Endothelial and perivascular cells maintain haematopoietic stem cells

Abstract

Several cell types have been proposed to create niches for haematopoietic stem cells (HSCs). However, the expression patterns of HSC maintenance factors have not been systematically studied and no such factor has been conditionally deleted from any candidate niche cell. Thus, the cellular sources of these factors are undetermined. Stem cell factor (SCF; also known as KITL) is a key niche component that maintains HSCs. Here, using Scf(gfp) knock-in mice, we found that Scf was primarily expressed by perivascular cells throughout the bone marrow. HSC frequency and function were not affected when Scf was conditionally deleted from haematopoietic cells, osteoblasts, nestin-cre- or nestin-creER-expressing cells. However, HSCs were depleted from bone marrow when Scf was deleted from endothelial cells or leptin receptor (Lepr)-expressing perivascular stromal cells. Most HSCs were lost when Scf was deleted from both endothelial and Lepr-expressing perivascular cells. Thus, HSCs reside in a perivascular niche in which multiple cell types express factors that promote HSC maintenance.

Figure 1. Scf^gfp is a strong loss-of-function allele and Scf is primarily expressed by perivascular cells in the bone marrow

a and b, Scf^gfp/gfp homozygous mice died perinatally and were severely anemic (n=4-20). c, Scf transcripts in livers from newborn mice by qRT-PCR (n=3). d and e, Newborn liver cellularity and HSC frequency (n=4). f, Irradiated mice (CD45.1+) were transplanted with 3×10⁵ newborn liver cells from Scf^gfp/gfp, Scf^gfp/+ or Scf^+/+ donor (CD45.2+) mice along with 3×10⁵ recipient (CD45.1+) bone marrow cells (3-4 experiments with 13-18 mice/genotype). g, Scf-GFP was expressed by rare non-haematopoietic stromal cells (n=8). h-j, GFP was primarily expressed by perivascular cells in the bone marrow of Scf^gfp/+ mice. Endothelial cells were stained with an anti-Endoglin antibody. k-n, GFP was not detected in bone-lining osteoblast lineage cells (Osteopontin) in the diaphysis (k-m) or in trabecular bone (n). o-q, Higher magnification images of a sinusoid. r-u, A CD150⁺CD48^-Lineage^- candidate HSC (arrow) localized adjacent to a GFP-expressing perivascular cell. Nuclei were stained with DAPI (in blue). All data represent mean±s.d. Two-tail student’s t-tests were used to assess statistical significance: ** p<0.01, *** p<0.001. Scale bars in (j), (m) and (n) are 50um. Scale bars in (q) and (u) are 20um.

Figure 2. Scf is required for adult HSC maintenance

a, Homozygous Scf^-/- mutant mice generated from germline recombination of the Scf^fl allele were perinatal lethal and anemic. b, Scf transcripts amplified by RT-PCR from the livers of newborn mice. c, Global deletion of Scf in Ubc-CreER; Scf^fl/fl mice led to anemia (n=5-6). d and e, Global deletion of Scf in adult mice significantly reduced cellularity and HSC frequency in bone marrow (two femurs and two tibias) and spleen (n=8-10). f, To perform a limit dilution analysis, three doses of donor bone marrow cells were competitively transplanted into irradiated mice. ELDA software (http://bioinf.wehi.edu.au/software/elda/) was used to calculate HSC frequency and statistical significance (two experiments). g, 3×10⁵ donor bone marrow cells were transplanted with 3×10⁵ recipient bone marrow cells into irradiated recipient mice (3 experiments with a total of 12-14 recipients/genotype). h, HSCs did not express Scf-GFP by flow cytometry. Δ, recombined Scf^fl allele; +, wild-type allele of Scf.

Figure 3. SCF from haematopoietic cells, osteoblasts, and Nestin-Cre-expressing stromal cells is not required for HSC maintenance

a, Vav1-Cre recombined the loxpEYFP reporter in virtually all HSCs, CD45+, and Ter119+ haematopoietic cells. b and c, Deletion of Scf from haematopoietic cells did not significantly affect bone marrow or spleen cellularity or HSC frequency (n=4). d, A competitive reconstitution assay with Vav-1-Cre; Scf^fl/-, Scf^+/- and Scf^+/+ bone marrow cells (two experiments with a total of 10 recipients/genotype). e, Col2.3-Cre recombined the loxpEYFP reporter in bone-lining osteoblast lineage cells. f, Nestin-Cre recombined the loxpEYFP reporter in rare stromal cells around larger blood vessels. g, Bone marrow and spleen cellularity and h, HSC frequency in Col2.3-Cre; Scf^fl/- mice relative to controls (n=5-6). i, A competitive reconstitution assay with Col2.3-Cre; Scf^fl/-, Scf^+/- and Scf^+/+ bone marrow cells (3-5 experiments with a total of 14-22 recipients/genotype). j, Bone marrow and spleen cellularity and k, HSC frequency in Nestin-Cre; Scf^fl/- mice relative to controls (n=5-7). l, 3×10⁵ donor bone marrow cells from Nestin-Cre; Scf^fl/- and Scf^+/+ mice gave similar levels of donor cell reconstitution in irradiated mice. Reconstitution levels from Scf^+/- cells were modestly but significantly lower (3-5 experiments with a total of 14-24 recipient mice/genotype). Δ, recombined Scf^fl allele; +, wild-type allele; −, germline deleted allele. NS, not significant. Scale bar is 100um in (e) and 50um in (f).

Figure 4. Deletion of Scf from endothelial cells depletes HSCs

a and b, Tie2-Cre recombined the loxpEYFP conditional reporter in VE-cadherin+ endothelial cells and in haematopoietic cells in the bone marrow. c, Bone marrow and spleen cellularity in Tie2-Cre; Scf^fl/- mice and littermate controls (n=4-7). d, HSC frequency in Tie2-Cre; Scf^fl/- mice and controls (n=4-7). e, HSC frequency in the liver of newborn Tie2-Cre; Scf^fl/- mice and controls (n=3-6). f, HSC frequency in one month-old Tie2-Cre; Scf^fl/- mice and controls (n=3-4). g, Bone marrow cells from Tie2-Cre; Scf^fl/- mice gave significantly lower levels of reconstitution relative to cells from Scf^+/- and Scf^+/+ mice (3-5 experiments with a total of 15-25 recipients/genotype).

Figure 5. Deletion of Scf from Lepr-Cre-expressing perivascular stromal cells depletes HSCs in the bone marrow

a-c, Lepr-Cre recombined the loxpEYFP reporter in perisinusoidal stromal cells in the bone marrow but not in bone-lining or haematopoietic cells. d, Lepr-Cre did not recombine in VE-cadherin+ endothelial cells. e, 0.013±0.009% (mean±s.d.; n=3) of bone marrow cells from Lepr-Cre; loxpEYFP mice were EYFP+. f, Spleen size and g, bone marrow and spleen cellularity (n=4-7). h, HSC frequency (n=4-7). i, Total HSC numbers (including bone marrow and spleen) in Lepr-Cre; Scf^fl/gfp mice (n=4-7). j, Limit dilution analysis of the frequency of long-term multilineage reconstituting cells in the bone marrow of Lepr-Cre; Scf^fl/gfp mice relative to controls (two experiments). k, HSC frequency in the newborn liver (n=4-11). l, HSC frequency in one month-old Lepr-Cre; Scf^fl/gfp mice and controls (n=3-6). m, Tie2-Cre; Lepr-Cre; Scf^fl/- mice had significantly reduced bone marrow cellularity and increased spleen cellularity compared to Scf^+/- or Tie2-Cre; Scf^fl/- controls (n=4-11). n, Deletion of Scf from endothelial and perivascular stromal cells in Tie2-Cre; Lepr-Cre; Scf^fl/- mice greatly depleted HSCs from adult bone marrow (n=4-11). o, Total HSC number was significantly reduced in Tie2-Cre; Lepr-Cre; Scf^fl/- mice compared to Tie2-Cre; Scf^fl/- or Lepr-Cre; Scf^fl/- mice (n=4-11). Scale bars are 50um.