Abstract
Background/Aim: The spliceosome pathway, including Splicing Factor 3b Subunit 4 (SF3B4), plays an important role in carcinogenesis and progression in various cancers; however, the clinical relevance of SF3B4 in esophageal squamous cell carcinoma (ESCC) remains unknown. Patients and Methods: SF3B4 expression was evaluated by real-time reverse transcription polymerase chain reaction in 80 ESCC patients. In order to explore the mechanism of SF3B4 in ESCC, the mRNA expression and copy number of SF3B4 were obtained from TCGA and we also implemented gene set enrichment analysis (GSEA). Results: The high SF3B4 expression group (n=33) showed significantly more lymphatic permeation and poorer prognosis than the low SF3B4 expression group (n=47). GSEA revealed that high SF3B4 expression was correlated with genes associated with the transcription factor E2F and the G2/M checkpoint. SF3B4 expression was positively correlated with SF3B4 DNA copy number. Conclusion: Over-expression of SF3B4 may play a crucial role in the lymphatic progression of ESCC.
Esophageal squamous cell carcinoma (ESCC) is one of the most frequent types of malignant cancer in East Asian countries, including Japan (1). Recent advancements in multidisciplinary therapy have improved clinical outcomes to some extent. However, the 5-year survival frequency of ESCC patients is still only 30-40% because of the development of lymph node metastasis and subsequent tumor invasion into adjacent critical organs (2-6). It is extremely important to clarify the molecular mechanisms underlying the progression of ESCC to improve patient outcome.
The spliceosome is composed of 5 small nuclear ribonucleoproteins (snRNPs) – U1, U2, U4, U5 and U6 – and multiple other proteins play an important role in the splicing process (7, 8). The involvement of alternative splicing factors has been reported in various diseases, including cancers (9). As a result, dysfunction of alternative splicing and aberrant production of specific splicing variants promote carcinogenesis, progression and chemo-resistance (10-12). Recently, bioinformatic studies have shown the contribution of the spliceosomal pathway to the progression of various cancers (13-17). In ESCC, upregulation of miR-196a was observed and target pathway analysis showed a strong association with the spliceosome pathway (17).
Splicing factor B subunit 4 (SF3B4) is a component of the U2 pre-mRNA spliceosomal complex. Loss of function mutations in SF3B4 are the major cause of Nager syndrome (18-20). Recently, Shen et al. have reported an intriguing finding concerning the early carcinogenic role of the SF3B4 gene in hepatoma. The upregulated SF3B4 accelerates the production of alternative splicing variants of the Kruppel-like factor 4 (KLF4) gene to be a nonfunctional skipped exon transcripts. That dysfunction leads to the inactivation of the p27 (CDKN1B) gene, which is involved in the activation of the SNAI2 gene, an epithelial to mesenchymal transition-related gene (21). We have also reported that SF3B4 is upregulated in hepatocellular carcinoma (HCC) and is associated with the prognosis of HCC patients (22). However, the relevance of SF3B4 in ESCC is unknown. Thus, in the current study, we aimed to clarify the clinicopathological and prognostic significance of SF3B4 expression in ESCC.
Patients and Methods
Patients. Primary ESCC samples and their corresponding adjacent epithelia were obtained from 80 patients who underwent esophagostomy at our institute (Kyushu University Beppu Hospital) and our affiliated hospitals between1998 and 2012 (Kyushu dataset). All patients had a histological diagnosis of ESCC and were closely followed at 3-month intervals after surgery. The median follow-up period was 3.9 years. All patients were treated in accordance with the Japan Esophageal Society Guidelines for the treatment of ESCC. Informed consent was obtained from all patients, and the Institutional Review Board of our university approved this study. Following collection, all resected ESCC and the adjacent normal esophageal epithelial samples were immediately frozen in liquid nitrogen and stored at −80°C until RNA extraction. All clinicopathological data, including patients' age, sex, histological grade, tumor depth of invasion, location, lymph node metastasis, lymphatic invasion and vascular invasion were obtained from the clinical and pathological records.
RNA preparation, reverse transcription (RT) reaction and quantitative polymerase chain reaction (qPCR). Total RNA from esophageal cancer or adjacent normal epithelial were extracted by the modified Acid Guanidinium Phenol Chloroform (AGPC) method using ISOGEN (Nippon Gene, Tokyo, Japan). RT was performed using 8 μg of total RNA with M-MLV reverse transcriptase according to the manufacturer's instructions (Invitrogen, Carlsbad, CA, USA).
Quantitative polymerase chain reaction (qPCR). qPCR was performed in a LightCycler 480 instrument (Roche Applied Science, Basel, Switzerland) using a LightCycler 480 Probes Master kit (Roche Applied Science) as described previously (22). In brief, PCR primer sequences for human SF3B4 were as follows: sense, 5’-AGACGGCGGGATCTCTTT-3’; antisense, 5’-CACGTACACAGTGGCATCCT-3’. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) primers, which served as the internal control to normalize the expression level of SF3B4, were as follows: sense, 5’-TTGGTATCGTGGAAGGACTCTCA-3’; antisense, 5’-TGTCATATTTGGCAGGTT-3’. The amplification conditions were as follows: 10 min at 95°C, followed by 45 cycles of 10 sec at 95°C and 30 sec at 60°C. The expression levels were expressed relative to the expression levels of Human Universal Reference Total RNA (Clontech, Palo Alto, CA, USA).
Public clinical dataset. We obtained paired SF3B4 expression profiles and survival data of 86 available ESCC cases from The Cancer Genome Atlas (TCGA) of the Broad Institute's Firehose (https://gdac.broadinstitute.org). Copy number data for 90 ESCC cases were also obtained from TCGA.
Gene set enrichment analysis (GSEA). We acquired ESCC expression profiles from the National Center for Biotechnology Information gene expression omnibus database (accession code GSE2533) and analyzed the correlations between SF3B4 expression and previously annotated gene expression signatures by applying GSEA (23, 24).
Statistical analysis. Either the X2 test or Fisher's exact test was used for comparisons between SF3B4 expression and clinicopathological findings. Survival curves were calculated by the Kaplan–Meier method and analyzed by the log-rank test. A comparison of SF3B4 expression in ESCC and normal epithelial tissue was evaluated using the Mann-Whitney's U-test. These results were obtained using R version 3.1.1 [R Core Team (2014). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria; https://www.r-project.org]. p-Values less than 0.05 were considered statistically significant.
Results
Comparison of SF3B4 expression levels between ESCC and normal epithelial tissue. We compared the SF3B4 expression levels between ESCC and the adjacent normal epithelial tissue by qPCR. SF3B4 expression was significantly higher in ESCC tumor tissues than in normal epithelial tissue (p=0.03; Figure 1).
Prognostic relevance of SF3B4 expression in ESCC. Next, to estimate the clinical significance of SF3B4 expression in ESCC, we divided the 80 patients in the Kyushu dataset into two groups: a high SF3B4 expression group and a low SF3B4 expression group, according to the ratio of SF3B4 expression in tumor tissue to that of normal epithelial tissue (T/N). The cut-off line was set to T/N=2. Patients in the SF3B4 high expression group had significantly poorer outcomes than those in the SF3B4 low expression group (p<0.01; Figure 2A). In addition, TCGA dataset also revealed that the high SF3B4 expression group had significantly poorer overall survival than the low SF3B4 expression group (p=0.04; Figure 2B).
Correlations between SF3B4 expression and clinicopathological factors. We examined the correlations between SF3B4 expression level and clinicopathological factors of patients in the Kyushu dataset (Table I). Lymphatic invasion was more frequently observed in the high SF3B4 expression group than in the low SF3B4 expression group (p=0.03). With respect to other clinicopathological factors, there were no significant differences between the 2 groups.
Expression of SF3B4 correlated with the cell cycle in ESCC. We investigated the function of SF3B4 in ESCC by applying GSEA to ESCC cases (GSE2533). The results showed that SF3B4 expression was positively correlated with E2F target genes' expression (p<0.01; Figure 3A). In addition, SF3B4 expression was positively correlated with the expression of genes associated with the G2/M checkpoint (p<0.01; Figure 3B).
Correlation between SF3B4 DNA copy number variation and SF3B4 expression. To examine the influence of DNA copy number variation on SF3B4, we analyzed the correlation between the copy number and expression levels of SF3B4 in TCGA dataset. SF3B4 expression was positively correlated with SF3B4 DNA copy number (R=0.47, p<0.01; Figure 4).
Discussion
Few studies have examined the biological role of SF3B4 in solid cancers. In our previous study, SF3B4 overexpression was a significant indicator of malignant outcomes in HCC (22), while another study showed that SF3B4 had a suppressive role in pancreatic cancer (25). Although, it has been reported that the spliceosome pathway is strongly involved in carcinogenesis and development via microRNA 196a in ESCC (17), the relationship between ESCC and SF3B4 has not been examined. Here we provide the first description of the clinicopathological significance of SF3B4 in ESCC through analysis of the Kyushu dataset and a public dataset. We showed that the expression of SF3B4 was upregulated in ESCC tissues compared with corresponding normal tissues. Moreover, elevated expression was associated with poor prognoses in ESCC, suggesting that SF3B4 may act as an oncogene in ESCC.
It is not clear why SF3B4 functions as an oncogene in ESCC. Shen et al. have reported that SF3B4 selectively modulated the epithelial-mesenchymal transition (EMT) by enhancing SNAI2 in HCC cells (21). EMT is significantly involved in carcinogenesis and progression of ESCC (26, 27). Our clinicopathological analysis revealed that SF3B4 expression was involved in lymphatic invasion and poor prognosis in ESCC. The molecular mechanism of lymphovascular invasion is closely related to EMT (28). Warzecha et al. have reported that changes in selective RNA splicing occur in specific genes in the EMT process as they switch from an epithelial system-specific isoform to a mesenchymal-specific isoform (29, 30).
SF3B4 triggers the production of alternative splicing variants of the tumor suppressor Kruppel-like factor 4 (KLF4), rendering its transcripts nonfunctional. It therefore promotes cell cycle progression through inactivation of p27Kip (21). Our GSEA study revealed that SF3B4 modulates the genes encoding cell cycle-related targets of E2F transcription factors. The E2F family is a target of Retinoblastoma (RB) proteins and plays an important role in cell cycle regulation, especially at the G1/S interphase (31). Thus, aberrant SF3B4 expression may regulate the cell cycle. We also showed that genes associated with the G2/M checkpoint were also associated with SF3B4 expression. Consistent with our previous report (22), a positive correlation between SF3B4 expression and DNA copy number was confirmed. In acute lymphoblastic leukemia and Burkitt's lymphoma, a duplicated copy of the chromosomal region containing SF3B4 was detected by FISH (32). Chromosomal aberrations may play a role in SF3B4 regulation. The detailed molecular mechanisms remain elusive and further analysis is required.
In conclusion, SF3B4 plays an important oncogenic role in the progression of ESCC. SF3B4 could be a therapeutic target, as well as novel prognostic factor in ESCC.
Acknowledgements
This work was supported by the following grants and foundations: Grants-in-Aid for Scientific Research of MEXT (24008081, 25430111, 25461953, 25861199, 25861200, 24592005 and 21229015); Funding Program for Next Generation World-Leading Researchers (LS094); Grants-in-Aid for Scientific Research on Innovative Areas of MEXT “Systems Cancer Research” (4201); The MEXT Strategic Programs on Innovative Research “Supercomputational Life Science.” This research used computational resources of the K computer provided by the RIKEN Advanced Institute for Computational Science through the HPCI System Research project (Project ID: hp140230). Computation time was also provided by the Supercomputer System, Human Genome Center, Institute of Medical Science, University of Tokyo (http://sc.hgc.jp/shirokane.html). The OITA Cancer Research Foundation 2014. Grant-in-Aid for Scientific Research of Ministry of Health, Labor and Welfare (MHLW) (14524362 and 14525288). Clinical samples and corresponding clinical information were provided by Oita Red Cross Hospital (Oita, Japan), Hiroshima Red Cross Hospital & Atomic-bomb Survivors Hospital (Hiroshima, Japan) and Iizuka Hospital (Fukuoka, Japan). The Authors appreciate the technical support of Ms. Kazumi Oda, Michiko Kasagi, Sachiko Sakuma and Tomoko Kawano.
Footnotes
Authors' Contributions
Shinya Kidogami: study concept and design, drafting of manuscript; Tomohiro Iguchi: study concept and critical revision of the manuscript; Kuniaki Sato: statistical analysis; Yukihiro Yoshikawa: data collection; Quingjang Hu: statistical analysis, Sho Nambara: experiment; Hisateru Komatsu: data collection; Masami Ueda: experiment; Yosuke Kuroda: data collection; Takaaki Masuda: study concept; Masaki Mori: critical revision of the manuscript; Yuichiro Doki; critical revision of the manuscript; Koshi Mimori: final approval of the manuscript.
Conflicts of Interest
The Authors declare no conflicts of interest regarding this study.
- Received March 23, 2020.
- Revision received April 6, 2020.
- Accepted April 7, 2020.
- Copyright© 2020, International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved