Human bone marrow (BM) contains a rare population of non-hematopoietic mesenchymal stem cells (BM-MSC) which can differentiate toward skeletal lineages such as osteoblasts, adipocytes, chondrocytes and hematopoiesis-supporting stromal cells. In vivo, BM-MSC are essential constituents of the hematopoietic stem cell niche, thus playing an important role in supporting, maintaining and controlling hematopoiesis.
We and others have previously shown that primary BM-MSC were exclusively found in the linneg/CD45neg/CD271pos cell fraction in human bone marrow, and we furthermore reported that expression of CD146 on BM-MSC correlated with in-situ localization (Tormin et al., Blood 2011,117:5067–5077). Although BM-MSC were highly enriched in linneg/CD45neg/CD271pos cells as reflected by CFU-F frequencies of about 1 in 20, there was still a considerable fraction of non-colony forming cells present in this population. Therefore, the current study aimed to identify novel MSC markers that would allow for a more precise definition of the candidate stromal stem cell population in human bone marrow.
Human bone marrow linneg/CD45neg cells were sorted based on CD271 expression and comparative gene expression profiling was performed using the Illumina Human HT-12 expression v4 BeadChip comprising 48,107 probes. In total, 215 genes were found to be significantly up-regulated in the linneg/CD45neg/CD271pos subset compared to linneg/CD45neg/CD271neg cells, whereas 97 genes were down-regulated. Twenty eight of the upregulated genes correlated to surface markers and expression of thirteen of them could be verified by FACS. Several of the surface markers identified by this approach, such as CD140b, CD10 and CD106 were previously described in the context of MSC isolation. However, the majority of them represented novel MSC markers including molecules such as CD151, CD81, IFNGR2, LEPR, TGFBR3, IL1R1, CD18, CD140a, and FGFR3.
FACS analysis of these markers on linneg/CD45neg/CD271pos cells revealed two staining patterns, i.e. A) marker expression either correlated directly with CD271 expression, or B) the novel maker was only expressed on a fraction of linneg/CD45neg/CD271pos cells. CD151 and CD106 are examples for pattern A markers and, as expected, CFU-F frequencies in sorted linneg/CD45neg/CD271pos/CD151pos and linneg/CD45neg/CD271pos/CD106pos cells were comparable with linneg/CD45neg/CD271pos cells. Furthermore, proliferation and in-vitro/in-vivo differentiation capacities were comparable. On the other hand, using CD140a (platelet-derived growth factor receptor α, PDGFRα) - one of the pattern B markers - allowed to clearly identify a population of linneg/CD45neg/CD271pos/CD140aneg cells which were highly enriched for CFU-F (24.15 ± 4.51 CFU-Fs per 100 plated cells, n=6) compared to linneg/CD45neg/CD271pos/CD140apos cells (1.13 ± 0.65 CFU-Fs per 100 plated cells, n=6). The high CFU-F frequency in CD140aneg cells was furthermore confirmed in single cell sorting and limiting dilution experiments. Quantitative RT-PCR of sorted primary CD140neg MSC showed considerably higher expression of ALPL, PPARγ, and ACAN as well as Oct4, Sox2 and Nanog compared to CD140apos cells, and multicolor FACS analysis revealed that linneg/CD45neg/CD271pos/CD140aneg cells co-expressed typical primary MSC markers (CD90, CD105, CD140b, STRO-1), but not CD31 and CD34. Furthermore, linneg/CD45neg/CD271pos/CD140aneg cells (bulk and single cell sorted) gave rise to typical cultured MSC (expression of standard surface markers, in-vitro differentiation capacity). Moreover, linneg/CD45neg/CD271pos/CD140aneg -derived stromal cells formed bone, adipocytes and hematopoietic stroma when transplanted s.c. into NOD-SCID mice.
Taken together, sorting of linneg/CD45neg/CD271pos cells based on CD140a (PDGFRα) expression enabled to isolate CFU-F with thus far unmet precision. Linneg/CD45neg/CD271pos/CD140aneg cells had typical BM-MSC properties, thus possibly representing a close to pure population of the candidate human primary mesenchymal stem/progenitor cells. These findings will enable to better characterize native BM-MSC and establish their physiological role in vivo.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.