Open Access Highly Accessed Research article

Comparative transcriptomics of human multipotent stem cells during adipogenesis and osteoblastogenesis

Marcel Scheideler1, Christian Elabd2, Laure-Emmanuelle Zaragosi2, Chiara Chiellini2, Hubert Hackl1, Fatima Sanchez-Cabo13, Sunaina Yadav1, Kalina Duszka1, Gerald Friedl4, Christine Papak1, Andreas Prokesch1, Reinhard Windhager4, Gerard Ailhaud2, Christian Dani2, Ez-Zoubir Amri2* and Zlatko Trajanoski1*

Author Affiliations

1 Institute for Genomics and Bioinformatics and Christian Doppler Laboratory for Genomics and Bioinformatics, Graz University of Technology, Petersgasse 14, 8010 Graz, Austria

2 ISBDC, Université de Nice Sophia-Antipolis, CNRS, 28 avenue de Valrose, 06100 Nice, France

3 Genomics Unit, Centro Nacional de Investiganciones Cardiovasculares, Madrid, Spain

4 Department of Orthopaedics, Medical University of Graz, Graz, Austria

For all author emails, please log on.

BMC Genomics 2008, 9:340  doi:10.1186/1471-2164-9-340

Published: 17 July 2008



A reciprocal relationship between bone and fat development in osteoporosis is clinically well established. Some of the key molecular regulators involved in this tissue replacement process have been identified. The detailed mechanisms governing the differentiation of mesenchymal stem cells (MSC) – the key cells involved – are however only now beginning to emerge. In an attempt to address the regulation of the adipocyte/osteoblast balance at the level of gene transcription in a comprehensive and unbiased manner, we performed a large-scale gene expression profiling study using a unique cellular model, human multipotent adipose tissue-derived stem cells (hMADS).


The analysis of 1606 genes that were found to be differentially expressed between adipogenesis and osteoblastogenesis revealed gene repression to be most prevalent prior to commitment in both lineages. Computational analyses suggested that this gene repression is mediated by miRNAs. The transcriptional activation of lineage-specific molecular processes in both cases occurred predominantly after commitment. Analyses of the gene expression data and promoter sequences produced a set of 65 genes that are candidates for genes involved in the process of adipocyte/osteoblast commitment. Four of these genes were studied in more detail: LXRα and phospholipid transfer protein (PLTP) for adipogenesis, the nuclear receptor COUP-TF1 and one uncharacterized gene, TMEM135 for osteoblastogenesis. PLTP was secreted during both early and late time points of hMADS adipocyte differentiation. LXRα, COUP-TF1, and the transmembrane protein TMEM135 were studied in primary cultures of differentiating bone marrow stromal cells from healthy donors and were found to be transcriptionally activated in the corresponding lineages.


Our results reveal gene repression as a predominant early mechanism before final cell commitment. We were moreover able to identify 65 genes as candidates for genes controlling the adipocyte/osteoblast balance and to further evaluate four of these. Additional studies will explore the precise role of these candidate genes in regulating the adipogenesis/osteoblastogenesis switch.