Share this article on:

Immunological Changes in Response to Exercise: Influence of Age, Puberty, and Gender


Medicine & Science in Sports & Exercise: February 2006 - Volume 38 - Issue 2 - pp 293-304
doi: 10.1249/01.mss.0000183479.90501.a0
BASIC SCIENCES: Original Investigations

Purpose: This study tested the hypothesis that exercise-induced perturbation and recovery of the immune system would vary with age, puberty, and gender in healthy children and adolescents.

Methods: Twelve-year-old girls (YG; N = 14) and boys (YB; N = 20), and 14-yr-old girls (OG; N = 11) and boys (OB; N = 13) cycled for 60 min at 70% V̇O2max. Blood was collected before, at 30 and 60 min of exercise, and at 30 and 60 min of recovery to measure total leukocytes, leukocyte and lymphocyte subsets, and cytokines. Age and pubertal (Tanner stage) effects within genders and gender effects within age and pubertal groups were determined.

Results: Exercise-induced increases in lymphocytes, CD3CD16+CD56+ counts, and IL-6 were approximately 83, 90, and 390% greater in OG versus YG (P < 0.05). Recovery leukocytosis and neutrophilia were approximately 56 and 35% greater in OB versus YB (P < 0.05). Pubertal stage did not have a statistically significant influence on responses in girls, but the lowest pubertal stage consistently showed smaller changes in lymphocytes and CD3CD16+CD56+ counts. Recovery neutrophilia was approximately 120% greater in postpubertal boys versus prepubertal or pubertal boys (P < 0.05). Responses of lymphocytes and CD3CD16+CD56+ counts, respectively, were approximately 120 and 82% greater in OG versus OB (P < 0.05), with no differences between YG and YB. Exercise-induced increases in total leukocytes, lymphocytes, and CD3CD16+CD56+ counts were at least 35% greater in girls versus boys of similar pubertal status (P < 0.05). Regardless of age, puberty, or gender, IL-8 levels were significantly higher during recovery versus rest (P < 0.05).

Conclusion: These results highlight the need to control for age, puberty, and gender when interpreting immunologic responses to exercise in a pediatric population.

1Children's Exercise and Nutrition Centre, 2Departments of Pediatrics and Medicine, and 3Department of Pathology and Molecular Medicine, McMaster University, Hamilton, Ontario, CANADA

Address for correspondence: Mark A. Tarnopolsky, M.D., Ph.D., FRCP(C), McMaster University Medical Centre, 1200 Main Street West, Hamilton, Ontario, Canada, L8N 3Z5; E-mail:

Submitted for publication April 2005.

Accepted for publication August 2005.

Physical inactivity among North American children and adolescents is cause for concern, with a persistent increase in the prevalence of childhood obesity (32,33) and an increased occurrence of adultlike health conditions such as type 2 diabetes mellitus (15). In this regard, the immune system is intimately linked to obesity and its associated complications (7), but to what extent the early onset of these adultlike disorders during childhood is linked to immune development is unknown. Indeed, the immune system undergoes considerable development in both cell number and function throughout childhood (3) and adolescence (4). Moreover, physical activity is a potent stimulus to the immune system in both adults (23) and children (5,10,24-26,30,31,34). That regular exercise during childhood can influence normal growth and development of the immune system, and, thus, the cause of disease, is of particular interest. However, our understanding of even basic immunological changes in response to acute exercise during childhood and adolescence and of how these responses may vary with age, puberty, and gender is deficient.

In adults, considerable evidence suggests that strenuous, prolonged exercise often causes a postexercise suppression of immune function, but that exercise of moderate intensity and duration tends to enhance several aspects of immunity (23). Similar exercise-induced perturbations to the immune system of children and adolescents however, can carry health implications unique to the growing individual. We have recently reported that under identical exercise conditions, changes in some cellular and soluble components of the immune system were smaller in pre- and early-pubertal boys versus adult men (30). Further recovery of immune perturbations following strenuous exercise was faster in the children than in the adults (30). Although our study was the first to compare children and adults under identical well-controlled experimental conditions, other investigators have reported immune changes following various forms of aerobic-type exercise in young boys and girls (5,10,24-26,31,34) and older adolescent boys and girls (5,16-18,20). No study, however, has systematically and simultaneously examined the influence of age, puberty, and gender on immunologic responses to exercise in a healthy pediatric population.

An improved understanding of how exercise impacts a child's immune system is of clinical importance. In children recovering from cancer, and in particular leukemia, moderate levels of exercise may improve or at least offset decrements in immune health caused by various treatment strategies. Consequently, the safe prescription of exercise for these children requires the distinction between normal and abnormal exercise-induced responses so that further impairment of their immune system is avoided. We, therefore, investigated the effects of age, puberty, and gender on changes in various cellular and soluble components of the immune system in response to standardized high-intensity aerobic exercise to better understand the "healthy" response. To minimize possible effects of previous antigenic experience on immune measures and confounding effects of age-associated factors on pubertal comparisons, children at two distinct chronologic ages were tested. Based on our previous observations (30), we hypothesized that perturbations to, and recovery of, the immune system would be smaller and faster, respectively, in younger versus older children and in less mature versus more mature individuals. Given the reported gender differences in components of the immune system among children (3) and adolescents (4), we further hypothesized that gender differences would exist in the immunologic changes induced by exercise.

Back to Top | Article Outline


Back to Top | Article Outline

Study design and subjects.

A total of 58 subjects volunteered to participate in this study, which was approved by the McMaster University research ethics review board. We recruited healthy boys (N = 33) and girls (N = 25) 12 and 14 yr of age who were at various stages of puberty. All subjects performed a preliminary session and an experimental session. By design, we made separate age and pubertal comparisons within each gender and gender comparisons within age and pubertal groups. Pubertal status of each subject was determined by self-assessment of pubic hair development (boys) or breast development (girls) according to Tanner (28). Self-assessment of pubertal status according to development of pubic hair in boys and breasts in girls has been shown to be valid and reproducible (14). Table 1 provides subject characteristics according to age and pubertal status. All 14-yr-old girls maintained regular menstrual cycles and were not taking oral contraceptive therapy (OCT), whereas three 12-yr-old girls reported having had their first menses, but had not yet developed regular cycles. The 14-yr-old girls were tested in the midfollicular phase of their menstrual cycle. No attempt was made to test the menarcheal 12-yr-old girls at a particular time because of the sporadic nature of their cycles. All subjects were healthy with no recent allergies or illness and none were taking medication. After the purpose, procedures, and risks of the study were explained, the children agreed verbally to participate and a parent then signed a written informed consent.

Back to Top | Article Outline

Preliminary session.

A preliminary visit was conducted to measure body height (SECA 216 Accu-Hite Stadiometer, Creative Health Products, Plymouth, MI), body mass (BWB-800, Tanita, Tokyo, Japan), and percent body fat (bioelectric impedance-101A, RJL Systems, Clinton Twp., MI) and to determine Tanner stage. Maximal O2 uptake (V̇O2max) was determined on a cycle ergometer (Ergomedic 818E, Monark, Sweden) using a progressive, continuous exercise test. Subjects began cycling at either 30 or 60 W, depending on age and estimated fitness level, with pedaling rate constant at 60 rpm. Work rate was increased by 30 W every 2 min for all subjects. A test was determined maximal when pedaling rate dropped below 50 rpm for 3 s, despite strong encouragement and the respiratory exchange ratio was >1.1. Other maximal criteria related to heart rate (HR) and V̇O2 were not used because equations of age-predicted maximal HR are not accurate during childhood and a plateau in V̇O2, despite increasing exercise intensity, is not consistently observed in children. During the test, HR was continuously monitored with a Polar HR monitor (Polar A1, Polar Electro, Kempele, Finland) and subjects breathed through a Hans Rudolph valve with an appropriately sized mouthpiece. Expired air was collected continuously and analyzed for O2 (Beckman O2 analyzer OM-11, Beckman Inc., CA) and carbon dioxide (HP47210A capnometer, Hewlett Packard, CA) with analyzers connected to a Vista PC interface with Turbofit software (VacuMed, Ventura, CA) on a personal computer. The highest 30-s V̇O2 was taken as the V̇O2max.

Back to Top | Article Outline

Experimental session.

Subjects were instructed to maintain their habitual diets, but to avoid "fast-food"type of meals, and to avoid excessive physical activity for the 2 d immediately before their experimental trial. All subjects complied with these instructions. To minimize circadian effects on immune measures, every experimental session was conducted in the morning with subjects arriving to the laboratory at either 0730 h or 0830 h in at least a 10-h fasted state. On arrival, subjects voided their bladder, were weighed in the nude using an electronic scale (Tanita), and then rested supine for approximately 10 min, after which time an indwelling venous catheter (Becton Dickinson, NJ) was placed in either an arm or a hand. After a further 10 min of supine rest, a resting, preexercise blood sample was drawn. Subjects then consumed a small, standardized breakfast, which served to standardize preexercise nutrition. Throughout the session, subjects were given flavored water to drink at a rate to maintain body hydration (30). At 40 min after the resting blood sample, subjects began cycling (Monark) at a power output equivalent to 70% of their predetermined V̇O2max, with the target intensity achieved in the first 5 min by analysis of expired gas (Beckman and Hewlett Packard). Exercise consisted of two 30-min bouts separated by a 5- to 7-min rest period. Additional expired gas samples were collected at steady state from minutes 11 to 15 and 26 to 30 of each exercise bout to ensure the proper work intensity, with the power output adjusted accordingly. Additional blood samples were collected after 30 and 60 min of exercise and at 30 and 60 min of recovery. Blood samples were drawn while subjects remained seated on the cycle ergometer or quietly in the laboratory. The catheter was kept patent by flushing with approximately 1.5 mL of sterile saline (0.9% NaCl) after each blood sampling. Consequently, the first 2 mL of blood at each sampling time was discarded. During the recovery period, subjects sat quietly and were allowed to empty their bladder if necessary.

Back to Top | Article Outline

Total leukocytes and leukocyte subsets.

Whole blood treated with EDTA was analyzed for total leukocytes, neutrophils, lymphocytes, and monocytes using an automated Coulter counter at the McMaster University Medical Centre Core Laboratory. Hemoglobin (Hb) and hematocrit (Hct) were also assessed in these samples to calculate changes in blood and plasma volume according to Dill and Costill (9) and all immune cell concentrations were corrected for exercise-induced changes in blood volume.

Back to Top | Article Outline

Lymphocyte subsets.

The EDTA-treated whole blood was used to determine lymphocyte subsets by direct immunofluorescence and flow cytometry. The following mouse antihuman monoclonal antibodies (Mab) and fluorochrome conjugates were used: CD3-PerCP, CD3-FITC, CD4-PerCP, CD8-PerCP, CD16-FITC, CD19-FITC, and CD56-PE. Lymphocyte subsets were classified as total T cells (CD3+), Thelper cells (CD3+CD4+), Tcytotoxic cells (CD3+CD8bright), B cells (CD3CD19+), and natural killer (NK) cells (CD3CD16+CD56+). All reagents were purchased from BD Biosciences and samples were stained as per the manufacturer's instructions within 6 h of collection. Well-mixed whole blood (100 μL) was added to 12 × 75-mm Falcon tubes containing an appropriate cocktail of Mab (10 μL each). Samples were vortexed and incubated for 20 min at room temperature (RT) in the dark. After adding 2 mL of BD Pharm Lyse solution to lyse red blood cells, samples were vortexed and incubated a further 10 min at RT. Samples were centrifuged (300 × g for 5 min at RT), washed with 2 mL of BD Pharmingen stain buffer, centrifuged (200 × g for 5 min at RT) and fixed with 0.5 mL of BD Cytofix buffer. Samples were stored at 3°C for no more than 48 h before run on a FACScan flow cytometer (Becton Dickinson, Mississauga, Canada) with CELLQuest software. The lymphocyte population was gated using forward-scatter versus side-scatter characteristics and 10,000 events per lymphocyte gate were collected. Analyses of lymphocyte subsets were performed "offline" with WinMDI 2.8 software (Joseph Trotter, The Scripps Research Institute, CA). Cell counts of each lymphocyte subset were calculated by multiplying the percentage of cells with appropriate fluorescence by the absolute number of lymphocytes. Because of logistical restraints, CD3+CD4+ and CD3+CD8bright cells were not determined in blood samples drawn at 30 min of exercise or at 30 min of recovery, whereas CD3CD19+ cells were not determined at 30 min of recovery.

Back to Top | Article Outline


Whole blood treated with EDTA was centrifuged at 2000 × g for 10 min, and the plasma was stored at −50°C until analyzed. ELISA kits (R&D Systems, Minneapolis, MN) were used to determine plasma concentrations of IL-6 (Cat. No. HS600B), IL-8 (Cat. No. D8000C), and TNF-α (Cat. No. HSTA00C) in duplicate. The sensitivities of these kits, as reported by the manufacturer, are 0.039 pg·mL−1 for IL-6, 3.5 pg·mL−1 for IL-8, and 0.12 pg·mL−1 for tumor necrosis factor (TNF)-α. In our hands, the intra- and interassay CV, respectively, are ≤5 and 9% for IL-6, ≤4 and 10% for IL-8, and ≤4 and 12% for TNF-α. Cytokines were determined in blood sample collected at rest, after 60 min of exercise and after 60 min of recovery. All postexercise cytokine concentrations were adjusted for changes in plasma volume (as above).

Back to Top | Article Outline

Statistical analyses.

Data are presented as means ± SEM, unless stated otherwise. To determine group differences in physical and fitness characteristics one-way ANOVA were used. Separate two-way ANOVA with one between factor (group) and one within factor (time) were used to analyze immune cell proportions and counts and cytokine concentrations for age, puberty, and gender comparisons. Where appropriate, a Tukey's post hoc test was used to determine significance among means. STATISTICA 5.0 (StatSoft, Tulsa, OK) was used for ANOVA. Pearson correlations (GraphPad Prism 4.03, GraphPad Software, San Diego, CA) were used to determine a possible influence of body size, in particular FFM, on inflammatory-related responses. In all cases, the threshold for statistical significance was set at P ≤ 0.05.

Back to Top | Article Outline


All subjects completed the exercise testing with no differences between groups (main effect age or puberty, P ≥ 0.90) in the relative exercise intensity, which averaged 68.0 ± 0.7% of V̇O2max or 29.2 ± 0.4 mL·kg−1 body mass·min−1.

Back to Top | Article Outline

Age effects.

Table 2 provides counts of total leukocytes and leukocyte subsets in 12-yr-old girls (YG) and boys (YB) and 14-yr-old girls (OG) and boys (OB) for age and pubertal comparisons. Total leukocytes and lymphocytes (interaction effect, P = 0.08 and P < 0.001, respectively), but not neutrophils or monocytes (interaction effects, P = 0.18 and P = 0.29, respectively), were greater or tended to be greater in OG versus YG at 30 and 60 min of exercise. In both OG and YG, total leukocytes remained elevated during the recovery period, but lymphocytes returned to resting levels by 30 min. Neutrophils were consistently higher in OG versus YG throughout the session (main effect age, P = 0.03). Total leukocyte and leukocyte subsets during exercise were not significantly different between YB and OB, but at 60 min of recovery total leukocytes (P = 0.01) and neutrophils (P = 0.05) were higher in OB versus YB.

Table 3 provides proportions and counts of lymphocyte subsets in YG, YB, OG, and OB for age and pubertal comparisons. Exercise caused reductions in the proportion of total CD3+ and CD3+CD4+ cells that were greater or tended to be greater in OG versus YG (interaction effect, P = 0.08 and P = 0.02, respectively). Cell proportions in OG and YG were not significantly different from resting levels at 60 min of recovery. In contrast, the proportion of CD3+CD8bright cells did not change with exercise OG or YG, but tended to be higher in OG versus YG at 60 min of recovery (interaction effect, P = 0.06). No age differences were seen within boys for responses of these cell proportions during exercise or into recovery (interaction effects, P ≥ 0.27).

The CD3+CD4+ and CD3+CD8bright cell counts increased with exercise (main effect time, P < 0.001 for both) similarly between YG and OG and between YB and OB. Values returned to resting levels by 60 min of recovery in all age groups. In contrast, the increase in total CD3+ cell counts was greater in OG versus YG at 30, but not at 60 min of exercise (interaction effect, P = 0.003), and recovery of CD3+ cell counts was complete by 30 min in OG and YG. Changes in total CD3+ cell counts during and following exercise were not significantly different between YB and OB (interaction effect, P = 0.98); however, total CD3+ cell counts in boys remained below resting levels at 30 and 60 min of recovery.

Although the proportion of CD3CD19+ cells decreased with exercise (main effect time, P < 0.001), no age differences were seen within girls or boys (interaction effects, P ≥ 0.69). By 60 min of recovery, CD3CD19+ proportions were higher than at rest in YG and YB (P < 0.001), but were not significantly different from resting levels in OG and OB (P = 0.08). In contrast to cell proportions, CD3CD19+ cell counts did not increase during exercise (P ≥ 0.10), but values at 60 min of recovery were higher than at rest in YG and YB (P < 0.02), but not in OG and OB (P = 0.93).

The proportion of CD3CD16+CD56+ cells increased with exercise (main effect time, P < 0.001), but no age differences were seen within girls or boys (interaction effects, P ≥ 0.25). The exercise-induced increase in CD3CD16+CD56+ cell counts was greater in OG versus YG (interaction effect, P = 0.002), but not in YB versus OB (interaction effect, P = 0.74). In all age groups, CD3CD16+CD56+ cell counts returned to resting levels by 30 min of recovery.

Table 4 provides concentrations of IL-6, TNF-α, and IL-8 in YG, YB, OG, and OB for age and pubertal comparisons. Exercise did not influence TNF-α levels in YG or OG (main effect time, P = 0.19), whereas concentrations gradually increased over time in OB, but remained stable in YB (interaction effect, P = 0.07). Within girls, IL-6 increased significantly in OG during and following exercise, but did not change in YG (interaction effect, P = 0.002). In both age groups of boys, IL-6 increased to the same extent during exercise and recovery (interaction effect, P = 0.38). Regardless of age, IL-8 remained stable immediately after exercise (P ≥ 0.37), but was significantly higher at 60 min of recovery versus rest (main effect time, P < 0.001).

Back to Top | Article Outline

Puberty effects in girls.

Because only four girls were classified at Tanner stage 2 (T2G), statistical comparisons were only made between girls at Tanner stage 3 (T3G) and girls at Tanner stages 4 and 5 (T4-5G). In general, no differences were seen between T3G and T4-5G for exercise-induced changes in, or recovery of, immune variables, with few exceptions. Exercise decreased the proportion of CD3+CD8bright cells in T4-5G, whereas it remained stable in T3G (interaction effect, P = 0.03), but no difference was noted between these pubertal groups at 60 min of recovery.

The exercise-induced increase in IL-8 in T3G tended to be greater than in T4-5G (interaction effect, P = 0.08). At 60 min of recovery, however, both pubertal groups demonstrated higher IL-8 levels compared with at rest (main effect time, P < 0.001).

Although we did not include T2G data in the statistical comparisons, it is noteworthy that the magnitude of change in both proportion and number of total CD3+ and CD3CD16+CD56+ cells and in total lymphocytes in T2G was consistently less than in T3G and T4-5G (see Figs. 3 and 4).

Back to Top | Article Outline

Puberty effects in boys.

Exercise-induced increases in total leukocytes and neutrophils were similar across pubertal stages in boys, but the recovery leukocytosis and neutrophilia were consistently greater in boys at Tanner stage 5 (T5B) compared with boys at Tanner stage 1 (T1B), Tanner stage 2 (T2B) and Tanner stages 3 and 4 (T3-4B) (interaction effect, P = 0.03 and P = 0.058 for leukocytosis and neutrophilia, respectively) (Fig. 1). No other immune cell changes were significantly different across Tanner stages in boys.

Similar to age effects on TNF-α in boys, TNF-α levels decreased slightly over time in T1B, but gradually increased in T5B (interaction effect, P = 0.03) (Fig. 2).

Back to Top | Article Outline

Gender effects in age groups.

No gender differences were seen between YG and YB for changes in any immune cell (interaction effects, P ≥ 21). In contrast, exercise-induced increases in total leukocytes and lymphocytes (interaction effect, P = 0.053 and P < 0.01, respectively), but not neutrophils or monocytes (interaction effect, P = 0.71 and P = 0.30, respectively), were greater in OG versus OB.

Gender differences between YG and YB and between OG and OB were found for the proportion of total CD3+ cells (interaction effect, P = 0.03 for both comparisons). The decrease in CD3+ proportion at 30 and 60 min of exercise was greater in YB versus YG, with full recovery in YB by 30 min. The decrease in CD3+ proportion was similar between OG and OB at 30 and 60 min of exercise, but OG had a greater rebound at 30 min of recovery compared with OB. In contrast, only a gender difference between YG and YB remained for the responses of total CD3+ cell counts (interaction effect, P = 0.02), because of lower counts at 30 and 60 min of recovery compared with resting values in YB, but not in YG.

No gender differences were found in the responses of the proportion or number of CD3CD19+, CD3+CD4+, or CD3+CD8bright cells between YG and YB or between OG and OB (interaction effects, P ≥ 0.27).

Although no gender differences were noted in the responses of CD3CD16+CD56+ cell proportions to exercise within age groups, when CD3CD16+CD56+ cell counts were considered, gender differences between OG and OB were found at 30 and 60 min of exercise (interaction effect, P = 0.03), but not between YG and YB (interaction effect, P = 0.71).

The only cytokine-related gender difference within age groups was for IL-6, which increased in YB, but remained stable in YG (interaction effect, P = 0.002). In contrast, the exercise-induced increase in IL-6 (main effect time, P < 0.001) was similar between OG and OB (interaction effect, P = 0.35).

Back to Top | Article Outline

Gender effects in pubertal groups.

Gender differences in total leukocyte responses were found between T3G and T3-4B (interaction effect, P = 0.04) and between T4-5G and T3-4B (interaction effect, P = 0.04) at 30 and 60 min of exercise, but not during recovery (Fig. 3). In contrast, gender differences between T4-5G and T5B (interaction effect, P = 0.03) were apparent only at 30 and 60 min of recovery. Gender differences in lymphocyte responses were also found between T3G and T3-4B (interaction effect, P = 0.047) and between T4-5G and T3-4B (interaction effect, P < 0.001) at 30 and 60 min of exercise, but not between T4-5G and T5B (interaction effect, P = 0.34) (Fig. 4). Lymphocyte counts in both genders returned to resting levels by 30 min of recovery.

Exercise caused decreases in total CD3+ proportions in T3G and T4-5G but not in T3-4B (interaction effect, P = 0.06 and P = 0.001, respectively). In T3-4B, the proportion of CD3+ cells did not fluctuate significantly from resting levels at any time point. A significant interaction was also found between T4-5G and T5B (interaction effect, P = 0.02), because of T4-5G having a higher CD3+ proportion at 30 min of recovery compared with T5B. When the total CD3+ cell count was considered, only differences between T4-5G and T3-4B were found at 30 min of exercise (interaction effect, P = 0.06).

Gender differences in CD3CD16+CD56+ cell proportions were found between T3G and T3-4B at 30 min of exercise (interaction effect, P = 0.03) and between T4-5G and T3-4B at 30 and 60 min of exercise (interaction effect, P = 0.049). These gender differences between pubertal groups in cell proportions were also found for CD3CD16+CD56+ cell counts (Fig. 5). Responses of CD3CD16+CD56+ cell proportions and counts were not significantly different between T4-5G and T5B (interaction effects, P ≥ 0.41).

The only cytokine-related gender difference within pubertal groups was for IL-8, which increased over time in T3G, but did not change in T3-4B (interaction effect, P = 0.05).

To assess whether the recovery of IL-6 and neutrophils, expressed as the change (▵) from their respective resting values, was related to subjects' FFM, Pearson correlations were calculated (Fig. 6). The relationship between IL-6 and FFM tended to be significant (r = 0.24, P = 0.07), whereas that between neutrophils and FFM achieved significance (r = 0.26, P = 0.049).

Back to Top | Article Outline


An improved understanding of normal perturbations to the immune system in healthy children and adolescents in response to physiologic stress carries significant health and clinical implications. To date, however, definitive investigations on this issue are clearly lacking. Therefore, we designed this study to provide a comprehensive analysis of immunologic changes in response to exercise in children and adolescents. Our results are descriptive in nature, but provide several novel findings with respect to the effects of age, puberty, and gender on exercise-induced changes in cellular and soluble components of the immune system.

Back to Top | Article Outline

Age effects.

We found that age-related differences in immunologic responses to exercise during childhood were more pronounced in girls than in boys. Although the literature contains studies of immunologic responses to exercise in 10- to 12-yr-old (10) and 14- to 16-yr-old (18) girls, to our knowledge, this study is the first to compare girls at different ages under identical experimental conditions. Notwithstanding significant methodologic differences in previous studies (laboratory (10) vs field (18)), their results suggested that age differences existed. In these previous studies, exercise increased total leukocytes by approximately 84% in older girls (18), but only by approximately 32% in younger girls (10). Likewise, the increase in NK cells (i.e., CD3CD16+CD56+), the most responsive cell type to exercise (23), was approximately 154% in older girls (18) and approximately 125% in the younger girls (10). In the girls in our study, similar age differences were observed for total leukocytes and NK cells insofar as the older girls had significantly greater changes. In addition, our results highlight the importance of reporting both relative proportions and absolute counts of immune cells insofar as greater decreases in the proportion of CD3+CD4+ cells in OG versus YG were offset by a greater overall lymphocytosis in the older girls. Thus, the exercise-induced increase in the absolute number of CD3+CD4+ cells was similar between OG and YG. In this regard, it is noteworthy that previous studies in children (5,10,24,26) have reported cell counts only.

Cytokine changes during and following exercise were also different between YG and OG in the present study. Specifically, IL-6 increased during exercise and into recovery in OG, but did not fluctuate from resting levels in YG, an age-related finding identical to our previous report comparing boys with men (30). IL-6 is an important cytokine that plays significant antiinflammatory and metabolic roles in adults (11). That IL-6 increased with exercise in adolescent girls is consistent with previous literature on this age group (18), but the literature is void of studies describing cytokine changes with exercise in groups of younger girls. Previous studies that have included young girls (25,31), however, have pooled their results with those of boys, making it impossible to discern gender responses. Importantly, we followed the cytokine response into the recovery period and that IL-6 continued to increase into recovery in OG suggests that age-related differences may be more related to inflammatory events than to metabolic regulation. Indeed, older girls also had a more pronounced neutrophilia during recovery from exercise, compared with younger girls, consistent with a greater inflammatory response. Previous work in female subjects has shown that, following exercise-induced muscle damage, adult women with regular menstrual cycles and not taking OCT significantly increased serum levels of creatine kinase (CK), an indirect marker of muscle damage, whereas premenarcheal girls showed no increase (2). It may be, therefore, that the younger girls in our study demonstrated smaller inflammatory-related responses because they experienced less trauma to the contracting skeletal muscle during the cycling task.

With respect to the boys in our study, the age-related difference in recovery leukocytosis and neutrophilia is also consistent with our previous observations in 9- and 10-yr-old boys, compared with men (30). We also found that TNF-α levels gradually increased into recovery in OB, but remained stable over time in YB, but in contrast to the girls, changes in IL-6 over time were similar between YB and OB. These results, in part, suggest that exercise may have induced a greater inflammatory response in the older versus younger boys, a similar observation to that found in the female subjects discussed above. Previous work in male subjects has also shown that boys, compared with men, respond to exercise designed to induce muscle damage with smaller increases in CK (27). Based on the Pearson correlations observed between recovery IL-6 and FFM and between recovery neutrophils and FFM in our study, it may be that an overall greater muscle mass contributed to the greater inflammatory-like responses in the older girls and boys. Although subjects cycled at the same percentage of their V̇O2max, the degree of muscle recruitment to sustain this exercise intensity may have been different between the groups, thus contributing to different degree of muscle damage. Regardless of why recovery kinetics of immune changes were faster in younger girls and boys, this observation is consistent with the notion that children generally recover more quickly following strenuous exercise.

Other immune-related responses to exercise in boys were not different between ages; NK cells increased by approximately 133% in YB and by approximately 177% in OB. The magnitude of these changes, however, is lower than previously reported in 14- to 18-yr-old boys (~238%) following a wrestling practice (16) and in 20- to 25-yr-old men (~200%) following identical exercise as used in the current study (30), but higher than in 9- and 10-yr-old boys (~110%) also using the current exercise protocol (30). The available data, therefore, support the notion of age differences in NK cell responses to exercise in male subjects, but the gap between age groups in the present study may have been too narrow to significantly highlight possible differences.

Back to Top | Article Outline

Puberty effects.

The findings of the present study show that when the physical maturity of children and adolescents is considered, many of the immunologic responses to exercise are more similar than dissimilar among pubertal groups. The lack of exercise effect on the proportion of CD3+CD8bright cells in T3 versus T4-5 girls, however, is consistent with less perturbation in the less mature girls. Our findings in the girls are somewhat limited because we only had appropriate numbers of subjects at two different pubertal stages for statistical comparisons. Notwithstanding the low number of subjects at Tanner stage 2 (N = 4), these girls consistently demonstrated much smaller exercise-induced increases in total lymphocyte and NK cell counts. Future work, therefore, needs to pursue possible pubertal effects on immune changes in larger numbers of girls representing more stages of puberty. In addition, the recovery of total leukocytes and neutrophils following exercise and changes in TNF-α levels maintained maturity-related effects in the boys.

Back to Top | Article Outline

Gender effects.

One of the most novel findings in this study was the significant gender difference in exercise-induced changes in various immune cells in adolescents, but not in younger children. Gender differences in adolescents were most notable in the overall lymphocyte and, more specifically, the NK cell response to exercise. Importantly, no gender differences were found in the proportion of NK cells, only the NK cell count, because of a greater overall lymphocytosis in the female adolescents, which further argues in favor of reporting both proportions and numbers of cells. Previous work from Cooper et al. (16) and Nemet et al. (18) reported increases in NK cells of approximately 238% following a wrestling practice in boys (16) and of approximately 154% following a water polo practice in girls (18). In contrast, increases in NK cells during exercise were greater in the girls versus the boys in the present study. It may be, however, that differences in exercise mode or intensity between the present and previous studies could help explain the contrasting gender-related results. In line with this hypothesis, we (29) and others (8) have shown that overall lymphocytosis in response to a cycling-specific protocol is greater in women than in men. Taken together, it seems that gender differences in lymphoid responses to exercise are revealed under controlled experimental conditions and begin to manifest sometime during adolescence. That gender differences in various immune measures were observed in adolescents, but not in younger children, and age differences were observed for girls, but not for boys, suggests that factors related to the presence of female sex hormones may be an important determinant of exercise-induced immune changes. We have recently shown, however, that exercise-induced changes in immune cells (total leukocytes, neutrophils, and lymphocytes) in women not taking OCT are not different between the midfollicular and midluteal phases of the menstrual cycle (29). Therefore, the presence of, and fluctuations in, endogenous sex hormones may not be overly important in exercise-induced perturbations to the immune system.

Another novel finding in this study was the IL-8 response to exercise in healthy children and adolescents. With few exceptions, IL-8 levels did not change after 60 min of exercise, but significantly increased to approximately 75% above resting values at 60 min of recovery. In adults, increases in systemic IL-8 levels appear to require prolonged strenuous exercise (19). A recent study, however, demonstrated that IL-8 protein expression is increased within skeletal muscle during and following exercise performed at a moderate intensity (1). Increased levels of this cytokine following exercise in children may represent a normal adaptation to physiologic stress and, therefore, be beneficial to overall growth and development in a number of ways. First, through its chemotactic effect on endothelial cells, IL-8 contributes to angiogenesis (13), which may result in an increase delivery of nutrients and anabolic mediators (e.g., insulinlike growth factor-1 (IGF-1)). Second, IL-8 can induce stem cell mobilization from the bone marrow (12), and recent evidence that bone marrow-derived cells mobilized with physical activity can incorporate into myofibers (22) suggests a potential link to muscle hypertrophy. Finally, Zaldivar et al. (35) recently presented data, in abstract form, showing that immune cells mobilized during exercise express increased levels of intracellular growth hormone (GH) and IGF-1. Given its potent chemoattractant effect on circulating immune cells (21), increased levels of IL-8 during and following exercise may serve to increase infiltration of GH- and IGF-1-producing immune cells into skeletal muscle, thus contributing to an anabolic response. Collectively, the above examples suggest that our novel observation in children and adolescents of exercise-induced increases in IL-8 may have particular relevance to the interaction of exercise and growth during childhood.

The results of our study may also have important implications in the pediatric clinic. In general terms, our data indicate that the age, puberty, and gender of a patient should be considered when interpreting their immune response to an acute bout of exercise. In this regard, a blunted immune response to acute exercise used as a provocation test to identify or monitor the effects of various treatment protocols (e.g., corticosteroids) in children with asthma, cancer, human immunodeficiency virus (HIV), or other immune-related diseases, for example, should not necessarily be viewed as pathologic, but in fact could reflect an appropriate response based on the patient's physical development. Our data also show that, unlike in adults where high intensity exercise often causes a reduction in immune cell counts during the recovery period (23), similar levels of intense exercise can be performed by young children without major perturbations to their immune system. Consequently, children with a compromised immune system recovering from organ transplants, cancer, or other diseases are likely to derive improvement in cardiorespiratory fitness with this type of exercise, but without further immune impairment. Likewise, the smaller inflammatory-related responses during the postexercise period in our younger subjects further suggest that children with inflammatory myopathies (e.g., juvenile dermatomyositis) can safely engage in this type of exercise without an acute exasperation of their condition. The above theories, although speculative, represent a fruitful area for future research and the potential of exercise immunology in the pediatric clinic.

Back to Top | Article Outline

Summary and perspectives.

The findings reported herein represent a comprehensive examination of age, puberty, and gender effects on immunologic changes in response to exercise in healthy children and adolescents. By design, we recruited children at two distinct chronologic ages to minimize possible effects of previous antigenic experience on immune measures and confounding effects of age-associated factors on pubertal comparisons. The data show that exercise performed at the same relative intensity results in smaller overall perturbations to the immune system in young boys and girls compared with older adolescent boys and girls, with differences more pronounced in the girls. In general, younger individuals also experienced faster recovery of these perturbations. Our data, however, cannot provide mechanisms for the observed age-, puberty-, and gender-related differences, and the clinical and biological significance of these findings remain to be determined. Recently, Cooper et al. (6) proposed that exercise-induced changes in immune and inflammatory mediators (e.g., cytokines) in children may have implications for overall growth and development. Studies from their laboratory suggest that cytokine changes to exercise tend to be smaller in children (25,31) versus adolescents (17,18) and that IGF-1 levels either decrease slightly in the children (25,31) or remain largely unchanged in the adolescents (17,18). Together, these results suggest that, in young children, anabolic mediators can be particularly sensitive to acute changes in inflammatory-related cytokines. Given the inhibitory effects of inflammatory-related cytokines (e.g., IL-6) on muscle protein synthesis (36), we suggest that one reason why young children are relatively resistant to major inflammatory responses during and following exercise may be to minimize disruption to anabolic mediators such as IGF-1, which is conducive to optimal adaptation and muscle growth.

We thank M. De Jonge, M. Kubacki, R. Trott, J-H. Lee, M. Hamadeh, and A. Mark for assistance with this experiment. We also thank Dr. Craig Horswill for valuable comments on an earlier version of this manuscript. The extraordinary effort and time of our subjects are gratefully acknowledged.

The financial support of the Gatorade Sports Science Institute and BD Biosciences is gratefully acknowledged.

Flow cytometry data were acquired from the McMaster Flow Cytometry Facility, supported by the Canadian Institutes of Health Research.

Sadly, Dr. Oded Bar-Or passed away during the review process of this paper; his contributions will never be forgotten.

Back to Top | Article Outline


1. Akerstrom, T., A. Steensberg, P. Keller, C. Keller, M. Penkowa, and B. K. Pedersen. Exercise induces interleukin-8 expression in human skeletal muscle. J. Physiol. 563:507-516, 2005.
2. Arnett, M. G., R. Hyslop, C. A. Dennehy, and C. M. Schneider. Age-related variations of serum CK and CK MB response in females. Can. J. Appl. Physiol. 25:419-429, 2000.
3. Bartlett, J. A., A. R. Goldklang, S. J. Schleifer, and S. E. Keller. Immune function in healthy inner-city children. Clin. Diagn. Lab. Immunol. 8:740-746, 2001.
4. Bartlett, J. A., S. J. Schleifer, M. K. Demetrikopoulos, B. R. Delaney, S. C. Shiflett, and S. E. Keller. Immune function in healthy adolescents. Clin. Diagn. Lab. Immunol. 5:105-113, 1998.
5. Boas, S. R., M. J. Danduran, S. A. McColley, K. Beaman, and M. R. O'Gorman. Immune modulation following aerobic exercise in children with cystic fibrosis. Int. J. Sports Med. 21:294-301, 2000.
6. Cooper, D. M., D. Nemet, and P. Galassetti. Exercise, stress, and inflammation in the growing child: from the bench to the playground. Curr. Opin. Pediatr. 16:286-292, 2004.
7. Dandona, P., A. Aljada, and A. Bandyopadhyay. Inflammation: the link between insulin resistance, obesity and diabetes. Trends Immunol. 25:4-7, 2004.
8. De Lanne, R., J. R. Barnes, and L. Brouha. Hematological changes during muscular activity and recovery. J. Appl. Physiol. 15:31-36, 1960.
9. Dill, D. B., D. L. Costill. Calculation of percentage changes in volumes of blood, plasma, and red cells in dehydration. J. Appl. Physiol. 37:247-248, 1974.
10. Eliakim, A., B. Wolach, E. Kodesh, et al. Cellular and humoral immune response to exercise among gymnasts and untrained girls. Int. J. Sports Med. 18:208-212, 1997.
11. Febbraio, M. A., B. K. Pedersen. Muscle-derived interleukin-6: mechanisms for activation and possible biological roles. FASEB J. 16:1335-1347, 2002.
12. Fibbe, W. E., J. F. Pruijt, G. A. Velders, et al. Biology of IL-8-induced stem cell mobilization. Ann. N. Y. Acad. Sci. 872:71-82, 1999.
13. Koch, A. E., P. J. Polverini, S. L. Kunkel, et al. Interleukin-8 as a macrophage-derived mediator of angiogenesis. Science 258: 1798-1801, 1992.
14. Matsudo, S. M. M., and V. K. R. Matsudo. Self-assessment and physician assessment of sexual maturation in Brazilian boys and girls: Concordance and reproducibility. Am. J. Hum. Biol. 6: 451-455, 1994.
15. Molnar, D. The prevalence of the metabolic syndrome and type 2 diabetes mellitus in children and adolescents. Int. J. Obes. Relat. Metab. Disord. 28(Suppl. 3):S70-S74, 2004.
16. Nemet, D., P. J. Mills, and D. M. Cooper. Effect of intense wrestling exercise on leucocytes and adhesion molecules in adolescent boys. Br. J. Sports Med. 38:154-158, 2004.
17. Nemet, D., Y. Oh, H. S. Kim, M. Hill, and D. M. Cooper. Effect of intense exercise on inflammatory cytokines and growth mediators in adolescent boys. Pediatrics 110:681-689, 2002.
18. Nemet, D., C. M. Rose-Gottron, P. J. Mills, and D. M. Cooper. Effect of water polo practice on cytokines, growth mediators, and leukocytes in girls. Med. Sci. Sports Exerc. 35:356-363, 2003.
19. Nieman, D. C., D. A. Henson, L. L. Smith, et al. Cytokine changes after a marathon race. J. Appl. Physiol. 91:109-114, 2001.
20. Nieman, D. C., M. W. Kernodle, D. A. Henson, G. Sonnenfeld, and D. S. Morton. The acute response of the immune system to tennis drills in adolescent athletes. Res. Q. Exerc. Sport 71: 403-408, 2000.
21. Opdenakker, G., W. E. Fibbe, and J. Van Damme. The molecular basis of leukocytosis. Immunol. Today 19:182-189, 1998.
22. Palermo, A. T., M. A. LaBarge, R. Doyonnas, J. Pomerantz, and H. M. Blau. Bone marrow contribution to skeletal muscle: a physiological response to stress. Dev. Biol. 279:336-344, 2005.
23. Pedersen, B. K., and L. Hoffman-Goetz. Exercise and the immune system: regulation, integration, and adaptation. Physiol. Rev. 80:1055-1081, 2000.
24. Perez, C. J., D. Nemet, P. J. Mills, T. P. Scheet, M. G. Ziegler, D. M. Cooper. Effects of laboratory versus field exercise on leukocyte subsets and cell adhesion molecule expression in children. Eur. J. Appl. Physiol. 86:34-39, 2001.
25. Scheett, T. P., P. J. Mills, M. G. Ziegler, J. Stoppani, and D. M. Cooper. Effect of exercise on cytokines and growth mediators in prepubertal children. Pediatr. Res. 46:429-434, 1999.
26. Shore, S., and R. J. Shephard. Immune responses to exercise and training: a comparison of children and young adults. Pediatr. Exerc. Sci. 10:210-226, 1998.
27. Soares, J. M., P. Mota, J. A. Duarte, and H. J. Appell. Children are less susceptible to exercise-induced muscle damage than adults: a preliminary investigation. Pediatr. Exerc. Sci. 8:361-367, 1996.
28. Tanner, J. M. Growth at Adolescence. Oxford, UK: Blackwell Scientific, 1962, pp. 32-37.
29. Timmons, B. W., M. J. Hamadeh, M. C. Devries, and M. A. Tarnopolsky. Influence of gender, menstrual phase, and oral contraceptive use on immunological changes in response to prolonged cycling. J. Appl. Physiol. 99:979-985, 2005.
30. Timmons, B. W., M. A. Tarnopolsky, and O. Bar-Or. Immune responses to strenuous exercise and carbohydrate intake in boys and men. Pediatr. Res. 56:227-234, 2004.
31. Tirakitsoontorn, P., E. Nussbaum, C. Moser, M. Hill, and D. M. Cooper. Fitness, acute exercise, and anabolic and catabolic mediators in cystic fibrosis. Am. J. Respir. Crit. Care Med. 164: 1432-1437, 2001.
32. Tremblay, M. S., and J. D. Willms. Secular trends in the body mass index of Canadian children. CMAJ. 163:1429-1433, 2000.
33. Troiano, R. P., and K. M. Flegal. Overweight children and adolescents: description, epidemiology, and demographics. Pediatrics 101:497-504, 1998.
34. Wolach, B., A. Eliakim, R. Gavrieli, et al. Aspects of leukocyte function and the complement system following aerobic exercise in young female gymnasts. Scand. J. Med. Sci. Sports 8:91-97, 1998.
35. Zaldivar, F., D. Nemet, J. Larsnon, and D. M. Cooper. Does exercise alter intracellular cytokine expression patterns in PMBCs? Med. Sci. Sports Exerc. 36:S87, 2004.
36. Zoico, E., and R. Roubenoff. The role of cytokines in regulating protein metabolism and muscle function. Nutr. Rev. 60:39-51, 2002.

Cited By:

This article has been cited 1 time(s).

Journal of Pediatric Gastroenterology and Nutrition
Relation Between Adiposity and Disease Risk Factors in Mexican American Children
McFarlin, BK; Johnston, CA; Tyler, C; O'Connor, DP; Strohacker, KA; Reeves, R; Jackson, AS; Foreyt, JP
Journal of Pediatric Gastroenterology and Nutrition, 49(4): 450-455.
PDF (389) | CrossRef
Back to Top | Article Outline


©2006The American College of Sports Medicine