MicroRNAs in colorectal cancer: A comparative analysis of circulating and tissue microRNA levels

Abstract

BACKGROUND

Colorectal cancer (CRC) is one of the most common cancers worldwide. The gold standard screening methods for early detection and monitoring are colonoscopy and stool-based tests. However, innovative and minimally invasive biomarkers need to be integrated into clinical practice.

AIM

To identify circulating microRNAs as potential CRC biomarkers through a comparative analysis of tissue and plasma samples from patients with CRC.

METHODS

This case-control study conducted a quantitative real-time polymerase chain reaction analysis of 84 microRNAs in tumoral and peritumoral tissues, and 179 microRNAs in plasma from 19 patients with CRC. A control cohort for the tissue analysis and another control cohort for the plasma analysis have been enrolled.

RESULTS

In total, 14 microRNAs were significantly differentially expressed in the tissue and plasma samples. Notably, five microRNAs (miR-26b-5p, miR-101-3p, miR-30d-5p, miR-107, and miR-21-5p) presented the same trend in terms of fold change in both types of biological samples. Significant associations between the circulating levels of miR-21-5p and miR-26b-5p and lymphovascular invasion were found.

CONCLUSION

These five microRNAs with significantly altered levels in plasma and tumoral tissue, could be good non-invasive CRC biomarkers candidates, enhancing screening, and supporting precision and individualized patient care.

Key Words: Colorectal cancer; MicroRNAs; Biomarker; Plasma; Colonic mucosa; Lymphovascular invasion

Core Tip: In this case-control study, we analyzed the levels of 84 microRNAs in tumoral and peritumoral colonic mucosa of 19 patients with colorectal cancer compared to those expressed in the normal mucosa from 18 controls. For each patient, we also investigated the levels of 179 circulating microRNAs in comparison with a cohort of 16 controls. A relationship between tissue and plasma levels has been established along with correlations with tumoral features. Five microRNAs (miR-26b-5p, miR-101-3p, miR-30d-5p, miR-107, and miR-21-5p) presented the same trend in terms of fold change in both biological samples, suggesting them as putative non-invasive biomarkers for monitoring colorectal cancer progression.

INTRODUCTION

Colorectal cancer (CRC) is one of the most prevalent cancers globally and the second leading causes of cancer-related mortality[1,2]. Although potentially preventable, most cases are diagnosed after clinical manifestations, meaning that diagnosed patients tend to be at an advanced stage of the disease and have an average survival rate of five-years[1,3]. As CRC develops and progress, multiple genetic and epigenetic modifications[4,5] occur that alter physiological processes, such as proliferation and apoptosis, while also giving pathological properties to the tumor, such as invasion, metastasis, and drug resistance[6,7].

A study of CRC in Romania between 2016 and 2018 found that the mortality rate attributed to this type of cancer in the country, particularly in the northern and central regions, was nearly twice the European average[8]. This high mortality rate could have been because Romania did not have a CRC screening program at the time. Unlike other European countries, Romania only began screening for CRC in 2020 through a pilot program[9]. Screening tests can help identify CRC in less advanced and potentially curable stages. Multiple diagnostic methods and tools exist, but their implementation via a screening strategy is difficult because of their invasiveness, variations in precision, and socio-economic implications[1,3,7].

At present, colonoscopy and stool-based tests are the gold standard screening methods for early CRC detection. However, the use of these tests remains low in most developed countries[10]. Therefore, innovative and minimally invasive biomarkers with prognostic and predictive value for patient stratification are needed. Various approaches involving biological fluids may provide better alternatives to conventional cancer screening and monitoring. These new methods can enhance patient compliance and adherence to screening programs and facilitate the monitoring of cancer evolution.

MicroRNAs are short non-coding RNAs that have been shown to impact gene expression. By modifying different cellular processes, they also play a role in tumor development and progression[1,3,7]. MicroRNAs can be isolated from cells and tissues as well as from fluids and secretions, such as urine and plasma[6]. Circulating microRNAs may result from processes such as apoptosis and selective secretion, and they can even be found in structures such as circulating exosomes and micro vesicles. Analyses of circulating microRNAs can provide important additional information regarding the dynamic character of malignancies[11].

In CRC, a broad number of microRNAs have abnormal expression. As such, they could be used as diagnostic and prognostic markers[2,7,12]. However, the interindividual variability and heterogeneity of tumor characteristics, the presence of confounding factors, and the lack of validation studies across different patient cohorts represent major problems for integrating microRNAs into clinical practice as biomarkers[1]. Our study aimed to identify a panel of circulating microRNAs with modified levels that could represent putative CRC biomarkers. A multistep analysis was performed to identify specific circulating microRNAs that could reflect the microRNA pattern in tumoral tissue. The expression of a large panel of microRNAs in the tumoral tissue and plasma of the same CRC patients was evaluated. Dysregulated microRNAs found in both tissue and plasma were identified and correlated with tumoral features.

MATERIALS AND METHODS

Patients and sample collection

The present study included a tissue case-control study and a plasma case-control study. Tumoral tissue (T_tissue) samples and peritumoral tissue (PT_tissue) samples, which consisted of adjacent neoplasm-free mucosal tissues were surgically collected from 19 CRC patients at the Fundeni Clinical Institute in Bucharest, Romania. The diagnosis was histologically confirmed by a pathologist. To serve as a control, normal colonic mucosa (CTRL_tissue) samples were collected from 18 individuals without CRC or inflammatory bowel disease during colonoscopy screenings. All the collected tissues were preserved in RNAprotect Tissue Reagent (Qiagen, Hilden, Germany) up to 72 hours and, after removing the reagent, the samples were stored at -80 °C until RNA extraction.

Before surgery, 2 mL of blood (CRC_plasma) were collected from each CRC patient using EDTA tubes. To serve as a control, 2 mL of blood (CTRL_plasma) were collected from a third group which was comprised of 16 individuals without CRC or a personal history of any type of cancer or inflammatory condition. The blood samples were processed as soon as possible but no later than 1 hour after collection, and a two-steps centrifugation protocol was applied: (1) 1900 × g for 10 minutes, 4 °C to isolate the plasma; and (2) 16000 × g for 10 minutes, 4 °C to remove the presence of any cellular debris. The plasma was aliquoted into nuclease-free cryovials and stored at -80 °C until total RNA isolation.

RNA isolation and microRNA expression analysis

Total RNA has been extracted and purified using miRNeasy Mini Kit (Qiagen, Germany) according to the manufacturer’s protocol. The quantity and quality of the total RNA was verified using NanoDrop 2000, Thermo Scientific and considering the 260/280 nm and 260/230 nm parameters > 1.8. Reverse transcription using a constant quantity of 20 ng of total RNA was performed with miRCURY LNA RT Kit (Qiagen, Hilden, Germany). By quantitative real-time polymerase chain reaction (qRT-PCR) analysis on ABI-7500 fast instrument (Thermo Fisher Scientific, Inc., MA, United States), we assessed the expression of 84 microRNAs available in the Human Cancer Focus YAHS-102 (Qiagen, Hilden, Germany) using the miRCURY LNA SYBR Green PCR Kit (Qiagen, Hilden, Germany). Four microRNAs (miR-149-3p, miR-202-3p, miR-205-5p and miR-206) were excluded from the analysis due to their Ct values, which were above 35. Each array contains three RNAs for the normalization of the array data (SNORD49A, SNORD38B, U6snRNA). The geometric mean of SNORD49A, and SNORD38B levels, was used for the normalization of Ct data, since the analysis performed by the RefFinder algorithm indicated these two RNAs as the most stable in the entire cohort[13].

Total RNA from plasma samples was isolated from a constant volume of 200 μL of plasma using an miRNeasy Serum/Plasma kit (Qiagen, Germany) according to the manufacturer’s protocol. RNA spike-in controls (RNA Spike-in kit; Qiagen, Germany) were added to the lysis buffer for checking the quality of RNA extraction. The isolated RNA was resuspended in 20 μL of RNase-free water and the reverse transcription was immediately performed using 6 μL of total RNA and, 1 μL of RNA synthetic spike-ins in a total reaction volume of 20 μL. We assessed the plasma levels of 179 microRNAs using the Human serum/plasma focus, miRCURY LNA miRNA Focus PCR panel YAHS-406Z (Qiagen, Germany) by qRT-PCR. The Ct values were normalized on the geometric mean of the 103 most expressed microRNAs, applying the global normalization strategy for microRNAs[14]. The microRNA expression data are shown as fold change (FC) and as 2^-ΔCt values.

DNA isolation and gene mutation detection

DNA was isolated using QIAamp DNA Mini Kit (Qiagen, Germany), according to the manufacturer’ protocol. The quantity and quality of the total DNA was verified using NanoDrop 2000, Thermo Scientific and considering the 260/280 nm and 260/230 nm parameters > 1.8. KRAS (codons 12, 13, 59, 61, 117, 146) and BRAF (codon 600) mutations were detected using EasyPGX^® ready KRAS kit and EasyPGX^® ready BRAF kit, respectively. Gene mutations were identified through real-time polymerase chain reaction according to the manufacturer’s protocols on AriaMx instrument (Agilent, CA, United States) and EasyPGX analysis software (Diatech Pharmacogenetics, Italy).

Statistical analysis

Statistical analysis was performed using the Statistical Package for Social Sciences (SPSS version 20.0). Continuous variables were tested using the t-test, while categorical variables by means of the χ² test. The Shapiro-Wilk test was conducted to detect the normality of the data distribution of each microRNA level. Since the levels were not normally distributed, the Mann-Whitney U test was used to assess the differences in microRNA levels between T_tissue and CTRL_tissue, CRC_plasma and CTRL_plasma, and between the two groups obtained following stratification based on tumoral characteristics. Furthermore, for the 19 paired tissues (T_tissue vs PT_tissue) analysis was performed using the Wilcoxon signed-rank test. MicroRNA levels changes were considered significant between the groups when the P value was < 0.05. Graphical representations were generated using GraphPad Prism 8.4.

RESULTS

A total of 19 patients with CRC (52.6% females and 47.4% males, with ages between 54 and 83 years old) were enrolled in the present study. The sociodemographic and clinicopathologic characteristics of the patients are reported in Table 1. The CTRL_tissue and CTRL_plasma groups did not differ in terms of age (tissue: 63.28 ± 8.40; 45-77; plasma: 57.81 ± 5.24; 53-74) and sex distribution (tissue: 55.6% female; 44.4% male; plasma: 50.0% female; 50.0% male) from the CRC group (P = 0.153; P = 0.110 for age and P = 0.858, χ² = 0.032; P = 0.877, χ² = 0.024, for sex).

Table 1 Sociodemographic, and clinical characteristics of the enrolled cohort, mean ± SD (minimum-maximum).

Features	Total (n = 19)
Age	67.58 ± 9.43 (51-83)
Sex (%, female; male)	52.6; 47.4
Localization	Ascending colon (n = 5); transverse colon (n = 2); descending colon (n = 1); sigmoid colon (n = 4); rectosigmoid junction (n = 4); rectum (n = 3)
Grading (%, G1; G2; G3)	G1 15.8; G2 52.6; G3 31.6
TNM classification	T2 N0 M0 (n = 2); T3 N0 M0 (n = 7); T3 N1 M0 (n = 7); T4A N1 M0 (n = 1); T4B N0 M0 (n = 1); T4B N1 M0 (n = 1)
Microsatellite instability (%, yes; no)	10.5; 89.5
BRAF mutation (%, yes; no)	5.3; 94.7
KRAS mutation (%, yes; no)	36.8; 63.2
Lymphovascular invasion (%, yes; no)	42.1; 57.9
Perineural invasion (%, yes; no)	26.3; 73.7
Hemoglobin (g/dL)	11.77 ± 2.84 (7.10-17.50)
White blood cells (× 103/μL)	7616.84 ± 2749.57 (4880.00-14840.00)
Platelets (× 103/μL)	287.79 ± 138.51 (152.00-777.00)
International normalized ratio	1.04 ± 0.08 (0.87-1.22)
Fibrinogen (mg/dL)	435.42 ± 129.52 (275.10-796.00)
Albumin (g/dL)	3.98 ± 0.69 (2.40-5.48)
Cholesterol (mg/dL)	161.46 ± 39.95 (73.00-227.00); available for 18 patients
Triglycerides (mg/dL)	94.19 ± 24.84 (36.00-134.00); available for 17 patients
Alanine aminotransferase (U/L)	23.86 ± 26.38 (5.62-120.00); available for 17 patients
Aspartate transaminase (U/L)	22.16 ± 11.53 (13.00-65.00); available for 18 patients
Urea (mmol/L)	36.01 ± 13.40 (15.00-59.50); available for 18 patients
Creatinine (μmol/L)	0.96 ± 0.22 (0.55-1.33); available for 18 patients
Uric acid (μmol/L)	4.91 ± 1.77 (2.30-8.91); available for 18 patients
Carcinoembryonic antigen (ng/mL)	41.18 ± 112.53 (0.65-444.00); available for 15 patients
Cancer antigen 19-9 (U/mL)	29.22 ± 50.82 (3.47-211.00); available for 16 patients
Alpha-fetoprotein (ng/mL)	3.62 ± 2.88 (1.07-9.40); available for 9 patients

TNM: Tumor-nodes-metastasis.

Comparison of microRNA levels in tissue and plasma

In the first analysis, we compared the microRNA levels in the T_tissue and PT_tissue samples and found that 42 microRNAs were differentially expressed (Supplementary Table 1). The tissues case-control analysis comparing T_tissue and CTRL_tissue identified a panel of 33 significant microRNAs (Supplementary Table 2), while the plasma case-control analysis revealed 45 significant microRNAs (Supplementary Table 3). Of the 42 microRNAs identified in the T_tissue vs PT_tissue comparison, 30 microRNAs were present in the array used to analyze the plasma, and 14 microRNAs were significantly dysregulated in both the tumoral tissue and the plasma. A graphical abstract summarizing the workflow of the study is depicted in Figure 1.

Figure 1 Graphical abstract summarizing the study workflow. qRT-PCR: Quantitative real-time polymerase chain reaction; T: Tumoral; CTRL: Control; PT: Peritumoral; DE: Differentially expressed; CRC: Colorectal cancer; LVI: Lymphovascular invasion.

Five microRNAs, miR-26b-5p, miR-101-3p, miR-30d-5p, miR-107 and miR-21-5p, presented the same trend of expression in the tissue and plasma samples. The last one (miR-21-5p) was less present in the PT_tissue than in CTRL_plasma (Table 2 and Figure 2). Nine microRNAs presented statistically significant results for the tissue and plasma samples with the opposite trend. The results are presented in Table 3.

Figure 2 The levels of miR-26b-5p, miR-101-3p, miR-30d-5p, miR-107 and miR-21-5p in tumoral vs peritumoral and plasma from colorectal cancer patients vs control plasma. Bar graphs represent the fold change values, and error bars represent the standard error of the mean. Statistical significance was calculated using the Wilcoxon signed-rank test in tumoral vs peritumoral analysis, and the Mann-Whitney test in colorectal cancer plasma vs control plasma analysis. T: Tumoral; CTRL: Control; PT: Peritumoral; CRC: Colorectal cancer. ^aP < 0.05 vs peritumoral, ^bP < 0.05 vs control plasma, ^cP < 0.01 vs peritumoral, ^dP < 0.01 vs control plasma, ^eP < 0.001 vs peritumoral, and ^fP < 0.001 vs control plasma.

Table 2 MicroRNAs differentially expressed in tissue and plasma, mean ± SD.

MicroRNA	2-∆Ct		FC	P value2	2-∆Ct		FC	P value3
	T	PT			CRC_plasma	CTRL_plasma
miR-26b-5p1	0.372 ± 0.29	0.636 ± 0.39	0.58	< 0.001	0.663 ± 0.33	0.901 ± 0.25	0.73	0.007
miR-101-3p1	1.125 ± 0.80	0.203 ± 0.13	0.61	0.01	1.177 ± 0.81	2.730 ± 0.90	0.64	0.001
miR-30d-5p1	0.270 ± 0.17	0.430 ± 0.32	0.63	0.007	1.677 ± 0.75	2.825 ± 1.02	0.59	< 0.001
miR-1071	0.505 ± 0.25	0.720 ± 0.41	0.70	0.014	1.195 ± 0.50	1.915 ± 1.10	0.62	0.011
miR-21-5p1	15.493 ± 8.49	8.140 ± 6.32	1.90	0.02	10.077 ± 19.87	4.185 ± 1.31	2.40	0.025
miR-145-5p	2.193 ± 1.95	26.580 ± 33.48	0.08	0.001	1.238 ± 0.825	0.350 ± 0.30	3.53	< 0.001
miR-143-3p	1.740 ± 1.69	10.532 ± 12.02	0.16	0.001	0.491 ± 0.35	0.888 ± 0.04	5.52	< 0.001
miR-125b-5p	0.650 ± 0.63	3.295 ± 3.46	0.19	0.001	1.078 ± 1.43	0.392 ± 0.23	2.74	0.002
let-7e-5p	0.240 ± 1.64	0.488 ± 0.46	0.49	0.02	0.118 ± 0.06	0.070 ± 0.28	1.69	0.009
let-7f-5p	0.395 ± 0.25	0.757 ± 0.44	0.52	0.002	0.132 ± 0.08	0.490 ± 0.02	2.71	< 0.001
miR-215-5p	1.690 ± 1.16	3.038 ± 3.00	0.55	0.036	1.122 ± 2.22	0.309 ± 0.11	3.62	0.006
miR-23b-3p	1.367 ± 0.98	2.414 ± 2.06	0.56	0.02	1.501 ± 1.17	0.760 ± 0.40	1.97	0.002
let-7c-5p	1.123 ± 1.73	1.792 ± 1.26	0.62	0.005	0.513 ± 0.21	0.270 ± 0.13	1.89	< 0.001
miR-18a-5p	0.708 ± 0.06	2.193 ± 1.95	2.65	0.049	0.340 ± 0.19	0.473 ± 0.13	0.71	0.009
miR-103a-3p	0.705 ± 0.40	1.025 ± 0.90	0.68	0.022	1.656 ± 0.87	2.730 ± 1.95	0.60	NS
miR-99a-5p	0.462 ± 0.47	0.284 ± 0.41	0.16	0.001	0.169 ± 0.17	0.135 ± 0.09	1.39	NS
miR-10b-5p	0.173 ± 0.11	0.601 ± 0.53	0.28	0.001	0.167 ± 0.08	0.141 ± 0.08	1.24	NS
miR-150-5p	0.308 ± 0.40	0.804 ± 0.90	0.38	0.001	2.715 ± 1.69	5.380 ± 8.72	0.56	NS
miR-26a-5p	1.340 ± 1.03	3.363 ± 2.57	0.39	0.001	0.647 ± 0.36	0.573 ± 0.40	1.16	NS
miR-30c-5p	0.227 ± 0.17	0.563 ± 0.40	0.40	0.001	0.500 ± 0.23	0.527 ± 0.21	0.95	NS
miR-130a-3p	0.036 ± 0.03	0.086 ± 0.07	0.41	0.004	0.496 ± 0.24	0.501 ± 0.24	0.92	NS
let-7a-5p	2.222 ± 1.69	5.198 ± 3.62	0.42	0.002	1.272 ± 0.53	1.140 ± 0.60	1.19	NS
let-7b-5p	1.078 ± 0.94	2.413 ± 1.82	0.44	0.004	7.238 ± 3.71	7.003 ± 1.50	1.03	NS
miR-126-3p	0.866 ± 0.80	1.747 ± 1.47	0.49	0.003	3.996 ± 2.55	3.193 ± 1.02	1.37	NS
miR-30b-5p	0.346 ± 0.17	0.650 ± 0.45	0.53	0.001	0.491 ± 0.24	0.536 ± 0.50	0.88	NS
let-7 g-5p	0.616 ± 0.57	1.078 ± 0.76	0.57	0.003	1.602 ± 0.55	1.995 ± 1.09	0.78	NS
miR-27b-3p	0.060 ± 0.61	1.780 ± 1.18	0.59	0.005	0.841 ± 0.76	0.461 ± 0.18	1.88	NS
miR-22-3p	0.168 ± 0.09	0.230 ± 0.15	0.73	0.014	2.028 ± 0.81	2.757 ± 1.14	0.74	NS
miR-29c-3p	0.800 ± 0.34	0.970 ± 0.64	0.82	0.04	0.914 ± 0.39	0.867 ± 0.38	1.02	NS
miR-16-5p	2.246 ± 2.64	2.561 ± 1.69	0.87	0.024	132.038 ± 56.12	164.355 ± 57.78	0.78	NS

¹The five microRNAs differentially modified both in tissue and plasma show the same trend in term of fold change.

²Tissue, paired analysis (19 tumoral tissue vs 19 peritumoral tissue).

³Plasma case-control (19 colorectal cancer plasma vs 16 control plasma).

T: Tumoral; PT: Peritumoral; CTRL: Control; CRC: Colorectal cancer; FC: Fold change; NS: Not significant.

Table 3 MicroRNAs differentially expressed in tissue and plasma with an opposite trend.

MicroRNA	Tissue paired (19 T_tissue vs 19 PT_tissue)	Plasma case-control (19 CRC_plasma vs 16 CTRL_plasma)
miR-145-5p	Decrease, FC = 0.08, P = 0.001	Increase, FC = 3.53, P < 0.001
miR-143-3p	Decrease, FC = 0.16, P = 0.001	Increase, FC = 5.52, P < 0.001
miR-125b-5p	Decrease, FC = 0.19, P = 0.001	Increase, FC = 2.74, P = 0.002
let-7e-5p	Decrease, FC = 0.49, P = 0.02	Increase, FC = 1.69, P = 0.009
let-7f-5p	Decrease, FC = 0.52, P = 0.002	Increase, FC = 2.71, P < 0.001
miR-215-5p	Decrease, FC = 0.55, P = 0.036	Increase, FC = 3.62, P = 0.006
miR-23b-3p	Decrease, FC = 0.56, P = 0.02	Increase, FC = 1.97, P = 0.002
let-7c-5p	Decrease, FC = 0.62, P = 0.005	Increase, FC = 1.89, P < 0.001
miR-18a-5p	Increase, FC = 2.65, P = 0.049	Decrease, FC = 0.71, P = 0.009

T: Tumoral; PT: Peritumoral; CTRL: Control; CRC: Colorectal cancer; FC: Fold change.

Association between circulating microRNA levels and tumoral characteristics

We conducted further analysis of the five microRNAs with the same trend in terms of FC in the tissue and plasma samples to find a possible association between their levels and tumoral characteristics. We stratified the patients based on whether a lymphovascular invasion (LVI) was present and found significant differences in the miR-21-5p and miR-26b-5p levels of patients with this invasion (LVI+) and those without (LVI-): FC = 0.31; P = 0.020 and FC = 1.77; P = 0.009, respectively (Figure 3). We also explored the association between microRNA levels and tumoral features, such as lymph node involvement, perineural invasion, and KRAS and BRAF mutations, but we did not find any significant results.

Figure 3 Levels of miR-21-5p and miR-26b-5p in plasma from patients with lymphovascular invasion (lymphovascular invasion+) vs non-vascular invasion (lymphovascular invasion-). Bar graphs represent the fold change values, and error bars represent the standard error of the mean. Statistical significance was calculated using the Mann-Whitney test. LVI: Lymphovascular invasion. ^aP < 0.05, and ^bP < 0.01.

DISCUSSION

CRC is recognized as the second most common cause of cancer-related mortality worldwide. It has been acknowledged that impaired microRNA expression plays an important role in CRC development and progression[1]. Although colonoscopy is the gold standard for CRC detection, less invasive and more affordable methods are under investigation for early detection and disease monitoring. The detection of circulating microRNAs could offer insight into solid tumors and serve as an innovative, non invasive, and low-risk alternative[6,15]. It has been reported that microRNAs are involved CRC development and progression by regulating signaling pathways, such as Hippo, Notch, and Wnt/β-catenin pathways[16]. Numerous research groups are investigating this hypothesis and finding aberrant microRNAs levels in tissues, blood, urine, and stool samples of CRC patients[12]. However, microRNAs have not been introduced as biomarkers in clinical practice.

In the present study, we conducted an extensive qRT-PCR analysis of 179 microRNAs in plasma and 84 microRNAs in tumoral tissue to identify a panel of circulating microRNAs that could mirror their expression in the tumor and represent putative CRC biomarkers. The results showed that 14 microRNAs were significantly differentially expressed in both tissue and plasma. Notably, five microRNAs (miR-26b-5p, miR-101-3p, miR-30d-5p, miR-107, and miR-21-5p) presented the same trend in terms of FC in tissue and plasma samples from the CRC patients. Specifically, miR-26b-5p, miR-101-3p, miR-30d-5p, and miR-107 levels were significantly lower in T_tissue than in PT_tissue and lower in CRC_plasma than in CTRL_plasma, while miR-21-5p levels were higher in the tumor condition.

It has been reported that miR-26b-5p acts as a tumor-suppressor microRNA in many types of cancer, such as hepatocellular carcinoma, human pulmonary cancer, and breast cancer[17-19]. It has been demonstrated that miR-26b can target the expression of the LEF1 gene, which is highly expressed in colon cancer cells. In fact, miR-26b is a potent candidate for gene therapy because it inhibits the proliferation of colon cancer cells by blocking the Wnt/β-catenin-dependent activation of LEF1[20]. Using qRT-PCR, Li et al[21] compared miR-26b expression in 38 CRC tissues and adjacent noncancerous tissues. In line with our results, they found significantly lower levels of this microRNA in the CRC tissue than in the noncancerous tissue. Interestingly, another group found that the exosomal levels of miR-26b-5p were significantly decreased in patients in stages II and III of CRC compared to the control group[22]. However, to our knowledge, the plasma levels in CRC patients have not yet been investigated by any research group.

Additionally, miR-101-3p plays an important role in CRC progression and has been reported to be notably downregulated in CRC cell lines and tumoral tissues, indicating that its overexpression inhibits the invasion, migration, and proliferation of cancer cells[7,23]. These results are supported by a study that investigated the interaction between miR-101-3p and circALPLP2 and found that the latter promoted CRC progression and metastasis by sponging the former[24]. These findings, as well as the finding that miR-101 targets the cAMP responsive element binding protein 1, which is expressed at higher levels in colon cancer tissue compared to adjacent normal tissue, may suggest a novel mechanism for CRC treatment based on targeting the miR-101/cAMP responsive element binding protein 1 pathway[23].

Meanwhile, miR-30d-5p was shown to display low levels in exosomes isolated from different CRC cell lines[25]. Both miR-30d-5p and miR-107 were downregulated in many cancers, including CRC, demonstrating their potential role as tumor suppressors[2,26,27]. Notably, miR-107 has also been shown to promote chemoresistance in CRC[28]. Along these lines, Zhang et al[29] suggested that miR-30d could be a potential clinical biomarker and therapeutic target since its overexpression can decrease cell autophagy and increase cell apoptosis, thus inhibiting CRC cells proliferation.

In line with our results, miR-21-5p has been found to be overexpressed in various human tumors, acting as an onco-microRNA and playing a role in tumor growth, invasion, and metastasis[30,31]. Its levels were also reported to be higher in serum samples from 80 patients with CRC compared with 50 healthy controls[30]. Furthermore, a CRC study that compared postoperative and follow-up samples collected three months after surgery found that the level of circulating miR-21-5p was significantly decreased[32]. Additionally, a comparative study of whole blood from CRC patients and healthy controls revealed that miR-21 overexpression was significantly associated with tumor characteristics, such as size, lymph node metastasis, liver metastasis, advanced clinical stage, and early onset[33]. These results highlight the clinical significance of miR-21 as a potential indicator of disease aggressiveness and as a diagnostic and disease monitoring biomarker.

Regarding LVI status, the level of miR-21-5p in plasma was significantly higher in LVI- patients compared to LVI+ patients. In the tissue samples, the level of miR-21-5p was higher in LVI+ patients (17.067 ± 9.84) compared to LVI- patients (14.348 ± 7.64), although it did not reach statistical significance. The inverse pattern of miR-21-5p expression in the tissue and plasma samples of the LVI+ patients might have been due to the altered secretion or retention of miR-21-5p by invasive tumors. This reflects the complex biology of microRNA dynamics and highlights the need to interpret tissue and plasma data as complementary, but distinct sources of information[3,34]. Meanwhile, the LVI- patients had lower levels of miR-26b-5p in tissue and plasma compared to the LVI+ patients.

The levels of one of the previously mentioned nine microRNAs, miR-145-5p, were lower in T_tissue than in PT_tissue and higher in CRC_plasma than in CTRL_plasma. Chen et al[35], also reported lower levels of this microRNA in CRC tissues compared to normal tissues, while another study found that elevated miR-145 levels in plasma were linked to an increased risk of developing cancer[36]. This finding suggests that there is a complex regulatory mechanism influencing the release of microRNA into circulation. It could also support findings in which microRNAs levels in plasma do not always mirror those found in the primary tumor[3].

Taken together, our study proves that when looking for CRC biomarkers, microRNA levels in both tissue and plasma should be examined. Still, certain microRNAs may exhibit differential expression in tissue or plasma, and both might be suggestive of malignancy. The absence of a correlation between microRNA levels in tumors and plasma has been explained by several theories. For example, circulating microRNAs have half-lives that vary from minutes to hours, which could account for variations in their concentrations in circulation and the potential lack of correlation between their expressions in the same patient’s tumor and plasma.

This study has two main limitations: (1) It focused on a relatively small cohort of patients. Therefore, an extensive study with a larger cohort is needed to validate the findings; and (2) Although qRT-PCR is commonly used to analyze microRNAs in tissue, microRNAs have a low abundance in plasma and are susceptible to degradation. Therefore, our plasma sample results should be validated using more sensitive methods, such as digital polymerase chain reaction.

CONCLUSION

Our findings suggest that miR-26b-5p, miR-101-3p, miR-30d-5p, miR-107, and miR-21-5p circulating levels could mirror their expression in tumors. These findings could indicate that these microRNAs are good non-invasive biomarker candidates for monitoring CRC.