Two of the main types of ncRNAs, miRNAs and lncRNAs, can suppress each other as ceRNAs and form a regulatory ceRNA network (lncRNA–miRNA–mRNA) to regulate target mRNAs of miRNAs. Comprehensive analysis concerning the ceRNA network has been established in ankylosing spondylitis [25], periodontitis [26, 27], rheumatoid arthritis [28], cervical cancer [29], and cutaneous melanoma [30]. Therefore, it is important to investigate the functional roles and regulatory mechanism of lncRNA-associated ceRNAs in the osteoblast differentiation of PDLSCs and explore their potential mechanism and treatment targets in periodontitis. In our study, the analysed data were obtained from our previous studies [19, 20] concerning lncRNA, mRNA and miRNA expression. Therefore, according to the ceRNA theory, the regulatory networks of the same miRNA could competitively bind with lncRNAs and mRNAs in one triplet. The successfully constructed network in our study indicated that lncRNA–miRNA–mRNA mechanisms might play a crucial role in the osteoblast differentiation of PDLSCs.
Among the 6 lncRNAs with upregulated expression in the ceRNA network, 4 were not reported (BC048201, LOC100302640, LOC728190, LINC00340), among which LOC100302640 had the highest expression level in the microarray data, and it was predicted that it may be related to the MAPK and TGF-beta pathways. In addition, LOC541471 was identified as an overall survivor-related lncRNA in glioblastoma multiforme [31]. Another study reported that LOC541471 is a novel prognostic biomarker for head and neck squamous cell carcinoma [32]. The leukaemia 2 gene (DLEU2) has been reported to play an important role in a variety of diseases [33, 34], especially in tumour diseases, including thyroid cancer, gastric cancer, sarcopenia, and laryngeal squamous cell carcinoma, and has been shown to play a role through the ceRNA mechanism [35–38].
In addition to these lncRNAs revealed in these paper, some classical lncRNAs, such as SNHG1, TUG1, GAS5, XIST, DANCR and FER1L4, were reported to be involved in the osteoblast differentiation of PDLSCs [17, 39–43], most of them function as ceRNAs (miRNA sponges) to regulate osteogenesis. In order to reveal a lncRNA-miRNA-mRNA network, a complete study includes procedures as follows. From RNA-sequencing data, we find out the differentially expressed genes at first. Secondly, based on the reliable database that involved in the field of interest, the enriched mRNAs were screened out by GO analysis. With predicted softwares, such as miRanda, we obtain related lncRNA and miRNA. Together with the mRNA, they compose the network. Furthermore, qRT-PCR, lentivirus transfection, rescue, pull-down, and luciferase assays are needed to be carried out to verify the network.
GO and pathway analyses were carried out to explore biological functions enriched among mRNAs with upregulated and downregulated expression. GO analysis was used as a controlled repertoire to investigate the function of mRNAs with upregulated and downregulated expression and describe mRNA and mRNA products that are distributed in any organism. Pathway analysis was conducted to identify genes with upregulated and downregulated expression based on KEGG analysis. In our research, the results of GO analysis showed that from 678 GO enriched terms were significant with P-value < 0.05 in the mRNAs with upregulated expression. These significant GO terms involved the transforming growth factor beta receptor signalling pathway, regulation of Wnt receptor signalling pathway, positive regulation of Notch signalling pathway, enzyme-linked receptor protein signalling pathway. In the mRNAs with downregulated expression, we found 714 GO enriched terms that were significant with a P-value < 0.05. These significant GO terms involved the BMP signalling pathway, Toll signalling pathway, mesenchymal cell differentiation, and regulation of cell differentiation. The most significant pathways, such as the TGF-beta signalling pathway, MAPK signalling pathway, p53 signalling pathway, and FoxO signalling pathway, are related to osteoblast differentiation. We further predicted the lncRNAs involved in the GO terms 0001649 (osteoblast differentiation), MAPK and TGF-β and found that LOC100302640, which had upregulated expression in the ceRNA network, was involved, suggesting that this new LOC100302640 may participate in the regulation of PDLSC osteogenic differentiation through a ceRNA mechanism. The function of LOC100302640 requires profound research, and the specific mechanism needs to be further explored.