Abstract

Background. Keloid has brought great trouble to people and currently has no uniformly successful treatment. It is urgent to find new targets to effectively prevent the progress of keloid. The current research mainly identifies the differentially expressed genes (DEGs) in keloid through high-throughput sequencing technology and bioinformatics analysis technology, to screen new therapeutic targets and potential biomarkers. However, due to the different samples, different control groups, and small sample sizes, the sequencing results obtained from different studies are quite different and lack reliability. It is necessary to analyze the existing datasets in a reasonable way. Methods. Datasets about keloid were filtered in Gene Expression Omnibus (GEO) and ArrayExpress databases according to the inclusion and exclusion criteria. The discovery datasets were used for summarizing significant DEGs, and the validation datasets were to validate the mRNA and miRNA expression levels. The Encyclopedia of RNA Interactomes (ENCORI) online platform was used to predict the interactions between miRNAs and their target mRNAs. Protein-protein interaction network (PPI network) analysis and functional enrichment analysis were conducted. miRNA-mRNA network was established by Cytoscape software and verified in keloid tissue () by RT-qPCR. miR-29a-3p mimic and inhibitor were transfected into keloid fibroblasts (KFs) to preliminary verify its targets, the prognostic value of which was estimated by the receiver operating characteristic (ROC) curve. Results. A total of 6 datasets involving 20 patients were included. 15 miRNAs and 12 target mRNAs were identified as potential biomarkers for keloid patients. The RT-qPCR results showed that miR-29a-3p, miR-92a-3p, and miR-143-3p were downregulated, and all their target mRNAs were upregulated in keloid tissue (). The expression of COL1A1, COL1A2, COL3A1, COL5A1, and COL5A2 decreased when miR-29a-3p was overexpressed but increased when miR-29a-3p w