Time series clustering of mRNA and lncRNA expression during osteogenic differentiation of periodontal ligament stem cells

Background: Long noncoding RNAs (lncRNAs) are regulatory molecules that participate in biological processes such as stem cell differentiation. Periodontal ligament stem cells (PDLSCs) exhibit great potential for the regeneration of periodontal tissue and the formation of new bone. However, although several lncRNAs have been found to be involved in the osteogenic differentiation of PDLSCs, the temporal transcriptomic landscapes of mRNAs and lncRNAs need to be mapped to obtain a complete picture of osteoblast differentiation. In this study, we aimed to characterize the time-course expression patterns of lncRNAs during the osteogenic differentiation of PDLSCs and to identify the lncRNAs that are related to osteoblastic differentiation.

Methods: We cultured PDLSCs in an osteogenic medium for 3, 7, or 14 days. We then used RNA sequencing (RNA-seq) to analyze the expression of the coding and non-coding transcripts in the PDLSCs during osteogenic differentiation. We also utilized short time-series expression miner (STEM) to describe the temporal patterns of the mRNAs and lncRNAs. We then performed Gene Ontology and Kyoto Encyclopedia of Genes and Genomes analyses to assess the biological relevance of genes in each profile, and used quantitative real-time PCR (qRT-PCR) to validate the differentially expressed mRNAs and lncRNAs that were associated with osteoblast differentiation. Lastly, we performed a knock down of two lncRNAs, MEG8, and MIR22HG, and evaluated the expression of osteogenic markers.

Results: When PDLSCs were differentiated to osteoblasts, mRNAs associated with bone remodeling, cell differentiation, and cell apoptosis were upregulated while genes associated with cell proliferation were downregulated. lncRNAs showed stage-specific expression, and more than 200 lncRNAs were differentially expressed between the undifferentiated and osteogenically differentiated PDLSCs. Using STEM, we identified 25 temporal gene expression profiles, among which 14 mRNA and eight lncRNA profiles were statistically significant. We found that genes in pattern 12 were associated with osteoblast differentiation. The expression patterns of osteogenic mRNAs (COL6A1, VCAN, RRBP1, and CREB3L1) and lncRNAs (MEG8 and MIR22HG) were consistent between the qRT-PCR and RNA-seq results. Moreover, the knockdown of MEG8 and MIR22HG significantly decreased the expression of osteogenic markers (runt-related transcription factor 2 and osteocalcin).

Discussion: During the osteogenic differentiation of PDLSCs, both mRNAs and lncRNAs showed stage-specific expression. lncRNAs MEG8 and MIR22HG showed a high correlation with osteoblastogenesis. Our results can be used to gain a more comprehensive understanding of the molecular events regulating osteoblast differentiation and the identification of functional lncRNAs in PDLSCs.