Prospective Identification and Characterization of Human Mesenchymal Stem Cell Subpopulations.

Mesenchymal stem cells (MSC) are multipotent cells that readily differentiate into a variety of cell types including osteoblasts, chondrocytes, and adipocytes. They support hematopoiesis (stroma function) and have potent immunomodulatory properties making them promising candidates for clinical use. Although MSC cultures are heterogeneous in morphology, no specific markers are available that would allow for the prospective isolation of distinct subpopulations. Therefore we investigated whether potential MSC subpopulation markers could be identified by gene expression profiling of MSC that were sorted based on functional parameters (i.e. proliferation characteristics). Human MSC were generated in standard cultures and stained with carboxyfluoresceine-succinimidylesther (CFSE) for cell division tracking. FACS analysis of CFSE-stained cells after 10 days culture showed that the majority of cells had undergone more than 3 cell divisions (51.2 ± 1.7%) whereas only 3.5 ± 0.9 % had not divided. Based on these results MSC were sorted by FACS according to their divisional history into non-proliferating cells (NP-MSC) and rapidly-proliferating cells (RP-MSC). Interestingly, the CFU-F frequencies of the NP-MSC were considerably (5-times) lower compared to RP-MSC. Furthermore, RP-MSC with low forward-/side-scatter properties had a higher CFU-F frequency than those with high scatter properties. Comparative oligonucleotide microarray analysis of FACS-sorted NP-MSC and RP-MSC populations identified a total of 112 differentially (≥4-fold) expressed genes. Fifty genes were significantly higher expressed in RP-MSC with 12 of them being cell cycle associated. Sixty-two genes showed a higher expression in NP-MSC and more than 20 of them were related to adhesion molecules, extracellular matrix proteins or cell senescence. Two of the differentially expressed genes (VCAM-1, FMOD) that corresponded to cell surface molecules were further tested for their potential to prospectively identify MSC subpopulations. Remarkably, FACS analysis using anti-VCAM-1 (CD106) and anti-fibromodulin (fmod) antibodies showed that approximately 40% and 3% of MSC were CD106^neg/fmod^neg and CD106^pos/fmod^pos, respectively. MSC were then sorted by FACS according to their expression of these markers and assayed for CFU-F frequency. Significant differences in CFU-F frequencies were observed between the double-negative and the double-positive MSC (12.3 ± 2.8 and 4.7 ± 2.1 colonies per 100 cells for CD106^neg/fmod^neg and CD106^pos/fmod^pos MSC, respectively). Experiments to test for differentiation potential differences are ongoing. Taken together, the data clearly demonstrate functional differences between MSC subpopulations. Furthermore, cell sorting based on proliferation characteristics and gene expression profiling enabled to identify surface markers that allowed for the first time to prospectively identify MSC subpopulations.