Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Feb 15.
Published in final edited form as: Cancer. 2017 Nov 7;124(4):785–796. doi: 10.1002/cncr.31062

Global and targeted circulating miRNA profiling of colorectal adenoma and colorectal cancer

Jinhua Zhang 1,2, Gottumakkala S Raju 3, David W Chang 1, Shu-Hong Lin 1, Zhinan Chen 2,4, Xifeng Wu 1,*
PMCID: PMC5801195  NIHMSID: NIHMS908065  PMID: 29112225

Abstract

Background

Circulating microRNAs (miRNAs) are emerging as promising biomarkers for cancer. The aim of this study was to investigate the potential of circulating cell-free miRNAs as biomarkers for colorectal cancer (CRC) and its precursor lesion, colorectal adenoma.

Methods

The serum levels of 800 miRNAs were assessed in a discovery set of 21 CRC patients, 19 adenoma patients, and 21 CRC healthy controls by Nanostring miRNA analysis platform. Significantly differentially expressed miRNAs were further examined in a validation cohort of 34 CRC patients, 33 adenoma patients, and 35 controls by Fluidigm qPCR assays.

Results

The ratios between the expression values of the differentially expressed miRNAs were computed. Three miRNA ratios (miR-17-5p/miR-135b, miR-92a-3p/miR135b, and miR-451a/miR-491-5p) were validated for discriminating adenoma and CRC patients from the healthy control group, and 5 miRNA ratios (let-7b/miR-367-3p, miR-130a-3p/miR-409-3p, miR-148-3p/miR-27b, miR-148a-3p/miR-409-3p, and miR-21-5p/miR-367-3p) were validated for discriminating CRC patients from adenoma and healthy control groups. The area under the receiver operating characteristic curve (AUC) values for the 3 miRNA ratios in discriminating adenomas from healthy controls were 0.831 and 0.735 in the discovery and validation sets, respectively. The AUC values for the 5 miRNA ratios in discriminating CRC from adenoma were 0.797 and 0.732 in the discovery and validation sets, respectively. Pathway analysis revealed that target genes regulated by the miRNAs from the miRNA ratios were mainly enriched in metabolism- and inflammation-related pathways.

Conclusions

Our data suggest that circulating miRNAs can distinguish CRC and adenoma patients and may represent novel biomarkers for early non-invasive detection of CRC.

Keywords: circulating miRNA, colorectal adenoma, colorectal cancer, early detection, biomarker

Introduction

Colorectal cancer (CRC) poses a significant threat to the health of global populations. In the United States, CRC is the second leading cause of cancer-related deaths with more than 50,000 deaths annually.1 CRC often develops in a progressive fashion, from normal colon epithelial cells to benign adenomas, and ultimately to malignant cancer lesions. The progression of CRC has been associated with sequential changes in genes such as KRAS, DCC, APC, and P53.2 The 5-year survival rate for patients with early stage CRC is nearly 90%, but only 39.6% of CRCs are diagnosed at an early stage.3 Thus, in most patients CRC could be cured if it were detected and resected at a precancerous or early stage. Therefore, early detection of CRC and colorectal adenoma is one of the main prerequisites for successful treatment and reduction of mortality from CRC.

Currently, there are few blood-based biomarkers suitable for population screening or early diagnosis of CRC. CRC screening tests that primarily detect cancer, which include the guaiac-based fecal occult blood tests (gFOBT) and the fecal immunochemical tests (FIT)-based fecal occult blood tests (FOBTs) and testing stool for exfoliated DNA (sDNA) were used frequently in population screening; nevertheless, these tests were insufficiently sensitive and specific for detecting the CRC in early stages. 4, 5 In recent years, methylated SEPT9 (mSEPT9) DNA was found as a sensitive and specific biomarker for detection of CRC from blood and available as the Epi proColon test which was clinically approved as an aid in CRC detection. 6 Nevertheless, the test's overall sensitivity of approximately 70% and specificity of 80% indicate there is still room for improvement. In addition, the endoscopic and radiologic examinations, including flexible sigmoidoscopy, colonoscopy, double contrast barium enema, and computed tomography (CT) colonography (or virtual colonoscopy) can detect cancer and advanced precursor lesions, but their widespread use is limited by their invasive nature and high costs. 4, 7 Therefore, there is a great need for a cost-effective and non-invasive biomarker-based test for early detection of CRC.

MicroRNAs (miRNAs) are small, 19- to 25-nucleotide noncoding RNAs, which negatively regulate the expression of their targeted genes at the post-transcriptional level by base-pairing to the complementary sites on the target mRNAs. They play an important regulatory role in a wide range of biological and pathological processes and are involved in cancer development.8, 9 MiRNAs are frequently dysregulated in cancer and have shown great potential as tissue-based markers for cancer classification and prognostication.10, 11 As some are also secreted into the blood as cell-free miRNAs, which can be detected in serum in highly stable form, circulating miRNAs are emerging as promising minimally invasive biomarkers for diagnosing and monitoring human cancers. 12, 13

In the past few years, researchers have investigated the potential of using miRNAs as novel serum/plasma-based biomarkers for early non-invasive or minimally invasive diagnosis of CRC. Several studies have examined the differences in circulating miRNA expression between CRC patients and healthy controls, with a few miRNAs consistently shown to be differentially expressed in CRC, such as upregulation of miRNA-21 and miR-92a-3p and downregulation of miR-422a. 14-16 In addition, some studies indicate the deregulated expression of selected miRNAs in patients with different types of adenoma.17-19 However, those studies were limited by one or more of the following factors: insufficient number of screened miRNAs, small sample size, lack of independent validation, and failure to differentiate adenoma from CRC and healthy controls.

In this two-phase study, we carried out global miRNA profiling using the Nanostring platform with validation by targeted real-time PCR Taqman assay to characterize the miRNA expression profile in large sets of serum specimens from patients with colon cancer or adenoma as well as from healthy controls. We aimed to identify miRNAs that could potentially serve as novel serum-based biomarkers for CRC screening or early detection.

Materials and Medthods

Patient cohorts and sample collection

Two separate patient cohorts were selected for this study: a discovery set (N=61) comprising 21 healthy control participants, 19 adenoma patients, and 21 CRC patients for miRNA profile screening, and a validation set (N=102) comprising 35 healthy controls, 33 adenoma patients, and 34 CRC patients. All adenoma patients did not have prior or incident diagnosis of CRC or familial form of adenomas. The adenoma and CRC patients had received no chemotherapy or radiotherapy; their disease was histologically confirmed at The University of Texas MD Anderson Cancer Center, and they were recruited at the center between 2009 and 2014. Control subjects with no cancer history were recruited from the general population of Houston, Texas by random telephone digit dialing during the same period. Controls were frequency-matched to CRC and adenoma patients by sex, age, ethnicity, and residence during the same period. Participants were selected at random for the discovery set versus validation set. The study was approved by the MD Anderson Institutional Review Board, and all study participants signed an approved informed consent form according to institutional guidelines.

Blood samples were collected according to standard phlebotomy procedures. A total of 10 mL of whole blood from each participant was collected into tubes and immediately placed on ice. Blood samples were subjected to centrifugation at 1000×g for 10 minutes at 4°C to spin down the blood cells. The serum supernatant was removed by pipette from the cellular material and then aliquoted and stored in liquid nitrogen tanks until assays were performed.

MiRNA profiling

Total RNAs were isolated from 750 μL serum samples using a miRNeasy Mini Kit (Qiagen, Germantown, MD). The concentration and the purity of each RNA sample was evaluated by NanoDrop ND-100 Spectrophotometer (Thermo Scientific, Wilmington, DE) Expression profiles of the top 800 highly curated human miRNAs preselected from miRBase 21 by the manufacturer were measured in the discovery set using Nanostring nCounter assays (NanoString Technologies, Seattle, WA) and performed by the Genomic and RNA Core Service of Baylor College of Medicine in accordance with the manufacturer's instructions. In brief, the prepared RNA samples were ligated with a specific DNA tag onto the 3′ end of each RNA species. These tags were designed to normalize the melting temperatures of the miRNAs and provided a unique identification for each miRNA species in the sample. After hybridization and the removal of excess capture and reporter probes, the purified ternary complexes were bound to the imaging surface before they were elongated and immobilized. The surface was then imaged using the nCounter digital analyzer. The expression levels were normalized by using the average of expression of the top 100 most-expressed miRNAs in all samples. MiRNAs detected in at least 80% of the serum samples were considered for further analysis. Ninety-three miRNAs significantly differentially expressed in healthy control, adenoma, and CRC groups in the discovery set. The ratios of the expression values of each value of a single miRNA compared with the values of all of the other miRNAs were computed. Of these, 191 miRNA ratios were significantly different between healthy control, adenoma, and CRC groups in the discovery set.

Validation of miRNA expression using qRT-PCR

Fifty-three differentially expressed miRNAs represented in the 191 miRNA ratios were evaluated in the validation set. We used Fluidigm 96.96 dynamic arrays (Fluidigm Corp., San Francisco, CA) for miRNA quantification by qRT-PCR in a Biomark HD system according to the manufacturer's protocol.

Purified RNA samples were reverse-transcribed with a TaqMan miRNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA), using 150 ng total RNA and pools of Megaplex RT Primers, followed by a pre-amplification step with Megaplex PreAmp Primers (all primers, Thermo Fisher Scientific, Waltham, MA). The expression levels of miRNAs in the validation set were normalized to the spiked-in cel-miR-39 and cel-miR-54, and the 2-ΔΔCt method was used for analysis 20, 21. The ratios of the expression values of each value of a single miRNA compared with the values of all of the other miRNAs were computed.

Pathway analysis

Potential target genes and the pathway enrichment analysis of the potential target genes was conducted with a web-based analytical tool, miRsystem22 (http://mirsystem.cgm.ntu.edu.tw/), which offers a comprehensive database of validated miRNA targets from TarBase and miRecords and predicted targets from 7 programs.

Statistical analysis

All statistical analyses were performed by STATA software 14.0 (Stata Corp, College Station, TX). The rank sum test was applied to evaluate the differences in median miRNA expression ratio levels between patients with CRC or adenoma and healthy controls. The combined miRNA risk score for each patient was derived by linear combination of the product of the level of miRNA ratio (1 for high, 0 for low) and its logistic regression corresponding coefficient. All patients were dichotomized by the median risk score, such that individuals with a risk score higher than the median were classified as high risk and those with a risk score lower than the median were classified as low risk. Association between risk score and adenoma or CRC risk was assessed by unconditional multivariable logistic regression models, adjusting for age and gender. Receiver operating characteristic (ROC) curves were constructed, and area under the ROC curve (AUC) was used to evaluate the diagnostic performance of the selected miRNA ratio panel. All tests were two-sided, and a P value less than 0.05 was considered statistically significant.

Results

Cohort characteristics

This study enrolled, overall, 163 individuals: 56 CRC patients, 52 adenoma patients, and 55 age- and sex-matched healthy controls. The demographic and, if pertinent, clinical characteristics of these groups, including sex, age, smoking status, clinical stage, polyp/tumor location and histology, stratified by the discovery set and validation set, are summarized in Table 1. There were no significant differences in demographic and clinical characteristics between the CRC patients, adenoma patients, and health controls in either validation or discovery groups. Discovery and validation sets were not different in host characteristics except for location of colorectal adenomas and tumor stage of CRC patients (Table 1 and Supplementary Table S1). The majority of the adenoma patients had tubular adenomas (TA) for histology (60-80%), while 50% of the adenoma patients in the validation set showed missing information for adenoma location.

Table 1. Selected characteristics of the study population.

Characteristics Discovery cohort Validation cohort
CRC (N = 21) Adenoma (N = 19) Control (N = 21) P value* CRC (N = 35) Adenoma (N = 33) Control (N = 34) P value*
Sex, N (%)
 Male 10 (47.6) 11 (57.9) 11 (52.4) 0.81 21 (60.0) 20 (60.6) 22 (64.7) 0.91
 Female 11 (52.4) 8 (42.1) 10 (47.6) 14 (40.0) 13 (39.4) 12 (35.3)
Age at diagnosis
 Mean (SD) 62.4 (9.1) 59.1 (7.8) 61. 7 (11.5) 0.79 59.5 (9.3) 60.5 (9.8) 61.7 (9.4) 0.27
Smoking status
 Never 9 (42.9) 8 (42.1) 13 (61.9) 0.47 17 (48.6) 16 (48.5) 20 (58.8) 0.89
 Former 9 (42.9) 6 (31.6) 6 (28.6) 13 (37.1) 13 (39.4) 11 (32.4)
 Current 3 (14.3) 5 (26.3) 2 (9.5) 5 (14.3) 4 (12.1) 3 (8.8)
Smoker
 Ever 12 (57.1) 11 (57.9) 8 (38.1) 0.35 18 (51.4) 17 (51.5) 14 (41.2) 0.62
 Never 9 (42.9) 8 (42.1) 13 (61.9) 17 (48.6) 16 (48.5) 20 (58.8)
TNM stage
 I 2 (9.5) 3 (8.6) 0.13
 II 4 (19.0) 8 (22.9)
 III 10 (47.6) 6 (17.1)
 IV 2 (9.5) 10 (28.6)
 Unknown 3 (14.3) 8 (22.9)
T stage
 0 5 (23.8) 12 (34.3) <0.001
 1 3 (14.3) 1 (2.9)
 2 13 (61.9) 3 (8.6)
 3 16 (45.7)
 4 3 (8.6)
Polyp location
 Ascending colon 5 6 0.01
 Cecum 3 2
 Descending colon 1 2
 Hepatic flexure 1 1
 Rectum 0 1
 Sigmoid colon 3 1
 Transverse colon 6 3
 Unknown 0 17
Tumor location
 Distal colon 4 (19.0) 5 (14.3) 0.23
 Unknown 2 (9.5) 8 (22.9)
 Proximal colon 7 (33.3) 16 (45.7)
 Rectum 8 (38.1) 6 (17.1)
Histology, N (%)
 SSA 1 (5.3) 7 (21.21) 0.31
 TA 15 (78.95) 22 (66.67)
 TVA 3 (15.79) 4 (12.12)
 Adenocarcinoma 21 (100.0) 35 (100.0) 1.00
Open in a new tab

Abbreviations: CRC, Colorectal cancer; SSA, sessile serrated adenoma; TA, tubular adenoma; TVA, tubulovillous adenoma

*

P values for sex, age, smoking status, and smoker were based on comparison of distribution across CRC, adenoma, and control subgroups. P values for clinical variables were based on comparison between discovery and validation sets.

Differentially expressed miRNA ratios in the discovery set

The use of miRNA ratios has been reported as an easily applicable method to develop clinically useful signatures based on circulating biomarkers due to the lack of a consistent internal standard for normalization.23 We therefore computed the ratios of the expression values of 93 miRNAs significantly differentially expressed in healthy control, adenoma, and CRC groups. Each value of a single miRNA was compared with the values of all of the other miRNAs; 2529 ratios were obtained and subsequently used to analyze group differences. Of these, 155 miRNA ratios were significantly different in CRC and adenoma samples than in healthy control samples (Supplementary Table S2). Thirty-six miRNA ratios were significantly different in CRC samples than in healthy control and adenoma samples (Supplementary Table S3).

Validation of miRNA ratios by qRT-PCR analysis

The 53 differentially expressed miRNAs represented in these 191 miRNA ratios were then validated in an independent set of serum samples. The ratios between the expression values of these miRNAs were computed and compared. Of these, 3 miRNA ratios, miR-17-5p/miR-135b, miR-92a-3p/miR135b, and miR-451a/miR-491-5p, were confirmed to be significantly higher in the adenoma and CRC groups than in the healthy controls (Figure 1). No significant differences were observed for these 3 miRNA ratios between the adenoma and CRC groups. Five miRNA ratios, let-7b/miR-367-3p, miR-130a-3p/miR-409-3p, miR-148-3p/miR-27b, miR-148a-3p/miR-409-3p, and miR-21-5p/miR-367-3p, were confirmed to be significantly higher in the CRC group than in both the adenoma group and the healthy controls (Figure 2). No significant differences were observed for these 5 miRNA ratios between adenoma and healthy control groups.

Figure 1.

Figure 1

Open in a new tab

Box plots of significant miRNA ratios (log10 scale on Y-axis) for miR-17-5p/miR-135b, miR-92a-3p/miR-135b, and miR-451a/miR-491-5p, showing elevated levels in CRC and adenoma patients compared to controls in (A) discovery and (B) validation sets. The lines inside the boxes denote the medians, and the boxes define values from the 25th percentile to the 75th percentile. The upper and lower bars mark the values within the 95% confidence interval. The rank sum test was performed for comparisons between groups. CRC, colorectal cancer.

Figure 2.

Figure 2

Open in a new tab

Box plots of significant miRNA ratios (log10 scale on Y-axis) for let-7b/miR-367-3p, miR-130a-3p/miR-409-3p, miR-148a-3p/miR-27b, miR-148a-3p/miR-409-3p, and miR-21-5p/miR-367-3p, showing elevated levels in CRC patients compared to adenoma and control subjects in (A) discovery and (B) validation sets. The lines inside the boxes denote the medians, and the boxes define values from the 25th percentile to the 75th percentile. The upper and lower bars mark the values within the 95% confidence interval. The rank sum test was performed for comparisons between groups. CRC, colorectal cancer.

Prediction of adenoma and CRC risks by miRNA ratio risk score

The combined effects of the 3 miRNA ratios miR-17-5p/miR-135b, miR-92a-3p/miR135b, and miR-451a/miR-491-5p on risk of adenoma and CRC were investigated by calculating risk scores. In both the discovery and validation sets, participants with a high-risk score exhibited significantly greater risks of adenoma (odds ratio [OR]=34.41, 95% confidence interval [CI], 3.08-384.82 for the discovery set and OR=6.69, 95% CI, 1.94–23.08 for the validation set) and of CRC (OR=16.03 95% CI, 2.39-107.34 for the discovery set and OR=3.66, 95% CI, 1.25–10.68 for the validation set) than those with a low-risk score (Table 2). The combined effects of the 5 miRNA ratios let-7b/miR-367-3p, miR-130a-3p/miR-409-3p, miR-148-3p/miR-27b, miR-148a-3p/miR-409-3p, and miR-21-5p/miR-367-3p on CRC risk compared to the adenoma or healthy control group were also investigated by calculating risk scores. Patients with a high-risk score exhibited significantly greater risk of progression from adenoma to CRC (OR=6.47, 95% CI, 1.38-30.18 for the discovery set and OR=5.14, 95% CI, 1.73–15.29 for the validation set) and greater risk of CRC (OR=5.12, 95% CI, 1.23-21.4 for the discovery set and OR=7.51, 95% CI, 2.18–25.85 for the validation set) than the low-risk group (Table 2).

Table 2. Serum miRNA markers risk scores associated with adenoma and CRC patients.

miRNA ratios differentiating adenoma and CRC from healthy controls:
Risk scorea Adenoma N (%) Control N (%) Adjusted ORa (95% CI) P value CRC N (%) Control N (%) Adjusted ORa (95% CI) P value
Discovery Low 1 (5.3) 12 (57.1) 1 (reference) 2 (9.5) 11(52.4) 1(reference)
High 18 (94.7) 9 (42.9) 34.41 (3.08-384.82) 0.004 19 (90.5) 10(47.6) 16.03 (2.39-107.34) 0.004
Validation Low 5 (15.2) 17 (50.0) 1 (reference) 8(22.9) 17(50.0) 1(reference)
High 28 (84.9) 17 (50.0) 6.69 (1.94-23.08) 0.003 27(77.1) 17(50.0) 3.66 (1.25-10.68) 0.018
miRNA ratios differentiating CRC from adenoma and healthy controls:
Risk scoreb CRC N (%) Adenoma N (%) Adjusted ORa (95% CI) Pvalue CRC N (%) Control N (%) Adjusted ORa (95% CI) P value
Discovery Low 3 (14.3) 10 (52.6) 1 (reference) 4 (19.0) 11(52.4) 1(reference)
High 18 (85.7) 9 (47.4) 6.47 (1.38-30.18) 0.018 17 (81.0) 10(47.6) 5.12 (1.23-21.4) 0.025
Validation Low 8 (22.9) 19 (57.6) 1 (reference) 5(14.3) 17(50.0) 1(reference)
High 27 (77.1) 14 (42.4) 5.14 (1.73-15.29) 0.003 30(85.7) 17(50.0) 7.51 (2.18-25.85) 0.001
Open in a new tab

Adjusted by age and sex.

a

Risk score from miR-17-5p/miR-135b, miR-92a-3p/miR-135b, and miR-451a/miR-491-5p ratios

b

Risk score from let-7b/miR-367-3p, miR-21-5p/miR-367-3p, miR-130a-3p/miR-409-3p, miR-148a-3p/miR-409-3p, and miR-148a-3p/miR-27b ratios

Discriminating CRC and adenoma from healthy controls using miRNA ratios

ROCs were generated to evaluate the discriminatory power of these miRNA ratios for differentiating between control and disease groups. As shown in Figure 3, AUC values for miR-17-5p/miR-135b, miR-92a-3p/miR-135b, and miR-451a/miR-491-5p in discriminating adenoma from healthy controls were 0.831 (95% CI, 0.706-0.956) for the discovery set and 0.735 (95% CI, 0.619-0.851) for the validation set. Similarly, AUC values for let-7b/miR-367-3p, miR-130a-3p/miR-409-3p, miR-148-3p/miR-27b, miR-148a-3p/miR-409-3p, and miR-21-5p/miR-367-3p in discriminating CRC from adenoma were 0.797 (95% CI, 0.659-0.935) for the discovery set and 0.732 (95% CI, 0.611-0.851) for the validation set.

Figure 3.

Figure 3

Open in a new tab

Receiver operating characteristic (ROC) curve analysis using serum miRNA ratios to differentiate adenoma patients from healthy controls and CRC patients. (A) Levels of 3 miRNA ratios, miR-17-5p/miR-135b, miR-92a-3p/miR-135b, and miR-451a/miR-491-5p, which differentiated adenoma patients from healthy controls. (B) Levels of 5 miRNA ratios, let-7b/miR-367-3p, miR-130a-3p/miR-409-3p, miR-148-3p/miR-27b, miR-148a-3p/miR-409-3p, and miR-21-5p/miR-367-3p, which differentiated CRC patients from adenoma patients.

Target genes and pathway analysis

To explore the biological mechanisms underlying the role of these miRNAs in CRC progression, we performed target prediction coupled with pathway analysis. Target genes of the 12 miRNAs comprising the validated miRNA ratios were identified by both prediction algorithms and experiment-supported databases. Pathway analysis showed that miR-17-5p, miR-135b, miR-92a-3p, miR-451a, and miR-491-5p from the 3 miRNA ratios for discriminating adenoma and CRC patients from healthy control group were mainly enriched in 14 subcategories and most of the pathways were related to metabolism, such as metabolism of amino acids and derivatives, glycerolipid metabolism, integration of energy metabolism, amino sugar and nucleotide sugar metabolism, and tyrosine metabolism. The predicted target genes also were involved in the calcium signaling pathway, the c-myb transcription factor network, and the lysosome pathway (Table 3).

Table 3. Enriched pathways of the 5 miRNAs from the 3 miRNA ratios to differentiate adenoma and CRC patients from healthy control group.

Database Pathways ID Genes Targets miRNAs Empirical P value
KEGG Calcium signaling pathway 4020 177 9 3 1.64E-04
PATHWAY INTERACTION DATABASE C-myb transcription factor network 200154 81 5 5 1.50E-02
REACTOME VIF-mediated degradation of APOBEC3G REACT 9453 54 4 3 1.77E-02
REACTOME Metabolism of amino acids and derivatives REACT 13 174 6 4 2.43E-02
REACTOME APC C CDH1 mediated degradation of CDC20 and other APC C CDH1 targeted proteins in late mitosis early G1 REACT 6761 69 4 3 2.93E-02
GO MOLECULAR FUNCTION TIER2 Protein binding transcription factor activity GO: 0000988 369 12 5 3.38E-02
KEGG Glycerolipid metabolism 561 49 2 2 3.54E-02
KEGG Lysosome 4142 121 4 3 3.55E-02
REACTOME Integration of energy metabolism REACT 1505 125 5 3 3.94E-02
KEGG Amino sugar and nucleotide sugar metabolism 520 47 3 2 4.22E-02
REACTOME Destabilization of MRNA by AUF1 (HNRNP D0) REACT 25325 54 4 2 4.30E-02
KEGG Oocyte meiosis 4114 112 4 3 4.39E-02
KEGG Tyrosine metabolism 350 41 2 2 4.56E-02
REACTOME Biological oxidations REACT 13433 139 4 4 4.59E-02
Open in a new tab

Empirical P values were compared with 1,000 random selections.

Analyzed with miR-17-5p, miR-135b, miR-92a-3p, miR-451 and miR-491-5p

Metabolism-related pathways are in bold font.

MiRNAs let-7b, miR-367-3p, miR-130a-3p, miR-409-3p, miR-148-3p, miR-27b, miR-409-3p, and miR-21-5p from the 5 miRNA ratios for discriminating CRC patients from adenoma and healthy control groups were mainly enriched in 26 subcategories, 4 of which were inflammation-related pathways, such as the chemokine signaling pathway, the chemokine and the chemokine receptors signaling pathway, and the cytokine-cytokine receptor interaction pathway (Table 4). The predicted target genes also were involved in several critical pathways such as protein processing, direct P53 effectors, the NKT (natural killer T cell) pathway, cell cycle checkpoints, and the DNA repair pathway.

Table 4. Enriched pathways of the 7 miRNAs of the 5 miRNA ratios to differentiate patients with CRC from adenoma and healthy control groups.

Database Pathways ID Genes Targets miRNAs Empirical P value
KEGG Protein processing in endoplasmic reticulum 4141 166 8 6 6.23E-03
KEGG Chemokine signaling pathway 4062 189 8 6 7.23E-03
REACTOME Chemokine receptors bind chemokines REACT 15344 54 4 3 8.36E-03
PATHWAY INTERACTION DATABASE ATF-2 Transcription factor network 200136 58 5 6 8.84E-03
PATHWAY INTERACTION DATABASE Direct P53 effectors 200120 137 6 6 9.07E-03
BIOCARTA Biocarta NKT pathway 28 4 2 1.02E-02
REACTOME GPCR ligand binding REACT 21340 410 11 5 1.17E-02
REACTOME G alpha (I) signaling events REACT 19231 200 6 4 1.53E-02
KEGG Dilated cardiomyopathy 5414 90 4 7 1.70E-02
KEGG Olfactory transduction 4740 388 1 2 1.72E-02
REACTOME Peptide ligand-binding receptors REACT 14819 186 6 5 1.78E-02
KEGG Cytokine-cytokine receptor interaction 4060 275 9 5 1.82E-02
REACTOME G alpha (S) signalling events REACT 19327 125 5 5 1.97E-02
REACTOME Class A 1 (Rhodopsin-like receptors) REACT 14828 305 8 5 2.11E-02
REACTOME Hemostasis REACT 604 467 11 7 2.50E-02
KEGG Nucleotide excision repair 3420 44 3 4 2.95E-02
REACTOME Endosomal sorting complex required for transport (ESCRT) REACT 27258 28 4 3 3.04E-02
PATHWAY INTERACTION DATABASE P75(NTR)-mediated signaling 200123 67 5 5 3.24E-02
REACTOME G ALPHA (Z) signaling events REACT 19333 45 3 4 3.52E-02
REACTOME Cell cycle checkpoints REACT 1538 117 5 6 3.59E-02
REACTOME Transcription-coupled NER (TC-NER) REACT 1628 44 3 4 3.70E-02
REACTOME Nucleotide excision repair REACT 1826 49 3 4 4.03E-02
REACTOME DNA repair REACT 216 108 4 4 4.24E-02
REACTOME PLC-Gamma1 signaling REACT 12079 34 2 3 4.43E-02
KEGG Neurotrophin signaling pathway 4722 127 5 5 4.89E-02
REACTOME APC C CDC20 mediated degradation of mitotic proteins REACT 6781 69 4 6 4.94E-02
Open in a new tab

Empirical P values were compared with 1,000 random selections.

Analyzed with let-7b, miR-367-3p, miR-130a-3p, miR-409-3p, miR-148-3p, miR-27b, and miR-21-5p.

Immune-related pathways are in bold font.

Discussion

In this study, we analyzed a cohort of 163 samples representing healthy control participants, patients with benign adenoma, and patients with CRC in independent discovery and validation sets. This analysis identified 3 miRNA ratios (miR-17-5p/miR-135b, miR-92a-3p/miR135b, and miR-451a/miR-491-5p) that were significantly upregulated in adenoma patients compared with the healthy control group. We also identified 5 miRNA ratios (let-7b/miR-367-3p, miR-130a-3p/miR-409-3p, miR-148-3p/miR-27b, miR-148a-3p/miR-409-3p, and miR-21-5p/miR-367-3p) that were significantly upregulated in CRC patients compared with the adenoma and healthy control groups and were effective at distinguishing patients with CRC from the benign adenoma and healthy control groups. Therefore, these circulating miRNA ratios are novel potential biomarkers for the detection and diagnosis of adenoma and CRC.

Previous studies have identified a spectrum of dysregulated miRNAs associated with the tumorigenesis and development of CRC.10, 24 However, these studies mainly focused on the miRNAs in tissues or cells. The reliance on surgical resection and invasive procedures for tissue sample collection limits the application of tissue miRNAs in cancer diagnosis. The goal, therefore, is to develop a method for comprehensive analysis of premalignant lesions and cancers through serum miRNA-based biomarkers without the need for biopsy, surgery, or other invasive methods.

Several circulating miRNAs, including miR-21-5p and miR-92a-3p, have been reported as potential biomarkers for the early detection of CRC.10-15 Since the normalization of miRNA data in serum samples is still a controversial issue, the use of miRNA ratios has been reported as an easily applicable method for developing clinically useful signatures based on circulating biomarkers.23 Several miRNA expression ratios have been identified as diagnostic markers for different cancers, such as lung cancer, bladder cancer, and head and neck cancer.23, 25-27 However, miRNA expression ratios have not been reported in CRC. Our study revealed several miRNA ratios that could act as early detection markers for CRC and adenoma.

Consistent with our findings, most of these miRNAs have previously been reported to be significantly dysregulated in CRC and to play important roles in tumor development. Expression levels of miR-17-5p, miR-135b, miR-92a-3p, miR-21-5p, and miR-148a-3p have been confirmed to be significantly higher in CRC tissues than in normal tissues.24 MiR-17-5p is an important regulator of the cell cycle, targeting CDKN1A (P21) and TP53.28, 29 Increased expression of miR-135b, like that of miR-17-5p, has previously been reported to be an important early event in colon carcinogenesis.30, 31 MiR-135b targets the 3′-untranslated region of APC, suppresses its expression, and acts as a downstream effector of oncogenic pathways in colon cancer.32 It has been reported that miR-92a-3p plays a central role in development and progression of CRC and in regulation of its aggressiveness.33 This miRNA downregulates the pro-apoptotic protein Bim, and its overexpression is significantly related to lymph nodes metastasis in CRC.33 Several studies reported that plasma miR-21-5p could be used as a biomarker for early detection of CRC.17 Levels of miR‐409-3p have been shown to be significantly lower in CRC tissues compared to adjacent non-tumor tissues, and reduced miR-409-3p expression was correlated with CRC metastasis.34 In addition, miR‐409-3p suppresses CRC cell proliferation, in part by inhibiting autophagy mediated by Beclin-1.35 MiR-491, like miR‐409-3p, functions as a tumor suppressor and is downregulated in CRC.24 In CRC cell line DLD-1, miR-491 inhibits Bcl-XL expression and induces apoptosis.36

To gain insight into the potential functional importance of the miRNAs comprising these miRNA ratios, we retrieved their target genes and analyzed their related pathways by miRsystem. The 5 miRNAs included in the 3 miRNA ratios that are capable of differentiating adenoma patients from healthy controls are likely key regulatory factors of oncogenic processes. Most of the pathways are related to metabolism, such as metabolism of amino acids and derivatives, glycerolipid metabolism and integration of energy metabolism (Table 3). Metabolic reprogramming is a major hallmark of cancer, which is characterized by upregulated glycolysis, glutaminolysis, lipid metabolism, the pentose phosphate pathway, and mitochondrial biogenesis, among others.37, 38 These metabolic programs provide cancer cells with not only energy sources but also vital metabolites to support large-scale biosynthesis, continuous proliferation, and other major processes of tumorigenesis.39, 40 MiRNAs regulate cell metabolic processes through complicated mechanisms, including direct targeting of key enzymes or transporters of metabolic processes and regulating transcription factors, and multiple oncogenic signaling pathways.41, 42 The 5 miRNAs' participation in several different metabolic pathways has been supported by several studies,43, 44 indicating their important regulatory role in CRC development, especially in the early stages of tumorigenesis. The functional mechanisms of these miRNAs in CRC tumorigenesis remain to be elucidated by biological experiments.

As for the 7 miRNAs included in the 5 miRNA ratios that could differentiate CRC patients from adenoma and healthy control groups, signal pathway analysis by miRsystem revealed that most of these miRNAs are involved in crucial signaling pathways, particularly inflammation-related pathways such as the chemokine signaling pathway and the cytokine-cytokine receptor interaction pathway (Table 4). It is widely believed that chronic inflammation creates a favorable environment for tumor initiation, promotion, and progression.37, 45, 46 Cytokines, chemokines and matrix-degrading enzymes produced during chronic inflammation may damage DNA and/or alter cell proliferation or survival, thereby promoting oncogenesis. Immune cells, which often infiltrate tumors and preneoplastic lesions, also produce a variety of cytokines and chemokines that form a localized inflammatory microenvironment and enhance premalignant cell growth and survival by activating signaling pathways such as NF-κB or MAPKs.47, 48 Several studies have reported some of these 7 identified miRNAs being involved in regulation of inflammation and cancer development.49, 50 Moreover, these miRNAs participate in several other cancer-related pathways, including direct P53 effectors, the NKT and DNA repair pathways, and cell cycle checkpoints, which indicates their important roles in CRC development.

Our study did have some limitations. We did not obtain history of colorectal adenoma or endoscopy checkup information for the control subjects; therefore, it is possible some participants might have adenomas or history of adenomas. Nevertheless, this confounding effect should bias the results towards the null. The relatively small cohorts of CRC, adenoma, and healthy control subjects prevented us from conducting stratified analyses by clinical and epidemiologic factors and adjusting for multiple comparisons. The small sample size of our discovery set also might limit our power to identify more differentially expressed miRNAs during the screening stage. In addition, due to the usage of different gene expression platforms, we could not merge the data from discovery and validation sets. Future studies using large, independent prospective cohorts are needed to validate our findings for clinical practice.

In conclusion, we conducted a systematic investigation to identify circulating miRNA biomarkers for early detection of CRC. We identified 3 miRNA ratios that effectively distinguish patients with precancerous adenoma from healthy controls and 5 miRNA ratios that distinguish patients with CRC from those with benign adenoma and healthy controls. The evaluation of miRNA signatures as ratios and functional enrichment analysis of their putative targets reveal potential etiologic mechanisms and support the miRNAs identified as candidate early detection markers for adenoma and CRC. Further functional investigations of miRNA gene targets in tumor tissues are also warranted to explore the mechanisms underlying the progression of CRC from normal tissue to precancerous adenoma and ultimately malignancy.

Supplementary Material

Supp TableS1

Footnotes

Conflict of interest: No conflicts of interests are declared.

Author contributions: Xifeng Wu obtained funding and contributed to study design and supervision; Jinhua Zhang collected, analyzed the data and drafted the manuscript; Hongshu Lin analyzed and interpreted the data; David W. Chang and Zhinan Chen analyzed the data and revised the manuscript for important intellectual content; Gottumakkala S. Raju provided clinical resources and/or technical support.

References

  • 1.Siegel RL, Miller KD, Jemal A. Cancer statistics, 2016. CA Cancer J Clin. 2016;66:7–30. doi: 10.3322/caac.21332. [DOI] [PubMed] [Google Scholar]
  • 2.Arends JW. Molecular interactions in the Vogelstein model of colorectal carcinoma. J Pathol. 2000;190:412–416. doi: 10.1002/(SICI)1096-9896(200003)190:4<412::AID-PATH533>3.0.CO;2-P. [DOI] [PubMed] [Google Scholar]
  • 3.Brenner H, Kloor M, Pox CP. Colorectal cancer. Lancet. 2014;383:1490–1502. doi: 10.1016/S0140-6736(13)61649-9. [DOI] [PubMed] [Google Scholar]
  • 4.Smith RA, Manassaram-Baptiste D, Brooks D, et al. Cancer screening in the United States, 2015: a review of current American cancer society guidelines and current issues in cancer screening. CA Cancer J Clin. 2015;65:30–54. doi: 10.3322/caac.21261. [DOI] [PubMed] [Google Scholar]
  • 5.Imperiale TF, Ransohoff DF, Itzkowitz SH, et al. Multitarget stool DNA testing for colorectal-cancer screening. N Engl J Med. 2014;370:1287–1297. doi: 10.1056/NEJMoa1311194. [DOI] [PubMed] [Google Scholar]
  • 6.Yan S, Liu Z, Yu S, Bao Y. Diagnostic Value of Methylated Septin9 for Colorectal Cancer Screening: A Meta-Analysis. Med Sci Monit. 2016;22:3409–3418. doi: 10.12659/MSM.900590. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Brenner H, Stock C, Hoffmeister M. Effect of screening sigmoidoscopy and screening colonoscopy on colorectal cancer incidence and mortality: systematic review and meta-analysis of randomised controlled trials and observational studies. Bmj. 2014;348:g2467. doi: 10.1136/bmj.g2467. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Bartel DP. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell. 2004;116:281–297. doi: 10.1016/s0092-8674(04)00045-5. [DOI] [PubMed] [Google Scholar]
  • 9.Lin S, Gregory RI. MicroRNA biogenesis pathways in cancer. Nat Rev Cancer. 2015;15:321–333. doi: 10.1038/nrc3932. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Okugawa Y, Grady WM, Goel A. Epigenetic Alterations in Colorectal Cancer: Emerging Biomarkers. Gastroenterology. 2015;149:1204–1225.e1212. doi: 10.1053/j.gastro.2015.07.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Hayes J, Peruzzi PP, Lawler S. MicroRNAs in cancer: biomarkers, functions and therapy. Trends Mol Med. 2014;20:460–469. doi: 10.1016/j.molmed.2014.06.005. [DOI] [PubMed] [Google Scholar]
  • 12.Zen K, Zhang CY. Circulating microRNAs: a novel class of biomarkers to diagnose and monitor human cancers. Med Res Rev. 2012;32:326–348. doi: 10.1002/med.20215. [DOI] [PubMed] [Google Scholar]
  • 13.Huang G, Li S, Yang N, Zou Y, Zheng D, Xiao T. Recent progress in circular RNAs in human cancers. Cancer Lett. 2017 doi: 10.1016/j.canlet.2017.07.002. [DOI] [PubMed] [Google Scholar]
  • 14.Clancy C, Joyce MR, Kerin MJ. The use of circulating microRNAs as diagnostic biomarkers in colorectal cancer. Cancer Biomark. 2015;15:103–113. doi: 10.3233/CBM-140456. [DOI] [PubMed] [Google Scholar]
  • 15.Zekri AR, Youssef AS, Lotfy MM, et al. Circulating Serum miRNAs as Diagnostic Markers for Colorectal Cancer. PLoS One. 2016;11:e0154130. doi: 10.1371/journal.pone.0154130. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Hur K, Toiyama Y, Okugawa Y, et al. Circulating microRNA-203 predicts prognosis and metastasis in human colorectal cancer. Gut. 2015 doi: 10.1136/gutjnl-2014-308737. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Yamada A, Horimatsu T, Okugawa Y, et al. Serum miR-21, miR-29a, and miR-125b Are Promising Biomarkers for the Early Detection of Colorectal Neoplasia. Clin Cancer Res. 2015;21:4234–4242. doi: 10.1158/1078-0432.CCR-14-2793. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Zheng G, Du L, Yang X, et al. Serum microRNA panel as biomarkers for early diagnosis of colorectal adenocarcinoma. Br J Cancer. 2014;111:1985–1992. doi: 10.1038/bjc.2014.489. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Aherne ST, Madden SF, Hughes DJ, et al. Circulating miRNAs miR-34a and miR-150 associated with colorectal cancer progression. BMC Cancer. 2015;15:329. doi: 10.1186/s12885-015-1327-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods. 2001;25:402–408. doi: 10.1006/meth.2001.1262. [DOI] [PubMed] [Google Scholar]
  • 21.Kroh EM, Parkin RK, Mitchell PS, Tewari M. Analysis of circulating microRNA biomarkers in plasma and serum using quantitative reverse transcription-PCR (qRT-PCR) Methods. 2010;50:298–301. doi: 10.1016/j.ymeth.2010.01.032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Lu TP, Lee CY, Tsai MH, et al. miRSystem: an integrated system for characterizing enriched functions and pathways of microRNA targets. PLoS One. 2012;7:e42390. doi: 10.1371/journal.pone.0042390. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Boeri M, Verri C, Conte D, et al. MicroRNA signatures in tissues and plasma predict development and prognosis of computed tomography detected lung cancer. Proc Natl Acad Sci U S A. 2011;108:3713–3718. doi: 10.1073/pnas.1100048108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Slattery ML, Herrick JS, Pellatt DF, et al. MicroRNA profiles in colorectal carcinomas, adenomas and normal colonic mucosa: variations in miRNA expression and disease progression. Carcinogenesis. 2016;37:245–261. doi: 10.1093/carcin/bgv249. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Neely LA, Rieger-Christ KM, Neto BS, et al. A microRNA expression ratio defining the invasive phenotype in bladder tumors. Urol Oncol. 2010;28:39–48. doi: 10.1016/j.urolonc.2008.06.006. [DOI] [PubMed] [Google Scholar]
  • 26.Avissar M, Christensen BC, Kelsey KT, Marsit CJ. MicroRNA expression ratio is predictive of head and neck squamous cell carcinoma. Clin Cancer Res. 2009;15:2850–2855. doi: 10.1158/1078-0432.CCR-08-3131. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Chen Z, Wu L, Lin Q, Shi J, Lin X, Shi L. Evaluation of miR-182/miR-100 Ratio for Diagnosis and Survival Prediction in Bladder Cancer. Arch Iran Med. 2016;19:645–651. [PubMed] [Google Scholar]
  • 28.Cai Z, Cao R, Zhang K, et al. Oncogenic miR-17/20a Forms a Positive Feed-forward Loop with the p53 Kinase DAPK3 to Promote Tumorigenesis. J Biol Chem. 2015;290:19967–19975. doi: 10.1074/jbc.M115.661504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Chen C, Lu Z, Yang J, et al. MiR-17-5p promotes cancer cell proliferation and tumorigenesis in nasopharyngeal carcinoma by targeting p21. Cancer Med. 2016;5:3489–3499. doi: 10.1002/cam4.863. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Khatri R, Subramanian S. MicroRNA-135b and Its Circuitry Networks as Potential Therapeutic Targets in Colon Cancer. Front Oncol. 2013;3:268. doi: 10.3389/fonc.2013.00268. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Li J, Liang H, Bai M, et al. miR-135b Promotes Cancer Progression by Targeting Transforming Growth Factor Beta Receptor II (TGFBR2) in Colorectal Cancer. PLoS One. 2015;10:e0130194. doi: 10.1371/journal.pone.0130194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Valeri N, Braconi C, Gasparini P, et al. MicroRNA-135b promotes cancer progression by acting as a downstream effector of oncogenic pathways in colon cancer. Cancer Cell. 2014;25:469–483. doi: 10.1016/j.ccr.2014.03.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Tsuchida A, Ohno S, Wu W, et al. miR-92 is a key oncogenic component of the miR-17-92 cluster in colon cancer. Cancer Sci. 2011;102:2264–2271. doi: 10.1111/j.1349-7006.2011.02081.x. [DOI] [PubMed] [Google Scholar]
  • 34.Liu M, Xu A, Yuan X, et al. Downregulation of microRNA-409-3p promotes aggressiveness and metastasis in colorectal cancer: an indication for personalized medicine. J Transl Med. 2015;13:195. doi: 10.1186/s12967-015-0533-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Tan S, Shi H, Ba M, et al. miR-409-3p sensitizes colon cancer cells to oxaliplatin by inhibiting Beclin-1-mediated autophagy. Int J Mol Med. 2016;37:1030–1038. doi: 10.3892/ijmm.2016.2492. [DOI] [PubMed] [Google Scholar]
  • 36.Nakano H, Miyazawa T, Kinoshita K, Yamada Y, Yoshida T. Functional screening identifies a microRNA, miR-491 that induces apoptosis by targeting Bcl-X(L) in colorectal cancer cells. Int J Cancer. 2010;127:1072–1080. doi: 10.1002/ijc.25143. [DOI] [PubMed] [Google Scholar]
  • 37.Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell. 2011;144:646–674. doi: 10.1016/j.cell.2011.02.013. [DOI] [PubMed] [Google Scholar]
  • 38.Sullivan LB, Gui DY, Vander Heiden MG. Altered metabolite levels in cancer: implications for tumour biology and cancer therapy. Nat Rev Cancer. 2016;16:680–693. doi: 10.1038/nrc.2016.85. [DOI] [PubMed] [Google Scholar]
  • 39.Schug ZT, Vande Voorde J, Gottlieb E. The metabolic fate of acetate in cancer. Nat Rev Cancer. 2016;16:708–717. doi: 10.1038/nrc.2016.87. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Martinez-Outschoorn UE, Peiris-Pages M, Pestell RG, Sotgia F, Lisanti MP. Cancer metabolism: a therapeutic perspective. Nat Rev Clin Oncol. 2017;14:11–31. doi: 10.1038/nrclinonc.2016.60. [DOI] [PubMed] [Google Scholar]
  • 41.Pinweha P, Rattanapornsompong K, Charoensawan V, Jitrapakdee S. MicroRNAs and oncogenic transcriptional regulatory networks controlling metabolic reprogramming in cancers. Comput Struct Biotechnol J. 2016;14:223–233. doi: 10.1016/j.csbj.2016.05.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Chen B, Li H, Zeng X, et al. Roles of microRNA on cancer cell metabolism. J Transl Med. 2012;10:228. doi: 10.1186/1479-5876-10-228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Izreig S, Samborska B, Johnson RM, et al. The miR-17 approximately 92 microRNA Cluster Is a Global Regulator of Tumor Metabolism. Cell Rep. 2016;16:1915–1928. doi: 10.1016/j.celrep.2016.07.036. [DOI] [PubMed] [Google Scholar]
  • 44.Godlewski J, Nowicki MO, Bronisz A, et al. MicroRNA-451 regulates LKB1/AMPK signaling and allows adaptation to metabolic stress in glioma cells. Mol Cell. 2010;37:620–632. doi: 10.1016/j.molcel.2010.02.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Coussens LM, Werb Z. Inflammation and cancer. Nature. 2002;420:860–867. doi: 10.1038/nature01322. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Ullman TA, Itzkowitz SH. Intestinal inflammation and cancer. Gastroenterology. 2011;140:1807–1816. doi: 10.1053/j.gastro.2011.01.057. [DOI] [PubMed] [Google Scholar]
  • 47.Grivennikov SI, Greten FR, Karin M. Immunity, inflammation, and cancer. Cell. 2010;140:883–899. doi: 10.1016/j.cell.2010.01.025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Berraondo P, Minute L, Ajona D, Corrales L, Melero I, Pio R. Innate immune mediators in cancer: between defense and resistance. Immunol Rev. 2016;274:290–306. doi: 10.1111/imr.12464. [DOI] [PubMed] [Google Scholar]
  • 49.Tili E, Michaille JJ, Croce CM. MicroRNAs play a central role in molecular dysfunctions linking inflammation with cancer. Immunol Rev. 2013;253:167–184. doi: 10.1111/imr.12050. [DOI] [PubMed] [Google Scholar]
  • 50.Iliopoulos D, Jaeger SA, Hirsch HA, Bulyk ML, Struhl K. STAT3 activation of miR-21 and miR-181b-1 via PTEN and CYLD are part of the epigenetic switch linking inflammation to cancer. Mol Cell. 2010;39:493–506. doi: 10.1016/j.molcel.2010.07.023. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp TableS1
NIHMS908065-supplement-Supp_TableS1.xlsx (51.7KB, xlsx)

RESOURCES