Identification of MicroRNA-Target Gene-Transcription Factor Regulatory Networks in Colorectal Adenoma Using Microarray Expression Data

Bioinformatics screening of colorectal-cancer causing molecular signatures through gene expression profiles to discover therapeutic targets and candidate agents

BACKGROUND

Detection of appropriate receptor proteins and drug agents are equally important in the case of drug discovery and development for any disease. In this study, an attempt was made to explore colorectal cancer (CRC) causing molecular signatures as receptors and drug agents as inhibitors by using integrated statistics and bioinformatics approaches.

METHODS

To identify the important genes that are involved in the initiation and progression of CRC, four microarray datasets (GSE9348, GSE110224, GSE23878, and GSE35279) and an RNA_Seq profiles (GSE50760) were downloaded from the Gene Expression Omnibus database. The datasets were analyzed by a statistical r-package of LIMMA to identify common differentially expressed genes (cDEGs). The key genes (KGs) of cDEGs were detected by using the five topological measures in the protein-protein interaction network analysis. Then we performed in-silico validation for CRC-causing KGs by using different web-tools and independent databases. We also disclosed the transcriptional and post-transcriptional regulatory factors of KGs by interaction network analysis of KGs with transcription factors (TFs) and micro-RNAs. Finally, we suggested our proposed KGs-guided computationally more effective candidate drug molecules compared to other published drugs by cross-validation with the state-of-the-art alternatives of top-ranked independent receptor proteins.

RESULTS

We identified 50 common differentially expressed genes (cDEGs) from five gene expression profile datasets, where 31 cDEGs were downregulated, and the rest 19 were up-regulated. Then we identified 11 cDEGs (CXCL8, CEMIP, MMP7, CA4, ADH1C, GUCA2A, GUCA2B, ZG16, CLCA4, MS4A12 and CLDN1) as the KGs. Different pertinent bioinformatic analyses (box plot, survival probability curves, DNA methylation, correlation with immune infiltration levels, diseases-KGs interaction, GO and KEGG pathways) based on independent databases directly or indirectly showed that these KGs are significantly associated with CRC progression. We also detected four TFs proteins (FOXC1, YY1, GATA2 and NFKB) and eight microRNAs (hsa-mir-16-5p, hsa-mir-195-5p, hsa-mir-203a-3p, hsa-mir-34a-5p, hsa-mir-107, hsa-mir-27a-3p, hsa-mir-429, and hsa-mir-335-5p) as the key transcriptional and post-transcriptional regulators of KGs. Finally, our proposed 15 molecular signatures including 11 KGs and 4 key TFs-proteins guided 9 small molecules (Cyclosporin A, Manzamine A, Cardidigin, Staurosporine, Benzo[A]Pyrene, Sitosterol, Nocardiopsis Sp, Troglitazone, and Riccardin D) were recommended as the top-ranked candidate therapeutic agents for the treatment against CRC.

CONCLUSION

The findings of this study recommended that our proposed target proteins and agents might be considered as the potential diagnostic, prognostic and therapeutic signatures for CRC.

ESRRG, ATP4A, and ATP4B as Diagnostic Biomarkers for Gastric Cancer: A Bioinformatic Analysis Based on Machine Learning

Based on multiple bioinformatics methods and machine learning techniques, this study was designed to explore potential hub genes of gastric cancer with a diagnostic value. The novel biomarkers were detected through multiple databases of gastric cancer–related genes. The NCBI Gene Expression Omnibus (GEO) database was used to obtain gene expression files. Three hub genes (ESRRG, ATP4A, and ATP4B) were detected through a combination of weighted gene co-expression network analysis (WGCNA), gene–gene interaction network analysis, and supervised feature selection method. GEPIA2 was used to verify the differences in the expression levels of the hub genes in normal and cancer tissues in the RNA-seq levels of Genotype-Tissue Expression (GTEx) and The Cancer Genome Atlas (TCGA) databases. The objectivity of potential hub genes was also verified by immunohistochemistry in the Human Protein Atlas (HPA) database and transcription factor–hub gene regulatory network. Machine learning (ML) methods including data pre-processing, model selection and cross-validation, and performance evaluation were examined on the hub-gene expression profiles in five Gene Expression Omnibus datasets and verified on a GEO external validation (EV) dataset. Six supervised learning models (support vector machine, random forest, k-nearest neighbors, neural network, decision tree, and eXtreme Gradient Boosting) and one semi-supervised learning model (label spreading) were established to evaluate the diagnostic value of biomarkers. Among the six supervised models, the support vector machine (SVM) algorithm was the most effective one according to calculated performance metrics, including 0.93 and 0.99 area under the curve (AUC) scores on the test and external validation datasets, respectively. Furthermore, the semi-supervised model could also successfully learn and predict sample types, achieving a 0.986 AUC score on the EV dataset, even when 10% samples in the five GEO datasets were labeled. In conclusion, three hub genes (ATP4A, ATP4B, and ESRRG) closely related to gastric cancer were mined, based on which the ML diagnostic model of gastric cancer was conducted.

Network-Based Approach and IVI Methodologies, a Combined Data Investigation Identified Probable Key Genes in Cardiovascular Disease and Chronic Kidney Disease

In fact, the risk of dying from CVD is significant when compared to the risk of developing end-stage renal disease (ESRD). Moreover, patients with severe CKD are often excluded from randomized controlled trials, making evidence-based therapy of comorbidities like CVD complicated. Thus, the goal of this study was to use an integrated bioinformatics approach to not only uncover Differentially Expressed Genes (DEGs), their associated functions, and pathways but also give a glimpse of how these two conditions are related at the molecular level. We started with GEO2R/R program (version 3.6.3, 64 bit) to get DEGs by comparing gene expression microarray data from CVD and CKD. Thereafter, the online STRING version 11.1 program was used to look for any correlations between all these common and/or overlapping DEGs, and the results were visualized using Cytoscape (version 3.8.0). Further, we used MCODE, a cytoscape plugin, and identified a total of 15 modules/clusters of the primary network. Interestingly, 10 of these modules contained our genes of interest (key genes). Out of these 10 modules that consist of 19 key genes (11 downregulated and 8 up-regulated), Module 1 (RPL13, RPLP0, RPS24, and RPS2) and module 5 (MYC, COX7B, and SOCS3) had the highest number of these genes. Then we used ClueGO to add a layer of GO terms with pathways to get a functionally ordered network. Finally, to identify the most influential nodes, we employed a novel technique called Integrated Value of Influence (IVI) by combining the network's most critical topological attributes. This method suggests that the nodes with many connections (calculated by hubness score) and high spreading potential (the spreader nodes are intended to have the most impact on the information flow in the network) are the most influential or essential nodes in a network. Thus, based on IVI values, hubness score, and spreading score, top 20 nodes were extracted, in which RPS27A non-seed gene and RPS2, a seed gene, came out to be the important node in the network.

Identification of the critical genes and miRNAs in hepatocellular carcinoma by integrated bioinformatics analysis

Hepatocellular carcinoma (HCC) is a global health problem with complex etiology and pathogenesis. Microarray data are increasingly being used as a novel and effective method for cancer pathogenesis analysis. An integrative analysis of genes and miRNA for HCC was conducted to unravel the potential prognosis of HCC. Two gene microarray datasets (GSE89377 and GSE101685) and two miRNA expression profiles (GSE112264 and GSE113740) were obtained from Gene Expression Omnibus database. A total of 177 differently expressed genes (DEGs) and 80 differently expressed miRNAs (DEMs) were screened out. Functional enrichment of DEGs was proceeded by Clue GO and these genes were significantly enriched in the chemical carcinogenesis pathway. A protein-protein interaction network was then established on the STRING platform, and ten hub genes (CDC20, TOP2A, ASPM, NCAPG, AURKA, CYP2E1, HMMR, PRC1, TYMS, and CYP4A11) were visualized via Cytoscape software. Then, a miRNA-target network was established to identify the hub dysregulated miRNA. A key miRNA (hsa-miR-124-3p) was filtered. Finally, the miRNA-target-transcription factor network was constructed for hsa-miR-124-3p. The network for hsa-miR-124-3p included two transcription factors (TFs) and five targets. These identified DEGs and DEMs, TFs, targets, and regulatory networks may help advance our understanding of the underlying pathogenesis of HCC.

Identification of significant alteration genes, pathways and TFs induced by LPS in ARDS via bioinformatical analysis

Background and aims

Acute respiratory distress syndrome (ARDS) or acute lung injury (ALI) is one of the most common acute thoracopathy with complicated pathogenesis in ICU. The study is to explore the differentially expressed genes (DEGs) in the lung tissue and underlying altering mechanisms in ARDS.

Methods

Gene expression profiles of GSE2411 and GSE130936 were available from GEO database, both of them included in GPL339. Then, an integrated analysis of these genes was performed, including gene ontology (GO) and KEGG pathway enrichment analysis in DAVID database, protein–protein interaction (PPI) network construction evaluated by the online database STRING, Transcription Factors (TFs) forecasting based on the Cytoscape plugin iRegulon, and their expression in varied organs in The Human Protein Atlas.

Results

A total of 39 differential expressed genes were screened from the two datasets, including 39 up-regulated genes and 0 down-regulated genes. The up-regulated genes were mainly enriched in the biological process, such as immune system process, innate immune response, inflammatory response, and also involved in some signal pathways, including cytokine–cytokine receptor interaction, Salmonella infection, Legionellosis, Chemokine, and Toll-like receptor signal pathway with an integrated analysis. GBP2, IFIT2 and IFIT3 were identified as hub genes in the lung by PPI network analysis with MCODE plug-in, as well as GO and KEGG re-enrichment. All of the three hub genes were regulated by the predictive common TFs, including STAT1, E2F1, IRF1, IRF2, and IRF9.

Conclusions

This study implied that hub gene GBP2, IFIT2 and IFIT3, which might be regulated by STAT1, E2F1, IRF1, IRF2, or IRF9, played significant roles in ARDS. They could be potential diagnostic or therapeutic targets for ARDS patients.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-021-06578-7.

Bioinformatics Analysis of Choriocarcinoma-Related MicroRNA-Transcription Factor-Target Gene Regulatory Networks and Validation of Key miRNAs

Objective

The aim of the current research was to construct a miRNA-transcription factor (TF)-target gene regulatory network in order to investigate the mechanism underlying choriocarcinoma and to verify the network through the overexpression or silencing of hub miRNAs in vitro.

Materials and Methods

A mRNA expression dataset and two miRNA expression datasets were analysed to identify differentially expressed genes (DEGs) and miRNAs (DEMs) between normal cells and choriocarcinoma cells. The top 400 upregulated and downregulated DEGs were identified as candidate DEGs, which were then mapped to construct protein–protein interaction (PPI) networks and select hub genes. Moreover, the DGIdb database was utilized to select candidate drugs for hub genes. Moreover, DEM target genes were predicted through the miRWalk2.0 database and overlaid with candidate DEGs to identify the differentially expressed target genes (DETGs). Furthermore, we established miRNA-TF-target gene regulatory networks and performed functional enrichment analysis of hub DEMs. Finally, we transfected mimics or inhibitors of hub DEMs into choriocarcinoma cells and assessed cell proliferation and migration to verify the vital role of hub DEMs in choriocarcinoma.

Results

A total of 140 DEMs and 400 candidate DEGs were screened from choriocarcinoma cells and normal cells. A PPI network of 400 candidate DEGs was established. Twenty-nine hub genes and 99 associated small molecules were identified to provide potential target drugs for choriocarcinoma treatment. We obtained 70 DETGs of DEMs derived from the intersection between predicted miRNA target genes and candidate DEGs. Subsequently, 3 hub DEMs were selected, and miRNA-TF-target gene regulatory networks containing 4 TFs, 3 TFs and 3 TFs for each network were constructed. The RT-PCR results confirmed that miR-29b-3p was highly expressed and that miR-519c-3p and miR-520a-5p were expressed at low levels in choriocarcinoma cells. The overexpression or silencing results suggested that 3 dysregulated hub DEMs jointly accelerated the proliferation and migration of choriocarcinoma.

Conclusion

Association of miRNA-TF-target gene regulatory networks may help us explore the underlying mechanism and provide potential targets for the diagnosis and treatment of choriocarcinoma.

Curcumin suppresses LGR5(+) colorectal cancer stem cells by inducing autophagy and via repressing TFAP2A-mediated ECM pathway

Abstract

Colorectal cancer stem cells (CSCs) have the potential for self-renewal, proliferation, and differentiation. And LGR5 is a stem cell marker gene of colorectal cancer. Curcumin can suppress oncogenicity of many cancer cells, yet the effect and mechanism of curcumin in LGR5(+) colorectal cancer stem cells (CSCs) have not been studied. In this study, we studied the effect of curcumin on LGR5(+) colorectal CSCs using the experiments of tumorsphere formation, cell viability and cell apoptosis. Then autophagy analysis, RNA-Seq, and real-time PCR were used to identify the mechanism responsible for the inhibition of LGR5(+) colorectal CSCs. Our results showed that curcumin inhibited tumorsphere formation, decreased cell viability in a dose-dependent manner, and also promoted apoptosis of LGR5(+) colorectal CSCs. Next, we found curcumin induced autophagy of LGR5(+) colorectal CSCs. When LGR5(+) colorectal CSCs were co-treated with curcumin and the autophagy inhibitor (hydroxychloroquine), curcumin-induced cell proliferation inhibition decreased. In addition, we also found that curcumin inhibited the extracellular matrix (ECM)-receptor interaction pathway via the downregulation of the following genes: GP1BB, COL9A3, COMP, AGRN, ITGB4, LAMA5, COL2A1, ITGB6, ITGA1, and TNC. Further, these genes were transcriptionally regulated by TFAP2A, and the high expression of TFAP2A was associated with poor prognosis in colorectal cancer. In conclusion, curcumin suppressed LGR5(+) colorectal CSCs, potentially by inducing autophagy and repressing the oncogenic TFAP2A-mediated ECM pathway.

Graphic abstract

Signaling pathways of genetic variants and miRNAs in the pathogenesis of myasthenia gravis

Background

Myasthenia gravis (MG) is a chronic autoimmune neuromuscular disorder causing muscle weakness and characterized by a defect in synaptic transmission at the neuromuscular junction. The pathogenesis of this disease remains unclear. We aimed to predict the key signaling pathways of genetic variants and miRNAs in the pathogenesis of MG, and identify the key genes among them.

Methods

We searched published information regarding associated single nucleotide polymorphisms (SNPs) and differentially-expressed miRNAs in MG cases. We search of SNPs and miRNAs in literature databases about MG, then we used bioinformatic tools to predict target genes of miRNAs. Moreover, functional enrichment analysis for key genes was carried out utilizing the Cytoscape-plugin, known as ClueGO. These key genes were mapped to STRING database to construct a protein-protein interaction (PPI) network. Then a miRNA-target gene regulatory network was established to screen key genes.

Results

Five genes containing SNPs associated with MG risk were involved in the inflammatory bowel disease (IBD) signaling pathway, and FoxP3 was the key gene. MAPK1, SMAD4, SMAD2 and BCL2 were predicted to be targeted by the 18 miRNAs and to act as the key genes in adherens, junctions, apoptosis, or cancer-related pathways respectively. These five key genes containing SNPs or targeted by miRNAs were found to be involved in negative regulation of T cell differentiation.

Conclusions

We speculate that SNPs cause the genes to be defective or the miRNAs to downregulate the factors that subsequently negatively regulate regulatory T cells and trigger the onset of MG.