Investigating miRNA-Driven DNA Methylation: Statistical Evidence of Gene-Specific Modulation

Investigating miRNA-Driven DNA Methylation: Statistical Evidence of Gene-Specific Modulation | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Investigating miRNA-Driven DNA Methylation: Statistical Evidence of Gene-Specific Modulation Seyeon Jeon, Ha Ra Jun, Ji-Young Lee, Chang Ohk Sung, SUNG-MIN CHUN This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5391278/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract DNA methylation plays a crucial role in the epigenetic regulation of gene expression and is closely associated with the development of cancer. Abnormal regulation of gene expression due to changes in DNA methylation patterns of specific genes has been frequently observed in tumor cells. However, the specific mechanisms underlying the induction of DNA methylation in certain genes have not been fully elucidated. This study aimed to investigate the potential of microRNAs (miRNAs) as statistically significant regulators guiding promoter methylation of specific genes. MiRNAs are known to specifically recognize DNA sequences and exhibit various degrees of complementarity. We performed Spearman's rank correlation between the expression levels of 734 microRNAs and the CpG island methylation levels of 20,587 genes, collected from 813 cell lines in the Cancer Cell Line Encyclopedia (CCLE) database. Subsequently, we validated the dependent relationship between the selected target microRNAs and gene clusters using linear regression analysis. We identified 25 target genes in which promoter methylation was induced by the expression of four target miRNAs (hsa-miR-200a, hsa-miR-200b, hsa-miR-200c, and hsa-miR-141) with statistically significant values. The correlations of the target pairs between methylation level of target genes and matched miRNAs were most pronounced in colorectal, gastric, lung, and ovarian cancers. Cancer-related genes, including ST14, OVOL1 , and EPCAM , were identified as the target genes, confirming the possibility that promoter methylation of these genes is regulated by miRNA. Using bioinformatics-based screening analysis, we discovered target pairs that exhibited statistically significant changes in promoter methylation patterns due to specific miRNA expression. Furthermore, we confirmed the potential for miRNAs to regulate the expression of cancer-related genes through miRNA-induced promoter methylation. This expands our understanding of the mechanism underlying tumor development through methylation and provides a new perspective on the utilization of microRNAs in the field of cancer treatment. DNA Methylation MicroRNA (miRNA) Epigenetic regulation CpG island Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction DNA methylation, orchestrated by the DNA methyltransferase (DNMT), is a pivotal epigenetic mechanism that finely regulates gene expression. It predominantly occurs at CpG sites within genomic regions known as CpG islands, which are densely packed with CpG dinucleotides and commonly located at the 5' end of gene regulatory regions. Preserving appropriate DNA methylation is crucial since CpG island hypermethylation can effectively silence genes, while hypomethylation may lead to genomic instability and result in aberrant gene expression. Consequently, DNA methylation exhibits tissue-specific patterns that are carefullu maintained and transmitted to the next generation of cells. 1) Aberrant DNA methylation patterns have been reported to be closely associated with various diseases, particularly cancer. Tumor cells often exhibit a 5–10% reduction in overall 5-methylcytosine levels compared to normal cells, while localized hypermethylation in many CpG islands. 2–3) For instance, promoter hypermethylation of RB (retinoblastoma), a gene that regulates the cell cycle, occurs in about 10% of sporadic unilateral retinoblastomas, leading to reduced expression of the RB protein. 4–5) Similarly, in sporadic colon tumors, promoter methylation of MLH1 , a DNA mismatch repair-related gene, leads to reduced MLH1 expression, and treatment with the demethylating agent 5-aza-2'-deoxycytidine was found to decrease MLH1 suppression. 6–7) Thus, hypermethylation of tumor suppressor genes in cancer cells can lead to their reduced expression, promoting cancer. Despite numerous reports of abnormal gene expression regulation through DNA methylation of specific genes in cancer, 8) a definitive explanation for the mechanism behind gene-specific DNA methylation remains elusive. DNA methylation is highly critical for gene expression regulation; hence, precise control of methylation patterns in numerous genes is essential. To achieve gene-specific DNA methylation regulation, an accurate regulatory mechanism must exist that can precisely target DNMT activity for each gene. Such a regulator should be capable of recognizing the nucleotide sequence of the target gene and have the versatility to match various genes. Additionally, it should be a substance that has been reported to show differential expression patterns in tumors. From this perspective, we have considered the possibility of microRNAs, which have already been reported in the thousands and are known to bind to the 3' untranslated regions of multiple target mRNAs, thereby inhibiting protein translation of the target genes 9–10) , as potential guider for DNMT regulation. MicroRNAs are small non-coding RNAs composed of 21–25 nucleotides that function primarily as post-transcriptional regulators in eukaryotic gene expression. 11–13) Unlike mRNA, which acts as an intermediary in protein expression, the function of microRNAs is based on their ability to recognize and bind to their own nucleotide sequence. In addition to its post-transcriptional regulation function, small RNAs have been shown to mediate DNA methylation in plants through a phenomenon known as the RNA-directed DNA methylation (RdDM) pathway. 14) The RdDM pathway has been observed in some yeasts 15) , and several studies reporting RdDM-like mechanisms have been reported in human cells. 16–18) Additionally, processes similar to RNA-directed DNA methylation (RdDM) have been reported in human cells, but they mostly involve self-feedback mechanisms, such as methylation of siRNA-derived repetitive sequences or de novo methylation of CpG islands by promoter antisense-derived small RNAs. 19) However, the molecular mechanism by which endogenous microRNAs in human tumor cells recognize specific target genes outside of microRNA-derived DNA sequences and induce methylation has not been elucidated. In this study, we aimed to explore the relationship between the expression levels of 734 miRNAs and the methylation status of 20,587 genes across 813 cell lines obtained from public databases through rigorous statistical analysis. By doing so, we seek to ascertain the potential of specific miRNAs in selectively regulating the methylation of CpG islands in target genes. Materials and methods Dataset This study was conducted based on the data collected from the Cancer Cell Lines Encyclopedia (CCLE) database ( https://sites.broadinstitute.org/ccle ), which provides large-scale genomic data from cancer cell lines. We obtained expression data for 734 microRNAs and methylation data for 54,531 promoter CpG clusters (corresponding to 20,587 genes). To perform linear regression analysis on the target genes selected through Spearman's rank correlation analysis of microRNA expression and gene cluster methylation, mRNA expression data were also obtained from CCLE. The analysis was conducted on a common set of 813 cell lines that were identified in all three datasets. Statistical analysis Spearman's rank correlation analysis between the miRNA expression and specific gene cluster methylation levels of each gene was performed for all cases, obtaining correlation coefficients, p -values, and the false discovery rate (FDR) for each pair. A correlation with an FDR value below 0.05 was considered significant. For the regression analysis of the final selected target pairs, we sequentially performed simple linear regression and multiple linear regression for the continuous dependent variable. One way analysis of variance (ANOVA) was conducted to compare differences among continuous variables. Post hoc tests were performed to verify differences between all groups. A p -value less than 0.05 was considered statistically significant in all statistical analyses. All statistical analyses and visualizations were conducted using software version 4.0.3. Filtering of promoter CpG island clusters We performed a selection process to identify gene clusters with potential promoter CpG islands, excluding non-coding RNA (RefSeq id: NM) gene clusters. Our selection criteria were based on a self-generated standard, taking into account the form and characteristics of the methylation data. Specifically, we defined CpG island clusters as those with a minimum of 20 CpG sites within a cluster and a ratio of CpG sites within the entire cluster of 10% or more. Ultimately, we identified 5,070 gene promoter CpG island clusters through this process. Cancer type grouping A total of 26 cancer types were identified in a set of 813 cell lines that shared common microRNA expression data sets, gene methylation data sets, and mRNA expression data sets. We performed clustering of the entire microRNA expression values using the NMF consensus tool in Gene Pattern ( https://www.genepattern.org/ ). This analysis was conducted under conditions ranging from k = 1 to k = 12. The microRNA expression patterns of the cell lines were clustered based on their similarity, and the k-value was set at the point where the cophenetic coefficient showed a sharp drop and clear clustering. we employed classification into a single group for cancer types only when they demonstrated similar microRNA expression patterns in two or more instances. Gene ontology analysis To investigate the interactions and biological characteristics of the selected genes, gene ontology analysis was performed using the Enrichr ( https://maayanlab.cloud/Enrichr/ ) database. Adjusted p -values of 0.05 or higher were considered statistically significant in the Biological Process, Molecular Function, and Cellular Component categories for the 25 selected genes. Sequence homology To verify the sequence homology between the selected microRNAs and gene clusters, we utilized BLASTn (version: 2.2.31+) from the Basic Local Alignment Search Tool. Since microRNAs are relatively short, approximately 22 bp, and we specifically focused on assessing homology within the promoter CpG island of specific genes, we employed the blastn-short program, which is suitable for identifying sequence homology for sequences with fewer than 30 bp. The nucleotide sequence of the target gene cluster was defined as the region between the first CpG site and the last CpG site. To convert these sequences into a database, the Makblastdb package was employed. Subsequently, the homology between the database and query sequence was checked by setting the microRNA as the query sequence. In the known mechanism of microRNA targeting in the mRNA 3' UTR, the seed sequence of 2–8 bp plays a crucial role. However, our study identified a different mechanism for the targeting of microRNAs. Therefore, we considered the entire mature microRNA sequence rather than limiting it to the seed sequence. We set the word size option, which represents the minimum matching length, to 7 bp. Results Identification of miRNA-Gene Promoter CpG Methylation Pairs with Confirmed Correlations To identify target pairs exhibiting a strong correlation between miRNA expression and gene promoter region methylation, and to investigate whether these target pairs have interdependent relationships that also influence mRNA expression, Spearman's rank correlation analysis followed by sequential regression analyses were performed (Figure 1) . Based on Spearman's rank correlation analysis between microRNA expression and gene promoter region methylation, 316,671 microRNA-gene cluster pairs exhibited significant correlations (FDR < 0.05). Among them, 50,813 pairs were predicted to correspond to gene clusters within a promoter CpG island (CGI). There were 929 pairs with a positive correlation (r ≥ +0.3, Figure 2a ) and 777 pairs with a negative correlation (r ≤ -0.4, Figure 2b ). Next, we implemented an alternative correlation coefficient threshold to address the imbalance between positive and negative correlation pairs in the overall correlation analysis results. This adjustment was necessary due to a considerable surplus of negative correlation pairs compared to positive ones. Following this, we carried out an additional selection process to identify target pairs of greater significance based on the correlation analysis outcomes. Based on the microRNA expression and correlation FDR values (log2 transformed), the positive and negative correlation pairs were each divided into four intervals. From these intervals, pairs with relatively high mean microRNA expression and low FDR values across all cell lines were selected ( Figure 2a and Figure 2b ). Consequently, 53 positive correlated pairs ( Figure 2c ) and 71 negative correlated pairs ( Figure 2d ) were identified within the selected intervals [positive: log2(miRNA expression mean) > 9.18, log2(FDR) 8.48, log2(FDR) < -211.82]. For future experimental validation, we then applied a criterion of relatively higher microRNA expression. We found that 12 gene clusters were positively regulated by these microRNAs based on their expression, while 52 gene clusters were negatively regulated. The final target microRNAs and gene clusters are presented in Table 1 . Sequential regression analyses of microRNA expression and promoter methylation effects on mRNA expression. To investigate whether the target pairs exhibiting strong correlations in microRNA expression and promoter methylation have a dependent relationship and also influence mRNA expression, we performed sequential three regression analyses. First, a simple linear regression (first SLR) analysis was performed on the gene promoter methylation influenced by microRNA expression. The results showed significant positive correlations for all 12 target pairs with β > 0 and p -value < 0.05. Bonferroni correction confirmed statistically significant results for all 12 positive target pairs ( Figure 3a ). Additionally, all 52 pairs with negative correlations showed statistically significant results with an β < 0 and p -value < 0.05. However, two pairs, has-miR-200b & DSP_ 2 and has-miR-200b & FAM83H_3 , did not show significant results after Bonferroni correction. Excluding these two pairs, significant results were observed for 50 pairs ( Figure 3b ). Subsequently, to validate the well-known theory of DNA promoter methylation-mediated mRNA expression regulation in our dataset, we performed a second simple linear regression analysis (second SLR) on the gene promoter methylation and the corresponding mRNA expression. We separated the gene clusters that showed positive and negative correlations with the target microRNAs into positive and negative groups, respectively, and examined mRNA expression for each microRNA ( Figure 3c and Figure 3d ). We found a significant negative correlation between promoter methylation and mRNA expression in both groups. In the SLR2 results for each target pair, excluding hsa-miR-141 & TJP2_3 and hsa-miR-200c & TJP2_3 from among the positive pairs, 10 pairs showed significant results with β < 0 and p -value < 0.05, while all negative pairs showed significant results with β < 0 and p -value < 0.05 ( Supplementary Table 1 ). Finally, we conducted a multiple linear regression (MLR) analysis, considering two independent variables together, to examine whether microRNA-mediated methylation regulation affects mRNA expression. Among the 12 target pairs with positive correlations, hsa-miR-141 & TJP2_3 and hsa-miR-200c & TJP2_3 did not show β < 0 for both microRNA-mRNA expression and methylation-mRNA expression, contrary to the hypothesis we aimed to test. Excluding these two pairs, β < 0 was observed in both microRNA-mRNA expression and methylation-mRNA expression for 10 pairs, and 8 pairs were significant with p-value 0 for microRNA-mRNA expression and β 0.05, were statistically significant ( Table 2 ). Comparing the significance of target pairs by cancer group We performed NMF clustering on the expression of all microRNAs (n = 734) in the dataset to explore the significance of target pairs across cancer types. A rebound was observed in the cophenetic coefficient at K = 5 and K = 6, but we proceeded with grouping using K = 5, which exhibited the most distinct clustering ( Figure 4a ). Among the 26 types of cancer, only those with similar clusters containing two or more types of cancer were grouped together. Consequently, the cancer types were categorized into seven groups. We also included fibroblasts, which are expected to exhibit patterns similar to normal cells, resulting in a total of eight groups ( Figure 4b ). Among the eight groups, groups 3, 6, 7, and 8 were observed to have low expression of the target microRNAs relative to the other cancer-type groups ( p < 2.2e-16 by one-way ANOVA, Figure 4c ). Using the same method, we performed simple linear regression analysis (SLR) of the target microRNA expression and the gene cluster methylation in each cancer group. Group 1 (colon cancer, gastric cancer, lung cancer, and ovarian cancer), which included the largest number of cancer types, showed SLR results that were the most similar to those of the overall cancer types, with 7 positive pairs and 45 negative pairs (81.2% of the target pairs, Figure 4d ). On the other hand, the SLR results of group 2 (breast cancer, endometrial/uterine cancer, and pancreatic cancer) and group 3 (brain cancer, liver cancer, and skin cancer), which included only three types of cancer each and with a similar number of cell lines, 103 and 109, respectively, exhibited markedly distinct trends ( Figure 4d ). The positive target pairs showed the same result as total cancer for two pairs (2/12) for both group 2 and group 3, while the negative target pairs showed the same result for 35 pairs (35/52) for group 2 but only one pair (1/52) for group 3. Hsa-miR-141 & ATP8B2_1 and hsa-miR-141 & ATP8B2_5 were the most frequent positive target pairs, exhibiting the same results as total cancer in three groups (i.e., group 1, group 3, and group 5). In contrast, hsa-miR-200b & ESRP2_2 was the most frequent negative target pair, demonstrating the same results as total cancer in all four groups (i.e., group 1, group 2, group 3, and group 5). Discussion In this study, we investigated the possibility that microRNA regulates gene promoter methylation using data collected from open-source databases. We examined the correlation between 734 microRNAs and various gene promoter CpG methylation using Spearman rank correlation analysis, identifying significant target microRNA and gene cluster pairs. In addition, we validated the dependency between the identified pairs through sequential linear regression analysis. Figure 1 depicts the workflow for the DNA methylation, microRNA and mRNA correlation, and the regression analysis. Through this process, we confirmed that the selected targets were indeed likely to be regulated by microRNA-mediated methylation of specific genes (Fig. 3 and Table 2 ). For the second simple linear regression (SLR2) analysis on gene promoter methylation and mRNA expression, we found that, regardless of the directionality and correlation between the microRNA and gene clusters, all target pairs showed a significant negative correlation between gene promoter CpG methylation and mRNA expression (β < 0, p < 0.05), as reported in previous studies. This was further confirmed by our data (Fig. 3 c, b and Supplementary Table 1 ). Multiple linear regression considering both microRNA expression and promoter methylation on mRNA expression showed that, except for some pairs, gene promoter methylation by microRNA could also affect mRNA expression (Table 2 ). However, mRNA expression is a complex mechanism that can be regulated by various factors, so promoter methylation alone is not sufficient to fully explain the expression patterns. The correlation explanatory power between the selected microRNA and gene cluster pairs, derived from the analysis encompassing all types of cancer, was found to be highest in cancer group 1, which includes colon cancer, gastric cancer, lung cancer, and ovarian cancer). Group 2 (breast cancer, endometrial/uterine cancer, and pancreatic cancer) and group 3 (brain cancer, liver cancer, and skin cancer), which had similar cell numbers, showed substantial differences in the simple linear regression results of the target microRNA expression and the gene cluster methylation. It is likely that this was due to differences in the expression of target microRNAs across cancer types, particularly between Group 2 and Group 3 (Fig. 4 c). Furthermore, we observed a lower proportion of statistically significant target pairs in the SLR results for Groups 3, 4, 6, 7, and 8, which had overall lower levels of target microRNA expression (Fig. 4 c and Fig. 4 d). To assess the enrichment of 25 target genes, gene ontology analysis was conducted and identified significant terms with a p -value < 0.05 in the Biological Process (BP), Molecular Function (MF), and Cellular Component (CC) categories (Fig. 5 ). DSP, CDH3, MARVELD2 , and TJP2 were found to be significantly associated with cell-cell junction organization (GO:0045216) with the lowest p -value (7.03e-04) in BP. In MF, the keratin filament binding (GO:1990254) of FAM83H and VIM showed the most significant association with a p -value of 2.24e-05. In CC, intermediate filament (GO:0005882) was significantly associated with FAM83H, DSP, VIM , and PPL , with the lowest p -value (4.21e-07). Notably, the genes MARVELD2, EPCAM , and TJP2 showed functional roles in bicellular tight junctions (GO:0005923), tight junctions (GO:0070160), and the apical junction complex (GO:0043296) in the CC. Promoter methylation of tumor suppressor genes is an important mechanism in tumorigenesis. 20) ST14 , a tumor suppressor gene, was identified as a target gene in this study and was found to be negatively regulated by promoter methylation mediated by hsa-miR-200b, hsa-miR-200c, and hsa-miR-141. According to Kim et al., the expression of Suppressor of tumorigenicity 14 ( ST14 ) or Serine protease 14 ( Prss14 ) genes are associated with a poor prognosis and increased invasiveness and metastatic potential in breast cancer patients. 21) These facts suggest that microRNA-mediated regulation of ST14 gene methylation at the 5' end, and the subsequent changes in mRNA expression, may function as a regulatory factor in tumor invasiveness and metastasis. Another target gene, OVOL1 , is a transcription factor that induces the reversal process of EMT (M-EMT, mesenchymal to epithelial transition) in human cancer cells, thereby inhibiting tumor invasion and metastasis, and affecting the growth, apoptosis, and mobility of cancer cells. 22) The target gene EPCAM (Epithelial cell adhesion molecule), frequently reported in the Gene Ontology analysis, is an important marker gene for cancer diagnosis and treatment, as it is overexpressed by various types of cancer cells, including colon and breast cancer. 23,24) The possibility that the promoter methylation of EPCAM can be regulated by microRNA suggests a new perspective on cancer diagnosis and treatment. Some previous studies have used bioinformatics to investigate the correlation between miRNA and CpG methylation. 25,26) Such studies have focused on the interaction between DNA methylation and miRNA expression in the context of the phenotype of breast cancer 25) or the mechanisms of racial heterogeneity in hepatocellular carcinoma. 26) They mainly discuss the regulation of microRNA expression by methylation or the inhibition of DNMT enzymes by microRNA. However, these studies have not specifically focused on gene promoter methylation mediated by microRNA. Therefore, our study can be considered one of the first to screen the targets of microRNA-mediated promoter methylation in 26 diverse cancer types. However, the results of the current analysis have limitations in proving the hypothesis that microRNA guides promoter methylation of specific genes. To more clearly confirm this hypothesis, direct evidence of an interaction between microRNA and DNMT is needed. We plan to perform additional experiments, such as microRNA knockdown using siRNA, microRNA transfection, and chromatin immunoprecipitation sequencing (ChIP-seq). Before conducting experimental validation, we randomly verified the sequence homology between the target gene clusters and the microRNAs using BLAST. The results are shown in Supplementary Table 2 and Supplementary Table 3 , and we were able to confirm sequence homology in some target pairs. However, due to the short query sequence length of approximately 22 base pairs and the limited matching of microRNA and gene cluster sequences with significant correlations, there was a notable disparity in the e-value and bit-score when compared to the general BLAST (the detailed methods can be found in the "sequence homology" section of the Materials and Methods ) Nevertheless, through screening using bioinformatics, we efficiently explored a large dataset to identify target pairs with strong correlations, and we validated the expression of mRNA for the target pairs using an mRNA database. This study is a foundational step in identifying the targets of microRNA-mediated promoter methylation phenomena. These results need further experimental validation. Our findings expand our understanding of the mechanisms underlying tumor formation through methylation and provide a new perspective on the use of microRNA in cancer treatment. Declarations Ethics approval and consent to participate Not applicable. This study utilizes open data from the CCLE, and therefore, no separate ethics approval and consent to participate is required. Patient consent for publication Not applicable. This study utilizes open data from the CCLE, and therefore, no patient consent for publication is required. Competing interests The authors declare no competing interests. Funding This study was supported by the grants (2019IP0836-1, 2023IP0085-2) from the Asan Institute for Life Sciences, Asan Medical Center, Seoul, Korea. Author Contribution S.J. performed most of the data analysis with assistance from S-M.C., C.O.S., J-Y.L. and H.R.J.; S.J. wrote the manuscript; H.R.J. validate the analyzed result; S-M.C. designed the study and edited the manuscript. Acknowledgments Not applicable Data Availability All of the datasets used in this study were downloaded from the Cancer Cell Lines Encyclopedia (CCLE) database: https://sites.broadinstitute.org/ccle. References Rakyan, Vardhman K., et al. DNA methylation profiling of the human major histocompatibility complex: a pilot study for the human epigenome project. PLoS biology. 2(12), e405 (2004). Feinberg, A. P., Gehrke, C. W., Kuo, K. C., & Ehrlich, M. Reduced genomic 5-methylcytosine content in human colonic neoplasia. Cancer research. 48(5), 1159–1161 (1988). Ehrlich, M. DNA hypomethylation in cancer cells. Epigenomics, 1 (2), 239–259 (2009). Greger V, Debus N, Lohmann D, Hopping W, Passarge E, Horsthemke B, Debus N, Lohmann D, Hopping W, Passarge E, Horsthemke B: Frequency and parental origin of hypermethylated RB1 alleles in retinoblastoma. Hum Genet. 94(5), 491–496 (1994). Ohtani-Fujita, N., et al. "CpG methylation inactivates the promoter activity of the human retinoblastoma tumor-suppressor gene." Oncogene. 8(4), 1036–1067 (1993). Kane, Michael F., et al. "Methylation of the hMLH1 promoter correlates with lack of expression of hMLH1 in sporadic colon tumors and mismatch repair-defective human tumor cell lines." Cancer research. 57(5), 808–811 (1997). Herman, James G., et al. Incidence and functional consequences of hMLH1 promoter hypermethylation in colorectal carcinoma. Proceedings of the National Academy of Sciences. 95(12), 6870–6875 (1998). Schinke, Carolina, et al. Aberrant DNA methylation in malignant melanoma. Melanoma research. 20(4), 253–265 (2010). Bousquet, M., Nguyen, D., Chen, C., Shields, L., & Lodish, H. F. MicroRNA-125b transforms myeloid cell lines by repressing multiple mRNA. Haematologica. 97(11), 1713 (2012). Ritchie, William, John EJ Rasko, and Stéphane Flamant. MicroRNA target prediction and validation. MicroRNA Cancer Regulation: Advanced Concepts, Bioinformatics and Systems Biology Tools. 39–53 (2013). Khvorova, A., Reynolds, A., & Jayasena, S. D. Functional siRNAs and miRNAs exhibit strand bias. Cell. 115(2), 209–216 (2003). Chendrimada, Thimmaiah P., et al. TRBP recruits the Dicer complex to Ago2 for microRNA processing and gene silencing. Nature. 436(7051), 740–744 (2005). Ameres, Stefan L., and Phillip D. Zamore. Diversifying microRNA sequence and function. Nature reviews Molecular cell biology. 14(8), 475–488 (2013). Wassenegger, Michael. RNA-directed DNA methylation. Plant molecular biology. 43, 203–220 (2000). Verdel, André, et al. RNAi-mediated targeting of heterochromatin by the RITS complex. Science. 303(5658), 672–676 (2004). Castanotto, Daniela, et al. Short hairpin RNA-directed cytosine (CpG) methylation of the RASSF1A gene promoter in HeLa cells. Molecular Therapy. 12(1). 179–183 (2005). Kawasaki, H., & Taira, K. Induction of DNA methylation and gene silencing by short interfering RNAs in human cells. Nature. 431(7005), 211–217 (2004). Morris, K. V., Chan, S. W. L., Jacobsen, S. E., & Looney, D. J. Small interfering RNA-induced transcriptional gene silencing in human cells. Science. 305(5688), 1289–1292 (2004). Tufarelli, Cristina, et al. Transcription of antisense RNA leading to gene silencing and methylation as a novel cause of human genetic disease. Nature genetics. 34(2), 157–165 (2003). Liyanage, Chamikara, et al. Promoter hypermethylation of tumor-suppressor genes p16INK4a, RASSF1A, TIMP3, and PCQAP/MED15 in salivary DNA as a quadruple biomarker panel for early detection of oral and oropharyngeal cancers. Biomolecules 9(4),148 (2019). Kim, S., Yang, J. W., Kim, C., & Kim, M. G. Impact of suppression of tumorigenicity 14 (ST14)/serine protease 14 (Prss14) expression analysis on the prognosis and management of estrogen receptor negative breast cancer. Oncotarget. 7(23), 34643 (2016). Roca, Hernan, et al. Transcription factors OVOL1 and OVOL2 induce the mesenchymal to epithelial transition in human cancer. PloS one. 8(10). e76773 (2013). Cimino, Ashley, et al. Epithelial cell adhesion molecule (EpCAM) is overexpressed in breast cancer metastases. Breast cancer research and treatment. 123, 701–708 (2010). Spizzo, Gilbert, et al. EpCAM expression in primary tumour tissues and metastases: an immunohistochemical analysis. Journal of clinical pathology. 64(5), 415–420 (2011). Aure, Miriam Ragle, et al. Crosstalk between microRNA expression and DNA methylation drives the hormone-dependent phenotype of breast cancer. Genome medicine. 13(1), 72 (2021). Varghese, Rency S., et al. Integrative analysis of DNA methylation and microRNA expression reveals mechanisms of racial heterogeneity in hepatocellular carcinoma. Frontiers in genetics. 12, 708326 (2021). Tables Table 1 and 2 are available in the Supplementary Files section. Additional Declarations No competing interests reported. Supplementary Files Table12.docx SupplementaryTables.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5391278","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":378366681,"identity":"1fb737ec-1799-4190-a769-a29ec68e2103","order_by":0,"name":"Seyeon Jeon","email":"","orcid":"","institution":"Department of Medical Science, Asan Medical Institute of Convergence Science and Technology, University of Ulsan College of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Seyeon","middleName":"","lastName":"Jeon","suffix":""},{"id":378366682,"identity":"14204529-0334-4c36-9967-98927cebb6c0","order_by":1,"name":"Ha Ra Jun","email":"","orcid":"","institution":"Department of Medical Science, Asan Medical Institute of Convergence Science and Technology, University of Ulsan College of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Ha","middleName":"Ra","lastName":"Jun","suffix":""},{"id":378366685,"identity":"d8f476db-b2ab-42c0-85ce-9fc12f31a079","order_by":2,"name":"Ji-Young Lee","email":"","orcid":"","institution":"Center for Cancer Genome Discovery, Asan Institute for Life Science, Asan Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Ji-Young","middleName":"","lastName":"Lee","suffix":""},{"id":378366686,"identity":"e97c905c-e01f-4f38-abdd-236419dbc616","order_by":3,"name":"Chang Ohk Sung","email":"","orcid":"","institution":"Department of Pathology, Asan Medical Center, University of Ulsan College of Medicine","correspondingAuthor":false,"prefix":"","firstName":"Chang","middleName":"Ohk","lastName":"Sung","suffix":""},{"id":378366689,"identity":"0cf5d4a4-e4e7-46e0-bfed-65763c67f576","order_by":4,"name":"SUNG-MIN CHUN","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyElEQVRIiWNgGAWjYHACNiC2gfNkGHiI05IG5/EQq+UwCVrkpx1+9ph3x3l53fYDzB8+VBwGajn7AK8Wg9tp5sa8Z24bbjuTwCY54wxQC2+7AX4t0jls0rxttxm33WBgY+ZtA2rhZyPgsNlgLefsgVqYP/8lRgvDbbCWA4lALQzSjCAtvG34dQD9YiY5ty05eduZxDbJnjPpPGw8xwg5LPmZxNs2O9ttxw8f/vCjwlqOnycNvxYkwNgApgj5ZBSMglEwCkYBEQAAP709QxlGlOAAAAAASUVORK5CYII=","orcid":"","institution":"Department of Pathology, Asan Medical Center, University of Ulsan College of Medicine","correspondingAuthor":true,"prefix":"","firstName":"SUNG-MIN","middleName":"","lastName":"CHUN","suffix":""}],"badges":[],"createdAt":"2024-11-05 01:38:27","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5391278/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5391278/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":69270324,"identity":"f42d2812-4f50-420a-ae99-c07f3facd83c","added_by":"auto","created_at":"2024-11-18 15:12:34","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":124176,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eOverview.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAn overview of the target screening and validation process. CCLE: Cancer Cell Line Encyclopedia. Pair: Correlated pairs of microRNAs and gene clusters. Common mRNA: Only genes with mRNA expression data were selected from among genes in gene clusters predicted to be CpG islands by Spearman correlation analysis.\u003c/p\u003e","description":"","filename":"OnlineFigure1.png","url":"https://assets-eu.researchsquare.com/files/rs-5391278/v1/47e52473bb1081bc88af47e9.png"},{"id":69270329,"identity":"49a30bfe-9540-4408-bd5e-9d49b3f98674","added_by":"auto","created_at":"2024-11-18 15:12:34","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":128559,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSelection of target microRNA and gene cluster pairs. (a)\u003c/strong\u003e Scatter plot of positive correlation pairs with Spearman's correlation coefficient (r) +0.3 or higher. The X-axis represents log2(FDR) values, and the Y-axis represents the log2-transformed average expression of microRNAs across all cell lines. The red box in the upper left corner indicates a region with relatively high microRNA expression and low FDR values. The middle lines of the plot correspond to x = -76.42 and y = 9.18. \u003cstrong\u003e(b)\u003c/strong\u003e Scatter plot of negative correlation pairs with correlation coefficient (r) of -0.4 or less, where the X-axis represents log2(FDR) and the Y-axis represents the log2-transformed average microRNA expression across all cell lines. The red box in the upper-left corner represents the region where the microRNA expression is relatively high and the FDR value is low. The middle line of the plot is x = -211.82 and y = 8.48. \u003cstrong\u003e(c)\u003c/strong\u003e An enlarged view of the red box area in the positive correlation plot (a). [log2(microRNA expression mean) \u0026gt; 9.18 and log2(FDR) \u0026lt; -76.42] \u003cstrong\u003e(d)\u003c/strong\u003eshows an enlarged view of the red box area in the negative correlation plot(b). [log2(microRNA expression mean) \u0026gt; 8.48 and log2(FDR) \u0026lt; -211.82]. The color of the dots represents the targeted microRNA in each plot.\u003c/p\u003e","description":"","filename":"OnlineFigure2.png","url":"https://assets-eu.researchsquare.com/files/rs-5391278/v1/bede6bd2a327ce991cbce379.png"},{"id":69270326,"identity":"5bf86c90-3c76-484f-9970-b1a767dea6b5","added_by":"auto","created_at":"2024-11-18 15:12:34","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":198099,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSimple linear regression analysis of gene cluster methylation regulated by target microRNA expression and mRNA expression regulated by gene promoter methylation. \u003c/strong\u003eFor the Manhattan plots, the x-axis represents target gene positions across the entire genome by chromosome and the y-axis is the negative logarithm \u003cem\u003ep\u003c/em\u003e-value: -log10(\u003cem\u003ep\u003c/em\u003e) of each simple linear regression \u003cem\u003ep\u003c/em\u003e-values. \u003cstrong\u003e(a)\u003c/strong\u003e Positive correlation pairs, \u003cstrong\u003e(b)\u003c/strong\u003e Negative correlation pairs. The red dashed line indicates the \u003cem\u003ep\u003c/em\u003e-value of 0.05, while the blue dashed line represents the \u003cem\u003ep\u003c/em\u003e-value threshold of the Bonferroni correction. Dots in color indicate the target microRNA. \u003cstrong\u003e(c)\u003c/strong\u003e Gene group positively correlated with the target microRNA, \u003cstrong\u003e(d) \u003c/strong\u003eGene group negatively correlated with the target miRNA. \u003cstrong\u003e(c)\u003c/strong\u003eand \u003cstrong\u003e(d)\u003c/strong\u003e are the results of simple linear regression analysis for gene cluster methylation and mRNA expression, and \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.05 was considered statistically significant.\u003c/p\u003e","description":"","filename":"OnlineFigure3.png","url":"https://assets-eu.researchsquare.com/files/rs-5391278/v1/da49ee86d9c9f5b35ece380d.png"},{"id":69270325,"identity":"4af7257d-8aa4-4c84-ba51-23c72daa2336","added_by":"auto","created_at":"2024-11-18 15:12:34","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":328739,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCancer grouping and comparison of regression analysis results across cancer groups.\u003c/strong\u003e \u003cstrong\u003e(a)\u003c/strong\u003e NMF consensus clusters of total microRNA expression data. Maximum cophenetic coefficients for k = 2 to 12 clusters and the consensus matrices for k = 2 to 12 are shown. The right bottom plot shows a comparison of the cophenetic coefficients among k clusters. \u003cstrong\u003e(b)\u003c/strong\u003e Cancer groups. Group 1 includes colon, gastric, lung, and ovarian cancers, with cell line 283 accounting for 40.4% of the total cancer types (283/701). Group 2 includes breast, endometrial, and pancreatic cancers, with cell line 103 accounting for 14.7% of the total cancer types (103/701). Group 3 includes brain, liver, and skin cancers, with cell line 109 accounting for 15.5% of the total cancer types (109/701). Group 4 includes bile duct, sarcoma, and thyroid cancers, with cell line 25 accounting for 3.6% of the total cancer types (25/701). Group 5 includes bladder, esophageal, and head and neck cancers, with cell line 71 accounting for 10.1% of the total cancer types (71/701). Group 6 includes neuroblastoma and rhabdoid cancers, with cell line 18 accounting for 2.6% of the total cancer types (18/701). Group 7 includes lymphoma and myeloma, with cell line 72 accounting for 10.3% of the total cancer types (72/701). Group 8 includes fibroblasts, with cell line 20 accounting for 2.9% of the total cancer types (20/701). \u003cstrong\u003e(c)\u003c/strong\u003eTargeted microRNA expression across the cancer groups (\u003cem\u003ep\u003c/em\u003e \u0026lt; 2.2e-16 by one-way ANOVA). \u003cstrong\u003e(d)\u003c/strong\u003e Bar plots of the first SLR results across the cancer groups. First SLR refers to simple linear regression analysis of the association between the microRNA expression and the gene cluster methylation. The blue bars (\"matched\") in the bar plot of (d) indicate the proportion of statistically significant pairs (\u003cem\u003ep\u003c/em\u003e-value \u0026lt; 0.05) with a correlation that matches the study hypothesis (positive correlation: β \u0026gt; 0, negative correlation: β \u0026lt; 0). The yellow bars (\"not significant\") represent the proportion of pairs with a correlation that matches the study hypothesis but are not statistically significant (\u003cem\u003ep\u003c/em\u003e-value \u0026gt; 0.05). The gray bars (\"not matched\") indicate the proportion of pairs with a correlation that does not match the study hypothesis (positive correlation: β \u0026lt; 0, negative correlation: β \u0026gt; 0).\u003c/p\u003e","description":"","filename":"OnlineFigure4.png","url":"https://assets-eu.researchsquare.com/files/rs-5391278/v1/7d36caf5f2953efe118224d9.png"},{"id":69270328,"identity":"edbcd44b-799a-4744-a16c-7fd3fc47425e","added_by":"auto","created_at":"2024-11-18 15:12:34","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":250823,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eGene ontology analysis of the target genes. \u003c/strong\u003eA barplot representing the results of the Gene Ontology (GO) analysis. The dark blue bar represents Biological Process (BP), the pink bar represents Molecular Function (MF), and the green bar represents Cellular Component (CC). The y-axis of all bar plots represents the GO term, while the x-axis represents -log10 (adjusted \u003cem\u003ep\u003c/em\u003e-value). The results of the GO analysis were considered significant when the adjusted \u003cem\u003ep\u003c/em\u003e-value was less than 0.05.\u003c/p\u003e","description":"","filename":"OnlineFigure5.png","url":"https://assets-eu.researchsquare.com/files/rs-5391278/v1/94ef683476f666f991775580.png"},{"id":69996812,"identity":"77572c5c-cca5-4d50-9420-b9d3c903878a","added_by":"auto","created_at":"2024-11-27 10:32:06","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2086361,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5391278/v1/df918db0-a99b-4ee5-a903-3c30b0734ed5.pdf"},{"id":69270323,"identity":"3fe3c1cc-9673-4b12-b9d4-351615f37140","added_by":"auto","created_at":"2024-11-18 15:12:34","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":40925,"visible":true,"origin":"","legend":"","description":"","filename":"Table12.docx","url":"https://assets-eu.researchsquare.com/files/rs-5391278/v1/991f83857f21339286d9a33d.docx"},{"id":69270942,"identity":"fcaae493-36e9-40a3-9dd4-9a225dfec5f0","added_by":"auto","created_at":"2024-11-18 15:20:34","extension":"docx","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":43683,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTables.docx","url":"https://assets-eu.researchsquare.com/files/rs-5391278/v1/297b9d2b2584d63bbefbfc34.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Investigating miRNA-Driven DNA Methylation: Statistical Evidence of Gene-Specific Modulation","fulltext":[{"header":"Introduction","content":"\u003cp\u003eDNA methylation, orchestrated by the DNA methyltransferase (DNMT), is a pivotal epigenetic mechanism that finely regulates gene expression. It predominantly occurs at CpG sites within genomic regions known as CpG islands, which are densely packed with CpG dinucleotides and commonly located at the 5' end of gene regulatory regions. Preserving appropriate DNA methylation is crucial since CpG island hypermethylation can effectively silence genes, while hypomethylation may lead to genomic instability and result in aberrant gene expression. Consequently, DNA methylation exhibits tissue-specific patterns that are carefullu maintained and transmitted to the next generation of cells.\u003csup\u003e1)\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eAberrant DNA methylation patterns have been reported to be closely associated with various diseases, particularly cancer. Tumor cells often exhibit a 5\u0026ndash;10% reduction in overall 5-methylcytosine levels compared to normal cells, while localized hypermethylation in many CpG islands.\u003csup\u003e2\u0026ndash;3)\u003c/sup\u003e For instance, promoter hypermethylation of \u003cem\u003eRB\u003c/em\u003e (retinoblastoma), a gene that regulates the cell cycle, occurs in about 10% of sporadic unilateral retinoblastomas, leading to reduced expression of the RB protein.\u003csup\u003e4\u0026ndash;5)\u003c/sup\u003e Similarly, in sporadic colon tumors, promoter methylation of \u003cem\u003eMLH1\u003c/em\u003e, a DNA mismatch repair-related gene, leads to reduced MLH1 expression, and treatment with the demethylating agent 5-aza-2'-deoxycytidine was found to decrease MLH1 suppression.\u003csup\u003e6\u0026ndash;7)\u003c/sup\u003e Thus, hypermethylation of tumor suppressor genes in cancer cells can lead to their reduced expression, promoting cancer.\u003c/p\u003e \u003cp\u003eDespite numerous reports of abnormal gene expression regulation through DNA methylation of specific genes in cancer, \u003csup\u003e8)\u003c/sup\u003e a definitive explanation for the mechanism behind gene-specific DNA methylation remains elusive. DNA methylation is highly critical for gene expression regulation; hence, precise control of methylation patterns in numerous genes is essential. To achieve gene-specific DNA methylation regulation, an accurate regulatory mechanism must exist that can precisely target DNMT activity for each gene. Such a regulator should be capable of recognizing the nucleotide sequence of the target gene and have the versatility to match various genes. Additionally, it should be a substance that has been reported to show differential expression patterns in tumors. From this perspective, we have considered the possibility of microRNAs, which have already been reported in the thousands and are known to bind to the 3' untranslated regions of multiple target mRNAs, thereby inhibiting protein translation of the target genes \u003csup\u003e9\u0026ndash;10)\u003c/sup\u003e, as potential guider for DNMT regulation.\u003c/p\u003e \u003cp\u003eMicroRNAs are small non-coding RNAs composed of 21\u0026ndash;25 nucleotides that function primarily as post-transcriptional regulators in eukaryotic gene expression.\u003csup\u003e11\u0026ndash;13)\u003c/sup\u003e Unlike mRNA, which acts as an intermediary in protein expression, the function of microRNAs is based on their ability to recognize and bind to their own nucleotide sequence. In addition to its post-transcriptional regulation function, small RNAs have been shown to mediate DNA methylation in plants through a phenomenon known as the RNA-directed DNA methylation (RdDM) pathway.\u003csup\u003e14)\u003c/sup\u003e The RdDM pathway has been observed in some yeasts\u003csup\u003e15)\u003c/sup\u003e, and several studies reporting RdDM-like mechanisms have been reported in human cells.\u003csup\u003e16\u0026ndash;18)\u003c/sup\u003e Additionally, processes similar to RNA-directed DNA methylation (RdDM) have been reported in human cells, but they mostly involve self-feedback mechanisms, such as methylation of siRNA-derived repetitive sequences or de novo methylation of CpG islands by promoter antisense-derived small RNAs.\u003csup\u003e19)\u003c/sup\u003e However, the molecular mechanism by which endogenous microRNAs in human tumor cells recognize specific target genes outside of microRNA-derived DNA sequences and induce methylation has not been elucidated.\u003c/p\u003e \u003cp\u003eIn this study, we aimed to explore the relationship between the expression levels of 734 miRNAs and the methylation status of 20,587 genes across 813 cell lines obtained from public databases through rigorous statistical analysis. By doing so, we seek to ascertain the potential of specific miRNAs in selectively regulating the methylation of CpG islands in target genes.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eDataset\u003c/h2\u003e \u003cp\u003eThis study was conducted based on the data collected from the Cancer Cell Lines Encyclopedia (CCLE) database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://sites.broadinstitute.org/ccle\u003c/span\u003e\u003cspan address=\"https://sites.broadinstitute.org/ccle\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e), which provides large-scale genomic data from cancer cell lines. We obtained expression data for 734 microRNAs and methylation data for 54,531 promoter CpG clusters (corresponding to 20,587 genes). To perform linear regression analysis on the target genes selected through Spearman's rank correlation analysis of microRNA expression and gene cluster methylation, mRNA expression data were also obtained from CCLE. The analysis was conducted on a common set of 813 cell lines that were identified in all three datasets.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eSpearman's rank correlation analysis between the miRNA expression and specific gene cluster methylation levels of each gene was performed for all cases, obtaining correlation coefficients, \u003cem\u003ep\u003c/em\u003e-values, and the false discovery rate (FDR) for each pair. A correlation with an FDR value below 0.05 was considered significant. For the regression analysis of the final selected target pairs, we sequentially performed simple linear regression and multiple linear regression for the continuous dependent variable. One way analysis of variance (ANOVA) was conducted to compare differences among continuous variables. Post hoc tests were performed to verify differences between all groups. A \u003cem\u003ep\u003c/em\u003e-value less than 0.05 was considered statistically significant in all statistical analyses. All statistical analyses and visualizations were conducted using software version 4.0.3.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eFiltering of promoter CpG island clusters\u003c/h3\u003e\n\u003cp\u003eWe performed a selection process to identify gene clusters with potential promoter CpG islands, excluding non-coding RNA (RefSeq id: NM) gene clusters. Our selection criteria were based on a self-generated standard, taking into account the form and characteristics of the methylation data. Specifically, we defined CpG island clusters as those with a minimum of 20 CpG sites within a cluster and a ratio of CpG sites within the entire cluster of 10% or more. Ultimately, we identified 5,070 gene promoter CpG island clusters through this process.\u003c/p\u003e\n\u003ch3\u003eCancer type grouping\u003c/h3\u003e\n\u003cp\u003eA total of 26 cancer types were identified in a set of 813 cell lines that shared common microRNA expression data sets, gene methylation data sets, and mRNA expression data sets. We performed clustering of the entire microRNA expression values using the NMF consensus tool in Gene Pattern (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.genepattern.org/\u003c/span\u003e\u003cspan address=\"https://www.genepattern.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). This analysis was conducted under conditions ranging from k\u0026thinsp;=\u0026thinsp;1 to k\u0026thinsp;=\u0026thinsp;12. The microRNA expression patterns of the cell lines were clustered based on their similarity, and the k-value was set at the point where the cophenetic coefficient showed a sharp drop and clear clustering. we employed classification into a single group for cancer types only when they demonstrated similar microRNA expression patterns in two or more instances.\u003c/p\u003e\n\u003ch3\u003eGene ontology analysis\u003c/h3\u003e\n\u003cp\u003eTo investigate the interactions and biological characteristics of the selected genes, gene ontology analysis was performed using the Enrichr (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://maayanlab.cloud/Enrichr/\u003c/span\u003e\u003cspan address=\"https://maayanlab.cloud/Enrichr/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e) database. Adjusted \u003cem\u003ep\u003c/em\u003e-values of 0.05 or higher were considered statistically significant in the Biological Process, Molecular Function, and Cellular Component categories for the 25 selected genes.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eSequence homology\u003c/h2\u003e \u003cp\u003eTo verify the sequence homology between the selected microRNAs and gene clusters, we utilized BLASTn (version: 2.2.31+) from the Basic Local Alignment Search Tool. Since microRNAs are relatively short, approximately 22 bp, and we specifically focused on assessing homology within the promoter CpG island of specific genes, we employed the blastn-short program, which is suitable for identifying sequence homology for sequences with fewer than 30 bp. The nucleotide sequence of the target gene cluster was defined as the region between the first CpG site and the last CpG site. To convert these sequences into a database, the Makblastdb package was employed. Subsequently, the homology between the database and query sequence was checked by setting the microRNA as the query sequence. In the known mechanism of microRNA targeting in the mRNA 3' UTR, the seed sequence of 2\u0026ndash;8 bp plays a crucial role. However, our study identified a different mechanism for the targeting of microRNAs. Therefore, we considered the entire mature microRNA sequence rather than limiting it to the seed sequence. We set the word size option, which represents the minimum matching length, to 7 bp.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003eIdentification of miRNA-Gene Promoter CpG Methylation Pairs with Confirmed Correlations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo identify target pairs exhibiting a strong correlation between miRNA expression and gene promoter region methylation, and to investigate whether these target pairs have interdependent relationships that also influence mRNA expression, Spearman\u0026apos;s rank correlation analysis followed by sequential regression analyses were performed \u003cstrong\u003e(Figure 1)\u003c/strong\u003e. Based on Spearman\u0026apos;s rank correlation analysis between microRNA expression and gene promoter region methylation, 316,671 microRNA-gene cluster pairs exhibited significant correlations (FDR \u0026lt; 0.05). Among them, 50,813 pairs were predicted to correspond to gene clusters within a promoter CpG island (CGI). There were 929 pairs with a positive correlation (r \u0026ge; +0.3, \u003cstrong\u003eFigure 2a\u003c/strong\u003e) and 777 pairs with a negative correlation (r \u0026le; -0.4, \u003cstrong\u003eFigure 2b\u003c/strong\u003e). Next, we implemented an alternative correlation coefficient threshold to address the imbalance between positive and negative correlation pairs in the overall correlation analysis results.\u0026nbsp;This adjustment was necessary due to a considerable surplus of negative correlation pairs compared to positive ones. Following this, we carried out an additional selection process to identify target pairs of greater significance based on the correlation analysis outcomes. Based on the microRNA expression and correlation FDR values (log2 transformed), the positive and negative correlation pairs were each divided into four intervals. From these intervals, pairs with relatively high mean microRNA expression and low FDR values across all cell lines were selected (\u003cstrong\u003eFigure 2a\u003c/strong\u003e and \u003cstrong\u003eFigure 2b\u003c/strong\u003e).\u0026nbsp;Consequently, 53 positive correlated pairs (\u003cstrong\u003eFigure 2c\u003c/strong\u003e) and 71 negative correlated pairs (\u003cstrong\u003eFigure 2d\u003c/strong\u003e) were identified within the selected intervals [positive: log2(miRNA expression mean) \u0026gt; 9.18, log2(FDR) \u0026lt; -76.42; negative: log2(miRNA expression mean) \u0026gt; 8.48, log2(FDR) \u0026lt; -211.82]. For future experimental validation, we then applied a criterion of relatively higher microRNA expression. We found that 12 gene clusters were positively regulated by these microRNAs based on their expression, while 52 gene clusters were negatively regulated. The final target microRNAs and gene clusters are presented in \u003cstrong\u003eTable 1\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSequential regression analyses of microRNA expression and promoter methylation effects on mRNA expression.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo investigate whether the target pairs exhibiting strong correlations in microRNA expression and promoter methylation have a dependent relationship and also influence mRNA expression, we performed sequential three regression analyses. First, a simple linear regression (first SLR) analysis was performed on the gene promoter methylation influenced by microRNA expression. The results showed significant positive correlations for all 12 target pairs with \u0026beta; \u0026gt; 0 and \u003cem\u003ep\u003c/em\u003e-value \u0026lt; 0.05. Bonferroni correction confirmed statistically significant results for all 12 positive target pairs (\u003cstrong\u003eFigure 3a\u003c/strong\u003e). Additionally, all 52 pairs with negative correlations showed statistically significant results with an \u0026beta; \u0026lt; 0 and \u003cem\u003ep\u003c/em\u003e-value \u0026lt; 0.05. However, two pairs, has-miR-200b \u0026amp;\u003cem\u003e\u0026nbsp;DSP_\u003c/em\u003e2 and has-miR-200b \u0026amp;\u003cem\u003e\u0026nbsp;FAM83H_3\u003c/em\u003e, did not show significant results after Bonferroni correction. Excluding these two pairs, significant results were observed for 50 pairs (\u003cstrong\u003eFigure 3b\u003c/strong\u003e). Subsequently, to validate the well-known theory of DNA promoter methylation-mediated mRNA expression regulation in our dataset, we performed a second simple linear regression analysis (second SLR) on the gene promoter methylation and the corresponding mRNA expression. We separated the gene clusters that showed positive and negative correlations with the target microRNAs into positive and negative groups, respectively, and examined mRNA expression for each microRNA (\u003cstrong\u003eFigure 3c\u003c/strong\u003e and\u003cstrong\u003e\u0026nbsp;Figure 3d\u003c/strong\u003e). We found a significant negative correlation between promoter methylation and mRNA expression in both groups. In the SLR2 results for each target pair, excluding hsa-miR-141 \u0026amp;\u003cem\u003e\u0026nbsp;TJP2_3\u003c/em\u003e and hsa-miR-200c \u0026amp; \u003cem\u003eTJP2_3\u003c/em\u003e from among the positive pairs, 10 pairs showed significant results with \u0026beta; \u0026lt; 0 and \u003cem\u003ep\u003c/em\u003e-value \u0026lt; 0.05, while all negative pairs showed significant results with \u0026beta; \u0026lt; 0 and \u003cem\u003ep\u003c/em\u003e-value \u0026lt; 0.05 (\u003cstrong\u003eSupplementary Table 1\u003c/strong\u003e). Finally, we conducted a multiple linear regression (MLR) analysis, considering two independent variables together, to examine whether microRNA-mediated methylation regulation affects mRNA expression. Among the 12 target pairs with positive correlations, hsa-miR-141 \u0026amp; \u003cem\u003eTJP2_3\u0026nbsp;\u003c/em\u003eand hsa-miR-200c \u0026amp; \u003cem\u003eTJP2_3\u003c/em\u003e did not show \u0026beta; \u0026lt; 0 for both microRNA-mRNA expression and methylation-mRNA expression, contrary to the hypothesis we aimed to test. Excluding these two pairs, \u0026beta; \u0026lt; 0 was observed in both microRNA-mRNA expression and methylation-mRNA expression for 10 pairs, and 8 pairs were significant with p-value \u0026lt; 0.05. For negative correlations, all pairs showed \u0026beta; \u0026gt; 0 for microRNA-mRNA expression and \u0026beta; \u0026lt; 0 for methylation-mRNA expression, consistent with our hypothesis, and 51 pairs, except for one pair with p-value \u0026gt; 0.05, were statistically significant (\u003cstrong\u003eTable 2\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eComparing the significance of target pairs by cancer group\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe performed NMF clustering on the expression of all microRNAs (n = 734) in the dataset to explore the significance of target pairs across cancer types. A rebound was observed in the cophenetic coefficient at K = 5 and K = 6, but we proceeded with grouping using K = 5, which exhibited the most distinct clustering (\u003cstrong\u003eFigure 4a\u003c/strong\u003e). Among the 26 types of cancer, only those with similar clusters containing two or more types of cancer were grouped together. Consequently, the cancer types were categorized into seven groups. We also included fibroblasts, which are expected to exhibit patterns similar to normal cells, resulting in a total of eight groups (\u003cstrong\u003eFigure 4b\u003c/strong\u003e). Among the eight groups,\u0026nbsp;groups 3, 6, 7, and 8 were observed to have low expression of the target microRNAs relative to the other cancer-type groups (\u003cem\u003ep\u003c/em\u003e \u0026lt; 2.2e-16 by one-way ANOVA, \u003cstrong\u003eFigure 4c\u003c/strong\u003e). Using the same method, we performed simple linear regression analysis (SLR) of the target microRNA expression and the gene cluster methylation in each cancer group. Group 1 (colon cancer, gastric cancer, lung cancer, and ovarian cancer), which included the largest number of cancer types, showed SLR results that were the most similar to those of the overall cancer types, with 7 positive pairs and 45 negative pairs (81.2% of the target pairs, \u003cstrong\u003eFigure 4d\u003c/strong\u003e). On the other hand, the SLR results of group 2 (breast cancer, endometrial/uterine cancer, and pancreatic cancer) and group 3 (brain cancer, liver cancer, and skin cancer), which included only three types of cancer each and with a similar number of cell lines, 103 and 109, respectively, exhibited markedly distinct trends (\u003cstrong\u003eFigure 4d\u003c/strong\u003e). The positive target pairs showed the same result as total cancer for two pairs (2/12) for both group 2 and group 3, while the negative target pairs showed the same result for 35 pairs (35/52) for group 2 but only one pair (1/52) for group 3. Hsa-miR-141 \u0026amp; \u003cem\u003eATP8B2_1\u003c/em\u003e and hsa-miR-141 \u0026amp; \u003cem\u003eATP8B2_5\u003c/em\u003e were the most frequent positive target pairs, exhibiting the same results as total cancer in three groups (i.e., group 1, group 3, and group 5). In contrast, hsa-miR-200b \u0026amp; \u003cem\u003eESRP2_2\u003c/em\u003e was the most frequent negative target pair, demonstrating the same results as total cancer in all four groups (i.e., group 1, group 2, group 3, and group 5).\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we investigated the possibility that microRNA regulates gene promoter methylation using data collected from open-source databases. We examined the correlation between 734 microRNAs and various gene promoter CpG methylation using Spearman rank correlation analysis, identifying significant target microRNA and gene cluster pairs. In addition, we validated the dependency between the identified pairs through sequential linear regression analysis. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e depicts the workflow for the DNA methylation, microRNA and mRNA correlation, and the regression analysis. Through this process, we confirmed that the selected targets were indeed likely to be regulated by microRNA-mediated methylation of specific genes (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). For the second simple linear regression (SLR2) analysis on gene promoter methylation and mRNA expression, we found that, regardless of the directionality and correlation between the microRNA and gene clusters, all target pairs showed a significant negative correlation between gene promoter CpG methylation and mRNA expression (β\u0026thinsp;\u0026lt;\u0026thinsp;0, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), as reported in previous studies. This was further confirmed by our data (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec, b and \u003cb\u003eSupplementary Table\u0026nbsp;1\u003c/b\u003e). Multiple linear regression considering both microRNA expression and promoter methylation on mRNA expression showed that, except for some pairs, gene promoter methylation by microRNA could also affect mRNA expression (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). However, mRNA expression is a complex mechanism that can be regulated by various factors, so promoter methylation alone is not sufficient to fully explain the expression patterns. The correlation explanatory power between the selected microRNA and gene cluster pairs, derived from the analysis encompassing all types of cancer, was found to be highest in cancer group 1, which includes colon cancer, gastric cancer, lung cancer, and ovarian cancer). Group 2 (breast cancer, endometrial/uterine cancer, and pancreatic cancer) and group 3 (brain cancer, liver cancer, and skin cancer), which had similar cell numbers, showed substantial differences in the simple linear regression results of the target microRNA expression and the gene cluster methylation. It is likely that this was due to differences in the expression of target microRNAs across cancer types, particularly between Group 2 and Group 3 (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec). Furthermore, we observed a lower proportion of statistically significant target pairs in the SLR results for Groups 3, 4, 6, 7, and 8, which had overall lower levels of target microRNA expression (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec and Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ed).\u003c/p\u003e \u003cp\u003eTo assess the enrichment of 25 target genes, gene ontology analysis was conducted and identified significant terms with a \u003cem\u003ep\u003c/em\u003e-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 in the Biological Process (BP), Molecular Function (MF), and Cellular Component (CC) categories (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e). \u003cem\u003eDSP, CDH3, MARVELD2\u003c/em\u003e, and \u003cem\u003eTJP2\u003c/em\u003e were found to be significantly associated with cell-cell junction organization (GO:0045216) with the lowest \u003cem\u003ep\u003c/em\u003e-value (7.03e-04) in BP. In MF, the keratin filament binding (GO:1990254) of \u003cem\u003eFAM83H\u003c/em\u003e and \u003cem\u003eVIM\u003c/em\u003e showed the most significant association with a \u003cem\u003ep\u003c/em\u003e-value of 2.24e-05. In CC, intermediate filament (GO:0005882) was significantly associated with \u003cem\u003eFAM83H, DSP, VIM\u003c/em\u003e, and \u003cem\u003ePPL\u003c/em\u003e, with the lowest \u003cem\u003ep\u003c/em\u003e-value (4.21e-07). Notably, the genes \u003cem\u003eMARVELD2, EPCAM\u003c/em\u003e, and \u003cem\u003eTJP2\u003c/em\u003e showed functional roles in bicellular tight junctions (GO:0005923), tight junctions (GO:0070160), and the apical junction complex (GO:0043296) in the CC.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003ePromoter methylation of tumor suppressor genes is an important mechanism in tumorigenesis.\u003csup\u003e20)\u003c/sup\u003e \u003cem\u003eST14\u003c/em\u003e, a tumor suppressor gene, was identified as a target gene in this study and was found to be negatively regulated by promoter methylation mediated by hsa-miR-200b, hsa-miR-200c, and hsa-miR-141. According to Kim et al., the expression of Suppressor of tumorigenicity 14 (\u003cem\u003eST14\u003c/em\u003e) or Serine protease 14 (\u003cem\u003ePrss14\u003c/em\u003e) genes are associated with a poor prognosis and increased invasiveness and metastatic potential in breast cancer patients.\u003csup\u003e21)\u003c/sup\u003e These facts suggest that microRNA-mediated regulation of \u003cem\u003eST14\u003c/em\u003e gene methylation at the 5' end, and the subsequent changes in mRNA expression, may function as a regulatory factor in tumor invasiveness and metastasis. Another target gene, \u003cem\u003eOVOL1\u003c/em\u003e, is a transcription factor that induces the reversal process of EMT (M-EMT, mesenchymal to epithelial transition) in human cancer cells, thereby inhibiting tumor invasion and metastasis, and affecting the growth, apoptosis, and mobility of cancer cells.\u003csup\u003e22)\u003c/sup\u003e The target gene \u003cem\u003eEPCAM\u003c/em\u003e (Epithelial cell adhesion molecule), frequently reported in the Gene Ontology analysis, is an important marker gene for cancer diagnosis and treatment, as it is overexpressed by various types of cancer cells, including colon and breast cancer.\u003csup\u003e23,24)\u003c/sup\u003e The possibility that the promoter methylation of \u003cem\u003eEPCAM\u003c/em\u003e can be regulated by microRNA suggests a new perspective on cancer diagnosis and treatment.\u003c/p\u003e \u003cp\u003eSome previous studies have used bioinformatics to investigate the correlation between miRNA and CpG methylation.\u003csup\u003e25,26)\u003c/sup\u003e Such studies have focused on the interaction between DNA methylation and miRNA expression in the context of the phenotype of breast cancer \u003csup\u003e25)\u003c/sup\u003e or the mechanisms of racial heterogeneity in hepatocellular carcinoma.\u003csup\u003e26)\u003c/sup\u003e They mainly discuss the regulation of microRNA expression by methylation or the inhibition of DNMT enzymes by microRNA. However, these studies have not specifically focused on gene promoter methylation mediated by microRNA. Therefore, our study can be considered one of the first to screen the targets of microRNA-mediated promoter methylation in 26 diverse cancer types. However, the results of the current analysis have limitations in proving the hypothesis that microRNA guides promoter methylation of specific genes. To more clearly confirm this hypothesis, direct evidence of an interaction between microRNA and DNMT is needed. We plan to perform additional experiments, such as microRNA knockdown using siRNA, microRNA transfection, and chromatin immunoprecipitation sequencing (ChIP-seq). Before conducting experimental validation, we randomly verified the sequence homology between the target gene clusters and the microRNAs using BLAST. The results are shown in \u003cb\u003eSupplementary Table\u0026nbsp;2\u003c/b\u003e and \u003cb\u003eSupplementary Table\u0026nbsp;3\u003c/b\u003e, and we were able to confirm sequence homology in some target pairs. However, due to the short query sequence length of approximately 22 base pairs and the limited matching of microRNA and gene cluster sequences with significant correlations, there was a notable disparity in the e-value and bit-score when compared to the general BLAST (the detailed methods can be found in the \"sequence homology\" section of the \u003cb\u003eMaterials and Methods\u003c/b\u003e)\u003c/p\u003e \u003cp\u003eNevertheless, through screening using bioinformatics, we efficiently explored a large dataset to identify target pairs with strong correlations, and we validated the expression of mRNA for the target pairs using an mRNA database. This study is a foundational step in identifying the targets of microRNA-mediated promoter methylation phenomena. These results need further experimental validation. Our findings expand our understanding of the mechanisms underlying tumor formation through methylation and provide a new perspective on the use of microRNA in cancer treatment.\u003c/p\u003e"},{"header":"Declarations","content":" \u003ch2\u003eEthics approval and consent to participate\u003c/h2\u003e \u003cp\u003eNot applicable. This study utilizes open data from the CCLE, and therefore, no separate ethics approval and consent to participate is required.\u003c/p\u003e \u003ch2\u003ePatient consent for publication\u003c/h2\u003e \u003cp\u003eNot applicable. This study utilizes open data from the CCLE, and therefore, no patient consent for publication is required.\u003c/p\u003e \u003ch2\u003eCompeting interests\u003c/h2\u003e \u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eThis study was supported by the grants (2019IP0836-1, 2023IP0085-2) from the Asan Institute for Life Sciences, Asan Medical Center, Seoul, Korea.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eS.J. performed most of the data analysis with assistance from S-M.C., C.O.S., J-Y.L. and H.R.J.; S.J. wrote the manuscript; H.R.J. validate the analyzed result; S-M.C. designed the study and edited the manuscript.\u003c/p\u003e\u003ch2\u003eAcknowledgments\u003c/h2\u003e \u003cp\u003eNot applicable\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eAll of the datasets used in this study were downloaded from the Cancer Cell Lines Encyclopedia (CCLE) database: https://sites.broadinstitute.org/ccle.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eRakyan, Vardhman K., et al. DNA methylation profiling of the human major histocompatibility complex: a pilot study for the human epigenome project. PLoS biology. 2(12), e405 (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFeinberg, A. P., Gehrke, C. W., Kuo, K. C., \u0026amp; Ehrlich, M. Reduced genomic 5-methylcytosine content in human colonic neoplasia. Cancer research. 48(5), 1159\u0026ndash;1161 (1988).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEhrlich, M. DNA hypomethylation in cancer cells. Epigenomics, \u003cem\u003e1\u003c/em\u003e(2), 239\u0026ndash;259 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGreger V, Debus N, Lohmann D, Hopping W, Passarge E, Horsthemke B, Debus N, Lohmann D, Hopping W, Passarge E, Horsthemke B: Frequency and parental origin of hypermethylated RB1 alleles in retinoblastoma. Hum Genet. 94(5), 491\u0026ndash;496 (1994).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOhtani-Fujita, N., et al. \"CpG methylation inactivates the promoter activity of the human retinoblastoma tumor-suppressor gene.\" Oncogene. 8(4), 1036\u0026ndash;1067 (1993).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKane, Michael F., et al. \"Methylation of the hMLH1 promoter correlates with lack of expression of hMLH1 in sporadic colon tumors and mismatch repair-defective human tumor cell lines.\" Cancer research. 57(5), 808\u0026ndash;811 (1997).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHerman, James G., et al. Incidence and functional consequences of hMLH1 promoter hypermethylation in colorectal carcinoma. \u003cem\u003eProceedings of the National Academy of Sciences.\u003c/em\u003e 95(12), 6870\u0026ndash;6875 (1998).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchinke, Carolina, et al. Aberrant DNA methylation in malignant melanoma. Melanoma research. 20(4), 253\u0026ndash;265 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBousquet, M., Nguyen, D., Chen, C., Shields, L., \u0026amp; Lodish, H. F. MicroRNA-125b transforms myeloid cell lines by repressing multiple mRNA. Haematologica. 97(11), 1713 (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRitchie, William, John EJ Rasko, and St\u0026eacute;phane Flamant. MicroRNA target prediction and validation. MicroRNA Cancer Regulation: Advanced Concepts, Bioinformatics and Systems Biology Tools. 39\u0026ndash;53 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKhvorova, A., Reynolds, A., \u0026amp; Jayasena, S. D. Functional siRNAs and miRNAs exhibit strand bias. Cell. 115(2), 209\u0026ndash;216 (2003).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChendrimada, Thimmaiah P., et al. TRBP recruits the Dicer complex to Ago2 for microRNA processing and gene silencing. Nature. 436(7051), 740\u0026ndash;744 (2005).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmeres, Stefan L., and Phillip D. Zamore. Diversifying microRNA sequence and function. Nature reviews Molecular cell biology. 14(8), 475\u0026ndash;488 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWassenegger, Michael. RNA-directed DNA methylation. Plant molecular biology. 43, 203\u0026ndash;220 (2000).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVerdel, Andr\u0026eacute;, et al. RNAi-mediated targeting of heterochromatin by the RITS complex. Science. 303(5658), 672\u0026ndash;676 (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCastanotto, Daniela, et al. Short hairpin RNA-directed cytosine (CpG) methylation of the RASSF1A gene promoter in HeLa cells. Molecular Therapy. 12(1). 179\u0026ndash;183 (2005).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKawasaki, H., \u0026amp; Taira, K. Induction of DNA methylation and gene silencing by short interfering RNAs in human cells. Nature. 431(7005), 211\u0026ndash;217 (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMorris, K. V., Chan, S. W. L., Jacobsen, S. E., \u0026amp; Looney, D. J. Small interfering RNA-induced transcriptional gene silencing in human cells. Science. 305(5688), 1289\u0026ndash;1292 (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTufarelli, Cristina, et al. Transcription of antisense RNA leading to gene silencing and methylation as a novel cause of human genetic disease. Nature genetics. 34(2), 157\u0026ndash;165 (2003).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiyanage, Chamikara, et al. Promoter hypermethylation of tumor-suppressor genes p16INK4a, RASSF1A, TIMP3, and PCQAP/MED15 in salivary DNA as a quadruple biomarker panel for early detection of oral and oropharyngeal cancers. Biomolecules 9(4),148 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim, S., Yang, J. W., Kim, C., \u0026amp; Kim, M. G. Impact of suppression of tumorigenicity 14 (ST14)/serine protease 14 (Prss14) expression analysis on the prognosis and management of estrogen receptor negative breast cancer. Oncotarget. 7(23), 34643 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRoca, Hernan, et al. Transcription factors OVOL1 and OVOL2 induce the mesenchymal to epithelial transition in human cancer. PloS one. 8(10). e76773 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCimino, Ashley, et al. Epithelial cell adhesion molecule (EpCAM) is overexpressed in breast cancer metastases. Breast cancer research and treatment. 123, 701\u0026ndash;708 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSpizzo, Gilbert, et al. EpCAM expression in primary tumour tissues and metastases: an immunohistochemical analysis. Journal of clinical pathology. 64(5), 415\u0026ndash;420 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAure, Miriam Ragle, et al. Crosstalk between microRNA expression and DNA methylation drives the hormone-dependent phenotype of breast cancer. Genome medicine. 13(1), 72 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVarghese, Rency S., et al. Integrative analysis of DNA methylation and microRNA expression reveals mechanisms of racial heterogeneity in hepatocellular carcinoma. Frontiers in genetics. 12, 708326 (2021).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTable 1 and 2 are available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"DNA Methylation, MicroRNA (miRNA), Epigenetic regulation, CpG island","lastPublishedDoi":"10.21203/rs.3.rs-5391278/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5391278/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eDNA methylation plays a crucial role in the epigenetic regulation of gene expression and is closely associated with the development of cancer. Abnormal regulation of gene expression due to changes in DNA methylation patterns of specific genes has been frequently observed in tumor cells. However, the specific mechanisms underlying the induction of DNA methylation in certain genes have not been fully elucidated. This study aimed to investigate the potential of microRNAs (miRNAs) as statistically significant regulators guiding promoter methylation of specific genes. MiRNAs are known to specifically recognize DNA sequences and exhibit various degrees of complementarity. We performed Spearman's rank correlation between the expression levels of 734 microRNAs and the CpG island methylation levels of 20,587 genes, collected from 813 cell lines in the Cancer Cell Line Encyclopedia (CCLE) database. Subsequently, we validated the dependent relationship between the selected target microRNAs and gene clusters using linear regression analysis. We identified 25 target genes in which promoter methylation was induced by the expression of four target miRNAs (hsa-miR-200a, hsa-miR-200b, hsa-miR-200c, and hsa-miR-141) with statistically significant values.\u003c/p\u003e \u003cp\u003eThe correlations of the target pairs between methylation level of target genes and matched miRNAs were most pronounced in colorectal, gastric, lung, and ovarian cancers. Cancer-related genes, including \u003cem\u003eST14, OVOL1\u003c/em\u003e, and \u003cem\u003eEPCAM\u003c/em\u003e, were identified as the target genes, confirming the possibility that promoter methylation of these genes is regulated by miRNA. Using bioinformatics-based screening analysis, we discovered target pairs that exhibited statistically significant changes in promoter methylation patterns due to specific miRNA expression. Furthermore, we confirmed the potential for miRNAs to regulate the expression of cancer-related genes through miRNA-induced promoter methylation. This expands our understanding of the mechanism underlying tumor development through methylation and provides a new perspective on the utilization of microRNAs in the field of cancer treatment.\u003c/p\u003e","manuscriptTitle":"Investigating miRNA-Driven DNA Methylation: Statistical Evidence of Gene-Specific Modulation","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-11-18 15:12:29","doi":"10.21203/rs.3.rs-5391278/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"dd68593b-88bd-4762-a80b-6745569925c9","owner":[],"postedDate":"November 18th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2024-11-27T10:23:51+00:00","versionOfRecord":[],"versionCreatedAt":"2024-11-18 15:12:29","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5391278","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5391278","identity":"rs-5391278","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}