Identifying circRNA-associated-ceRNA networks in juvenile spondyloarthropathies patients

The NCBI GEO is a public functional genomics data repository that contains high-throughput gene expression data, chips, and microarrays [10]. We conducted a search using the keywords “juvenile idiopathic arthritis” in the GEO Database, resulting in 805 records. After narrowing down the search to include only “Homo sapiens,“ the number of records reduced to 801. The objective of this study is to investigate circRNA-associated-ceRNA networks in patients with juvenile spondyloarthropathies. To obtain the mRNA dataset, we specifically selected the study type “Expression profiling by array,“ which yielded 35 remaining records. We carefully examined these 35 links and identified GSE58667 (11 JSpA patients and 4 controls) [11] as the gene expression dataset for juvenile spondyloarthritis, making it our selection. For the miRNA dataset, we utilized the study type “Non-coding RNA profiling by array,“ resulting in four remaining records. Among these, GSE79481 (8 enthesitis-related arthritis patients and 8 controls) [12] was chosen as it provided global miRNA profiling in patients with enthesitis-related arthritis. Regarding the circRNA dataset, we performed a search using the terms “spondylitis” and “circRNA,“ along with the filters “Homo sapiens” and “Non-coding RNA profiling by array,“ resulting in two remaining links. GSE178408 (6 ankylosing spondylitis patients and 3 controls) [7] was selected as it contained circRNA expression profiles from ankylosing spondylitis patients and healthy controls. The children included in the GSE58667 dataset were diagnosed with JSpA based on the criteria set forth by the International League of Associations for Rheumatology (ILAR). The patients with ERA in the GSE79481 dataset met the ILAR criteria for diagnosis. Lastly, the patients in the GSE178408 dataset were diagnosed with ankylosing spondylitis (AS) according to the modified New York criteria established in 1984.

We utilized the GEO2R web-based tool [13] to detect DEGs in the samples from diseased individuals and control subjects. The DEcircRNAs were identified based on the criteria of |log2(fold-change)| > 2.0 and P-value < 0.05. Similarly, significant DEmiRNAs and DEmRNAs were recognized using the standards of |log2(fold-change)| > 1.0 and P-value < 0.05.

To predict the target miRNAs of DEcircRNAs, we referred to the Circbank database [14]. The target mRNAs of miRNAs were retrieved from the miRDB databases [15]. To ensure the credibility of the findings, we identified the final miRNAs by selecting overlapping miRNAs in the circbank database and DEmiRNAs from GSE79481. Similarly, the final mRNAs were obtained by selecting overlapping mRNAs in the miRDB database and DEmRNAs from GSE58667. Further investigation was carried out solely on the final miRNA and mRNA targets.

We established the circRNA-miRNA-mRNA regulatory network by integrating the potential circRNA-miRNA pairs and miRNA-mRNA pairs. The Cytoscape 3.8.2 software [16] was employed to visualize the ceRNA network.

DAVID is an internet-based biological information database utilized for annotation, visualization, and integrated discovery [17]. KEGG is an integrated database resource for biological interpretation of genome sequences and other high-throughput data. Meanwhile, Gene Ontology (GO) is a fundamental bioinformatics tool utilized to annotate genes and scrutinize their biological processes [18].

Using the online search tool, Search Tool for the Retrieval of Interacting Genes (STRING), a protein-protein interaction (PPI) network of the differentially expressed mRNAs (DEmRNAs) was constructed. The PPI network was then presented visually through the utilization of Cytoscape (version 3.8.2). Furthermore, the identification of hub genes was performed through the use of the Molecular Complex Detection (MCODE) application in Cytoscape.

In order to evaluate alterations in immune cell expression and to obtain the distribution of different immune cell types, we applied the CIBERSORTx algorithm online [19]. To conduct this analysis, we downloaded GSE58667 series matrix files in txt format from the NCBI GEO website. We compared immune cell differences between healthy control subjects and JSpA patients.

To explore potential therapeutic agents for JSpA, we employed the CMap tool [20]. Specifically, we utilized the online L1000 platform of the Connectivity Map tool to query the 299 identified DEmRNAs. By submitting a list of 167 upregulated and 132 downregulated hub genes, we obtained connectivity scores, which reflect the proximity of the gene expression profiles on a scale from − 1 to 1. A positive score indicates a promoted effect, while a negative score suggests an inhibited effect.

We performed microarray analysis to determine the expression levels of circRNA, miRNA, and mRNA, as summarized in Table 1. Using a pre-set threshold, we identified 225 DEcircRNAs in the GSE178408 dataset, consisting of 45 upregulated and 180 downregulated circRNAs, 23 DEmiRNAs in the GSE79481 dataset, consisting of 8 upregulated and 15 downregulated miRNAs, and 1324 DEmRNAs in the GSE58667 dataset, consisting of 621 upregulated and 703 downregulated mRNAs. To further analyze the relationship between circRNAs, miRNAs, and mRNAs, we utilized the Circbank and miRDB databases to predict potential interactions. After taking the intersection of the results, we identified 7 DEmiRNAs (1 upregulated and 6 downregulated) and 5 circRNAs that were co-expressed. Finally, we identified 299 co-expressed mRNAs (167 upregulated and 132 downregulated) through the intersection of the 7 miRNAs and 4221 mRNAs obtained from miRDB databases. A flowchart of our analysis is presented in Fig. 1, while the volcano plots of our results are shown in Fig. 2.

In our analysis, we identified 10 interactions between circRNA and miRNA, which consisted of 5 circRNAs (hsa_circ_0006387, hsa_circ_0005070, hsa_circ_0001517, hsa_circ_0005303, and hsa_circ_0102537) and 7 miRNAs (hsa-miR-1225-3p, hsa-miR-1825, hsa-miR-193a-5p, hsa-miR-300, hsa-miR-485-3p, hsa-miR-548c-5p, and hsa-miR-628-3p). We merged the predicted mRNAs of these 7 miRNAs from miRDB with the DEmRNAs retrieved from the GEO database, resulting in a total of 299 target mRNAs. Finally, we integrated the 5 circRNAs, 7 miRNAs, and 299 target mRNAs into a circRNA-miRNA-mRNA network for further investigation, as depicted in Fig. 3. The 5 overlapped circRNAs are summarized in Table 2.

Functional and pathway enrichment analyses were conducted using DAVID to obtain a better understanding of the biological roles of the 299 DEmRNAs. Figure 4 displays the enriched KEGG pathways and GO terms. KEGG pathway analysis revealed that the DEGs were mainly associated with the JAK-STAT signaling pathway. Additionally, GO biological process analysis indicated that these DEmRNAs were significantly involved in the regulation of transcription from RNA polymerase II promoter, regulation of transcription and DNA-templated, signal transduction, and positive regulation of cell proliferation. Changes in cellular component were mainly related to the nucleus, cytosol, and cytoplasm. Changes in molecular function were primarily associated with protein binding.

The PPI network analysis of the 299 DEmRNAs resulted in a complex network consisting of 299 nodes and 307 edges, as depicted in Fig. 5A. To identify the most crucial nodes in the network, we employed Cytoscape MCODE, which revealed 10 hub genes. The partial genes are shown in Fig. 5B.

We can conclude that neutrophils accounted for the majority of all enriched cells, followed by monocytes, NK cells and T cells. Low percentage of B cell and plasma cell infiltration. While, no differential expression proportion of immune- enriched cells in the JSpA and normal groups is shown in Fig. 6.

It is important to note that while these compounds (Table 3) may show potential in targeting the DEmRNAs associated with JSpA, further research is needed to determine their safety and efficacy in treating the disease. In addition, it is necessary to conduct extensive pre-clinical and clinical studies before these compounds can be used as a treatment option for JSpA patients.