Stressful exercise has been found to reduce the proinflammatory cytokine response to lipopolysaccharide (LPS). In this study, we aimed to determine whether receptor antagonists for corticosterone or catecholamines would increase the LPS-induced tumor necrosis factor-α (TNF-α) response after exhaustive exercise. Female F344 rats were randomly assigned to one of five groups: control (vehicle), RU-486 (glucocorticoid receptor antagonist, 30 mg/kg)-, propranolol [nonselective β-adrenergic receptor (AR) blockade, 30 mg/kg]-, atenolol (β1-AR blockade, 30 mg/kg)-, or ICI 118551 (β2-AR blockade, 30 mg/kg)-treated groups. Each antagonist was given intraperitoneally 30 min prior to exercise or control period. Exercised rats ran until exhaustion on a treadmill at gradually increasing speeds, from 10 to 36 m/min at 15% grade. Immediately postexercise or control period all rats were injected with LPS (1 mg/kg, i.v.). Plasma TNF-α was reduced by prior exercise to approximately 10% of that of sedentary controls (p < 0.01). Plasma TNF-α concentration in exercised RU-486-treated rats was significantly different than that of nonexercised rats (19.2%, p < 0.01) and not different from exercised rats. However, pretreatment of rats with the nonselective β-AR blocker propranolol almost completely reversed the exercise-induced suppression of plasma TNF-α in response to LPS. β1-AR pretreatment almost completely attenuated the exercise-induced suppression of LPS-induced plasma TNF-α while β2-antagonism had a partial effect. These results indicate that exercise-induced catecholamines, acting through β-ARs (especially the β1-AR), are responsible for the exercise-induced suppression of plasma TNF-α after LPS administration.