DNA methylation and miRNA expression in colon adenomas compared with matched normal colon mucosa and carcinomas

Abstract

Dysregulation of DNA methylation patterns and non‐coding RNA, including miRNAs, has been implicated in colon cancer, and these changes may occur early in the development of carcinoma. In this study, the role of epigenetics as early changes in colon tumorigenesis was examined through paired sample analysis of patient‐matched normal, adenoma and carcinoma samples. Global methylation was assessed by genomic 5‐methyl cytosine (5‐mC) and long interspersed nuclear element‐1 (LINE‐1) promoter methylation by pyrosequencing. KRAS mutations were also assessed by pyrosequencing. Expression of miRNA, specifically, two microRNA genes—miR‐200a and let‐7c—was analysed using RT‐qPCR. Differences in global methylation in adenomas were not observed, compared with normal tissue. However, LINE‐1 methylation was decreased in adenomas (p = .056) and carcinomas (p = .011) compared with normal tissue. Expressions of miRNA, miR‐200a and let‐7c were significantly higher in adenomas than normal tissues (p = .008 and p = .045 respectively). Thus the significant changes in LINE‐1 methylation and microRNA expression in precancerous lesions support an early role for epigenetic changes in the carcinogenic process. Epigenetic characteristics in adenomas may provide potential diagnostic and prognostic therapeutic targets early in cancer development at the adenoma stage.

1. INTRODUCTION

Colon cancer results from genetic and epigenetic changes that contribute to the malignant cell phenotype. Colon cancers are categorized into specific phenotypes based on: (a) chromosomal instability (CIN) manifested as aneuploidy, chromosomal rearrangements and accumulation of somatic mutations in oncogenes such as KRAS and BRAF ¹ , ² , ³ ; (b) the presence of microsatellite instability (MSI) driven, characterized by DNA mismatch repair defects; and (c) MLH1 promoter methylation and the CpG island methylator phenotype (CIMP). ⁴ Fearon and Vogelstein first described the accumulation of somatic mutations driving normal ‐ adenoma ‐ carcinoma progression. ⁵ Modifications have been made to the model with the inclusion of epigenetic processes such as loss of gene expression due to hypermethylation of gene promoters as seen in CIMP, ⁶ decreased methylation of intergenic DNA (hypomethylation) ⁷ and aberrant gene regulation by microRNA. ⁸

Genomic DNA methylation patterns are altered in most human carcinomas, with discrete DNA segments becoming hypermethylated and intergenic sequences becoming hypomethylated. ⁹ Hypomethylation, particularly in transcriptional control areas of transposable elements, promotes genomic instability including deletions, insertions and recombination. ¹⁰ The most abundant retrotransposon, long interspersed nuclear element‐1 (LINE‐1), compromises approximately 17% of the human genome. ¹¹ In normal cells, heavy DNA methylation suppresses transcription of the RNA intermediates required for the movement of LINE‐1 elements, which leads to genomic instability. Hypomethylation of LINE‐1 elements has been observed in colon cancer cells and precancerous lesions supporting its use as a biomarker for the assessment of malignancy or potential malignancy. ¹²

MicroRNAs are components of a post‐transcriptional epigenetic regulation system that inhibits protein translation by base‐pairing with target mRNAs. ¹³ Aberrations in miRNA biogenesis pathways and/or regulation can contribute to numerous diseases including cancer. ¹⁴ , ¹⁵ miRNA genes have been categorized into groups, or families, based on the mature miRNA, sequence and structure of their precursor species (pre‐miRNAs).

The first known human miRNA, let‐7, and its family members are highly conserved across species in sequence and function. Genes encoding the let‐7 family map to 21q21.1, one of several chromosomal regions recurrently deleted in cancers. ¹⁴ The let‐7c miRNA may be involved in the regulation of oncogenes such as the KRAS gene, which is frequently mutated in CIN colon cancer, including over 30% of colon carcinomas. ¹⁶ The oncogene, CDC25A, required for progression from G1 to the S phase of the cell cycle, may also be a target of let‐7c. ¹⁷ It has been proposed that let‐7c regulates cancer stem cell maintenance by repressing self‐renewal and promoting differentiation in both normal development and cancer. ¹⁸

The microRNA‐200 (miR‐200) family also has strong potential for regulating carcinoma progression. ¹⁹ The miR‐200 family is composed of five members, miR‐200a/b and miR‐429 (on chromosome 1), and miR‐200c and miR‐141 on chromosome 12. ²⁰ , ²¹ The miR‐200 family is upregulated in cancer. ²² MiR‐200a regulates important signalling pathways in certain cancers. ²³ Expression of miR‐200c has been proposed to regulate stem cell renewal, providing a molecular link between normal and cancerous stem cells. ²⁴

The current study investigates the role of epigenetics including DNA methylation and miRNA in early colon cancer tumorigenesis by analysis in paired tissue types. The results support the potential for epigenetic biomarkers in early colon cancer detection.

2. METHODS

2.1. Patients

Archival formalin‐fixed paraffin‐embedded (FFPE) tissue samples were obtained from the Rush University Medical Center Pathology Department, and the study was approved by the Rush University Medical Center Institutional Review Board (IRB# 12121202). Seven 5‐micron sections were cut from separate blocks of paraffin‐embedded tissue for each tissue type and placed on individual glass slides. One slide was stained with haematoxylin and eosin (H&E) for pathological assessment of normal, adenoma and carcinoma for each patient. The criteria for inclusion of tissue in the study were the source of the tissue (normal colon, adenoma and carcinoma) and at least 2 × 2 mm² in size. Only tubular adenomatous adenomas were included, and sessile serrated adenomas were excluded. Tissue described as normal was selected from independent uninvolved tissue or non‐malignant tissue at least 100 mm from the involved tissue on the same section when separate uninvolved tissue sections were not available. The TNM stages ranged from T1NO to T4N2b and were classified into two categories (pT1 to pT2 and pT3 to pT4) for statistical analysis. Immortalized colon cancer cells, strain SW48 [SW‐48], were acquired from the American Type Culture Collection (ATCC^® CCL‐231™).

2.2. RNA analysis

RNA was extracted from unstained tissue sections using the RecoverAll™ Total Nucleic Acid Isolation Kit (Thermo Fisher Scientific) according to the manufacturer's recommendations. The isolated RNA was converted into complementary DNA (cDNA) using Moloney murine leukaemia virus reverse transcriptase (Invitrogen).

miRNA expression relative to beta‐2 microglobulin (β2M) internal control was measured from the cDNA using qRT‐PCR with TaqMan^® probes (Thermo Fisher Scientific). The primer sequences used were as follows: miR‐200a forward (5′‐GCGGGTCACCTTTAACA‐3′) and miR‐200a reverse (5′‐GTGAGCATCTTACCGGACAG‐3′), and miR‐200a probe sequence (5′‐/56‐FAM/CGTTACCAG/ZEN/ACAGTGTTAGAGTCAAGCTG/31ABkFQ/‐3′); let‐7c forward (5′‐GCTCCAAAGGAAAGCTAGAAGG‐3′), let‐7c reverse (5′‐CATCCGGGTTGAGGTAGT‐3′) and let‐7c probe (5′‐/56‐FAM/CTCCCAGGG/ZEN/TGTAACTCTAAACCATACAAC/31ABkFQ/‐3′); β2M forward (5′‐ACCTCCATGATGCTGCTTAC‐3′), β2M reverse (5′‐GGACTGGTCTTTCTATCTCTTGTAC‐3′) and β2M probe (5′‐/56‐FAM/CTGCCGTGT/ZEN/GAACCATGTGACTTTG/31ABkFQ/‐3′).

Amplification was performed on a Quant 6 Studio thermal cycler (Life Technologies) using the universal program. β2M and test gene cycle threshold numbers (C _T values) were collected for each sample. C _T was converted to number of transcripts by linear regression to a standard curve of serially diluted synthetic oligomers of known molarity (IDT Technologies). Samples showing no expression of β2‐microglobulin were excluded from analysis. Results were expressed as per cent of miR‐200a or let‐7c transcripts relative to number of transcripts of the β2M internal control. The miRNA expression levels were compared between normal and adenomas; normal and carcinoma; and adenomas and carcinomas.

2.3. DNA analysis

Genomic DNA was extracted from unstained 4‐micron FFPE tissue sections for global methylation and LINE‐1 hypomethylation studies. Using the corresponding reviewed H & E slides, the tissue was microdissected by scalpel blade into 30–150 µl of 1.0 mg/ml proteinase K in 10 mM Tris/1 mM EDTA/mM KCl and incubated overnight at 57°C. The proteinase K was then inactivated through a 5‐min heat treatment at 95°C. The resulting crude lysate concentration was measured by Nanodrop LITE (Thermo Scientific) for ELISA and PCR procedures.

2.3.1. Global methylation

Global methylation was measured by enzyme‐linked immunoassay (ELISA) of 5‐methyl cytosine (5‐mC) detected in 100–300 ng sample DNA. The assay was performed using DNA isolated from normal, adenoma and carcinoma archived tissues and SW48 cells (Methylated DNA Quantification Kit; EpiGentek) according to manufacturer's protocol. After the colorimetric signal was read on a microplate reader, the data were used to quantify the relative and absolute amounts of 5‐mC. The equation used to calculate the relative amount of 5‐mC was

$$\text{5 - mC\%}=\frac{(\text{sample}\text{OD - NC}\text{OD})/\mathrm{S}}{((\text{PC}\text{OD - NC}\text{OD})/\mathrm{P})\times 100\%}$$

where OD is the optical density (absorbance) obtained from the microplate reader, S is the amount of sample DNA used in ng, and P is the amount of positive control in ng, and NC and PC are negative control and positive control respectively.

2.3.2. LINE‐1 hypomethylation

Bisulphite conversion of DNA extracted from the fixed tissue samples was performed using the reaction protocol outlined in EZ DNA Methylation™ kit (Zymo Research, Inc.). Ten microliters of crude lysate DNA was used. Converted DNA was desulphonated to avoid strand cleavage, cleaned by column washing and eluted for use.

The LINE‐1 retrotransposon assessed in this study is located on chromosome 22q11‐q12; genomic coordinates (GRCh38): 22:15,000,000–37,200,000; the primer sequences were based on repeat elements (locus X58075:111‐358). Analysis was based on LINE‐1 sequence with GenBank accession number ONS374723. The bisulphite‐converted DNA was amplified and subjected to pyrosequencing on a Pyromark Q24 pyrosequencer (Qiagen), programmed with the following sequence to analyse: TYGATTTTTTAGGTGYGTTYGTTA. The average of the relative per cent C (methylated) vs. T (unmethylated) at each of three CpG sites was reported. Non‐CpG cytosines were included in each sequence to confirm complete conversion.

2.3.3. KRAS mutation analysis

KRAS gene mutations were assessed by pyrosequencing using the KRAS Pyro Kit (Qiagen, Inc.) according to manufacturer's protocol. Five microliters of extracted DNA was amplified for sequencing. A 15‐µl aliquot of each PCR product was used for pyrosequencing to analyse KRAS codons 12, 13 and 61.

2.4. Cell culture methods

The SW48 (ATCC^® CCL‐231) colon cancer cell line was authenticated using the PowerPlex Hs15 STR system (Promega) and analysed in GeneMapper to verify the short tandem repeat genotype (CSF1PO 9/10, D13S317 11/12, D16S539 11/13, D5S818 10/14, D7S820 9/10, THO1 6/9.3, TPOX 8 and VWA 18/20/21). After authentication, the cells were grown in McCoy's 5A culture medium (Mediatech, Inc.).

MirVana miRNA mimics, has‐miR‐200a‐5p and has‐let‐7c‐5p (Ambion, Life Technologies Corp.) were introduced into the cells according to the DharmaFECT1 Transfection Protocol (courtesy of Dr Carl Maki, Rush University), and the cells were cultured to near confluence in antibiotic free McCoy's 5A (1×) medium using 6‐well plates. Five microliters of siRNA, 800 µl of medium and 2.5 µl of DharmaFECT1 reagents were added into each well. The cells were incubated at 37°C with 5% CO₂. The phenotype of the transfected cells was compared to that of negative control transfected cells (scrambled siRNA and untreated cells) at 24‐, 48‐ and 72‐h time points. Decreased expression of the genes targeted for the knockdown, miR‐200a and let‐7c was confirmed by qRT‐PCR of RNA isolated from the cell cultures.

2.5. Statistical analysis

Global methylation results were analysed using Wilcoxon signed rank‐sum paired two‐sample test (IBM spss Inc., Software Version 19). Results with p < .05 were considered statistically significant. Basic summary statistics were calculated for per cent methylation of the LINE‐1 promoter at CpG sites. The difference in methylation levels among groups was assessed by two‐sample t‐tests and Mann–Whitney tests.

For RNA expression, per cent (miR‐200a copies/β2M copies) and per cent (let‐7c copies/β2M copies) were analysed using paired sample and mixed linear models in spss. The relationship between miR‐200a or let‐7c and tissue type was evaluated by Spearman rank correlation. The relationships between miR‐200a or let‐7c and race, miR‐200a or let‐7c and gender (male/female), and miR‐200a or let‐7c and age at diagnosis were examined using Mann–Whitney 2‐independent samples test. Expression differences of miR‐200a or let‐7c were assessed against clinico‐pathological features (pT1‐pT4) using the Mann–Whitney test.

3. RESULTS

3.1. Patient demographics

A total of 148 tissue samples (55 normal, 52 tubular adenomas and 42 colon cancers) from 55 patients were included in the study (Table 1). Where cancer stage was known, the majority of carcinomas (78%) were later stage (pT3 to pT4). All adenomas were tubular adenoma in type. The degree of adenoma dysplasia was not included. The patients were equally represented by gender and race, and 62% of the patients were over 65 years of age.

TABLE 1.

Patient demographics

Characteristics	Number of patients
Total	55
Age
<65	21
>65	34
Gender
Male	29
Female	26
Ethnicity
White	30
Black	25
Carcinoma pathologic stage
pT1 to pT2	9
pT3 to pT4	31
Other/Unknown	15

3.2. Genomic DNA and LINE‐1 methylation levels

Genomic methylation was measured by ELISA. Levels of 5‐mC per sample ranged from 31.9 to 511.3 ng, 5.9 to 780.0 ng and 41.6 to 559 ng in normal, adenomas and cancer tissue DNA respectively. Relative per cent 5‐mC levels measured by Wilcoxon signed rank test were higher in adenomatous tissue than normal and cancer tissue (Table 2). A trend towards increased median relative methylation in normal to adenomas to carcinoma tissue was observed when comparisons were made as independent samples. There was considerable variability in the 5‐mC levels within groups, however (Figure 1A). Although comparisons did not reach significance compared as pairwise or independent samples, the results are consistent with intermediate methylation levels in adenomas compared with normal and cancer tissue.

TABLE 2.

Global methylation as measured by levels of 5‐mC relative to positive control

Tissue	Mean ± SD	p‐Value a
Normal Adenomas	12.5 ± 18.3 14.1 ± 17.3
		.339
Normal Carcinoma	12.5 ± 18.3 10.4 ± 8.5
		.469
Adenomas Carcinoma	14.1 ± 17.3 10.4 ± 8.5
		.475

^a Wilcoxon signed rank test.

FIGURE 1.

(A) Global methylation levels, relative to a positive control, in normal, adenoma and cancer tissue as measured by ELISA. Although there seems to be a trend, there was no significant difference in relative methylation levels across the sample types. (B) LINE‐1 methylation is significantly lower in carcinoma tissue compared to normal tissue (p = .011). LINE‐1 methylations in adenomas was marginally lower than that in normal tissue (p = .056)

To address intergenic methylation, LINE‐1 methylation levels were measured in the three groups. Independent t‐tests revealed significant LINE‐1 hypomethylation in carcinoma tissue compared with normal tissue (p = .011; Figure 1B). LINE‐1 methylation was decreased in adenomas compared with normal tissue (p = .056) and not significantly different from cancer tissue (p = .316). The data are consistent with LINE‐1 hypomethylation associated with malignant cellular phenotype and the possibility of early hypomethylation present in adenomas. The intermediate LINE‐1 hypomethylation in adenomas is consistent with this process occurring in the precancerous state.

3.3. miRNA expression

If dysregulation of miRNA has a role in tumorigenesis, and if that change occurs early in the process, then a change in expression of miR‐200a and let‐7c should occur in adenomas and carcinoma tissue compared with normal tissue. To investigate this possibility, the expression levels of miR‐200a in matched normal, adenomas and carcinoma samples were measured by qRT‐PCR. Expression was calculated as per cent of miRNA transcripts relative to number of transcripts of the β2M internal control.

Statistical comparisons as pairwise samples in log‐transformed data were performed using linear mixed models. As shown in Figure 2A, log expression of miR‐200a in adenomas was significantly higher than in normal tissues (p = .008), while log expression of miR‐200a in carcinoma tissues compared with normal tissue was variable and not statistically significant (p = .220; p = .204 respectively). Significant, possibly transient, aberrations in miR‐200a expression in adenomas support miRNA expression changes occurring in the precancerous adenomas.

FIGURE 2.

(A) miR‐200a expression levels in normal, adenoma and cancer tissue normalized to beta‐2‐microglobulin. The miR‐200a expression level was significantly higher in adenomas compared to normal but not cancer tissue. (B) let‐7c expression was significantly higher in adenomas and carcinoma compared to normal

Statistical comparisons of let‐7c expression measured as paired samples by linear mixed model analysis of log‐transformed data revealed log expression of let‐7c in adenoma samples significantly higher than in normal samples (p = .045; Figure 2B). Log expression of let‐7c in carcinoma samples was significantly higher than normal tissue (p = .019; Table 2). There was no significant difference between adenoma and cancer samples (p = .757). This is an example of dysregulation of miRNA occurring in precancerous adenomas and persisting in cancerous tissue.

The effect of confounding factors such as age of the patients at the time of diagnosis, gender and reported race was assessed. There was no significant association between age of the patient and the expression of miR‐200a/let‐7c; gender and miR‐200a/let‐7c expression; race and miR‐200a/let‐7c; and pathologic staging and miR‐200a/let‐7c expression across the three sample types. There was a marginal difference in expression of miR‐200a by race (p = .071). Comparison of the relationship of the expression of miR‐200a and let‐7c using Spearman rank correlation showed no statistically significant difference between the two (r = .544). The expression levels of miR‐200a and let‐7c in carcinoma samples were not significantly related to TNM staging.

3.4. KRAS gene mutation status

Genes controlled by miR‐200a include KRAS, which is frequently mutated and dysregulated in colon cancer. KRAS codons 12, 13 and 61 were assessed in a subset of these samples. There were 0%, 14% and 38% positive results for non‐malignant, adenoma and carcinoma samples respectively (Table 3). A relationship between lower incidence of KRAS mutations in adenomas where miR‐200a and let‐7c expression is altered is consistent with induction of miR‐200a and let‐7c expression as earlier events, with miRNA expression changes preceding KRAS mutations in this patient group. Even though this interpretation is supportive of an early event hypothesis, any conclusions have to take into account the levels of variability for miRNA‐200a expression in these groups, partly, owing to the source of the nucleic acid used in the study. Under these circumstances, complete demonstration of subtle changes will require increased sample numbers or yet to be developed assays and isolation procedures.

TABLE 3.

KRAS mutation analysis (codons 12 and 13)

Tissue	Mutated/Total	% Positive results
Normal	0/12	0%
Adenoma	2/14	14%
Carcinoma	5/13	38%

3.5. Effect of loss of miRNA expression

In order to observe the effect of dysregulation of expression of let‐7c and miR‐200a on proliferation, division and global methylation in colon cancer cells, miR‐200a and let‐7c gene expression was mechanistically decreased in immortalized colon cancer cells, strain SW48 (CCL‐231). Separate monolayers were transfected on 12‐well plates with anti‐miR‐200a siRNA, anti‐let‐7c siRNA or scrambled siRNA, and a control with no transfection. The expression levels of the miRNA genes were measured at 24 h, 48 h and 72 h from the time of transfection by siRNA. Knockdown of the genes was successfully achieved as confirmed by the apparent decrease in the miRNA expression of the siRNA transfected genes by RT‐qPCR compared with controls. Neither gene, however, was completely silenced by this method with the expression of miR‐200a decreased by 83% and that of let‐7c was decreased by 37% compared with scrambled control (Table 4).

TABLE 4.

Relative microRNA expression in cell line 24 h after transfection with siRNA

siRNA	miR‐200a/β‐2M%	let‐7c/β‐2M%
None	2260	2394
Scrambled	2537	3017
Complementary	423	1900

Cells were monitored up to 72 h after transfection. No discernible effect on malignant cell growth was observed microscopically at 24, 48 and 72 h. As confirmed by levels of β2‐microglobulin and β‐actin (constitutively expressed genes), the cancer cells continued to grow across the 72‐h incubation period—despite the knockdown of the target genes. We also measured the expression of a central tumor suppressor gene, cyclin‐dependent kinase inhibitor 2A (CDKN2A). This gene encodes the p16 protein, which is a cell cycle regulator that is frequently lost in human cancer. We found sixfold decreased expression of this gene from 24 to 72 h in the untreated and control cancer cells, while loss of let‐7c resulted in only threefold loss of CDKN2A expression there was no decrease in expression upon loss of miR‐200a. These results suggest that continued expression of at least 17% (for miR‐200a) may be sufficient. Alternatively, redundant factors may maintain cell growth in the absence of either of these micro RNAs.

4. DISCUSSION

This study utilized matched normal, adenomatous and malignant colon cancer tissues from individual patients to investigate changes in DNA methylation and miRNA expression in early stages of colon cancer progression. Global methylation, LINE‐1 hypomethylation, miR‐200a expression and let‐7c expression were measured in these matched samples.

Altered methylation patterns have been observed in colon cancer as well as in histologically normal tissue within the same colon. ²⁵ Based on reports of changes in global DNA methylation in cancer cells, ²⁶ , ²⁷ levels of 5‐methyl cytosine were measured in the matched tissue DNA by ELISA. No significant differences in 5‐mC levels among matched normal, adenoma and carcinoma samples were observed. Global methylation status is a combination of intergenic and gene regions. These areas may have counter effects to one another, depending on the cell state. A more detailed analysis of methylation using advanced bioinformatic methods may uncover specific or regional differences with high resolution.

LINE‐1 hypomethylation is found in carcinoma cells compared with normal cells. We and others have previously observed that adenomas removed with synchronous carcinomas have significantly lower LINE‐1 methylation than adenomas in the absence of synchronous carcinomas. ¹² , ²⁶ , ²⁷ Investigation of the methylation status of LINE‐1 in the three sample types studied here (normal, adenoma and carcinoma), when grouped as independent samples, revealed decreased methylation of these sequences in adenomas compared to normal tissue and further decrease in methylation in carcinoma tissue compared to adenomas. Hence, although the global DNA methylation results were not conclusive, LINE‐1 hypomethylation occurred in adenomas supporting intergenic hypomethylation as an early event. With increased miRNA expression in adenomas discussed below, miRNA dysregulation may be concurrent with intergenetic methylation changes. ²⁶ , ²⁷

Differential expression of specific miRNAs between non‐malignant (normal) and malignant colon tissue tissues has previously been reported. ²⁸ Here, we extended those studies to the expression of two specific miRNAs, miR‐200a and let‐7c, using nucleic acid extracted from matched normal, adenomas and carcinoma tissue. The current approach of using paired analysis including adenomatous adenomas also differs from that of other studies that utilized non‐paired normal and adenomatous adenomas. ²⁹ , ³⁰ , ³¹ , ³² , ³³

The results showed significantly higher miR‐200a expression levels in adenomas than in normal tissues. The miRNA, let‐7c, showed a similar pattern to that of miRNA‐200a expression. We found significantly higher expression of let‐7c in adenomas and carcinomas than in normal tissues. Conversely, there was no significant difference in neither miR‐200a nor let‐7c expression between adenoma and colon cancer. This supports the concept of these cancer‐related aberrations occurring in the precancerous tissue and persisting in the malignant cells.

The expression levels of miR‐200a and let‐7c in carcinoma samples were not significantly related to TNM staging. This does not rule out that other miRNA or epigenetic markers may be associated with progression and metastasis. For example, elevated expression of microRNAs, miR‐99a, miR‐100 and miR‐125b has been observed in mesenchymal‐like colon cancer cells associated with metastasis and poor survival. ³⁴

Higher expression of let‐7c in adenoma than in normal tissue was unexpected because let‐7c is described in previous studies as a tumor suppressor, albeit in tissues other than colon. ³⁵ Dysregulation of let‐7c, therefore, would be expected to result in loss of expression as tumorigenesis proceeds. Thelet‐7 family has many members, let‐7c among them, that could have distinct biological characteristics. The mechanistic role of let‐7c, therefore, as an actual oncogene is not defined and deserves further study.

The results reported here suggest that changes in miR‐200a and let‐7c occur in adenoma development rather than in the invasive cancer stage. Epigenetic events might precede other factors, including accumulation of genetic variants such as KRAS mutations. ⁵ KRAS mutations frequently occur after mutations activating the Wnt/β‐catenin signalling pathway in conventional colon cancer progression. BRAF mutations precede this activation in serrated progression. ³⁶ In the current study, KRAS mutations were present in 38% of the cancer tissues compared to 14% of the adenomas and absent in normal tissue. Since the expression of the miRNA in the genes investigated is significantly increased in adenomas compared with normal tissues, it might be inferred that miRNA expression changes preceded KRAS mutations in this patient group. Increased expression of both miR‐200a and let‐7c was observed in adenomas (and carcinomas) with KRAS mutations over those without, but the comparisons did not reach significance by independent t‐test analysis. The two adenomas showing KRAS mutations may have been further along in the transformation process. Unlike carcinomas, adenomas are not subject to TNM staging, so this would be difficult to confirm pathologically.

To further study the effect of these miRNA genes on cancer cell growth, knockdown of these two genes was performed in cultured colon cancer cells using the siRNA transfection protocol. The impact of the knockdown was not appreciably visible in the cells. There was no difference in the expression of two housekeeping genes, beta‐2‐microglobulin and beta‐actin. A loss of CDKN2A expression with loss of let‐7c would be consistent with the potential activity of let‐7c as a tumor suppressor; however, the complexity of miRNA function as well as the inability to completely suppress expression preclude strong conclusions from these observations.

Cell growth is likely regulated by a combination of factors including multiple microRNAs. A single miRNA can target several genes and a single gene can be regulated by different miRNAs, so changes in gene expression are complex following under‐ or over‐expression of any given miRNA. Another complication is that siRNA‐based knockdown of the genes does not achieve a complete silencing of the genes, as opposed to gene knockout protocols. Guided enzymatic (CRISPR) or other gene deletion studies may further define the immediate effects of complete loss of these miRNA genes, or further reveal redundant functions provided by other genes.

Epigenetic events are reversible and that could give rise to novel treatment strategies for colon cancer patients. A suggested area of colon cancer therapy is the application of anti‐miRs that competitively inhibit microRNAs. ³⁷ There are ongoing trials investigating miRNAs as a therapy for certain cancers. ³⁸ Further confirmation of epigenetic factors would require prospective studies on fresh or frozen tissues, as it is impossible to completely avoid the degradation of nucleic acid obtained from archival sources. Survival of miRNA in blood and plasma raises the opportunity to include key miRNA in liquid biopsies. ³⁹ Early overexpression is a valuable characteristic for such screening molecules.

CONFLICT OF INTEREST

There are no conflicts of interest.

Gebrekiristos M, Melson J, Jiang A, Buckingham L. DNA methylation and miRNA expression in colon adenomas compared with matched normal colon mucosa and carcinomas. Int J Exp Path. 2022;103:74–82. doi: 10.1111/iep.12432