Induction of multilineage differentiation of MSCs by Dex into osteoblasts, adipocytes, skeletal muscle cells, and chondroblasts is accepted widely[1]. However, this diverse induction of Dex in MSCs is depended on the concentration and exposure time[5–8]. Shifting the focus from multilineage differentiation of MSCs, recent studies are paying more attention to the pro-apoptotic effect of Dex. It was reported that long-term exposure to a high dosage of Dex (10− 6 mol/L) induced apoptosis and inhibited proliferation of BMSCs[5, 6], which may contribute to the pathogenesis of skeletal and metabolic disorders, including OP, SONFH, fragility fracture. Nevertheless, the exact mechanism of Dex-induced apoptosis of MSCs remains equivocal, despite reports that Dex can regulate gene expression of MSCs[37].
Emerging evidence has revealed that a large number of signaling pathways, such as PI3K/Akt/mTOR[23], RAF-MEK-MAPK/ERK[38], NF-kB[26] and p53-dependent signaling pathway[39], are implicated in apoptosis. These signaling pathways are regulated by a variety of transcripts, including coding and noncoding RNAs. The lncRNA is a kind of noncoding RNA with a length ofmore than 200 nucleotides and has been reported to regulate apoptosis, proliferation, migration, and differentiation of MSCs by interfering with DNA, mRNA, or protein[40–42]. However, it is still unclear which crucial genes are involved in the apoptosis of Dex-induced MSCs.
Here, we utilized microarray to identify differentially expressed lncRNA and mRNA in Dex-induced apoptosis in hBMSCs and identified 137 differentially expressed mRNA (90 up-regulated and 47 down-regulated). The GO enrichment analysis demonstrated that these differentially expressed mRNA were enriched in the regulation of cell cycle, cell division, cell proliferation- processes closely related to apoptosis. Moreover, pathway enrichment analysis identified a total of 71 significantly differential signaling pathways mainly involved in the cell cycle, signaling by Rho GTPases, polo-like kinase-mediated events, Cyclin B2 mediated events, and cytokine-cytokine receptor interaction were closely related to the regulation of apoptosis[20–22]. Furthermore, Path-Net analysis highlighted signaling pathways mediated by mTOR, Ras, HIF-1, NF-kappa B, and TGF-beta may play critical roles in Dex-induced apoptosis of hBMSCs; consistent with previous reports[23–27]. Notably, many signaling pathways pointed to the regulation of apoptosis by cross-talk.
Also, 90 differentially expressed lncRNA (61 up-regulated and 29 down-regulated) were identified in our microarray assay. Pathway enrichment analysis identified a total of 6 significantly differential signaling pathways. Among them, inactivation of Cdc42 and Rac, signaling by the Robo receptor, Rho GTPase cycle, and PDGF signaling pathway has been previously reported to be associated with regulation of apoptosis. For example, both Cdc42 and Rac, subgroup members of Rho GTPases, suppressed apoptosis through the regulation of cell cycle progression[28, 30]. Moreover, the Robo receptor also has been reported to regulate various cellular processes, including cell proliferation, apoptosis, adhesion, and migration[29]. The PDGF family is divided into four subtypes: PDGF-A, PDGF-B, PDGF-C, and PDGF-D; of these, PDGF-D plays a key role in the regulation of proliferation, apoptosis, migration of cancer cells[31, 43] as well as apoptosis in hepatic stellate cells[44].
A CNC network constructed to investigate the interaction between differentially expressed mRNA and lncRNA in Dex-induced apoptosis of hBMSCs identified some key mRNAs, such as SAMHD1, CDK1, GINS4, CDH11 and CIT, which were closely associated with regulation of cell proliferation and apoptosis. In addition, some important lncRNA identified included, ENSG00000233901.1, ENSG00000251018.2, ENSG00000255733.1, ENSG00000226605.1, ENSG00000233901.1, ENSG00000230921.1, SETMAR. Most importantly, there was a significant correlation between these key transcripts. For instance, GINS4 was positively correlated with ENSG00000251018.2, but negatively with ENSG00000233901.1. CDH11 was positively correlated with SETMAR, and negatively with ENSG00000230921.1. Besides, SAMHD1, CDK1,, and CIT were negatively correlated with ENSG00000255733.1, ENSG00000226605.1, and ENSG00000233901.1, respectively. Although the relationship between these lncRNAs involved in CNC network and apoptosis has not been previously reported, they may regulate apoptosis by interacting with mRNA, due to the correlation between them.
Next, we selected 6 differentially expressed mRNA and 4 differentially expressed lncRNA, which have been reported in previous studies, to confirm the reliability of the microarray data and facilitate further analysis. Gene expression analysis by qRT-PCR confirmed that both GINS4, CIT, and CDK1 were up-regulated, and BCL2-L11, SAMHD1, and CDH11 were down-regulated. Besides, lncRNAs STXBP5-AS1, IFNG-AS1, and MIR210HG were up-regulated, while ZFHX4-AS1 was down-regulated. GINS4, also known as SLD5, is an important player in the initial stages of DNA replication. Down-regulation of GINS4 promoted cell cycle arrest, growth inhibition, and apoptosis in colorectal cancer cells[32]. Inactivation of CIT, a serine/threonine kinase, increased apoptosis, suppressed proliferation, and arrested cell cycle via the regulation of Cyclophilin A in PDAC cells[33]. Similarly, inhibition of CDK1, a mitosis-promoting factor, promoted apoptosis of cancer cells through G2/M arrest [34]. BCL2-L11, a member of the Bcl-2 family, was pro-apoptotic in cardiomyocytes due to its interaction with Bcl-2[45, 46]. SAMHD1, a mammalian dNTP hydrolase (dNTPase), increased apoptosis and reduced the proliferation of cancer cells by regulating the G1/G0 phase [35]. CDH11 belonging to the cadherin family induced apoptosis, inhibited cell proliferation by arresting cell cycle in the G0/G1 phase in colorectal cancer cell lines[36].
The lncRNA IFNG-AS1 has been reported to inhibit apoptosis, promote proliferation, invasion, and migration of HP75 cells [47]. Likewise, lncRNA ZFHX4-AS1 repressed apoptosis of breast cancer cells by regulating the Hippo signaling pathway[48]. In addition, the anti-apoptotic effects of MIR210HG in cancer cells by promoting proliferation and invasion in cervical cancer[49]. On the other hand, STXBP5-AS1 promoted apoptosis, suppressed proliferation, and invasion of MCF-7 cancer cells [50]. It can be stated that both mRNA (GINS4, CIT, and CDK1) and lncRNA (IFNG-AS1, ZFHX4-AS1, and MIR210HG) can be anti-apoptotic, whereas BCL2-L11, SAMHD1, and CDH11 and STXBP5-AS1 can be pro-apoptotic in specific cell lines. It is interesting to note that up-regulation of STXBP5-AS1 (apoptosis activators) and down-regulation of ZFHX4-AS1 (apoptosis inhibitors) in Dex-induced hBMSCs were consistent in both qRT-PCR and microarray analysis. Contrarily, the expression of the other 8 genes was inconsistent with the pro-apoptotic effect of Dex on hBMSCs. To our knowledge, the regulation of apoptotic genes is cell-type dependent. For instance, previous studies demonstrated that the lncRNA H19 inhibited MC apoptosis[51], but promoted apoptosis in cardiomyocytes and hippocampal neurons[52, 53]. Hence, given the cell-type specificity of Dex induced responses, the roles and underlying mechanisms of the genes identified in the present study need further elucidation.
Collectively, we identified differentially expressed lncRNA and mRNA in Dex-induced apoptosis of hBMSCs by microarray. Bioinformatic analysis such as GO enrichment analysis, pathway enrichment analysis, and CNC network construction further revealed the role of these differentially expressed genes. The present study provides evidence in support of further studies to reveal the exact mechanisms of Dex-induced apoptosis of BMSCs.