In the bone marrow (BM), osteoblasts and endothelium constitute functional niches providing positive or negative signals for hematopoietic stem cell (HSC) self-renewal. In addition to hematopoietic cells, endothelial cells, and osteoblasts, adult BM contains numerous adipocytes. Interestingly, the number of adipocytes correlates inversely with the gross hematopoietic activity of the marrow. Whether adipocytes have a direct effect on hematopoietic progenitors or whether they act as mere space-fillers in this context remains unclear. To determine the potential role of bone marrow adipocytes in hematopoiesis, we induced bone marrow-derived OP9 mesenchymal cells to differentiate into either osteoblastic or adipocytic stroma, and then tested their capacity to serve as surrogate HSC niches during in vitro hematopoietic culture in the absence of exogenous growth factors. We found that the presence of BM-derived adipocytes suppresses the expansion of short-term hematopoietic progenitors by at least two fold, as measured by the number of CD45+ cells expanded, the number of colony forming unit cells (CFU-Cs) and the competitive repopulation units (CRUs) during the first two months post-transplant. As an in vivo correlate, we compared the hematopoietic activity within the BM of the adipocyte-poor thoracic vertebrae and the adipocyte-rich proximal tail vertebrae. Indeed, we found that the percentage of HSCs (ckit+Lin-Sca1+), CMPs, GMPs and MEPs was decreased by two fold in the adipocyte-rich BM as determined phenotypically by FACS and functionally by short-term and long-term competitive repopulation assays. Mechanistically, oligonucleotide expression microarrays and conditioned media experiments on the OP9 co-cultures suggest that the inhibition of the progenitor compartment in adipocyte-rich environments is due to the loss of supportive factors (Notch ligands, N-cadherin, Angiopoietin-1, SCF, BMPs and Wnt5a) in addition to the presence of an active inhibitor. Finally, we found bone marrow adipocytes to accumulate in great numbers upon bone marrow ablation, a process that is hindered in genetically adipocyte-deficient mice. Since early BM transplant survival depends on the rapid accumulation of short term hematopoietic progenitors, whose replication we found hindered in adipocyte-rich BM, we were interested to explore whether the absence of adipocytes in the context of BM transplantation would foster faster recovery of lethally irradiated mice. As predicted, circulating leukocyte counts on the third week post-transplant were 3–5 times higher on the recovering fatless mice. Accordingly, the percentage of CFU-Cs, HSCs, CMPs, GMPs and MEPs on day 17 post-transplant was doubled on the adipocyte-depleted mice as compared to their wildtype littermates. We therefore conclude that, as seen in vitro, the presence of adipocytes in the recovering HSC niche is detrimental to the rapid hematopoietic expansion required to reconstitute blood production. We are currently examining the pharmacologic modulation of adipocyte formation for its effects in BM transplantation.

Author notes

Disclosure: No relevant conflicts of interest to declare.