To establish the role of hepatic nerves in hepatic glycogenolytic and gluconeogenic regulation during exercise, dogs underwent a laparotomy during which the hepatic nerves were either left intact (C; n = 8) or cut (DN; n = 5). At least 17 days after surgery, dogs were studied during 150 min of treadmill exercise (12% grade, 100 m/min). Glucose production (Ra) and gluconeogenesis (GNG) were assessed by combining [3-3H]glucose, [U-14C]alanine, and indocyanine green infusions with arterial, portal vein, and hepatic vein sampling. Glucagon and insulin were similar at rest and exercise in both groups. Norepinephrine rose from 145 +/- 10 to 242 +/- 32 pg/ml by 150 min of exercise in C and from 150 +/- 25 to 333 +/- 83 pg/ml in DN. Epinephrine rose from 66 +/- 7 pg/ml at rest to 108 +/- 10 and 148 +/- 24 pg/ml after 30 and 150 min of exercise in C and from 90 +/- 15 pg/ml at rest to 185 +/- 33 (P less than 0.05 compared with C) and 194 +/- 36 pg/ml after 30 and 150 min of exercise in DN. Plasma glucose fell gradually from 108 +/- 2 and 106 +/- 3 mg/dl at rest to 96 +/- 4 and 92 +/- 8 by the end of exercise in C and DN, respectively. Ra was similar in C and DN rising from 3.2 +/- 0.2 to 8.7 +/- 0.6 and 2.6 +/- 0.2 to 7.5 +/- 1.1 mg.kg-1.min-1, respectively, by the end of exercise. Minimum and maximum rates of GNG from alanine, glycerol, and lactate were elevated in DN compared with C during rest and exercise. However, the exercise-induced changes in GNG were similar in both groups. In conclusion, nerves to the liver are not essential to the increased Ra and glucose homeostasis during moderate-intensity exercise.