Characterization

MSC are isolated by their ability to adhere to plastic when plating either whole bone marrow or mononuclear cells separated on a Ficoll density gradient. Within the bone marrow, it has been calculated that their frequency is 1/10⁴ to 1/10⁵ of total mononuclear cell numbers (Kadiyala, 1997; Suva, 2004; Wexler, 2003). They are

characterized by a spindle-shaped, fibroblast-like morphology (Colter, 2000;

Friedenstein, 1976). This isolation procedure bears the disadvantage of possible contamination with endothelial cells, fibroblasts and macrophages. During culture, macrophages are eliminated by passages, but fibroblasts and endothelial cells can survive and also expand in culture.

To phenotypically characterize these cells, flow cytometry has been used extensively, using monoclonal antibodies to surface antigens. MSC are not characterized by the expression of a unique marker, but by a panel of surface antigens. MSC are characterized by the presence of CD29, CD44, CD54 low, CD73, CD90 (thy-1), CD105, and CD106, and absence of CD11b, CD14, CD31, CD34, CD45, CD117 (c-kit) and CD133. Thus, hematopoietic stem cell- and endothelial cell markers are absent on MSC (Peister, 2004; Pittenger, 1999; Reyes, 2001a; Suva, 2004). Differences in cell surface antigen expression can be observed within different studies and might be explained by differences in culture conditions or by factors secreted by accessory cells during early passages (da Silva Meirelles, 2006; Grove, 2004; Ho, 2008; Kemp, 2005; Markov, 2007).

Recently, some new markers have been identified which are expressed on human MSC but not on fibroblasts: neural ganglioside GD2 (Martinez, 2007), STRO-1 with CD106 (Gronthos, 2008), and SSEA-4, a marker for embryonic stem cells (Gang, 2007), markers which may help to distinguish MSC from fibroblasts. Gene expression analysis of MSC and fibroblasts has further identified molecular markers in order to differentiate both cells types (Ishii, 2005). This technique is however not suitable for ex vivo purification, as most markers identified are not expressed on the surface of cells (Ishii, 2005). Furthermore, the precise barrier between MSC and fibroblasts is not clear as MSC colonies were isolated from primary cultured fibroblasts (Sudo, 2007).

Because of the lack of specific markers for MSC, to date demonstration of multilineage differentiation into mesenchymal tissues or cells is necessary for MSC to be called as such. In vitro differentiation into adipocytes, osteoblasts and chondrocytes confirms their pluripotency and multilineage potential (Dominici, 2006).

Dominici et al claimed the need for a more homogenous description of MSC and a consensus in the scientific community (Dominici, 2006).

Established protocols for differentiation of MSC into osteoblasts, chondrocytes and adipocytes.

Differentiation of MSC into osteoblasts in vitro is obtained by adding different components such as ascorbic acid, beta-glycerophosphate and dexamethasone to culture medium on confluent MSC for 3 weeks (Pittenger, 1999). Osteoblast differentiation is demonstrated by the formation of bone nodules stained by histochemistry with Alizarin red (Stenderup, 2001), showing alkaline phosphatase activity (D'Ippolito, 1999), or calcium deposit accumulation with von Kossa technique (Friedenstein, 1976; Marom, 2005; Pittenger, 1999). Bone morphogenic protein (BMP)-2 in conjunction with fibroblast growth factor (FGF)-2 strengthen bone forming pathways in vitro (Knippenberg, 2006; Partridge, 2002). In vivo, BMP-2 and BMP-9 induce osteogenesis of implanted MSC (Kang, 2008). How endogenous MSC intervene precisely in bone homeostasis needs further clarification (Kang, 2008), but interaction between HSC and MSC seem to be important (Jung, 2008).

To induce differentiation into adipocytes, confluent MSC culture medium is supplemented with dexamethasone or hydrocortisone, isobutyl methyl xanthine (IBMX), indomethacin, horse or rabbit serum and insulin (Janderova, 2003; Koch, 2007; Noth, 2002; Pittenger, 1999; Suva, 2004). After 3 weeks, there is an accumulation of lipid containing vacuoles that fuse to form pre-adipocytes and adipocytes (Pittenger, 1999). Triglycerides within these lipid vacuoles can be stained by histochemistry using Oil-red-O to demonstrate adipogenic differentiation (Hausman, 1981; Pittenger, 1999). Analysis of expression of adipocyte genes such as lipoprotein lipase, peroxysome proliferation-activated receptor gamma-2 (PPARγ2), and GLUT4 are commonly used to confirm differentiation into adipocytes (Neubauer, 2004; Pittenger, 1999). In vivo, little is known about adipogenic differentiation of MSC. Bone marrow of elderly patients contains more fat tissue (Gimble, 2006). Fatty degeneration of bone marrow cells or adipocyte invasion of the bone marrow may be underlying this phenomenon. Whether MSC contribute actively to this degeneration or to bone marrow infiltration has not been clearly established. A proposed model is that adipogenic cells (possible MSC) or adipocytes secrete toxic fatty acids and adiponectine to inhibit osteoblast differentiation, and PPAR gamma expression in order to activate osteoclasts, thereby inhibiting bone formation and inducing bone resorbtion (Duque, 2008; Kirkland, 2002).

Differentiation into chondrocytes can be induced in vitro, by pellet culture of MSC in medium containing, TGFbeta3 (Pittenger, 1999) or pyruvate, ascorbic acid, dexamethasone and TGFbeta1 (Johnstone, 1998; Suva, 2004). Using this method, the pelleted cells form a round dense mass which can be dislodged and differentiation finalized in suspension culture for 3 weeks (Pittenger, 1999; Suva, 2004). Pellet sections are then stained by histochemistry with Alcian blue (Carlberg, 2001) or Masson’s trichrome (Estrada, 2001) to stain acid mucopolysaccharides and glycosaminoglycans or collagen fibrils contained in cartilage respectively.

Immunohistochemistry for type 2 collagen is commonly used to confirm differentiation into hyaline cartilage (Pittenger, 1999; Suva, 2004). Differentiation pathways and potential improvements of differentiation protocols have been studied extensively, as cartilage replacement in injured joints or osteoarthritis seems a reachable clinical application for stem cell technology. One major technical issue is however that stem cells-derived tissues are not as biomechanically resistant as endogenous/native cartilage (Djouad, 2006; Raghunath, 2005). Cartilage homeostasis and cartilage repair have been evaluated in rats and rabbits. These studies showed that MSC rapidly invade lesions and that regeneration of hyaline cartilage and subchondral bone is FGF-2 dependent (Anraku, 2008; Mizuta, 2006).

In summary, according to recent guidelines, MSC are characterized as such by first identification of a panel of surface markers and second the demonstration of multilineage differentiation into osteoblasts, chondrocytes and adipocytes in vitro (Dominici, 2006).