Human mesenchymal stromal cells (MSCs) are spindle-shape, plastic-adherent cells with capacity to differentiate to bone, cartilage, and fat. MSCs express fibroblast, endothelial, and lymphocyte antigens, e.g. CD105, CD73, CD90, and CD166 which are the cornerstone of phenotypic characterization of these cells. We recently showed that MSCs are the only bone marrow cell to express GD2, a neural ganglioside. Now, for the first time we show that GD2 may serve as the single, unique, and definitive marker of marrow and adipose derived MSCs that can be used to isolate GD2+ MSCs, which possess important biologic properties justifying prospective isolation. MSCs expression of GD2 is uniformly high on freshly isolated and culture-expanded cells. Using the Miltenyi AutoMACS® device and a monoclonal antibody recognizing GD2 (clone 14.G2A) we prospectively isolated a highly enriched MSC population from bone marrow MNCs. The selected fraction was >98% pure for GD2+ cells determined by flow cytometry. Light microscopy showed that the GD2-selected cells were smaller, thinner, and more spindle-like when attached to plastic compared to unselected MSCs which spread wider along the surface of the culture flask, the so-called “fried egg” appearance. The doubling time of GD2-selected MSCs was 30 hrs compared to 90 hrs for unselected cells representing a 3-fold greater growth rate. Cell cycle analysis by flow cytometry showed ∼80% of cells were in G0/G1 and ∼20% were in S/G2/M phases of the cell cycle in both populations. With the shorter doubling time, this data indicates that GD2-selected MSCs move through the cell cycle more rapidly than unselected cells. In accordance with this finding, electron microscopy showed few organelles in the GD2-selected cells, but increase lamellar bodies indicating overall less complexity, but consistent with a greater membrane turnover rate (cell division) than unselected MSCs. Moreover, flow cytometric analysis revealed an increased expression of receptors for bFGF and EFG, known mitogenic factor receptors for MSCs, compared to unselected MSCs. In vitro differentiation of GD2-selected MSCs showed a more robust osteoid matrix formation (osteoblast) and proteoglycan formation (chondroblast) assayed by semi-quantitative Alizarin Red and Alcian blue staining, respectively. Additionally, more GD2-selected MSCs differentiated to adipocytes than among unselected cells. Surprisingly, GD2 expression persisted on the in vitro human MSC-differentiated osteoblasts, chondroblasts, and adipocytes, in contrast to human bone-derived osteoblasts, adipose tissue, and cartilage which lacked GD2 expression. We conclude that

  • GD2 is a unique, stably expressed surface MSC marker which can be used to prospectively isolate MSCs from marrow,

  • GD2-selcted cells have a more robust in vitro proliferation and differentiation potential which may be valuable for cell therapy, and

  • biologically, in vitro isolated MSCs may not represent the in vivo progenitor for bone, fat, or cartilage.

Author notes

Disclosure: No relevant conflicts of interest to declare.