D. B. Lacy

Similar Dose Responsiveness of Hepatic Glycogenolysis and Gluconeogenesis to Glucagon In Vivo

This study was undertaken to determine whether the dose-dependent effect of glucagon on gluconeogenesis parallels its effect on hepatic glycogenolysis in conscious overnight-fasted dogs. Endogenous insulin and glucagon secretion were inhibited by somatostatin (0.8 μg · kg−1 · min−1), and intraportal replacement infusions of insulin (213 ± 28 (αU kg−1 · min−1) and glucagon (0.65 ng · kg−1 · min−1) were given to maintain basal hormone concentrations for 2 h (12 ± 2 and 108 ± 23 pg/ml, respectively). The glucagon infusion was then increased 2-, 4-, 8-, or 12-fold for 3 h, whereas the rate of insulin infusion was left unchanged. Glucose production (GP) was determined with 3-[3H]glucose, and gluconeogenesis (GNG) was assessed with tracer (U-[14C]alanine conversion to [14C]glucose) and arteriovenous difference (hepatic fractional extraction of alanine, FEA) techniques. Increases in plasma glucagon of 53 ± 8, 199 ± 48, 402 ± 28, and 697 ±149 pg/ml resulted in initial (15- 30 min) increases in GP of 1.1 ± 0.4 (N = 4), 4.9 ± 0.5 (N = 4), 6.5 ± 0.6 (N = 6), and 7.7 ± 1.4 (N = 4) mg kg−1 · min−1, respectively; increases in GNG (∼3 h) of 48 ± 19, 151 ± 50, 161 ± 25, and 157 ± 7%, respectively; and increases in FEA (3 h) of 0.14 ± 0.07, 0.37 ± 0.05, 0.42 ± 0.04, and 0.40 ± 0.17, respectively. In conclusion, GNG and glycogenolysis were similarly sensitive to stimulation by glucagon in vivo, and the dose-response curves were markedly parallel.

Role of Gluconeogenesis in Sustaining Glucose Production During Hypoglycemia Caused by Continuous Insulin Infusion in Conscious Dogs

The roles of glycogenolysis and gluconeogenesis in sustaining glucose production during insulin-induced hypoglycemia were assessed in overnight-fasted conscious dogs. Insulin was infused intraportally for 3 h at 5 mU · kg−1 · min−1 in five animals, and glycogenolysis and gluconeogenesis were measured by using a combination of tracer [(3-3H]glucose and [U-14C]alanine) and hepatic arteriovenous difference techniques. In response to the elevated insulin level (263 ± 39 μU/ml), plasma glucose level fell (41 ± 3 mg/dl), and levels of the counterregulatory hormones glucagon, epinephrine, norepinephrine, and cortisol increased (91 ± 29 to 271 ± 55 pg/ml, 83 ± 26 to 2356 ± 632 pg/ml, 128 ± 31 to 596 ± 81 pg/ml, and 1.5 ± 0.4 to 11.1 ± 1.0 μg/dl, respectively; for all, P < .05). Glucose production fell initially and then doubled (3.1 ± 0.3 to 6.1 ± 0.5 mg · kg−1 · min−1; P < .05) by 60 min. Net hepatic gluconeogenic precursor uptake increased ∼ eightfold by the end of the hypoglycemic period. By the same time, the efficiency with which the liver converted the gluconeogenic precursors to glucose rose twofold. Five control experiments in which euglycemia was maintained by glucose infusion during insulin administration (5.0 mU · kg−1 · min−1) provided baseline data. Glycogenolysis accounted for 69–88% of glucose production during the 1st h of hypoglycemia, whereas gluconeogenesis accounted for 48–88% of glucose production during the 3rd h of hypoglycemia. These data suggest that gluconeogenesis is the key process for the normal counterregulatory response to prolonged and marked hypoglycemia.

Effect of prior exercise on the partitioning of an intestinal glucose load between splanchnic bed and skeletal muscle.

Exercise leads to marked increases in muscle insulin sensitivity and glucose effectiveness. Oral glucose tolerance immediately after exercise is generally not improved. The hypothesis tested by these experiments is that after exercise the increased muscle glucose uptake during an intestinal glucose load is counterbalanced by an increase in the efficiency with which glucose enters the circulation and that this occurs due to an increase in intestinal glucose absorption or decrease in hepatic glucose disposal. For this purpose, sampling (artery and portal, hepatic, and femoral veins) and infusion (vena cava, duodenum) catheters and Doppler flow probes (portal vein, hepatic artery, external iliac artery) were implanted 17 d before study. Overnightfasted dogs were studied after 150 min of moderate treadmill exercise or an equal duration rest period. Glucose ([14C]glucose labeled) was infused in the duodenum at 8 mg/kg x min for 150 min beginning 30 min after exercise or rest periods. Values, depending on the specific variable, are the mean +/- SE for six to eight dogs. Measurements are from the last 60 min of the intraduodenal glucose infusion. In response to intraduodenal glucose, arterial plasma glucose rose more in exercised (103 +/- 4 to 154 +/- 6 mg/dl) compared with rested (104 +/- 2 to 139 +/- 3 mg/dl) dogs. The greater increase in glucose occurred even though net limb glucose uptake was elevated after exercise (35 +/- 5 vs. 20 +/- 2 mg/min) as net splanchnic glucose output (5.1 +/- 0.8 vs. 2.1 +/- 0.6 mg/kg x min) and systemic appearance of intraduodenal glucose (8.1 +/- 0.6 vs. 6.3 +/- 0.7 mg/kg x min) were also increased due to a higher net gut glucose output (6.1 +/- 0.7 vs. 3.6 +/- 0.9 mg/kg x min). Adaptations at the muscle led to increased net glycogen deposition after exercise [1.4 +/- 0.3 vs. 0.5 +/- 0.1 mg/(gram of tissue x 150 min)], while no such increase in glycogen storage was seen in liver [3.9 +/- 1.0 vs. 4.1 +/- 1.1 mg/(gram of tissue x 150 min) in exercised and sedentary animals, respectively]. These experiments show that the increase in the ability of previously working muscle to store glycogen is not solely a result of changes at the muscle itself, but is also a result of changes in the splanchnic bed that increase the efficiency with which oral glucose is made available in the systemic circulation.

Exercise-Induced Fall in Insulin and Increase in Fat Metabolism During Prolonged Muscular Work

The role of the exercise-induced fall in insulin in fat metabolism was studied in dogs during 150 min of treadmill exercise alone (controls) or with insulin clamped at basal levels by an intraportal infusion to prevent the normal fall in insulin concentration (ICs). To counteract the suppressive effect of insulin on glucagon release, glucagon was supplemented by an intraportal infusion in ICs. In all dogs, catheters were placed in a carotid artery and in the portal and hepatic veins for sampling and in the vena cava and the splenic vein for infusion purposes. Glucose levels were clamped in ICs to recreate the glycemic response evident in controls. In controls, insulin fell by 7 ± 1 μxU/ml but was unchanged from basal levels in ICs (0 ± 2 μU/ml). Glucagon, norepinephrine, epinephrine, and cortisol rose similarly in controls and ICs. Arterial free-fatty acid (FFA) levels rose by 644 ± 126 μeq/L in controls but did not increase in ICs (−12 ± 148 μeq/L). Arterial glycerol levels rose by 337 ± 43 and 183 ± 19 (μM in controls and ICs. Hepatic FFA delivery and fractional extraction increased by 17 ± 3 and 0.06 ± 0.02 μmol · kg−1 · min−1, respectively, in controls. In ICs, hepatic FFA delivery increased by only 1 ± 2 (μmol · kg−1 · min−1, whereas hepatic fractional extraction fell slightly (−0.03 ± 0.03). Consequently, net hepatic FFA uptake rose by 4.8 ± 1.5 μxmol · kg−1 · min−1 in controls but decreased slightly in ICs (−0.5 ± 1.1 μmol · kg−1 · min−1). At least partly because of the differences in hepatic FFA uptake, arterial β-hydroxybutyrate (50 ± 14 vs. 23 ± 8 μM) and net hepatic β-hydroxybutyrate output (2.2 ± 0.7 vs. 0.4 ± 0.3 μmol · kg−1 · min−1) rose more in controls than in ICs. In summary, the exercise-induced fall in insulin 7) is essential to the increase in FFA levels during prolonged muscular work, 2) facilitates hepatic FFA uptake by enhancing the delivery of FFAs to the liver and their extraction by the liver, and 3) enhances the net β-hydroxybutyrate output by the liver, at least in part through the effects described in 1 and 2. Hence, the exercise-induced fall in insulin is essential for the transition to the increased rate of FFA metabolism that occurs as work duration progresses.

Effects of morphine on glucose homeostasis in the conscious dog.

This study was designed to assess the effects of morphine sulfate on glucose kinetics and on glucoregulatory hormones in conscious overnight fasted dogs. One group of experiments established a dose-response range. We studied the mechanisms of morphine-induced hyperglycemia in a second group. We also examined the effect of low dose morphine on glucose kinetics independent of changes in the endocrine pancreas by the use of somatostatin plus intraportal replacement of basal insulin and glucagon. In the dose-response group, morphine at 2 mg/h did not change plasma glucose, while morphine at 8 and 16 mg/h caused a hyperglycemic response. In the second group of experiments, morphine (16 mg/h) caused an increase in plasma glucose from a basal 99 +/- 3 to 154 +/- 13 mg/dl (P less than 0.05). Glucose production peaked at 3.9 +/- 0.7 vs. 2.5 +/- 0.2 mg/kg per min basally, while glucose clearance declined to 1.7 +/- 0.2 from 2.5 +/- 0.1 ml/kg per min (both P less than 0.05). Morphine increased epinephrine (1400 +/- 300 vs. 62 +/- 8 pg/ml), norepinephrine (335 +/- 66 vs. 113 +/- 10 pg/ml), glucagon (242 +/- 53 vs. 74 +/- 14 pg/ml), insulin (30 +/- 9 vs. 10 +/- 2 microU/ml), cortisol (11.1 +/- 3.3 vs. 0.9 +/- 0.2 micrograms/dl), and plasma beta-endorphin (88 +/- 27 vs. 23 +/- 6 pg/ml); all values P less than 0.05 compared with basal. These results show that morphine-induced hyperglycemia results from both stimulation of glucose production as well as inhibition of glucose clearance. These changes can be explained by rises in epinephrine, glucagon, and cortisol. These in turn are part of a widespread catabolic response initiated by high dose morphine that involves activation of the sympathetic nervous system, the endocrine pancreas, and the pituitary-adrenal axis. Also, we report the effect of a 2 mg/h infusion of morphine on glucose kinetics when the endocrine pancreas is clamped at basal levels. Under these conditions, morphine exerts a hypoglycemic effect (25% fall in plasma glucose, P less than 0.05) that is due to inhibition of glucose production (by 25-43%, P less than 0.05). The hypoglycemia was independent of detectable changes in insulin, glucagon, epinephrine and cortisol, and was not reversed by concurrent infusion of a slight molar excess of naloxone. Therefore, we postulate that the hypoglycemic effect of morphine results from the interaction of the opiate with non-mu receptors either in the liver or the central nervous system.

Regulation of Glucose Uptake and Metabolism by Working Muscle: An In Vivo Analysis

To assess the mechanisms whereby muscular work stimulates glucose uptake and metabolism in vivo, dogs were studied during rest (–40–0 min), moderate exercise (0–90 min), and exercise recovery (90–180 min) with plasma glucose clamped at 5.0, 6.7, 8.3, and 10.0 mM (n = 5 at 5.0 mM and n = 4 at all other levels) using a variable glucose infusion. Basal insulin was maintained with somatostatin and insulin replacement. Whole-body glucose uptake, limb glucose uptake, and oxidative and nonoxidative glucose plus lactate metabolism, were assessed with tracers ([3H]glucose and [14C]glucose) and arteriovenous differences. The combined effects of glucose and exercise on the increment above resting values for limb glucose uptake, arteriovenous glucose difference, LGO, LGNO, and rate of glucose disappearance were synergistic (∼ 112, 90, 125, 76, and 90% > the additive values, respectively). Neither exercise nor recovery affected the Km for limb glucose uptake (4.7 ± 1.1, 4.8 ± 0.4, and 5.2 ± 0.3 mM during rest, exercise, and recovery, respectively), but both conditions increased the Vmax (44 ± 16, 217 ± 30, and 118 ± 14 mumol/min during rest, exercise, and recovery, respectively). Similarly, the Km for arteriovenous glucose differences were unaffected by exercise recovery (4.9 ± 0.6, 5.0 ± 0.4, and 5.3 ± 0.3 mM during rest, exercise, and recovery, respectively), but the maximum rose (272 ± 50, 650 ± 78, and 822 ± 111 microM during rest, exercise, and recovery, respectively). The LGO was unchanged by glycemia at rest (15 ± 4 mumol/min at 10.0 mM). The Km for LGO during exercise was 5.1 ± 0.3 mM, and the Vmax was 163 ± 15. The capacity for LGO returned to basal during recovery. LGNO increased gradually with increasing glycemia during rest, exercise, and recovery and did not approach saturation (38 ± 13, 105 ± 36, and 132 ± 45 mumol/min during rest, exercise, and recovery, respectively, at 10.0 mM). In general, the LGNO was elevated at every glucose level during exercise (∼ twofold) and recovery (∼ threefold) compared with rest. Arterial free fatty acid and glycerol levels decreased with increasing glycemia within all periods. Free fatty acids were suppressed by a greater amount during exercise compared with rest and recovery. This study shows that 1) the combined effects of exercise and increased glucose level act synergistically on glucose uptake and metabolism; 2) exercise increases the Vmax for limb glucose uptake and arteriovenous glucose difference without altering the Km for these variables; 3) the capacity for LGNO predominates at rest, whereas the capacity for LGO predominates during exercise; 4) during recovery the capacity for LGO returned to basal, whereas that for LGNO remained elevated; and 5) glucose-induced suppression of free fatty acid levels was greatest during exercise. In conclusion, an increase in circulating glucose within the physiological range, which has only minor effects at rest, profoundly increases muscle glucose metabolism and decreases free fatty acid availability during exercise.

Prior exercise increases net hepatic glucose uptake during a glucose load

The aim of these studies was to determine whether prior exercise enhances net hepatic glucose uptake (NHGU) during a glucose load. Sampling catheters (carotid artery, portal, hepatic, and iliac veins), infusion catheters (portal vein and vena cava), and Doppler flow probes (portal vein, hepatic and iliac arteries) were implanted. Exercise (150 min; n = 6) or rest (n = 6) was followed by a 30-min control period and a 100-min experimental period (3.5 mg. kg-1. min-1 of glucose in portal vein and as needed in vena cava to clamp arterial blood glucose at approximately 130 mg/dl). Somatostatin was infused, and insulin and glucagon were replaced intraportally at fourfold basal and basal rates, respectively. During experimental period the arterial-portal venous (a-pv) glucose gradient (mg/dl) was -18 +/- 1 in sedentary and -19 +/- 1 in exercised dogs. Arterial insulin and glucagon were similar in the two groups. Net hepatic glucose balance (mg. kg-1. min-1) shifted from 1.9 +/- 0.2 in control period to -1.8 +/- 0.2 (negative rates represent net uptake) during experimental period in sedentary dogs (Delta3.7 +/- 0.5); with prior exercise it shifted from 4.1 +/- 0.3 (P < 0.01 vs. sedentary) in control period to -3.2 +/- 0.4 (P < 0.05 vs. sedentary) during experimental period (Delta7.3 +/- 0.7, P < 0.01 vs. sedentary). Net hindlimb glucose uptake (mg/min) was 4 +/- 1 in sedentary animals in control period and 13 +/- 2 during experimental period; in exercised animals it was 7 +/- 1 in control period (P < 0. 01 vs. sedentary) and 32 +/- 4 (P < 0.01 vs. sedentary) during experimental period. As the total glucose infusion rate (mg. kg-1. min-1) was 7 +/- 1 in sedentary and 11 +/- 1 in exercised dogs, approximately 30% of the added glucose infusion due to prior exercise could be accounted for by the greater NHGU. In conclusion, when determinants of hepatic glucose uptake (insulin, glucagon, a-pv glucose gradient, glycemia) are controlled, prior exercise increases NHGU during a glucose load due to an effect that is intrinsic to the liver. Increased glucose disposal in the postexercise state is therefore due to an improved ability of both liver and muscle to take up glucose.The aim of these studies was to determine whether prior exercise enhances net hepatic glucose uptake (NHGU) during a glucose load. Sampling catheters (carotid artery, portal, hepatic, and iliac veins), infusion catheters (portal vein and vena cava), and Doppler flow probes (portal vein, hepatic and iliac arteries) were implanted. Exercise (150 min; n = 6) or rest ( n = 6) was followed by a 30-min control period and a 100-min experimental period (3.5 mg ⋅ kg-1 ⋅ min-1of glucose in portal vein and as needed in vena cava to clamp arterial blood glucose at ∼130 mg/dl). Somatostatin was infused, and insulin and glucagon were replaced intraportally at fourfold basal and basal rates, respectively. During experimental period the arterial-portal venous (a-pv) glucose gradient (mg/dl) was -18 ± 1 in sedentary and -19 ± 1 in exercised dogs. Arterial insulin and glucagon were similar in the two groups. Net hepatic glucose balance (mg ⋅ kg-1 ⋅ min-1) shifted from 1.9 ± 0.2 in control period to -1.8 ± 0.2 (negative rates represent net uptake) during experimental period in sedentary dogs (Δ3.7 ± 0.5); with prior exercise it shifted from 4.1 ± 0.3 ( P < 0.01 vs. sedentary) in control period to -3.2 ± 0.4 ( P < 0.05 vs. sedentary) during experimental period (Δ7.3 ± 0.7, P < 0.01 vs. sedentary). Net hindlimb glucose uptake (mg/min) was 4 ± 1 in sedentary animals in control period and 13 ± 2 during experimental period; in exercised animals it was 7 ± 1 in control period ( P < 0.01 vs. sedentary) and 32 ± 4 ( P < 0.01 vs. sedentary) during experimental period. As the total glucose infusion rate (mg ⋅ kg-1 ⋅ min-1) was 7 ± 1 in sedentary and 11 ± 1 in exercised dogs, ∼30% of the added glucose infusion due to prior exercise could be accounted for by the greater NHGU. In conclusion, when determinants of hepatic glucose uptake (insulin, glucagon, a-pv glucose gradient, glycemia) are controlled, prior exercise increases NHGU during a glucose load due to an effect that is intrinsic to the liver. Increased glucose disposal in the postexercise state is therefore due to an improved ability of both liver and muscle to take up glucose.

Stimulation of glucose production through hormone secretion and other mechanisms during insulin-induced hypoglycemia

To assess the role of counterregulatory hormones per se in the response to continuous insulin infusion, overnight-fasted dogs were given 5 mil · kg−1 · min−1 insulin intraportally either alone (INS, n = 5), with glucose to maintain euglycemia (INS + GLU, n = 5), or with glucose and hormone replacement [i.e., glucagon, epinephrine, norepinephrine, and cortisol infusions (INS + GLU + HR, n = 6)]. The increases in counterregulatory hormones that occurred during insulin-induced hypoglycemia were simulated in the latter group. In this way, it was possible to separate the effects of hypoglycemia per se from those due to the associated counterregulatory hormone response. Glycogenolysis and gluconeogenesis were measured with a combination of tracer ([3-3H]glucose and [U-14C]alanine) and hepatic arteriovenous (AV) difference techniques during a 40-min control and a 180-min experimental period. Insulin levels increased similarly in all groups (to ≃250 μU/ml), whereas plasma glucose levels decreased in INS (115 ± 3 to 41 ± 3 mg/dl; P < .05) and rose slightly in both INS + GLU (108 ± 2 to 115 ± 4 mg/dl; P < .05) and INS + GLU + HR (111 ± 3 to 120 ± 3 mg/dl; P < .05) due to glucose infusion. Glucagon, epinephrine, norepinephrine, and cortisol were replaced in INS + GLU + HR so that the increments in their levels were 102 ± 6, 106 ± 14, 117 ± 9, and 124 ± 37%, respectively, of their increments in INS. At no time was there a significant difference between the hormone levels in INS and INS + GLU + HR. The rise in the counterregulatory hormones per se accounted for only half (53 ± 9% by the AV difference method and 54 ± 10% by tracer method) of the glucose production associated with hypoglycemia resulting from insulin infusion. The rate and efficiency of alanine conversioto glucose in the hormone-replacement studies were only 29 ± 10 and 50 ± 27% of what occurred during hypoglycemia induced by insulin infusion. In conclusion, the counterregulatory hormones alone (i.e., without accompanying hypoglycemia) can account for only 50% of the glucose production that is present during insulin-induced hypoglycemia. The remaining 50%, therefore, must result from effects of hypoglycemia other than its ability to trigger hormone release.

Effects of an Acute Increase in Epinephrine and Cortisol on Carbohydrate Metabolism During Insulin Deficiency

This study was undertaken to investigate the effects of an acute increase in the plasma epinephrine level, with or without an accompanying increase in the plasma cortisol level, during selective insulin deficiency on glycogenolysis and gluconeogenesis in conscious overnight-fasted dogs. Experiments consisted of an 80-min tracer and dye equilibration period, a 40-min basal period, and a 180-min experimental period. In all protocols, selective insulin deficiency was created during the experimental period by infusing somatostatin peripherally (0.8 μg · kg−1 · min−1) with basal replacement of glucagon intraportally (0.65 ng · kg−1 · min−1). In EPI+SAL (n = 6), an additional infusion of epinephrine (0.04 μg · kg−1 · min−1) was infused during the experimental period along with saline. In EPI+CORT (n = 6), hydrocortisone (3.0 μg · kg−1 · min−1) was infused in addition to epinephrine during the experimental period. In SAL+CORT (n = 5), hydrocortisone was infused during the experimental period. In SALINE (n = 5), neither epinephrine nor cortisol was infused. [3-3H]glucose, [U-14C]alanine, and indocyanine green dye were used to assess glucose production (rate of appearance [Ra]) and gluconeogenesis using tracer and arteriovenous difference techniques. During selective insulin deficiency in SALINE, the arterial plasma glucose level increased from 6.0 ± 0.1 to 15.8 ± 1.1 mmol/l; Ra increased from 14.7 ± 0.7 to 24.9 ± 1.7 μmol· kg−1 · min−1. Gluconeogenic efficiency and the conversion of alanine and lactate to glucose increased to 300 ± 55 and 355 ± 67% of basal. In EPI+SAL and EPI+CORT, plasma glucose increased from 6.2 ± 0.1 to 19.8 ± 0.9 mmol/l and from 6.3 ± 0.1 to 19.5 ± 0.9 mmol/l. In EPI+SAL and EPI+CORT, Ra increased from 16.5 ± 1.1 to 29.3 ± 3.2 μmol · kg−1 · min−1 and from 15.4 ± 1.3 to 28.3 ± 2.5 μmol.kg−1 · min−1. The rise in gluconeogenic efficiency was similar to the rise that occurred in SALINE, but gluconeogenic conversion increased 17-fold in each of the two epinephrine groups. During the epinephrine infusion, gluconeogenesis accounted for a maximum of 55% of total glucose production as opposed to 31% during insulin deficiency alone. An increase in cortisol alone during insulin deficiency (SAL+CORT) had no effect on glucose level, glucose production, or gluconeogenesis. These results suggest that small increases in the plasma epinephrine level during insulin deficiency can significantly worsen the resulting hyperglycemia through stimulation of both glycogenolysis and gluconeogenesis. The addition of an acute rise in plasma cortisol does not affect glycogenolysis or gluconeogenesis or further worsen the hyperglycemia during insulin deficiency.

Exercise-Induced Rise in Glucagon and Ketogenesis During Prolonged Muscular Work

These experiments examined the role of the exercise-induced increment in glucagon in the control of ketogenesis during prolonged moderate-intensity (100 m/min, 12% grade) treadmill exercise. Dogs were studied during 150 min of exercise with saline infusion alone (C; n = 6) with the glucagon levels clamped at basal values (somatostatin infusion with basal glucagon replacement and the normal fall in insulin simulated; BG; n = 5) or with the normal exercise-induced rise in glucagon simulated (somatostatin infusion with the rise in glucagon and the fall in insulin simulated; SG; n = 5). Glucose was infused as needed in SG and BG to maintain the glycemic response seen in C. In all dogs, catheters were inserted into the carotid artery and the portal and hepatic veins for blood sampling and the vena cava and the splenic vein for infusions. Glucagon rose from 62 ± 5 and 57 ± 4 pg/ml at rest to 104 ± 20 and 120 ± 12 pg/ml during exercise in C and SG but did not deviate from basal in BG (56 ± 3 pg/ml). Insulin fell similarly from rest to the end of exercise in C (13 ± 2 to 5 ± 1 μU/ml), SG (11 ± 1 to 6 ± 1 μU/ml), and BG (10 ± 1 to 6 ± 1 μU/ml). In C, SG, and BG, free-fatty acid (FFA) levels rose from 941 ± 81,1240 ± 155, and 938 ± 36 μeq/L at rest to 1615 ± 149, 1558 ± 175, and 1391 ± 160 μeq/L with exercise. In C, SG, and BG, net hepatic FFA uptake was 2.9 ± 0.4, 4.6 ± 0.6, and 3.6 ± 0.3 μmol · kg−1 · min−1 at rest and rose to 7.8 ± 1.8, 7.2 ± 0.8, and 7.1 ± 1.3 μmol · kg−1 · min−1 with exercise. Arterial blood β-hydroxybutyrate levels rose from 20 ± 3 to 69 ± 25 μM and from 21 ± 4 to 66 ± 16 μM by the end of exercise in C and SG but from only 16 ± 2 to 27 ± 5 μM in BG. Net hepatic β-hydroxybutyrate output rose by 2.3 ± 0.6 and 2.9 ± 1.0 μmol · kg−1 · min−1 with exercise in C and SG compared to an increase of only 0.6 ± 0.2 μmol · kg−1 · min−1 in BG. Arterial blood acetoacetate did not increase appreciably from resting levels of 86 ± 7, 78 ± 7, and 60 ± 5 μM during exercise in C, SG, or BG. In C and SG, net hepatic acetoacetate output increased from 0.4 ± 0.1 and 0.9 ± 0.2 μmol · kg−1 · min−1 at rest to 1.7 ± 0.5 and 1.5 ± 0.3 μmol · kg−1 · min−1 by the end of exercise. In BG, acetoacetate production increased from 0.5 ± 0.1 to only 0.8 ± 0.2 μmol · kg−1 · min−1with exercise. The intrahepatic ketogenic efficiency was 0.13 ± 0.02, 0.10 ± 0.01, and 0.09 ± 0.01 at rest in C, SG, and BG. This variable was 0.16 ± 0.06 and 0.19 ± 0.04 by the end of exercise in C and SG but only 0.07 ± 0.02 in BG. These data indicate that, in normal dogs, the exercise-induced increment in glucagon is essential for the full rise in hepatic ketone output during muscular work and that this effect of the increase in glucagon is mediated by intrahepatic mechanisms.