During "nondamaging" exercise, skeletal muscle markedly releases interleukin (IL)-6, and it has been suggested that one biological role of this phenomenon is to inhibit the production of tumor necrosis factor (TNF)- alpha, which is known to cause pathogenesis such as insulin resistance and atherosclerosis. To test this hypothesis, we performed three experiments in which eight healthy males either rested (CON), rode a bicycle for 3 h (EX), or were infused with recombinant human IL-6 (rhIL-6) for 3 h while they rested. After 2.5 h, the volunteers received a bolus of Escherichia coli lipopolysaccharide endotoxin (0.06 ng/kg) i.v. to induce low-grade inflammation. In CON, plasma TNF-alpha increased significantly in response to endotoxin. In contrast, during EX, which resulted in elevated IL-6, and rhIL-6, the endotoxin-induced increase in TNF-alpha was totally attenuated. In conclusion, physical exercise and rhIL-6 infusion at physiological concentrations inhibit endotoxin-induced TNF-alpha production in humans. Hence, these data provide the first experimental evidence that physical activity mediates antiinflammatory activity and suggest that the mechanism include IL-6, which is produced by and released from exercising muscles.

Prolonged exercise results in a progressive decline in glycogen content and a concomitant increase in the release of the cytokine interleukin‐6 (IL‐6) from contracting muscle. This study tests the hypothesis that the exercise‐induced IL‐6 release from contracting muscle is linked to the intramuscular glycogen availability. Seven men performed 5 h of a two‐legged knee‐extensor exercise, with one leg with normal, and one leg with reduced, muscle glycogen content. Muscle biopsies were obtained before (pre‐ex), immediately after (end‐ex) and 3 h into recovery (3 h rec) from exercise in both legs. In addition, catheters were placed in one femoral artery and both femoral veins and blood was sampled from these catheters prior to exercise and at 1 h intervals during exercise and into recovery. Pre‐exercise glycogen content was lower in the glycogen‐depleted leg compared with the control leg. Intramuscular IL‐6 mRNA levels increased with exercise in both legs, but this increase was augmented in the leg having the lowest glycogen content at end‐ex. The arterial plasma concentration of IL‐6 increased from 0.6 ± 0.1 ng l−1 pre‐ex to 21.7 ± 5.6 ng l−1 end‐ex. The depleted leg had already released IL‐6 after 1 h (4.38 ± 2.80 ng min−1 (P < 0.05)), whereas no significant release was observed in the control leg (0.36 ± 0.14 ng min−1). A significant net IL‐6 release was not observed until 2 h in the control leg. This study demonstrates that glycogen availability is associated with alterations in the rate of IL‐6 production and release in contracting skeletal muscle.