Perivascular support of human hematopoietic stem/progenitor cells

Abstract

Hematopoietic stem and progenitor cells (HSPCs) emerge and develop adjacent to blood vessel walls in the yolk sac, aorta-gonad-mesonephros region, embryonic liver, and fetal bone marrow. In adult mouse bone marrow, perivascular cells shape a "niche" for HSPCs. Mesenchymal stem/stromal cells (MSCs), which support hematopoiesis in culture, are themselves derived in part from perivascular cells. In order to define their direct role in hematopoiesis, we tested the ability of purified human CD146(+) perivascular cells, as compared with unfractionated MSCs and CD146(-) cells, to sustain human HSPCs in coculture. CD146(+) perivascular cells support the long-term persistence, through cell-to-cell contact and at least partly via Notch activation, of human myelolymphoid HSPCs able to engraft primary and secondary immunodeficient mice. Conversely, unfractionated MSCs and CD146(-) cells induce differentiation and compromise ex vivo maintenance of HSPCs. Moreover, CD146(+) perivascular cells express, natively and in culture, molecular markers of the vascular hematopoietic niche. Unexpectedly, this dramatic, previously undocumented ability to support hematopoietic stem cells is present in CD146(+) perivascular cells extracted from the nonhematopoietic adipose tissue.

Figure 1

In situ expression of hematopoietic niche markers by human perivascular cells. (A-D) Immunohistochemistry performed on paraffin-embedded sections of 17-week-old human FBM. Pericytes surrounding microvessels express (A) CD146, (B) nestin, (C) CXCL12, and (D) leptin receptor (Lep-R) (original magnification, ×63). (E-S) Immunohistochemistry performed on human adipose tissue cryosections. (E-G) VWF-positive endothelial cells (green) are surrounded by perivascular cells expressing (E) nestin, (F) CXCL12, and (G) Lep-R. (H-S) Triple-staining immunohistochemistry performed on human adipose tissue cryosections shows coexpression of CD146 with (K-M) nestin, (N-P) CXCL12, and (Q-S) LepR. (H-J) Single staining with anti-VWF antibody followed by incubation with conjugated IgG controls revealed the lack of autofluorescence (original magnification, ×40). IgG, immunoglobulin G.

Figure 2

Isolation and culture of MSCs and stromal subsets from lipoaspirate. (A) SVF was obtained from human lipoaspirate specimens (n = 4 donors). An aliquot of SVF was directly seeded in tissue-culture plates for the isolation of conventional MSCs by plastic adherence. Another aliquot of SVF was processed for FACS sorting of DAPI⁻CD45⁻CD34⁻CD146⁺ perivascular cells and DAPI⁻CD45⁻CD34⁺CD146⁻ cells. (B) FACS analysis of cultured fat-derived MSCs, CD146⁺ perivascular cells, and CD146⁻ cells. After 9 passages in culture, MSCs retain a low percentage of CD146⁺ cells, while purified CD146⁺ perivascular cells and CD146⁻ cells retain a stable phenotype homogeneously positive and negative for CD146, respectively.

Figure 3

Cultured CD146⁺ perivascular cells express markers of hematopoietic perivascular niche cells. (A-B) Ex vivo–expanded CD146⁺ perivascular cells purified from fat and FBM similarly express higher levels of mRNA of perivascular cell markers when compared with MSCs and CD146⁻ cells (n = 2 donors for each tissue). (C-N) Fat and FBM-derived CD146⁺ perivascular cells similarly and almost exclusively express (C-F) nestin and (G-J) CXCL12 in culture compared with CD146⁻ cells. (K-N) No difference in Lep-R expression was observed between CD146⁺ and CD146⁻ cells from either fat and FBM (original magnification, ×20). mRNA, messenger RNA.

Figure 4

CD146⁺ perivascular cells promote ex vivo maintenance of undifferentiated HSPCs. (A) In the absence of cytokines and stromal cell feeder layer (No feeder), CD45⁺ hematopoietic cells cultured in RN-treated wells rapidly died within the first 2 weeks of culture. At any time of culture, the total number of CD45⁺ cells recovered from CD146⁺ cell cocultures was significantly higher when compared with MSC cocultures (n = at least 5 independent experiments for each time point, each experiment was performed in triplicate; **P < .01, ***P < .001). (B) A similar pattern was observed for the total number of CD34⁺ cells (n = at least 5 independent experiments for each time point, each experiment was performed in triplicate; ***P < .001). (C) Representative FACS analysis after 2 weeks of coculture of CB CD34⁺ cells with MSCs or CD146⁺ cell cocultures. After gating on CD45⁺CD10⁻CD19⁻ cells, CD34⁺33⁻ cells were defined as CD34⁺Lin⁻ cells (black box). (D) The absolute number of CD34⁺Lin⁻ cells was significantly higher in CD146⁺ cell cocultures, compared with MSC cocultures, at any time of culture (n=at least 5 independent experiments for each time point, each experiment was performed in triplicate; **P < .01, ***P < .001). (E-F) Coculture of CB CD34⁺ cells with MSCs led to a significantly higher frequency of CD14⁺ myeloid cells after 2 weeks (E) (40.24% ± 2.723% vs 26.67% ± 2.075%. n = 10 independent experiments, each experiment was performed in triplicate; ***P < .0001) and a higher frequency of CD10⁺/CD19⁺ lymphoid progenitors or mature cells after 4 weeks of coculture (F) (5.155% ± 1.918% vs 0.9541% ± 0.2564%, n = 8 independent experiments, each experiment was performed in triplicate; *P < .05). No difference in the absolute numbers of myeloid and lymphoid cells was observed between CD146⁺ cell and MSC cocultures. All data are presented as mean ± SEM.

Figure 5

CD146⁺ perivascular cells but not MSCs sustain functional HSPCs with engraftment potential and self-renewal ability. (A) CFU assay revealed significantly higher number of CFUs in CD146⁺ cell cocultures after 1, 2, 4, and 6 weeks of coculture as compared with MSC cocultures (n = 3 independent experiments, each experiment was performed in triplicate; *P < .05, ***P < .001). (B) Representative flow cytometry analysis for the detection of human CD45⁺HLA⁺ cells in bone marrow of NSG mice 6 weeks posttransplantation with phosphate-buffered saline, or with the same number of CD45⁺ cells (10⁵) harvested after 2 weeks of CB CD34⁺ cell coculture with MSCs or CD146⁺ cells. (C) All mice injected with CD45⁺ cells obtained from CD146⁺ cell cocultures showed human engraftment whereas no engraftment was ever detected (ND) in mice that received MSC cocultures (n = 3 independent experiments, n = 11 mice per group; ***P < .0001). (D) Frequency of CD34⁺ progenitors, CD19⁺ lymphoid, and CD14⁺ myeloid cells within the CD45⁺HLA⁺ population of cells in the bone marrow of chimeric mice. (E) Human CD45⁺HLA⁺ hematopoietic cells were also detected 6 weeks posttransplantation in the contralateral tibia of mice injected with HSPCs cocultured with CD146⁺ perivascular cells. (F-I) Representative flow cytometry analysis of secondary host bone marrow. (F) Bone marrow from primary hosts transplanted with MSC coculture was injected in secondary hosts as a negative control. (G) Human engraftment was observed 4 weeks after secondary transplantation of bone marrow from chimeric mice transplanted with CD146⁺ cell coculture (n = 3 engrafted mice of 4). (H) Both CD19⁺ lymphoid and CD33/CD14/CD15⁺ myeloid cells were detectable within the human CD45⁺ engrafted hematopoietic cells in secondary hosts. (I) Quantification of the level of chimerism in secondary mice. All data are presented as mean ± SEM.

Figure 6

CD146⁺ perivascular cells induce Notch activation in hematopoietic cells. (A) Immunocytochemical staining for Jagged-1 (JAG1, red), CD146 (green) and nuclei (DAPI, blue) performed on fat-derived CD146⁺ perivascular cells and MSCs (original magnification, ×20). White arrows indicate clusters of cells within the MSC that coexpress JAG1 and CD146. (B) Western blot analysis showing significantly higher expression of Jagged-1 in CD146⁺ perivascular cells compared with MSCs derived from fat. (C) qPCR analysis revealed that fat-derived MSCs and CD146⁺ perivascular cells express multiple Notch ligands. (D-E) Quantification of (D) CD45⁺ and (E) CD34⁺ hematopoietic and progenitor cells with activated Notch pathway (CD45⁺NICD⁺) after 1 week of coculture with fat or FBM-derived CD146⁺ perivascular cells, MSCs, and CD146⁻ cells (n = 3 independent experiments, n = 40 random fields analyzed; ***P < .0001). Data are presented as mean ± SEM.

Figure 7

Notch inhibition affects survival and B-cell differentiation of HSPCs. Inhibition of Notch was achieved by addition of 10μM DAPT to CD146⁺ perivascular cells and CB CD34⁺ cell coculture every other day. Vehicle (DMSO) was added to control cocultures. (A-B) Total number of CD45⁺ cells and CD34⁺Lin⁻ cells was significantly reduced after 2 weeks of coculture with DAPT (5.03 ± 0.54 × 10⁴ vs 3.02 ± 0.37 × 10⁴ CD45⁺ cells, n = 4 independent experiments, each experiment was performed in triplicate, **P < .01; 1.5 ± 0.16 × 10⁴ vs 0.82 ± 0.12 × 10⁴ CD34⁺Lin⁻ cells, n = 4 independent experiments, each experiment was performed in triplicate, **P < .01). (C) Similarly, the total number of CFUs was significantly reduced after 4 weeks of coculture with DAPT (478.3 ± 112.4 vs 191.0 ± 43.28, n = 3 independent experiments, each experiment was performed in triplicate; * P < .05). (D) Flow cytometry viability analysis revealed a significantly higher frequency of PI⁺ dead cells in coculture performed in the presence of DAPT (13.08% ± 1.13% vs 19.94 ± 1.31, n = 4 independent experiments, each experiment was performed in triplicate, *** P < .0001). (E) Notch inhibition also significantly increased B-cell development (0.13 ± 0.04 × 10³ vs 1.72 ± 0.55 × 10³ of lymphoid cells, n = 3 individual experiments, each experiment performed in triplicate; **P < .01). Data are presented as mean ± SEM.