Abstract Currently, the niche for long-term hematopoietic stem cells (HSCs) is thought to consist conceptually of two parts: the endosteal surface (the osteoblastic niche) and a sinusoidal endothelium (the vascular niche), and a subset of osteoblasts functions as a key component of the hematopoietic stem cell niche. However, it is still unclear that the precise cellular and molecular contribution of osteoblastic cells on the HSC supportive microenvironment. In this study, we try to characterize the osteoblastic cells and investigate the property of osteoblastic niche cells. For isolation of osteoblastic cells, we treated the bone fragments of femur and tibiae with collagenase following flush-out of the bone marrow (BM). Non-hematopoietic and non-endothelial cells were then enriched by magnetic cell sorting of the CD45-CD31-Ter119- population, and expression of Sca-1 and platelet derived growth factor receptor α (PDGFRα) was analyzed. FACS analysis showed that CD45-CD31-TER119- cells were subdivided into three fractions: Sca-1+PDGFRα+, Sca-1-PDGFRα-, Sca-1-PDGFRα+. First we examined the multilineage differentiation potential of three populations. Although Sca-1- fractions efficiently differentiated into the osteoblastic lineage and showed calcium deposition, these cells hardly differentiated into adipocytes. In contrast to the Sca-1- cells, we found that Sca-1+PDGFRα+ cells can differentiate into osteoblastic and adipocytic lineages, suggesting that Sca-1+ cells have multi-potency. Next we examined the expression of osteoblastic marker expression by quantitative RT-PCR analysis, and found that Sca-1- populations expressed Runx2 and OB-cadherin. Alkaline phosphatase (ALP) staining of freshly isolated cells showed that Sca-1- fractions expressed ALP, while Sca-1+ cells did not express ALP. These data suggest that Sca-1- populations were the cell fractions, which were already committed to osteoblastic lineage. In addition, osteocalcin was expressed in PDGFRα+ fraction in Sca-1- cells, indicating that Sca-1-PDGFRα+ cells are more mature osteoblastic cells than Sca-1-PDGFRα-cells. Furthermore, N-cadherin was specifically expressed in Sca-1-PDGFRα+ cells, suggesting that N-cadherin was up-regulated with the maturation of osteoblastic cells. In addition, N-cadherin expression was up-regulated in Sca-1-PDGFRα+ cells with the postnatal development of BM. Interestingly, in the freshly isolated cells, we found that Sca-1+PDGFRα+ cells showed higher expression of Angiopoietin-1 (Ang-1), compared to Sca-1- fractions. Ang-1 expression was up-regulated in Sca-1-PDGFRα+ cells after over night incubation. Next we investigated the ability of these fractionated cells to support hematopoiesis. We examined the capacity of these fractionated cells on maintenance of colony formation ability of BM linage-Sca-1+c-Kit+ cells after 5 days of co-culture. Although CFU-C formation was supported Sca-1+PDGFRα+ cells, Sca-1-PDGFRα+ cells maintained CFU-Mix formation compared to the Sca-1+PDGFRα+ and Sca-1-PDGFRα-cells. From these data above, we hypothesize that multiple osteoblastic populations form a “niche complex” and collaborate with other supporting cells, such as CXCL12-abundant reticular (CAR) cells, to support HSCs, and that N-cadherin+ osteoblastic cells provide a foothold for anchoring of quiescent HSCs. Now we are investigating the gene expression profiles of these three populations and are tying to clarify the changes of characteristics of osteoblastic cells during postnatal BM development.