Introduction
Although the incidence of gastric cancer (GC) is declining, it is the fifth most common cancer after cancers of the lung, breast, colon and prostate, and GC is the third leading cause of cancer deaths (1–3). Although palliative chemotherapy and surgery are beneficial, the outcome of GC remains dismal (4–6). The high mortality of GC is explained by our poor understanding of its mechanism of progression and the lack of suitable diagnostic markers that hinder diagnosis before the disease reaches an advanced stage (7,8). GC represents a biologically and genetically heterogeneous group of tumours that are induced by multiple factors that deregulate cell signalling pathways, which leads to the acquisition of malignant phenotypes such as increased cell proliferation, inhibition of apoptosis and enhanced invasiveness (9–11). Identification of novel molecules is therefore required to improve diagnosis and therapy.
Protein arginine methyltransferase 5 (PRMT5) catalyses the symmetrical dimethylation of arginine residues of histone and non-histone substrates (12–14). Because PRMT5 is implicated in diverse cellular and biological processes such as transcriptional regulation, RNA metabolism, ribosome biogenesis, Golgi apparatus homeostasis and cell cycle regulation, PRMT5 is generating increased interest in the field of cancer research (12,15,16). For example, PRMT5 is expressed at higher levels compared with normal tissues in leukemias, lymphomas and gliomas as well as in ovarian, breast, prostate and lung cancers (13,14,17–20). In contrast, the expression levels of PRMT5 in gastroenterological malignancies and the biological function of PRMT5 remain to be determined.
To address these questions, we hypothesized that PRMT5 is linked to the malignant phenotype of GC. The aim of the present study was to determine whether PRMT5 acts as an oncogene in GC. This study provides insight into the contribution of PRMT5 to the malignant behavior of GC, and provides a rationale for targeting this enzyme in GC.
Materials and methods
Sample collection
Ten GC cell lines (AGS, KATOIII, MKN1, MKN45, MKN74, N87, NUGC2, NUGC3, NUGC4 and SC-6-JCK) and the non-tumourigenic epithelial cell line (FHs74) were obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA), Tohoku University, Japan, or were established at our institute. Cell lines were cultured at 37°C in Dulbecco’s modified Eagle’s medium (DMEM; Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS) in an atmosphere containing 5% CO2 (21). Primary GC tissues and the corresponding non-cancerous adjacent tissues were collected from 179 patients who underwent gastric resection for GC without neoadjuvant therapy at Nagoya University Hospital between 2001 and 2010. The tissue samples were immediately frozen in liquid nitrogen and stored at −80°C. The specimens were classified histologically according to the 7th edition of the Union for International Cancer Control (UICC) classification system. Since 2006, adjuvant chemotherapy using S-1 (an oral fluorinated pyrimidine) is used to treat all patients with UICC stage II/III GC unless contraindicated by the patient’s condition (22–24). This study conformed to the ethical guidelines of the World Medical Association Declaration of Helsinki, Ethical Principles for Medical Research Involving Human Subjects. Written informed consent for the use of clinical samples and data, as required by the institutional review board at Nagoya University, Japan, was obtained from all patients.
Analysis of PRMT5 mRNA expression
PRMT5 mRNA levels were determined using a quantitative real-time reverse-transcription polymerase chain reaction (qRT-PCR) assay. Total RNAs (10 μg per sample), which were isolated from 12 cell lines and 179 primary GC tissues as well as the corresponding non-cancerous adjacent tissues, were used to generate cDNAs. The cDNAs were amplified using PCR primers specific for PRMT5 as follows: sense 5′-TCTCATGGTTTCCCATCCTC-3′ in exon 16 and antisense 5′-CCTTCTTGGAATTGCTGCAT-3′ in exon 17, which amplify a 102-bp product. The RT-PCR amplification reaction was performed as follows: initial denaturation at 95°C for 10 min, 40 cycles at 95°C for 10 sec and at 60°C for 30 sec. All samples were tested in triplicate, and samples without template were included in each PCR plate as negative controls. A SYBR-Green PCR Core reagents kit (Applied Biosystems, Foster City, CA, USA) was used to perform qRT-PCR, and real-time detection of the SYBR-Green fluorescence emission intensity was conducted using an ABI StepOnePlus Real-Time PCR system (Applied Biosystems). The levels of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA were quantified in each sample for standardization. The expression level of each sample was calculated as the value of PRMT5 mRNA divided by that of GAPDH mRNA (25).
PCR array analysis of differential gene expression
To analyse gene expression by 10 GC cell lines and the FHs74 cell line, we used the Human Epithelial to Mesenchymal Transition (EMT) RT2 Profiler PCR Array (Qiagen, Hilden, Germany) of 84 key genes, including those that encode the proteins as follows: transcription factors, extracellular matrix proteins as well as proteins involved in the EMT, cell differentiation, morphogenesis, growth, proliferation, migration, cytoskeleton and major signalling pathways (26).
Knockdown of PRMT5 expression using a small interfering RNA (siRNA)
NUGC3 and AGS cells were cultured in a 24-well plate (5×104 cells ml−1). Cells were transiently trans-fected the next day with either 30 nM of an siRNA specific for PRMT5 (siPRMT5; 5′-CAGCCACUGAUGGACAAUCUGGAAU-3′ and 5′-CCGGCUACUUUGAGACUGUGCUUUA-3′) or a control siRNA (siControl) using LipoTrust EX Oligo (Hokkaido System Science, Sapporo, Japan). After transfection, cells were cultured in serum-free DMEM for 72 h and used in the western blot and functional analyses. Western blot analysis using a rabbit anti-PRMT5 polyclonal antibody (Cell Signaling Technology, Beverly, MA, USA) diluted 1:1,000 was performed as previously described (27).
Cell proliferation assay
Cell proliferation was evaluated using the Premix WST-1 Cell Proliferation assay system (Takara Bio Inc., Shiga, Japan). NUGC3 and AGS cells (5×103 cells/well) were seeded into 96-well plates in DMEM supplemented with 2% FBS for 7 days. The optical density of the solution in each well was measured on days 1, 3 and 5 for NUGC3 cell, and 1, 3, 5 and 7 for AGS cell following addition of 10 μl of WST-1. In addition, the % decrease by siPRMT5 in proliferation was calculated as [1 − (fold change of siPRMT5/fold change of untransfected)] × 100 on every measurement day.
Cell invasion assay
The ability of GC cells to invade Matrigel was determined using BioCoat Matrigel invasion chambers (BD Biosciences, Bedford, MA, USA) according to the manufacturer’s protocol. NUGC3 and AGS cells (2.5×104 cells/well) were seeded into the upper well of the chamber in serum-free DMEM. After 48 h, cells on the lower surface of the membrane were fixed, stained and counted using a microscope (x200 magnification, five randomly selected fields).
Wound-healing assay
The migration of GC cells was evaluated using wound-healing assays. NUGC3 and AGS cells (2×104 cells/well) were seeded into 12-well plates in serum-free DMEM using the ibidi Culture insert method (ibidi GmbH, Martinsried, Germany) to establish wound gaps of a defined width. After 24 h, the insert was removed, and the width of the wound was measured at 200-μm intervals (10 per well, x40 magnification) according to cell-dependent time intervals.
Statistical analysis
The Chi-square and Mann-Whitney tests were used to compare the differences between the two groups. The significance of a correlation between two variables was assessed using Spearman’s rank correlation coefficient. Risk factors for positive peritoneal lavage cytology were evaluated using binomial logistic analysis. Overall survival rates were calculated using the Kaplan-Meier method, and the difference in survival curves was analysed using the log-rank test. We performed multivariable regression analysis to detect prognostic factors using the Cox proportional hazards model, and variables with a P-value of <0.05 were entered into the final model. All statistical analyses were performed using JMP v.10 software (SAS Institute, Inc., Cary, NC, USA). A P-value of <0.05 was considered statistically significant.
Results
Differential gene expression by GC cell lines
Expression levels of PRMT5 mRNA were heterogeneous among GC cell lines. The levels of PRMT5 mRNA were >2-fold higher in AGS, KATOIII, MKN1, MKN45 and NUGC3 cells compared with the control FHs74 cells, whereas N87 cells showed lower PRMT5 expression level than FHs74 cells (Fig. 1A). PRMT5 mRNA levels did not differ according to the extent of differentiation of the GC cells. PCR array analysis revealed that mRNAs encoding gem (nuclear organelle) associated protein 2 (GEMIN2), signal transducer and activator of transcription 3 (STAT3) and transforming growth factor beta 3 (TGFB3) were expressed at levels that correlated significantly with those of the mRNA encoding PRMT5 (Fig. 1B).
Effect of PRMT5 knockdown on the malignant phenotype of GC cells
Inhibition of PRMT5 expression using a specific siRNA was conducted to evaluate the function of PRMT5 in GC cells. NUGC3 (undifferentiated) and AGS (differentiated) cells were selected as PRMT5-overexpressed cells from the qRT-PCR results. The effect of PRMT5 knockdown was confirmed using western blotting assay (Fig. 2A). We evaluated the proliferation, invasion and migration of NUGC3 and AGS cells. Knockdown of PRMT5 expression significantly decreased the proliferation ability of NUGC3 cell with 82 and 83% decrease on day 1 and 3, respectively (Fig. 2B). Invasion and migration abilities of NUGC3 cells were also reduced by inhibition of PRMT5 compared with the untransfected and siControl-transfected cells (Fig. 2B). Similarly, the proliferation ability of AGS cells was significantly decreased by PRMT5 knockdown with 23 and 27% decrease on day 5 and 7, respectively (Fig. 3). The invasion and migration of AGS cells were also significantly decreased after inhibition of PRMT5 expression (Fig. 3).
Clinical implications of PRMT5 expression in tumour tissues
The mean level of PRMT5 mRNA was significantly higher in 179 GC tissues compared with those of the corresponding adjacent normal tissues (Fig. 4A). Patients were classified into high and low PRMT5 expression groups according to the median value of PRMT5 mRNA levels. PRMT5 mRNA levels were independent of tumour depth, differentiation and lymph node metastasis, whereas they were significantly and specifically associated with peritoneal lavage cytology (Table I). To investigate further the relationship between high PRMT5 mRNA levels in GC tissues and positive peritoneal lavage cytology, multivariate binomial logistic analysis was conducted and revealed that high PRMT5 mRNA levels were significantly associated with positive peritoneal lavage cytology (odds ratio, 3.90, 95% confidence interval 1.59–10.2, P=0.003; Table II). Patients with high levels of PRMT5 mRNA (n=90) were more likely to survive for shorter times compared with those with low levels (n=89), and their 5-year survival rates were 51 and 60%, respectively (P=0.019; Fig. 4B). In the multivariate analysis, lymph node metastasis and positive lavage cytology were identified as independent prognostic factors, but PRMT5 expression was not (hazard ratio 1.54, 95, confidence interval 0.94–2.54, P=0.086).
Discussion
Arginine methylation is an important posttranslational modification of nuclear and cytoplasmic proteins and plays a vital role in cellular function (14,28). The human genome encodes 11 PRMT isoforms that covalently modify arginine residues in histone and nonhistone proteins that contribute to diverse cellular regulatory networks (20). Types I and II PRMTs catalyse monomethylation at the ω-NH2 group of arginine; however, they differ in their ability to add the second methyl group, either asymmetrically (type I) or symmetrically (type II) (13,14,29). PRMT5 is a type II PRMT that catalyses the transfer of methyl groups from S-adenosyl methionine to the arginine residues of histones or non-histone proteins and is involved in numerous cellular processes (19,30). Because PRMT5 possesses multiple cellular functions, it is an important determinant of oncological properties of various malignancies (12–14,17,18). In the present study, analyses of the expression levels of PRMT5 and their effects on the phenotypes of GC cell lines, patient characteristics and outcomes were performed to assess the potential of PRMT5 as a novel biomarker and therapeutic target for patients with GC.
Ibrahim et al (18) found that PRMT5 is likely involved in the EMT of lung adenocarcinomas; however, the roles of PRMT5 in the EMT are unknown. In the present study, we conducted PCR array analysis to evaluate the involvement of PRMT5 in the oncological processes, particularly in the EMT that occurs in GC cells. GEMIN2, STAT3 and TGFB3 were identified as genes that were expressed in concert with PRMT5. GEMIN2 (synonym, SIP1) encodes a zinc-finger transcription factor targeting the E2-box on the E-cadherin promoter and acts as a direct transcriptional repressor of E-cadherin (31,32). GEMIN2 acts downstream in EMT-inducing signal transduction pathways activated by growth factors as well as in integrin engagement and hypoxia (33). STAT3 was discovered as a latent transcription factor that is robustly activated by interleukin-6 and epidermal growth factor (34,35). Numerous studies indicate that STAT3 modulates the expression of important EMT transcription factors that integrate signals generated by multiple extracellular stimuli that influence EMT phenotypes (34–36). Mammalian TGF isoforms (TGF-β1, TGF-β2 and TGF-β3) share 97% amino acid sequence identity and signal through activation of TGF-β receptors (28). TGF-β isoforms play major roles in tumourigenesis mediated by the EMT through regulating cell growth, migration, invasion and metastasis (28,37,38). Although pathway analyses should be mandatory to further identify the involvement PRMT5 in EMT process, our present results might suggest that PRMT5 partially participate in EMT programs through its coordinate expression with other EMT-inducing molecules in GC cells.
There are no reports, to the best of our knowledge, that identify the function of PRMT5 in the malignant phenotypes of GC. We show here that PRMT5 contributed to the cell proliferation in addition to the migration and invasion abilities of GC cell lines. In NUGC3 cells, a significant decrease of proliferation ability by knockdown of siPRMT5 was shown on day 3, whereas the differences in proliferation ability were exhibited from day 5 in the AGS cells. The possible explanations of this time lag were differences in cell-dependent efficacy, quickness and duration of action of siRNA transfection. Although analyses of apoptosis, cell cycle and interaction with the components of intracellular signalling pathways may contribute to understanding the functions of PRMT5 in the malignant phenotypes of GC, our findings support the hypothesis that PRMT5 acts as an oncogene when expressed at high levels and serves as a candidate target for the therapy of GC. To verify our in vitro results, we evaluated the expression of PRMT5 and its clinical implications using surgically resected gastric tissues. PRMT5 was significantly overexpressed in GC tissues compared with the corresponding adjacent normal tissues. Further, patients with high levels of PRMT5 mRNA were more likely to survive for shorter times compared with those without.
We were intrigued by our finding here that PRMT5 expression in patients’ GC tissues was strongly associated with positive peritoneal lavage cytology. Peritoneal metastasis is a most frequent and dismal condition in patients with advanced GC and is diagnosed only when macroscopic disseminated nodules are found during staging laparoscopy or from findings of positive peritoneal lavage cytology (6,39). Therefore, the availability of specific biomarkers that predict or enable early detection of peritoneal metastasis is anticipated to enhance management of GC. As an important step toward achieving this goal, we show here that the level of PRMT5 mRNA in GC tissues is an independent risk factor for positive peritoneal lavage cytology. Linitis plastica, serosal invasion (T4) and lymph node metastasis have been reported to be risk factors for peritoneal metastasis of gastric cancer (40–42). In this study, high PRMT5 expression was strongly associated with positive cytology, but independent of these well-known risk factors. This reveals the unique predictive value of PRMT5 for peritoneal metastasis and indicates that GC patients with increased expression of PRMT5 can be categorized into a high-risk group with peritoneal metastasis regardless of tumor types, T and N status. The value of PRMT5 expression as a tool to screen for peritoneal metastasis will be enhanced by the development of assays to detect PRMT5 expression in serum and peritoneal fluid.
The present study has certain limitations. First, extensive expression analyses of proteins, particularly ones related to EMT process, that potentially interact with PRMT5 as well as apoptosis assays must be conducted to further understand the biological functions of PRMT5 in GC. Second, this study was limited by the relatively small sample size. To determine the usefulness of PRMT5 expression as a biomarker for GC, analysis of a larger cohort using multiple clinical samples, such as gastric tissues, ascites fluids and serum samples, will be required to deepen our knowledge on clinical significance of PRMT5.
Taken together, our findings indicate that PRMT5 acts as an oncogene in GC by enhancing the malignant phenotype of a cancer cell line. PRMT5 expression in gastric tissues may represent a promising biomarker for identification of patients at high risk, particularly for peritoneal metastasis.