Aerobic exercise in humans mobilizes HSCs in an intensity-dependent manner

Here we demonstrate for the first time that mobilization of hematopoietic stem cells (HSC) through exercise is intensity dependent, with the greatest mobilization occurring immediately after high-intensity exercise. As well, we show that exercise is a general stimulus for mobilization: increases in specific HSC populations are reliant on general mononuclear cell mobilization. Finally, we demonstrate no differences in mobilization between groups with different aerobic fitness.

Abstract

Hematopoietic stem and progenitor cells are necessary to maintain, repair, and reconstitute the hematopoietic blood cell system. Mobilization of these cells from bone marrow to blood can be greatly increased under certain conditions, one such being exercise. The purpose of this study was to identify the importance of exercise intensity in hematopoietic mobilization, to better understand the mobilization kinetics postexercise, and to determine if exercise is capable of mobilizing several specific populations of hematopoietic cells that have clinical relevance in a transplant setting. Healthy individuals were exercised on a cycle ergometer at 70% of their peak work rate (WR_peak) until volitional fatigue and at 30% of their WR_peak work matched to the 70% WR_peak bout. Blood was collected before, immediately post, and 10, 30, and 60 min postexercise. Total blood cells, hematocrit, and mononuclear cells isolated by density gradient centrifugation were counted. Specific populations of hematopoietic stem cells were analyzed by flow cytometry. Mononuclear cells, CD34⁺, CD34⁺/CD38⁻, CD34⁺/CD110⁺, CD3⁻/CD16⁺/CD56⁺, CD11c⁺/CD123⁻, and CD11c⁻/CD123⁺ cells per millilter of blood increased postexercise. Overall, the 70% WR_peak exercise group showed greater mobilization immediately postexercise, while there was no observable increase in mobilization in the work matched 30% WR_peak exercise group. Mobilization of specific populations of hematopoietic cells mirrored changes in the general mobilization of mononuclear cells, suggesting that exercise serves as a nonspecific mobilization stimulus. Evidently, higher intensity exercise is capable of mobilizing hematopoietic cells to a large extent and immediately postexercise is an ideal time point for their collection.

NEW & NOTEWORTHY Here we demonstrate for the first time that mobilization of hematopoietic stem cells (HSCs) through exercise is intensity dependent, with the greatest mobilization occurring immediately after high-intensity exercise. As well, we show that exercise is a general stimulus for mobilization: increases in specific HSC populations are reliant on general mononuclear cell mobilization. Finally, we demonstrate no differences in mobilization between groups with different aerobic fitness.

the hematopoietic system consists of mature blood cells, found mostly in circulation, as well as multiple types of stem and progenitor cells (HSCs), found predominantly in the bone marrow. The roughly 10 billion new blood cells required per day to maintain homeostasis are generated through the self-renewal, proliferation, and differentiation of these specific hematopoietic stem and progenitor cells (1). Hematological malignancies that disrupt this system are cured through HSC transplantation. Nearly 20,000 HSC transplants are performed each year in the USA (44). In recent years it has become increasingly evident that the quality of a blood graft extracted from a donor is critical for the success of the transplant outcome (26). All hematopoietic progenitor cell types must be adequately represented in a donor graft to properly restore hematopoiesis in a recipient.

In humans, HSCs are generally identified as side population cells from the bone marrow (45) or as CD34⁺ cells in peripheral blood (5). While HSCs in general can be detected with CD34, identifying specific subpopulations of CD34⁺ HSCs requires additional surface markers. For example, while total CD34⁺ content in a graft may be high, CD34⁺/CD110⁺ content, the cell population responsible for the development of platelets, may be low. The result of peripheral blood stem cell (PBSC) transplants with such grafts are recipient individuals who have an inability to properly regenerate platelets (34), resulting in bleeding, bruising, and susceptibility to infection, all of which are complications that can lead to death. Thus the development of mobilization regimes able to adequately mobilize both sufficient numbers and specific types of HSCs from bone marrow to blood is of incredible importance.

HSCs are retained in their niche through specific adhesion molecules and chemokine gradients (30). Although the majority of HSCs are found in the bone marrow, a small fraction of HSCs can be found in blood (41). In adult mice, it is estimated that ~1–5% of HSCs circulate between bone marrow and blood each day (4). However, depending on certain conditions, the number of mobilized cells can be greatly increased. In early development HSCs migrate from the fetal liver and spleen and populate the bone marrow (29). HSCs in adults migrate toward sites of injury or inflammation to aid in tissue repair (19). In a clinical setting, in lieu of extracting cells directly from bone marrow, HSCs are mobilized into circulation with granulocyte-colony-stimulating factor (G-CSF) to provide sufficient cells for a graft (16). G-CSF is a colony-stimulating hormone released endogenously by various cells throughout the body to stimulate production of granulocytes and it also has potent mobilization properties. While G-CSF is mainly used alone, many interleukins, chemokines, or hematopoietic growth factors can also be used in conjunction with G-CSF to bring about mobilization with various efficiencies and kinetics. However, there may be important limitations to pharmacological interventions. Considering that there can be negative side-effects from long-term G-CSF dosing (18) or that some clinical populations, such as individuals with cancer (31) or diabetes (12), may not respond well to G-CSF substitute and/or synergistic strategies should be investigated.

In this capacity, a lesser studied stimulus for HSC mobilization is exercise. Very little data exist on the effects of acute exercise or exercise training on mobilization of human HSCs (10). Exercise seems to acutely increase peripheral blood HSC quantity; however, firm conclusions are difficult to make when comparing studies due to the wide range of subject characteristics, lack of proper markers used to define HSCs, and timing of analysis. Bonsignore and colleagues compared the quantity of HSCs in peripheral blood, identified as CD34⁺ cells, between individuals competing in a 1.5-km field test (5). CD34⁺ cells as well as burst-forming unit erythroid (BFU-E) and colony-forming unit-granulocyte macrophage (CFU-GM), erythrocyte and granulocyte macrophage colonies measured using cell culture assays, increased in circulation after the field test (5). Morici and colleagues evaluated the quantity of HSCs in peripheral blood immediately before and following an all-out rowing test in competitive rowers (28). Although HSCs, identified as CD34⁺ cells, were increased in peripheral blood immediately following exercise in both males and females tested, the most primitive HSC population, CD34⁺/CD38⁻ cells, were unaffected by the acute exercise stimulus (28). Wardyn and colleagues compared HSC content in peripheral blood in previously sedentary individuals vs. habitual exercisers following an acute treadmill test (45). The authors concluded that HSC quantity in peripheral blood, identified there as side population cells, increased irrespective of training status and gender immediately postexercise (45). Zaldivar and colleagues examined the response of early vs. late pubertal boys to 20 min of high-intensity cycling and observed equal increases in peripheral CD34⁺ cells in both groups (50). Thijssen and colleagues observed a similar relative increase in circulating HSCs following an acute exercise stimulus in young and elderly individuals (40). When taken together, comparing the modalities of exercise used in these studies, it would seem that the acute increase in peripheral HSC quantity following exercise is intensity dependent.

Little knowledge exists, however, on the ability of exercise to mobilize specific populations of HSCs that are clinically relevant. We set out to address this. The purpose of this study was to further understand how exercise intensity effects HSC mobilization, to establish the kinetics of mobilization postexercise, to understand how HSC mobilization differs in individuals of differing aerobic capacity, and to determine the specificity of the stimulus by assaying several specific subpopulations of HSCs.

METHODS

Participants.

Eleven healthy young men (age 23.5 ± 2.9 yr) were recruited to participate in this study. Participants were asked to not strenuously exercise or remain recreationally active during their participation in the study. Exclusion criteria included smoking, diabetes, the use of nonsteroidal anti-inflammatory drugs (NSAIDs) and/or statins, and history of respiratory disease and/or any major orthopedic disability. The study was approved by the Hamilton Health Sciences Integrated Research Ethics Board and conformed to the guidelines outlined in the Declaration of Helsinki. Participants gave their informed written consent before their inclusion to the study. For subject characteristics, see Table 1.

Table 1.

Subject characteristics

Subject Characteristics	Means ± SD
Age	23.5 ± 2.9 yr
Height	178 ± 4 cm
Weight	77.3 ± 2.2 kg
V̇o2peak	54.3 ± 9.5 ml/(kg⋅min)
Average time to exhaustion (70% WRpeak trial)	17.8 ± 4 min
Worked-matched time (30% WRpeak trial)	41.4 ± 3 min

WR_peak, peak work rate.

Exercise protocol.

Each participant reported to the laboratory to perform a ramp incremental exercise test (50 W baseline for 3 min followed by a 30 W/min ramp) on an electronically braked cycle ergometer (Excalibur Sport V2.0; Lode, Groningen, The Netherlands) for determination of peak V̇o₂ (V̇o_2peak) and peak work rate (WR_peak). Expired gas and ventilatory parameters were collected for the determination of V̇o_2peak (Moxus Metabolic System; AEI Technologies, Pittsburgh, PA). Subjects were asked to maintain a cadence between 60 and 70 rpm during the test, and the test was terminated upon volitional fatigue. Following a 1-wk washout period, participants exercised at a workload pertaining to 70% of their WR_peak until volitional fatigue. Following another 1-wk washout period, participants performed a bout of cycling at 30% WR_peak, with the work output matched to the 70% WR_peak bout. These two intensities were chosen as they would likely fall within different very different exercise domains (46).

Blood collection and processing.

Fifteen milliliters of blood were collected via venous catheter before (Pre), immediately after (Post), and 10, 30, and 60 min postexercise using heparinized vacutainers (0268795, Fisher Scientific Canada).Whole blood total cell counts were done using a countess automated cell counter (Invitrogen, Carlsbad, CA). Hematocrit was measured with microhematocrit tubes (22-274-913; Beckman Coulter). Mononuclear cells were then isolated from the remainder of the blood using Ficoll-Paque Plus (14) (17-1440-02; GE Healthcare Life Sciences), were resuspended in 1% bovine serum albumin in PBS, and were counted using a Countess automated cell counter.

Flow cytometry.

Mononuclear cells were analyzed via flow cytometry (CyFlow Space, Partec) immediately after their isolation from blood. The following antibodies were used in the analysis: CD34 (FAB7227G; R&D Systems), CD110 (FAB1016A; R&D Systems), CD38 (FAB2404A; R&D Systems), CD41 (FAB7616A; R&D Systems), CD4 (FAB3791F; R&D Systems), CD8 (FAB1509A; R&D Systems), CD11c (FAB1777N; R&D Systems), CD123 (FAB301P; R&D Systems), and CD3/CD(15+56) cocktail (319101; BioLegend). Antibody concentrations used reflect manufacturer recommendations. At least 1 × 10⁶ cells were stained for each antibody, with at least 5 × 10⁵ events captured and analyzed. Unstained and single stain controls were used for compensation and gating. Final gates were based on FSC/SSC and two parameter plots.

Cell culture.

Isolated mononuclear cells were also used for the long-term culture initiating cell (LTC-IC) assay, a limiting dilution assay described in more detail elsewhere (43). In brief, following collection, mononuclear cells were plated on a feeder layer of the FBMD-1 cell line, grown for 35 days with weekly media changes, switched to Human Methylcellulose Complete Medium (HSC003; R&D Systems), and then grown for a further 14 days. The presence of colonies was then scored and LTC-IC frequency was calculated using L-Calc (Stem Cell Technologies, British Colombia, Canada).

Statistical analysis.

Statistical analysis was performed using Sigma Stat 11.0 analysis software (Systat Software, Chicago, IL). Data and graphs are expressed as means ± SE with P ≤ 0.05 considered significant. Statistical differences between time points and groups were determined using one and two way repeated measures ANOVA. Tukey corrections were applied to account for multiple comparisons. LTC-IC data analyzed with a one-tailed t-test, justified as the culture experiments were performed after the majority of the flow cytometry had been analyzed. Statistical differences in V̇o_2peak subgroups were analyzed by t-test. Differences in mononuclear mobilization between V̇o_2peak subgroups were analyzed with a three way ANOVA.

RESULTS

Exercise did not alter total blood cell counts or hematocrit.

Total blood cell counts and hematocrit were measured for each time point Pre and Post exercise. No significant differences were detected at any time point Pre or Post exercise at either intensity (data not shown).

Mononuclear cell numbers per milliliters of blood increased Post exercise in an intensity-dependent manner.

Although blood and hematocrit were not measurable different, mononuclear cell number was greatly impacted by exercise. Mononuclear cells were isolated via density gradient centrifugation. Peripheral blood mononuclear cells consist of monocytes, dendritic cells, lymphocytes, and a tiny fraction of multipotent and more differentiated HSCs. In the 70% WR_peak trial mononuclear cells increased by 139% from Pre to immediately Post exercise (P < 0.05) (Fig. 1). No significant differences in mononuclear cell number were observed in the 30% WR_peak trial, at any time point. Mononuclear cell number was increased by 96% at Post in the 70% WR_peak trial when compared with Post in the 30% WR_peak trial (P < 0.05).

Fig. 1.

Mononuclear cells. Mononuclear cell counts, as isolated through density gradient centrifugation, expressed per milliliters of blood for the 70% WR_peak and 30% WR_peak exercise sessions across the time course (where WR_peak is peak work rate). *P < 0.05, between groups and from before (Pre) exercise with respect to time.

Exercise acted as a general stimulus; many specific populations of HSCs and leukocytes are mobilized in the same magnitude as mononuclear cells.

Several specific populations of HSCs and white blood cells were measured with flow cytometry. Many of these populations were chosen as they represent specific populations needed in a blood cell graft to fully reconstitute the blood cell pool of a transplant recipient postmyeloablation.

CD34⁺ cells represent general hematopoietic progenitors and are used as the standard measure of graft quality. In the 70% WR_peak trial CD34⁺ cells per milliliters of blood increased from Pre to immediately Post by 98.4% (P < 0.05) (Fig. 2). No differences were detected between Pre, Post, or any other time point postexercise following the 30% WR_peak trial. Finally, CD34⁺ cells per milliliters of blood at Post in the 70% WR_peak trial were 170% higher than at Post in the 30% WR_peak trial (P < 0.05).

Fig. 2.

CD34⁺ cells per milliliters blood. CD34⁺ cell counts, analyzed with flow cytometry, expressed per milliliters of blood for the 70% WR_peak and 30% WR_peak exercise sessions across the time course. *P < 0.05, between groups and from Pre exercise with respect to time.

CD34⁺/CD38⁻ represent a more primitive and undifferentiated CD34⁺ population. In the 70% WR_peak trial CD34⁺/CD38⁻ cells per milliliters of blood increased by 88% (P < 0.05) from Pre to immediately Post exercise (Fig. 3A). Again, no differences were detected following the 30% WR_peak trial at any time point. Finally, CD34⁺/CD38⁻ cells per milliliters of blood at Post in the 70% WR_peak trial were 106% higher than at Post in the 30% WR_peak trial (P < 0.05).

Fig. 3.

Specific hematopoietic stem cell (HSC) populations. Five different blood cell populations, analyzed by flow cytometry, known to positively correlate with successful outcomes post blood cell graft. A: CD34⁺/CD38⁻. B: CD3⁻/CD(16+56)⁺. C: CD4⁺/CD8⁺ ratio. D: CD34⁺/CD41⁺. E: CD34⁺/CD110⁺. Each expressed as positive cells per milliliters of blood for the 70% WR_peak and 30% WR_peak exercise sessions, across the time course. *P < 0.05,between groups and from Pre exercise with respect to time.

CD3⁻/CD(16+56)⁺ characterize natural killer cells. CD3⁻/CD(16+56)⁺ cells per milliliters of blood increased by 291% (P < 0.05) from Pre to Post exercise following the 70% WR_peak trial (Fig. 3B) and then returned to Pre exercise levels. No differences were detected at any other time point in the 70% WR_peak trial or at any time point in the 30% WR_peak trial. As well, CD3⁻/CD(16+56)⁺ cells per milliliters of blood at Post in the 70% WR_peak trial were 293% higher than at Post in the 30% WR_peak trial (P < 0.05).

The CD4⁺/CD8⁺ lymphocyte ratio is a general measure of immune function, and here, a predictor of graft quality. The calculated CD4⁺/CD8⁺ cell ratio was maintained around an ideal 2 at all measured time points (average 70% WR_peak trial: 2.36 ± 0.44; average 30% trial: 1.83 ± 0.65) and did not drop below that level (Fig. 3C).

CD34⁺/CD41⁺ and CD34⁺/CD110⁺ represent megakaryocyte progenitors; cell types that work to reconstitute platelets in a recipient. While numbers of circulating CD34⁺/CD41⁺ cells did not change (Fig. 3D), circulating CD34⁺/CD110⁺ cells increased. CD34⁺/CD110⁺ cells per milliliters of blood increased from Pre to Post in the 70% WR_peak trial by 169% (P < 0.05) (Fig. 3E). No changes were detectable in the 30% WR_peak trial. As well, CD34⁺/CD110⁺ cells per milliliters of blood at Post in the 70% WR_peak trial were 129% higher than at Post in the 30% WR_peak trial (P < 0.05).

CD11c⁺/CD123⁻ and CD11c⁻/CD123⁺ demarcate dendritic type 1 and type 2 cells, respectively. CD11c⁺/CD123⁻ cells per milliliters of blood increased by 149% (P < 0.05) from Pre to Post exercise following the 70% WR_peak trial (Fig. 4A). Following the established trend, no differences were detected following the 30% WR_peak trial at any time point. As well, CD11c⁺/CD123⁻ cells per milliliters of blood at Post in the 70% WR_peak trial were 77% higher than at Post in the 30% WR_peak trial (P < 0.05).

Fig. 4.

Dendritic cell populations. Two dendritic cell populations, analyzed by flow cytometry, known to positively correlate with successful outcomes post blood cell graft. A: CD11c⁺/CD123⁻. B: CD11c⁻/CD123⁺. Each expressed as positive cells per milliliters of blood for the 70% WR_peak and 30% WR_peak exercise sessions, across the time course. *P < 0.05, between groups and from Pre exercise with respect to time.

CD11c⁻/CD123⁺ cells per milliliters of blood increased by 165% (P < 0.05) from Pre to Post exercise following the 70% WR_peak trial (Fig. 4B) and remained significantly higher than Pre until 30 min postexercise. Again, no differences were detected following the 30% WR_peak trial at any time point. Finally, CD11c⁻/CD123⁺ cells per milliliters of blood at Post in the 70% WR_peak trial were 168% higher than at Post in the 30% WR_peak trial (P < 0.05) and were also significantly higher at 10 and 30 min following exercise between the 70% WR_peak and 30% WR_peak trials.

The same number of mononuclear cells from each exercise session and each collection time point were used for the flow cytometry analysis. In all cases, there were no detectable differences in the raw number of positive cells (data not shown). Only when expressing the populations as positive number of cells per milliliters of blood, taking into account the increased mononuclear cell numbers in blood, were any statistical differences detected. The differences are thus highly dependent on the general mobilization of mononuclear cells, rather than the mobilization of specific subpopulations of cells.

Long-term culture initiating cell frequency increased in blood in an intensity-dependent manner.

LTC-IC were grown from mononuclear input cells and enumerated following well established tissue culture protocols (43). These cells represent primitive HSCs and are quantified based on their ability to produce colonies in culture after 5 wk. Again, the results here follow a similar trend to the results mentioned previous to this point. LTC-IC number per milliliters of blood increased by 713% (P < 0.05) from Pre to immediately Post exercise following the 70% WR_peak trial (Fig. 5). No differences were detected between Pre and Post exercise following the 30% WR_peak trial.

Fig. 5.

Long-term culture initiating cell frequency (LTC-IC), expressed as number of cells per milliliters blood, Pre and immediately Post exercise for the 70% WR_peak and 30% WR_peak exercise sessions. *P < 0.05, between groups and between Pre and immediately after (Post) exercise with respect to time.

Fitness level had no bearing on the ability to mobilize mononuclear cells to the blood cell pool.

Of the 11 total subjects, four had a significantly higher V̇o_2peak (P > 0.05) than the average of the entire group. These four subjects [average V̇o_2peak of 66.375 ± 4.35 ml/(kg·min)] were split from the remaining seven subjects [average V̇o_2peak of 49 ± 4.94 ml/(kg·min)] to determine if fitness level had an impact on the magnitude of mobilization. Though the subgroup with the higher fitness level performed more total work than the subgroup with the lower fitness level during both stages (70% WR_peak and 30% WR_peak work matched) of the test, no significant differences were detected between groups in total blood cell counts (data not shown), hematocrit (data not shown), mononuclear cell counts (Fig. 6), or specific HSC populations at any time point (data not shown). The only detectable differences were as before; Pre and immediately Post exercise in the 70% WR_peak trial for each fitness subgroup.

Fig. 6.

Fitness comparison. A comparison between mononuclear cell counts, as isolated by density gradient centrifugation, expressed per milliliters of blood for the 70% WR_peak and 30% WR_peak exercise sessions, including the two fitness subgroups, across the time course. *P < 0.05, within fitness level subgroups for intensity, but no differences were detected between the 2 fitness subgroups.

DISCUSSION

Here we demonstrate that high-intensity endurance exercise is capable of mobilizing hematopoietic cells from bone marrow to blood and that the ideal time point for their collection is immediately following exercise. Lower intensity exercise work matched to the higher intensity test failed to mobilize hematopoietic cells in all tested measures, suggesting that intensity of exercise is far more important than total work. As well, we demonstrate that exercise mediated mobilization seems to be a general process; increases in mobilization of specific populations of HSCs were dependent on overall mononuclear cell mobilization. Finally, we demonstrate that between two differing fitness levels no differences in the extent of exercise-mediated mobilization could be determined.

In general, the observed lack of mobilization in the 30% WR_peak work-matched trial, when compared with the 70% WR_peak until volitional fatigue trial, demonstrates that intensity may be a key factor in exercise induced mobilization. While the precise mechanism of exercise induced hematopoietic mobilization remains largely unknown, that intensity plays role is no surprise. Exercise can have dramatic effects on the body. For example, increases of ~8,000-fold in circulating levels of IL-6 have been observed in athletes after finishing a marathon race (24). Large increases in TNF-α, IL-1β, and IGF-binding protein 1, which stimulates inflammatory cytokine production, have also been observed (33). Increased production of cytokines and mobilizing factors both within the bone marrow niche and in circulation in response to exercise has been suggested as a mechanism for exercise-induced hematopoietic mobilization (11). Consequently, increased exercise intensity, leading to the subsequent release of greater amounts of these factors, followed by increased mobilization is no surprise. Increased blood flow within bone during high-intensity exercise (36, 37) may also lead to increase HSC efflux from marrow. As mobilization was seen to peak immediately postexercise, presumably when many of these factors are at their highest and bone marrow blood flow is most increased, the results make sense. However, while the purpose of the present study was to observe how exercise intensity affects HSC mobilization, future studies should aim to address the notion that fatiguing exercise vs. nonfatiguing exercise may impact mobilization, regardless of intensity.

Ensuring that clinically relevant HSC subpopulations are present in mobilized PBSC grafts is important for HSC transplant success. Exogenous treatment of clinical donors with G-CSF for several days before cell collection can raise circulating CD34⁺ cells anywhere from two- to fourfold (21, 23) to as much as 20-fold (6), though the extent of mobilization varies greatly on an individual basis. Unfortunately, G-CSF treatment regimens often fail. When considering CD34⁺ content alone, mobilization regimes have a 30% failure rate among healthy individuals and a greater than 60% failure rate in high-risk patients, such as those with lymphoid malignancies, prior exposure to chemotherapeutic agents, or radiation exposure (15, 27, 42). Poor CD34⁺ G-CSF induced mobilization is also observed with increasing donor age (38). Furthermore, G-CSF treatment can also fail to mobilize other clinically relevant HSC populations, such as long-term CD34⁺/CD38⁻ cells (13) and B and T lymphocytes (21). An additional method for mobilizing HSCs, such as the high-intensity exercise intervention presented here, could potentially synergize with more traditional strategies. Here we demonstrate that high-intensity exercise induced a twofold increase in circulating CD34⁺ cells (Fig. 2). However, when comparing the magnitude of this increase to the increase in circulating CD34⁺ cells in healthy individuals mobilized with G-CSF [average 15-fold increase (39) or an average 23-fold increase (20)], exercise falls on the low end of the reported G-CSF spectrum.

Mobilization of clinically relevant HSC subpopulations in a donor can mean the difference between transplant success or transplant failure. The more cells that are mobilized, the easier it is for clinicians to harvest threshold doses necessary for transplant success. While CD34⁺ cell dose is largely considered an important determining factor for transplant success (35), higher graft content of more specific HSC subpopulations has been shown to correlate with better PBSC transplant outcomes. PBSC grafts with a greater content of more primitive populations of HSCs, identified as CD34⁺/CD38⁻ cells, have been shown to correlate to greater overall recipient survival (17). Here, we demonstrate that exercise is capable of mobilizing these cells to a large extent (Fig. 3A). Immune reconstitution posttransplant is critically important, helping myeloablated recipients to fight off infections and stay healthy. CD3⁻/CD(16+56)⁺ natural killer cell graft content also correlates with successful immune reconstitution (32). Again, high-intensity exercise appears to be capable of mobilizing this population (Fig. 3B). Increased lymphocyte dosage during a transplant (CD4⁺ and CD8⁺ cells) correlates with increased recipient neutrophil recovery and decreased infection rates (25). A high CD4/CD8 ratio, however, can predict adverse immune-related outcomes posttransplantation (49). Here we demonstrate exercise induced increased in numbers of CD4⁺ and CD8⁺ cells maintained at or near an ideal ratio of 2 (Fig. 3C). Finally, increased graft content of another CD34⁺ subset, CD34⁺/CD110⁺ platelet progenitors, correlates with increased platelet reconstitution posttransplant (34). As well, increased CD34⁺/CD41⁺ graft content can predict increased platelet and neutrophil recovery posttransplant (9, 26). In this capacity, high-intensity exercise appears to result in a nonsignificant trend toward mobilization of more CD34⁺/CD41⁺ cells over time and a significant increase in mobilization of CD34⁺/CD110⁺ cells (Fig. 3, D and E). As well, high-type one dendritic cell content (CD11c⁺/CD123⁻), high-type two dendritic cell content (CD11c⁻/CD123⁺), and total dendritic cell content have all been shown to be predictive of overall survival posttransplantation (8); interestingly, we observe increases in both CD11c⁺/CD123⁻ and CD11c⁻/CD123⁺ cells following high-intensity exercise (Fig. 4, A and B). In addition, several cell culture colony forming assays can be used to predict recipient survival. Increased numbers of LTC-ICs in a given graft has been correlated with increased recipient survival (3) and we have demonstrated the capacity of exercise to mobilize more cells of this type (Fig. 5).

Oftentimes hematopoietic donors are in poor health, have a relatively lower fitness level, and the exercise tolerance of such individuals is questionable. Determining how fitness level effects mobilization is therefore important and clinically relevant. To this end, subjects were split into a relatively higher fitness level [V̇o_2peak: 66.375 ± 4.35 ml/(kg·min)] and relatively lower fitness level [V̇o_2peak: 49 ± 4.94 ml/(kg·min)] in postanalysis. Given previous findings that long-term endurance exercise can greatly increase the HSC pool (2, 7), it was expected that the higher fitness level group would show a larger magnitude of mobilization than the lower fitness level group. This was not the case (Fig. 6). As well, a correlation analysis between V̇o_2peak and circulating mononuclear cell number also revealed no relationship (data not shown). Differences in fitness levels appear to have little bearing on the magnitude of mononuclear cell mobilization. This result carried over to the analysis of the specific HSC populations (data not shown). Although the average V̇o_2peak of our subjects was higher than those of sedentary individuals, and individuals at a relatively lower fitness level may mobilize differently due to their health, what this suggests is that fitness level may be unimportant. Unhealthy individuals, with a fitness level lower than average, may mobilize HSCs to the same extent as normal individuals. Again, the fitness level of these two subgroups was higher than average; this may not necessarily translate to individuals with a lower than average V̇o_2peak. However, populations of the elderly, sick, or otherwise impaired needing to donate blood for a PBSC graft may potentially be able to exercise at a high level of intensity, relative to their low V̇o_2peak and, in combination with traditional mobilization strategies, mobilize sufficient HSCs to collect a successful graft. Indeed, the safety and feasibility of exercise in patients undergoing HSC transplant have previously been demonstrated (47). As well, higher intensity exercise protocols have been well tolerated in a number of clinical populations (22, 48), suggesting that these protocols are viable in nonhealthy individuals.

In conclusion, high-intensity endurance exercise may serve as an aid to more traditional HSC mobilization strategies. It is often the case that some patients mobilize poorly; in such situations, additional chemokines and cytokines are used with traditional mobilizing agents (42). While these optimizations help in some situations, developing additional methods will always be useful for healthcare providers. In isolation, the magnitude of exercise-induced HSCs mobilization falls on the low end of G-CSF-based mobilization. Although leukapheresis postexercise, repeated over the course of several days, could in theory harvest enough HSCs for a suitable graft, combining exercise with traditional mobilization strategies would be a better option. Based on our findings, although limited by a relatively low sample size, exercise could be considered as a means to enhance HSC content in a donor’s blood graft for patients undergoing PBSC transplantation.

GRANTS

This work was supported by an NSERC Discovery grant (1455843) awarded to G. Parise.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

J.M.B. and J.P.N. performed experiments; J.M.B., J.P.N., and G.P. analyzed data; J.M.B., J.P.N., and G.P. interpreted results of experiments; J.M.B. prepared figures; J.M.B. and G.P. drafted manuscript; J.M.B., J.P.N., and G.P. edited and revised manuscript; J.M.B., J.P.N., and G.P. approved final version of manuscript.