Bone Marrow Cell Aggregates: A Way of Collecting the Adult Human Hematopoietic Stem Cell Niche

Hematopoietic stem cell (HSC) function is critical in maintaining hematopoiesis continuously throughout the lifespan of an organism and any change in their ability to self-renew and/or to differentiate into blood cell lineages induces severe diseases. Postnatally, HSC are mainly located in bone marrow where their stem cell fate is regulated through a complex network of local influences, thought to be concentrated in the bone marrow (BM) niche. Despite more than 30 years of research, the precise location of the HSC niche in human BM remains unclear because most observations were obtained from mice models. BM harvesting collects macroscopic coherent tissue aggregates in a cell suspension variably diluted with blood. The qualitative interest of these tissue aggregates, termed hematons, was already reported (first by I. Blaszek's group (Blaszek et al., 1988, 1990) and by our group (Boiret et al., 2003)) yet they remain largely unknown. Should hematons really be seen as elementary BM units, they must accommodate hematopoietic niches and must be a complete ex vivo surrogate of BM tissue. In this study, we analyzed hematons as single tissue structures.

Biological samples were collected from i) healthy donor bone marrow (n= 8); ii) either biological samples collected for routine analysis by selecting bone marrow with normal analysis results (n=5); or iii) from spongy bone collected from the femoral head during hip arthroplasty (n=4). After isolation of hematons, we worked at single level, we used immunohistochemistry techniques, scanning electronic microscopy, confocal microscopy, flow cytometry and cell culture.

Each hematon constitutes a miniature BM structure organized in lobular form around the vascular tree. Hematons are organized structures, supported by a network of cells with numerous cytoplasmic expansions associated with an amorphous structure corresponding to the extracellular matrix. Most of the adipocytes are located on the periphery, and hematopoietic cells can be observed as retained within the mesenchymal network. Although there is a degree of inter-donor variability in the cellular contents of hematons (on average 73 +/- 10 x10³ cells per hematon), we observed precursors of all cell lines in each structure. We detected a higher frequency of CD34⁺ cells than in filtered bone marrow, representing on average 3% and 1% respectively (p<0.01). Also, each hematon contains CFU-GM, BFU-E, CFU-Mk and CFU-F cells. Mesenchymal cells are located mainly on the periphery and seem to participate in supporting the structure. The majority of mesenchymal cells isolated from hematons (21/24) sustain in vitro hematopoiesis. Interestingly, more than 90% of the hematons studied contained LTC-ICs. Furthermore, when studied using confocal microscopy, a co-localization of CD34⁺ cells with STRO1⁺ mesenchymal cells was frequently observed (75% under 10 µm of the nearest STRO-1⁺ cell, association statistically highly significant; p <1.10^-16).

These results indicate the presence of one or several stem cell niches housing highly primitive progenitor cells. We are confirming these in vitro data with an in vivo xenotransplantation model. These structures represent the elementary functional units of adult hematopoietic tissue and are a particularly attractive model for studying homeostasis of the BM niche and the pathological changes occurring during disease.