Abstract
Background: MicroRNAs (miRNAs) such as miR-17-5p, miR-21, miR-106a and miR-106b are reported to be highly expressed in gastric carcinoma (GC) tissues. Recently, we reported that these miRNAs were consistently detectable in plasma and reflected tumor dynamics of GC. We hypothesized that these plasma miRNA concentrations could be used as prognostic markers in patients with GC. Materials and Methods: Between 2008 and 2009, preoperative plasma samples were collected from 69 consecutive patients with GC at our hospital. We retrospectively examined the association between plasma miRNA concentrations and prognosis. Results: The postoperative cause-specific survival rate of patients with high plasma miR-21 concentration was significantly poorer than those with a low concentration (p=0.0451), as was that of those with high plasma concentration of miR-106a (p=0.1132). There were no prognostic differences according to the plasma concentration of miR-17-5p and miR-106b. Those with high miR-21 concentration had also a slightly higher incidence of vascular invasion (p=0.0311). Multivariate analysis revealed that the presence of a high miR-21 concentration in plasma was an independent prognostic factor (p=0.0133, hazard ratio: 13.4 (95% CI: 1.72-104.4)). Conclusion: The level of circulating miR-21 could be a reliable prognostic marker in the plasma of patients with GC. These findings contribute to the stratification of patients in order to identify those who need meticulous follow-up for early detection of recurrence and additional or alternative treatments of GC.
Recent advances in diagnostic methods, less invasive treatment techniques, and perioperative management have increased the early detection of gastric cancer (GC) and reduced the mortality and morbidity rates. As finding molecular targets for GC treatment might help improve survival, numerous studies have attempted to identify biological factors which confer a poor prognosis of GG (1). In the clinical setting, however, few molecules have been assayed as therapeutic or prognostic biomarkers.
MicroRNAs (miRNAs), which are small and non-coding RNAs, regulate the translation of specific protein-coding genes. Since their discovery in 1993 (2), altered expressions of miRNAs have been associated with several diseases and tumor miRNAs are involved in tumor genesis and the development of various types of cancer. Recently, several studies demonstrated that miRNAs are detectable in plasma/serum (3-6). Tumor-derived miRNAs are resistant to endogenous ribonuclease activity because these may be protein-bound, such as to the argonaut-2 protein and high-density lipoproteins (7, 8) or packaged by secretory particles, including apoptotic bodies and exosomes, in plasma/serum (9-12). Secretory miRNAs are up-regulated in the plasma of patients bearing tumors. Thus, detection and monitoring of tumors is now becoming possible by the evaluation of tumor-derived secretory miRNAs, although the secretory mechanisms and biological functions of extracellular miRNAs remain unclear.
Recently, we reported that miR-17-5p, miR-21, miR-106a and miR-106b, which are highly expressed in GC tissues (13-16), were consistently detectable in plasma and reflected tumor dynamics of GC (17). Specifically, the plasma concentrations of miR-17-5p, miR-21, miR-106a and miR-106b were significantly higher in patients with GC than in healthy volunteers. In each of these miRNAs, comparison between expressions in plasma and primary GC tissue samples showed similar tendencies. Moreover, the concentrations of these miRNAs in plasma were significantly reduced in postoperative samples compared to the levels in the pre-operative samples (17).
In this study, we assayed the expression of miR-17-5p, miR-21, miR-106a and miR-106b. Out of these, miR-21 detection in cancer tissues was reported to have a poor prognosis in patients with GC, and also in those with head and neck squamous cell carcinoma and digestive system cancer (18, 19). The overexpression of miR-17-5p was frequently detected in primary pancreatic cancer tissues and was significantly associated with poor overall survival (20). Some reports provided evidence that the plasma expression of several miRNAs was significantly associated with overall and/or disease-free survival in different types of cancer (21, 22). Therefore, this study was designed to examine that these plasma miRNAs concentrations could be used as prognostic markers in patients with GC.
Materials and Methods
Patients and samples. Between October 2008 and July 2010, 69 preoperative plasma samples were collected from consecutive patients with GC, treated at the Kyoto Prefectural University of Medicine. There were 39 patients with TNM stage I, 13 with stage II, 12 with stage III, and five with stage IV. After obtaining informed consent, 7 ml of peripheral blood was obtained from each patient before surgery. Immediately after collection, the blood samples were subjected to isolation of cell-free nucleic acids using a three-spin protocol (1500 rpm for 30 min, 3000 rpm for 5 min, 4500 rpm for 5 min) to prevent contamination by cellular nucleic acids. Plasma samples were then stored at -80°C until further processing. The resected GC specimens were fixed in buffered formalin and embedded in paraffin for pathological examination by standard methods. Macroscopic and microscopic classification of tumors was based on the UICC/TMN staging system (23). Institutional Review Board (IRB) approval was obtained for this study.
RNA extraction from plasma samples. Total RNA was extracted from 400 μl of plasma using mirVana PARIS Kit (Ambion, Austin, TX, USA), and finally eluted into 100 μl of pre-heated (95°C) Elution Solution according to the manufacturer's protocol.
Protocols for the detection of miRNAs. We selected four miRNAs, namely miR-17-5p, miR-21, miR-106a and miR-106b, which were found to be frequently expressed in GC tissues (13-16), as candidates for this plasma assay. The amounts of miRNAs were quantified in duplicate by qRT-PCR using human TaqMan MicroRNA Assay Kits (Applied Biosystems, Foster City, CA, USA). Reverse-transcription was carried out with the TaqMan MicroRNA Reverse Transcription Kit (Applied Biosystems) in 15 μl containing 5 μl of RNA extract, 0.15 μl of 100 mM dNTPs, 1 μl of Multiscribe Reverse Transcriptase (50 U/μl), 1.5 μl of 10×Reverse Transcription Buffer, 0.19 μl of RNase inhibitor (20 U/μl), 1 μl of gene-specific primer and 4.16 μl of nuclease-free water. For the synthesis of cDNA, the reaction mixtures were incubated at 16°C for 30 min, at 42°C for 30 min, at 85°C for 5 min and then held at 4°C. Then, 1.33 μl of the cDNA solution was amplified using 10 μl of TaqMan 2xUniversal PCR Master Mix with no AmpErase UNG (Applied Biosystems), 1 μl of gene-specific primers/probe, and 7.67 μl of nuclease-free water in a final volume of 20 μl. Quantitative PCR was run on a 7300 Real-time PCR system (Applied Biosystems) and the reaction mixtures were incubated at 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. The cycle threshold (Ct) values were calculated with SDS 1.4 software (Applied Biosystems). The amounts of plasma miRNAs were calculated from a standard curve constructed using synthetic microRNAs, mirVana miRNA Reference Panel (Ambion). The standard reference miRNAs were amplified for each reaction.
Statistical analysis. Mann-Whitney U-test was performed to examine the association of each plasma miRNA concentration with clinicopathological factors in GC. Cause-specific death was recorded when the cause of death was specified as recurrent GC. The cumulative cause-specific overall survival rates were calculated using the Kaplan Meier method, and log-rank test was used for assessment of differences between clinical factors. Multivariate analysis using Cox regression model was performed to identify significant contributors that were associated independently with death on univariate analysis; multivariate risk ratios are presented with 95% confidence intervals (CI). A p-value less than 0.05 was considered significant.
Results
Evaluation of prognostic impact of the miR-17-5p, miR-21, miR-106a and miR-106b in plasma of patients with GC. The cause-specific survival curves according to plasma miR-17-5p, miR-21, miR-106a and miR-106b concentrations are shown in Figure 1. In order to dichotomize the data for plasma miRNA concentrations into high and low expression groups, we used the cut-off value of the median concentration of the miRNA in plasma of healthy volunteers, which we previously reported (data not shown) (17). The cut-off concentrations for miR-17-5p, miR-21, miR-106a and miR-106b were 0.0484, 0.0326, 0.0144, and 0.0491 amol/μl, respectively.
As a result, patients with a high miR-21 concentration in plasma presented a significantly poorer prognosis than those with a low miR-21 concentration (3-year survive rate: high vs. low: 64.9% and 85.9 %, p=0.0451). The survival rate of patients with a high plasma miR-106a concentration tended to be poorer than that of the low concentration group (3-year survival rate: high vs. low: 66.0% and 91.7%, p=0.1132), while there were no prognostic differences according to the plasma concentration of miR-17-5p (3-year survival rate: high vs. low: 81.0% and 66.6%, p=0.4961) or miR-106b (3-yearr survival rate: high vs. low: 83.1% and 68.2%, p=0.6444).
Association between plasma miR-21 concentration and clinicopathological features in patients with GC. We examined the association between the plasma concentrations of miR-21 and clinicopathological factors (Table I). As a result, patients with high plasma miR-21 concentration more frequently had a vascular invasion (p=0.0311). However, there was no significant correlation between plasma miR-21 concentrations and other clinicopathological factors, such as sex, age, the presence of lymphatic invasion, the extent of lymph node metastasis, depth of tumor invasion, pathological stage or recurrence including distant metastasis.
Evaluation of prognostic impact of the miR-17-5p, miR-21, miR-106a and miR-106b in plasma of patients with gastric cancer. The cause-specific survival curves according to plasma miR-17-5p (A), miR-21(B), miR-106a (C) and miR-106b (D) concentrations are shown. There were no prognostic differences according to the plasma concentration of miR-17-5p and miR-106b (A, D). However, the postoperative cause-specific survival rate of patients with a high plasma miR-21 concentration was significantly poorer than that of those with a low concentration (p=0.0451) (B), as tended to be the rate of patients with high plasma miR-106a concentration (p=0.1132) (C).
Investigation of the potential utility of miR-21 as a prognostic biomarker in the plasma of patients with GC. Table II shows univariate and multivariate analyses for cause-specific survival in 69 consecutive patients with GC. Univariate analysis showed that lymphatic invasion, venous invasion, the extent of lymph node metastasis, depth of tumor invasion and the plasma miR-21 concentration were significant prognostic factors. Multivariate analysis using the Cox regression procedures revealed that the presence of high miR-21 concentration in plasma was an independent factor predicting a poor prognosis (p=0.0133, hazard ratio: 13.4 (95% CI: 1.72-104.4)).
Discussion
Non-coding RNAs, so-called miRNAs, have been demonstrated to regulate gene expression by targeting mRNAs for translational repression or cleavage. miRNAs have been proven to contribute to carcinogenesis and the progression of various carcinomas (3, 24-26), and consequently miRNAs have recently become known as new factors in oncogenesis and have utility as biomarkers and therapeutic targets for different type of cancer. In fact, accumulating reports also suggest the potential role of miRNAs in the early detection of patients with several malignancies (3, 24-26).
For many decades, cell-free nucleic acids have been known to be present in peripheral blood. Several studies have identified tumor-specific and/or tumor-associated alterations in the circulating nucleic acids of patients with various types of cancer. In particular, plasma/serum miRNAs can be present in a remarkably stable form (3-6) and the expression level of serum miRNAs is reproducible and consistent among individuals (5, 9). Many studies have demonstrated the presence of circulating miRNAs and their potential use as novel biomarkers of malignancy, such as leukemia (27), and prostate (9), oral (28), pancreatic (29), colorectal (30), ovarian (31), lung (22) and breast (32), including our reports in esophageal (33), pancreatic (34) and gastric cancer (17, 35). To our knowledge, however, there have been no the reports on the prognostic value of circulating miRNAs in the plasma of patients with GC.
Correlation between plasma miR-21 concentration and clinicopathological features of patients.
Recently, we reported that circulating miR-17-5p, miR-21, miR-106a and miR-106b were consistently detectable in plasma and reflected tumor dynamics of GC (17). However, we were unable to confirm the prognostic value of these miRNAs because the follow-up periods were short. In the present study, we confirmed that the high miR-21 concentration in plasma reflected a significantly poorer prognosis (Figure 1). We then showed that a high miR-21 concentration in plasma was significantly correlated with venous invasion. We also reported a similar tendency for plasma miR-21 expression of patients with esophageal cancer (33). Furthermore, the presence of high miR-21 concentration in plasma was demonstrated to be an independent prognostic factor (Table II). These findings highlight the potential utility of miR-21 as a prognostic biomarker. Its use could also facilitate clinical decision-making to select for prospective patients needing meticulous follow-up for early detection of recurrence and additional or alternative treatments, such as postoperative chemotherapy or neo-adjuvant chemotherapy of GC.
Univariate and multivariate analyses for cause-specific survival of patients with gastric cancer.
Concerning the other miRNAs, miR-17-5p, miR-106a and miR-106b, we found no prognostic value in GC. The miR-106a expression in cancer tissues and its prognostic value is different among cancer types. miR-106a expression in GC and breast cancer is higher than that in normal tissue (16, 36). In contrast, miR-106a expression in brain tumor is low and its low expression is associated with a poor prognosis (37). In this study, high plasma miR-106a tended to be associated with a poorer prognosis (p=0.1132). This finding might indicate the putative oncogenic role of miR-106a in GC. However, detailed analyses in vitro and studies with large numbers of patients with long follow-up periods are needed. Regarding miR-17-5p and miR-106b, these miRNAs have already been reported to have an oncogenic role in various types of cancer (38, 39). Nonetheless, our study did not demonstrate their prognostic value in GC. However, the 3-year cause-specific survival of patients with a high concentration of these miRNAs was poorer the one of patients those with low concentrations (Figure 1). These findings might also warrant further study of these two miRNAs.
We previously found that a tumor suppressor, let-7a, had a lower plasma expression level in patients with GC, contrary to our expectation (17). As shown in another of our studies, the concentration of miR-375 was significantly lower in patients with esophageal cancer than in healthy volunteers (33). Since circulating miRNAs are considered to be released from normal tissues as well as from cancer tissues, the majority of these miRNAs are expected to have originated from normal tissues. Our findings on the expression of some tumor suppressor miRNAs in patients with cancers seemed strange from this perspective. Kosaka and colleagues proposed a novel hypothesis that miRNAs might be involved in the maintenance of the surveillance system against cancer progression (12). During the initial stage of tumorigenesis, down-regulation of miRNA expression in cancer cells is compensated for by surrounding cells, with supply exosomes containing such miRNAs. However, once the surrounding cells can no longer meet this demand, cancer cells enter an advanced stage of development. Interestingly, this theory could explain our clinical findings in patients with esophageal cancer because patients with high expression of tumor-suppressive miRNA, such as miR-375, in plasma had a better prognosis than those with depleted levels (40). Unfortunately, this theory did not appear to apply to our data of patients with GC. Namely, there were no significant prognostic differences between high and low let-7a patients (p=0.6427) (data not shown). However, many factors, such as miRNA species, cancer type, and follow-up periods might affect prognostic outcomes.
The level of circulating miR-21 could play a pivotal role as a prognostic biomarker. Our results have provided evidence that the plasma levels of miR-21 might contribute to the stratification of GC patients.
- Received October 13, 2012.
- Revision received November 6, 2012.
- Accepted November 7, 2012.
- Copyright© 2013 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
