Institute of Cellular and Molecular Radiation Biology, Department of Genetic Instability, Recombination and Repair, Commissariat à l'Energie Atomique, Fontenay-aux-Roses, France and UMR 217 (UMR CEA-Centre National de la Recherche Scientifique)Commissariat à l'Energie Atomique, Fontenay-aux-Roses, France
Author contributions: A.J.S.: Conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing; J.N., P.V.: Conception and design, collection and/or assembly of data, data analysis and interpretation; V.B.: Provision of study material or patients; P.L.: Manuscript writing; D.T.-L.: Conception and design, data analysis and interpretation, manuscript writing.
First published online in STEM CELLSExpress March 19, 2009
Ionizing radiation (IR) exposure causes rapid and acute bone marrow (BM) suppression that is reversible for nonlethal doses. Evidence is accumulating that IR can also provoke long-lasting residual hematopoietic injury. To better understand these effects, we analyzed phenotypic and functional changes in the stem/progenitor compartment of irradiated mice over a 10-week period. We found that hematopoietic stem cells (HSCs) identified by their repopulating ability continued to segregate within the Hoechst dye excluding “side population (SP)” early after IR exposure. However, transient phenotypic changes were observed within this cell population: Sca-1 (S) and c-Kit (K) expression levels were increased and severely reduced, respectively, with a concurrent increase in the proportion of SPSK cells positive for established indicators of the presence of HSCs: CD150 and CD105. Ten weeks after IR exposure, expression of Sca-1 and c-Kit at the SP cell surface returned to control levels, and BM cellularity of irradiated mice was restored. However, the c-Kit+Sca-1+Lin−/low (KSL) stem/progenitor compartment displayed major phenotypic modifications, including an increase and a severe decrease in the frequencies of CD150+Flk2− and CD150−Flk2+ cells, respectively. CD150+ KSL cells also showed impaired reconstituting ability, an increased tendency to apoptosis, and accrued DNA damage. Finally, 15 weeks after exposure, irradiated mice, but not age-matched controls, allowed engraftment and significant hematopoietic contribution from transplanted congenic HSCs without additional host conditioning. These results provide novel insight in our understanding of immediate and delayed IR-induced hematopoietic injury and highlight similarities between HSCs of young irradiated and old mice. STEM CELLS 2009;27:1400–1409
Hematopoietic stem cells (HSCs) are a rare self-renewing bone marrow (BM) cell population that maintains blood and the immune system throughout life [1, 2]. Under nonstress conditions, young adult murine HSCs lack expression of several cell surface lineage markers normally found on differentiating or mature blood cells while displaying high levels of Sca-1 and c-Kit. They are part of the stem/progenitor compartment commonly referred to as c-Kit+Sca-1+Lin−/low (KSL) . However, only few KSL cells are functional HSCs with the capacity to sustain long-term hematopoiesis upon transplantation into lethally irradiated recipients. The recent use of additional markers, including Endoglin (CD105), Slamf1 (CD150), and Flk2 (CD135) improved the purity of HSCs isolated from BM cells by cell-surface staining [4–8].
Alternative schemes to purify HSCs are based on their unique dye efflux properties. The ATP-binding cassette (ABC) transporter Bcrp1/ABCG2 enables partial exclusion of Hoechst 33342 fluorescent dye, which makes stem cells appear in a specific pattern, called side population (SP), when analyzed at two distinct wavelengths by flow cytometry [9, 10].
Under steady-state conditions, HSCs cycle slowly in an asynchronous manner to ensure BM homeostasis [11–13]. However, their proliferation rates can be increased to replenish hematopoietic or peripheral blood (PB) cell compartments after insults such as bleeding, cytotoxic agent exposure, or irradiation [13–17].
In contrast to other hematological stresses, γ-ray radiation exposure causes immediate damage to all intermediate subsets of the hematopoietic process, with various effects respective to their status in the hematopoietic hierarchy [18–20]. Although HSCs appear more resistant than most committed progenitors and precursors [20, 21], it was recently shown that ionizing radiation (IR) can induce early apoptosis and long-lasting senescence among KSL stem/progenitor cells . Intriguingly, in this latter study, no KSL cells were detected in the BM of sublethally irradiated mice 3 days after exposure, which is paradoxical because the hematopoietic system can recover from such nonlethal insults. This result requires further investigations to identify and characterize radioresistant cells with replenishment ability early after irradiation.
Delayed consequences of IR exposure are major concerns for patients undergoing radiotherapy and for individuals who are irradiated accidentally. Comprehensive understanding of the HSC response to IR exposure may be valuable for identifying the mechanisms underlying long-lasting deleterious hematopoietic effects induced by radiation exposure. At a more fundamental level, IR, through perturbation of the whole hematopoietic system, may reveal a precious tool to gain further insight into HSC biology. This outcome is best exemplified by the identification and characterization of colony-forming units-spleen by McCulloch and Till  while investigating IR sensitivity of BM cells.
In the present study, we took advantage of the SP assay to identify a radioresistant subset of HSCs early after IR exposure. We demonstrated transient phenotypic changes at their surface, including an increase in the Sca-1 expression level and a dramatic decrease in that of c-Kit. Ten weeks after IR exposure, although BM recovery was seemingly complete, a phenotypic imbalance within the KSL stem/progenitor compartment was observed: CD150+Flk2− and CD150−Flk2+ cell frequencies were increased and dramatically reduced, respectively. CD150+ HSCs also showed impaired reconstituting ability, an increased tendency for apoptosis, and accrued DNA damage accumulation. Finally, transplanted HSCs harvested from nonirradiated congenic donors could engraft and significantly contribute to host hematopoiesis in irradiated recipients several weeks after exposure but not in control recipients. Together, these results allow better understanding of the early and delayed hematopoietic response to the radiation insult and widen the spectrum of long-lasting IR-induced injuries to HSCs.
MATERIALS AND METHODS
Mice and IR Exposure Procedure
C57BL/6-Ly5.2 and C57BL/6-Ly5.1 mice were purchased from Charles River Laboratories (Les Oncins, France, http://www.criver.com) and maintained at the CERFE and Commissariat à l'Energie Atomique (CEA) facilities (Evry and Fontenay-aux-Roses, France). C57BL/6-TgN β-actin enhanced green fluorescent protein (eGFP) mice  were bred and maintained at the Fontenay-aux-Roses CEA facility. Mice were exposed to various doses of total body irradiation (TBI) in an IBL 137Cs gamma-irradiator at an average rate of 0.5 Gy/minutes. All mice were used between 8 and 12 weeks of age, in accordance with French regulations and with consent of the local ethics committee.
Preparation and Staining of BM Cells
BM cells were obtained by flushing the femurs and tibias of young adult mice. Erythrocytes were eliminated using Red Blood Cell Lysing Buffer (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com). The remaining cells were filtered through 70-μm nylon mesh and then stained with the indicated combinations of antibodies. Staining of SP cells was performed as described previously . Antibody staining was done for 15 minutes on ice; after the final round of staining and washing, cells were resuspended in phosphate-buffered saline (PBS) or Hanks' balanced salt solution (HBSS) containing 1 μg/ml 7-aminoactinomycin D or 1 μg/ml Hoechst 33342 (Sigma-Aldrich) to allow for dead cell exclusion.
Antibodies, Phenotypic Analysis, and Isolation of BM Mononuclear Cells by Flow Cytometry
Biotin-conjugated lineage-specific antibodies included anti-CD3ε (145-2C11), anti-B220 (RA3-6B2), anti-Mac-1 (M1/70), anti-Gr1 (RB6-8C5), and anti-Ter 119 (BD Pharmingen, San Diego, http://www.bdbiosciences.com/index_us.shtml). To characterize HSCs and their progenies, we used anti-Sca-1 (E13-161.7 and D7), anti-Flk2/CD135 (A2F10.1), anti-c-Kit/CD117 (2B8), anti-endoglin/CD105 (MJ7/18), anti-Slamf1/CD150 (TC15-12F12.2), anti-CD62P (RB40.34), anti-CD45 (30-F11), anti-CD45.1 (A20), and anti-CD45.2 (104). Antibodies were biotinylated or conjugated to fluorescein isothiocyanate, phycoerythrin (PE), PE-Cy5, PE-Cy5.5, PE-Cy7, or allophycocyanin (APC), depending on the staining strategy. Secondary reagents were APC-Cy7- and PE-conjugated streptavidin. All reagents were from BD Pharmingen, BioLegend (San Diego, CA, http://www.biolegend.com), and eBioscience, Inc. (San Diego, CA, http://www.ebioscience.com).
Samples were analyzed and sorted using a Cyan MLE analyzer and a MoFlo cell sorter (Dako, Glostrup, Denmark, http://www.dako.com). BD CompBeads (BD Pharmingen) and the Summit Autocomp application (Dako) were used for compensation and levels were adjusted manually. The fluorescence-activated cell sorter (FACS) profiles displayed are representative of at least three independent experiments.
For competitive long-term reconstitution assays, recipients were lethally irradiated (10 Gy, single dose) 24 hours before transplantation and were maintained on antibiotic water (Baytril 10%; Bayer AG, Leverkusen, Germany, http://www.bayer.com) for at least 8 weeks. Double FACS-sorted cells from donor mice were injected retro-orbitally into the irradiated recipients. Blood was collected from the retro-orbital sinus of the recipient mice at different time points as specified, erythrocytes were lysed, and leukocytes were stained for lineage markers. Anti-CD19 was used to identify B cells, anti-CD3ε to identify T-cells, and anti-Mac-1 and anti-Gr-1 to identify myeloid cells. Unless specified, all recipients survived for at least 20 weeks after transplantation.
For noncompetitive short-term transplantation assays, 6-Gy irradiated mice were transplanted 6 hours or 15 weeks after TBI with 10 × 106 unfractionated BM cells harvested from nonirradiated CD45.1 donor mice. The donor contribution was analyzed in recipient PB 5 weeks after transplantation.
Cell Cycle Analysis
Double FACS-sorted cells were incubated for 40 minutes at 37°C with 20 μg/ml Hoechst 33342 in HBSS medium containing 10% fetal calf serum, 20 mM Hepes, 2 mM L-glutamine (Gibco, Grand Island, NY, http://www.invitrogen.com), and 25 μg/ml verapamil (Sigma-Aldrich) as described previously . Pyronin Y (Sigma-Aldrich) was then added at 1 μg/ml, and the cells were incubated for another 15 minutes at 37°C and immediately analyzed by flow cytometry.
A caspase 3&7 FLICA Kit (Immunochemistry Technologies, LLC, Bloomington, MN, http://www.immunochemistry.com) was used according to the manufacturer's protocol. Briefly, about 5 million whole BM cells were incubated for 1 hour at 37°C in 300 μl of 1× FLICA reagent, washed twice, resuspended in 400 μl of washing buffer, and finally analyzed by flow cytometry.
HSCs were double FACS-sorted in droplets of PBS on polylysine-coated slides and incubated at 37°C for 10 minutes. Cells were immediately fixed and permeabilized using a Cytofix/Cytoperm Kit (BD Biosciences, San Diego, http://www.bdbiosciences.com). Staining of γ-H2AX foci was done with anti-phospho-histone H2A.X (Ser139), clone JBW301 antibody (Millipore, Billerica, MA, http://www.millipore.com). After being washed, cells were stained with a secondary Cy-3-conjugated antibody (Amersham Biosciences, Piscataway, NJ, http://www.amersham.com) and 4,6-diamidino-2-phenylindole (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com). Foci were quantitated by fluorescence microscopy using an Axioplan 2 imaging microscope with a ×63 oil objective (Carl Zeiss, Jena, Germany, http://www.zeiss.com).
Data are expressed as mean ± SD in all panels. Statistical comparisons were performed according to Student's t-test. Significance is indicated on the figures using the following convention: *, p < .05; **, p < .01; ***, p < .001; NS, nonsignificant.
SP Cells Show Phenotypic and Cell-Cycle Status Modifications After IR Exposure
To analyze early IR-induced effects on hematopoietic cells and more specifically on HSCs, we first examined BM cellularity and BM frequency of KSL stem/progenitor cells. Two days after 3- and 6-Gy TBI, KSL cell frequency and BM cellularity were decreased in a dose-dependant manner (Fig. 1A, 1B), consistent with a previous report . KSL cells were hardly detectable for the 6-Gy irradiated mice at this time; this finding is intriguing because BM cellularity can recover from such sublethal IR exposure. Because the KSL approach seemed inappropriate for identifying putative radioresistant HSCs among BM cells, we looked for the presence of SP cells able to exclude the fluorescent dye Hoechst 33342. The SP fraction is enriched for HSCs and shows remarkable purity and homogeneity regarding the expression levels of several cell surface markers among which are the two HSC canonical markers, Sca-1 and c-Kit [25, 26]. Two days after 3- and 6-Gy TBI, an SP fraction persisted in the BM of irradiated animals (Fig. 1C). However, its original phenotype was changed: Sca-1 and c-Kit expression levels were increased and severely reduced, respectively, both in a dose-dependant manner (Fig. 1D, 1E).
To further characterize the phenotype of this Sca-1+/brightc-Kit+/low SP (SPSK) population, we examined the expression levels of two markers associated with more primitive HSCs: Endoglin and CD150. Endoglin (CD105) is an ancillary transforming growth factor-β receptor that has been reported to be a functional marker of long-term repopulating HSCs (LT-HSCs) . Endoglin is highly expressed in LT-HSCs at the mRNA and protein levels; in short-term HSCs (ST-HSCs), Endoglin expression levels are about half that of LT-HSCs . CD150 is a member of the SLAM family, whose expression, as for Endoglin, has been associated with LT-HSCs [6, 7], although the question of whether all LT-HSCs are CD150+ remains controversial [28, 29]. We analyzed the expression levels of Endoglin and CD150 at the SPSK cell surface. These two markers were revealed to be redundant in subdividing the SPSK subset of nonirradiated mice into two distinct populations: 60% of SPSK cells constituted the CD105interCD150− fraction, whereas the remaining SPSK cells displayed positive expression for CD105 and CD150 (Fig. 1F, left panel). Two days after IR exposure, the proportions of SPSK cells in the CD105interCD150− and CD105+CD150+ subsets were decreased and increased, respectively, in a dose-dependant manner (Fig. 1F, middle and right panels). This result proves that although Sca-1 and c-Kit expression is modified, the majority of the persisting SPSK cells after radiation exposure express high levels of Endoglin and CD150, two markers associated with primitive HSCs.
To further characterize these cells, we examined their cell cycle status. Before irradiation, 88.6 ± 1.8% of SPSK cells were in the quiescent G0 state, whereas 10.1 ± 1.6 and ∼1% of them were in the G1 and S/G2/M phases, respectively (Fig. 1G, left panel). Two days after 3-Gy TBI, the majority of SPSK cells were outside the G0 phase, with 65.4 ± 5.7 and 9.5 ± 2.9% of them in the G1 and S/G2/M phases, respectively (Fig. 1G, right panel).
Together, the phenotype and the cell cycle status of SPSK cells suggest that this fraction may be enriched for radioresistant HSCs involved in BM replenishment early after exposure.
The SPSK Subset Contains All BM Reconstitution Potential 2 Days After IR Exposure
To test whether SPSK cells include radioresistant HSCs early after irradiation, a competitive reconstitution assay was performed. We transplanted 500 SPSK cells from control or 3-Gy irradiated eGFP donor mice into lethally irradiated nontransgenic recipients. Donor contribution was monitored over a 20-week period by analyzing leukocyte chimerism in recipient PB. Throughout the investigation, we observed multilineage donor contribution from irradiated donor SPSK cells (Fig. 2), although their participation in recipient PB chimerism was lower than that of control SPSK cells. This finding demonstrates that 2 days after 3-Gy TBI, the persisting SPSK subset contains cells with durable self-renewing ability and multilineage differentiation capacity, the two hallmarks of HSCs.
To examine whether other BM cell populations apart from the SP fraction possess long-term repopulating ability after IR exposure, we transplanted 500,000 double-sorted non-SP cells, isolated from 3-Gy irradiated eGFP donor mice, in competition with the same amount of untreated control BM cells into lethally irradiated recipients. Ten weeks after the transplant, we detected no donor contribution in the recipient PB myeloid cell population. PB chimerism was only observable in the B- and T-cell subsets and reached 0.07 and 0.2% for CD19+ and CD3ε+ cells, respectively, plausibly corresponding to long-lived lymphocytes. This assay proves that there is no long-term multilineage reconstitution potential in the BM of 3-Gy irradiated mice outside the SP fraction 2 days after TBI.
Taken together, these two transplantation assays prove that early after IR exposure, some HSCs persist in the BM of irradiated mice. These radioresistant HSCs segregate in the SP fraction but their Sca-1 and c-Kit cell surface expression levels are increased and severely decreased, respectively.
IR Induces Long-Term Phenotypic Alterations in the Stem/Progenitor Compartment
To determine whether changes in Sca-1 and c-Kit expression at the SP cell surface were transient or permanent, we monitored their expression levels over a 10-week period after 3-Gy TBI. The cytometric analysis revealed that the respective increase and decrease in Sca-1 and c-Kit expression levels were gradual and transient, with optima being attained on day 2 after irradiation (Fig. 3). Ten weeks after exposure, the expression levels of these two markers at the SP cell surface were similar for irradiated and control mice.
Because IR exposure can provoke latent residual hematopoietic injury, we examined whether IR-induced effects persisted in the BM of irradiated mice 10 weeks after exposure. At this time, BM cell counts were similar for irradiated (3 and 6 Gy) and control mice. Yet, the BM frequency of KSL cells was decreased in a dose-dependant manner for irradiated mice (Fig. 4A). Interestingly, the proportion of SP cells within the KSL compartment was increased for irradiated mice (Fig. 4B). In addition, a higher proportion of CD150+ cells were observed within the SPSK fraction: 49.5 ± 8.7 and 76.8 ± 7.2% of SPSK cells were positive for CD150 expression 10 weeks after 3- and 6-Gy TBI, respectively, versus 35.2 ± 4.6% for nonirradiated age-matched mice. To further characterize the stem/progenitor cell population of irradiated mice, the two markers CD150 and Flk2 were used to resolve the KSL compartment into three phenotypic subsets: the CD150+Flk-2−, CD150−Flk2−, and CD150−Flk-2+ subfractions, which in young adult mice correspond at the functional level to cell populations enriched for LT-HSCs, ST-HSCs, and multipotent progenitor cells (MPPF), respectively [4, 8, 30]. The gain of Flk2 expression at the surface of HSCs marks the loss of self-renewal in HSC maturation [4, 8]. In young adult mice, CD150−Flk-2+ cells contribute to approximately 60% of the KSL compartment. This proportion was dramatically decreased for irradiated mice 10 weeks after IR exposure; this reduction was accompanied by an increase in the KSL frequency of CD150+Flk-2− and CD150−Flk2− cells (Fig. 4C, 4D). We also examined the BM frequency of these three cell populations. We found the frequencies of CD150−Flk-2− KSL and CD150−Flk-2+ KSL cells to be reduced among BM cells; in contrast, BM frequency of CD150+Flk-2− KSL cells was similar for control and 3-Gy irradiated mice 10 weeks after exposure and slightly higher for the 6 Gy radiation dose (Fig. 4E).
Together, these results show that the KSL compartment of irradiated mice displays dose-dependent and long-lasting modifications, including reduced BM frequency and increased SP cell content. The phenotypic resolution of the KSL subset using CD150 and Flk2 is also profoundly changed. However, IR exposure does not diminish the BM frequency of CD150+ KSL cells.
IR Provokes Long-Lasting Functional Damage to the Stem Cell Compartment
To determine whether IR modifies the functional properties of the CD150+ KSL fraction, we first assessed the cell cycle status of these cells. We found that ∼85% of CD150+ KSL cells purified from irradiated or control mice were in the G0 quiescent state 10 weeks after exposure (data not shown). We next examined the function of CD150+ SPSK cells, corresponding to the SP fraction of the CD150+ KSL cell population, by competitively transplanting 200 double-sorted CD150+ SPSK cells from 3- or 6-Gy irradiated and control donors into lethally irradiated recipients. We monitored recipient PB chimerism over a 20-week period. CD150+ SPSK cells from irradiated donor mice were less efficient than control counterparts in contributing to recipient PB over the whole investigation time course, with a 2.8- and a 4.2-fold decrease in the level of donor contribution to PB leukocytes 20 weeks after transplantation for the 3- and 6-Gy radiation doses, respectively (Fig. 5A). Moreover, recipient PB chimerism was lower for irradiated donor cells in all lineages 5 and 20 weeks after IR exposure (Fig. 5B).
The Stem/Progenitor Compartment of Irradiated Mice Resembles That of Aged Mice
The phenotypic and functional alterations we highlighted in the KSL compartment of irradiated mice several weeks after exposure resemble observations reported in studies performed with aged mice and transgenic mice with DNA damage repair deficiencies [31–33]. To reveal possible additional similarities, we assessed stem/progenitor cells from 3- and 6-Gy irradiated mice for P-selectin cell-surface expression, genomic damage accumulation measured by γ-H2AX foci, and frequency of apoptotic events.
We analyzed the P-selectin expression level at the surface of SPSK cells of irradiated mice 10 weeks after exposure. It has been recently reported that P-selectin expression continuously increases at HSC surface as mice get older . Applying the same staining protocol, we observed that the proportion of cells displaying positive P-selectin expression was higher for irradiated mice than for control counterparts (9 and 20% after 3- and 6-Gy TBI, respectively, vs. ∼5% for control mice) (Fig. 6A). Genomic integrity was also examined, because accumulation of DNA damage has been suggested to play an important role in limiting HSC function [31, 32]. To this end, double-sorted CD150+ SPSK cells from irradiated and age-matched control mice were assessed for the presence of γ-H2AX foci. An accrued number of foci per cell were observable in nuclei of cells purified from irradiated mice, with a 2.1- and a 3.4-fold increase for the 3- and 6-Gy radiation doses, respectively (Fig. 6B). Moreover, the analysis of apoptotic cells using a substrate of activated caspase 3 and 7 revealed a higher proportion of apoptotic events among CD150+ KSL cells with increasing radiation dose (Fig. 6C).
Taken together, these data show that IR exposure causes long-lasting modifications and alterations in the stem/progenitor compartment of irradiated mice, including increased cell surface P-selectin expression, accrued DNA damage accumulation, and a higher frequency of apoptotic events.
BM Cells from Nonirradiated Donors Transplanted into Irradiated Recipients Several Weeks After Sublethal Exposure Do Engraft and Significantly Contribute to Host Hematopoiesis
We finally examined HSC engraftment in irradiated recipients transplanted several weeks after sublethal exposure, once hematopoietic homeostasis was restored. To this end, 10 × 106 mononuclear BM cells from nonirradiated young adult CD45.1 donors were injected in CD45.2 recipients that had been exposed to 6-Gy TBI 15 weeks before transplantation or in CD45.2 age-matched nonirradiated recipients. A control group of 6-Gy irradiated mice transplanted on the day of exposure was also included. As expected, 5 weeks after transplantation, limited donor contribution was observed among PB leukocytes of nonirradiated recipients (0.5 ± 0.5%), whereas donor cells contributed to 80 ± 6.2% of PB leukocytes in irradiated mice transplanted on the day of exposure, consistent with a previous study . Interestingly, irradiated recipients transplanted 15 weeks after exposure displayed donor contribution to PB leukocytes reaching 4.1 ± 1.6%. The level of chimerism among recipient Gr-1+/Mac-1+ PB cells was also determined, as donor contribution to PB myeloid lineage has been shown to be a good indicator of stem cell activity [32, 36]. We found that donor-derived cells contributed to ∼0.2 (corresponding to the detection level), 78.3 ± 9, and 8.3 ± 4.1% of PB myeloid lineage in nonirradiated recipients, irradiated recipients transplanted on the day of exposure, and irradiated recipients transplanted 15 weeks after exposure, respectively (Fig. 6D). This result indicates that donor HSCs can engraft and significantly contribute to host myeloid and lymphoid lineages even if BM transplantation is performed several weeks after sublethal TBI of the recipient.
We have examined the effects of IR exposure on stem/progenitor cells over a period of time covering both BM suppression and recovery. Early after irradiation, we could identify a population of radioresistant HSCs that belonged to the SP fraction. This cell population presented transient phenotypic changes that paralleled the BM suppression and replenishment phases. Ten weeks after IR exposure, BM cell counts were similar for irradiated mice and age-matched controls. However, phenotypic and functional alterations were identified in the KSL compartment of irradiated mice, accounting for long-lasting hematopoietic injury.
Early after irradiation, HSCs identified by their repopulating ability continue to segregate within the Hoechst dye-excluding SP population, but their Sca-1 and c-Kit expression levels are increased and severely reduced, respectively, with a concurrent increase in the proportion of SPSK cells positive for established indicators of the presence of HSCs: CD150 and CD105. Variations in Sca-1 and c-Kit expression levels explain the inability of the KSL approach to isolate HSCs early after IR exposure. Moreover, because these variations are gradual and transient over the investigation period, it is likely that IR exposure itself causes stem cells to partially lose expression of c-Kit and upregulate that of Sca-1, rather than revealing a previously uncharacterized radioresistant population.
Downregulation of c-Kit expression accompanying cell cycle entry or mobilization of stem/progenitor cells has been reported after granulocyte colony-stimulating factor (G-CSF) administration . In addition to transcriptional controls, exocytosis, and ligand-induced internalization, decreased c-Kit expression after G-CSF treatment has been attributed to direct proteolytic cleavage by neutrophil and macrophage proteases, including matrix metalloproteinase-9 (MMP9), released in the BM of treated mice . Similarly, reduced c-Kit expression at the HSC surface was observed in the first days after 5-fluorouracil (5-FU) treatment [16, 38]. Moreover, downregulation of c-Kit expression was also observed during the regeneration of the megakaryocytic lineage in transgenic mice with conditional megakaryocytic lineage suppression . Finally, in MMP9−/− mice, impaired release of Kit ligand after 5-FU injection or IR exposure was identified, leading to a failure or delay in hematopoietic recovery [40, 41]. Taken together, these observations suggest that downregulation of c-Kit expression at the HSC surface occurs in response to various hematological insults or stresses.
Upregulation of Sca-1 expression at the HSC surface is more intriguing. Sca-1 is widely believed to have an important role in cell-cell adhesion and signaling , although little is known about its biochemical function. A possible homing defect and impaired repopulating ability were reported for Sca-1−/− progenitor cells . Moreover, Sca-1−/− mice exhibit deregulated c-Kit expression at the surface of HSCs after 5-FU administration . Our data, showing inversed and almost symmetrical regulation of Sca-1 and c-Kit expression at the surface of cycling SP cells, support a possible interdependence of these two markers.
The expression of Sca-1 and c-Kit at the SP cell surface returned to normal levels several weeks after exposure. In contrast, major persistent alterations were observed within the stem/progenitor compartment of irradiated mice at this time: (a) the BM frequency of KSL cells was reduced; (b) the proportion of SP cells within the KSL fraction was increased; (c) the phenotypic composition of the KSL compartment was modified, with increased proportions of CD150+Flk-2− and CD150−Flk-2− cells, whereas that of MPPF-like CD150−Flk-2+ cells was dramatically reduced; (d) the proportion of CD150+ cells among SPSK cells was dose dependently increased; (e) the repopulating capacity of CD150+ SPSK cells was altered when assessed in a competitive transplantation; (f) the P-selectin expression level was increased at the surface of SPSK cells; (g) an accrued number of γ-H2AX foci were observable in the nuclei of CD150+ SPSK cells from irradiated mice compared with control counterparts; (h) the frequency of apoptotic events in the CD150+ KSL cell population was higher in irradiated animals than in control mice; and (i) irradiated recipients transplanted several weeks after sublethal exposure with BM cells from nonirradiated congenic donors, but not age-matched controls, showed engraftment and significant contribution of donor HSC. These changes were highly reproducible, pronounced, often dose-dependent, and persistent: they may be seen as an IR-induced long-lasting “imprint” and prove that the hematopoietic system keeps some kind of memory of the radiation insult.
The higher frequency of apoptotic events among CD150+ KSL cells of irradiated mice several weeks after exposure in conjunction with an accrued number of damage-induced γ-H2AX foci agrees with a recent report suggesting that HSCs can accumulate genomic damage in the G0 quiescent state. The DNA damage may be repaired; otherwise the cells are eliminated by apoptosis while progressing through the cell cycle [2, 32].
When comparing the repopulating ability of HSCs from irradiated and control mice, we found IR to diminish the capacity of SPSK cells and CD150+ SPSK cells to competitively reconstitute lethally irradiated recipients when harvested 2 days and 10 weeks after radiation exposure, respectively. Shortly after exposure, this reduced repopulating capacity may partly be due to (a) the cell cycle status of SPSK cells, as cycling HSCs have reduced engraftment potential compared with quiescent HSCs ; (b) the transient downregulation of c-Kit at the surface of HSCs, because c-Kit seems to be required for the extensive self-renewing divisions that HSCs must undergo after transplantation into myeloablated recipients ; or (c) the reduced homing efficiency of HSCs of irradiated mice. The decrease observed in the repopulating ability of CD150+ SPSK cells of irradiated mice several weeks after exposure may be the consequence of a reduced HSC homing capacity, as was shown for HSCs of old mice  and/or an increase in the proportion of senescent cells in the CD150+ SPSK fraction as was previously suggested by Wang et al. .
Overall, our study allows better understanding of immediate and delayed hematopoietic responses to the radiation insult and provides novel insight in the complex spectrum of long-lasting IR-induced injuries to HSCs. The vast majority of persistent phenotypic and functional alterations we could associate to stem/progenitor cells of irradiated mice were previously reported in studies performed with aged mice and transgenic mice with deficiencies in DNA damage repair [28, 31–34, 47]. This proves that IR can induce within weeks persistent alterations in the stem/progenitor compartment of young adult mice that normally take months or years to gradually appear in nonirradiated individuals; it is therefore plausible that the mechanisms underlying long-lasting deleterious effects induced by radiation exposure on HSCs are, at least partly, similar to those governing the process of aging.
We thank Chrisophe Joubert, Laurent Vial, Claire Chauveau, and colleagues for excellent animal care, Daniel Lewandowsky, Ghida Harfouche, and Marion Brenot for advice and technical help, and Elisabeth Jajolet for precious administrative support. This work was supported by funds from the Commissariat à l'Energie Atomique and Electricité de France.
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.