Integrated Analysis of lncRNA-miRNA-mRNA ceRNA Network in Human Diarrhea Irritable Bowel Syndrome


 Background:Many studies on long chain non-coding RNAs (lncRNAs) are published in recent years. But the roles of lncRNAs in diarrhea irritable bowel syndrome (IBS-D) are still unclear and should be further examined. The present work focused on determining the molecular mechanisms underlying lncRNAs regulation in IBS-D on the basis of the lncRNA-miRNA-mRNA competing endogenous RNA (ceRNA) network.Methods:This study collected the mRNAs (GSE36701) expression data within human tissue samples with IBS-D group and normal group based on Gene Expression Omnibus (GEO) database and collected the differentially expressed lncRNAs (DELs) and differentially expressed miRNAs (DEmiRs) based on PubMed.Functional enrichment analysis of DEGs was performed on the DAVID database. Then the interaction network was constructed and visualized using STRING database and Cytoscape.Results: This study identified 3192 DEmRNAs (1437 with up-regulation and 1755 with down-regulation),29 DEmiRs (18 upregulated and 11 downregulated)and 2 DELs(one upregulated and one downregulated) between IBS-D and control samples.Furthermore,we constructed a lncRNA-miRNA-mRNA network through two DELs (lncRNA TUG1 with up-regulation and lncRNA H19 with down-regulation), four DemiRs (hsa-miR-148a-3p,hsa-miR-342-3p,hsa-miR-149-5p with up-regulation and hsa-miR-219a-5p with down-regulation)and 24 DEGs (4 with up-regulation and 20 with down-regulation) with 42 axes. Simultaneously, we conducted functional enrichment and pathway analyses on genes within the as-constructed ceRNA network. According to our PPI/ceRNA network and functional enrichment analysis results, two critical genes were found (BCL2L11 and QKI). Conclusion:In conclusion, the ceRNA interaction axis we identified is a potentially critical target for treating IBS-D.BCL2L11 axis(LncH19-hsa-miR-148a-3p-BCL2L11) may via interaction with PI3K/AKT pathways in IBS-D.Our results shed more lights on the possible pathogenic mechanism in IBS-D using a lncRNA-associated ceRNA network.


Introduction
Irritable bowel syndrome (IBS) is a common chronic functional bowel disease, characterized by no organic lesions, but recurrent abdominal pain, abdominal distension, and (or) changes in stool characteristics [1] .It is the most commonly diagnosed gastroenterology disorder which affects over 10% of the population globally and the prevalence varies regionally [2] ,signi cantly affect quality of life.IBS patients can be categorized into four distinct subtypes according to the ROME criteria which are based on predominant stool pattern: IBS with diarrhoea (IBS-D), constipation (IBS-C), mixed (IBS-M) and an unde ned pattern of abnormal stool (IBS-U) [3] .As the most common clinical subtype of IBS,diarrheapredominant irritable bowel syndrome (IBS-D) is accompanied by mental and emotional abnormalities that affect the living quality of patients, except the shared features [4] .The pathophysiology of IBS has not been clearly understood yet, but everal factors have been involved such as alterations in gastrointestinal motility , visceral hypersensitivity , small intestinal bacterial overgrowth (SIBO), environmental factors,low-grade in ammatory [5][6] .The current treatment options for IBS, which predominantly target the individual symptoms, are limited and make adequate treatment of global IBS symptoms a signi cant challenge.
For decades, research has focused on the 2% of the human genome that codes for proteins [7][8] . In recent years,researchers have found that the remaining 98% of the genome that was once considered as nonfunctional "junk" includes noncoding RNAs (ncRNAs) that play important roles in a wide range of biological processes such as growth,development, and organ function. Non-coding RNAs have been classi ed as small non-coding RNAs, including microRNAs,and long non-coding RNAs (lncRNAs) [9] .CHAO et al. [10] found that the down-regulation of LncRNA H19 resulted in the expression changes of AQP1 and AQP3 may play an important role in the occurrence and development of IBS-D.ZHAO et al. [11] con rmed lncRNA TUG1 attenuated TNF-α-caused apoptosis and in ammatory response in ICC by down-regulating miR-127 and then inactivating NF-κB and Notch pathways in patients with IBS-D.Meanwhile,It has been proved that microRNAs (miRs) have been veri ed to be related with the progression and development of IBS-D [11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28] .Such as hsa-miR-150 and hsa-miR-342-3p link to pain and in ammatory pathways both of which are thought to be dysregulated in IBS [12] ;The cxolonic mucosal microRNAs, microRNA-219a-5p and microRNA-338-3p are downregulated in Irritable Bowel Syndrome and are associated with barrier function and MAPK signaling [13] ;Serum exosomes derived from Irritable Bowel Syndrome patient increase cell permeability via regulating miR-148b-5p/RGS2 signaling in human colonic epithelium cells [14] .To sum up, lncRNA-miRNA-mRNA regulatory network exerts a vital part in IBS-D genesis and progression.At the same time, there are few reports about the ceRNA regulation mechanism of lncRNA-miRNA related to IBS-D and the interaction between ncRNAs.
In this study, lncRNAs together with the corresponding mechanisms of action within human tissue from IBS-D cases were explored by means of bioinformatic analysis.Firstly, this study applied Gene Expression Omnibus (GEO) database and PubMed for obtaining the IBS-D-related differentially expressed genes (DEGs), differentially expressed miRNAs (DEmiRs), and differentially expressed lncRNAs (DELs). Cytoscape 3.7.2 was utilized to construct a lncRNA-miRNA-mRNA network, followed by the construction of a PPI network.
The present study aims to further screen the key lncRNA-miRNA-mRNA ceRNA axis within IBS-D with data collected from public databases and bioinformatics methods. Besides, our results shed more lights on the IBS-D molecular mechanism, thus providing a novel direction for targeted therapy of IBS-D. To further examine the key gene regulatory axis, we established a lncRNA-miRNA-mRNA regulatory subnetwork.The data ow mechanism diagram is displayed in (Figure1).

Page 4/29
According to the cutoff criteria of GSE36701 (log|FoldChange|>0.176, P.Value<0.05), volcano map and heat map for differential expressed mRNAs were obtained(Figure2-3). In which 3192 DEmRNAs (1437 with up-regulation and 1755 with down-regulation) were discovered through the comparison of gene expression data between IBS-D and control samples.
Compared with lncRNA prediction,mRNAs combined with the above 5miRNAs were predicted by miRTarBase, TargetScan, StarBase and miRDB databases, and 214 mRNAs were obtained by intersection of the above 4 databases(Figure4B).We acquired the common mRNAs through the intersection of 214mRNAs with 3192 DemiRs of GSE36702,a total of 29 mRNAs were found(Figure4C).
lncRNA-miRNA-mRNA ceRNA network construction According to the negative regulatory relationship between miRNAs and lncRNAs,miRNAs and mRNAs,lncRNA-miRNA-mRNA ceRNA network was constructed.Cytoscape (version 3.  connections with other nodes, so they are of great signi cance in PPI network,we predicted QKI and BCL2L11 were key genes.

Functional enrichment and differential expression analysis
In the present study, the 24 DEGs of ceRNA network used to analyze GO/KEGG were imported into DAVID website. A total of 7 GO enrichment items were obtained, including four BP entries,one CC entry and two MF entries (Figure7A,7B,Table5). Four signal pathways were obtained by enrichment analysis of KEGG pathways (Figure7C,7D,Table6).In order to show a smaller subset of high-dimensional data, use sangerbox software to draw the chord diagram of GO and KEGG,as shown in gure 8.To accurately explore the differential expression of key genes in the dataset, two key genes were visualized in( Figure  9).In the GSE36701 dataset, QKI and BCL2L11 were all down-regulated, further revealing the underlying mechanisms for key genes.  According to the negative regulatory relationship between miRNA and mRNA,miRNA and lncRNA, 9 nodes in the above PPI network were selected to construct the core lncRNA-miRNA-mRNA regulatory model diagram related to the progress of IBS-D, as shown in Figure 10.

Discussion
IBS is one of the most commonly diagnosed gastrointestinal (GI) conditions. Although plenty of proteincoding genes have been identi ed as IBS-D-related genes, these genes couldn't perfectly explain how IBS-D occurs and develops. Recently, a growing body of research has focused on the epigenetic regulation and the roles of ncRNAs in IBS-D pathogenesis. In this study, we constructed the ceRNA network based on the expression pro les of whole rectal mucosa of healthy controls and IBS-D patients.
Then we performed the functional analyses of GO enrichment and KEGG pathways based on the genes in the ceRNA network.A total of 7 GO enrichment items were obtained, including four BP entries,one CC entry and two MF entries.BP is enriched in negative regulation of apoptotic process,negative regulation of myoblast differentiation,positive regulation of gene expression,glucose transport;CC is enriched in cortical actin cytoskeleton;MF is enriched in protein binding,protein kinase activity. Four signal pathways were obtained by enrichment analysis of KEGG pathways,that are Pathways in cancer,Central carbon metabolism in cancer,Renal cell carcinoma,PI3K-Akt signaling pathway.
The Phosphatidylinositol 3-kinase (PI3K)/phosphorylated protein kinase B (AKT) pathway is an important pathway in the regulation of cell proliferation and apoptosis [29] . Akt is a serine/threonine kinase and an important target of PI3K.Fei and WANG's [30] study revealed that miR-495 upregulation can reduce visceral sensitivity in IBS-D via inhibition of PI3K/AKT signaling pathway by targeting PKIB.SONG et al. [31] found overexpression of miR-340 or the silencing of SPP1 inhibited GC cell proliferation, invasion, migration, and EMT process through inhibition of the PI3K/AKT signaling pathway, but promoted apoptosis of GC cells.
When uploading the 24 DEGs identi ed in ceRNA network to STRING website, 8 edges were found to be present in 9 DEGs incorporated in constructing a PPI network. QKI and BCL2L11 have high degree values and have more than 3 direct connections with other nodes, so they are of great signi cance in PPI network,we predicted QKI and BCL2L11 were key genes.
Quaking (QKI) is a member of the signal transduction and activator of RNA metabolism (STAR) and hnRNP Khomology-type family of RNA-binding proteins [32] .Substantial research implicated QKI RNAbinding protein function in many more cell types than initially anticipated.Like many mRNA regulators, quaking-related proteins regulate the expression of diverse mRNA targets by various mechanisms and have essential roles in cell cycle and differentiation regulation [33][34][35] .Among the four known splice variants of the human QKI gene (QKI-5, QKI-6, QKI-7, and QKI-7b) [32] , QKI-5 is the variant most abundantly expressed in normal human colon tissues, and preferentially down regulated in human colorectal cancer(CRCs) [36] .In this reseach,QKI was also down-regulated in IBS-D.QKI has been demonstrated in the regulation of cellular processes such as cell cycle, apoptosis, and differentiation [37] .CHU et al. [38] found QKI could promote the proliferation, metastasis, and invasion of pancreatic cancer through activating broblasts surrounding pancreatic cancer and accelerating EMT and increasing the autophagy in pancreatic cancer.Mukohyama Junko et al. [39] found that overexpression of QKI-5 suppressed organoidforming capacity and tumorigenic capacity of colorectal carcinoma PDX cells .In addition to this studies [40][41] found that QKI could inhibit the proliferation and tumorigenesis in multiple cancer types such as lung adenocarcinoma, and renal cell carcinoma.
BCL2L11 (also known as BIM) is a member of BCL-2 family, inducing apop-tosis and inhibiting autophagy by inactivating BCL2 or by activating BAX-BAK1 and by bridging BECN1 or DYNLL1, respectively [42] .According to previous reports, BCL2L11 is involved in biological pro-cesses in a variety of solid tumors such as ovarian cancer, endometrial adenocarcinoma, prostate tumor and gastric cancer [43][44][45] .PAN et al.' [46] data demonstrated that lncRNA ACTA2-AS1 could suppress colon adenocarcinoma progression via sponging miR-4428 to regulate BCL2L11 expression.ZHANG et al.' [47] studies from both in vitro and in vivo shown that miR-24 regulates BCL2L11 expression by directly binding with 3'UTR of mRNA, thus promoting gastric cancer cell growth, migration while inhibiting cell apoptosis.YUAN et al.' [48] results showed that downregulation of LncRNA H19 could promote cell proliferation, inhibit cell apoptosis, and suppress multiple in ammatory cytokine expressions in HK-2 cells by modulating the miR-130a/BCL2L11 pathway.As a disease of digestive system,IBS-D was speculated to be closely related to the above key genes QKI and BCL2L11 and PI3K/AKT signaling pathway.
Certain limitations should be noted in the present work. The way of action has been predicted based on the measured RNA network, however, which has not been con rmed (dual luciferase reporter gene analysis, gene overexpression or gene knockout). Although several related genes have been screened out in the present study, further in vitro clinical research and in vivo experiments should be carried out to con rm its expression and functional mechanism in terms of IBS-D.
Currently, little research is conducted to explore the lncRNA mechanism in IBS-D. The present work has its certain strengths. For instance,for the rst time construct the lncRNA-miRNA-mRNA network based on GEO database and PubMed. Nonetheless, our ndings were just obtained from bioinformatics analysis.

Conclusion
In summary, the ceRNA interaction axis we identi ed is a potentially critical target for treating IBS-D.We speculate that the BCL2L11 axis(LncH19-hsa-miR-148a-3p-BCL2L11) may via interaction with PI3K/AKT pathways in IBS-D.Our results shed more lights on the possible pathogenic mechanism in IBS-D using a lncRNA-associated ceRNA network.

Materials And Methods
Microarray data and data processing One dataset was obtained by setting the screening criteria for the species type as"Homo sapiens", from GEO database of National Center for Biotechnology Information (NCBI) (https://www.ncbi.nlm.nih.gov/geo/), with such keywords being searched as"Diarrhea irritable bowel syndrome"(IBS-D). Then select the study type as " Expression pro ling by array", so that we get the series matrix les. One dataset was included in this study, namely GSE36701. As for mRNA expression data , relative to normal rectal mucosa in control group, the experimental group collected diarrhoea predominant IBS subjects to identify differentially expressed genes. Simultaneously, we applied GEO database for downloading series matrix les and expressive data. DEGs,DEmiRs,DELs were classi ed as the up-regulation or down-regulation group.These data would be used in the following ceRNA network construction and protein interaction network construction.
Prediction of miRNA-mRNA and lncRNA-miRNA pairs Interaction relationships between lncRNAs and miRNAs were predicted with Starbase database.By using jveen(http://jvenn.toulouse.inra.fr/app/example.html) to output all the predicted results of DElncRNAs, and intersecting the output results with DEmiRs, the interaction pair of lncRNAs and miRNAs can be obtained. Similarly, we use datebases to predict the miRNAs in the lncRNA-miRNA interaction pair, and intersect the prediction result with the DEmRNAs. As a result,the miRNA-mRNA interaction pair can be obtained.The difference is that we use four databases when predicting mRNAs, we take the intersection of four databases as our prediction result.Interactions between lncRNAs and miRNAs were analyzed on the basis of lncRNA target prediction databases shown below: evaluated by Degree to screen hub-gene. The higher the Degree of a node is, the greater its signi cance in PPI network is.
Functional enrichment (DEGs of ceRNA network) and differential expression analysis (key genes) Gene ontology (GO) functional annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis were done for analyzing DEGs by using DAVID online tool (version 6.8)(https ://david -d.ncifc rf.gov/), P<0.05 indicated that the pathways or GO biological process terms were signi cantly enriched.The present work adopted sangerbox(http://www.sangerbox.com/Signin) to visualize key DEGs in GSE36701. Flow diagram of data processing. The difference expression of mRNAs of GSE36701 and lncRNAs,miRNAs of PubMed were analyzed, and then the intersection was selected; Construction of CeRNA network, protein-protein interaction network and functional enrichment analysis, and nally determine the key genes Heatmap of differentially expressed analysis of mRNAs of GSE36701.  The lncRNA-miRNA-mRNA ceRNA network.The diamond: lncRNA, the triangle: miRNA, and the ellipse: mRNA, Red: up-regulation, blue: down-regulation