Analyzing the interactions of circRNAs, miRNAs, and mRNAs to predict ceRNA networks in human acute type A aortic dissection

Background: Acute type A aortic dissection(ATAAD) is a life-threatening vascular disease. However, its underlying mechanism is still not well understood. Here, circular RNAs(circRNAs) were shown to function as competitive endogenous RNAs (ceRNAs) to regulate the effect of microRNAs(miRNAs) on their target genes may play a critical role in ATAAD. However, comprehensive identification and integrated analysis of the circRNA-miRNA-mRNA network in ATAAD have not been performed. Results: The gene expression profile of circRNAs, miRNAs, and mRNAs was performed between 6 ATAAD patients and 6 age-matched normal ascending aortic wall tissues patients were analyzed using the Arraystar human RNAs microarray. We identified that the expression of 12576 circRNAs,1603 miRNAs, and 14596 mRNAs were found to be differentially expressed(DE). Gene ontology(GO) and Kyoto Encyclopedia of Genes and Genomes pathway analyses(KEGG) were performed on these DE mRNAs and miRNA-mediated target genes of circRNAs. Furthermore, we used a multi-step computational framework and several bioinformatics methods to construct a ceRNA network containing circRNAs, miRNAs, and mRNAs based on co-expression analysis between the DE genes. The constructed ceRNA regulatory network containing 25 circRNAs, 17 miRNAs and 72 mRNAs. In the whole ceRNA network. We identified that plenty of key genes, such as hsa_circRNA_404522, hsa_circRNA_0022920, hsa_circ_0075881, hsa-miRNA1285-3p, hsa-miRNA-1285-3p, hsa-miRNA-637, hsa-miRNA-650, TINAGL1, JPH4, PLXNA2, TGFBR1, and THSD4. Furthermore, we also integrated the circRNA-miRNA-mRNA regulatory modules of the key genes. Conclusions: This study found a profile of dysregulated circRNAs, miRNA and mRNAs, and competitive circRNA-miRNA–mRNA regulatory networks were comprehensively integrated and predicted to be involved in ATAAD by GO and KEGG pathway analysis. It might be prospective clinical markers associated with ATAAD, and it is worthwhile to perform further studies to reveal the underlying link between these key genes and the molecular mechanisms of AD. KEGG pathway analysis was conducted to predict the molecular interactions and reaction networks associated with differentially expressed genes. Data were analyzed by two-sided Fisher’s exact test, and the FDR was calculated to correct the -log10 (p-value). A cutoff of -log10 ≤ 0.05 was set for statistical significance. RNA sequencing; qRT-PCR: real-time polymerase chain reaction; DE: differentially expressed; FDR: false discovery rate; FC: fold change; DAVID: Database for Annotation, Visualization and Integrated Discovery; BP: biological process; CC: cellular component; MF: molecular function; ARVC: arrhythmogenic right ventricular cardiomyopathy; HCM: Hypertrophic cardiomyopathy; TGFbeta: transforming growth factor-beta; ERK: extracellular regulated protein kinase; AT1R: angiotensin II type I receptor.

protein serine/threonine kinase, regulation of calcium-mediated signaling, and actin-myosin filament sliding. While upregulated DEGs were shown to be concerned with immune response to stress, defense response, cell cycle process, and inflammatory response. This conforms to the knowledge that regulation of actin-mediated cell contraction, transmembrane receptor protein serine/threonine kinase, cell cycle process, and inflammatory responses are main mechanisms of ATAAD development and progression [3,[41][42][43][44][45].
Through KEGG pathway analysis, mRNAs of ceRNA networks were predicted to be involved in the DNA replication, dilated cardiomyopathy, arrhythmogenic right ventricular cardiomyopathy (ARVC), Hypertrophic cardiomyopathy (HCM), adrenergic signaling in cardiomyocytes, cell cycle, vascular smooth muscle contraction, regulation of actin cytoskeleton, and TGF-beta signaling pathway. It is well-known that the vascular smooth muscle contraction, regulation of actin cytoskeleton, and TGFbeta signaling pathway were confirmed to associated with proliferation, migration and apoptosis of vascular smooth muscle cells in aortic dissection, and focal adhesion and regulation of actincytoskelet [45][46][47][48][49][50][51].
In the present study, we found dysregulated circRNAs, miRNAs and mRNA in the aortic wall of the ATTAD patients and predicted their ceRNA network. A total of 105 circRNA-miRNA-mRNA network was generated. According to the analysis of the degree of node connection, the top 5 host gene of the circRNAs were INADL, FIBP, LRRC16A, ADARB1, and LOC100506499, it is may be the key circRNAs during disease development and progression. FIBP may play an important role in proteolysis, inflammation, and apoptotic processes in the disease of abdominal aortic aneurysms [52]. In addition, the top 5 miRNA was hsa-miR-1285-3p, hsa-miR-637, hsa-miR-650,hsa-miR-485-5p, and hsa-miR-509-5p, the top 5 target genes were TINAGL1, JPH4, PLXNA2, TGFBR1, and THSD4. As well as we known, only a small portion of the DE gene in the ceRNA network have been annotated. For example, inhibition of extracellular regulated protein kinase(ERK) phosphorylation or blockade of angiotensin II type I receptor (AT1R) prevented aneurysmal degeneration of TGFBR1 deficient aorta, loss of SMC-TGFBR1 triggers multiple deleterious pathways, including abnormal TGFBR2, ERK, and AngII/AT1R signals that disrupt aortic wall homeostasis to cause aortic aneurysm formation [53]. CircRNA_010567 could regulate miR-141 and its target TGF-β1 expression and may mediate fibrosis-associated protein resection [36]. TINAGL1, JPH4, PLXNA2, and TGFBR1 is the key mRNAs in our study, it could play an important role in the AD, the separate ceRNA network is shown in Fig. 10. It is worthwhile to perform further studies to reveal the underlying link between these key genes and the molecular mechanisms of AD.
The limitation of this study is that we detected circRNAs, circRNAs and miRNA changes, only selected eight DE circRNAs for qRT-PCR experiments. What excites us is that the results of qRT-PCR experiments were consistent with the normalized expression of RNA-seq. However, we did not validate the miRNA and mRNA, and further research is urgently needed. Maybe these circRNAs can interact with miRNAs and alter the expression of miRNAs and/or its target genes may have great significance in the pathogenesis research for AD and may point out a potential direction of future treatment as well.

Conclusions
We found a profile of dysregulated circRNAs, miRNAs and mRNAs, and competitive circRNA-miRNA-mRNA regulatory networks were comprehensively integrated and predicted to be involved in ATAAD by GO and KEGG pathway analysis. It might be prospective clinical markers associated with the development of ATAAD, and it is worthwhile to perform further studies to reveal the underlying link between these key genes and the molecular mechanisms of AD. Tables Table 1 The primers of GAPDH 5 and circRNAs.   Figure 1 Flow-chart of data analysis.  The DE circRNAs type. a: histogram; b: Pie chart.

Figure 4
The top10 Enrichment score of the BP, CC, and MF terms, and KEGG pathway analysis. a: Up-regulated mRNA; b: Down-regulated mRNA.

Figure 6
The top10 Enrichment score of the host genes of circRNA of the BP terms and KEGG pathway(Red, GO BP; yellow, KEGG pathway; X-length indicates the number of enriched genes, and black polyline indicates the p-value).  qRT-PCR validation of dysregulated DE circRNAs in ATAAD compared with age-matched patient with normal ascending aortic wall tissues * represents P < 0.05; ** represents P < 0.01).