Colon cancer versus colitis: a microRNA and gene expression analysis via system biology tools Authors: Fatemeh TajikRostami, Alireza Omidi Farzin, Babak Arjmand, Zahra Razzaghi, Majid Rezaei-Tavirani, Mostafa Rezaei-Tavirani, Behzad Hatami Published: 2025 ABSTRACT Aim:The molecular mechanism of this relationship can be detected that promote the transition from colitis to colon cancer via system biology tools is the main aim of this study. Background: There is a correlation between colitis and associated colon analysis. Methods:The differentially expressed microRNA and gene profiles of samples with colitis and colon cancer are extracted from the Gene Expression Omnibus (GEO) database and assessed using the GEO2R program and the CluePedia plugin of Cytoscape software to identify crucial actors and controlled genes or microRNAs.Results:Analyses revealed microRNA analysis was not suitable tool to detect molecular mechanism of colitis to associated colon cancer in mouse model but the gene expression profile assessment leads to introduce CLASP1, CENPM, KIF18A, PRS7, CENPI, ZWINT, SPC24, and CENPK as the critical genes that promote this transition in fibroblast cells of colon cancer.Conclusion:Regulation of ribosome and kinetochore was suggested as a possible treatment to prevent colitis – colon cancer transition. Introduction Ulcerative colitis (UC) is characterized by a chronic inflammatory condition of the colon, which increases the risk of colectomy and colorectal cancer in patients (1). Prevalence and incidence of UC as a worsening and dispatching disease are increasing in communities. Application of 5-aminosalicylate acid medications as the first-line treatment is suggested for patients with mild to moderate disease; however, some patients may need surgery (2). There are many documents that colitis can be considered a risk factor of colorectal cancer (CRC). It is reported that, colitis increases risk of CRC 2.4-fold (3, 4). Investigations revealed that the greasepaint of colitis-associated colorectal cancer (CAC) has similarities and differences with sporadic colorectal carcinoma (sCRC). The mouse model for CAC is often provided via administration of azoxymethane/dextran sodium sulfate (AOM/DSS). Experiments indicate that, this model can not display exactly greasepaint of CAC. Hence, scientists try to achieve a new model to study CAC cases (5). microRNAs (miRNAs), as the small non-coding RNAs, can regulate several genes and are involved in many processes and biological functions, such as the promotion of cancers (6). Results of many investigations revealed that miRNAs control gene expression in cells. For example, miR-142-3p and miR-128-3p target beta-Catenin and FOXO4 in HCT116 cells, respectively (7). Crucial roles of miRNAs in cancer promotion are widely investigated and confirmed. Some miRNAs are oncogenic, while the tumor-suppressor miRNAs play a role in preventing cancer development (8). miRNA and gene expression analysis are pointed out as useful methods to explore the molecular mechanisms of diseases. These analyses need advanced methods such as bioinformatic assessments to yield beneficial findings about diseases and disorders (9-11). PPI network analysis is a valuable method to evaluate interactions between a large set of genes to find the crucial individuals. The molecular mechanism of many diseases is assessed via PPI network analysis (12-14). Differentiation between colitis and CAC via miRNA expression analysis and gene expression evaluation is the main aim of this study. Data are retrieved from the GEO database and assessed via the GEO2R program and directed PPI network analysis. Methods Data collection and pre-evaluation analysis There are an expression profiling by array including 20 mouse colon miRNA expression profiles in the GEO database (https://www.ncbi.nlm.nih.gov/geo/geo2r/?acc=GSE304769), including controls, mice with chronic colitis induced by dextran sulfate sodium (DSS), and 7 individuals with tumor induced by azoxymethane (AOM) followed by DSS administration; the AOM-DSS group. These profiles were assessed via the GEO2R program to find the significant DEGs. Results were analyzed via a Venn diagram and a UMAP plot. Since the colon cancer-associated fibroblasts are part of the tumor microenvironment and involved in colon cancer development, and the colitic-associated fibroblasts secrete pro-inflammatory cytokines and play a considerable role in the promotion of inflammation of the colon. Gene expression changes of this cell type were searched in the GEO database (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE92945). An expression profiling by high throughput sequencing of gene expression profiles of colon cancer and colitis-associated fibroblast were downloaded and evaluated by GEO2R via UMAP plot and Venn diagram to find the significant DEGs. Directed PPI network analysis The gene significantly DEGs of the compared fibroblast cells were assessed via directed PPI network analysis by using the CluePedia plugin of Cytoscape software. The network was formed via activation, inhibition, and expression actions from CluePedia_STRING.ACTIONS_v11.09606_27.02.2 source. The created PPI network was analyzed by the “Network analyzer” application of Cytoscape software. The distribution of nodes based on out-degree and in-degree was assessed, and the top nodes were considered for further investigation. The top nodes and the related first neighbors were extracted from the main connected component and visualized to find a better understanding of the critical core of the created network. Statistical analysis The significant DEGs and differentially expressed miRNA were selected based on an adjusted p-value of < 0.05. Protein-protein network analysis and hub gene selection We further predicted the protein-protein interaction (PPI) network using all the DEGs in each dataset, including upregulated and downregulated genes from the STRING (https://string-db.org) database, aiming to provide more comprehensive information for the key genes. The visualization and analysis of the constructed network were performed by Cytoscape software version 3.7.0. Then, the top key genes with the highest degree index were identified through the Cytohubba plugin. To further evaluate internal interaction within commonly selected hub genes shared between two diseases, we predicted a co-expression network by using the GeneMANIA (https://genemania.org/) platform. UMAP plot of comparison between DSS, AOM-DSS, and control miRNA profiles. Venn diagram of comparison between DSS, AOM-DSS, and control miRNA profiles. UMAP plot of comparison between colon cancer and colitis-associated fibroblast gene expression profiles. Venn diagram of comparison between colon cancer and colitis-associated fibroblast gene expression profiles. The main connected component of the directed PPI network of comparison between colon cancer and colitis-associated fibroblast gene expression profiles. Green, red, and yellow refer to activation, inhibition, and expression actions, respectively. The top nodes based on out-degree and in-degree values are labeled in black and blue, respectively. Classification of the nodes of the main connected component of the PPI network based on out-degree value. An interval amount of 4 is considered based on the out-degree value. Construction of miRNA‑hub gene regulatory network for common hubs The miRNet (www.mirnet.ca) serves as a comprehensive online platform aimed at facilitating the integrative analysis of miRNAs-centric regulatory networks. This database aggregates a broad range of experimentally and predicted miRNA-target interactions to construct a network. In this analysis, we predicted a miRNA-hub gene regulatory network using the miRNet database (https://mirnet.ca) that targets common shared hub genes between GDM and NAFLD. Results UMAP plot of comparison between the DSS, AOM-DSS, and control miRNA profiles is shown in Figure 1. As depicted in Figure 1, the DSS and AOM-DSS groups are completely separated from the control group. However, DSS and AOM-DSS groups are divided weakly. Venn diagram of analysis is presented in Figure 2. As it is administered in Figure 2, AOM-DSS and DSS groups are discriminated from controls by a large number of miRNAs, but both treated groups are distinguished from each other by a few miRNAs. It seems, the miRNA profile of severe colitis closely mimics that of colitis-associated cancer in this specific mouse model. Analyses administered that the gene expression profiles of colon cancer-associated fibroblasts are discriminated from gene expression profiles of colitis-associated fibroblasts by 1214 DEGs (see Figures 3 and 4). Among the 1214 queried DEGs, 290 individuals were included in the components of the network. The main connected component of the directed PPI network is presented in Figure 5. The top node based on out-degree value, which was characterized by an out-degree of 22. The nodes were classified into groups with an interval of 4 based on out-degree value (see Figure 6). Ignoring the nodes with an out-degree of zero, 191 nodes clustered in six groups. As depicted in Figure 6, the top group includes 3 DEG, while the first group contains 130 genes. In a similar manner, the nodes of the main connected component were classified into five groups based on in-degree value (Figure 7). The top nodes based on out-degree and in-degree values are presented in Figure 8. The key genes and their centrality properties are shown in Table 1. Classification of the nodes of the main connected component of the PPI network based on in-degree value. An interval amount of 4 is considered based on the in-degree value. The critical actor (labeled in black) and controlled (shown in blue) genes and their first neighbors from the main connected component of the studied PPI. Discussion miRNA analysis showed colitis and CAC are differentiated from control samples completely via miRNA expression changes, but there was no difference between colitis and CAC (except for two dysregulated miRNAs). This may be from the weak differentiation of CAC from colitis or the occurrence of the early stage of CAC. However, gene expression analysis revealed significant differences between colitis and CAC diseases. Evolutionary genomic analysis revealed precancer clones, existing wide single-nucleotide alterations, and copy number alterations in precancer CAC via evolutionary genomic analysis, as reported by Baker AM et al (15). Gene expression analysis showed that colitis into CAC transition is accompanied by rigorous gene expression changes. The top three actors and five controlled genes of transition from colitis to ACA are shown in table 1. As depicted in Figure 8, except for RPS7, all critical genes are presented in a unique domain of the main connected component. The top actor is CLAPS1, which I characterized with an out-degree of 22 and is a unique up-regulated critical gene. Investigation indicates that there are two proteins (CLASP1 and CLASP2) in human cells that are essential for poleward flux via supporting elongation of tied microtubule ends under strength (16). Overall, CLASPs are involved in regulating microtubule regulation, which is required for promoting fundamental cellular processes (17). The second actor gene is CENPM, which is known as a crucial part of the constituent centromere-associated network. The essential role of this gene in the assembly of kinetochores and also chromosome organization is highlighted in the literature (18). Correlation between expression and several malignancies is confirmed by researchers. It is reported that overexpression of CENPM improves survival in colon adenocarcinoma. It plays a significant role in regulating the cell cycle and the microenvironment of the tumor (19). The third regulator gene is KIF18, which is involved in the occurrence and progression of tumors (20). It is suggested that KIF18A is a suitable drug target in colorectal cancer (21). It is pointed out that KIF18 up-regulates CENPE, which leads to activation of the PI3K/AKT pathway; the pathway that accelerates the progression of cervical squamous cell carcinoma (22). As shown in Table 1, CENPK and CENPI, the other members of the CENP family, are regulated by actor genes. Li et al introduced CENPK as a risk factor in colorectal cancer (23). Based on a report by Ding N et al, CENPI is up-regulated in colorectal cancer relative to normal samples and plays role in regulating cell migration and invasion (24). Both SPC24 and ZWINT are controlled genes that are related to the kinetochore complex. Kinetochores are specialized protein complex that facilitate interaction between spindle microtubules and chromosomes; the principal process in chromosome segregation (25). It is reported that hnRNP K is involved in the regulation of the NDC80 kinetochore complex in HCT116 colon cancer cells to stable spindle microtubule-kinetochore attachment (26). The last regulated critical gene is PRS7. There are documents about different patterns of ribosomal protein regulation between colorectal carcinomas and colorectal adenoma or normal mucosa. Up-regulation of some ribosomal proteins, besides down-regulation of others, is reported in this regard. It is confirmed that the expression patterns of these proteins depend on progression, differentiation, and metastasis of colorectal carcinomas. Since the involvement of ribosomal proteins in essential functions such as transcription and repair, DNA replication, cell growth and proliferation, RNA splicing and modification, cellular transformation, apoptosis, and development regulation is investigated and emphasized (27), the important role of these proteins' dysregulation in colorectal cancer occurrence and progression is an acceptable idea. Table 1List of key genes and their centrality properties. The nodes are ranked based on out-degree value. FC refers to fold change.No.GeneDescriptionLog2(FC)Out-degreeIn-degree1CLASP1Cytoplasmic linker associated protein 10.942232CENPMCentromere protein M-2.812113KIF18AKinesin family member 18A-2.352114PRS7ribosomal protein S7-1.411275CENPIcentromere protein I-2.274196ZWINTZW10 interacting kinetochore protein-2.072187SPC24SPC24 component of NDC80 kinetochore complex-3.001208CENPKcentromere protein K-2.04120 List of key genes and their centrality properties. The nodes are ranked based on out-degree value. FC refers to fold change. Conclusion In conclusion, the transition between colitis and associated colon cancer in fibroblast cells is accompanied by dysregulation of essential cellular functions such as cell division; the processes that are regulated by kinetochores, ribosomes, and the related genes. It can be suggested that CLASP1, some members of CENP, KIF18A, PRS7, ZWINT, and SPC24 be considered as possible drug targets in future investigations to prevent colitis-colon cancer transition. As this study relies solely on bioinformatics data from the GEO database, the lack of experimental validation (such as RT-PCR or Immunohistochemistry on patient samples) is a limitation.