F4/80 Identifies a Subset of Non-Mobilizable Bone Marrow HSCs Involved in Stress-Induced Hematopoiesis

Hematopoietic stem cell (HSCs) are commonly isolated by using cell surface markers in order to study their hierarchy and functional properties. However, even with the most rigorous methods of isolation, subsets of HSCs likely exhibit functional heterogeneity. We have found that F4/80, an adhesion G-protein coupled receptor well known as a macrophage marker, was expressed on an HSC subpopulation (50-60% of Lin⁻ Sca1⁻ cKit⁻ CD150⁺ CD34⁻ CD135⁻ cells) in the bone marrow. Interestingly, F4/80 was not expressed on HSCs that have egressed in the blood and on only a small fraction (3%) of splenic HSCs. To evaluate the function of the HSC subset expressing F4/80, we transplanted competitively 200 F4/80⁺ or F4/80⁻ HSCs using CD45.1/2 congenic system. We found that F4/80⁺ HSCs exhibited a lower engraftment potential compared to F4/80⁻ HSCs at 16 weeks after transplantation (19.7±4.8% and 37.1±4.5% donor contribution, respectively, P=0.025), although both HSC subsets were able to sustain mixed chimerism for myeloid, B and T cells without significant alteration in lineage bias. Since F4/80⁺ HSCs were not found in extramedullary tissues (blood or spleen), we tested whether they could be mobilized following G-CSF or CXCR4 antagonist (AMD3100) treatment. Interestingly, F4/80⁺ HSCs were still retained in the bone marrow after either G-CSF- or AMD3100-induced mobilization whereas F4/80⁻ HSCs were mobilized efficiently. However, the frequency of bone marrow F4/80⁺ HSCs was reduced in G-CSF-treated animals (3.3-fold, p<0.0001) while the F4/80⁻ HSC frequency was increased (2.2-fold, p<0.0001). F4/80 could act directly to retain HSCs in the bone marrow microenvironment or it could mark a non-mobilizable pool of stem cells. To test the later possibility, we crossed transgenic mice expressing Cre recombinase knocked in the CD169 locus, a marker of bone marrow macrophage, with ROSA26-loxP-stop-loxP-tdTomato (CD169/tomato). We found that CD169/tomato selectively labelled a large fraction of F4/80⁺ HSCs (31.7±8.4%) by contrast to F4/80⁻ HSCs which were, by and large, not labelled (2.1±1.2%). We next induced HSC mobilization with G-CSF. Strikingly, we found that in a manner similar to F4/80⁺ HSCs, CD169/tomato⁺ HSCs were not mobilized in the blood while their numbers were reduced in the bone marrow after G-CSF (5-fold, p=0.014). Since macrophage depletion can induce HSC mobilization (Chow et al., JEM 2011), we tested the effect of macrophage depletion using clodronate liposomes. Interestingly, F4/80⁺ HSCs in wild-type mice or CD169/tomato⁺ HSCs were not mobilized following macrophage depletion but were depleted from the bone marrow while F4/80⁻ HSCs underwent a 3-fold expansion. The reduced HSC numbers in marrow upon mobilization by G-CSF suggested a role in survival or proliferation. To evaluate proliferation status at steady state, we performed cell cycle analyses and BrdU incorporation assay which revealed that F4/80⁺ HSCs significantly more proliferative than F4/80⁻ HSCs either by cell cycle analyses (3.7-fold increase in non-G0 phase, p=0.024) or by BrdU incorporation (2-fold increase, p=0.026). We next assessed the impact of 5-flurouracil (5FU) administration a chemotherapeutic agent that kills cycling cells and induce stress hematopoietic recovery. While both F4/80⁺ and F4/80⁻ HSCs were depleted 4 days after 5FU treatment, we observed that F4/80⁺ HSCs expanded dramatically compared to F4/80⁻ HSCs (10-fold and 4.5-fold over F4/80⁻ HSCs at 8 days and 12 days post-5FU, respectively) during recovery phase. Taken together, these results identify F4/80⁺ HSCs as a strictly resident subset of bone marrow HSCs involved in rapid recovery after hematopoietic stress.