Systems analysis of human mesenchymal stem cells during differentiation into osteoblasts

To better understand human Mesenchymal Stem Cells (hMSCs) ex vivo and their use in therapy, we have studied the molecular phenotype of telomerised hMSC (hMSC-TERT) using four different approaches. Firstly, we used RNASeq to examine changes in expression for a set of skeletally-related genes during ex vivo osteoblast differentiation of hMSC-TERT. From this set, we propose a set of ex vivo differentiation-stage-specific and tissue specific markers (n=21). We also identified a subset of genes (n=20) to predict the bone forming capacity of hMSC and another set (n=20) associated with the ex vivo phenotype of hMSC obtained from osteoporotic patients. These results address the current lack of stage-specific osteoblastic markers available to the bone community. Secondly, we compared hMSC-TERT with primary hMSC using gene expression microarray analysis. We showed that CD markers, osteoblast markers and transcription factors had a high degree of similarity between the two cell populations. When compared quantitatively, some differences were observed. We also show that telomerisation of the hMSC resulted in functional enrichment of skeletal system development, cell cycle and immune response pathways. Our results show that telomerisation maintains the molecular phenotype of primary hMSCs and enhances certain characteristics for their clinical use. Thirdly, we combined gene expression microarrays, clustering with Self Organising Maps and literature analysis, to identify three novel osteogenic transcription factor candidates: ZNF25, ZNF608 and ZBTB38. Functional studies in vitro showed that of the three, ZNF25 had the most significant effect on osteoblast differentiation. siRNA knockdown experiments for ZNF25 in hMSC-TERT cells during osteoblast differentiation resulted in significant suppression of ALP activity and extracellular mineralized matrix formation. Our results indicate that ZNF25 is involved in matrix mineralisation and maturation process of OB cells.Finally, we combined RNASeq data with a domain-domain interaction (DDI) network to explore how alternatively spliced isoforms of hMSCs interact with each other during osteoblast differentiation. The DDI network identified switching of isoform interactions during osteoblast differentiation. Switching isoforms were enriched for biological processes transcription, cytoskeleton and phosphorylation. Proteins displaying isoform switching included known osteogenic transcription factors as well as members of the MAPK signalling pathway