Abstract
Background: Primary tumor spread to the liver is the major cause of disease progression and death in patients with colorectal cancer (CRC). MicroRNAs (miRNAs) are small non-coding RNAs that are involved in cancer development and progression, but their role in metastasis has not been extensively investigated. Materials and Methods: Firstly, expression profiling of 752 miRNAs in 20 primary tumors and their corresponding liver metastases was performed. Secondly, validation of the results was carried out on an independent cohort of 66 patients with metastatic CRC using reverse transcription–quantitative polymerase chain (RT-qPCR) reaction. Results: In total, 33 miRNAs were found to be significantly deregulated in liver metastases compared to their primary tumors. Fifteen miRNAs were chosen for subsequent validation, which confirmed significantly reduced expression of miR-143, miR 10b, and miR-28-5p, and increased expression of miR 122, miR-122*, and miR 885-5p in the tissue of liver metastases. Conclusion: These results indicate that miRNAs could serve as new therapeutic targets in patients with metastatic CRC.
Colorectal cancer (CRC) is the third most common malignancy and the fourth leading cause of cancer-related deaths worldwide. Despite improvements in diagnosis and treatment, more than 50% of patients diagnosed with CRC will eventually die from their disease. Approximately 90% of CRC-related mortalities are caused by direct or indirect effects of metastatic dissemination (1). Metastasis is a multistep process of cell spread from the primary tumor to non-adjacent organs and the formation of a secondary tumor. Although cancer cell dissemination has been considered a late step during tumor development, recent evidence suggests that it could be an earlier event that manifests only several years from diagnosis (2). Nevertheless, each step in metastasis requires specific genetic and epigenetic changes that are largely unknown. Therefore, it is of critical importance to understand the key molecular switches involved in the metastatic process of CRC and identify new biomarkers enabling prediction of prognosis and treatment outcome, as well as the metastasizing propensity of the primary tumor.
Multiple studies have been conducted to investigate the role of genes and their products in metastasis (3). Recently, small non-coding RNAs called microRNAs (miRNAs) were found to be attractive candidates as novel biomarkers for different types of cancer as they display specific expression patterns and can be detected in tissues, as well as in different body fluids (4, 5). They can function as tumor suppressors as well as oncogenes and are involved in multiple biological processes, including cell cycle, proliferation, differentiation, and apoptosis (6). They have also emerged as important regulators of the metastatic process (7); however, CRC metastasis-related miRNAs and their biological roles remain to be identified.
Currently, expression profiling of miRNAs in primary tumors is predominantly being performed. However, due to high genomic instability, the genomic make-up of primary tumors and their metastases may change over time. Moreover, specific clones with distinct molecular characteristics may undergo selection due to the adaptation of metastasized tumor cells to their specific microenvironment. Therefore, we systematically analyzed the expression of miRNAs in resected liver metastases and their corresponding primary tumors in order to identify those associated with CRC metastasis.
Materials and Methods
Patients and tissue samples. In total, 86 samples of primary tumors and 86 samples of corresponding liver metastases from patients with pathologically verified metastatic CRC (54 males, 32 females) who had undergone resection from April 1996 through March 2010 at the Department of Surgery (University Hospital in Pilsen, Czech Republic) were used. All patients were of the same ethnicity (European descent). The ages of patients ranged between 29 and 80 years, with a median of 63 years. Enrolled patients did not receive any neoadjuvant treatment. Written informed consent was obtained from all patients and the study was approved by the local Ethical Board at University Hospital in Pilsen.
Extraction of microRNAs. Isolation of total RNA enriched for small RNAs was performed using formalin-fixed paraffin embedded (FFPE) samples. Paraffin sections 4-μm thick were stained with hematoxylin and eosin, then microscopically verified by a pathologist and examined to identify sites with cancer cells suitable for macrodissection. All samples were deparaffinized, then treated with DNAse I and proteinase K, and RNA was extracted from 10-μm sections of tissue using miRNeasy FFPE Kit (Qiagen, Hilden, Germany) as described previously (8). Concentration and purity of RNA were determined spectrophotometrically by measuring optical density (A260/280 >2.0, A260/230 >1.8) using Nanodrop ND-1000 (Thermo Fisher Scientific, Waltham, MA, USA).
Expression profiling of microRNAs. Expression profiling of miRNAs was performed using microRNA Ready to Use PCR Panels (Exiqon, Vedbaek, Denmark). A set of two cards (Human Panel I+II, V3.R) enabling quantification of 752 human miRNAs and six endogenous controls for data normalization was used. First-strand cDNA synthesis was performed according to the standard protocol using Universal cDNA Synthesis Kit (Exiqon) and 40 ng of total RNA. Subsequently, real-time PCR amplification was carried out using ExiLENT SYBR Green master mix and PCR amplification primers that are microRNA-specific and optimized with LNA™ (Exiqon). The final measurement was performed using a Roche LC480 instrument (Roche, Basel, Switzerland) and initial data were analyzed using the second-derivative method.
Real-time quantification of microRNAs. Complementary DNA was synthesized from total RNA according to the TaqMan MicroRNA Assay protocol (Applied Biosystems, Foster City, CA, USA) using a T100™ Thermal Cycler (Bio-Rad, Hercules, CA, USA). Real-time PCR was performed according to the standard protocol using the TaqMan MicroRNA Assay kit and the Applied Biosystems 7500 Sequence Detection System (both Applied Biosystems).
Data normalization and statistical analysis. The threshold cycle data were calculated by SDS 2.0.1 software (Applied Biosystems). All real-time PCR reactions were run in triplicates. The average expression levels of all measured miRNAs were normalized using miR-103 (Assay No. 000439; Applied Biosystems) and subsequently analyzed by the 2−ΔCt method. miR-103 was selected as an endogenous control through combination of standard geneNorm and NormFinder algorithms from six possible genes included on the microRNA Ready-to-Use PCR Panels. Statistical differences between the levels of analyzed miRNAs in primary tumor and corresponding metastatic tissue were evaluated by two-tailed non-parametric Wilcoxon test for paired samples. All calculations were performed using GraphPad Prism version 5.00 (GraphPad Software, San Diego, CA, USA). p-Values of less than 0.05 were considered statistically significant.
MicroRNAs differentially expressed in liver metastases compared to corresponding primary tumors (miRNAs in bold were selected for the validation phase of the study).
Results
Metastasis-associated miRNA signature from matched primary tumors and liver metastases. To identify miRNAs whose deregulated expression is associated with metastatic disease, the expression profiles of 752 miRNAs were analyzed in tissues of 20 patients with metastatic CRC using total RNA isolated from primary tumors and corresponding liver metastases. Unsupervised hierarchical clustering was carried out and only miRNAs with differences in expression with p<0.0001 were considered statistically significant. Applying this criterion, a set of 33 miRNAs differentially expressed between primary tumors and corresponding metastatic tissues was gained using the package of linear models for microarray data analysis (Table I). From this set, 22 miRNAs were found to be significantly up-regulated in metastatic tissue, while 11 miRNAs were significantly down-regulated (p<0.0001; Figure 1).
Hierarchical clustering of 20 patients with metastatic colorectal cancer stratified according to the expression profile of 33 miRNAs differentially expressed between primary tumors (blue) and liver metastases (yellow) (p<0.0001).
Validation of selected miRNAs by quantitative real-time PCR. Based on the significance of the difference (fold change, p-value), previous observations, biological plausibility (according to the putative targets and PubMed hits when a particular miRNA was combined with the keyword ‘cancer’), and favorable expression levels (Ct value <32), 15 miRNAs (miR-143, miR-34a, miR-22, miR-22*, miR-122, miR-122*, miR-885-5p, miR-10b, miR-708, miR 361-3p, miR-28, miR-338-3p, miR-424, miR-365, and miR-193b*) were chosen for further validation on an independent cohort of 66 patients with metastatic CRC by RT-qPCR. The data were normalized using miR-103, which was selected as being the best endogenous control through combination of standard geneNorm and NormFinder algorithms from six possible genes (miR-103, miR-191, miR 423-5p, small nucleolar RNA, C/D box 49A (SNORD49A), small nucleolar RNA, C/D box 38B (SNORD38B), and U6B small nuclear RNA (RNU6B) included on microRNA Ready-to-Use PCR Panels. Using two-tailed non-parametric Wilcoxon test for paired samples, decreased expression levels of miR-143 (p<0.0001; Figure 2A), miR-10b (p<0.0001; Figure 2B), and miR 28 5p (p<0.0001; Figure 2C), and increased expression of miR-122 (p<0.0001; Figure 2D), miR-885-5p (p<0.0001; Figure 2E), and miR-122* (p=0.0001; Figure 2F) were confirmed in tissues of liver metastases compared to corresponding primary tumors. The expression of further validated miRNAs (miR-34a, miR-22, miR-22*, miR-708, miR-361-3p, miR-338-3p, miR-424, miR-365, and miR 193b*) showed them not to be significantly deregulated in metastatic tissue (p>0.05).
Normalized expression of validated miRNAs in primary tumors and liver metastases. Expression of miR-143 (A), miR-10b (B), and miR-28-5p (C) was significantly reduced and that of miR-122 (D), miR-885-5p (E) and miR-122* (F) was significantly increased in metastatic tissue compared to primary tumors.
Discussion
Primary tumor spread to the liver is the major cause of disease progression in patients with CRC and the most common cause of death that may occur through organ damage caused by growing lesions, treatment complications or paraneoplastic syndrome (9). Over the past few years, molecular genetic studies have been carried out to identify genes and their products responsible for the metastatic process, which can be characterized into several specific tumor cell properties – loss of cellular adhesion, increased invasiveness, intravasation, survival in the vascular system, extravasation, and survival and proliferation at a new site (10). However, due to the heterogeneity of this process, the role of specific genes in the onset of metastases and their diagnostic and prognostic value are still limited.
Recently, gene-expression analyses based on genome-wide profiling has been widely used to characterize human cancer, and comparison of tumor tissues with the corresponding normal tissues revealed significant deregulation of miRNA expression in different types of cancer, including CRC (11-13), strongly suggesting an important role of these molecules in cancerogenesis. Moreover, several studies showed that an miRNA signature can be used to determine the clinical and biological features of the tumors, as well as their metastatic potential (14-16).
The aim of this study was to identify a global miRNA pattern for patients with metastatic CRC that would enable differentiation between primary tumors and their corresponding liver metastases. Using genome-wide expression profiling, 33 miRNAs with significantly deregulated expression in metastatic tissue were found, of which 15 were further validated by RT-qPCR. It was confirmed that expression of miR-122, miR-122*, and miR-885-5p was significantly higher in metastatic tissue compared to primary tumors, whereas expression of miR-143, miR-10b, and miR-28-5p was significantly reduced in liver metastases.
miR-143 is an important tumor suppressor and one of the most studied miRNAs in CRC pathogenesis. Its decreased expression in tumor tissue was first described in 2003 by Michael et al. (17). Several targets of miR-143 have since been identified, including important oncogenes such as Kirsten rat sarcoma viral oncogene homolog (KRAS) (18), DNA methyltransferase 3A (DNMT3A) (19), metastasis-associated in colon cancer-1 (MACC1) (20), and toll-like receptor 2 (TLR2) (21). MACC1 has been described as a key activator of the metastasis inducing hepatocyte growth factor receptor signaling pathway promoting proliferation, invasiveness and migration of CRC cells. Zhang et al. showed that the expression level of miR-143 significantly correlated with the expression of MACC1 and restoring expression of miR-143 in SW620 cells led to inhibition of cell proliferation, migration and invasion (20). Besides MACC1, TLR2 was recently reported to be involved in CRC pathogenesis. Knockdown of expression of this protein significantly inhibited the invasive and migratory activities of SW620 and HCT116 cells; in addition, using western blot analyses and luciferase assay, TLR2 was shown to be a direct target of miR-143 (21). These data indicate the essential involvement of miR-143 in the metastatic process.
Recently, strict classification of cancer-associated miRNAs as oncogenes or tumor suppressors was shown to be oversimplified (22). As demonstrated in the case of miR-10b, this miRNA is strongly associated with migration and invasion in esophageal cancer (23) and breast cancer (24) cell lines, thus suggesting its oncogenic character. However, different results were recently reported in CRC. Firstly, down-regulation of miR-10b was found not only in tumor tissue compared to healthy tissue but also in metastatic tissue compared to tumor tissue (25). Similarly, Hur et al. generated a metastasis-specific miRNA profile for patients with CRC and found decreased levels of this miRNA in liver metastases compared to primary tumors; interestingly, high miR-10b expression in primary tumors was associated with worse prognosis and predicted the presence of distant metastases (26).
miR-122 and miR-885-5p are known to be highly expressed in the liver and are associated with acute liver injury (25, 27, 28). Vliegenthart et al. described elevated levels of these miRNAs bound to Argonaut RISC catalytic component 2 (AGO2) in blood serum in patients with liver toxicity (27). Similarly, high expression of serum miR-885-5p was found in individuals with chronic hepatitis B, liver cirrhosis, or hepatocellular cancer, and its measurement enabled healthy controls to be differentiated from patients with liver pathologies with high sensitivity and specificity; in addition it showed better diagnostic performance than alanine aminotransferase (28). Like us, Hur et al. identified higher expression of miR-885-5p in liver metastasis compared to primary tumors; moreover, high serum levels were associated with worse prognosis, lymph node metastases, and distant metastases (26). Similarly, overexpression of miR-122 in liver metastases was described and correlated with the expression of cationic amino acid transporter 1, which plays an important role in NO production and in supply of cells with extracellular arginine, lysine, and ornithine (29). These data suggest that not only tissue miRNAs but also circulating miRNAs might be able to predict the presence of liver metastases in patients with CRC. On the other hand, the elevated expression of liver-specific miR-122 (and probably also miR-122*) and miR-885-5p in metastatic tissue suggest the residual presence of normal liver tissue in samples of metastases. Pizzini et al. confirmed that the presence of only 5% of healthy liver cells in a sample significantly increased the expression of miR-122 (25); therefore, the contribution of normal liver to the metastatic transcriptome cannot be completely ruled out.
In conclusion, 6 miRNAs (miR-122, miR-122*, miR-885-5p, miR-10b, miR-143, and miR 28-5p) were found to be differently expressed between primary tumors of patients with CRC and their matched liver metastases. Although the involvement of these miRNAs in particular signaling pathways has not been elucidated in detail and further validation on larger independent cohorts is needed, they could hopefully serve as new candidates for targeted therapeutic interventions in patients with metastatic CRC.
Acknowledgements
This work was supported by grant project GACR 16-182575, Ministry of Health of the Czech Republic, grant nr. 16-31765A and project by the National Sustainability Program I (NPU I) Nr. LO1503 and the project CEITEC 2020 (LQ1601) provided by the Ministry of Education Youth and Sports of the Czech Republic.
- Received February 22, 2016.
- Revision received April 4, 2016.
- Accepted April 6, 2016.
- Copyright© 2016, International Institute of Anticancer Research (Dr. John G. Delinasios), All rights reserved
