Introduction: The existence of peripheral blood (PB) derived mesenchymal stem cells (PB-MSCs) have been documented in different mammalian species including young and adult human. However, the number of PB-MSCs is low in normal adult human blood. We have demonstrated previously that there was an increase in the number of PB-MSCs following long bone fracture and in the patients suffering from fracture non-union. The present study was to compare the biological characteristics of the PB-MSCs from fracture non-union patients, with human bone marrow derived MSCs (BM-MSCs).

Methods: 200 mls PB was collected from 9 patients suffering from fracture non-union. The mononuclear cells (MNCs) were isolated by density gradients centrifugation and cultured in á-MEM containing 15% FBS. The PB-MNCs from normal donors (n=8) and BM-MSCs from patients underwent total hip replacement were used as controls. The colony forming efficiency (CFE) of the PB-MSCs was calculated, and the phenotypes of PB-MSCs and BM-MSCs were compared using immunocytochemistry and flow cytometry methods. Their multipotent differentiation potentials into osteoblasts, chondrocytes, adipocytes, neurogenic and angiogenic cells were examined under specific inductive culture media. The in vivo osteogenic potential of PB-MSCs was examined by implanting the HA-TCP blocks seeded with PB-MSCs into the SCID mice for 12 weeks.

Results: After 28 days in culture, fibroblastic colonies were formed in the PB-MNCs cultures in 5 of 9 fracture non-union patients, with CFE ranging from 2.08–2.86 per 10^8 MNCs. No fibroblastic colony was seen in PB-MNCs cultures of the 8 normal donors. Under flow cytometry examination, PB-MSCs and BM-MSCs were CD34 (low) and CD105+, but PB-MSCs were CD29-, CD44-, and ALP (low), whereas BM-MSCs were CD29+, CD44+, and ALP (high). Under specific differentiation inductions, the PB-MSCs differentiated into osteoblastic cells (ALP+, type I collagen+, osteocalcin+ and Alizarin red+; chondrocytes (type II collagen+ and Alcian Blue nodules formation); adipocytes (Oil red-O positive lipid accumulation). Neurogenic differentiation was confirmed by positive neuro-filament staining, and differentiation into endothelial cells was evident with tube formation in 2D culture, and positive staining for VW factor and CD31. After implantation in the SCID mice for 12 weeks, newly formed woven bones were found in the biomaterials seeded with PB-MSCs, and they were positive for human osteocalcin immunostaining.

Discussion: This study indicated that there were more PB-MSCs in the peripheral circulation of the fracture non-union patients than that in the normal subjects. This may be due to a continous systemic response for recruiting MSCs from remote bone marrow sites, with attempt to repair the fracture(s). The PB-MSCs were clearly multi-potential cells, which had shared some common phenotypic markers with BM-MSCs, as well as many distinguishable makers from the BM-MSCs. The recruitment of the PB-MSCs through circulation might be a general phenomenon of systemic responses in many pathological conditions, such as fracture or wound healing and other systemic diseases. Further understanding the roles of PB-MSCs in diseases and repair may lead to novel therapeutic strategies.