AbstractVarious extrinsic signals tightly control hematopoietic stem cell quiescence. Our recent study showed that hematopoietic stem cells are regulated by a special FoxP3+ regulatory T-cell population with high expression of a hematopoietic stem cell marker, CD150. Extracellular adenosine generated via a cell-surface ectoenzyme CD39 on CD150high regulatory T cells maintained hematopoietic stem cell quiescence. It remains unclear how conventional T cells and the other cell-surface ectoenzyme, CD73, contribute to regulation of hematopoietic stem cells. This work shows that CD150high regulatory T cells as well as unique CD150high CD4+ conventional T cells regulate hematopoietic stem cells via CD73. Global CD73 deletion increased the numbers of hematopoietic stem cells, cycling stem cell frequencies, and levels of reactive oxygen species in hematopoietic stem cells. In vivo antioxidant treatment inhibited the increase of hematopoietic stem cells in CD73 knockout mice, suggesting that CD73 maintains stem cell quiescence by preventing oxidative stress. High levels of CD73 expression were frequently found on CD150high regulatory T cells and CD150high FoxP3−CD4+ T cells within the bone marrow. Transfer of these CD150high regulatory T cells and CD150high CD4+ conventional T cells abolished the increase of hematopoietic stem cells in CD73 knockout mice. In addition, the increase of stem cells in CD73 knockout mice was also inhibited by pharmacological activation of adenosine receptor 2A which is highly expressed by hematopoietic stem cells. Taken together, these results suggest that CD73 of CD150high regulatory T cells and CD150high CD4+ conventional T cells protects hematopoietic stem cells from oxidative stress, maintaining stem cell quiescence via adenosine receptor 2A.
The bone marrow (BM) microenvironment provides various cues to regulate hematopoietic stem cell (HSC) quiescence, self-renewal, and multilineage differentiation,41 and to protect HSC from various stresses, such as oxidative stress5 and toxic substances.6 Different mesenchymal subsets and megakaryocytes form a specialized regulatory zone for HSC residence, called the niche, within the BM.41 It is thought that tight control of HSC quiescence and function helps to prevent HSC exhaustion and genetic mutation. Due to a growing demand for clinical BM transplantation, understanding how the BM microenvironment regulates HSC remains important.
Our recent study demonstrated that HSC were regulated by a unique population of regulatory T cells (Treg) with high expression of a HSC marker, CD150.7 These CD150 Treg frequently localized adjacent to HSC.7 Treg-mediated HSC regulation depended on a cell-surface ectoenzyme, CD39, which was highly expressed by CD150 Treg.7 CD39 is known to convert extracellular adenosine triphosphate (ATP) and adenosine diphosphate (ADP) into adenosine monophosphate (AMP) which is further hydrolyzed by the other cell-surface ectoenzyme, CD73, into extracellular adenosine, a purine nucleotide with various tissue-protective effects.8 The results of our study using conditional knockout (KO) of CD39 in Treg suggested that extracellular adenosine generated via CD39 on Treg protected HSC from oxidative stress, maintaining HSC quiescence.7 It remains unclear how the other cell-surface ectoenzyme, CD73, contributes to HSC regulation and which BM cell populations regulate HSC via CD73. In addition, while the BM serves as a reservoir of memory T cells,109 little is known about the role of these conventional T cells in HSC regulation.
This work identified unique CD150CD4FoxP3 conventional T cells (nonTreg) which highly expressed CD73 and CD39, like CD150 Treg. Our observations in CD73 KO mice into which these CD150 T-cell populations were transferred suggest that CD73 of CD150 CD4 nonTreg and CD150 Treg maintain HSC quiescence and abundance.
C57BL/6J mice, SJL mice, BALB/c mice, CD73 KO mice, FoxP3YFP mice, and Lep-cre mice (Jackson Laboratory, Bar Harbor, ME, USA) were housed in a specific pathogen-free environment. CD39-flox mice were kindly provided by Dr. Simon C. Robson (Harvard Medical School). Seven-week old CD73 KO mice were analyzed. The mice were sacrificed by CO2 inhalation and cervical dislocation. Studies were conducted with approval from Institutional Review Boards and Animal Care and Use Committees at Columbia University.
Antibodies and reagents
We used FITC-conjugated Lineage monoclonal antibodies (B220, Mac1, GR-1, CD2, CD3a, CD8a, CD4, CD19 and Ter119), APC-780-conjugated cKit monoclonal antibodies, PECy7- or APC-conjugated CD39 monoclonal antibodies, FITC-conjugated FoxP3 monoclonal antibodies, PE-conjugated Ki67 monoclonal antibodies (all purchased from eBioscience), APC/Cy7-conjugated NK1.1 monoclonal antibodies, BV510-conjugated CD3 monoclonal antibodies, PE/Cy7-, or BV605-conjugated CD4 monoclonal antibodies, Alexa700- or Pacific blue-conjugated CD48 monoclonal antibodies, PerCP/Cy5.5-, APC-, or BV605-conjugated CD73 monoclonal antibodies, PE- or PE/Cy7-conjugated CD150 monoclonal antibodies (all from Biolegend), and BV605-conjugated Sca-1 mono clonal antibodies (from BD Pharmingen or Biolegend).
N-acetyl-L-cysteine (A9165) was purchased from Sigma-Aldrich.
Competitive reconstitution assay
SJL (CD45.1) mice were irradiated at 475 cGy twice (950 cGy in total) at least 2 h apart. Two hours after the last irradiation, donor BM cells (CD45.2), together with competitor BM cells (SJL) (3 × 10/each), were injected into the tail veins of SJL recipients. Peripheral blood samples were analyzed periodically. Red blood cells were lysed with an ammonium chloride potassium buffer. The antibodies used to analyze donor chimerism were anti-CD45.1, anti-CD45.2, anti-GR1, anti-CD11b, anti-B220, and anti-TCR-β (all from Biolegend).
Flow cytometric analysis
BM cells were isolated by crushing tibiae and femora. Following treatment with a red blood cell lysis buffer (Biolegend), the cell suspension (2×10 cells) was plated onto 96-well plates and incubated with culture media containing 2 μM CellROX Deep Red (Invitrogen) for 30 min. Flow cytometry was performed using an LSRII (BD Biosciences), LSRFortessa (BD Biosciences), or FACSCanto (BD Biosciences) cytometer followed by analysis using FlowJo software (Tree Star Inc.).
BM cells (2.0×10 cells/each well) were plated in six-well plates (Corning, NY, USA) containing 1 mL MethoCult (M3234, Stemcell Technologies Inc.) supplemented with 1% penicillin/streptomycin (Gibco), stem cell factor (50 ng/mL), interleukin-3 (15 ng/mL), interleukin-6 (20 ng/mL), and granulocyte-macrophage colony-stimulating factor (15 ng/mL). Colonies were maintained at 37°C in humidified incubators. Colony formation was scored on day 10.
Flow cytometry following intracellular staining of FoxP3
Intracellular FoxP3 staining was performed according to the manufacturer’s protocol (eBioscience).
Stromal cell analysis
For analysis of stromal cells, long bones were gently crushed using Hanks balanced saline solution to harvest the BM cells. Whole bone marrow was digested with collagenase IV (200 U/mL) and DNase I (200 U/mL) at 37°C for 30 min. Following treatment with a red blood cell lysis buffer (Biolegend), the cell suspension (2×10 cells) was plated onto 96-well plates and then stained with antibodies. Anti-CD140a (APA5), anti-CD140b (APB5), anti-CD45 (30F-11), anti-CD31 (390) and anti-Ter119 antibodies (all from Biolegend) were used to stain perivascular stromal cells.
T-cell transfer assay
HSC numbers were determined in CD73 KO mice 7 days after intravenous injection of CD150 BM Treg, CD150 BM Treg, CD150 BM nonTreg, or CD150 BM nonTreg (30,000 cells/mouse). Data were pooled from three independent experiments (4-10 mice/group).
Adenosine 2A receptor agonist treatment
CD73 KO mice were given PSB0777, a potent adenosine 2A receptor (A2AR) agonist, daily for 7 consecutive days (25 μg/mouse, intraperitoneally). Total HSC numbers in one tibia and one femur were analyzed 1 day after the final injection.
Statistical analyses were performed with GraphPad Prism software (version 6.0). Statistical significance was determined using a two-tailed t-test or one-way analysis of variance (ANOVA) with a Bonferroni post-test correction. P values less than 0.05 were considered to be statistically significant. All data are presented as mean ± standard deviation (SD).
CD73 deletion increased hematopoietic stem cell pool size
The effect of global CD73 deletion on hematopoiesis was first analyzed. CD73 KO mice showed increases in BM cellularity (Figure 1A). CD73 deletion significantly increased the frequencies of cycling cKitSca1Lin-hematopoietic stem and progenitor cells (HSPC) and CD150CD48cKitSca1Lin- HSC, as well as numbers of HSPC and HSC (Figure 1A-C, Online Supplementary Figure S1A,B and Online Supplementary Table S1). Consistently, the numbers of colonies formed following in vitro culture of BM cells isolated from CD73 KO mice were significantly higher than those from wild-type mice (Online Supplementary Figure S1C). The numbers of other BM cell populations were not significantly altered by CD73 deletion (Online Supplementary Figure S1D). To assess the size of the pool of functional HSC and HSPC, we performed competitive BM transplantation assays to evaluate the reconstituting potential of BM cells. BM cells of CD73 KO mice or control wildtype mice (B6 CD45.2; 3 × 10 cells/mouse) were intravenously injected into lethally-irradiated B6 SJL mice (CD45.1), together with competitor SJL BM (CD45.1; 3 × 10 cells/mouse). CD45.2 donor blood chimerism from CD73 KO BM cells remained significantly higher than that from control BM cells for 6 months after transplantation, suggesting that CD73 deletion increased functional HSC and HSPC frequencies (Figure 1D). Additionally, no myeloid skewing was observed in donor hematopoietic cells derived from CD73KO BM (Online Supplementary Figure S1E). Taken together, these results indicate that CD73 maintains quiescence and pool size of HSPC and HSC.
CD73 maintains hematopoietic stem cell pool size by preventing oxidative stress
CD73 KO mice showed slight but significant increases in the levels of reactive oxygen species (ROS) in HSPC and HSC but not in Lin cells (Figure 2A, Online Supplementary Figure S2A), suggesting that CD73 prevents oxidative stress against HSC and HSPC but not against mature cells. To test whether CD73 maintains HSC quiescence in a ROS-dependent manner, we used treatment with an antioxidant, N-acetylcysteine (NAC), which reversed the increases in HSC numbers and reconstituting potential of BM cells in CD73 KO mice (Figure 2B,C). These results indicate that CD73 maintains HSC quiescence by preventing oxidative stress.
This ROS-mediated expansion of HSC pool size in CD73 KO mice was consistently observed in our recently reported study, using two models: (i) mice with conditional deletion of CD39 in Treg; and (ii) mice with reduction of BM Treg achieved by CXCR4 deletion in Treg.7 In contrast, some previous studies showed that increased ROS levels in HSC led to loss of HSC quiescence, and HSC exhaustion.1311 HSC proliferation in our models is likely explained by moderate increases in ROS levels (1.2- to 1.5-fold) compared to greater increases in other models (3- to 5-fold).1311 Indeed, some studies showed that a moderate increase in ROS levels induced HSC proliferation.1514 Moreover, the peripheral blood of CD73 KO mice showed a non-significant trend toward myeloid skewing (Online Supplementary Figure S2B), which may also reflect a moderate increase in ROS levels in HSC.
CD73high cells were frequently found in CD150high regulatory T cells and CD150highCD4+ non-regulatory T cells
To identify cell populations which play important roles in CD73-mediated HSC regulation, flow cytometric analysis was performed to measure CD73 and CD39 expression levels on hematopoietic cell populations within the BM. Intermediate to high expression of CD39 was found on the following hematopoietic cell populations: HSC; HSPC; CD11bGr1 cells; CD11bGr1 cells; B220 B cells; CD4FoxP3 cells (CD4 nonTreg); CD4FoxP3 Treg; CD8 T cells; CD4CD3NK1.1 NKT cells; and NK cells (Online Supplementary Figure S3A). In contrast, high levels of CD73 expression were mainly observed within CD4 T cells (Treg, CD4 nonTreg) (Figure 3A,B). While CD8 T cells, CD11bGr1 cells, and CD4 NKT cells showed intermediate levels of CD73 expression, CD73 was not expressed by HSC, B cells, or CD11bGr1 myeloid cells (Figure 3A,B).
As previously reported,7 CD73 cells among Treg were predominantly CD150, showing equivalent levels of expression of CD150 as those of HSC (Figure 3B-D). These CD150 Treg also highly expressed CD39 (Figure 3D, Online Supplementary Figure S3A). Notably, there were also CD150 fractions among BM CD4 nonTreg, which highly expressed CD39 and CD73 as compared to the rest of the CD4 nonTreg (CD150) (Figure 3B-D, Online Supplementary Figure S3B). CD150 populations comprised 20% of CD4 nonTreg and 40% of Treg (Figure 3E). CD150 nonTreg frequently showed a CD44CD62L effector memory phenotype as compared to the rest of the CD4 nonTreg (CD150) (Figure 3F; Online Supplementary Figure S3C), which is consistent with our previous observations in CD150 Treg.7 The frequencies of CD39 and CD73 cells among CD150 nonTreg were equivalent to those among CD150 Treg, and higher than those in BM CD150 Treg, BM CD150CD4 nonTreg, and lymph node CD4 nonTreg (Figure 3B,D, Online Supplementary Figure S3A,D).
CD39 and CD73 expression among CD45 mesenchymal cells was further analyzed following division of the CD45 cells into the following three populations; CD45-CD31 vasculature; CD45CD31CD140aCD140b cells; and CD45CD31CD140aCD140b cells (Online Supplementary Figure S3E). Previous studies suggested that the former two populations were putative cellular constituents of the HSC niche.32 CD45CD140aCD140b cells were shown to overlap exclusively with leptin receptor-positive (lepr) perivascular niche cells.32 CD39 and CD73 expression was observed on CD31 vasculature and CD45-CD31CD140aCD140b cells, but not on CD45CD31-CD140aCD140b cells (Figure 3G,H, Online Supplementary Figure S3F). The frequencies of CD39 and CD73 cells within CD31 vasculature and CD45CD140aCD140b cells were comparable to those in CD150 Treg, and lower than those in CD150 Treg and CD150 nonTreg (Figure 3B,G-H, Online Supplementary Figure S3A). Taken together, these observations indicate that, while various BM cell populations expressed CD39 and/or CD73, CD73 cells were frequently found within CD150 Treg and CD150 nonTreg.
The increase of hematopoietic stem cells in CD73 knockout mice was reversed by transfer of CD150high regulatory T cells and CD150high non-regulatory T cells and by in vivo adenosine receptor agonist treatment
To assess the role of CD150 Treg and CD150 CD4 nonTreg in CD73-mediated HSC regulation, we analyzed how transfer of these T-cell populations influences HSC. Transfer of CD150 Treg and of CD150 CD4 nonTreg significantly reversed the increase of HSC in CD73 KO mice relative to control wildtype mice. In contrast, the size of the HSC pool was not significantly altered by transfer of CD150 nonTreg or CD150 Treg (Figure 4A). These observations are consistent with our previous observations that transfer of CD150 Treg reversed the increase of HSC in mice with conditional deletion of CD39 in Treg.7 These results suggest that CD150 Treg and CD150 nonTreg contribute largely to CD73-mediated HSC regulation.
The downstream signaling of CD73 in HSC regulation was further assessed. HSC showed higher levels of expression A2AR than various other BM cell populations (Figure 4B). In vivo A2AR agonist treatment significantly reversed the increase of HSC in CD73 KO mice (Figure 4C). These observations suggest that A2AR signaling plays an important role in CD73-mediated HSC regulation.
This study identified CD73 and CD150 nonTreg as important regulators of HSC quiescence and abundance in the adult BM. Global CD73 deletion increased ROS levels in HSC, and HSC numbers. This increase of HSC was reversed by anti-oxidant treatment, suggesting that CD73 maintains HSC quiescence by preventing oxidative stress. Because HSC did not express CD73 but CD39, CD73-mediated HSC regulation is driven by the microenvironment. While various BM cells showed intermediate levels of CD73 expression, CD73 cells were frequently found within unique CD150 Treg and CD150 nonTreg. Transfer of these CD150 Treg and CD150 nonTreg, but not of CD150 Treg or CD150 nonTreg, reversed the increase of HSC in CD73 KO mice. Additionally, pharmacological activation of A2AR, highly expressed by HSC, reversed the increase of HSC in CD73 KO mice. Taken together, these results suggest that CD73 of CD150 Treg and CD150 nonTreg regulates HSC quiescence and abundance via A2AR.
To the best of our knowledge, this is the first study showing the role of conventional T cells in HSC regulation. This work is complemented by our recent study7 showing that CD39 on CD150 Treg played a critical role in maintaining HSC quiescence. As both CD150 Treg and CD150 nonTreg frequently displayed an effector memory T-cell phenotype,7 the observations of our current and previous studies7 suggest that BM CD4 memory T cells and memory Treg coordinate each other to generate extracellular adenosine via CD39 and CD73, maintaining HSC quiescence. As the BM is known to be a site to which memory T cells frequently home and in which they are maintained,109 memory T cells and Treg generated following infection may play important roles in protecting BM HSC from oxidative and inflammatory stresses, controlling hematopoiesis. As CD150 Treg frequently localized adjacent to HSC, a future histological analysis is warranted to identify the spatial distribution of CD150 nonTreg (CD3CD4NK1.1FoxP3) with respect to HSC (CD150CD48CD41Lin), although such a study is technically challenging because of the requirement of multiple colors.
Our study does not rule out the possibility that HSC are regulated by other adenosine receptors or by P2 receptors that bind ATP metabolized by CD39. HSC expression of CD39, but not of CD73, may reflect the possibility that tight control of ATP/adenosine ratios is required for the maintenance of HSC quiescence. Indeed, a previous study showed that global P2YR deletion abrogated the radioresistance of HSC.16 However, under normal conditions, P2YR KO mice did not show significant alteration of HSC numbers,16 suggesting that the observed phenotypes in CD73 KO mice and FoxP3 CD39 mice under normal conditions were not attributable to P2YR.
CD150 nonTreg and CD150 Treg are likely to generate adenosine in concert with various CD39 or CD73 BM cell populations, including HSC and the following two niche constituents: CD31 vasculature and CD140aCD140bCD45 mesenchymal cells which exclusively overlap with lepr perivascular cells.32 Nevertheless, these two niche constituents are unlikely to be the major source of adenosine, because the frequencies of CD39 and CD73 cells in these mesenchymal cells were comparable to those in CD150 Treg and transfer of these latter cells did not alter HSC numbers in CD73 KO mice. Indeed, our additional study using leprcre CD39 mice showed that conditional deletion of CD39 in lepr cells did not alter HSC number or reconstituting potential of BM cells (Online Supplementary Figure S4A-C). This observation further supports the important role of CD150 Treg and CD150 nonTreg in adenosine-mediated HSC regulation.
In summary, this work showed that CD150 Treg and CD150 nonTreg maintain HSC quiescence via CD73. An examination of the roles of adenosine and memory T cells in human hematopoiesis and transplantation is warranted.
This work was supported by NIH NHLBI R01HL129506 (JF), an ASH Junior Faculty Scholar Award (JF), a Schaefer Research Scholar Award (JF) and an Uehara Memorial Foundation Research Fellowship Award (YH). Research reported in this publication was performed in the CCTI Flow Cytometry Core, supported in part by the Office of the Director, National Institutes of Health under award S10OD020056. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
- ↵Check the online version for the most updated information on this article, online supplements, and information on authorship & disclosures: www.haematologica.org/content/104/6/1136
- Received May 23, 2018.
- Accepted December 10, 2018.
- Kunisaki Y, Bruns I, Scheiermann C. Arteriolar niches maintain haematopoietic stem cell quiescence. Nature. 2013; 502(7473):637-643. PubMedhttps://doi.org/10.1038/nature12612Google Scholar
- Ding L, Saunders TL, Enikolopov G, Morrison SJ. Endothelial and perivascular cells maintain haematopoietic stem cells. Nature. 2012; 481(7382):457-462. PubMedhttps://doi.org/10.1038/nature10783Google Scholar
- Ding L, Morrison SJ. Haematopoietic stem cells and early lymphoid progenitors occupy distinct bone marrow niches. Nature. 2013; 495(7440):231-235. PubMedhttps://doi.org/10.1038/nature11885Google Scholar
- Itkin T, Gur-Cohen S, Spencer JA. Distinct bone marrow blood vessels differentially regulate haematopoiesis. Nature. 2016; 532(7599):323-328. PubMedhttps://doi.org/10.1038/nature17624Google Scholar
- Jang YY, Sharkis SJ. A low level of reactive oxygen species selects for primitive hematopoietic stem cells that may reside in the low-oxygenic niche. Blood. 2007; 110(8):3056-3063. PubMedhttps://doi.org/10.1182/blood-2007-05-087759Google Scholar
- Arai F, Hirao A, Ohmura M. Tie2/angiopoietin-1 signaling regulates hematopoietic stem cell quiescence in the bone marrow niche. Cell. 2004; 118(2):149-161. PubMedhttps://doi.org/10.1016/j.cell.2004.07.004Google Scholar
- Hirata Y, Furuhashi K, Ishii H. CD150(high) bone marrow tregs maintain hematopoietic stem cell quiescence and immune privilege via adenosine. Cell Stem Cell. 2018; 22(3):445-453. https://doi.org/10.1016/j.stem.2018.01.017Google Scholar
- Jacobson KA, Gao ZG. Adenosine receptors as therapeutic targets. Nat Rev Drug Discov. 2006; 5(3):247-264. PubMedhttps://doi.org/10.1038/nrd1983Google Scholar
- Mazo IB, Honczarenko M, Leung H. Bone marrow is a major reservoir and site of recruitment for central memory CD8+ T cells. Immunity. 2005; 22(2):259-270. PubMedhttps://doi.org/10.1016/j.immuni.2005.01.008Google Scholar
- Di Rosa F, Gebhardt T. Bone marrow T cells and the integrated functions of recirculating and tissue-resident memory T cells. Front Immunol. 2016; 7:51. Google Scholar
- Abbas HA, Maccio DR, Coskun S. Mdm2 is required for survival of hematopoietic stem cells/progenitors via dampening of ROS-induced p53 activity. Cell Stem Cell. 2010; 7(5):606-617. PubMedhttps://doi.org/10.1016/j.stem.2010.09.013Google Scholar
- Miyamoto K, Araki KY, Naka K. Foxo3a is essential for maintenance of the hematopoietic stem cell pool. Cell Stem Cell. 2007; 1(1):101-112. PubMedhttps://doi.org/10.1016/j.stem.2007.02.001Google Scholar
- Liu J, Cao L, Chen J. Bmi1 regulates mitochondrial function and the DNA damage response pathway. Nature. 2009; 459(7245):387-392. PubMedhttps://doi.org/10.1038/nature08040Google Scholar
- Juntilla MM, Patil VD, Calamito M. AKT1 and AKT2 maintain hematopoietic stem cell function by regulating reactive oxygen species. Blood. 2010; 115(20):4030-4038. PubMedhttps://doi.org/10.1182/blood-2009-09-241000Google Scholar
- Lewandowski D, Barroca V, Duconge F. In vivo cellular imaging pinpoints the role of reactive oxygen species in the early steps of adult hematopoietic reconstitution. Blood. 2010; 115(3):443-452. PubMedhttps://doi.org/10.1182/blood-2009-05-222711Google Scholar
- Cho J, Yusuf R, Kook S. Purinergic P2Y(1)(4) receptor modulates stress-induced hematopoietic stem/progenitor cell senescence. J Clin Invest. 2014; 124(7):3159-3171. PubMedhttps://doi.org/10.1172/JCI61636Google Scholar