Journal Logo


Effects of Obesity and Exercise on Bone Marrow Progenitor Cells after Radiation


Author Information
Medicine & Science in Sports & Exercise: June 2019 - Volume 51 - Issue 6 - p 1126-1136
doi: 10.1249/MSS.0000000000001894
  • Free


There are currently 14 million cancer survivors in the United States, and this number is expected to increase in the coming years (1). Radiation therapy is commonly used in cancer treatment, with two-thirds of patients receiving radiation therapy (1). Despite recent technological advances in radiation therapy and improvements in radiation shielding, injury to healthy tissue remains an unavoidable side effect of radiation therapy. Further, with nearly 60% of the U.S. population being overweight/obese, and with a cancer diagnosis often accompanied by a decrease in physical activity levels, the number of obese and inactive cancer survivors is rapidly increasing (2). Thus, evaluating the long-term effects of obesity and exercise on the late effects of radiation therapy will have direct effects on cancer survivors and people exposed to higher doses of radiation.

Hematopoietic stem and progenitor cells (HSPC), the most primitive cells of the hematopoietic system from which all mature blood cells are derived, are particularly susceptible to oxidative stress (3). Radiation induces DNA damage in HSPC that can be passed down to their progeny, resulting in the progression of cellular damage to cells that were never exposed to radiation (3). In addition to the direct effects of radiation on HSPC, radiation also induces long-term alterations to their microenvironment in the bone marrow. Within the bone marrow, HSPC reside in a complex niche that consists of a variety of cell types that regulate hematopoiesis. Endothelial cells, mesenchymal stromal cells (MSC), and their progeny including adipogenic and osteogenic lineage cells populate the bone marrow and are major contributors to the HSPC niche (4). MSC represent a heterogeneous cellular population that expresses platelet-derived growth factor receptor-α (PDGFRα), leptin receptor, nestin, CD51, and Sca-1 (4). MSC regulate HSPC fate primarily via paracrine mechanisms (4) and are relatively resistant to radiation (5). Importantly, long-term proinflammatory alterations to stromal cell populations in the bone marrow after radiation exposure has been linked to impaired hematopoiesis long after radiation exposure (6). Together, the long-term inflammatory effects of radiation exposure on the bone marrow, the contribution of MSC to inflammation, the important role MSC play in regulating hematopoiesis, and the long-term persistence of MSC populations after radiation exposure suggest that further investigation into the lasting effects of MSC radiation exposure is warranted.

Obesity and exercise have been suggested to have differential effects on hematopoiesis and the HSPC niche (7). Obesity increases HSPC proliferation resulting in decreases in more primitive HSPC populations (8), whereas exercise training results in HSPC expansion without a reduction in repopulating capacity (9,10). These effects on hematopoiesis have been linked to alterations in the HSPC niche. Specifically, exercise training prevents the increase in marrow adipose tissue (MAT) induced by high-fat diet (11), in part by promoting osteogenic differentiation of MSC (8,9,11). Further, exercise training reduces inflammatory cytokine production. It remains unknown, however, whether obesity and exercise differentially influence HSPC populations and the bone marrow microenvironment after radiation exposure. This information would be valuable for devising long-term strategies to mitigate the late effects of radiation therapy on the hematopoietic system. Thus, the present study was designed to investigate how exercise training and obesity modulate HSPC and the bone marrow environment after exposure to radiation. On the basis of the current knowledge, we hypothesized that obesity exacerbates the responses after radiation, including alterations in HSPC and stromal cell populations, and increases bone marrow inflammation. Conversely, we predicted that exercise training would alleviate the damage in bone marrow induced by radiation and obesity.



All protocols were approved by the Illinois Institutional Animal Care and Use Committee. Male CBA (Jackson Laboratories, Bar Harbor, ME) mice were maintained in a 12-h light–12-h dark schedule with food and water provided ad libitum. CBA mice develop acute myeloid leukemia 18 months after radiation exposure (12). Mice were received at 4 wk old and allowed 1 wk to acclimate to the facility. Mice were randomly divided between control (CON; n = 20) and high-fat diet groups (HF; n = 20). CON mice received the control diet with 16% kcal from fat (AIN-93 M; Research Diets, New Brunswick, NJ), whereas HF mice were given a high-fat diet with 45% kcal from fat (D12451; Research Diets, New Brunswick, NJ) throughout the study.

Body weight and composition

Body weight was assessed weekly in experimental mice using a standard scale. Body composition was assessed monthly using an EchoMRI-100 (Echo Medical Systems, Houston, TX) according to previous protocols (13,14). Mice were placed in a specialized tube and not anesthetized during body composition analysis.

Exercise protocol

We used a progressive treadmill exercise program that has been previously shown to expand the HSPC pool and decrease MAT (9,10,14,15). At 9 wk of age, mice from both the CON and the HF groups were further randomly divided into sedentary (SED, n = 10) and exercise (EX, n = 10) groups (Fig. 1A). Briefly, mice in CON-EX and HF-EX groups were exercise trained 3 d·wk−1 (M/W/F) progressing to 1 h·d−1 until sacrifice. All mice were exercised during the day cycle at the same time of day throughout the study. Each exercise session began with a 5-min warm-up period at 8 m·min−1, followed by a 30- to 45-min training period and concluded with a 5-min cool-down period at 8 m·min−1. Treadmill speed during the training period progressively increased starting at 10 m·min−1 for 25 min to a maximum of 23 m·min−1 for 45 min. Exercise sessions occurred at the same time of day, and mice were encouraged to run by stimulation with bristles of a paintbrush as we have found this method to be more effective and humane than electric shock. Nonexercised mice were exposed to treadmill noise, placed in the confined treadmill lanes, and manually manipulated in the same manner/time as exercise-trained mice to control for any stress associated with the treadmill apparatus. One CON + EX mouse passed after irradiation during an exercise session for unknown causes.

High-fat diet–induced obesity is not attenuated with exercise. A, A representative schematic of the study design. Mouse body weight (B) and body fat percentage (C) measured throughout the study. Data are presented as mean ± SEM (n = 9–10 per group). *P < 0.05 significant difference between CON + EX compared with HF + SED and HF + EX. B P < 0.05 significant difference of CON + SED compared with HF + SED and HF + EX. HbA1c percentage (D), HbA1c concentration (E), and hemoglobin concentration (F) were compared at end point. Data are presented as mean ± SEM (n = 6 per group), with *P < 0.05 indicating a main effect of exercise and γ indicating a main effect of diet. CON + SED, control, sedentary; CON + EX, control, exercise trained; HF + SED, high-fat diet, sedentary; HF + EX, high-fat diet, exercise trained; Hb, hemoglobin.

Total body irradiation

At 13 wk of age, all mice were administered a 3-Gy dose of γ-radiation. The radiation was administered as previously described with minor modifications (10,15). Briefly, radiation was administered with a Theratron 780 Cobalt 60 radiation unit at a fixed dose rate of 37.7 cGy·min−1 for 8 min. Mice were anesthetized with ketamine/xylazine (65/4 mg·kg−1) by intraperitoneal injection to prevent them from moving during radiation treatment. A uniform dose distribution across the radiation field was applied. After total body irradiation, mice were provided Jell-O for 1 wk in addition to their diet to ensure sufficient calories. Mice were evaluated daily during recovery. Mice continued with their respective interventions after radiation until sacrifice at 4 wk after radiation exposure. As such, this study design allowed investigations of the early changes in the bone marrow that could contribute to radiation-induced leukemia development.

Serum sample preparation and biochemical parameters

Whole blood was allowed to clot for at least 30 min at room temperature, then centrifuged at 2000g for 30 min at 4°C. After centrifuge, supernatants were immediately separated as serum samples and transferred to a new 1.5-mL microcentrifuge tube at −80°C until use. Commercial enzyme-linked immunosorbent assay kits were used to measure the concentration of glycated hemoglobin (HbA1c; Aviva Systems Biology Corp., San Diego, CA) and total hemoglobin (Hb; Abcam, Cambridge, MA) in mouse serum, according to manufacturer’s protocols. Optimal serum sample dilutions were determined from preliminary experiments. Percentage of HbA1c (HbA1c %) was calculated by dividing the concentration of plasma HbA1c by the concentration of plasma Hb.

Bone marrow cellularity

Four weeks after radiation exposure, mice were euthanized via CO2 asphyxiation followed by decapitation. Both femurs and one tibia were quickly removed and cleared of muscle and connective tissue. Subsequently, mouse bones were flushed in 1 mL of phosphate buffered saline (PBS) to retrieve bone marrow cells as previously described (15,16). An aliquot of the bone marrow cell suspension was separated, diluted 1:100 in PBS, and mixed with trypan blue to identify dead cells. All viable cells were counted with a Countess Automated Cell Counter (Life Technologies, Carlsbad, CA). The remaining bone marrow cells were centrifuged at 400g at 4°C. The supernatant from the flushed bone marrow samples (i.e., bone marrow supernatant) was separated from the cellular fraction, snap frozen, and stored at −80°C for future analysis. Concurrently, bone marrow cells were resuspended in 1 mL of 5% fetal bovine serum/PBS and placed on ice for flow cytometric analysis.

HSPC and MSC quantifications in bone marrow

Mouse bone marrow cells were collected for the characterization of HSPC and stromal cell populations. Flushed bones were mechanically and enzymatically digested as previously described (16). Briefly, femurs and tibias were gently crushed using mortar and pestle in Dulbecco’s modified Eagle’s medium. Bone fragments were further processed by cutting with scissors and suspended into a 0.2% collagenase solution for 1 h at 37°C. Cells and bone fragments were filtered through a 40-μm cell strainer. Cells liberated from bone digestion were recombined with cells previously flushed. MSC were stained from whole bone marrow isolates. For HSPC staining, bone marrow cells were enriched through magnetic cell sorting (MACS) using the EasySep Mouse Hematopoietic Progenitor Cell Negative Enrichment Kit (Stemcell Technologies, Vancouver, Canada) per manufacturer’s instruction. Cells (5 × 106) were stained using the HSPC and MSC markers (Table 1).

After staining, cells were fixed in 10% formalin and permeabilized in PBS 0.01% triton. Nestin+ MSC went through the same staining protocol as all MSC populations but were stained with nestin after cell fixation. After phenotypic staining, cells were incubated with CellROX® Green Reagent (Life Technologies) for 30 min at 37°C for oxidative stress quantification. Cells were washed, suspended, and placed on ice for evaluation via the Attune Acoustic Flow Cytometer (Life Technologies). Compensation and gating strategies were derived from unstained and single stained controls. Total cell quantities of HSPC and subpopulations were determined by measuring the percentage of detected cells × the total number of cells obtained from MACS. MSC and stromal cell populations were calculated by measuring the percentage of detected cells × total number of cells obtained from bone marrow flushing and crushing.

Quantification of MAT

Osmium tetraoxide staining was used for the quantification of MAT as previously described (17). Briefly, one tibia was fixed with 10% neutral buffered formalin for 24 h. Subsequently, tibias were washed in water and decalcified in 14% EDTA/PBS at room temperature for 2 wk. The bones were then washed, and the distal joint was removed. MAT staining was performed with 2 mL of 1% osmium tetroxide/2.5% potassium dichromate for 48 h at room temperature. Tibias were washed in water and stored in PBS until imaging. MAT imaging was performed with the MicroXCT-400 (Zeiss, Oberkochen, Germany). Tibias were aligned to center frame of view and analyzed at a voltage of 40 kV and 8 W. Tibias were exposed for 1 s and analyzed for 360° at 1° × 1°. Images were reconstructed and analyzed using Amira imaging software (Fei Biotechnologies, Carlsbad, CA). Total MAT was expressed as a percentage of total bone volume. During analysis, the researcher was blinded to experimental conditions and treatment groups.

Cytokine array analysis

Bone marrow supernatants previously obtained during bone marrow cell isolation were analyzed via a C-Series Mouse Inflammation Array C1000 (RayBiotech, Norcross, GA) per the manufacturer’s protocol. Assays were visualized using 201 ChemiDoc XRS camera (BioRad, Hercules, CA) and quantified using ImageJ (National Institutes of Health, Rockville, MD). Pixel density was subtracted from background and expressed as fold change from CON samples. Data were then normalized using log2‐fold change versus CON + SED. Cytokines with log2‐fold changes greater than 50% (i.e., <−1 or >1) were considered biologically relevant. Only cytokines that were detectable are reported.

Leukemia blast viability assay

KG-1a myeloblasts (ATCC), originally derived from a patient with acute myeloid leukemia, were used as an in vitro model of leukemia blast growth (18). KG-1a cells were expanded in growth media (20% fetal bovine serum/1% penicillin–streptomycin/IMDM) and used at passage 11 for analysis. For viability analysis, KG-1a cells were plated in equal numbers in 96-well plates with serum-free media, KG-1a growth media, or pooled bone marrow supernatant from each group. Cells were incubated in each condition for 1 (i.e., baseline) to 72 h. The 72-h time point resulted in cell detachment due to overconfluence and was discarded. After incubation for each time point, viability was assessed using the MTT Cell Growth Assay Kit (Sigma-Aldrich, St. Louis, MO) per manufacturer’s instructions and as previously described (19,20). Data are expressed as fold change from baseline within each condition.

Statistical analysis

The Shapiro–Wilk test was used to test for normality within each data set. Nonnormal data or unequal variance were log transformed. Body weight and composition were analyzed via a repeated-measures three-factor ANOVA (diet–exercise–time) in SPSS. Cell populations, HbA1c, Hb, MAT, and leukemia blast viability were analyzed by two-factor ANOVA (diet–exercise) in GraphPad Prism. For cellular populations, statistically significant interactions were evaluated by Tukey’s multiple comparison test. For the cytokine array, heatmap analysis was performed using the “ggplot2” package in R 3.3.4. All analyses were performed by investigators blinded to experimental group.


High-fat diet is effective at inducing an obesity phenotype that is not attenuated with exercise

A schematic overview of the present study is presented in Figure 1A. Mice on the 45% high-fat diet (i.e., both HF + SED and HF + EX) had a significant higher body weight when compared with their respective control groups at weeks 6–13 (CON + SED and CON + EX, P < 0.007; Fig. 1B). No significant differences were observed between HF + SED and HF + EX mice throughout the study. Similarly, percent body fat was significantly higher in HF + SED and HF + EX mice compared with CON + EX mice at week 4 and by week 8 compared with CON + SED and CON + EX mice (P < 0.01; Fig. 1C). At week 12, both HF + SED and HF + EX mice had significantly higher body fat compared with mice consuming the control diet. At sacrifice, there was no significant difference in the percentage of glycated hemoglobin (HbA1c) between groups; however, a trend for a significant interaction was identified (P = 0.064; Fig. 1D). Before radiation, HbA1c percentage was significantly higher in high-fat diet–fed mice compared with control diet–fed mice (data not shown). Exercise-trained mice had significantly lower levels of plasma HbA1c compared with sedentary mice (P < 0.05; Fig. 1E). Main effects for both exercise training and high-fat diet were observed for hemoglobin concentrations (both P < 0.05; Fig. 1F).

Exercise and obesity have opposite effects on HSPC content after radiation

Total HSPC pool (lineageSca-1+cKit+ [LSK]), subpopulations, and progenitor cells such as long-term hematopoietic stem cell (LT-HSC; LSK, CD150+/CD48), short-term hematopoietic stem cell (ST-HSC; LSK, CD150/CD48), multipotent progenitor (MPP; LSK, CD150/CD48+), common lymphoid progenitor (CLP; lineageSca1+), common myeloid progenitor (CMP; lineagecKit+), and reactive oxygen species in HSPC were measured via flow cytometry. There were significantly more HSPC in exercise-trained mice and significantly fewer HSPC in obese mice (P < 0.05 main effect of exercise and diet; Fig. 2A). Oxidative stress was significantly lower in HSPC from exercise-trained mice (P < 0.05 exercise main effect; Fig. 2B). There were significantly more LT-HSC, ST-HSC, MPP, and CLP in exercise-trained mice (all P < 0.05 exercise main effect; Fig. 2C–F). There were significantly fewer MPP and CLP in obese mice (both P < 0.05 diet main effect; Fig. 2E and F). No differences were observed in CMP in either diet or exercise condition (Fig. 2G).

Exercise increases bone marrow cellularity and HSPC after sublethal irradiation regardless of diet. Quantification of LSK (A), median fluorescence intensity of reactive oxygen species levels in LSK (B), LT-HSC (C), ST-HSC (D), MPP (E), CLP (F), and CMP (G). HSPC and progenitor cells were quantified from cells obtained from MACS depletion of lineage negative cells via flow cytometry. Data are presented as mean ± SEM (n = 9–10 per group). *P < 0.05 main effect of exercise. #P < 0.05 main effect of diet.

Exercise and obesity have opposite effects on MAT and bone marrow cellularity after radiation

We quantified how exercise training and obesity modulate MAT accumulation after the sublethal IR exposure (Fig. 3A). Treadmill exercise training decreased tibial MAT in both CON and HF groups (P < 0.05 exercise main effect; Fig. 3B) compared with respective sedentary mice after radiation exposure. Conversely, obesity increased MAT in both HF groups (P < 0.05 diet main effect; Fig. 3B) compared with mice consuming the CON diet. Similarly, exercise training significantly increased bone marrow cellularity, whereas obesity reduced it (P < 0.05 exercise and diet main effects; Fig. 3C).

MAT is decreased in exercise-trained mice and increased in obese mice after sublethal irradiation exposure. Quantification of MAT via osmium tetroxide staining and microCT. A, Representative μCT images of osmium tetroxide stained tibias for MAT quantification. B, Graphical representation of MAT expressed as a percentage of whole bone volume. Data are presented as mean ± SEM (n = 5 per group). C, Quantification of bone marrow cellularity. Data are presented as mean ± SEM (n = 9–10 per group). *P < 0.05 exercise main effect, #P < 0.05 diet main effect.

Exercise training and obesity differentially influence bone marrow stromal cell content and oxidative stress after radiation

MSC represent a heterogeneous population within the bone marrow that possesses immunomodulatory properties. We quantified multiple stromal cell populations, including nestin+ MSC (CD45Ter119Nestin+) (21), PDGFRα+ MSC (CD45Ter119 CD31PDGFRα+) (22), osteoblasts (CD45Ter119CD31CD51+), adipocyte progenitors (CD45Ter119CD31CD51Sca1+) (23), and endothelial progenitors (CD45Ter119CD51CD31+Sca1+) (24) by flow cytometry. We further examined oxidative stress in each of the above cell populations via flow cytometry. CON + SED mice had significantly fewer nestin+ MSC compared with all other groups (P < 0.05; Fig. 4A), whereas CON + EX had significantly more nestin+ MSC compared with all other groups (P < 0.05; Fig. 4A). Nestin+ MSC from CON + EX had significantly higher oxidative stress than all other groups (P < 0.05; Fig. 4B). There were significantly fewer PDGFRα+ MSC in obese mice (P < 0.05; Fig. 4C) with significantly lower levels of oxidative stress (P < 0.05; Fig. 4D). There were significantly more osteoblasts in exercise-trained mice (P < 0.05; Fig. 4E) with significantly higher levels of oxidative stress (P < 0.05; Fig. 4F). There were significantly more adipocyte progenitors in CON + EX mice compared with all other groups (P < 0.05; Fig. 4G) with no differences in levels of oxidative stress in this cell population between groups (Fig. 4H). There were significantly more endothelial progenitors in exercise-trained mice and significantly fewer endothelial progenitors in obese mice (both P < 0.05; Fig. 4I). Endothelial progenitors from CON + EX had significantly higher levels of oxidative stress compared with all other groups (P < 0.05; Fig. 4J).

Exercise training and obesity differentially influence stromal cell population quantity and reactive oxygen species levels after sublethal irradiation. Quantification of nestin+ MSC (A), PDGFRα+ MSC (C), osteoblasts (E), adipocyte progenitors (G), and endothelial progenitors (I). Total bone marrow cell numbers were quantified in whole bone marrow isolates. Quantification of reactive oxygen species median fluorescence intensity in nestin+ MSC (B), PDGFRα+ MSC (D), osteoblasts (F), adipocyte progenitors (H), and endothelial progenitors (J). Data are presented as mean ± SEM (n = 9–10 per group). *P < 0.05 main effect of exercise, #P < 0.05 main effect of diet, and γ P < 0.05 vs all other groups.

Proteomic analysis of bone marrow environment

Supernatant samples collected from flushed bone marrow were characterized for relative cytokine content using a targeted proteomics approach. Pooled bone marrow supernatants from the same treatment group were analyzed using a cytokine array and expressed as log2‐fold change from CON + SED (Fig. 5A). Relative to CON + SED, CON + EX had 20 (12 up- and 8 downregulated), HF + SED had 22 (11 up- and 11 downregulated), and HF + EX had 20 (10 up- and 10 downregulated) differentially expressed cytokines (Fig. 5B). Comparing exercise-trained groups (i.e., CON + EX vs HF + EX), all 16 of the cytokines that were differentially expressed in each group were altered in the same direction (i.e., either up- or downregulated) (Fig. 5B). Comparing HF + SED and CON + EX, eight of the differentially expressed cytokines in each group were altered in the same direction, whereas six were altered in different directions (Fig. 5B). Comparing HF + SED and HF + EX, seven of the differentially expressed cytokines were altered in the same direction, whereas eight of the differentially expressed cytokines were altered in the opposite direction (Fig. 5B). Among cytokines differentially expressed within the three groups, seven were altered in the same direction, whereas five were altered in the opposite direction (Fig. 5B). Of these, five cytokines that were altered in the opposite direction, changes in CON + EX and HF + EX were the same, and differences between these two groups and HF + SED accounted for all the oppositely altered cytokines. Interestingly, both exercise-trained groups (i.e., CON + EX and HF + EX) had the same four cytokines that showed the greatest increase in expression: CXCL4, LIX, TNF-RI, and MIP-1γ (Fig. 5A). Two of these cytokines, LIX and MIP-1γ, were also in the top four of highest content in HF + SED (Fig. 5A). IL-5 and LTN were among the cytokines with the lowest relative content in all groups (Fig. 5A).

Obesity and exercise modulate soluble factors in the bone marrow after sublethal irradiation. Supernatant collected from flushed bone marrow and pooled for each group was analyzed using a targeted, semiquantitative proteomic array. Data were normalized to CON + SED and expressed as log2‐fold change. Fold changes ≥50% (i.e., log2‐fold change ≤−1 or ≥1 were considered biologically relevant). A, Heatmap with two-way hierarchical clustering comparing the unscaled log2‐fold changes from CON + SED of all detected cytokines. B, Venn diagram depicting the number of cytokines that were relatively higher (red) or relatively lower (green) within each group compared with CON + SED. The black numbers indicate the number of cytokines with relative differences from CON + SED in the same direction between overlapping circles. The gray numbers indicate the number of cytokines with relative differences from CON + SED in the opposite direction between overlapping circles. C, KG-1a leukemia blast viability at each time point expressed as a fold change relative to the baseline measures within each group. #Main effect of diet (n = 8 per group).
Antibodies used for HSPC and bone marrow stromal cell analysis.

Obese bone marrow environment accelerates leukemia blast expansion

Bone marrow supernatant from each group was applied to proliferating KG-1a leukemia blasts to functionally characterize the in vitro effects of exercise and/or obesity on leukemia blast viability. Leukemia blast viability, as measured by MTT assay, was significantly higher when cultured with bone marrow supernatant from obese mice at 24 h, with no effect of exercise (P < 0.05, main effect for diet; Fig. 5C). No significant differences between groups were detected after 48 h in culture.


Radiation exposure damages the hematopoietic system, leading to HSPC death, bone marrow inflammation, and increased risk of developing cancer (25). In this study, we observed that exercise training increased the quantity of bone marrow HSPC and reduced long-term oxidative stress in HSPC whereas obesity reduced several HSPC populations. In addition, exercise training remodels the HSPC niche to be more conducive to hematopoiesis via increased quantity of bone marrow endothelial progenitors and MSC, while reducing MAT content, which was also opposite to the effects of obesity. Interestingly, we observed increased levels of ROS within most stromal cell populations in exercise-trained mice, contrary to what was observed in HSPC. Our cytokine analysis revealed complete overlap in the direction of the changes (i.e., up- or downregulation) in bone marrow cytokines in CON + EX and HF + EX, whereas exercise training reversed several of the alterations in bone marrow cytokines induced by obesity (i.e., HF + EX vs HF + SED). Lastly, obesity promoted early leukemia blast expansion with no effect of exercise training. Taken together, these data suggest that exercise training and obesity differentially modulate the cellular and soluble constituents of the bone marrow after radiation exposure. These findings could have implications for exercise and dietary interventions in long-term cancer survivors who have undergone radiation therapy.

After these sublethal radiation doses, persistent oxidative stress delays full recovery of bone marrow cellularity until 28–56 d (26). We observed a higher quantity of bone marrow cells and HSPC in exercise-trained mice compared with sedentary mice, along with a reduction of reactive oxygen species in HSPC, suggesting a reduction in long-term hematopoietic stress. Conversely, obesity resulted in persistent bone marrow hypocellularity and decreased HSPC content in our study. These results are similar to what has been observed previously in lean and obese mice without previous radiation exposure (8,11) and in lean, exercise-trained mice exposed to radiation (10,15,27). In obese, sedentary mice exposed to a fractionated radiation dose, cell death was not exacerbated; however, bone marrow cell proliferation was reduced (28). Interestingly, acute exercise is known to induce HSPC proliferation (16). Thus, differential HSPC content in exercise-trained versus obese mice exposed to radiation could be due to increased HSPC protection and/or proliferation in exercise-trained mice with reduced protection and/or proliferation in obese.

Several factors in the HSPC niche regulate HSPC survival, maintenance, and proliferation. Negative correlations between MAT and bone marrow cellularity have been observed after radiation exposure in mice (29). Further, radiation-induced damage to endothelial cells and/or MSC reduces HSPC quantity (29). In the present study, exercise-trained mice had lower MAT content and higher numbers of endothelial progenitors, osteoblasts, and MSC. These effects were blunted by obesity. An interesting observation made from the current study is that exercise training and high-fat diet differentially affected two different phenotypically identified MSC populations. Nestin+ MSC were highest in exercise-trained mice on the control diet, but they were lowest in sedentary mice on the control diet. Conversely, PDGFRα+ MSC were significantly reduced in obesity. These data align with previous studies suggesting that nestin+ MSC and PDGFRα+ MSC are distinct cell populations (30). Nestin+ MSC are generally associated with bone marrow vessels; thus, higher nestin+ MSC content in exercise-trained mice may be related to the higher content of endothelial progenitors observed in the present study and higher content of HSPC in the vascular niche of exercise-trained mice (10). PDGFRα+ MSC highly overlap with leptin receptor+ MSC that are precursors for marrow adipocytes (30). Interestingly, we observed an increase in the quantity of preadipocytes and osteoblasts in exercise-trained mice consuming the control diet despite the reduction in MAT in exercise-trained mice. Unlike mature adipocytes, immature adipocyte progenitors and osteoblasts have been shown to promote hematopoiesis (23). We speculate that obesogenic conditions promote PDGFRα+ MSC that differentiate along the adipogenic lineage, resulting in higher MAT as previously observed (30). Conversely, exercise training may maintain preadipocytes in an undifferentiated state, which is associated with promoting hematopoiesis and reduced MAT.

Contrary to what was observed in HSPC, we observed an increase in reactive oxygen species in most stromal cell populations from exercise-trained mice. Oxidative stress in bone marrow stromal cells has been linked to markers of stromal cell dysfunction such as increased proinflammatory cytokines such as IL-1β, increased MAT, and increased oxidative stress in HSPC (6). We did not observe any of these markers to be increased in exercise-trained mice, suggesting that despite higher levels of reactive oxygen species, exercise-trained stromal cells were not experiencing oxidative stress. It is possible that the increased reactive oxygen species in exercise-trained mice could be the result of stromal cells from exercise-trained mice having directly experienced and survived the radiation exposure, whereas stromal cells in sedentary mice could be progeny of cells exposed to radiation. An alternative hypothesis may be that bone marrow stromal cells are reactive oxygen species scavengers. Indeed, bone marrow stromal cells have been shown to import reactive oxygen species from HSPC via gap junctions (31). This could explain the reduced reactive oxygen species in HSPC from exercise-trained mice.

Persistent alterations in paracrine signaling has been suggested to be a mechanism responsible for long-term hematopoietic defects after radiation exposure (6). Using a targeted, semiquantitative proteomics approach, we investigated the effects of exercise training and obesity on changes to bone marrow cytokine levels after radiation exposure. In agreement with our MAT data, obese, sedentary mice had higher levels of leptin, which was reduced by exercise training. Leptin is secreted from adipocytes, and elevated leptin signaling has been implicated in biasing MSC into adipogenic lineages (30). Interestingly, from all cytokines altered in both exercise groups (i.e., CON + EX and HF + EX), a complete overlap between the direction of change (i.e., up- or downregulated) was observed, suggesting that the effects of exercise on the bone marrow cytokine environment are highly conserved despite obesogenic conditions. Conversely, obese, exercise-trained mice had several cytokines whose levels were altered in the opposite direction of obese, sedentary mice, suggesting that exercise training could specifically reverse some of the cytokine alterations induced by obesogenic conditions. Specifically, higher relative levels of M-CSF and MIP-2 in exercise-trained mice were relatively decreased in obese, sedentary animals. M-CSF and MIP-2 are potent activators of hematopoiesis (32), and unlike GM-CSF or G-CSF, M-CSF promotes myelopoiesis without long-term HSPC exhaustion (32). As such, exercise training increased the lower levels of these prohematopoietic cytokines seen in obesity. Conversely, exercise-trained mice had lower relative levels of MIP-1α, MIP-3α, and TNFα, all of which were relatively increased in obese, sedentary mice. MIP-1α inhibits expansion of healthy HSPC, whereas leukemia blasts are resistant to this inhibition (33). Further, MIP-3α, a proinflammatory cytokine, is increased in the bone marrow of patients with multiple myeloma, and multiple myeloma cells increase MIP-3α from osteoblasts in a mechanism that involves TNFα and IL-1β (34), which was also increase in obese, sedentary conditions. Similarly, TNFα is a marker of stromal damage and has been shown to inhibit hematopoiesis (35). Exercise training reversed the higher levels of inhibitory cytokines seen in obesity. Three cytokines, SDF-1, TNF-R1, and MCP-5, were altered in exercise-trained mice but not obese mice. SDF-1 and TNF-R1 were both higher in exercise-trained mice. Although the role of non-membrane-bound TNF-R1 remains to be fully elucidated, TNF-R1 binds TNFα in circulation and could decreases its biological activity (36). This would further enhance the effects of lower TNFα levels in exercise-trained mice. CXCL12/SDF-1 is necessary for HSPC maintenance and retention in the niche (37). A recent report demonstrated that CXCL12 also plays an important role in reducing reactive oxygen species in HSPC and promoting their long-term health and survival after exposure to an oxidative insult (38). As such, higher CXCL12 levels in exercise-trained mice could explain, in part, the reduced oxidative stress observed in HSPC. Lastly, MCP-5 was lower in exercise-trained mice and unchanged in obese, sedentary mice. MCP-5 promotes local inflammation by acting as a chemoattractant for CCR2-expressing inflammatory cells (39), suggesting reduced inflammation in bone marrow from exercise-trained mice. Obese, sedentary mice had higher levels of IL-1β, which contributes to both persistent inflammation and disruption of steady-state hematopoiesis (6). Together, our cytokine data suggest that exercise training is associated with a prohematopoietic, anti-inflammatory environment within the bone marrow after radiation exposure, whereas obesity is associated with higher levels of factors that inhibit hematopoiesis.

Given the increase in hematopoietic cells and the prohematopoietic environment in exercise-trained mice, we were interested in investigating if exercise training had similar growth-promoting effects on leukemia cells. Because radiation-induced leukemia develops in 18 months in CBA mice exposed to 3‐Gy radiation (12), we used an in vitro model where leukemia blasts were treated with bone marrow supernatant from each respective group. There were no effects of exercise training on leukemia blast viability, but obesity significantly increased leukemia blast viability after 24 h in culture. The mechanisms for this effect remain unclear; however, our proteomic array does reveal some clues. IGFBP-3 was relatively higher in CON + EX, HF + SED, and HF + EX; however, the relative expression in CON + EX was approximately twice as high as HF + SED and HF + EX. A recent study revealed that reduced levels of IGFBP-3 in vivo led to increased IGF-1-mediated leukemia blast proliferation (40). IGF-1 was not included in our cytokine array; however, the relatively higher expression of IGFBP-3 in CON + EX could explain the reduced proliferation of leukemia blasts treated with bone marrow supernatant from this group. IL-4, which was relatively higher in CON + EX and unchanged in HF + SED and HF + EX, has also been shown to specifically induce leukemia blast apoptosis without affecting healthy HSPC growth (41). These data suggest that the prohematopoietic environment in exercise-trained mice did not enhance leukemia blast viability in vitro; however, the bone marrow environment in obesity creates proleukemic conditions, perhaps mediated by relatively lower expression of IGFBP-3 and IL-4.

A limitation of the present study is the lack of nonirradiated control group. The effects of radiation on HSPC and the bone marrow environment have been extensively examined (29,42). The focus of the present study was to examine the effects of host factors, such as obesity and exercise, on the late effects of radiation. Including a nonirradiated control group would have increased the complexity of the question and detracted from the primary aim of the study. As such, investigating the effects of radiation independent of obesity and exercise were beyond the scope of the present study.

Overall, the results from this study demonstrate that exercise training and high-fat diet–induced obesity differentially influence HSPC content and the bone marrow environment after radiation exposure. Specifically, our results suggest that exercise training creates an environment more conducive to hematopoiesis, likely by increased protection of HSPC and stromal cells without promoting leukemia blast viability. From a translational perspective, these data suggest that cancer survivors with obesity may be at greater risk of hematopoietic complications after radiation therapy. Further, a moderate-intensity, aerobic, exercise training intervention should be considered to mitigate the late effects of radiation therapy on the hematopoietic system in cancer survivors, regardless of body mass index. Future clinical studies should further investigate if exercise training interventions can be a viable therapeutic option to maintain healthy hematopoiesis and bone marrow function in cancer survivors with or without obesity who have undergone radiation therapy.

Funding was graciously provided by the American Institute for Cancer Research (M. D.), the American College of Sports Medicine (M. D. and R. E.), and the National Sciences and Engineering Research Council (NSERC) of Canada (M. D.). The authors thank Dr. Leilei Yin of the Imaging Technology Group at the Beckman Institute for instruction and use of the Xardia microCT. M. D. and H. C. conceived the study design and direction. R. E., M. N., G. X., and D. H.-S. performed the experiments and analyzed the data. R. E. and M. D. wrote the manuscript. R. E., M. N., G. X., D. H.-S., H. C., and M. D. edited the manuscript and approved final submission.

The authors have no financial conflicts of interest to disclose. The results of the present study do not constitute endorsement by the American College of Sports Medicine. The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.


1. American Cancer Society. Cancer Facts & Figures. 2018; Available from
2. Zhang FF, Saltzman E, Kelly MJ, et al. Comparison of childhood cancer survivors’ nutritional intake with US dietary guidelines. Pediatr Blood Cancer. 2015;62(8):1461–7.
3. Wang Y, Schulte BA, LaRue AC, Ogawa M, Zhou D. Total body irradiation selectively induces murine hematopoietic stem cell senescence. Blood. 2006;107(1):358–66.
4. Birbrair A, Frenette PS. Niche heterogeneity in the bone marrow. Ann N Y Acad Sci. 2016;1370(1):82–96.
5. Singh S, Kloss FR, Brunauer R, et al. Mesenchymal stem cells show radioresistance in vivo. J Cell Mol Med. 2012;16(4):877–87.
6. Lorimore SA, Mukherjee D, Robinson JI, Chrystal JA, Wright EG. Long-lived inflammatory signaling in irradiated bone marrow is genome dependent. Cancer Res. 2011;71(20):6485–91.
7. Emmons R, Niemiro GM, De Lisio M. Hematopoiesis with obesity and exercise: role of the bone marrow niche. Exerc Immunol Rev. 2017;23:82–95.
8. van den Berg SM, Seijkens TT, Kusters PJ, et al. Diet-induced obesity in mice diminishes hematopoietic stem and progenitor cells in the bone marrow. FASEB J. 2016;30(5):1779–88.
9. Baker JM, De Lisio M, Parise G. Endurance exercise training promotes medullary hematopoiesis. FASEB J. 2011;25(12):4348–57.
10. De Lisio M, Parise G. Characterization of the effects of exercise training on hematopoietic stem cell quantity and function. J Appl Physiol (1985). 2012;113(10):1576–84.
11. Styner M, Thompson WR, Galior K, et al. Bone marrow fat accumulation accelerated by high fat diet is suppressed by exercise. Bone. 2014;64:39–46.
12. Rivina L, Schiestl R. Mouse models of radiation-induced cancers. Adv Genet. 2013;84:83–122.
13. Kelly SA, Nehrenberg DL, Hua K, Garland T Jr, Pomp D. Exercise, weight loss, and changes in body composition in mice: phenotypic relationships and genetic architecture. Physiol Genomics. 2011;43(4):199–212.
14. Emmons R, Xu G, Hernandez-Saavedra D, et al. Effects of obesity and exercise on colon cancer induction and hematopoiesis in mice. Am J Physiol Endocrinol Metab. 2019;316(2):E210–20.
15. De Lisio M, Baker JM, Parise G. Exercise promotes bone marrow cell survival and recipient reconstitution post-bone marrow transplantation, which is associated with increased survival. Exp Hematol. 2013;41(2):143–54.
16. Emmons R, Niemiro G, Owolabi O, De Lisio M. Acute exercise mobilizes hematopoietic stem and progenitor cells and alters the mesenchymal stromal cell secretome. J Appl Physiol (1985). 2016;120(6):624–32.
17. Scheller EL, Troiano N, Vanhoutan JN, et al. Use of osmium tetroxide staining with microcomputerized tomography to visualize and quantify bone marrow adipose tissue in vivo. Methods Enzymol. 2014;537:123–39.
18. Saland E, Boutzen H, Castellano R, et al. A robust and rapid xenograft model to assess efficacy of chemotherapeutic agents for human acute myeloid leukemia. Blood Cancer J. 2015;5(3):e297.
19. De Lisio M, Jensen T, Sukiennik RA, Huntsman HD, Boppart MD. Substrate and strain alter the muscle-derived mesenchymal stem cell secretome to promote myogenesis. Stem Cell Res Ther. 2014;5(3):74.
20. Niemiro GM, Raine LB, Khan NA, et al. Circulating progenitor cells are positively associated with cognitive function among overweight/obese children. Brain Behav Immun. 2016;57:47–52.
21. Méndez-Ferrer S, Michurina TV, Ferraro F, et al. Mesenchymal and haematopoietic stem cells form a unique bone marrow niche. Nature. 2010;466(7308):829–34.
22. Houlihan DD, Mabuchi Y, Morikawa S, et al. Isolation of mouse mesenchymal stem cells on the basis of expression of Sca-1 and PDGFR-α. Nat Protoc. 2012;7(12):2103–11.
23. Ambrosi TH, Scialdone A, Graja A, et al. Adipocyte accumulation in the bone marrow during obesity and aging impairs stem cell-based hematopoietic and bone regeneration. Cell Stem Cell. 2017;20(6):771–784.e6.
24. Winkler IG, Barbier V, Wadley R, Zannettino AC, Williams S, Lévesque JP. Positioning of bone marrow hematopoietic and stromal cells relative to blood flow in vivo: serially reconstituting hematopoietic stem cells reside in distinct nonperfused niches. Blood. 2010;116(3):375–85.
25. Williams JP, Calvi L, Chakkalakal JV, Finkelstein JN, O’Banion MK, Puzas E. Addressing the symptoms or fixing the problem? Developing countermeasures against normal tissue radiation injury. Radiat Res. 2016;186(1):1–16.
26. Simonnet AJ, Nehmé J, Vaigot P, Barroca V, Leboulch P, Tronik-Le Roux D. Phenotypic and functional changes induced in hematopoietic stem/progenitor cells after gamma-ray radiation exposure. Stem Cells. 2009;27(6):1400–9.
27. De Lisio M, Phan N, Boreham DR, Parise G. Exercise-induced protection of bone marrow cells following exposure to radiation. Appl Physiol Nutr Metab. 2011;36(1):80–7.
28. Vares G, Wang B, Ishii-Ohba H, Nenoi M, Nakajima T. Diet-induced obesity modulates epigenetic responses to ionizing radiation in mice. PloS One. 2014;9(8):e106277.
29. Green DE, Rubin CT. Consequences of irradiation on bone and marrow phenotypes, and its relation to disruption of hematopoietic precursors. Bone. 2014;63:87–94.
30. Zhou BO, Yue R, Murphy MM, Peyer JG, Morrison SJ. Leptin-receptor-expressing mesenchymal stromal cells represent the main source of bone formed by adult bone marrow. Cell Stem Cell. 2014;15(2):154–68.
31. Taniguchi Ishikawa E, Gonzalez-Nieto D, Ghiaur G, et al. Connexin-43 prevents hematopoietic stem cell senescence through transfer of reactive oxygen species to bone marrow stromal cells. Proc Natl Acad Sci U S A. 2012;109(23):9071–6.
32. Kandalla PK, Sarrazin S, Molawi K, et al. M-CSF improves protection against bacterial and fungal infections after hematopoietic stem/progenitor cell transplantation. J Exp Med. 2016; jem.20151975.
33. Baba T, Mukaida N. Role of macrophage inflammatory protein (MIP)-1α/CCL3 in leukemogenesis. Mol Cell Oncol. 2014;1(1):e29899.
34. Giuliani N, Lisignoli G, Colla S, et al. CC-chemokine ligand 20/macrophage inflammatory protein-3α and CC-chemokine receptor 6 are overexpressed in myeloma microenvironment related to osteolytic bone lesions. Cancer Res. 2008;68(16):6840–50.
35. Pronk CJ, Veiby OP, Bryder D, Jacobsen SE. Tumor necrosis factor restricts hematopoietic stem cell activity in mice: involvement of two distinct receptors. J Exp Med. 2011;208(8):1563–70.
36. Speeckaert MM, Speeckaert R, Laute M, Vanholder R, Delanghe JR. Tumor necrosis factor receptors: biology and therapeutic potential in kidney diseases. Am J Nephrol. 2012;36(3):261–70.
37. Sugiyama T, Kohara H, Noda M, Nagasawa T. Maintenance of the hematopoietic stem cell pool by CXCL12-CXCR4 chemokine signaling in bone marrow stromal cell niches. Immunity. 2006;25(6):977–88.
38. Zhang Y, Dépond M, He L, et al. CXCR4/CXCL12 axis counteracts hematopoietic stem cell exhaustion through selective protection against oxidative stress. Sci Rep. 2016;6:37827.
39. Charo IF, Ransohoff RM. The many roles of chemokines and chemokine receptors in inflammation. N Engl J Med. 2006;354(6):610–21.
40. Zhang J-M, Wang C-C, Zhang G-C, et al. ADAM28 promotes tumor growth and dissemination of acute myeloid leukemia through IGFBP-3 degradation and IGF-I-induced cell proliferation. Cancer Lett. 2018;442:193–201.
41. Peña-Martínez P, Eriksson M, Ramakrishnan R, et al. Interleukin 4 induces apoptosis of acute myeloid leukemia cells in a Stat6-dependent manner. Leukemia. 2018;32(3):588–96.
42. Shao L, Wang Y, Chang J, Luo Y, Meng A, Zhou D. Hematopoietic stem cell senescence and cancer therapy-induced long-term bone marrow injury. Transl Cancer Res. 2013;2(5):397–411.


Copyright © 2019 by the American College of Sports Medicine