Glucagon, in Concert With Insulin, Supports the Postabsorptive Plasma Glucose Concentration in Humans

Fourteen healthy individuals (eight women and six men) gave their written informed consent to participate in this study, which was approved by the Washington University Human Research Protection Office and conducted in the outpatient facilities of the Washington University General Clinical Research Center. Their mean (±SD) age was 29 ± 5 years. Their mean BMI was 23.0 ± 2.9 kg/m². All had negative medical histories and normal physical examinations as well as normal fasting plasma glucose and creatinine concentrations, hematocrits, and electrocardiograms.

Subjects were studied in the morning after an overnight fast and in the supine position throughout, on four separate occasions. To permit estimation of glucose kinetics, a primed (22.5 μmol/kg), continuous (0.25 μmol · kg⁻¹ · min⁻¹) intravenous infusion of [6,6-²H₂]glucose was started at −180 min and continued through 300 min. Arterialized venous blood samples, from a hand vein with that hand kept in a ∼55°C plexiglas box, were drawn at 15-min intervals from −30 through 300 min for plasma glucose and glucose-enrichment determinations and from −15 through 240 min and at 270 and 300 min for the other analytes detailed below. Heart rate and blood pressure were determined at those time points, and the electrocardiogram was monitored throughout.

Octreotide acetate (Sandostatin; Novartis Pharmaceuticals, East Hanover, NJ) was infused intravenously in a dose of 30 ng · kg⁻¹ · min⁻¹ (9), and human growth hormone (Genotropin; Pharmacia and Upjohn, Kalamazoo, MI) was infused intravenously in a dose of 3.0 ng · kg⁻¹ · min⁻¹, from 0 through 240 min on all four study occasions. Hormone infusates were prepared in saline containing 0.5 g/dl albumin. In random sequence on the four occasions the infusions were octreotide (with growth hormone) plus saline; octreotide plus regular human insulin (Novolin; Novo Nordisk, Bagsvaerd, Denmark) in a dose of 0.1 mU · kg⁻¹ · min⁻¹; octreotide plus glucagon (Eli Lilly, Indianapolis, IN) in a dose of 1.0 ng · kg⁻¹ · min⁻¹; and octreotide plus both insulin and glucagon.

Plasma glucose concentrations were measured with a glucose oxidase method (Yellow Springs Analyzer 2; Yellow Springs Instruments, Yellow Springs, OH). Plasma insulin (14), C-peptide (14), glucagon (15), pancreatic polypeptide (16), growth hormone (17), and cortisol (18) concentrations were measured with radioimmunoassays. The insulin, C-peptide, glucagon, and pancreatic polypeptide assays were performed with materials purchased from Linco Research (St. Louis, MO), and the cortisol assay with materials was purchased from Diasorin (Stillwater, MN). An antibody provided by the National Institutes of Health was used for the growth hormone assay. Plasma epinephrine and norepinephrine concentrations were measured with a single-isotope derivative (radioenzymatic) method (19). Serum nonesterified fatty acids (20) and blood lactate (21) were measured with enzymatic techniques.

Plasma proteins were precipitated with ice-cold acetone, and lipids were extracted with hexane. The aqueous phase was dried by Speed-Vac centrifugation (Savant Instruments, Farmingdale, NY). Samples were derivatized with 10% heptafluorobutyric anhydride in ethyl acetate (30 min at 70°C). The tracer-to-tracee ratio (TTR) of heptafluorobutyric-to-glucose was measured by gas chromatography–mass spectrometry using electron impact ionization (ions of mass/charge ratio 519 and 521 for natural and [6,6-²H₂]glucose, respectively) on an Agilent 5973 system equipped with a 30-m × 0.25-mm HP-5MS column. Instrument response was calibrated using prepared glucose standards of known isotopic enrichment. Non–steady-state kinetic analysis to obtain the rates of appearance (R_a) and disappearance (R_d) of plasma glucose was performed (22,23) as follows: R_a(t) = (F − pV × C × dE/dt)/E and R_d(t) = R_a(t) − pV × dC/dt, where R_a(t) and R_d(t) are the rates of appearance and disappearance of unlabeled glucose, respectively, as functions of time (μmol · kg⁻¹ · min⁻¹); F is the infusion rate of [6,6-²H₂]glucose (μmol · kg⁻¹ · min⁻¹); pV is the effective glucose volume of distribution (assumed to be 40 ml/kg); C is the concentration of unlabeled glucose (mmol/l); E is the isotopic enrichment (TTR); dE/dt is the rate of change of TTR; and dC/dt is the rate of change of unlabeled glucose concentration. Plasma glucose concentrations and TTR values were smoothed using a loess local polynomial smoothing function (Mathcad 11; Mathsoft Engineering and Education, Cambridge, MA) before calculations and to obtain the rates of change of TTR and concentration.

Data are expressed as the mean ± SE, except where the SD is specified. Condition- and time-related data were analyzed by mixed-model repeated-measures ANOVA. P values <0.05 were considered to indicate significant differences.

Plasma C-peptide concentrations were suppressed during octreotide (with growth hormone) infusions on all four study occasions (Table 1). Peripheral plasma insulin concentrations were suppressed during octreotide (with growth hormone) infusions in the absence of insulin replacement (e.g., 3 ± 0 to 2 ± 0 μU/ml [18 to 12 pmol/l] at 240 min) and raised approximately twofold during insulin replacement (e.g., 4 ± 1 to 8 ± 1 unit/ml [24 to 48 pmol/l] at 240 min) (P < 0.0001) (Fig. 1). Peripheral plasma glucagon concentrations were suppressed during octreotide (with growth hormone) infusions in the absence of partial glucagon replacement (e.g., 89 ± 4 to 72 ± 4 pg/ml [26 ± 1 to 21 ± 1 pmol/l] at 240 min) and not raised during partial glucagon replacement (e.g., 98 ± 4 to 83 ± 3 pg/ml [28 ± 1 to 24 ± 1 pmol/l] at 240 min) (Fig. 1). Despite growth hormone infusions with octreotide, plasma growth hormone concentrations were suppressed on all four study occasions (Table 2).

During octreotide (with growth hormone) infusion (combined insulin and glucagon deficiency) mean (±SE) plasma glucose concentrations decreased from 83 ± 1 mg/dl (4.6 ± 0.1 mmol/l) at 0 min to a nadir of 63 ± 2 mg/dl (3.5 ± 0.1 mmol/l) at 60 min and then increased progressively to 89 ± 4 mg/dl (4.9 ± 0.2 mmol/l) at 240 min (Fig. 2). During octreotide (with growth hormone) plus insulin replacement (isolated glucagon deficiency), plasma glucose concentrations decreased and remained low, reaching 55 ± 2 mg/dl (3.1 ± 0.1 mol/l) at 240 min (P < 0.0001) (Fig. 2). Compared with during octreotide (with growth hormone) alone (combined insulin and glucagon deficiency) during octreotide (with growth hormone) plus partial glucagon replacement (insulin and partial glucagon deficiency), plasma glucose concentrations increased to higher levels (P = 0.0469), reaching 103 ± 10 mg/dl (5.6 ± 0.6 mmol/l) at 240 min (Fig. 2). Compared with during octreotide (with growth hormone) plus insulin (glucagon deficiency) during octreotide (with growth hormone) plus insulin replacement plus partial glucagon replacement (partial glucagon deficiency), glucose concentrations increased to higher levels (P = 0.0090), reaching 62 ± 4 mg/dl (3.4 ± 0.2 mmol/l) at 240 min (Fig. 2).

During octreotide (with growth hormone) infusion, rates of glucose appearance (R_a) decreased initially to a nadir at 45 min and were less than rates of glucose disappearance (R_d) (Fig. 3), corresponding to the initial decrease in the plasma glucose concentrations (Fig. 2). Then glucose R_a increased, transiently exceeding glucose R_d, corresponding to the subsequent increase in the plasma glucose concentrations. During octreotide (with growth hormone) plus insulin replacement, glucose R_a decreased initially and then rose slightly but remained low (P = 0.0009) thereafter and matched glucose R_d (Fig. 3), corresponding to the low-plateau plasma glucose concentrations (Fig. 2). During octreotide (with growth hormone) plus partial glucagon replacement (without and with insulin replacement), after the initial decrease glucose R_a rose to higher rates relative to glucose R_d (Fig. 3), corresponding to the higher plasma glucose concentrations (Fig. 2); glucose R_a tended to be higher, albeit not significantly, compared with that during octreotide (with growth hormone) alone and with octreotide (with growth hormone) plus insulin replacement, respectively (Table 3).

Plasma epinephrine concentrations were not altered significantly (despite a small apparent increase at 60 min) during octreotide (with growth hormone) infusions in the absence of insulin replacement (Fig. 4). However, plasma epinephrine levels increased when insulin was replaced and hypoglycemia developed (P < 0.0001, P = 0.0025) (Fig. 4). Plasma norepinephrine concentrations also increased (P = 0.0016, P = 0.0006) under those conditions (data not shown). Plasma pancreatic polypeptide concentrations were unaltered during octreotide (with growth hormone) infusions, even when insulin was replaced and hypoglycemia developed (data not shown). Plasma cortisol concentrations declined during the studies; they appeared to be higher when insulin was replaced and hypoglycemia developed (Table 3).

Serum nonesterified fatty acid concentrations increased during octreotide (with growth hormone) infusions without insulin replacement but not with insulin replacement (P < 0.0001, P = 0.0015) (Table 4). However, nonesterified fatty acid levels were not suppressed below baseline during insulin replacement. Blood lactate concentrations tended to increase during octreotide (with growth hormone) infusions with insulin replacement (P = 0.0633) and increased significantly with insulin and partial glucagon replacement (P = 0.0368) (data not shown).

There were no significant changes in heart rate on the four study occasions (data not shown). Compared with during octreotide (with growth hormone) infusions, systolic blood pressure was slightly lower during insulin replacement (P = 0.0013) and during insulin plus partial glucagon replacement (P = 0.0156), but diastolic blood pressure did not differ (data not shown).

Given that postabsorptive glucoregulation primarily or even exclusively by insulin is plausible (2), and the increasing interest in glucagon antagonism as a potential treatment of diabetes, we tested the hypothesis that glucagon, in concert with insulin, supports the postabsorptive plasma glucose concentration in humans. To do so, we suppressed endogenous insulin and glucagon secretion (and growth hormone secretion but with growth hormone partially replaced on all occasions) with the somatostatin analog octreotide (9) over 4 h in healthy individuals and assessed the impact of superimposed infusions of insulin, glucagon, and both insulin and glucagon in doses intended to replace (insulin) or partially replace (glucagon) hepatic portal venous insulin and glucagon concentrations (10).

During combined insulin and glucagon deficiency (octreotide with growth hormone but without insulin or glucagon replacement), plasma glucose concentrations decreased initially but then increased. This previously documented biphasic glycemic pattern (3–6) suggests that glucagon is involved but that insulin is the predominant determinant of the postabsorptive plasma glucose concentration. During isolated glucagon deficiency (octreotide with growth hormone plus insulin replacement), plasma glucose concentrations decreased to hypoglycemic levels, as evidenced not only by the glucose levels but also by activation of adrenomedullary epinephrine secretion. Indeed, glucose levels would have undoubtedly fallen to lower levels were it not for activation of glucose counterregulatory systems, specifically increased epinephrine secretion (7). Since this dose of insulin, in the absence of octreotide, does not cause hypoglycemia (10), this finding provides direct evidence that glucagon supports the postabsorptive plasma glucose concentration in humans. During endogenous insulin and glucagon deficiency, partial portal glucagon replacement (octreotide with growth hormone plus partial portal glucagon replacement) raised plasma glucose concentrations to levels higher than those during combined insulin and glucagon deficiency. Similarly, during endogenous insulin and glucagon deficiency with portal insulin replacement, partial portal glucagon replacement (octreotide with growth hormone plus portal insulin replacement plus partial portal glucagon replacement) raised plasma glucose concentrations to levels higher than those during glucagon deficiency. These two findings provide additional direct evidence that glucagon supports the postabsorptive plasma glucose concentration. Thus, these data indicate that glucagon, in concert with insulin, supports the postabsorptive plasma glucose concentrations in humans.

The observed changes in the plasma glucose concentrations were the result of sequential changes in glucose production rather than in glucose utilization. Initially during endogenous insulin and glucagon deficiency, under all four study conditions, the rates of glucose appearance (R_a) decreased to those lower than the rates of glucose disappearance (R_d) and the plasma concentrations decreased. Then glucose R_a rose to different rates. It rose during combined insulin and glucagon deficiency, transiently greater than glucose R_d, resulting in an increase in plasma glucose concentrations. During isolated glucagon deficiency, glucose R_a increased to a lesser extent and was matched to glucose R_d, resulting in low plateau plasma glucose concentrations. During insulin and partial glucagon deficiency and during partial glucagon deficiency, glucose R_a increased, transiently exceeding glucose R_d, seemingly to a greater extent than during insulin and glucagon deficiency and glucagon deficiency, respectively, resulting in higher plasma glucose concentrations.

It is interesting that after an initial decrease plasma glucose concentrations increased but then appeared to plateau within the postabsorptive physiological range during 4 h of octreotide-induced suppression of insulin and glucagon secretion in these healthy individuals. Indeed, although plasma glucose concentrations have been observed to increase above fasting levels during somatostatin infusion (5,6), a 12-h somatostatin infusion did not cause sustained fasting hyperglycemia (24). Thus, in the short-term, factors in addition to these hormones must be involved in maintenance of the postabsorptive glucose level. Nonetheless, chronic insulin and glucagon deficiencies (e.g., in pancreatectomized humans who have low insulin levels and little or no circulating biologically active 3,500-Da glucagon [25–29]) cause diabetes. That indicates that among the pancreatic islet hormones, insulin plays a predominant role in maintenance of the postabsorptive plasma glucose concentration and further that elevated glucagon levels are not a requisite condition for the development of diabetes.

During octreotide infusions, the measured plasma C-peptide concentrations decreased to levels that approached the assay detection limit, indicating nearly complete suppression of insulin secretion. Interestingly, however, plasma C-peptide levels were slightly, but significantly, lower when insulin was replaced and plasma glucose concentrations remained low. That finding suggests that subphysiological plasma glucose concentrations, even those not low enough to cause symptoms of hypoglycemia, suppress insulin secretion to a greater extent than a pharmacological dose (9) of the somatostatin analog octreotide.

In contrast to the C-peptide findings, the measured plasma glucagon concentrations decreased by only 20–25% during octreotide infusions. That is not a new finding (4–6). Assuming comparable suppression of β-cell insulin and α-cell glucagon secretion by octreotide, it suggests that the antibody used to measure glucagon recognizes species in addition to biologically active 3,500-Da glucagon, a phenomenon well documented with earlier antibodies (e.g., 25–29). Clearly, however, given lower rates of glucose production and plasma glucose concentrations during octreotide infusion with insulin replacement, octreotide produced biological glucagon deficiency.

We conclude that these data provide direct evidence that glucagon, in concert with insulin, supports the postabsorptive plasma glucose concentration in humans. They do not distinguish the relative roles of insulin and of glucagon, but the fact that chronic insulin and glucagon deficiencies cause hyperglycemia (25–29) indicates that insulin is the predominant determinant of the postabsorptive glucose level.

This study was supported, in part, by National Institutes of Health Grants R37 DK27085, MO1 RR00036, P30 DK56341, and P60 DK20579 and a fellowship award from the American Diabetes Association.

The authors acknowledge the assistance of the staff at the Washington University General Clinical Research Center in the performance of this study; the technical assistance of Krishan Jethi, Gene Wade Sherrow, Michael Morris, Zina Lubovich, Sharon Travis, Sharon O’Neil, Freida Custodio, Jennifer Shew, and Dr. Adewole Okunade; and the assistance of Janet Dedeke in the preparation of this manuscript.