OS is the most common malignant bone disease primarily localized to the long bones and is characterized by a high propensity to metastasize[25]. Baicalin has been widely used to treat various diseases in traditional Chinese medicine, and it has been reported to exert anticancer functions[26]. However, the exact mechanism of its anticancer effects against human OS remains unclear. Moreover, it is vital to identify potential molecular diagnostic markers and/or therapeutic targets to combat human OS.
In recent decades, the complexity of the human genome has been revealed by advanced RNA sequencing analyses, and numerous studies have demonstrated that thousands of lncRNAs are expressed in different kinds of human cancers. Certain lncRNAs behave like oncogenes or tumor suppressors, displaying an important function in cancer initiation, progression, metastasis, and recurrence[27, 28]. To date, only a few lncRNAs have been experimentally verified, but their roles in regulating gene expression remain to be deciphered. To the best of our knowledge, this is the first study to identify lncRNAs, miRNAs, and mRNAs to reveal regulatory pathways with regard to baicalin-induced apoptosis in MG63 cells.
Our findings indicated that baicalin could inhibit the proliferation and induce apoptosis of MG63 cells. With |log2 (fold change)| > 2 and p-value < 0.05 thresholds, a total of 58 lncRNAs, 31 miRNAs, and 2136 mRNAs with significant differential expression were identified in the baicalin-treated MG63 cells compared with expression in the untreated cells. Our study has allowed for a deeper understanding of the antitumor effects in OS. Bioinformatics analysis was performed to explore interactions among the DEmRNAs. The most enriched GO terms of DEmRNAs contained BPs, MFs, and CCs, which included transport, cellular nitrogen compound metabolism, macromolecule metabolic process, intracellular part, intracellular and nucleic acid binding. These GO terms could participate in DNA replication, damage detection, and regulation of the activity of cyclin-dependent protein serine/threonine kinases. Through KEGG pathway analysis, we detected a number of cancer-related pathways, including alpha-linolenic acid metabolism, linoleic acid metabolism, biosynthesis of unsaturated fatty acids, the VEGF signaling pathway, and choline metabolism in cancer. Studies have shown that targeting the VEGF-VEGFR pathway seems to be the best approach in hepatic epithelioid hemangioendothelioma[29].
Abnormal choline metabolism continues to be identified in multiple cancers[30]. All these above views demonstrate that our ceRNA network reflects vital mechanisms of anticancer effects. Furthermore, we then constructed a PPI network to identify hub DEmRNAs. Proteins that corresponded to genes were used to build the PPI network. A PPI network including 148 proteins and 597 edges was constructed to reveal the relationships among baicalin-treated proteins in MG63 cells. Two core networks of a total of 64 hub genes were identified in this PPI network by using MCODE v1.5.1. GO enrichment analysis showed that the hub genes might function in proteasome-mediated ubiquitin-dependent protein catabolic processes (1) and nuclear-transcribed mRNA catabolic processes (2), indicating that these processes are present in OS. These reports, together with our findings, suggest that these key regulators might play key roles in regulating baicalin-induced apoptosis in human OS cells.
Subsequently, to provide a possible explanation for the baicalin-induced altered RNA expression levels, we established a ceRNA network of lncRNA-miRNA-mRNA according to bioinformatics analysis. The crosstalk between 2 lncRNAs, 3 miRNAs, and 18 mRNAs revealed a complex mechanism in baicalin-treated MG63 cells. The ceRNA hypothesis explains a new mechanism of RNA interaction and provides important clues and theoretical guidance for further understanding the tumorigenesis mechanism[31]. We observed that lncRNAs sponge several miRNAs, while miRNAs could regulate more than one mRNA. Based on the ceRNA network, we concluded that the lncRNAs with changed expression were linked with baicalin-treated MG63 cells by sponging the related miRNAs. Such a sponging effect may reflect the regulatory potential of ncRNAs in MG63 cells. In recent years, growing experimental evidence indicates that ceRNA is closely related to the development of cancers[32–34].
LncRNAs, the widest ncRNA subgroup, are RNA molecules of more than 200 bases in length, transcribed by RNA polymerase II, and capped and polyadenylated at their 5’ and 3’ ends, respectively[35]. LncRNAs exert both beneficial and detrimental functions by acting at the transcriptional, posttranscriptional, or epigenetic level. In our study, we also identified two DElncRNAs (ENST00000607286 and ENST00000449500) in the ceRNA networks, which were rarely studied in previous studies. ENST00000607286 is a transcript of two exons with a length of 2204 nucleotides and is located on chromosome 2q13. Our results showed interactions between ENST00000607286 and miR-625-5p and miR-1908-3p. Studies have shown that aberrant expression of miR-625-5p in cancers might be a potential risk factor. The mechanisms underlying the antitumor function of miR-625-5p in different cancer types have been demonstrated in previous studies. LINC00958 facilitates cervical cancer cell proliferation and metastasis by sponging miR-625‐5p to upregulate LRRC8E expression, which provides a novel biomarker for experimenters to discover better treatments for cervical cancer patients[36]. LINC009581 elevated the expression of CPSF7 by acting as a miR-625-5p sponge, which accelerated the development and progression of lung adenocarcinoma, thereby demonstrating that LINC009581 might be utilized as a promising therapeutic target for lung adenocarcinoma[37]. LINC00511 is a tumor promoter that sponges miR-625-5p by targeting NFIX in gastric cancer cells, which could be considered a brand new target for gastric cancer treatment[38]. Novel hypoxia-associated circDENND2A enhances the migration and invasion of glioma cells by directly sponging miR-625-5p[39].
In addition, ENST00000449500 is located on chromosome 20 with a total length of 801 bp and two exons. Our results also showed intimate interactions between ENST00000449500 and miRNAs, such as miR-486-3p and miR-1908-3p. Unlike lncRNAs, miRNAs, a group of endogenous, evolutionarily conserved nonprotein-coding RNA molecules with a typical length of 20–24 nucleotides, essentially regulate gene expression via posttranscriptional regulation. ceRNAs can competitively bind to MREs, revealing that miRNAs are at the center of ceRNA networks. It has been reported that dysregulated miRNAs play various roles in the initiation, progression, invasiveness, and metastasis of tumors[40]. miRNAs are involved in multiple roles during carcinogenesis. Among them, miR-486-3p has been shown to exert a regulatory role in tumor progression. Chou et al.[41] demonstrated that miRNA-486-3p functions as a tumor suppressor in oral cancer by targeting DDR1. Another study on cervical cancer suggested that c-Myc could upregulate lncRNA-PVT1 expression, which subsequently releases the inhibition of ECM1 by sponging miR-486-3p, thus enhancing the proliferation and viability of cervical cancer cells. The c-Myc/lncRNA-PVT1/miR-486-3p/ECM1 axis might serve as a new target for more efficient diagnosis and treatment of cervical cancer[42]. In laryngeal squamous cell carcinoma (LSCC), circFLNA functions in LSCC migration by sponging miR-4863p which downregulates FLNA protein expression. Targeting the circFLNA/miR4863p/FLAN axis provides a potential therapeutic target for aggressive LSCC[43]. Moreover, miR-486-3p may serve as a biomarker for the detection of oral tongue squamous cell carcinoma[44]. To date, no study has reported any association of miR-1908-3p with cancer. Moreover, our studies have reported that the apoptosis-related miRNAs miR-130a, miR-222, miR-195, miR-29c, miR-92a-1, and miR-216a were aberrantly expressed in baicalin-induced apoptosis in MG63 cells. Interestingly, among them, miR-195, miR-29c, and miR-92a-1 were reported in a previous study of osteosarcoma[45–47], indicating the high degree of confidence in our sequencing results. This is the first study to show aberrant expression of ENST00000449500 and ENST00000607286 in MG63 cells and indicates a potential prognostic role of this 2-lncRNA signature in MG63 cells. In addition, bioinformatics-based investigations of lncRNAs will be helpful in future experimental studies.
Eventually, the selected RNAs in the ceRNA network were verified by qRT-PCR to confirm the reliability and validity of the above bioinformatics results. The 18 verified RNA expression levels were in line with the sequencing results, indicating the high degree of confidence of this network. We verified the independent RNAs, and the results supported our findings and provided a better understanding of lncRNA-related ceRNAs and their important role in baicalin-induced apoptosis of MG63 cells.
ENST00000449500 is also called lncRNA melanoma highly expressed noncoding RNA (MHENCR); Chen et al[48] found that MHENCR promotes melanoma progression by regulating the miR-425/489-mediated PI3K-Akt pathway. According to our research, lncRNA MHENCR might affect the OS process by regulating miR-1908-3p/miR-468-3p, indicating a new mechanism of MHENCR in OS. Another lncRNA, ENST00000607286, named SLC9A3-AS1 on NCBI, was one of thirteen transcripts. Bai et al[49] found that SLC9A3-AS1 was highly expressed in the peripheral blood of lung cancer patients, but there is no research on its function so far.
To our knowledge, our work is the first expression profile that analyzes and validates the ceRNA-mediated mechanism to signify how baicalin exerts its antitumor effects in OS. Although the findings of our study have important clinical implications, the limitations must also be noted. First, a comparative analysis of the two groups of ceRNA networks showed only two DElncRNAs and three DEmiRNAs. Second, ceRNA binding experiments also need to be further investigated. Further experiments are needed to support the identification of functional roles.