Prolonged running results in an increase in the circulating concentrations of the cytokines tumor necrosis factor (TNF)-α, interleukin-1β (IL-1β), IL-6, and IL-1 receptor antagonist (ra) (23,24,37). These changes occur in a cascade-like manner, with modest increases in TNF-α and IL-1β and a marked increase in IL-6, peaking at the end of exercise, preceding a marked increase in IL-1ra, which peaks at 1–2 h after exercise (22–24,37).
Contracting skeletal muscle is a major source of IL-6 during exercise (34). Its appearance in the circulation is decreased by glucose ingestion (19) and endurance training (12), whereas low preexercise muscle glycogen results in lower blood glucose (4), a greater transcriptional activity of the IL-6 gene in muscle, and higher IL-6 concentrations during subsequent exercise (4,17). Resting muscle glycogen concentrations are correlated negatively with postexercise IL-6 messenger RNA (12) and IL-6 release from muscle during exercise (14).
Recombinant human IL-6 increases hepatic glucose production (36) and fasting blood glucose concentrations (38), and stimulates the systemic release and oxidation of fatty acids (27,40). In addition, during exercise, recombinant IL-6 increases endogenous glucose production, whole-body glucose disposal and metabolic clearance rate (8). Augmented hepatic glucose output, increased uptake of glucose by other tissues and fat oxidation in skeletal muscle suggests that, during exercise, IL-6 might be a CHO sensor mobilizing substrates and/or augmenting substrate delivery to the working muscle.
Helge et al. (14) report a greater release of IL-6 with one-legged, knee extension exercise at 85% Wmax compared with 65% Wmax but with no significant difference in venous IL-6 concentrations at the end of exercise. One study has demonstrated a measurable difference in the IL-6 concentrations, with higher IL-6 concentrations immediately after running at 85% V˙O2max compared with 60% V˙O2max (26).
In contrast, the one- to twofold increase in TNF-α and IL-1β concentrations during running (23,24,37) seems to be specific to this mode of exercise, with no changes in these cytokines reported with other forms of endurance exercise (31,34). The reason for this apparently running-specific response remains unclear, but it might be related to muscle damage, which results in neutrophil infiltration and proinflammatory cytokine accumulation (10) and occurs with running (22) but not cycling nor knee extensor exercise (4,35).
Unlike in fever and sepsis, monocytes are not the source of TNF-α during running (32). Both TNF-α and IL-1β are expressed in skeletal muscle (4,19,24,34), and this expression may increase with prolonged running (19,24). However, at least in the case of TNF-α, there is no evidence of a net release from muscle into the circulation during exercise (9,34). While the source of the increases in TNF-α and IL-1β during exercise remains unconfirmed, the factors that regulate their appearance in the circulation during exercise seem to differ from those that regulate IL-6.
The purpose of this study was to examine the effect of increasing exercise intensity on the time course of changes in circulating cytokines with acute running. We hypothesized that running would increase TNF-α, IL-1β, and IL-6 concentrations and that the IL-6 response would be enhanced with increasing exercise intensity. In addition, as increased IL-6 during exercise is followed by an increase in IL-1ra (20,22–24,37) and an IL-6 infusion stimulates IL-1ra production (33), a secondary hypothesis was that the increase in IL-1ra with exercise would also be enhanced by exercise intensity.
Ten physically active men were recruited for the study, which was approved by the QinetiQ Research Ethics Committee. Participants were nonsmokers, were free from musculoskeletal injury, and did not suffer from any condition or take any medication known to affect inflammation. Participants who satisfied the inclusion criteria provided their written informed consent.
Participants completed two preliminary visits for medical screening, habituation with trial procedures, and measurement of V˙O2max. Participants then completed three counterbalanced 8-d experimental conditions, separated by a minimum of 1 wk, during which they refrained from unprescribed physical activity and ate a controlled diet. On days 1–3, participants refrained from all physical activity and followed their prescribed diet. On day 4, participants performed a single 60-min bout of treadmill running at 55%, 65%, and 75% of V˙O2max, followed by 3 h of recovery. On days 5–8, participants refrained from all physical activity, followed their prescribed diet, and attended the laboratory for follow-up analysis (FU1–FU4). Participants were asked to report any symptoms of illness or fever in the days leading up to the study. If such an episode occurred, the trial was postponed until participants had been clear of symptoms for a minimum of 5 d.
Participants completed a 3-d food diary consisting of two weekdays and one weekend day to calculate habitual daily energy intake and macronutrient composition. Participants were issued with a set of calibrated weighing scales to measure food intake and received both verbal and written instructions. Food diaries were analyzed using nutritional analysis software (Microdiet V2; Downlee Systems Limited, Chapel-en-le-Frith, UK).
Assessment of cardiorespiratory responses to level running and aerobic power.
To establish the association between oxygen uptake (V˙O2) and running velocity during level running, participants completed a 20-min submaximal run on a treadmill (XELG 70 ERGO; Woodway, Waukesha, WI), consisting of four, 5-min stages. Sixty-second samples of expired air (inspiration to inspiration) were collected in the final minute of each stage. After a 30-min rest, participants completed a discontinuous, incremental exercise test to exhaustion to establish V˙O2max using a modified Taylor protocol (13). HR was monitored continuously throughout the test (Vantage NV; Polar Electro Oy, Kempele, Finland). The results of the two tests were used to estimate the treadmill velocity corresponding to 50%, 55%, 65%, and 75% V˙O2max during level running based on the regression line of V˙O2 and velocity.
Experimental dietary provision.
A diet consisting of approximately 55% CHO, 30% fat, and 15% protein, and isocaloric with their habitual diet, was designed for each participant based on individual dietary habits. Participants were provided with three menus that were given in a 3-d cyclic order with menu A on days 1 and 5, menu B on days 2 and 6, and menu C on days 3 and 7 (see the next section for details of diet on day 4). During the experimental period, participants provided their own food and were given both verbal and written instructions concerning the quantity and preparation of their meals and timings. Participants received a set of calibrated weighing scales to aid them with food preparation.
After an overnight fast, participants arrived at the laboratory at 7:30 a.m., voided, and had their nude body mass measured (ID7; Mettler-Toledo, Giessen, Germany). Participants subsequently adopted a semirecumbent position on a bed and had a cannula (18-gauge, 1.2 × 45 mm; Becton Dickinson, Franklin Lakes, NJ) inserted into a prominent forearm vein, where it remained until the final blood sample was collected. The cannula was kept patent with 5 mL of an isotonic saline solution (0.9% NaCl).
A fasting baseline blood sample was collected at 8:00 a.m. for measurement of all biochemical markers and exercise commenced at 8:15 a.m. (Fig. 1). Exercise bouts consisted of 60 min of treadmill running preceded by a 5-min warm-up at 50% V˙O2max, separated by 5 min for volitional stretching. Sixty-second samples of expired air and RPE (21) were collected after 18, 38, and 58 min of exercise. HR was recorded continuously and water was consumed ad libitum during exercise (0.27 ± 0.14, 0.28 ± 0.10, and 0.34 ± 0.18 L in the 55%, 65%, and 75% V˙O2max conditions). The ambient room temperature during exercise was 20°C ± 1°C in all conditions.
On completion of exercise, participants dried off, had their nude body mass measured, and rested in a semirecumbent position for a further 3 h. The difference between pre- and postexercise body mass was calculated (0.72 ± 0.27, 0.80 ± 0.16, and 0.93 ± 0.24 kg in the 55%, 65%, and 75% V˙O2max conditions), and participants were provided with 1.5 mL of plain water during the recovery period for every gram change in body mass. Blood samples were collected at identical time points in the three conditions: at 8:00 a.m. (BASE), after 20, 40, and 60 min of exercise (EX20–EX60) and 0.5, 1.0, 2.0, and 3.0 h of recovery (R0.5–R3.0) (Fig. 1).
Participants consumed a standardized diet (13.2 MJ, 53% CHO, 32% fat, and 15% protein) divided into three meals. The first meal of the day was eaten in the laboratory immediately after the final blood sample at 3 h after exercise. The further two meals were taken home by participants to be consumed at 4:00 p.m. and 7:30 p.m. (Fig. 1). Only plain water was allowed after 9:00 p.m. The same diet was provided in all three conditions.
Days 5–8 (FU1–FU4).
After an overnight fast, participants voided and had their nude body mass measured. A fasting blood sample was drawn by venepuncture at 8:00 a.m. Only plain water was allowed after 9:00 p.m. on days before laboratory visits.
Samples of expired air were collected into evacuated Douglas bags. The bags were emptied through a flow controller and volume counter and were analyzed for fractions of oxygen and carbon dioxide (Series 1400; Servomex Ltd., Sussex, UK) calibrated using certified reference gases (100% N, 16% O2, 5% CO2; BOC Gases, Surrey, UK).
For the measurement of TNF-α, IL-1β, IL-6, IL-1ra, and cortisol, blood was transferred into precooled tubes containing 15%, 0.12 mL of K3 EDTA (Vacutainer System; Becton Dickinson), gently inverted 8–10 times, and centrifuged immediately. For measurement of creatine kinase (CK), blood was transferred into precooled standard tubes (Vacutainer System; Becton Dickinson) and left to clot at room temperature for 60 min. All tubes were centrifuged at 2000 rpm and 5°C for 10 min, samples were separated, and aliquots were stored at −70°C until analyses. For the measurement of glucose and lactate, whole blood was transferred into precooled tubes containing fluoride oxalate. Tubes were gently inverted 8–10 times and analyzed immediately in duplicate (2300 Stat Plus; Yellow Springs Instruments, Inc., Yellow Springs, OH).
TNF-α, IL-1β, IL-6, and IL-1ra were measured using commercial, solid-phase, enzyme-linked, immunosorbent assay (Quantikine HS; R&D Systems Ltd., Abingdon, UK). The TNF-α assay has a detection limit of 0.12 pg·mL−1 and an inter/intra-assay coefficient of variation (CV) of <14% across the range 0.5–32.0 ng·L−1. The IL-1β assay has a detection limit of 0.1 ng·L−1 and an inter/intra-assay CV of <12% across the range 0.5–8.0 ng·L−1. The IL-6 assay has a detection limit of 0.039 pg·mL−1 and an inter/intra-assay CV of <10% across the range 0.15–10 ng·L−1. The IL-1ra assay has a detection limit of 2 ng·L−1 and an inter/intra-assay CV of <8% across the range 50–3000 ng·L−1. Cortisol was measured in plasma using an electrochemiluminescence immunoassay on a Roche Modular E170 (Roche, Lewes, UK). The assay has a sensitivity of 8 nmol·L−1 established from precision profiles (22% CV of duplicates) and a CV of <6% from 16 to 1750 nmol·L−1. CK was measured using standard reagents and method on a P module analytical system (Roche). The inter/intra-assay CV across the measuring range (10–1000 U·L−1) is <5%.
All data are presented as mean ± SD unless otherwise stated. Statistical significance was accepted at an α level of P < 0.05. Paired-samples t-tests were used to compare habitual with experimental dietary data. A one-way ANOVA was used to compare variables relating to exercise performance and baseline concentrations of cortisol, TNF-α, IL-6, IL-1ra, and CK in the 55%, 65%, and 75% V˙O2max conditions, with the Student–Newman–Keuls (SNK) post hoc tests performed where appropriate.
All biochemical data and body mass were analyzed using a linear mixed model (LMM) ANOVA with the factors time (sampling) and condition (55% vs 65% vs 75% V˙O2max) included and participants as a random within-condition factor. The assumptions of the LMM were investigated by examining the distribution of residuals and the pattern of residuals versus fitted values. Where nonnormality or nonconstant variance was observed, a transformation was applied to the data so that the assumptions were satisfied. Only cortisol data did not require transformation. The assumptions of the LMM were satisfied after log transformations for glucose, lactate, TNF-α, IL-6, and IL-1ra, whereas CK required a log-square root transformation.
Where there was a significant main effect of time but no significant condition × time interaction, each subsequent time point was compared against BASE using a pooled mean using the Dunnett test with BASE as the “control.” When the condition × time interaction was significant, within each group, subsequent time points were compared against BASE using the Dunnett test with BASE as the control and groups were compared with each other at all time points using the SNK test.
Pearson correlation coefficient was calculated to examine the relationship between postexercise (EX60) IL-6 concentrations and postexercise glucose and lactate concentrations and RER in the final minute of exercise. Correlation coefficients were also calculated between absolute concentrations of, and fold increase in, IL-6 and IL-1ra with exercise in the 55%, 65%, and 75% V˙O2max conditions. All statistical analyses were performed with the SPSS v15 (SPSS, Inc., Chicago, IL) with the exception of the Dunnett and SNK tests, which were performed with Statistica (StatSoft, Inc., Tulsa, OK).
Participant characteristics are shown in Table 1.
The energy content of the experimental diets was not significantly different from habitual energy intake (P = 0.440), and there were no significant differences for any other dietary variable (data not shown).
There was a significant main effect of time (P < 0.05) but no condition × time interaction (P = 0.970) for body mass data. Measured from day 4 (BASE) to day 8 (FU4), pooled, mean data showed a decrease (approximately 0.5%) in body mass from BASE that was significant (P < 0.05) at FU4 only (data not shown).
Oxygen uptake, % V˙O2max, HR, and RPE were significantly higher (P < 0.001) at 65% V˙O2max compared with 55% V˙O2max, and at 75% V˙O2max, were higher (P < 0.001) than at both 55% and 65% V˙O2max (Table 2). RER was not significantly different between 55% and 65% V˙O2max but was higher at 75% V˙O2max compared with 55% (P < 0.001) and 65% (P < 0.01) V˙O2max.
Glucose, Lactate, and Cortisol
There was no effect of exercise intensity on blood glucose concentrations (condition × time interaction, P = 0.449). Pooled, mean concentrations were significantly increased (P < 0.001) from BASE throughout exercise (Fig. 2A). Concentrations increased from 4.3 mmol·L−1 at BASE to 4.5 mmol·L−1 at EX60 with exercise at 55% and 65% V˙O2max and to 5.3 ± 0.8 mmol·L−1 at 75% V˙O2max. After exercise, glucose concentrations were decreased from BASE at R1.0 (P < 0.001) and R3.0 (P < 0.01).
Analysis of blood lactate concentrations showed a main effect of time (P < 0.001) and a condition × time interaction (P < 0.001). Concentrations were significantly increased (P < 0.001) from BASE at EX20 in the 55% V˙O2max condition, throughout exercise at 65% V˙O2max (P < 0.001), and throughout exercise and up to R1.0 at 75% V˙O2max (P < 0.01) (Fig. 2B). The increase at 65% V˙O2max was greater than that at 55% V˙O2max, resulting in higher (P < 0.001) concentrations at EX20, EX40, and EX60, whereas lactate concentrations at 75% V˙O2max were higher throughout exercise and up to R1.0 (P < 0.001) than at both 55% and 65% V˙O2max.
There were no significant differences between baseline cortisol concentrations in the three conditions (P = 0.618). Analysis of cortisol concentrations showed a main effect of time (P < 0.001) and a condition × time interaction (P < 0.001). Cortisol concentrations at 55% V˙O2max were significantly decreased from BASE throughout exercise (P < 0.05), with a trend being shown at R0.5 (P = 0.06) (Fig. 2C). Subsequently, cortisol concentrations decreased further and were significantly decreased from BASE at R1.0 (P < 0.001), R2.0 (P < 0.001), and at R3.0 (P < 0.001), where concentrations were reduced to 65% ± 17% of BASE levels. At 65% V˙O2max, concentrations were not significantly different from BASE during exercise but were significantly lower than BASE at R0.5 (P < 0.05), R1.0 (P < 0.001), R2.0 (P < 0.001), and at R3.0 (P < 0.001), where concentrations were reduced to 51% ± 18% of BASE levels. At 75% V˙O2max, cortisol concentrations increased during exercise and, by EX60, were significantly increased from BASE (+30% ± 37%, P < 0.01), resulting in higher cortisol concentrations compared with 55% V˙O2max at EX20 (P < 0.05) and EX40 (P < 0.001), and with both 55% V˙O2max (P < 0.001) and 65% V˙O2max (P < 0.001) at EX60. Concentrations were not significantly different from BASE from R0.5 to R2.0 but remained higher than both 55% V˙O2max and 65% V˙O2max at R0.5 (P < 0.001), R1.0 (P < 0.001), and R2.0 (P < 0.01). At R3.0, concentrations were lower (P < 0.001) than BASE but were no longer significantly different from 55% V˙O2max or 65% V˙O2max. Concentrations were not different from BASE in any condition from FU1–FU4.
There were no significant differences between baseline TNF-α concentrations in the three conditions (P = 0.381). There was a significant main effect of time (P < 0.01) but no significant group × time interaction (P = 0.214) for TNF-α. Pooled, mean concentrations were significantly (P < 0.01) increased from BASE at EX60 (18%–27%) and remained increased (P < 0.05) at R0.5 (11%–24%; Fig. 3A). Concentrations were not significantly different from BASE at R1.0, and there were no further differences thereafter.
IL-1β was undetectable in six participants and another one was considered to be an outlier because his baseline IL-1β concentrations (2.83–4.56 ng·L−1) were 7–11 times higher than the mean of the remaining three participants (0.37–0.41 ng·L−1). Owing to the small (n = 3) data set, no statistical analysis was performed on IL-1β data (data not shown).
There were no significant differences between baseline IL-6 concentrations in the three conditions (P = 0.381). There was a both significant main effect of time (P < 0.001) and a significant group × time interaction (P = 0.001) for IL-6. In the 55% V˙O2max condition, IL-6 concentrations were significantly increased (P < 0.001) from BASE at EX40 and remained so throughout exercise and the first 3 h of recovery (Fig. 3B). Peak concentrations occurred at R3.0, increased 530% from BASE. In the 65% V˙O2max condition, concentrations were also increased (P < 0.001) from BASE at EX40 through to R3.0, with peak concentrations occurring at EX60, increased 640%. There were no significant differences between the 55% and 65% V˙O2max conditions at any time point. In the 75% V˙O2max condition, concentrations were increased (P < 0.05) from BASE at EX20 and had increased by 500% at EX40 (P < 0.001), where they were higher than in the 55% V˙O2max (P < 0.01) and 65% V˙O2max conditions, although this did not reach the assigned level of statistical significance (P = 0.07). Peak concentrations occurred at EX60, where they were increased (P < 0.001) 1200% from BASE and higher (P < 0.001) than in both the 55% and 65% V˙O2max conditions. Concentrations remained higher (P < 0.001) than BASE in the first 3 h of recovery and higher than at 55% (P < 0.001) and 65% (P < 0.01) V˙O2max at R0.5 but not thereafter. From FU1 to FU4, IL-6 was not significantly different from BASE in any group.
There were no significant differences between baseline IL-1ra concentrations in the three conditions (P = 0.488). There was a significant main effect of time (P < 0.001) and a significant group × time interaction (P < 0.05) for IL-1ra. At 55% V˙O2max, there was no significant change in IL-1ra concentrations (Fig. 3C). At 65% V˙O2max, concentrations were significantly increased (P < 0.05) at EX40 and remained increased (P < 0.01) up to R1.0 where concentrations peaked, increased 56% from BASE. At 75% V˙O2max, concentrations were higher than BASE at EX20 (P < 0.05), EX40 (P < 0.001) and EX60 (P < 0.001). Peak concentrations occurred at R1.0, where they were increased (P < 0.001) 210% from BASE and higher (P < 0.05) than at 55% and 65% V˙O2max. Concentrations remained increased (P < 0.001) from BASE at R2.0 (+209%) and R3.0 (+190%) and, at R2.0, were higher than at 55% (P < 0.05) and 65% (P < 0.01) V˙O2max. IL-1ra concentrations were not significantly different from BASE in any group at FU1 or thereafter.
There were no significant differences between baseline CK concentrations in the three conditions (P = 0.556). There was a significant main effect of time (P < 0.001) but no significant group × time interaction (P = 0.592) for CK concentrations. Pooled, mean concentrations were significantly increased from BASE at FU1 (P < 0.001) and FU2 (P < 0.01) but not thereafter, with peak concentrations occurring at FU1 in all three groups (55% V˙O2max, +52%; 65% V˙O2max, +80%; 75% V˙O2max, +163%) (Fig. 4).
Across all three conditions, postexercise IL-6 concentrations correlated with postexercise glucose (r = 0.477, P < 0.01) and lactate (r = 0.740, P < 0.001) concentrations and RER during the final minute of exercise (r = 0.470, P < 0.01). At 75% V˙O2max, the correlation between the fold increase in IL-6 at EX60 and the fold increase in IL-1ra at R2.0 was significant (r = 0.782, P < 0.01), as was the correlation between the absolute concentrations at these time points (r = 0.629, P = 0.05). In the 65% V˙O2max condition, fold increase in IL-6 at EX60 correlated with the fold increase in IL-1ra at R1.0 (r = 0.843, P < 0.01), whereas the IL-6 concentration at R1.0 correlated with the IL-1ra concentration at R2.0 (r = 0.828, P < 0.01). In the 55% V˙O2max condition, there were no significant correlations between the fold increases in IL-6 and IL-1ra, although the correlation between IL-6 concentrations at EX60 and IL-1ra concentrations at R0.5 and R1.0, both approached significance (r = 0.570, P = 0.086 and r = 0.551, P = 0.096).
This is the first study to examine the effects of increasing exercise intensity on both pro- and anti-inflammatory cytokine concentrations and showed that: 1) 60 min of acute running at 55%–75% V˙O2max increased TNF-α, IL-6, and IL-1ra concentrations, but only IL-6 and IL-1ra were affected by exercise intensity, with higher concentrations at 75% V˙O2max compared with 55% and 65% V˙O2max; 2) the intensity-dependent effect on IL-6 was accompanied by a similar effect on RER values; and 3) changes in IL-1ra were correlated with changes in IL-6, particularly at higher exercise intensities.
Compared with exercise at 55% and 65% V˙O2max, we observed a significantly greater increase in IL-6 at 75% V˙O2max. Contracting skeletal muscle seems to be a major source of IL-6 during exercise (9,34,35). The time course of increase in circulating IL-6—becoming more rapid toward the end of exercise—mimics the increase in net IL-6 release from muscle (34,35), whereas muscle release can account for the overall increase during acute, endurance exercise (35). Therefore, although we did not take measures across working muscle, based on the evidence described above and the pattern of increase in circulating IL-6, it is likely that release by the active skeletal muscle was a major contributor to the increased plasma concentrations shown.
Reduced muscle glycogen availability results in higher plasma IL-6 concentrations during exercise (4,17), and preexercise muscle glycogen concentration correlates inversely with postexercise muscle IL-6 messenger RNA (12). Accordingly, increasing exercise intensity, which increases CHO metabolism through augmented muscle glycogen turnover and uptake of circulating glucose by working muscle (29,39), results in a greater release of IL-6 from muscle, with the release late in exercise correlating with postexercise glycogen concentration (14). Helge et al. (14) observed no differences in plasma IL-6 concentrations with leg extensor exercise, but subsequently, a greater increase in IL-6 concentrations was reported during 60 min of running at 85% V˙O2max compared with 60% V˙O2max (26). We observed an association between postexercise IL-6 and lactate concentrations, suggesting that the higher IL-6 concentrations with increasing exercise intensity might be the result of greater metabolic demands. The results of Peake et al. (26) indicate a threshold above which there is an increase in the sensitivity of the change in IL-6 to increasing exercise intensity, and our findings extend those of Peake et al. by confirming the presence of this threshold between 65% and 75% V˙O2max.
As well as confirming the presence of a threshold, our results also demonstrate that the greater increase in IL-6 concentrations was accompanied by a significantly higher RER value, indicative of a greater reliance on CHO as a substrate, as well as an overall correlation between postexercise IL-6 and RER. Our results also demonstrate that differences in IL-6 concentrations were dependent on exercise duration: after 20 min of running, there was no effect of exercise intensity, whereas after 40 min, concentrations at 75% V˙O2max were higher compared with 55% V˙O2max, and, by the end of exercise, higher compared with both 55% and 65% V˙O2max. In addition, only at 75% V˙O2max were concentrations increased from baseline after 20 min of exercise. A greater reliance on CHO will result in a more rapid muscle glycogenolysis and, with increasing exercise duration, an increased need for alternative fuel sources to maintain substrate delivery to working muscle. It is possible, therefore, that the association between postexercise IL-6 and glucose concentrations in the present study reflects a direct and/or indirect effect of IL-6 on endogenous glucose production (38). Together, our results seem consistent with data showing that IL-6 release during exercise acts to increase hepatic glucose output and uptake of glucose by other tissues (8) and fat oxidation in skeletal muscle (40).
In contrast to IL-6, the increase in TNF-α concentrations was not different between 55% and 75% V˙O2max. The source of the increased TNF-α with exercise remains uncertain. Unlike with inflammation, monocyte activity is either unchanged (31) or reduced (32) with exercise, making monocytes an unlikely contributor to the increased concentrations. Circulating lipopolysaccharides (LPS) are reported to increase during exercise (2,3,15), which are known stimulators of TNF-α. In animals, Kupffer cells are a major source of circulating TNF-α after LPS challenge (16), making the liver a possible source of the running-specific increase in TNF-α. However, TNF-α concentrations are unchanged after a triathlon despite increased LPS (15), while even the presence of increased LPS concentrations during exercise in humans remains contentious (7).
Ultrastructural damage to muscle is associated with neutrophil infiltration and proinflammatory cytokine accumulation (10). Exercise-induced muscle damage might, therefore, explain the accumulation of TNF-α in skeletal muscle after running (19). Because neither cycling nor knee extensor exercise results in muscle cell disruption (4,35), such a mechanism might also explain unchanged muscle TNF-α concentrations (4,34) with these modes of activity. TNF-α is not released from contracting muscle during either cycling (9) or knee extensor exercise (34), but this has never been evaluated with running. Because our exercise protocol resulted in both increased circulating TNF-α and muscle cell disruption, as evidenced by increased CK at 1 and 2 d after exercise, we cannot exclude muscle as a possible source of the increased plasma TNF-α. In animals, acute exercise also increases TNF-α expression in lung (5) and adipose tissue (30), making both possible contributory sources to increases in circulating concentrations.
If TNF-α is released from the skeletal muscle during running as a result of muscle damage, higher-intensity exercise may produce more marked increases in TNF-α as a result of greater damage. However, despite a tendency for higher CK concentrations at FU1 and FU2 with increasing exercise intensity, indicative of greater muscle cell disruption, we observed no effect of exercise intensity on TNF-α concentrations. Our results do not rule out the possibility of an exercise intensity threshold above which TNF-α concentrations are increased, as one study reports increased (fourfold) TNF-α at 70% V˙O2max but not 50% V˙O2max with 30 min of cycling (18). The authors attribute this effect to greater damage to myocytes at 70% V˙O2max, although no measures of muscle cell damage were taken. This finding also differs from other studies that report no change in either TNF-α (31) or CK (4,31) with 60 min of cycling at 70% of peak V˙O2 (V˙O2peak).
The exercise intensity–dependent effect of running on IL-1ra concentrations both supports and extends findings from Peake et al. (25), who reported a greater increase in IL-1ra at 85% compared with 60% V˙O2max during 60 min of running. Our findings provide further information regarding the relationship between exercise intensity and changes in IL-1ra. Specifically, the findings point to a threshold for stimulation of IL-1ra between 55% and 65% V˙O2max and, as with IL-6, an increase in the sensitivity of IL-1ra response to increasing exercise intensity between 65% and 75% V˙O2max.
IL-6 can stimulate IL-1ra (33), and the IL-1ra response to exercise is both enhanced (25,26) and attenuated (12) concomitantly with the changes in IL-6. We observed significant positive correlations between changes in IL-6 and IL-1ra, which concur with previous findings (22,25). We also showed similar effects of exercise intensity on IL-6 and IL-1ra concentrations, with higher concentrations after exercise at 75% V˙O2max, and together, these findings support the theory that the increase in IL-6 during exercise mediates the subsequent increase in IL-1ra.
Although correlated, the magnitude of changes in IL-6 and IL-1ra was markedly different, with IL-6 increased 12-fold at 75% compared with only a 2-fold increase in IL-1ra. In addition, at 55% V˙O2max, there was no change in IL-1ra despite increased IL-6 and no correlation between the two. These discrepancies are, however, consistent with the effects of an IL-6 infusion (33) and exercise studies (22-24), which report larger increases in IL-6 compared with IL-1ra. Together, these findings might indicate a threshold above which circulating IL-6 concentrations stimulate IL-1ra production. Stimulation of IL-1ra after acute exercise may play a role in suppressing the long-term effects of elevated proinflammatory cytokines, which have been associated with poorer outcome in conditions such as cardiovascular disease and type 2 diabetes (28). A threshold for IL-1ra release could have important health implications, with the magnitude of the increase in IL-6 during exercise determining, at least in part, the subsequent anti-inflammatory response.
IL-6 can also increase cortisol (33), which both assists in maintaining blood glucose (6) and has anti-inflammatory properties (1). As with IL-6, we observed an exercise intensity–dependent effect on cortisol, with higher concentrations at 75% V˙O2max compared with 55% and 65% V˙O2max. The increased cortisol occurred after the increase in IL-6, suggesting a possible stimulatory effect of IL-6 on cortisol during exercise, meaning that the increase in IL-6 might partly explain the increase in cortisol. However, Fischer (11) suggests that such an effect might require plasma IL-6 concentrations of ≥50 pg·mL−1. Peak IL-6 concentrations in our study were only 7.4 pg·mL−1, which might indicate that other factors also contributed to the exercise-induced activation of the hypothalamic–pituitary–adrenal axis.
In conclusion, compared with 55% and 65% V˙O2max, 60 min of treadmill running at 75% V˙O2max results in a greater increase in circulating IL-6 and IL-1ra but not TNF-α concentrations. The higher RER at 75% V˙O2max supports the notion that the exercise intensity–dependent effect on IL-6 might be related to a higher rate of CHO utilization. A greater increase in IL-1ra at 75% V˙O2max and significant correlations between changes in IL-6 and IL-1ra suggest that changes in IL-1ra are mediated, in part, by the preceding changes in IL-6.
This work was funded by the Human Capability Domain of the UK Ministry of Defence Scientific Research Programme.
The authors would like to acknowledge Mrs. Anne Wright and Mr. Enhad Chowdhury for their assistance with the collection of data and all the participants, without whose considerable effort, the study would not have been possible. J.P.R. Scott acknowledges the Royal Commission for the Exhibition of 1851 for an Industrial Fellowship in support of his research.
The authors report no conflict of interest.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. Barnes PJ. Anti-inflammatory actions of glucocorticoids: molecular mechanisms. Clin Sci (Lond). 1998; 94 (6): 557–72.
2. Bosenberg AT, Brock-Utne JG, Gaffin SL, Wells MT, Blake GT. Strenuous exercise causes systemic endotoxemia. J Appl Physiol. 1998; 65 (1): 106–8.
3. Brock-Utne JG, Gaffin SL, Wells MT, et al.. Endotoxemia in exhausted runners after a long-distance race. S Afr Med J. 1998; 73 (9): 533–6.
4. Chan MH, Carey AL, Watt MJ, Febbraio MA. Cytokine gene expression in human skeletal muscle during concentric contraction: evidence that IL-8, like IL-6, is influenced by glycogen availability. Am J Physiol Regul Integr Comp Physiol. 2004; 287 (2): R322–7.
5. Colbert LH, Davis JM, Essig DA, Ghaffar A, Mayer EP. Tissue expression and plasma concentrations of TNFα, IL-1β, and IL-6 following treadmill exercise in mice. Int J Sports Med. 2001; 22 (4): 261–7.
6. Cryer PE. Glucose counterregulation: prevention and correction of hypoglycaemia in humans. Am J Physiol Endocrinol Metab. 1993; 264 (2 Pt 1): E149–55.
7. Davis JM, Nieman DC, Murphy AE. Unusually low plasma concentrations of lipopolysaccharide following 160-km race (reply). Brain Behav Immun. 2007; 21 (4): 515.
8. Febbraio MA, Hiscock N, Sacchetti M, Fischer CP, Pedersen BK. Interleukin-6 is a novel factor mediating glucose homeostasis during skeletal muscle contraction. Diabetes. 2004; 53 (7): 1643–8.
9. Febbraio MA, Steensberg A, Starkie RL, McConell GK, Kingwell BA. Skeletal muscle interleukin-6 and tumor necrosis factor-α release in healthy subjects and patients with type 2 diabetes at rest and during exercise. Metabolism. 2003; 52 (7): 939–44.
10. Fielding RA, Manfredi TJ, Ding W, Fiatarone MA, Evans WJ, Cannon JG. Acute phase response in exercise: III. Neutrophil and IL-1β accumulation in skeletal muscle. Am J Physiol. 1993; 265 (1 Pt 2): R166–72.
11. Fischer CP. Interleukin-6 in acute exercise and training: what is the biological relevance? Exerc Immunol Rev. 2006; 12: 6–33.
12. Fischer CP, Hiscock NJ, Penkowa M, et al.. Supplementation with vitamins C and E inhibits the release of interleukin-6 from contracting human skeletal muscle. J Physiol. 2004; 558 (Pt 2): 633–45.
13. Froelicher VF Jr, Brammell H, Davis G, Noguera I, Stewart A, Lancaster MC. A comparison of three maximal treadmill exercise protocols. J Appl Physiol. 1974; 36 (6): 720–5.
14. Helge JW, Stallknecht B, Pedersen BK, Galbo H, Kiens B, Richter EA. The effect of graded exercise on IL-6 release and glucose uptake in human skeletal muscle. J Physiol. 2003; 546 (Pt 1): 299–305.
15. Jeukendrup AE, Vet-Joop K, Sturk A, et al.. Relationship between gastro-intestinal complaints and endotoxaemia, cytokine release and the acute-phase reaction during and after a long-distance triathlon in highly trained men. Clin Sci (Lond). 2000; 98 (1): 47–55.
16. Jiang Q, DeTolla L, Kalvakolanu I, et al.. Febrile range temperature modifies early systemic TNFa expression in bacterial endotoxin-challenged mice. Infect Immun. 1999; 67 (4): 1539–46.
17. Keller C, Steensberg A, Pilegaard H, et al.. Transcriptional activation of the IL-6 gene in human contracting skeletal muscle: influence of muscle glycogen content. FASEB J. 2001; 15 (14): 2748–50.
18. Kimura H, Suzui M, Nagao F, Matsumoto K. Highly sensitive determination of plasma cytokines by time-resolved fluoroimmunoassay; effect of bicycle exercise on plasma level of interleukin-1α (IL-1α), tumor necrosis factor α (TNFα), and interferon γ (IFNγ). Anal Sci. 2001; 17 (5): 593–7.
19. Nieman DC, Davis JM, Henson DA, et al.. Carbohydrate ingestion influences skeletal muscle cytokine mRNA and plasma cytokine levels after a 3-h run. J Appl Physiol. 2003; 94 (5): 1917–25.
20. Nieman DC, Henson DA, Smith LL, et al.. Cytokine changes after a marathon race. J Appl Physiol. 2001; 91 (1): 109–14.
21. Noble BJ, Borg GA, Jacobs I, Ceci R, Kaiser P. A category–ratio perceived exertion scale: relationship to blood and muscle lactates and heart rate. Med Sci Sports Exerc. 1983; 15 (6): 523–8.
22. Ostrowski K, Hermann C, Bangash A, Schjerling P, Nielsen JN, Pedersen BK. A trauma-like elevation of plasma cytokines in humans in response to treadmill running. J Physiol. 1998; 513 (Pt 3): 889–94.
23. Ostrowski K, Rohde T, Asp S, Schjerling P, Pedersen BK. Pro- and anti-inflammatory cytokine balance in strenuous exercise in humans. J Physiol. 1999; 515 (Pt 1): 287–91.
24. Ostrowski K, Rohde T, Zacho M, Asp S, Pedersen BK. Evidence that interleukin-6 is produced in human skeletal muscle during prolonged running. J Physiol. 1998; 508 (Pt 3): 949–53.
25. Peake JM, Suzuki K, Hordern M, Wilson G, Nosaka K, Coombes JS. Plasma cytokine changes in relation to exercise intensity and muscle damage. Eur J Appl Physiol. 2005; 95 (5–6): 514–21.
26. Peake JM, Wilson G, Hordern M, et al.. Changes in neutrophil receptor expression, degranulation and respiratory burst activity after moderate and high intensity exercise. J Appl Physiol. 2004; 97 (2): 612–8.
27. Petersen EW, Carey AL, Sacchetti M, et al.. Acute IL-6 treatment increases fatty acid turnover in elderly humans in vivo
and in tissue culture in vitro
. Am J Physiol Endocrinol Metab. 2005; 288 (1): E155–62.
28. Petersen AM, Pedersen BK. The anti-inflammatory effect of exercise. J Appl Physiol. 2005; 98 (4): 1154–62.
29. Romijn JA, Coyle EF, Sidossis LS, et al.. Regulation of endogenous fat and carbohydrate metabolism in relation to exercise intensity and duration. Am J Physiol. 1993; 265 (3 Pt 1): E380–91.
30. Rosa Neto JC, Lira FS, Oyama LM, et al.. Exhaustive exercise causes an anti-inflammatory effect in skeletal muscle and a pro-inflammatory effect in adipose tissue in rats. Eur J Appl Physiol. 2009; 106 (5): 697–704.
31. Starkie RL, Hargreaves M, Rolland J, Febbraio MA. Heat stress, cytokines, and the immune response to exercise. Brain Behav Immun. 2005; 19 (5): 404–12.
32. Starkie RL, Rolland J, Angus DJ, Anderson MJ, Febbraio MA. Circulating monocytes are not the source of elevations in plasma IL-6 and TNF-α levels after prolonged running. Am J Physiol Cell Physiol. 2001; 280 (4): C769–74.
33. Steensberg A, Fischer CP, Keller C, Moller K, Pedersen BK. IL-6 enhances IL-1ra, IL-10, and cortisol in humans. Am J Physiol. 2003; 285 (2): E433–7.
34. Steensberg A, Keller C, Starkie RL, Osada T, Febbraio MA, Pedersen BK. IL-6 and TNF-α expression in, and release from, contracting human skeletal muscle. Am J Physiol. 2002; 283 (6): E1272–8.
35. Steensberg A, van Hall G, Osada T, Sacchetti M, Saltin B, Klarlund-Pedersen B. Production of interleukin-6 in contracting human skeletal muscles can account for the exercise-induced increase in plasma interleukin-6. J Physiol. 2000; 529 (Pt 1): 237–42.
36. Stouthard JM, Romijn JA, Van Der PT, et al.. Endocrinologic and metabolic effects of interleukin-6 in humans. Am J Physiol. 1995; 268 (5 Pt 1): E813–9.
37. Toft AD, Thorn M, Ostrowski K, et al.. N-3 polyunsaturated fatty acids do not affect cytokine response to strenuous exercise. J Appl Physiol. 2000; 89 (6): 2401–6.
38. Tsigos C, Papanicolaou DA, Defensor R, Mitsiadis CS, Kyrou I, Chrousos GP. Dose effects of recombinant human interleukin-6 on pituitary hormone secretion and energy expenditure. Neuroendocrinology. 1997; 66 (1): 54–62.
39. Wahren J. Human forearm muscle metabolism during exercise. IV. Glucose uptake at different work intensities. Scand J Clin Lab Invest. 1970; 25 (2): 129–35.
40. Wolsk E, Mygind H, Grøndahl TS, Pedersen BK, van Hall G. IL-6 selectively stimulates fat metabolism in human skeletal muscle. Am J Physiol Endocrinol Metab. 2010; 299 (5): E832–40.