Comprehensive Analysis of Circular RNA Expression in ceRNA Networks and Identification of the Effects of hsa_circ_0006867 in Keloid Dermal Fibroblasts

Abstract

Circular RNAs (circRNAs) play a crucial role in the pathogenesis of various fibrotic diseases, but the potential biological function and expression profile of circRNAs in keloids remain unknown. Herein, microarray technology was applied to detect circRNA expression in four patient-derived keloid dermal fibroblasts (KDFs) and normal dermal fibroblasts (NDFs). A total of 327 differentially expressed (DE) circRNAs (fold change > 1.5, p < 0.05) were identified with 195 upregulated and 132 downregulated circRNAs. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses showed that the upregulated circRNAs were mainly enriched in the cytoskeleton and tight junctions, while the downregulated circRNAs were related to morphogenesis of the epithelium and axonal guidance. To explore the function of DE circRNAs, a circRNA-miRNA-mRNA network, including five circRNAs, nine miRNAs, and 235 correlated mRNAs, was constructed using bioinformatics analyses. The expression of five DE circRNAs was validated by qRT-PCR in 18 pairs of KDFs and NDFs, and hsa_circ_0006867 showed promising regulatory function in keloids in vitro. Silencing hsa_circ_00006867 suppressed the proliferation, migration, and invasion of keloid fibroblasts. RNA-binding protein immunoprecipitation (RIP) assays indicated that hsa_circ_00006867 may serve as a platform for miRNA binding to Argonaute (AGO) 2. In addition, hsa-miR-29a-5p may be a potential target miRNA of hsa_circ_00006867. Taken together, our research provided multiple novel clues to understand the pathophysiologic mechanism of keloids and identified hsa_circ_0006867 as a biomarker of keloids.

Keywords: ceRNA network; circular RNA; dermal fibroblast; keloid; miRNA sponge; microarray analysis.

FIGURE 1

Detailed flowchart of the entire study. First, microarray analysis was performed to explore DE circRNAs in KDFs and NDFs. Then, GO and KEGG pathway analyses showed the potential functions of the DE circRNAs. six circRNAs were then verified by qRT–PCR, and a circRNA-miRNA-mRNA network was established for the validated circRNAs. Finally, we explored the cellular function and potential mechanism of hsa_circ_0006867.

FIGURE 2

Expression profiles and characteristics of circRNAs in KDF and NDF samples as examined by microarray. (A) The box plot shows the distributions of the expression values after quantile normalization. (B) The scatter plot shows the difference in circRNA expression between KDFs and normal controls. The red dots above the green line and blue dots below the bottom indicate a more than 1.5-fold change (FC) in circRNAs between the two compared groups. (C) Volcano plot of differentially expressed circRNAs (fold change > 1.5 and p < 0.05). The red and blue dots represent upregulated and downregulated circRNAs, respectively. (D) The circRNA classification statistics show the percentage of each type among all differentially expressed circRNAs. (E) The Circos diagram depicts the location of all differentially expressed circRNAs on chromosomes. The orange/red and green lines represent upregulated and downregulated circRNAs, respectively. The segment length indicates the FC value.

FIGURE 3

Heatmap of circRNA profiles from the microarray data showing distinguishable circRNA expression patterns. Red and blue indicate high and low expression, respectively, and the color scales represent expression values. The columns represent samples, and the rows represent circRNAs. The type of circRNA and p values of each analysis are shown in the first and second columns, respectively.

FIGURE 4

The top 10 GO and KEGG pathway enrichment terms for the predicted target genes of DE circRNAs. GO analysis covered three domains, namely biological processes, cellular components, and molecular functions. The KEGG plot shows the top enrichment score value of the significantly enriched pathway. (A, B) The upregulated - and (C, D) downregulated circRNAs were analyzed. Abbreviations: ORF, open reading frame.

FIGURE 5

Characteristics and verification of six candidate circRNAs. (A) Structural patterns and (B) qRT–PCR results of upregulated circRNAs (left), including hsa_circ_0001055, hsa_circ_0006867, and hsa_circ_0073244, as well as three downregulated circRNAs, including hsa_circ_0004662, hsa_circ_0000479, and hsa_circ_0029998.

FIGURE 6

Establishment of the circRNA-miRNA-mRNA interaction network. The network contained five dysregulated circRNAs, nine miRNAs, and 235 correlated mRNAs. Edges with T-shaped arrows represent directed relationships. Nodes with purple, yellow, and pink color represent circRNAs, miRNAs, and mRNAs, respectively.

FIGURE 7

Inhibition of hsa_circ_0006867 suppresses keloid cell proliferation, migration, and invasion. (A) Effect of siRNAs targeting hsa_circ_0006867. (B) CCK-8 assays. (C) Scratch wound-healing assays. (D) Transwell migration and invasion assays. *p < 0.05, **p < 0.01, and ***p < 0.001.

FIGURE 8

Exploration of the potential regulatory mechanism of hsa_circ_0006867 in keloids. (A) Nuclear and cytoplasmic fractionation assays verified the localization of hsa_circ_0006867 in KDF1 and KDF2 cells. (B) The five highest ranking target miRNAs of hsa_circ_0006867 and annotation of their circRNA/miRNA interactions. (C) The dual-luciferase reporter assay was performed using 293T and KDF1 cells after cotransfection. (D) The expression of miR-29a-5p was significantly upregulated after silencing hsa_circ_0006867. (E) The RIP assay showed that hsa_circ_0006867 and miR-29a-5p were enriched in the presence of AGO2 antibody. (F) The Venn plot showed the overlap of the predicted mRNA target of miR-29a-5p based on the datasets from miRDB, TargetScan, and miRWalk 3.0.