cancer-associated protein and regulated by ADAR1), which was subjected to further study in large sample cohorts and functional analyses, and the results demonstrated that BLCAP RNA over-editing may contribute to hepatocarcinogenesis. Materials and methods HCC tissue specimens All HCC tissue specimens were obtained from patients who underwent surgical resection of their tumors and provided informed consent prior to liver surgery. The primary tumor specimens were immediately frozen at −80°C until DNA/RNA extraction. Specimens (approximately 1 cm3) of both tumor and adjacent liver tissue were taken from each patient, and the diagnosis of HCC was confirmed by pathological examination. The HCC specimens presented in this study were grouped according to differentiation grades II–III following the Edmondson–Steiner grading system. The clinical characteristics of the patients and tumors are summarized in Appendix: Supplementary Table S1. This project and its protocols involving human and animal tissues were approved by the ethics committee of the Chinese National Human Genome Center. Laboratory methods See the Supporting Materials and Methods section for detailed experimental procedures.
Results Editome analysis of six samples using whole-genome and transcriptome sequencing data By carrying out next-generation DNA sequencing, we obtained a total of 1,810,665,526–1,910,437,344 high-quality reads (Appendix: Supplementary Table S2) from six samples including two pairs of HCC tissues and 2 HCC cell lines used in this study (Appendix: Supplementary Table S1). These reads were aligned to the human genome reference sequence (hg19) using Burrows–Wheeler Alignment (BWA) software; an average of 92.34% of the sequence reads were clean enough to be used for further analysis and the average coverage depth reached 51.31-fold for each DNA sample. In parallel, using RNA sequencing, we obtained a total of 371,011,349– 640,637,068 high-quality reads from the same six samples. We used BWA software to map these reads to the refseq gene database and human genome (hg19), with an average mapping ratio of 57.81%. Thus, our sequencing coverage was adequate to sensitively and reliably detect various somatic alterations and sequence differences between DNA and RNA at the whole-genome level. And then to explore the RNA editome of human tissues, we first removed the somatic substitute variations. And then we compared the genotypes derived from DNA-Seq with those from RNASeq using the six specimens listed above (Appendix: Supplementary Table S1) according to strict criteria (Fig. 1). Additionally, to accurately obtain the site of RNA editing, we removed any potential RNA editing sites in the first five and last five positions of the sequencing reads according to previous guidelines published in scientific journals [15–17]. The final data showed that the RNA editing sites were randomly located in the 80-bp-length reads (i.e., between the 6th and the 85th positions of the sequencing reads) (Appendix:
511
Supplementary Fig. S1a and S1b). In total, 52,792 potential RNA editing sites (Supplemental_data_1.xlsx, ftp://183.62.232.83/) (21,886 in T311, 18,048 in N321, 23,723 in T273, 11,442 in N283, 8235 in LM6, and 8124 in M97L) were identified (Fig. 2a). Of these, only 900 RNA editing sites were detected among all six specimens. An average of 5379 RNA editing sites (range from 2001–10,492) were found specifically in the individual samples (Appendix: Supplementary Fig. S2). Furthermore, we found that approximately 40% of the RNA editing sites were located in the DARNED database [18], and 60% of the RNA editing sites were first reported in this study (Fig. 2a). These data also showed that 12 types of differences were found in each of the 6 samples (Fig. 2b). A total of 82.15% of the RNA editing sites were A-to-G changes, which may have been the result of deamination by adenosine deaminases acting on RNA (ADAR). An additional 12.09%, 1.70%, and 0.98% of the RNA editing sites involved T-to-C, G-to-A, and C-to-T alterations, respectively; the other eight types accounted for only 3.07% of the changes (Fig. 2b and Appendix: Supplementary Table S3). The relative proportion of each type across individuals was similar to previous findings [10]. To experimentally validate our calls, we randomly sequenced 123 potential RNA editing sites in T273, N283, T311, and N321, using an Ion Torrent sequencer. These results validated 66.0%, 81.0%, 87.0%, and 90.0% of the editing sites in T273, N283, T311, and N321, respectively (Appendix: Supplementary Table S4), which suggested that these identified RNA editing sites have high credibility and that these data could be used for further analysis. Over-editome analysis of HCCs Interestingly, we found that the number of RNA editing sites was significantly different in the six samples and that the number of RNA editing sites was greater in HCC tissue than adjacent liver tissue (nonHCC) (p < 0.05, Appendix: Supplementary Fig. S3). No differences between the LM6 and 97L cell lines were observed (p > 0.05, Appendix: Supplementary Fig. S3), although these two cell lines were also derived from the same genetic background (one male patient suffers from HCC) [6]. In addition, we found that hepatoma cell lines display less RNA editing sites compared to non tumoral liver. The potential reasons may be related with that cell line homogenized under subculture, whilst hepatoma cells have frequent heterogeneity in tumor tissue. To identify how many genes are involved in RNA over-editing events, we first statistically analyzed the differential expression of genes with RNA editing events in HCC samples compared to nonHCC samples according to a previously described criterion [19]. The editing sites with significant differences between HCC and nonHCC samples at the RNA level were defined as RNA over-editing sites based on the following criteria: fold change ≥2 and p ≤ 0.01, Q ≤ 0.1 by DEGseq. We then identified 3509 RNA editing sites (Supplemental_data_2.xlsx, ftp://183.62.232.83/) with significantly different expression levels in the two pairs of HCC tissues (p < 0.01). Of these, 2101 and 1690 RNA editing sites were found in T273/N283 and T311/N321, respectively; there were also 282 sites that were RNA edited in both samples. Among the 282 shared RNA over-editing sites, we found only 95 nonsynonymous sites (Appendix: Supplementary Table S5) associated with 84 genes (Appendix: Supplementary Fig. S4a). Seven of these 84 genes with RNA editing events displayed high-frequency A→G transcript over-editing in 2 pairs of HCC samples, leading to an amino acid substitution. Given the low frequency of transcripts with RNA editing events, Sequenom MassARRAY technology and Ion Torrent sequencing were used to validate the 7 RNA over-editing sites in the DNA and cDNA samples from the two pairs of HCC tissue. These results showed that 6 genes were confirmed to be RNA overediting genes including BLCAP (Appendix: Supplementary Fig. S4b), AZIN1 (Appendix: Supplementary Fig. S4c), FDPS (Appendix:
512
X. Hu et al./Cancer Letters 357 (2015) 510–519
Fig. 1. A flowchart for the process of screening and identifying variants and RNA editing sites using high-throughput sequence data. (RE indicates RNA editing.).
X. Hu et al./Cancer Letters 357 (2015) 510–519
513
Fig. 2. Characterization of the RNA editome of six HCC specimens. (a) The number of RNA editing sites was identified in six HCC specimens, including known RNA editing sites deposited in the DARNED database and novel RNA editing sites identified in this study. (b) Twelve types of RNA editing were identified in the transcriptome of the six specimens.
Supplementary Fig. S4d), CCNI, PLXNB, and MLL4 (Appendix: Supplementary Table S4). BLCAP RNA over-editing contributes to hepatocarcinogenesis To study the role of RNA editing in liver carcinogenesis, we selected BLCAP gene as a model to observe whether the over-editing gene may play roles in hepatocarcinogenesis in some proof of principle experiments, Using Sequenom MassARRAY technology, we found that the BLCAP gene demonstrated a high frequency of
nonsynonymous A→I transcript editing, leading to a Tyr→Cys amino acid substitution in the two pairs of HCC samples. To confirm this result, we investigated the status of BLCAP editing in 179 matched primary HCC and non-HCC samples using a Sequenom MassARRAY. Our results showed that the expression of RNA-edited BLCAP was significantly higher in HCCs than non-HCCs (p < 0.0001, Mann– Whitney Test, Fig. 3a, Appendix: Supplementary Table S6) and that 40.1% of the primary HCC specimens had BLCAP gene over-editing, as defined by a 2-fold increase in RNA-edited BLCAP in HCCs compared to non-HCCs (Fig. 3a, Appendix: Supplementary Table S6). The
Fig. 3. BLCAP over-editing is strongly associated with HCC pathogenesis. (a) BLCAP editing was detected in 179 paired HCC and non-HCC samples using Sequenom MassARRAY. The p values shown were calculated using a Mann–Whitney U test. The upper and lower edges of each box represent the 75th and 25th percentile, respectively; the upper and lower bars indicate the highest and lowest values determined, respectively. (b) Dot Box showing BLCAP editing in HCCs. The p values shown were calculated by Mann– Whitney U test. The HCC specimens were subdivided into six categories according to the tumor size, tumor number and according to the presence or absence of PVTT. T peak indicates wild type BLCAP and C peak indicates RNA editing BLCAP.
514
X. Hu et al./Cancer Letters 357 (2015) 510–519
Table 1 The statistical results of BLCAP overediting status and the clinical characteristics of 179 samples in this study. Clinical characteristics
Clinical variable
No. of patients
BLCAP_RE
Contents lists available at ScienceDirect
Cancer Letters j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / c a n l e t
Original Articles
RNA over-editing of BLCAP contributes to hepatocarcinogenesis identified by whole-genome and transcriptome sequencing Xueda Hu a,b,1, Shengqing Wan b,1, Ying Ou c,1, Boping Zhou a,d,1, Jialou Zhu b, Xin Yi b, Yanfang Guan b, Wenlong Jia b, Xing Liu c, Qiudao Wang c, Yao Qi c, Qing Yuan c, Wanqiu Huang e, Weijia Liao f, Yun Wang c, Qinghua Zhang c, Huasheng Xiao c, Xinchun Chen a,d, Jian Huang a,c,d,* a
Shenzhen Key Laboratory of Infection and Immunity, Shenzhen Third People’s Hospital, Guangdong Medical College, Shenzhen 518112, China BGI-Shenzhen, Shenzhen 518083, China c Shanghai-MOST Key Laboratory for Disease and Health Genomics, Chinese National Human Genome Center and National Engineering Center for Biochip at Shanghai, Shanghai, China d Guangdong Key Laboratory of Diagnosis & Treatment for Emerging Infectious Disease, Shenzhen Third People’s Hospital, Guangdong Medical college, Shenzhen 518112, China e Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China f Hepatology Institute of Guilin Medical University, Guilin, Guangxi Zhuang Autonomous Region, China b
A R T I C L E
I N F O
Article history: Received 2 October 2014 Received in revised form 26 November 2014 Accepted 2 December 2014 Keywords: RNA over-editing BLCAP gene Hepatocarcinogenesis Whole-genome and transcriptome sequencing
A B S T R A C T
Hepatocellular carcinoma (HCC) is one of the most common cancers worldwide, although the treatment of this disease has changed little in recent decades because most of the genetic events that initiate this disease remain unknown. To better understand HCC pathogenesis at the molecular level and to uncover novel tumor-initiating events, we integrated RNA-seq and DNA-seq data derived from two pairs of HCC tissues. We found that BLCAP is novel editing gene in HCC and has over-editing expression in 40.1% HCCs compared to adjacent liver tissues. We then used RNA interference and gene transfection to assess the roles of BLCAP RNA editing in tumor proliferation. Our results showed that compared to the wild-type BLCAP gene, the RNA-edited BLCAP gene may stably promote cell proliferation (including cell growth, colony formation in vitro, and tumorigenicity in vivo) by enhancing the phosphorylation of AKT, mTOR, and MDM2 and inhibiting the phosphorylation of TP53. Our current results suggest that the RNA over-editing of BLCAP gene may serve as a novel potential driver in advanced HCC through activating AKT/mTOR signal pathway. © 2014 Elsevier Ireland Ltd. All rights reserved.
Introduction Hepatocellular carcinoma (HCC) is a highly malignant tumor with a poor clinical outcome and represents the third most common cause of cancer-related deaths worldwide [1]. Many predisposing environmental factors can contribute to liver cancer, such as infection with hepatitis B virus or hepatitis C virus, chronic exposure to Aflatoxin B1, and alcoholic cirrhosis [2]. In addition, previous studies have indicated that the development of HCC is a multistep process characterized by the accumulation of genetic and epigenetic alterations, and many of the genetic abnormalities contributing to HCC, such as potential oncogenes and tumor suppressor genes, are well known. Furthermore, with the rapid development of genomic technology, several studies have demonstrated the possibility of
* Corresponding author. Tel.: +86 21 51320142; fax: +86 21 51320142. E-mail address: [email protected] (J. Huang). 1 These authors contributed equally to this work. http://dx.doi.org/10.1016/j.canlet.2014.12.006 0304-3835/© 2014 Elsevier Ireland Ltd. All rights reserved.
characterizing the genomic profile of HCC [3–7]; these efforts have mainly focused on identifying the somatic DNA variations in tumors and investigating the relationships between these mutations and various clinical features. The transfer of genetic information, as described by the central dogma, has always been considered to be faithful and deterministic. Therefore, cancer genomic studies have paid the most attention to DNA sequence variations, the ultimate source of genetic information, as the basis of individual differences in oncogenesis. Moreover, most studies that have examined mRNA and protein levels have only analyzed their expression differences rather than the sequence differences between individuals. In addition, there are known exceptions to the one-to-one relationship between DNA and mRNA sequences. RNA editing is the post- or co-transcriptional modification of RNA nucleotides (nt) from their complementary DNA sequences [8]. In humans, the most frequent form of RNA editing is the conversion of adenosine to inosine; the transcription machinery subsequently recognizes these inosines as guanosines [9]. With DNA-seq and RNA-seq as well as next-generation sequencing platforms becoming routine tools in molecular research, several
X. Hu et al./Cancer Letters 357 (2015) 510–519
bioinformatic methods have been established to identify RNA editing sites on a genome-wide basis via high-throughput sequencing data [8,10–12]. Additionally, some genome-wide studies of human RNA editing have been reported [13,14], although these studies were mostly performed at the population genetics level and examined healthy individuals, and very few studies have examined cancers. To analyze the RNA editome in HCC and further understand its potential role in hepatocarcinogenesis, we integrated DNA-Seq and RNA-Seq analysis of two pairs of HCC tissues, their matched nontumor tissues, and 2 HCC-related cell lines. Our analysis focused on the RNA editing of BLCAP (bladder cancer-associated protein and regulated by ADAR1), which was subjected to further study in large sample cohorts and functional analyses, and the results demonstrated that BLCAP RNA over-editing may contribute to hepatocarcinogenesis. Materials and methods HCC tissue specimens All HCC tissue specimens were obtained from patients who underwent surgical resection of their tumors and provided informed consent prior to liver surgery. The primary tumor specimens were immediately frozen at −80°C until DNA/RNA extraction. Specimens (approximately 1 cm3) of both tumor and adjacent liver tissue were taken from each patient, and the diagnosis of HCC was confirmed by pathological examination. The HCC specimens presented in this study were grouped according to differentiation grades II–III following the Edmondson–Steiner grading system. The clinical characteristics of the patients and tumors are summarized in Appendix: Supplementary Table S1. This project and its protocols involving human and animal tissues were approved by the ethics committee of the Chinese National Human Genome Center. Laboratory methods See the Supporting Materials and Methods section for detailed experimental procedures.
Results Editome analysis of six samples using whole-genome and transcriptome sequencing data By carrying out next-generation DNA sequencing, we obtained a total of 1,810,665,526–1,910,437,344 high-quality reads (Appendix: Supplementary Table S2) from six samples including two pairs of HCC tissues and 2 HCC cell lines used in this study (Appendix: Supplementary Table S1). These reads were aligned to the human genome reference sequence (hg19) using Burrows–Wheeler Alignment (BWA) software; an average of 92.34% of the sequence reads were clean enough to be used for further analysis and the average coverage depth reached 51.31-fold for each DNA sample. In parallel, using RNA sequencing, we obtained a total of 371,011,349– 640,637,068 high-quality reads from the same six samples. We used BWA software to map these reads to the refseq gene database and human genome (hg19), with an average mapping ratio of 57.81%. Thus, our sequencing coverage was adequate to sensitively and reliably detect various somatic alterations and sequence differences between DNA and RNA at the whole-genome level. And then to explore the RNA editome of human tissues, we first removed the somatic substitute variations. And then we compared the genotypes derived from DNA-Seq with those from RNASeq using the six specimens listed above (Appendix: Supplementary Table S1) according to strict criteria (Fig. 1). Additionally, to accurately obtain the site of RNA editing, we removed any potential RNA editing sites in the first five and last five positions of the sequencing reads according to previous guidelines published in scientific journals [15–17]. The final data showed that the RNA editing sites were randomly located in the 80-bp-length reads (i.e., between the 6th and the 85th positions of the sequencing reads) (Appendix:
511
Supplementary Fig. S1a and S1b). In total, 52,792 potential RNA editing sites (Supplemental_data_1.xlsx, ftp://183.62.232.83/) (21,886 in T311, 18,048 in N321, 23,723 in T273, 11,442 in N283, 8235 in LM6, and 8124 in M97L) were identified (Fig. 2a). Of these, only 900 RNA editing sites were detected among all six specimens. An average of 5379 RNA editing sites (range from 2001–10,492) were found specifically in the individual samples (Appendix: Supplementary Fig. S2). Furthermore, we found that approximately 40% of the RNA editing sites were located in the DARNED database [18], and 60% of the RNA editing sites were first reported in this study (Fig. 2a). These data also showed that 12 types of differences were found in each of the 6 samples (Fig. 2b). A total of 82.15% of the RNA editing sites were A-to-G changes, which may have been the result of deamination by adenosine deaminases acting on RNA (ADAR). An additional 12.09%, 1.70%, and 0.98% of the RNA editing sites involved T-to-C, G-to-A, and C-to-T alterations, respectively; the other eight types accounted for only 3.07% of the changes (Fig. 2b and Appendix: Supplementary Table S3). The relative proportion of each type across individuals was similar to previous findings [10]. To experimentally validate our calls, we randomly sequenced 123 potential RNA editing sites in T273, N283, T311, and N321, using an Ion Torrent sequencer. These results validated 66.0%, 81.0%, 87.0%, and 90.0% of the editing sites in T273, N283, T311, and N321, respectively (Appendix: Supplementary Table S4), which suggested that these identified RNA editing sites have high credibility and that these data could be used for further analysis. Over-editome analysis of HCCs Interestingly, we found that the number of RNA editing sites was significantly different in the six samples and that the number of RNA editing sites was greater in HCC tissue than adjacent liver tissue (nonHCC) (p < 0.05, Appendix: Supplementary Fig. S3). No differences between the LM6 and 97L cell lines were observed (p > 0.05, Appendix: Supplementary Fig. S3), although these two cell lines were also derived from the same genetic background (one male patient suffers from HCC) [6]. In addition, we found that hepatoma cell lines display less RNA editing sites compared to non tumoral liver. The potential reasons may be related with that cell line homogenized under subculture, whilst hepatoma cells have frequent heterogeneity in tumor tissue. To identify how many genes are involved in RNA over-editing events, we first statistically analyzed the differential expression of genes with RNA editing events in HCC samples compared to nonHCC samples according to a previously described criterion [19]. The editing sites with significant differences between HCC and nonHCC samples at the RNA level were defined as RNA over-editing sites based on the following criteria: fold change ≥2 and p ≤ 0.01, Q ≤ 0.1 by DEGseq. We then identified 3509 RNA editing sites (Supplemental_data_2.xlsx, ftp://183.62.232.83/) with significantly different expression levels in the two pairs of HCC tissues (p < 0.01). Of these, 2101 and 1690 RNA editing sites were found in T273/N283 and T311/N321, respectively; there were also 282 sites that were RNA edited in both samples. Among the 282 shared RNA over-editing sites, we found only 95 nonsynonymous sites (Appendix: Supplementary Table S5) associated with 84 genes (Appendix: Supplementary Fig. S4a). Seven of these 84 genes with RNA editing events displayed high-frequency A→G transcript over-editing in 2 pairs of HCC samples, leading to an amino acid substitution. Given the low frequency of transcripts with RNA editing events, Sequenom MassARRAY technology and Ion Torrent sequencing were used to validate the 7 RNA over-editing sites in the DNA and cDNA samples from the two pairs of HCC tissue. These results showed that 6 genes were confirmed to be RNA overediting genes including BLCAP (Appendix: Supplementary Fig. S4b), AZIN1 (Appendix: Supplementary Fig. S4c), FDPS (Appendix:
512
X. Hu et al./Cancer Letters 357 (2015) 510–519
Fig. 1. A flowchart for the process of screening and identifying variants and RNA editing sites using high-throughput sequence data. (RE indicates RNA editing.).
X. Hu et al./Cancer Letters 357 (2015) 510–519
513
Fig. 2. Characterization of the RNA editome of six HCC specimens. (a) The number of RNA editing sites was identified in six HCC specimens, including known RNA editing sites deposited in the DARNED database and novel RNA editing sites identified in this study. (b) Twelve types of RNA editing were identified in the transcriptome of the six specimens.
Supplementary Fig. S4d), CCNI, PLXNB, and MLL4 (Appendix: Supplementary Table S4). BLCAP RNA over-editing contributes to hepatocarcinogenesis To study the role of RNA editing in liver carcinogenesis, we selected BLCAP gene as a model to observe whether the over-editing gene may play roles in hepatocarcinogenesis in some proof of principle experiments, Using Sequenom MassARRAY technology, we found that the BLCAP gene demonstrated a high frequency of
nonsynonymous A→I transcript editing, leading to a Tyr→Cys amino acid substitution in the two pairs of HCC samples. To confirm this result, we investigated the status of BLCAP editing in 179 matched primary HCC and non-HCC samples using a Sequenom MassARRAY. Our results showed that the expression of RNA-edited BLCAP was significantly higher in HCCs than non-HCCs (p < 0.0001, Mann– Whitney Test, Fig. 3a, Appendix: Supplementary Table S6) and that 40.1% of the primary HCC specimens had BLCAP gene over-editing, as defined by a 2-fold increase in RNA-edited BLCAP in HCCs compared to non-HCCs (Fig. 3a, Appendix: Supplementary Table S6). The
Fig. 3. BLCAP over-editing is strongly associated with HCC pathogenesis. (a) BLCAP editing was detected in 179 paired HCC and non-HCC samples using Sequenom MassARRAY. The p values shown were calculated using a Mann–Whitney U test. The upper and lower edges of each box represent the 75th and 25th percentile, respectively; the upper and lower bars indicate the highest and lowest values determined, respectively. (b) Dot Box showing BLCAP editing in HCCs. The p values shown were calculated by Mann– Whitney U test. The HCC specimens were subdivided into six categories according to the tumor size, tumor number and according to the presence or absence of PVTT. T peak indicates wild type BLCAP and C peak indicates RNA editing BLCAP.
514
X. Hu et al./Cancer Letters 357 (2015) 510–519
Table 1 The statistical results of BLCAP overediting status and the clinical characteristics of 179 samples in this study. Clinical characteristics
Clinical variable
No. of patients
BLCAP_RE
