Enhancing Hematopoietic Stem Cell Transplantation Efficacy by Mitigating Oxygen Shock

Abstract

Hematopoietic stem cells (HSCs) reside in hypoxic niches within bone marrow and cord blood. Yet, essentially all HSC studies have been performed with cells isolated and processed in non-physiologic ambient air. By collecting and manipulating bone marrow and cord blood in native conditions of hypoxia, we demonstrate that brief exposure to ambient oxygen decreases recovery of long-term repopulating HSCs and increases progenitor cells, a phenomenon we term extraphysiologic oxygen shock/stress (EPHOSS). Thus, true numbers of HSCs in the bone marrow and cord blood are routinely underestimated. We linked ROS production and induction of the mitochondrial permeability transition pore (MPTP) via cyclophilin D and p53 as mechanisms of EPHOSS. The MPTP inhibitor cyclosporin A protects mouse bone marrow and human cord blood HSCs from EPHOSS during collection in air, resulting in increased recovery of transplantable HSCs. Mitigating EPHOSS during cell collection and processing by pharmacological means may be clinically advantageous for transplantation.

Figure 1. BM harvest and processing in hypoxia (3%O₂) or in ambient air

A) BM was harvested and processed in a hypoxic chamber (3%O₂, 5%CO₂, N₂ balance) or ambient air (∼21% O₂). Ai) Flow cytometric density dot-plots are representative of 6 independent experiments. LT-HSC are defined as CD34^-CD48^-LSK. Aii) Relative frequency histograms. Bar indicates CD34 positive or negative staining based on isotype control antibody. Aiii) Number of LT-HSCs collected, per 10⁶ BM cells, when harvested and processed in air (A) or in the hypoxic chamber (H); mean±SE for 6 independent experiments; statistics determined by Mann-Whitney method. B)LT-HSCs, collected from BM harvested in hypoxic chamber and divided into two; one was removed from the chamber and immediately exposed to air for 60 min (H→A), the other was left in the chamber for 60 min (H→H) before further staining and processing. Chart bars are mean±SE for 3 independent experiments. C) Length of air exposure on numbers of LT-HSCs, which were collected after harvest in hypoxic chamber (zero time) and aliquots exposed to air for 30, 60, or 90 min before staining and processing (1experiment). Logistically, it was difficult to do a time point much less than 30 min. D) Flow cytometric density dot-plots of CD34 and CD150 expression in CD48-CD41-LSK cells. With this antibody combination, LT-HSCs are defined as CD150+CD34-CD48-CD41-LSK (Oguro, et al., 2013). Density plots represent 2 similar experiments. Ei) ROS levels in LT-HSC harvested and processed in ambient air (A→A) or in hypoxia (H→H) (3 independent harvests on the same day; mean±SD). Eii) ROS levels in CD48^-LSK; mean±SD for 3 independent experiments (1 mouse per collection per experiment on different days). Fi) Mitochondrial activities, measured by Mitotracker Green FM mean fluorescent intensity in CD48^-LSK BM cells collected and processed in ambient air (A→A) or hypoxia (H→H); mean±SD for 3 independent experiments. Fii) Percentage of CD48^-LSK cells with hyperpolarized mitochondrial membrane potential using JC-1 probe in BM collected and processed in ambient air or hypoxia; mean±SD for 3 independent experiments. G) Numbers of LT-HSC, Short-term HSC (ST-HSC; CD34⁺CD41^-CD48^-LSK), or multi-potent progenitors (MPP; CD48^-Lin^-Sca1^-c-kit⁺); mean±SD for 6 independent experiments (one mouse per experiment each harvested, processed, and analyzed on different days). H) Absolute numbers per femur (upper graphs) and cycling status (% in S-phase as determined by high specific activity tritiated thymidine kill assay; lower graphs) of HPC for BM cells collected and processed in hypoxia (H→H), or collected under hypoxia and placed in air for 60 min. (H → A) and then cultured in hypoxia (5% O₂). Mean ±SE for 3 mice each from 1 experiment. *p<0.05; **p<0.001. Results were reproduced in 4 additional experiments with 3 mice/group each. Ii) Representative flow cytometric contour plots of human LT-HSC (Lin^-CD34⁺CD38^-CD45RA^-CD90⁺CD49f⁺ as per Notta et al., 2011) from CB collected under hypoxic conditions with half left in hypoxia for processing (H→H) and half left in air for 60 min. (H→A). Iii) Number of phenotyped human CB-derived HSC per 10⁶ total cells; mean ± SD for 4 independent CB harvests. See also Figure S1.

Figure 2. Competitive mouse HSC repopulation, HSC homing, Apoptosis, and CXCR4 expression

A) Competitive HSC engraftment. Donor cells were CD45.2⁺, competitors were CD45.1⁺ (competitor cells all collected in air and injected either in hypoxia or air immediately after donor cells were injected), and infused into 950 cGy irradiated dual CD45.2⁺/CD45.1⁺ F1 recipients at 150K donor and 150K competitor cells. Ai) BM cells were harvested in hypoxia and cells split so that half were processed and injected in hypoxia, and half subjected to air for 60 min. prior to processing and injecting in air. Mean ± SE of CD45.2⁺ (donor cell) chimerism (1 experiment) for numbers of mice evaluated. Open circles (hypoxia (H)→Air (A)) and closed circles (H→H). Aii) Combined results of 2-3 separate engrafting studies as noted in text. Aiii) CRUs calculated from LDA as per Antonchuk, et. al., 2002 (n=3-4 mice per group at each cell concentration infused for each of two experiments) for month 3 peripheral blood and month 7/8 for BM. P value is based on Poisson statistics. B) Homing of cells collected in hypoxia, and then left to be processed and injected under hypoxia (H→H for input), or collected under hypoxia and then exposed to room air for 60 minutes before processing and injecting cells in air (H→A for input). Cells were analyzed 24 hours after injection by removing 2 femurs plus 2 tibias from 10 mice under hypoxia and splitting the cells into two, one half which was left at hypoxia for staining and assessment (H→H for output) and one half placed into air for 60 min. for staining and assessment (H→A for output). Lin- BM cells were injected and % LSK cells homed (output to input LSK) showed no significant differences in any of the groups by AN OVA analysis Ci) Representative relative frequency histograms of CXCR4 surface staining intensity in LT-HSCs. Cii) Mean fluorescence staining intensity of 4 BM harvests (mean ± SD). D) Flow cytometric assessment of apoptosis using antibody to activated caspase-3. Data represent 2 similar experiments. See also Figure S2.

Figure 3. MPTP in EPHOSS: effects of cyclosporine A

A) Effects of BM harvest in air in presence of 50μg/ml cyclosporine A (CSA-harvest) or DMSO control on CD34 expression levels in CD48-LSK cells. For CSA group, mice used for cell collection were injected with CSA (Experimental Procedures). Mice for control group were injected with control medium. Representative flow cytometric density dot-plot for 3 independent harvests. LT-HSCs are noted in circles. B) Dose-response of CSA on LT-HSC collection (N=1 expt). Ci) Relative frequency histogram representative of 3 independent harvests in DMSO or 50ug/ml CSA. Cii) Average LT-HSCs collected after DMSO or CSA-harvest (3 independent experiments; mean ±SD). Ciii) Effect of time of air exposure on BM harvested in the presence of either DMSO or CSA (1 experiment). D) Effects of CSA or DMSO harvest on GM-CSF plus SCF-induced CFU-GM colony formation (mean±SD from 6 mice each in a total of 2 experiments). Effects of FK506- (E) or carboxyatractylate- (CAT-) (F) harvests shown as a relative frequency histogram from flow cytometric data. Data in E and F are representative of 2 independent experiments each with similar results. Bar charts are quantitation (mean ± range) for 2 experiments. G) Effect of DMSO-, CSA-, or CAT-harvest on ROS levels in CD48-LSK cells (mean of 2 independent experiments ± range). Hi) Effect of CSA-harvest, compared to DMSO (control; C)-harvest, on HSC engraftment in competitive repopulation transplant assay (mean % CD45.2⁺ donor cell chimerism ±SE for numbers of mice shown at each point. Results for months 1-4, and 8 for primary mice are for 2 experiments, while that at 1,3 months for secondary mice are for 1 experiment. Hii) CRUs calculated by LDA analysis respectively at 3 months for peripheral blood (PB) and at 8 months for BM (n=3-5 mice per group at each cell concentration infused for each of two experiments). P value is based on Poisson statistics. See also Figures S3 and S4.

Figure 4. Effect of human CB CSA and mouse BM cyclophilin D (CypD) -/- collections

A) Effect of CSA collection on CB CD34⁺ cells, LT-HSCs, and MPPs. Results are from 5 different CB collections. B) CRUs calculated from LDA (n-3-5 mice per group at each cell concentration for each of 2 separate CB collections. Bi) Combined results of 2 expts at 2 months for PB. Bii) Results of one of the CB collections at 3.5 months for BM. Biii) Results of the other CB collection at 4.5 months for BM (Bii and Biii results for BM were not combined as the percent chimerism for both was largely different precluding averaging the results for LDA). Ci) Effect of CypD gene deletion on phenotyped LT-HSC recovery when BM is harvested in air (N=7 experiments). Cii) ROS levels in LT-HSC cells harvested from 5 of the 7 different experiments shown in Ci. D) Recovery of multi-cytokine-stimulated CFU-GM, BFU-E, and CFU-GEMM (Average of 6 WT and 10 CypD -/- mice in a total of 2 experiments expressed as mean ±SE). E) GM-CSF plus SCF-induced CFU-GM colony formation (mean ± SE) in BM from same animals as in D. F) Percent PB chimerism of CypD -/- BM cell engraftment at limiting dilution analysis for months 1-3 in PB and at 4.5 months in BM (4-5 mice/group at each cell concentration for 1 expt.). Fii) CRUs as calculated for month 2 and 3 PB and month 4.5 BM by LDA for CypD -/- BM cell engraftment. See also Figures S4 and S5.

Figure 5. Effect of p53, miR210 or hif-1α gene deletion

A) Effect of p53 gene deletion on hypoxic (H→H) and air (H→A) BM harvest on phenotyped LT-HSC recovery for independent BM harvests from six p53-/- and three littermate control wild type (WT) mice. BM from each mouse was harvested and maintained in the hypoxic chamber for 60 min before staining and fixation (H→H) or was exposed to air for 60 min before staining and fixation (H→A). Mean LT-HSC per 10⁶ nucleated BM cells ± SD for each group. B) Effect of p53 gene deletion on multi-cytokine stimulated progenitor cell recovery (left) or cell cycle status (right; high specific activity tritiated thymidine kill assay). BM was from the same animals harvested in E and cells were cultured in hypoxia (5%O₂). C) Phenotyped LT-HSC, ST-HSC, and MPP, recovery for miR210-/- (i), or hif-1α-/- (ii) BM. Cells from one femur were collected and processed in air (A) and cells from the contralateral femur were collected and processed in hypoxia (H). One of two experiments with similar results for (i), and one experiment for (ii). D) HPC recovery from miR210-/- (i) or hif-1α -/- (ii) BM harvested and processed in air (A), or hypoxia (H), and cultured in hypoxia (5%O₂). 1 of 3 reproducible experiments with 3 mice/group for each experiment for the miR210 -/- mice in which one femur was collected in air and the contralateral femur collected in hypoxia, with one of the other two experiments done in a similar manner, and one experiment done with collection in hypoxia and processing in air vs. collection and processing in hypoxia. 1 of 2 reproducible experiments for hif-1α -/- mice with harvest in air vs. hypoxia. a=significant (p<0.01) compared to WT A control, b=not significant (p>0.05) compared to hif-1α-/- or miR210-/-.