Fuel metabolism in Canada geese: effects of glucagon on glucose kinetics

Abstract

During prolonged fasting, birds must rely on glucose mobilization to maintain normoglycemia. Glucagon is known to modulate avian energy metabolism during prolonged fasting, but the metabolic effects of this hormone on long-distance migrant birds have never been investigated. Our goal was to determine whether glucagon regulates the mobilization of the main lipid and carbohydrate fuels in migrant birds. Using the Canada goose (Branta canadensis) as a model species, we looked for evidence of fuel mobilization via changes in metabolite concentrations. No changes could be found for any lipid fraction, but glucagon elicited a strong increase in glucose concentration. Therefore, we aimed to quantify the effects of this hormone on glucose kinetics using continuous infusion of 6-[³H]-d-glucose. Glucagon was found to cause a 50% increase in glucose mobilization (from 22.2 ± 2.4 μmol·kg⁻¹·min⁻¹ to 33.5 ± 3.3 μmol·kg⁻¹·min⁻¹) and, together with an unchanged rate of carbohydrate oxidation, led to a 90% increase in plasma glucose concentration. This hormone also led to a twofold increase in plasma lactate concentration. No changes in plasma lipid concentration or composition were observed. This study is the first to demonstrate how glucagon modulates glucose kinetics in a long-distance migrant bird and to quantify its rates of glucose mobilization.

glucagon regulates the mobilization of carbohydrate reserves and plays a key role in metabolic fuel selection in mammals (23). In birds, however, the endocrine regulation of fuel metabolism is much less well understood. They are known to use glucagon to modulate energy metabolism during prolonged fasting and endurance exercise (16, 29), and this hormone is often considered as their main activator of lipolysis (5, 11, 21, 29). The Canada goose (Branta canadensis) is a long-distance migrant, making annual round trips of more than 7,000 km to the Arctic, including single, nonstop flights reaching 1,000 km (38). Therefore, it must seasonally endure multiple fasting events during long flights and at stopover sites when snow cover hinders foraging. Current knowledge about fuel metabolism in migrant birds is mainly based on measurements of metabolite concentrations (reviewed in Ref. 36), although a few studies using indirect calorimetry or substrate kinetics are also available. Changes in metabolite concentrations provide valuable information but cannot be used to draw conclusions about rates of fuel mobilization or utilization (25). Indirect calorimetry shows that migrant birds use lipids as a dominant fuel to support fasting and endurance exercise (27, 28, 47, 50, 51). The only study reporting substrate kinetics in a long-distance migrant reveals that, relative to other vertebrates, ruff sandpipers have a high lipolytic rate (48). Only one previous study deals with the effects of glucagon on avian fuel kinetics, and it was carried out on the king penguin: a nonflying species. It shows that this hormone stimulates lipid fluxes during fasting in king penguins and causes large increases in circulating levels of glucose, glycerol, and nonesterified fatty acids (5). The metabolic effects of glucagon on flying migrant birds have never been investigated in vivo. However, an in vitro study on the isolated abdominal fat pads of white-crowned sparrows reported that this hormone had no effect on glucose uptake but reduced lipid synthesis (20). Lipoproteins may play a role in transporting energy from lipid stores to flight muscles (17, 22, 31), but it is unclear whether glucagon is involved in this process. Therefore, the goal of this study was to measure whether glucagon regulates the mobilization of the main lipid and carbohydrate fuels in migrant birds using the Canada goose as a model species. More specifically, 1) we monitored the plasma levels of phospholipids and triacylglycerol (two main components of lipoproteins), as well as nonesterified fatty acids, glycerol, glucose, and lactate during glucagon infusion; 2) we examined the fatty acid composition of plasma phospholipids, triacylglycerol, and nonesterified fatty acids to characterize the lipid profile of Canada goose plasma and to determine whether glucagon affects specific fatty acids; and 3) because glucagon did not affect plasma lipid concentration and composition, but caused a strong response on plasma glucose concentration, we quantified the effects of this hormone on glucose kinetics to determine glucose mobilization. In addition, tissue carbohydrate content was measured to estimate how long this source of fuel could sustain resting metabolic rate. Even though migrant birds are known for their extraordinary capacity to mobilize and utilize lipids, our results showed no changes in plasma lipid concentration, whereas a large increase in glucose concentration was observed. Therefore, we hypothesized that glucagon stimulates the rate of glucose mobilization in this long-distance migrant bird, to provide more carbohydrates for oxidation.

MATERIALS AND METHODS

Animals.

Wild Canada goose eggs (Branta canadensis maxima) were collected on Varennes Islands (QC, Canada; Canadian Wildlife Service permit SC-19). They were artificially incubated, and the hatchlings were imprinted on E. Vaillancourt to facilitate data collection on calm, resting animals. During their first month, the goslings were raised indoors, fed duck starter feed [Ritchie Feed and Seed (RFS), Ottawa, ON, Canada] and freshly cut clover, and exercised outdoors for 2–4 h daily. The goslings were then raised to adulthood on a local farm (Gatineau, QC, Canada) where they ate grass and were supplemented with turkey feed (14% protein; 7% lipid; 4% fiber; RFS). Because the geese lived outdoors, from September to March, they were also provided with cracked corn (RFS) to build and maintain lipid reserves to survive through the winter. Before the experiments, which took place between September and December 2012, the animals therefore had started to build up their winter lipid reserves. Eight imprinted adult Canada geese (three males and five females) were transferred indoors from this outdoor captive colony. The birds were kept at 20°C, two individuals at a time, for 1–2 wk, in a windowless room (6.1 × 4.2 × 2.1 m) with an artificial light regime mimicking the current natural photoperiod. During the acclimation period, the birds had ad libitum access to turkey feed and running water (67×42×11-cm basin) and were acclimated to these conditions for at least 2 days before undergoing surgery. Postsurgery, the turkey feed was softened using tap water to facilitate deglutition. The transfer from the outdoor captive colony to the indoor facility lasted 15–30 min, acclimation was allowed for 2–9 days, surgery (from induction to awakening) lasted 2–4 h, 2 days were allowed for postsurgical recovery, and experiments lasted 3 h. All procedures were approved by the Animal Care Committee of the University of Ottawa and adhered to the guidelines established by the Canadian Council on Animal Care for the use of animals in research.

Catheterizations.

Two days before measuring glucose kinetics, occlusive catheters were placed in the right jugular vein and left carotid artery for in vivo measurement of glucose kinetics. During surgery, the animals were placed on their left side, rather than on their back, to prevent air sacs from collapsing (as observed in previous experiments; 48). After tracheal intubation, catheterization was performed under 1–2% isoflurane anesthesia. Both catheters were Chronic-Cath-treated with Hydrocoat (Norfolk Medical, Skokie, IL) and had a female luer-lock, an injection cap, and two fixed beads. The venous catheter (CC-3.5H CUSTOM) was a 3.5 Fr (0.6 mm ID × 1.1 mm OD), 40.6 cm in length, with the fixed beads at 5.1 cm and 6.4 cm from the insertion end. The arterial catheter (CC-5H CUSTOM) was a 5 Fr (1.0 mm ID × 1.7 mm OD), 45.7 cm in length, with the fixed beads at 10.2 cm and 11.4 cm from the insertion end. The catheters were fed into their respective vessels until both beads were inserted, sutured in place, and exteriorized in the interscapular area. They were filled with saline containing heparin (20 U/ml) and penicillin G (125,000 U/ml) and were flushed twice a day to maintain patency. The catheters were coiled and secured in a neoprene pocket sewn on the interior of a custom-made jacket developed with the assistance of Lomir Biomedical (Notre-Dame-de-l'Île-Perrot, QC, Canada). Particular care was taken to avoid injecting heparinized saline into the circulation. This was accomplished by withdrawing the heparinized saline until blood came out of each catheter, and then refilling each catheter using fresh heparinized saline with the catheter volumes determined during surgery. Surgical success rate was low because birds are very sensitive to anesthesia, and maintaining double catheter patency was particularly challenging. Continuous tracer infusion and blood sampling were only completed successfully in five of the eight birds (4.0 ± 0.1 kg). Three individuals failed to recover quickly from surgery or lost catheter patency.

Indirect calorimetry.

Food was withheld for 8 h before starting measurements to obtain reliable measurements of baseline metabolic rate (by removing the heat increment of feeding) and to facilitate postexperimental equipment cleanup (2). Rates of oxygen consumption (Ṁ_O2) and carbon dioxide production (Ṁ_CO2) were then measured with a calibrated Oxymax system (Columbus Instruments, Columbus, OH) (see Ref. 52 for details) connected to a custom-made, dual-purpose Lexan respirometer supplied with 20°C air at a rate of 8–12 l/min. Preliminary experiments revealed that geese remain quieter while standing on a slightly cooled surface. Therefore, the detachable bottom of the respirometer was removed to allow the birds to stand directly on a custom-made, 60 × 60-cm thermostated aluminum plate kept at 15°C using a water bath. Substituting the flat Lexan bottom with the flat aluminum plate did not change the volume of the respirometer and did not affect how it operated.

Glucose kinetics.

The rate of glucose mobilization was measured by continuous infusion of radiolabeled glucose (24). Luered extension lines reaching outside the respirometer were added to the catheters (ES-18-M/F; Norfolk Medical, Skokie, IL). The infusate was freshly prepared immediately before each infusion by mixing 4.44 MBq of 6-[³H]-d-glucose (Amersham, Oakville, ON, Canada; ∼28 GBq/mmol) to a final volume of 4 ml with sterile saline. Continuous infusion of 6-[³H]-d-glucose was performed through the venous catheter at 1 ml/h with a calibrated syringe pump (Harvard Apparatus, South Natick, MA). Average infusion rate was 2.32 ± 0.27 × 10³ Bq·kg⁻¹·min⁻¹ (n = 5). We set our time scale (t = 0 min) at the start of glucagon infusion (0.125 μg glucagon·kg⁻¹·min⁻¹). A priming dose of labeled glucose equivalent to 30 min of infusion was administered before starting the actual tracer infusion at t = −60 min. This protocol ensured the large glucose pool was labeled rapidly and, therefore, that isotopic steady state was reached in less than 40 min. Blood samples (1 ml each) were drawn from the arterial catheter at t = −20 min, −10 min, and 0 min to determine baseline glucose kinetics. Additional blood samples were taken every 30 min (t = 30 min and 60 min) during the glucagon infusion, as well as for 60 min during recovery from the glucagon infusion (t = 90 min and 120 min). A preliminary experiment, in which glucagon was administered at increasing rates [0.000 (vehicle only), 0.025, 0.050, and 0.125 μg glucagon·kg⁻¹·min⁻¹, based on the rates used in two bird studies (5, 21)] for 60 min each, revealed that the infusion rate of 0.125 μg glucagon·kg⁻¹·min⁻¹ was the lowest rate to cause an increase in blood glucose levels (from 9.5 to 19.7 mmol/l). At the lower infusion rates (vehicle, 0.025 and 0.050 μg glucagon·kg⁻¹·min⁻¹), blood glucose levels were 9.2–9.7 mmol/l, a range expected for resting birds. We were satisfied with the increase in blood glucose levels observed with the glucagon infusion rate of 0.125 μg·kg⁻¹·min⁻¹ and decided not to use greater infusion rates to ensure the procedure would not harm our birds. For this reason, the physiological effects of glucagon (or lack thereof) presented in this study may not necessarily be representative of other glucagon infusion rates. Continuous infusion of glucagon was also performed through the venous catheter with a calibrated syringe pump. Both luered syringes (glucagon and labeled glucose) were connected to the venous catheter extension using a luered three-way stopcock acting as a Y-connector. Blood was centrifuged immediately after sampling to separate the plasma that was stored at −20°C until analyses.

Tissue sampling.

Nine further imprinted adult Canada geese (four males and five females) were anesthetized at the farm in December 2012 using an intramuscular injection of ketamine and xylazine and, while anesthetized, were euthanized using an overdose of intraperitoneally administered pentobarbital sodium. Approximately 5 g of tissue were rapidly excised and freeze-clamped using aluminum tongs precooled in liquid nitrogen. Pectoralis samples were taken along the anterior part of the keel, along the whole thickness of the muscle. All tissues were sampled and frozen within 10 min following death. The pectoralis and liver were then dissected out of the carcasses, placed in labeled plastic bags, and brought back to the laboratory where organ weight (including the freeze-clamped samples) for both tissues was recorded to calculate carbohydrate reserves. The samples were stored at −80°C until assayed for carbohydrate concentrations. For analysis, the brittle freeze-clamped pectoralis samples were shattered in small pieces and randomly sampled, thereby allowing measurements representative of the whole thickness of the muscle.

Glucose and glycogen concentration.

Measurement procedures for tissue concentration of glucose and glycogen were adapted from Fournier and Weber (14). Briefly, ∼1 g of tissue (pectoralis, liver) was finely ground in liquid nitrogen with a precooled mortar and pestle. Each frozen sample was weighed and placed in 4 volumes of ice-cold 6% perchloric acid (PCA). Following homogenization, the samples were centrifuged at 2,800 g for 10 min. The supernatants were then aliquoted in 1.5-ml centrifuge tubes and stored at −20°C until assayed. Glucose concentration was determined according to Bergmeyer (4). Glycogen concentration was determined using amyloglucosidase (A-7095; Sigma-Aldrich, St. Louis, MO) to hydrolyze glycogen, followed by the addition of 6% PCA to stop the reaction. Glucose and glycogen standards and negative controls were used in each series of assay, because the amyloglucosidase contained significant amounts of glucose and/or glycogen. The tissue glycogen concentrations were corrected for this contribution by the contaminated enzyme and for the free glucose already present before glycogen hydrolysis.

Plasma analyses.

Plasma lipids were analyzed as in Vaillancourt and Weber (48). Briefly, heptadecanoate (17:0; 0.30 mg/ml; a fatty acid that has a very low abundance in birds) was added to plasma as an internal standard for subsequent analysis of nonesterified fatty acids (NEFA) by gas chromatography. Total plasma lipids were extracted twice in chloroform:methanol (2:1 vol/vol) (13). The aqueous phase was discarded. The organic phase (containing the lipids) was dried at 70°C under N₂ and resuspended in hexane:isopropanol (3:2 vol/vol). Triacylglycerol (TAG), NEFA, and phospholipids (PL) were separated by sequential elution from solid-phase extraction tubes (Supelclean, 1 ml, 100 mg LC-NH₂; Sigma-Aldrich). TAG was eluted with chloroform:isopropanol (2:1 vol/vol), NEFA with isopropyl ether:acetic acid (98:2 vol/vol) and PL with methanol. Because the 17:0 initially added to the plasma eluted with the NEFA fraction, 17:0 was added to the eluted TAG and PL fractions. After methylation (NEFA) or acid transesterification with acetyl chloride in methanol (TAG and PL) (1), the fatty acid composition of each fraction was analyzed by gas chromatography. Individual fatty acid methyl esters were separated on a 6890N gas chromatograph (Agilent Technologies, Mississauga, ON, Canada) equipped with a flame-ionization detector (FID) and a fused silica capillary column (DB-23, 60 m, 0.25 mm I.D., 0.25 μm film thickness; Agilent Technologies). Hydrogen was the carrier gas. The injector port was at 220°C and the FID at 240°C. Column temperature was kept at 130°C for 1 min, raised to 170°C at a rate of 6.5°C/min, immediately raised to 215°C at a rate of 2.75°C/min and maintained at 215°C for 12 min, and finally raised to 230°C at a rate of 40°C/min and maintained at 230°C for 12 min. Exact retention times of individual fatty acids were determined with pure standards (Sigma-Aldrich). Plasma glucose, lactate, and glycerol concentrations were determined at 340 nm on a SpectraMax Plus 384 spectrophotometer (Molecular Devices, Sunnyvale, CA), according to Bergmeyer (4). Radioactivity was measured by liquid scintillation counting (Beckman Coulter LS 6500, Fullerton, CA) in Bio-Safe II scintillation fluid (RPI Corp., Mount Prospect, IL). Specific activity was calculated by dividing glucose activity by glucose concentration.

Calculations and statistics.

Rates of carbohydrate and lipid oxidation were calculated from Ṁ_O2 and Ṁ_CO2, as described previously (48), and were corrected for protein oxidation assuming a rate of nitrogen excretion of 0.197 mg nitrogen·kg⁻¹·min⁻¹. This value was derived from the nitrogen excretion of ruff sandpipers [0.534 mg nitrogen·kg⁻¹·min⁻¹ (47)] allometrically scaled from a 110-g ruff sandpiper to a 4.0-kg Canada goose (33), based on the resting metabolic rate of nonpasserine birds. Following this correction, the relative contribution of protein oxidation to Ṁ_O2 was 18.4 ± 0.5%, a value in agreement with the protein content (14%) of the turkey feed provided during the postsurgical recovery period. The rate of glucose mobilization was calculated with the steady-state equation of Steele (44). Statistical comparisons for plasma metabolite concentrations, Ṁ_O2, Ṁ_CO2, respiratory exchange ratio (RER), rates of lipid and carbohydrate oxidation, relative contribution to Ṁ_O2 of lipid and carbohydrate oxidation, glucose specific activity, glucose mobilization, and relative contribution of glucose flux to Ṁ_O2 and to glucose oxidation were performed in a two-tiered approach using one-way, repeated-measures ANOVA or Friedman repeated-measures ANOVA on ranks when the assumptions of normality and homoscedasticity were not met. In the first-tier analysis, all time points were compared with t = 0 min. This allowed us 1) to determine whether baseline values had reached steady state and 2) to use t = 0 min as the control value if steady state had not been reached. The second-tier analysis was performed when baseline had reached steady state. Baseline values (t ≤ 0 min) were averaged, and this baseline average was used as the control value when comparing values obtained during glucagon infusion and recovery. For both tiers, when statistically significant changes were detected, Bonferroni's post hoc test was used to determine which means were different from baseline values. Statistical comparisons for glucose and glycogen concentration in the pectoralis and liver were performed using paired t-test. All percentages were transformed to the arcsine of their square root before analysis. Significance threshold was set at P < 0.05, and all of the values are presented as means ± SE.

RESULTS

Plasma metabolite concentrations.

Baseline concentrations of PL, NEFA, and TAG were 3.2 ± 0.1, 1.6 ± 0.2, and 0.29 ± 0.02 mmol/l, respectively, and all remained unaffected by glucagon (Fig. 1A; P > 0.05). Similarly, glycerol concentration did not change from its baseline level of 0.35 ± 0.04 mmol/l throughout the experiment (Fig. 1B; P > 0.05). By contrast, baseline glucose concentration was 9.6 ± 0.6 mmol/l, increased to 17.9 ± 1.4 mmol/l with glucagon, and remained elevated throughout recovery (Fig. 1B; P < 0.001). Baseline lactate concentration was 1.5 ± 0.1 mmol/l and increased to 3.1 ± 0.9 mmol/l within the first 30 min of glucagon infusion (Fig. 1B; P < 0.001), before returning to baseline within 30 min of recovery. The relative abundance of individual fatty acids in NEFA, TAG, and PL did not change with glucagon (P > 0.05), and average values are shown in Table 1.

Fig. 1.

Plasma nonesterified fatty acid (A), phospholipid (A), triacylglycerol (A), glycerol (B), glucose (B), and lactate (B) concentration in adult Canada geese before, during, and after glucagon infusion. *Significant difference from baseline (P < 0.05). **Significant difference from baseline (P < 0.01). ***Significant difference from baseline (P < 0.001). Values are expressed as means ± SE (n = 5).

Table 1.

Relative contributions (mass percent) of individual fatty acids to total nonesterified fatty acids, triacylglycerol, and phospholipids in plasma of Canada geese

	NEFA	TAG	PL
16:0	22.2 ± 0.4	12.6 ± 0.1	22.4 ± 0.2
16:1 (n-7)	1.6 ± 0.0	t.a.	t.a.
18:0	28.9 ± 1.0	19.1 ± 0.2	25.5 ± 0.2
18:1 (n-9)	37.2 ± 1.1	25.7 ± 0.3	10.6 ± 0.0
18:1 (n-6)	1.8 ± 0.5	1.2 ± 0.1	1.3 ± 0.0
18:2 (n-6)	5.8 ± 0.2	9.4 ± 0.1	7.4 ± 0.1
18:3 (n-6)	1.4 ± 0.1	t.a.	t.a.
20:4 (n-6)	1.1 ± 0.1	20.4 ± 0.3	29.2 ± 0.2
21:0	n.d.	7.1 ± 0.5	t.a.
22:1 (n-9)	n.d.	2.1 ± 0.1	n.d.
22:5 (n-6)	n.d.	t.a.	1.1 ± 0.0
22:6 (n-3)	t.a.	2.3 ± 0.1	2.5 ± 0.0
SFA	51.1 ± 1.3	38.9 ± 0.2	47.9 ± 0.1
MUFA	40.6 ± 1.1	31.3 ± 0.3	14.4 ± 0.0
PUFA	8.3 ± 0.2	29.8 ± 0.3	37.7 ± 0.1
DBI	1.2 ± 0.1	3.4 ± 0.0	3.2 ± 0.0
Chain length	17.5 ± 0.0	18.6 ± 0.0	18.3 ± 0.0
Total concentration, μmol FA/ml plasma	1.644 ± 0.018	0.855 ± 0.007	6.385 ± 0.016

Values are mass percent of the fatty acid methyl esters expressed as means ± SE (n = 5). NEFA, nonesterified fatty acids; TAG, triacylglycerol; PL, phospholipids; SFA, saturated fatty acids; MUFA, monounsaturated fatty acids; PUFA, polyunsaturated fatty acides; DBI, double-bond index; n.d., not detected; t.a., trace amounts. Only fatty acids contributing >1% of total fatty acid concentration in at least one fraction are tabulated. For each fraction, the total concentration (in micromoles of FA per milliliter plasma), the relative contributions of SFA, MUFA, and PUFA, the DBI, and the mean carbon chain length are also indicated. The relative abundance of individual fatty acids in each lipid fraction did not change with glucagon infusion; thus, average values for the whole experiment are presented here.

Gas exchange.

Baseline values for these resting Canada geese averaged 270 ± 22 μmol·kg⁻¹·min⁻¹ (Ṁ_O2; Fig. 2A), 213 ± 17 μmol·kg⁻¹·min⁻¹ (Ṁ_CO2; Fig. 2B), and 0.80 ± 0.06 (RER; Fig. 2C). Except for a weak and transient increase in Ṁ_O2 to 338 ± 26 μmol O₂·kg⁻¹·min⁻¹ early in recovery (P < 0.05), these parameters remained unaffected by glucagon (P > 0.05).

Fig. 2.

Rates of oxygen consumption (Ṁ_O2) (A), carbon dioxide production (Ṁ_CO2) (B), and respiratory exchange ratio (RER) (C) of adult Canada geese before, during, and after glucagon infusion. *Significantly different from baseline (P < 0.05). Values are expressed as means ± SE (n = 5).

Metabolic fuel oxidation.

Baseline rates of lipid and carbohydrate oxidation were 145 ± 59, and 72 ± 48 μmol O₂·kg⁻¹·min⁻¹, respectively. These parameters remained unaffected by glucagon (P > 0.05). The baseline relative contributions of lipid and carbohydrate oxidation to Ṁ_O2 were 51.4 ± 22.3% and 28.7 ± 21.4%, respectively. Both parameters remained unaffected by glucagon (Fig. 3B; P > 0.05).

Fig. 3.

Rates (A) and relative contribution to Ṁ_O2 (B) of lipid and carbohydrate oxidation (CHO) in adult Canada geese before, during, and after glucagon infusion. Values are expressed as means ± SE (n = 5).

Glucose kinetics.

Baseline glucose specific activity at isotopic steady state was 107 ± 11 Bq/μmol. Specific activity decreased within 30 min of initiating the glucagon infusion and remained below baseline during recovery (Fig. 4A; P < 0.001). Baseline rate of glucose mobilization [or rate of hepatic production or rate of appearance or turnover rate (R_t)] was 22.2 ± 2.4 μmol·kg⁻¹·min⁻¹. Glucose mobilization was already stimulated after 30 min of glucagon infusion and remained ∼50% above baseline until the end of the experiment (Fig. 4B; P < 0.001). It reached a maximal value of 33.5 ± 3.3 μmol glucose·kg⁻¹·min⁻¹ after 30 min of recovery.

Fig. 4.

Plasma glucose specific activity (A) and rate of mobilization (B) in adult Canada geese before, during, and after glucagon infusion. Note that, at steady state, glucose mobilization is synonymous with glucose flux, glucose turnover rate (R_t), hepatic glucose production, and rate of appearance (R_a) of glucose. ***Significant differences from baseline (P < 0.001). Values are expressed as means ± SE (n = 5).

Relative contribution of glucose as an oxidative fuel.

Calculations were made to determine the potential contribution of glucose oxidation to total metabolic rate (Ṁ_O2) using two different assumptions: either that 50% or that 100% of glucose flux was oxidized. Mammalian studies show that, at rest, only about half of total glucose flux is oxidized [rats: 43–45% (8); dogs: 30–50% (39); humans: 40–60% (19, 32)] and, for this reason, the 50% value was selected. The assumption that 100% of glucose flux is oxidized is probably unrealistic, but it provides upper-bound values for the maximal possible contribution of glucose oxidation to Ṁ_O2. Assuming 50%, this fuel would account for 25.0 ± 2.2% of Ṁ_O2 under baseline conditions and reach a maximum of 34.8 ± 3.4% of Ṁ_O2 after 30 min of glucagon infusion (P < 0.01). Assuming 100% of glucose flux was oxidized, this fuel alone would be responsible for 50.1 ± 4.4% of Ṁ_O2 under baseline conditions and reach 69.7 ± 6.9% of Ṁ_O2 during glucagon infusion (P < 0.01). Using these same assumptions, we have determined the potential contribution of glucose to carbohydrate oxidation. Assuming 50% of glucose flux was oxidized, the contribution from this substrate would be 53.3 ± 6.2% of carbohydrate oxidation at t = 0 min, increase to 89.0 ± 9.7% during glucagon infusion and late recovery (t = 30 min, 60 min, and 120 min; Fig. 5B; P < 0.001), and reach a maximum of 160 ± 16% in early recovery (t = 90 min; Fig. 5B; P < 0.001). Assuming 100% of glucose flux was oxidized, this substrate would account for 107 ± 12% of carbohydrate oxidation at t = 0 min, increase to an average of 178 ± 19% during glucagon infusion and late recovery (t = 30 min, 60 min, and 120 min; Fig. 5B; P < 0.001), and reach a maximum value of 319 ± 32% in early recovery (t = 90 min; Fig. 5B; P < 0.001). Because these calculations often yield values greater than 100%, the assumption that 100% of glucose flux was oxidized is obviously unrealistic. This shows that glucose mobilization exceeded the need for oxidation, and that nonoxidative disposal must have occurred. In the case of the 319% value (t = 90 min; Fig. 5B), its reciprocal indicates that a maximum of 31% of glucose flux was actually oxidized and that at least 69% of glucose flux was accounted for by nonoxidative disposal.

Fig. 5.

Relative contribution of glucose to total energy expenditure (A) and carbohydrate oxidation (B) of adult Canada geese before, during, and after glucagon infusion, assuming 100% (▲) or 50% (○) of the glucose undergoing turnover is oxidized. *Significant difference from baseline (P < 0.05). ***Significant difference from baseline (P < 0.001). Values are expressed as means ± SE (n = 5).

Carbohydrate reserves.

The concentration of glucose was 4.06 ± 0.24 μmol/g wet tissue in the pectoralis and 42.6 ± 4.1 μmol/g wet tissue in liver (data not shown; P < 0.001). The concentration of glycogen was 30.1 ± 5.5 μmol glucosyl units/g wet tissue in the pectoralis muscle and 64.7 ± 8.9 μmol glucosyl units/g wet tissue in liver (data not shown; P < 0.05). Glucose and glycogen concentrations were greater in the liver than in the pectoralis muscle, and hepatic carbohydrate reserves could, if they were the only metabolic fuel, sustain the resting metabolic rate measured here in Canada geese for 30 min. This value was calculated using the average weight of the liver (51.0 ± 2.7 g) and yielded hepatic carbohydrate reserves of 5.5 ± 0.7 mmol glucosyl units.

DISCUSSION

This in vivo study is the first to provide measurements of glucose mobilization in a long-distance migrant bird, and it shows that glucagon modulates the glucose kinetics of Canada geese. This hormone caused the baseline rate of glucose mobilization (22.2 μmol·kg⁻¹·min⁻¹) to increase by 50% and, together with an unchanged rate of carbohydrate oxidation, led to a 90% increase in plasma glucose concentration. Glucagon also caused a twofold increase in plasma lactate concentration. Even though glucagon is often considered as the main lipolytic hormone, at the infusion rate used here, it did not modulate the concentration of any lipid fraction.

Glucagon increases glucose mobilization.

Although plasma glucagon levels were not measured in wild geese, at the infusion rate used in this study, glucagon increased glucose mobilization (Fig. 4B) and concentration (Fig. 1B). Because the rate of carbohydrate oxidation remained unaffected (Fig. 3A), glucagon may be important to maintain normoglycemia when the birds are forced to fast (e.g., during very long flights, and during northward, spring stopovers when their foraging grounds are covered with snow). In a previous study, fasting penguins administered one-fifth of the glucagon given to our geese also had an increase in plasma glucose concentration, and the effect of glucagon also lasted at least an hour postinfusion, even though plasma glucagon concentration did return to near-baseline levels (5). In domestic geese infused with glucagon at rates of 0.05–0.50 μg·kg⁻¹·min⁻¹ (21, 43), blood glucose levels were still elevated 35–45 min after the end of the glucagon infusion. These observations made in penguins and domestic geese are in agreement with our results in Canada geese. Unfortunately, plasma glucagon concentrations were not measured in our study, but on the basis of a study by Sitbon and Mialhe (43), it can be argued that the glucagon infusion used here led to plasma glucagon levels that are physiologically relevant for birds. In that study, domestic geese were infused with 0.050 μg glucagon·kg⁻¹·min⁻¹ (an infusion rate 2.5 times lower than the rate used in our study), and plasma glucagon concentrations are reported at baseline, during glucagon infusion, and during recovery. Glucagon levels were 1.1 ng/ml at baseline, increased to 1.7 ng/ml during glucagon infusion, and returned to baseline values within 5 min into recovery. Assuming the increased plasma glucose levels had completely suppressed endogenous glucagon production, exogenous glucagon would have been responsible for the totality of the plasma glucagon concentration (1.7 ng/ml) observed during glucagon infusion. In this scenario, we can estimate that our Canada geese would have had a plasma glucagon concentration of 4.3 ng/ml. Assuming the increased plasma glucose levels had no effect on endogenous glucagon production, exogenous glucagon would have been responsible only for the increase in plasma glucagon concentration (+0.6 ng/ml) observed during glucagon infusion. In this scenario, we can estimate that our Canada geese would have had a 1.5 ng/ml increase in glucagon levels above the 1.1 ng/ml baseline, for a total plasma glucagon concentration of 2.6 ng/ml during the infusion. This theoretical range of plasma glucagon concentrations (2.6–4.3 ng/ml) is within the physiological range of the species for which we could find data [garden warblers: 2.5–3.4 ng/ml (46); domestic fowl: 2.2–4.9 ng/ml (7); domestic duck: 1.5–6.5 ng/ml (45)]. The resting rate of glucose mobilization observed here in Canada geese (22.2 μmol·kg⁻¹·min⁻¹) was lower than those reported in domestic fowls (41 μmol glucose·kg⁻¹·min⁻¹) and American goldfinches (228 μmol glucose·kg⁻¹·min⁻¹) (6, 35). The lower rate of glucose mobilization in Canada geese, compared with domestic fowls, may be explained by 1) the greater reliance of long-distance migrant birds on lipids than on carbohydrates during flight and 2) the low stress levels experienced by imprinted geese [the baseline Ṁ_O2 of 270 μmol O₂·kg⁻¹·min⁻¹ measured here is below the theoretical resting value of 344 μmol O₂·kg⁻¹·min⁻¹ predicted for our geese (33)]. However, allometric scaling for this nonpasserine (33), 3.1-kg species does not explain these differences, because the resting rate of glucose mobilization of a 4-kg domestic fowl would still be 38.2 μmol·kg⁻¹·min⁻¹, a value almost twice that observed in Canada geese. Compared with the American goldfinches, the much lower rate of glucose mobilization in Canada geese may be explained by body size effects. According to the allometric equation of Lasiewski and Dawson (33) for the resting metabolic rate (RMR) of passerine birds, the weight-specific metabolic rate of a 13.4-g American goldfinch is expected to be 8 times greater than that of a 4-kg Canada goose. Allometrically scaled, the resting rate of glucose mobilization of a hypothetical 4-kg American goldfinch would be 29 μmol·kg⁻¹·min⁻¹, a value comparable to the rate observed in Canada geese. Similarly, the resting rate of glucose mobilization values measured in Canada geese are comparable to those observed in resting mammals [dogs: 18–23.4 μmol·kg⁻¹·min⁻¹ (37, 53); goats: 25–30 μmol·kg⁻¹·min⁻¹ (53); young pigs: 17.3 μmol·kg⁻¹·min⁻¹ (30); humans: 13.2 μmol·kg⁻¹·min⁻¹ (49)]. However, it is anticipated that the values for the rate of glucose mobilization reported here are at the lower end of the range achievable by Canada geese, because 1) glycogenolysis increases in the liver of migrant birds in preparation for migration (40) and 2) the imprinted, relatively sedentary animals used here may not have been physiologically prepared for migration (41). Even though they were occasionally allowed outside of their protective enclosure to make short bouts of exercise, the frequency, duration, and intensity of these flights are dwarfed by those performed by their wild counterparts.

Glucagon increases plasma lactate concentration.

Glucagon infusion caused a twofold increase in plasma lactate concentration (Fig. 1B). It was probably unrelated to experimental stress, because this increase was only observed during glucagon infusion, and lactate levels promptly returned to their low baseline levels during recovery. Possibly, glucagon or a greater availability of substrate caused by an increased glucose concentration (Fig. 1B) might have stimulated glycolysis, thereby making more pyruvate available for lactate production and regeneration of NAD⁺ required for glycolysis.

Rates of total carbohydrate and lipid oxidation remain unaffected by glucagon.

Glucagon had no significant effect on total lipid and carbohydrate oxidation (Fig. 3A) or on the relative contributions of lipids and carbohydrates to Ṁ_O2 (Fig. 3B). The greater relative contribution of lipid oxidation to sustain the metabolic rate observed in Canada geese (Figs. 2C and 3B) is expected for this long-distance migrant bird species. Resting barnacle and bar-headed geese, pigeons, ruff sandpipers, and European starlings are also known to strongly rely on lipids (9, 12, 28, 47, 48, 51). In contrast, the relatively large contribution of carbohydrate oxidation to Ṁ_O2 is surprising. However, considering the natural history of our model species, it is possible that an 8-h fast does not cause fat reserves to be used exclusively. A longer fasting period appears to be required to elicit a greater contribution of lipids at rest. In fact, the birds we used for liver glycogen measurements were fasted overnight, and they still had very large reserves of carbohydrates, even though they were housed outside in the cold. In addition, relatively large contributions of carbohydrate oxidation to Ṁ_O2 have been reported in other bird species at rest [sunbird: RER = 0.85 (27); zebra finch: RER = 0.87 (26); European starling: RER = 0.82 (51)]. Therefore, we consider the Canada goose to be a model representative of other migrant bird species, and we expect this species would rely more heavily on lipids during prolonged flight.

Glucagon does not affect lipid concentration and fatty acid composition.

Curiously, glucagon had no effect on the concentration of any circulating lipid (NEFA, TAG, PL; Fig. 1A) or on their fatty acid composition (Table 1). This is in sharp contrast with the major increase in glycerol and NEFA concentrations observed in penguins infused with glucagon at a much lower rate (5). This might be explained by 1) differences in feeding states [although because of their respective natural history, penguins fasted for 7–9 days, and geese fasted for only a few hours, both are expected to be in phase I of fasting (10, 34)], 2) interspecific differences in exercising habits [whereas king penguins swim by alternating 3-min dives with 6-min recovery periods (42), Canada geese can make uninterrupted flights reaching 1,000 km (38)], 3) interspecific differences in lipid store function [geese use their energy-dense fat stores mainly as a source of fuel for flight, whereas king penguins depend on blubber for energy and for insulation/thermoregulation], and 4) possible differences in plasma glucagon concentrations between the studies. The lack of change in TAG and PL concentrations in Canada geese suggests the average lipoprotein size remained unaffected by glucagon. The fatty acid concentrations for the various lipid fractions indicate that, in their physiological state, Canada geese transported their circulating fatty acids mainly as lipoproteins (81% of total plasma fatty acids), while NEFA played a much minor role (19%). These values would likely be different in the postprandial state or during flight. The fatty acid concentration for the NEFA, PL, and TAG fractions in Canada geese (1.6, 3.2 and 0.3 mmol/l; results from present study), ruff sandpiper [1.6, 8.1, and 1.0 mmol/l; (48)] and Western sandpiper [0.8, 4.9, and 6.1 mmol/l; (22)] yield PL/TAG ratios of 11.0, 8.1, and 0.8, respectively. This indicates that, in their specific physiological state, the average lipoprotein size in Canada geese was smaller (i.e., had a greater density) than in the other two species and consisted mostly of high-density lipoproteins (HDL) (3). The decreasing PL/TAG ratios for Canada geese, ruff sandpipers, and Western sandpipers could possibly be explained by decreasing fasting durations prior to experimentation. Whereas Canada geese and ruff sandpipers were fasted for 8 h (our study) and 1 h (48), respectively, Western sandpipers were captured and sampled after a 6- to 10-h feeding period in the mudflats (22). The fatty acid composition differed between plasma NEFA and lipoprotein (PL and TAG) fractions and between long-distance migrant species. In Canada geese, the most abundant free fatty acids (FFAs) were 16:0, 18:0, and 18:1 in the NEFA fraction, 18:0, 18:1, and 20:4 in the TAG fraction, and 16:0, 18:0, and 20:4 in the PL fraction. In ruff sandpipers, 16:0, 18:0, and 18:1 were also dominant in the NEFA fraction, 18:0, 18:1, and 22:1 abounded in the TAG fraction, and 16:0, 18:0, 20:4, and 22:6 were the most abundant FFAs in the PL fraction (48). In Western sandpipers, during the fall migration, the NEFA fraction was dominated by 18:0, 18:1, 20:5, and 22:6, whereas 16:0, 18:0, and 18:1 were most prevalent in the TAG and PL fractions (22). The greater abundance of long-chain, polyunsaturated fatty acids (e.g., 20:5 and 22:6) in sandpipers can be explained by differences in diets; geese feed on grass and grains (18), whereas sandpipers feed mainly on benthic invertebrates and other arthropods (15).

Hepatic carbohydrate reserves could sustain RMR for 30 min.

Hepatic carbohydrate reserves could, if they were the sole metabolic fuel, sustain the resting metabolic rate measured here in Canada geese for a half hour. However, Fig. 3B shows that carbohydrate oxidation was responsible for roughly 50% of the baseline metabolic rate. To prevent the depletion of carbohydrate reserves and maintain normoglycemia, lipid stores likely had to contribute glycerol for gluconeogenesis. The glycogen concentration found in the pectoralis muscle of Canada geese during fall (30.1 μmol glucosyl units/g wet tissue) is comparable to that of American goldfinches captured in winter and maintained at 30°C for several hours [25.7 μmol glucosyl units/g wet tissue (35)]. However, the concentration of glycogen found in the liver of Canada geese (64.7 μmol glucosyl units/g wet tissue) is 28-fold greater than that found in goldfinches captured in winter [2.33 μmol glucosyl units/g wet tissue (35)]. This large difference could reflect the stress induced by capture, fasting prior to and during the experiments, or diurnal variations. The authors noted that the concentrations of glycogen measured in the liver of their subjects were lower than those observed in some other small passerine species (35). In goldfinches captured with full crops and euthanized within 1 min of capture (n = 3), concentrations of liver glycogen [31.7 μmol glucosyl units/g wet tissue (35)] were much closer to those measured in the liver of Canada geese.

Perspectives and Significance

For wild geese forced to fast during very long flights and during their spring stopovers, glucose mobilization is essential to maintain normoglycemia. This study is the first to demonstrate how glucagon modulates the glucose kinetics of a long-distance migrant bird and quantifies the rates of glucose mobilization in Canada geese. Contrary to expectation, glucagon had no effect on the plasma concentration of any lipid fraction, but had a strong effect on glucose mobilization. Canada geese had a baseline rate of glucose mobilization of 22.2 μmol·kg⁻¹·min⁻¹, and glucagon caused this rate to increase 1.5-fold. This glucagon-induced increase in glucose mobilization led to a 1.9-fold increase in plasma glucose concentration but had no effect on total carbohydrate oxidation. A twofold increase in plasma lactate concentration was also observed. At the dose administered here, glucagon strongly stimulated hepatic glucose production but had no effect on lipid metabolism, even though it is often considered as the main avian lipolytic hormone. Determining whether different doses of glucagon modulate the fuel metabolism of Canada geese similarly strikes us as an exciting challenge for future work.

GRANTS

This research was supported by an National Sciences and Engineering Research Council (NSERC) discovery grant to J.-M.W, as well as by an NSERC Alexander Graham Bell Canada Graduate Scholarship, an Ontario Graduate Scholarship, a University of Ottawa Excellence Scholarship, a Prix Acfas-Desjardins, doctorat, an Arctic Institute of North America's Grant-in-Aid and Sigma Xi Grants-in-Aid of Research (Grant G2009101616 and G2009151055) to E.V.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Author contributions: E.V. and J.-M.W. conception and design of research; E.V. performed experiments; E.V. analyzed data; E.V. and J.-M.W. interpreted results of experiments; E.V. prepared figures; E.V. drafted manuscript; E.V. and J.-M.W. edited and revised manuscript; E.V. and J.-M.W. approved final version of manuscript.