Increased reactive oxygen species and exhaustion of quiescent CD34-positive bone marrow cells may contribute to poor graft function after allotransplants

Poor graft function (PGF) is a fatal complication following allogeneic haematopoietic stem cell transplantation. However, the underlying mechanism is unclear. Effective cross-talk between haematopoietic stem cells (HSCs) and bone marrow microenvironment is important for normal haematopoiesis. Normal HSCs reside in a hypoxic bone marrow microenvironment that protects them from oxidative stress that would otherwise inhibit their self-renewal and results in bone marrow failure. Whether an increased level of reactive oxygen species (ROS) causes PGF following allotransplant is unclear. Using a prospective case-pair study, we identified increased levels of ROS in CD34+ bone marrow cells in subjects with PGF. Elevated ROS levels was associated with an increased frequency of DNA strand breaks, apoptosis, exhaustion of quiescent CD34+ cells and defective colony-forming unit plating efficiency, particularly in the CD34+CD38− fraction. Up-regulated intracellular p53, p21, caspase-3 and caspase-9 levels (but not p38) were detected in CD34+ cells, particularly in the CD34+CD38− fraction. To further study the potential role of ROS levels in post-transplant haematopoiesis, CD34+ bone marrow cells from subjects with good graft function were treated with H2O2. This increased ROS levels resulting in defective CD34+ cells, an effect partially reversed by N-acetyl-L-cysteine. Moreover, CD34+ bone marrow cells from the donors to subjects with poor or good graft function exhibited comparable haematopoietic reconstitution capacities in the xeno-transplanted NOD-PrkdcscidIL2rgnull mice. Thus, even if the transplanted donors' bone marrow CD34+ cells are functionally normal pre-transplant, ROS-induced apoptosis may contribute to the exhaustion of CD34+ bone marrow cells in subjects with PGF following allotransplant.


INTRODUCTION
Poor graft function is an important, often fatal complication following allogeneic haematopoietic stem cell transplant [1][2][3][4]. The etiology of poor graft function is complex and includes many factors such as bone marrow toxic drugs, infections, graft-versus-host disease (GvHD) and an impaired bone marrow microenvironment [2,3,[5][6][7][8][9]. One or more of these factors may operate in different persons or the same person with poor graft function post-allotransplant.
Interactions between haematopoietic stem and progenitor cells and the bone marrow microenvironment are important in maintaining normal haematopoiesis [10][11][12][13]. We recently reported that the transplanted donor CD34-positive cells were quantitatively normal in subjects with poor graft function, but the frequency of bone marrow CD34-positive cells were dramatically reduced and bone marrow endosteal, vascular microenvironment were impaired in subjects with poor graft function compared with those with good graft function posttransplant [2,3], raising the question whether bone marrow CD34-positive cells in subjects with poor graft function are functionally impaired posttransplant, or the transplanted donors' CD34-positive cells are already defective pretransplant. Moreover, what drives the bone marrow CD34-positive cells impairing functionally posttransplant and its molecular mechanisms remain to be elucidated in poor graft function.
Reactive oxygen species (ROS) are free radicals derived from diatomic oxygen and exhibit diverse reactivities. ROS affect cell cycle progression, cell motility and growth factor signaling in many cell types, including haematopoietic stem and progenitor cells [14,15]. Haematopoietic stem and progenitor cells may occupy a hypoxic niche in the bone marrow microenvironment that protects them from oxidative stress [16,17]. In Atmor FoxOdeficient mice, haematopoietic stem cells are depleted due to increased ROS levels. This effect is reversible by treatment with the anti-oxidative drug N-acetyl-L-cysteine [18]. We hypothesized that increased levels of ROS in the bone marrow microenvironment post-allotransplant result in the depletion of donor haematopoietic stem and progenitor cells, leading to poor graft function.

Reduced quiescent cells and increased levels of DNA double-strand breaks and apoptosis in CD34-positive bone marrow cells from subjects with poor graft function
Subjects with poor or good graft function were evaluated at comparable intervals posttransplant to minimize bias. Subjects with poor graft function had significantly lower numbers of CD34-positive cells ( Figure 1A Apoptosis was also markedly increased in CD34positive bone marrow cells from subjects with poor graft function ( Figure 1C, 24.48%±2.66% vs. 12.19%±2.08%; P=0.001), particularly in the CD34-positive, CD38-negative fraction ( Figure 1C, 48.55%±4.38% vs. 18.65%±1.85%; P<0.0001).
Comet assay was performed to confirm the DNA damage of bone marrow CD34-positive cells. As shown in Figure 1D, the bone marrow CD34-positive cells from normals are round-shaped without tails. By contrast, the CD34-positive cells from subjects with poor graft function showed longer tails than those with good graft function. Moreover, both the frequencies of comet cells and tail DNA in bone marrow CD34-positive cells of subjects with poor graft function were significantly higher than those with good graft function and normals ( Figure 1E, 1F).

Defective CFU plating-efficiency of CD34positive bone marrow cells from subjects with poor graft function
CD34-positive bone marrow cells from subjects with poor graft function had significantly lower plating efficiencies (CFU counts/2×10E+3 CD34-positive cells; Figure 2)  Increased levels of ROS and p53 are associated with exhaustion of CD34-positive bone marrow cells Intracellular levels of ROS, p53, phospho-p53, p21, phospho-p38, caspase-3 and caspase-9 were analyzed in subjects with poor or good graft function and normals. Subjects with poor graft function had significantly higher intracellular ROS levels ( Figure 3B) in CD34www.impactjournals.com/oncotarget

CD34-positive bone marrow cells from the donors of poor or good graft function subjects demonstrated comparable haematopoietic reconstitution capacity in the xeno-transplanted NOD-Prkdc scid IL2rg null mice
Bone marrow and spleen from all xeno-grafted NOD-Prkdc scid IL2rg null mice contained large numbers of human haematopoietic cells which correlated with the dose of cells injected (Figure 7). Engraftment levels were

DISCUSSION
Poor graft function following allotransplant remains a life-threatening complication with limited effective treatment options [1][2][3][4]. Several risk factors, including the number of transplanted CD34-positive cells, disease state, drug-induced toxicity, GvHD, infections, and an impaired bone marrow microenvironment have been reported to be associated with the occurrence of poor graft function [2,3,[5][6][7][8][9]. However, no study has focused on the functional characterization of bone marrow CD34-positive cells in subjects with poor graft function following allotransplant or in donors to subjects with poor graft function pretransplant. Moreover, the effect of increased ROS on bone marrow CD34-positive cells in subjects with poor graft function following allotransplant is unclear.
Oxidative stress, in particular ROS, regulates haematopoietic stem and progenitor cells function in mice [14][15][16][17][18] and Drosophia [19]. Cells with low ROS levels have better long-term repopulating capacity compared with those with high ROS levels which are mostly involved with short-term repopulation [15]. Under normal conditions, haematopoietic stem and progenitor cells are found in hypoxic bone marrow microenvironment, a setting which protects them from oxidative stress [15][16][17]. In contrast, exceedingly high ROS production occurs under various pathological conditions, which can inhibit haematopoietic stem and progenitor cells self-renewal and induce DNA damage and apoptosis resulting in premature exhaustion of haematopoietic stem and progenitor cells and haematopoietic dysfunction [18,20,21].
Appropriate control of quiescence is crucial for normal haematopoietic stem and progenitor cells function [22][23][24]. Cell cycle changes affect the repopulating ability of murine stem cells [25][26][27]. We found haematopoietic stem and progenitor cells are functionally impaired in subjects with poor graft function and had a significantly lower fraction of quiescent bone marrow-derived CD34positive cells compared with subjects with good graft function and with normals. However, it should be noted that the median age of the normal cohort is younger than those in the cohorts of poor graft function and good graft function in the current study. Our data are consistent with the hypothesis that poor graft function is associated with a defect in maintenance of haematopoietic stem and progenitor cells quiescence, which is in accordance with the worldwide practice that the administration of a CD34positive selected stem cell boost is an effective option for improving graft function [1,[28][29][30].
Our data indicate impaired haematopoietic stem and progenitor cells function is associated with increased intracellular levels of ROS. This was associated with increased levels of p53, p21 but not p38, in contrast to the results of previous study [18]. Although whether the ROS elevation is the cause or consequence of poor graft function and the underlying molecular mechanisms remain to be clarified, our data provide evidence that elevated intracellular ROS lead to increased DNA damage and apoptosis via the p53-p21 pathway.
The sources and regulation of abnormal intracellular ROS in bone marrow CD34-positive cells from subjects with poor graft function have yet to be elucidated. Effective cross-talk between haematopoietic stem and progenitor cells and the bone marrow microenvironment is important for the regulation of haematopoiesis [10][11][12][13]. At the junction of these types of regulation, ROS produced endogenously via cellular respiration or nicotinamide adenine dinucleotide phosphate-oxidase activity (haematopoietic stem and progenitor cell-derived) [31,32] as well as after exposure to exogenous stress (bone marrow microenvironment-derived) [16][17][18]33] play important roles in regulating haematopoietic stem and progenitor cell functions. We previously reported that the bone marrow endosteal and vascular microenvironment are impaired in poor graft function post-transplant [2,3]. In the current study, CD34-positive bone marrow cells from the donors to subjects with poor or good graft function exhibited comparable haematopoietic reconstitution capacities in the xeno-transplanted NOD-Prkdc scid IL2rg null mice. Based on our data and previous reports [2,3,[16][17][18][33][34][35], it is conceivable that the preconditioning and some posttransplant events, such as GvHD, CMV reactivation, and some cytotoxic agents, may trigger abnormally increased ROS in bone marrow microenvironment, which may subsequently lead to defective haematopoietic stem and progenitor cells and the occurrence of poor graft function post-allotransplant. We are aware, however, that further studies are needed to determine whether the potential sources of the ROS are intrinsic to haematopoietic stem and progenitor cells, the bone marrow microenvironment, or both in the future.
In conclusion, although requiring further functional validation, our data demonstrated for the first time that even if the transplanted donors' bone marrow CD34positive cells are functionally normal pretransplant, elevated ROS in bone marrow CD34-positive cells postallotransplant may result in increased levels of DNA strand breaks, apoptosis, up-regulated levels of p53, p21 (but not p38), and exhaustion of quiescence in subjects with poor graft function. Moreover, H 2 O 2 -induced increases in ROS resulted in defective CD34-positive cells is partially reversed by N-acelyl-L-cysteine. Thus, our data suggest

Subjects and normals
Three cohorts were analyzed, subjects with poor or good graft function posttransplant and normals (transplant donors). Transplant recipients were identified from consecutive subjects receiving an allotransplant for a haematologic neoplasm April 1, 2014 to March 31, 2015 at Peking University Institute of Hematology and willing to participate in the study. 15 subjects developing poor graft function were eligible. For each case, 2 matched transplant recipients with good graft function were randomly-selected from the same cohort after matching for age, pretransplant disease state and interval posttransplant ("risk-set sampling") [36]. Variables of subjects and controls are summarized in Table 1. Bone marrow samples from donors (n=45) were controls. The normal cohort comprised 25 males and 20 females, ages 18-55 years (median, 29 years). The study was approved by the Ethics Committee of Peking University People's Hospital, and written informed consent was obtained from all subjects compliant with the Declaration of Helsinki.
Chimerism analyses were done by DNA fingerprinting for short tandem repeats in blood samples and/or by chromosome fluorescent in situ hybridization of bone marrow samples. Complete donor chimerism was defined as no recipient haematopoietic or lymphoid cells detected (sensitivity >0.1% recipient signals) [37].
Diseases were categorized as standard-or high-risk. Standard-risk was defined as 1 st or 2 nd complete remission (CR1 or CR2) of acute leukemia or myelodysplastic syndrome (MDS). All other subjects were classified as high-risk. Haematologic relapse was defined as blasts >5% in the blood, bone marrow or extra-medullary site. GvHD was scored as acute or chronic as previously described [2,3,37].

Transplantation protocols
Donor selection, HLA-typing, graft harvesting, conditioning therapy and GvHD prophylaxis were done as reported [2,3,37,38]. The subjects were screened pre-transplant for cytomegalovirus (CMV) infection by serology. Weekly real-time quantitative PCR was used to detect CMV reactivation in blood samples. CMV infections were treated with ganciclovir or foscarnet. After allotransplants, rhG-CSF (5 μg/kg/day) was administered to recipients of HLA-mismatched related transplants from day +6 until neutrophils were >0.5 × 10E+9/L for 3 consecutive days. rhG-CSF was not administered to recipients of HLA-identical sibling transplants, except in cases where neutrophils were <0.5 × 10E+9/L until day +21. The subjects received RBCs if their hemoglobin concentrations were ≤70 g/L, or following platelet transfusion if their platelets were ≤20 × 10E+9/L.

Cell cycle and apoptosis analyses
Bone marrow mononuclear cells were isolated by density centrifugation using lymphocyte separation medium (GE Healthcare, Milwaukee, WI, USA). Cell cycle analyses were performed by incubating with 10 μg/ ml Hoechst 33342 (Thermo Fisher Scientific, Waltham, MA, USA ) at 37°C for 45 min, and 1.0 μg/ml Pyronin Y (Sigma, St. Louis, MO, USA) at 37°C for an additional 15 min. Cells were stained with mouse anti-human CD34-PerCP-Cy5.5 and CD38-APC-conjugated monoclonal antibodies (Becton Dickinson) at room temperature for 15 min.
To detect apoptosis, cells were incubated with CD34-PE, CD38-APC and CD45-V500 and then incubated for 15 min with Annexin-V-FITC and 7-aminoactinomycin D (7-AAD) apoptosis detection kit (Becton Dickinson) according to the manufacturer's instruction. Multi-parameter flow cytometric analyses were done using a BD LSRFortessa (Becton Dickinson). Aliquots of unstained samples were used as negative controls. Data were analyzed using BD LSRFortessa software (Becton Dickinson).

Single-cell gel electrophoresis (the comet assay)
Sorted CD34-positive bone marrow cells were suspended in 25μl of 0.6% (w/v) low melting point agarose in PBS, and immediately pipetted onto a frosted glass microscope slide precoated with a layer of 0.8% (w/v) normal melting point agarose similarly prepared in PBS. The slides immersed in lysis solution at 4°C overnight to remove cellular protein. The slides were placed in electrophoresis tank containing 0.3 M NaOH and 1 mM EDTA for 30 min before electrophoresis at 28 V for 25 min at an ambient temperature of 4°C. The slides were then washed 3 times with 0.4 M Tris-HCl, pH 7.5, at 4°C before staining with 8μl DAPI. One hundred comets from each slide were electronically captured using fluorescence microscope, and the images were analyzed with CASP1.2.3 software (Institute of Theoretical Physics, University of Wroclaw, Wroclaw, Poland)

Treatment of CD34-positive bone marrow cells with H 2 O 2 and N-acetyl-L-cysteine
CD34-positive bone marrow cells were cultured with StemSpan™ SFEM (Stem Cell Technologies) supplemented with 50 ng/ml cytokines(SCF, FLT-3, and TPO) (PeproTech, Locky Hill, NJ, USA) and maintained at 37°C in a 5% CO 2 atmosphere. There were too few cells from subjects with poor graft function for these experiments.

Characteristics
Poor Graft Function cases a (n=15) To further investigate the effect of oxidative stress on normal posttransplant haematopoiesis, CD34-positive bone marrow cells from subjects with good graft function were treated with 100 μM H 2 O 2 (Sigma), for 24 h at 37°C with or without 1 mM N-acetyl-L-cysteine (Sigma). Quantification of the apoptosis, intracellular ROS levels and cell-cycle state of different groups were analyzed by flow cytometry (see above).
Evaluation of the hematopoietic reconstituting activity of donor CD34-positive bone marrow cells in NOD-Prkdc scid IL2rg null mice NOD-Prkdc scid IL2rg null mice were purchased from Beijing Vitalstar Biotechnology Co., Ltd., Beijing, China. The mice were maintained in specific pathogenfree conditions and fed sterilized water and food at the animal facility of Peking University People's Hospital. The mice were 4-6 weeks old at the time of transplantation. All animal experiments were approved by the Ethics Committee of Peking University People's Hospital.
CD34-positive cells were isolated from the bone marrow mononuclear cells of donors to subjects with poor or good graft function using a CD34 MicroBead Kit (Miltenyi Biotec). The hematopoietic reconstituting activity of the donor CD34-positive bone marrow cells was evaluated using a NOD-Prkdc scid IL2rg null xenograft assay by intra-bone marrow injection [39,40]. Briefly, mice received 1.5 Gy of total body irradiation from a 60 Co source for 90 s. Within 24 h, intra-bone marrow injections of 1×10E+4, 1×10E+5 or 4×10E+5 CD34-positive bone marrow cells from donors to subjects with poor (n=5) or good graft function (n=5) were performed in triplicate.

Statistical analyses
Statistical analyses were done using one-way ANOVA to compare the three groups. Subject variables were compared using the χ 2 test for categorical variables and the Mann-Whitney U test for continuous variables. Intracellular ROS levels and cell cycle status in subjects with poor or good graft function were compared with those in their paired donors' bone marrow cells by paired-T test. Correlations between hematologic parameters and ROS levels in CD34-positive bone marrow cells in patients posttransplant were determined using Pearson correlation coefficients. Analyses were performed using SPSS 22.0 (IBM, Armonk, NY, USA) and GraphPad Prism 6.0 software (GraphPad Software, La Jolla, CA, USA) packages, and P-values < 0.05 were considered statistically significant.