Novel Insights Into Effects of Cortisol and Glucagon on Nocturnal Glucose Production in Type 2 Diabetes

Abstract

Context

The effect of physiological changes in night-time cortisol and glucagon on endogenous glucose production (EGP) and nocturnal glycemia are unknown.

Objective

To determine the effects of changes in cortisol and glucagon on EGP during the night.

Design

Two overnight protocols were conducted. In Protocol 1, endogenous cortisol was blocked with metyrapone and hydrocortisone infused either at constant (constant) or increasing (variable) rates to mimic basal or physiological nocturnal cortisol concentrations. In Protocol 2, endogenous glucagon was blocked with somatostatin and exogenous glucagon was infused at either basal or elevated rates to mimic nocturnal glucagon concentrations observed in nondiabetic (ND) and type 2 diabetes (T2D) individuals. EGP was measured using [3-³H] glucose and gluconeogenesis estimated with ²H₂O in all studies.

Setting

Mayo Clinic Clinical Research Trials Unit, Rochester, MN, US.

Participants

In Protocol 1, 34 subjects (17 ND and 17 T2D) and in Protocol 2, 39 subjects (21 ND and 18 T2D) were studied.

Main Outcome Measures

Endogenous glucose production

Results

EGP, gluconeogenesis, and glycogenolysis were higher with variable than with constant cortisol at 7 am in T2D subjects. In contrast, nocturnal EGP did not differ in ND subjects between variable and constant cortisol. While elevated glucagon increased EGP, glycogenolysis, and gluconeogenesis in ND, the data in T2D subjects indicated that EGP and gluconeogenesis but not glycogenolysis were higher during the early part of the night.

Conclusion

Nocturnal hyperglucagonemia, but not physiological rise in cortisol, contributes to nocturnal hyperglycemia in T2D due to increased gluconeogenesis.

Type 2 diabetes (T2D) is characterized by fasting hyperglycemia due to elevated rates of endogenous glucose production (EGP) (1-3). The overnight postabsorptive period is a significant component of overall glycemic exposure. In a recent study of nocturnal glucose metabolism, we have demonstrated (4) elevated glucose concentrations throughout the night in T2D compared to nondiabetic (ND) subjects due to elevated EGP that, in turn, was because of higher rates of gluconeogenesis (GNG) and glycogenolysis (GGL). While circulating overnight glucagon concentrations were higher in T2D subjects, plasma cortisol concentrations rose gradually and comparably through the night in both T2D and ND subjects.

Whereas the role of cortisol (5,6) and glucagon (7-10) on fasting and postprandial glucose metabolism has been reported extensively in humans, the relative roles and contributions of abnormalities in each of these hyperglycemia causing hormones on glucose metabolism throughout the night, have not been elucidated. To do so, we designed 2 sets of experimental protocols to determine the role and effect sizes of the (i) normal nocturnal physiological rise in cortisol or (ii) persistent nocturnal hyperglucagonemia, on EGP and its components (GNG and GGL) in T2D subjects. We studied matched ND subjects under similar experimental conditions as controls.

Research Design and Methods

After approval by the Mayo Clinic Institutional Review Board and the Radioactive Drug Research Committee, subjects reported to the Mayo Clinic Clinical Research Trials Unit (CRTU) for a screen visit in the morning after an overnight fast. After written informed consent, a history and physical examination was performed and blood and urine samples drawn for analyses to ensure that subjects met enrollment criteria. Body composition was measured by Lunar iDXA software v. 15.0 (GE Healthcare Technologies, Chicago, IL, US). Subjects with sleep apnea were excluded using the Sleep Questionnaire. T2D subjects were either on mono or combination antidiabetic therapy with metformin and/or a sulfonylurea. They were asked to stop the drug(s) 2 weeks prior to the study visit and to monitor their finger stick glucoses regularly during this period. If 3 successive finger stick glucose values exceeded 300 mg/dL, then they resumed their medication and were withdrawn from the study. Subjects were asked to refrain from unaccustomed physical exercise and to maintain their weight within 2% between screen and study visits. Every subject completed 2 study visits in random order.

Subjects meeting the enrollment criteria were admitted to the CRTU at 4 pm and ate a standard 10 kcal/kg meal (50% carbohydrate, 20% protein, and 30% fat) at 5 pm. In the evening, a pulse oximeter was placed to track oxygen saturation levels throughout the night to rule out sleep apnea. The purpose of this first night was to acclimatize the subject with the CRTU environment and minimize sleep disruptions during the study period on the second night. Participants received 3 standardized caffeine-free meals at 7:00 am, 12:00 pm,and 5:00 pm, prepared by the CRTU metabolic kitchen, each of which consisted of 33% of the subject’s total estimated calorie intake based on the Harris Benedict + 20% calorie requirements. Activity for all subjects was monitored by an accelerometer, and each subject was asked to walk on a treadmill periodically at a predetermined slow speed (~1.2 miles/h) to mimic activities of daily living (11). Activity was minimized after the evening meal. The subjects then remained fasting until the end of the study.

Protocol 1 (constant vs variable cortisol)

On the second evening of each visit 2 gm of metyrapone was given orally (0.5 g every 4 h from 6 pm to 6:00 am) to block endogenous cortisol production (5,6,12). An Investigational New Drug exemption was granted by the US Food and Drug Administration for use of metyrapone in this study. Subjects consumed deuterated water (²H₂O; 5 g/kg total body water) at 6:00 pm, 8:00 pm, and 10:00 pm for estimation of GNG (4,13,14). Dose of water was split (~1.7 g/kg total body water) to avoid side effects. Subjects were permitted small sips of water containing ²H₂O upon request throughout the night.

At 8:00 pm, an arterial line was inserted (under local anesthesia and aseptic precautions) into a radial artery to obtain arterial samples periodically throughout the night for measurement of analytes (4). In all subjects, an intravenous cannula was inserted into a vein on the contralateral forearm for tracer infusions. In the evening, a pulse oximeter was placed as on day 1. An intravenous infusion of hydrocortisone was started at 10:00 pm. The order (constant vs variable cortisol) of study visit was randomized. For 1 visit, this was infused at a constant rate (~0.15 µg/kg/min) to achieve constant plasma cortisol. On the other, the rate was adjusted periodically to mimic the nocturnal rise in plasma cortisol, the variable cortisol visit (5,6). Primed continuous infusion of [3-³H] glucose was started at 10:00 pm (0.06 µci/min) and continued until the end of the study to measure glucose turnover.

Protocol 2 (basal vs elevated glucagon)

This protocol was similar to protocol 1 except, on the second night of each visit, instead of metyrapone, an intravenous infusion of somatostatin (60 ng/kg/min) and insulin (0.25 mU/kg/min) was started at 10:00 pm and continued until the end of the study. Also, instead of hydrocortisone, an intravenous infusion of glucagon was started at 10:00 pm. On 1visit, glucagon infusion rate was 0.65 ng/kg/min to achieve basal glucagon concentrations expected in the liver in people without diabetes. On the other visit, the glucagon infusion rate was 1.5 ng/kg/min to mimic the nocturnal hyperglucagonemia (elevated) observed in T2D subjects (4). Additionally, an infusion of 50% dextrose labeled with [3-³H] glucose was also started at 10:00 pm, if required, at variable rates to maintain euglycemia (~5.5 mM) during the somatostatin pancreatic clamp. This infusion was necessary predominantly in ND subjects and not in T2D subjects since the glucose concentrations during the clamp were higher than 5.5 mM in the latter.

Analytical techniques

Plasma samples were placed on ice, centrifuged at 4°C, separated, and stored at −20°C until assay. Plasma glucose concentrations were measured using a glucose oxidase method (YSI, Inc., Yellow Springs, OH, US). Plasma [3-³H] glucose specific activity was measured by liquid scintillation (15). [5-²H] glucose and [2-²H] method was estimated by the hexamethylenetetramine method (14,16,17). Insulin was analyzed using the DxI 800 (Beckman Instruments, Chaska, MN, US). C-peptide and glucagon were analyzed using a radioimmunoassay (C-peptide: Roche Diagnostics, Indianapolis, IN, US; glucagon: Millipore Corporation, Billerica, MA, US). Cortisol was measured with the competitive binding immunoenzymatic assay on the DxI 800 (Beckman Instruments).

Subject characteristics

In Protocol 1, 17 ND and 17 T2D subjects were enrolled. In Protocol 2, 21 ND and 18 T2D subjects were enrolled. ND subjects were in good health and did not have a history of diabetes in first-degree family members. T2D subjects did not have micro- or macrovascular complications apart from mild background retinopathy and had an hemoglobin A1c (HbA1c) of <9%. All subjects had normal kidney (creatinine ≤ 1.4 mg/dL in women and ≤ 1.5 mg/dL in men) and liver functions, and none of the subjects had any evidence of cardiac, renal, pulmonary, or hepatic disorders. Subjects that were taking medications that could affect glucose metabolism such as corticosteroids, tricyclic antidepressants, benzodiazepines, opiates, barbiturates, and anticoagulants or with a history or cerebrovascular disease, anemia, history of alcoholism, or substance abuse were excluded. All subjects were required to be at a stable weight for at least 2 weeks prior to the study.

Calculations

Rates of EGP were estimated at 1:00 am, 4:00 am, and 7:00 am using steady state equations (4). Rates of GNG were calculated by multiplying the plasma C5 glucose to C2 glucose ratio by EGP while rates of GGL were calculated by subtracting the rate of GNG from the rate of EGP (14,16,17).

Statistics

Due to a limitation of no prior studies examining nocturnal EGP, the sample size estimates for this set of studies considered our measurements of variability obtained during the basal period (18) because that was considered most similar to what we would expect around the nocturnal period of the current study. Based on this study, the mean EGP was expected to range from 14.5 (ND subjects) to 17.1 (T2D) μmol/kg/min with a common standard deviation of 1.9. A 15% change in EGP, which represents approximately the difference observed in ND subjects and T2D, was defined to be the minimum clinically relevant difference. Thus, our minimum clinically relevant effect size was 1.14 (=14.5*0.15/1.9) μmol/kg/min. We anticipated correlations in the statistical tests, so planned to use the effective dimensionality of the data to determine the per-comparison error rate (the overall omnibus test was α = 0.05), but for study planning purposes, we set our per-test error rate at 0.01 (to allow for 5 statistically independent tests). Seventeen participants were needed per group to obtain a power of 80% with this level of significance.

Data are described as means and standard deviation or frequency and percentage for sample characteristics. EGP was estimated as the mean of the sample obtained 10 min prior and at the top of the hour at 1:00 am, 4:00 am, and 7:00 am and is expressed per kg fat free mass per min. The primary hypothesis for protocol 1 that the EGP in T2D under the variable (rising) cortisol infusion would be different at 7:00 am than the constant cortisol infusion was tested using a mixed-model framework. For protocol 2, the main comparison was whether higher EGP was observed with elevated glucagon in comparison to basal glucagon in T2D. A model with all main effects and interactions of study cohort (ND vs T2D), protocol-specific infusion types (cortisol or glucagon infusion types), and measurement time (1:00 am, 4:00 am, and 7:00 am) was applied. Model contrasts of least squares (model-based) means were constructed to test the primary hypothesis along with secondary tests of interest (eg, various post hoc comparisons of the estimated means). This same analytical framework was used for other measurements in the study, including GNG and GGL. For the primary data analysis, all available data were used. One participant did not complete a study visit. When EGP was estimated to be negative, GNG and GGL were not estimated. As an additional exploratory analysis, we analyzed the differences in glucose levels in T2D participants within study protocols (ie, variable vs. constant cortisol; elevated vs basal glucagon) at 1:00 am, 4:00 am, and 7:00 am using paired t-tests. All P values reported are 2-sided and have not been adjusted for multiple comparisons. P values < 0.05 were considered statistically significant. Statistical analyses were conducted using R version 3.4.2.

Results

Protocol 1—cortisol

In brief 17, ND (9 men, age 61 ± 14 years and body mass index (BMI) 30 ± 2 kg/m², HbA1c 36 ± 2 mmol/mol or 5.3 ± 0.2 %) and 17 T2D (9 men, age 62 ± 20 years, BMI 31 ± 3 kg/m², HbA1c 59 ± 12 mmol/mol or 7.6 ± 1.0%, duration of T2D 11 ± 5 years) subjects completed the study (Table 1).

Table 1.

Baseline characteristics of the subjects

	Protocol 1 (Cortisol)		Protocol 2 (Glucagon)
Subject Characteristics	No Diabetes (N = 17)	Type 2 Diabetes (N = 17)	No Diabetes (N = 21)	Type 2 Diabetes (N = 18)
Age (years)	61 ± 14	62 ± 20	58 ± 8	59 ± 12
Gender (M:F)	9:8	9:8	11:10	11:7
Body mass index (kg/m2)	30 ± 2	31 ± 3	29 ± 3	31 ± 3
Fat Free Mass (kg)	55 ± 11	53 ± 9	54 ± 10	56 ± 10
HbA1c (%) mmol/mol	5.3 ± 0.2/36 ± 2	7.6 ± 1.0/59 ± 12*	5.3 ± 0.4/34 ± 4	7.2 ± 0.7/56 ± 8*
Fasting plasma glucose (mg/dL)	90 ± 5	161 ± 49*	94 ± 8	146 ± 38*

Data are mean ± standard deviation.

*P < 0.001 vs ND

Plasma glucose, insulin, C-peptide, and cortisol concentrations.

Plasma glucose concentrations were higher (P < 0.0001) in the T2D subjects (~ 12 mM) at each time point than the ND subjects (~ 5 mM) with minimal changes in either group during the night during both constant and variable cortisol visits. Plasma insulin concentrations were not different in the T2D than the ND subjects overnight in either study visit. Plasma C-peptide concentrations followed the pattern observed with plasma insulin throughout the night in both groups. In T2D subjects, glucose concentrations were higher during the variable vs constant cortisol at 4:00 am (P = 0.022) and 0700 (P = 0.001) but not at 1:00 am (P = 0.70) (Fig. 1A-D, left panel).

Figure 1.

Mean (standard deviation) concentrations of plasma (A) glucose, (B) insulin, (C) c-peptide, and (D) cortisol (left) and glucagon (right), observed between 10:00 pm and 7:00 am for ND and subjects with T2D in protocol 1 cortisol (left panel) and protocol 2 glucagon (right panel).

The variable hydrocortisone infusion resulted in a comparable gradual increase in nocturnal plasma cortisol concentrations in both T2D (11, 13, 18 µg/dL at 1:00 am, 4:00 am, and 7:00 am, respectively) and ND (12, 14, and 19 µg/dL at 1:00 am, 4:00 am, and 7:00 am, respectively) subjects during the variable cortisol visit. Similarly plasma cortisol concentrations were comparable during the constant cortisol visit in T2D (6, 7, and 7 µg/dL at 1:00 am, 4:00 am, and 7:00 am, respectively) and ND (6, 6, and 7 µg/dL at 1:00 am, 4:00 am, and 7:00 am, respectively) groups. Plasma cortisol concentrations were higher during variable vs constant cortisol study visits at 1:00 am, 4:00 am, and 7:00 am in both groups.

Plasma ACTH measured during the study were as expected in the normal range in T2D (13.2 ± 1.3 vs. 13.5 ± 1.5 pg/mL) as well as in ND (12.4 ± 1.6 vs. 12.8 ± 1.7 pg/mL) subjects during the variable vs constant visits, respectively.

Specific activity.

[3-³H] glucose tracer/tracee ratio (specific activity) remained constant in both T2D and ND subjects throughout the night permitting use of steady state equations for calculation of glucose turnover (data not shown).

Endogenous glucose production, gluconeogenesis, and glycogenolysis.

Within group comparison revealed no effects of variable vs constant cortisol concentrations on EGP in ND subjects at 1:00 am and 4:00 am. At 7:00 am, EGP was numerically but not statistically higher, and GNG was statistically higher in ND subjects. Likewise, in T2D subjects, EGP was higher during variable than during constant cortisol only at 7:00 am (due to higher rates of GNG and GGL) but not at 1:00 am or 4:00 am (Fig. 2).

Figure 2.

Box-plot graphs showing contribution of (A) Endogenous glucose production, (B) gluconeogenesis, and (C) glycogenolysis in ND and T2D subjects during constant and variable cortisol visits at 1:00, 4:00, and 7:00 am. The horizontal solid black line in the box represents the mean of the study groups.

Between group comparison demonstrated higher EGP in the T2D than the ND subjects at 1:00 am, 4:00 am, and 7:00 am during both variable and constant cortisol (P < 0.006) visits indicating that nocturnal rates of EGP were higher in T2D than ND subjects regardless of variable (ie, rising) or constant cortisol concentrations. Further analyses showed that the higher EGP in T2D at 1:00 am was due to higher rates of GGL but not GNG, while at 4:00 am and 7:00 am, higher EGP was due to higher rates of GGL and GNG during both constant and variable cortisol visits.

Protocol 2—Glucagon

In brief, 21 ND (11 men, age 58 ± 8 years and BMI 29 ± 3 kg/m², HbA1c 34 ± 4 mmol/mol or 5.3 ± 0.4 %) and 18 T2D (11 men, age 59 ± 12 years, BMI 31 ± 3 kg/m², HbA1c 56 ± 8 mmol/mol or 7.2 ± 0.7%, duration of T2D 12 ± 4 years) subjects completed the study (Table 1).

Plasma glucose, insulin, C-peptide and glucagon concentrations.

Plasma glucose concentrations were higher throughout the night in T2D than ND subjects during both basal and elevated glucagon visits. While plasma glucose concentrations in ND subjects did not differ between basal and elevated glucagon visits by design, they were higher during elevated than basal glucagon visit in T2D subjects. The glucose values in T2D at 1:00 am, 4:00 am, and 7:00 am were collectively higher under the elevated glucagon infusion protocol than the basal protocol (P ≤0.002 for each time point) (Fig. 1A-D, right panel).

Plasma insulin concentrations at 1:00 am, 4:00 am, and 7:00 am did not differ between T2D and ND subjects, nor between basal and elevated glucagon study visits. C-peptide concentrations were comparably suppressed in both groups during all study visits implying comparable suppression of pancreatic β-cell function by somatostatin. By design, plasma glucagon concentrations were higher during elevated than basal glucagon visits in both groups. However, glucagon concentrations were comparable between ND and T2D subjects during basal and elevated glucagon visits.

Rates of glucose infusion.

The glucose infusion rates required to maintain euglycemia were lower at 7:00 am for T2D vs ND in basal (1.1 ± 1.4 vs 3.2 ± 1.6 µmol/kg/min; P < 0.0006) as well as elevated (0.4 ± 1.0 vs 2.8 ± 2.0 µmol/kg/min; P < 0.0003) glucagon visits, indicating insulin resistance in T2D. Glucose infusion rates also were lower throughout the night at 1:00 am (basal: 0.13 ± 0.5 vs 0.87 ± 1.0 µmol/kg/min; P < 0.0001; elevated: 0.02 ± 0.1 vs 0.32 ± 0.6 µmol/kg/min; P < 0.0001) than at 4:00 am (basal: 0.34 ± 0.7 vs 1.9 ± 1.8 µmol/kg/min; P < 0.0001; elevated: 0.06 ± 0.2 vs 1.6 ± 1.6 µmol/kg/min; P < 0.0001) in T2D when compared to ND subjects (Fig. 3).

Figure 3.

This figure shows rates of glucose infusion required to maintain euglycemia in protocol 2. Fifty percent dextrose was infused starting at 10:00 amto maintain plasma glucose ~95 mg/dL. The glucose infusion rates were significantly higher in ND as compared to T2D during both basal and elevated glucagon visits at 1:00, 4:00, and 7:00 am.

Endogenous glucose production, gluconeogenesis, and glycogenolysis

In contrast to protocol 1, within-group comparison revealed higher rates of EGP with elevated than basal glucagon in ND at 1:00 am, 4:00 am, and 7:00 am and at 1:00 am and 4:00 am but not at 7:00 am in T2D subjects. In ND subjects, this was due to higher rates of GGL (at 1:00 am, 4:00 am, and 7:00 am) and GNG (at 4:00 am and 7:00 am ) while in T2D subjects, this was because of higher rates of GNG (1:00 am, 4:00 am, and 7:00 am) but not GGL (Fig. 4).

Figure 4.

Box-plot graphs showing contribution of (A) endogenous glucose production, (B) gluconeogenesis, and (C) glycogenolysis in ND and T2D subjects during basal and elevated glucagon visits at 1:00, 4:00, and 7:00 am. The horizontal solid black line in the box represents the mean of the study groups.

Discussion

The current protocols were designed to demonstrate separately the relative contributions of early morning physiological rise in cortisol or mild nocturnal hyperglucagonemia (as observed in T2D from our prior study (4) on plasma glucose concentrations, rates of EGP, GNG, and GGL throughout the night and early morning in T2D individuals. The results show that the physiological rise in cortisol, compared to basal cortisol, increased glucose concentrations slightly but significantly at 0700 in T2D subjects. It did so by increasing EGP via increased GNG, and GGL. However, the physiological rise in cortisol did not have significant effects on any of these parameters in ND subjects throughout the night or early morning. In contrast, mild nocturnal hyperglucagonemia, which was experimentally created in protocol 2 to match what we have previously observed in the overnight state in people with T2D (4) resulted in higher nocturnal glucose concentrations in T2D. This was due to higher EGP in T2D as a result of increased GNG but not GGL. In contrast, in ND subjects, comparable mild hyperglucagonemia raised EGP throughout the night via increased GNG and GGL. Taken together, the data suggest mild nocturnal hyperglucagonemia, rather than a physiological rise in early morning cortisol, has a greater contribution to night-time glucose control in T2D.

Glucocorticoids play an important role in the regulation of carbohydrate metabolism. While glucocorticoid excess results in hyperglycemia and insulin resistance, its deficiency is associated with increased insulin sensitivity (19). Chronic hypercortisolemia in the ranges of 30 to 40 μg/dL, as observed during chronic stress or illnesses, impair both insulin suppression of EGP (by stimulating GNG) and insulin enhancement of glucose uptake (20,21). Under conditions of daily living, cortisol concentrations rise during sleep to ~15 to 20 μg/dL in the early morning. This physiological rise in cortisol leads to fasting and post-breakfast hyperglycemia due to both impaired suppression of EGP and reduced glucose uptake in healthy and type 1 diabetes individuals (5,6). Our current study extends the prior observations by further defining the effects of gradual rise in nocturnal cortisol concentrations on glucose turnover throughout the night. As we have recently reported (4), cortisol concentrations rise gradually and continuously but comparably in T2D and ND subjects from 10 pm to 7:00 am. Metyrapone was utilized to block endogenous cortisol secretion and then hydrocortisone infused at a variable rate (5,6) to mimic as close as possible the normal nocturnal pattern of cortisol concentrations observed in T2D and ND subjects (4). By design, there were clear differences in plasma cortisol concentrations between the basal and variable cortisol study visits in both groups but matched cortisol concentrations between ND and T2D. However, glucose concentrations were only slightly higher in T2D subjects during variable than basal cortisol at 4:00 am and 7:00 am when the physiological rise of cortisol was created with the hydrocortisone infusion. T2D subjects had higher rates of EGP than ND subjects throughout the night during both cortisol study visits. During the variable cortisol visit at 1:00 am, this was due to higher rates of GGL but not GNG, while at 4:00 am and 7:00 am, this was due to higher rates of both GNG and GGL in T2D than in ND subjects. Taken together, the results imply that the normal nocturnal rise in cortisol concentrations has minimal impact on overnight hepatic glucose metabolism in T2D and in mitigating the inherent impairments in hepatic glucose production, when compared to anthropometrically matched ND subjects. The effect size of inhibiting nocturnal cortisol concentration (from ~17.6 μg/dL to ~7.0 μg/dL at 7:00 am) observed in T2D subjects when EGP was lowered via reduced rates of GNG and GGL translated to a difference in plasma glucose concentrations of ~1.1 mM (12.6 vs 11.5 mM: variable vs constant) or ~ 20 mg/dL implying an HbA1c reduction of approximately 0.5% over 3 to 4 months if this average difference was maintained.

On the other hand, lowering of plasma glucagon concentrations overnight, from mild hyperglucagonemia (~150 pg/mL; observed in T2D subjects) (4), to basal concentrations (~90 pg/mL; observed in ND subjects), resulted in lowering of plasma glucose concentrations throughout the night between 2 and 3 mM (36-54 mg/dL), which, in turn, implies an HbA1c reduction of ~1% to 1.5% over 3 to 4 months if this average difference is maintained. This effect was achieved by a reduction in overnight rates of EGP via lowered rates of GNG but not GGL in T2D subjects. In ND subjects, plasma glucose concentrations were comparable between elevated and basal glucagon study visits by design since exogenous glucose was infused during both study visits to maintain euglycemia. However, as anticipated, the glucose infusion rates required to maintain euglycemia were higher during the basal than during the elevated glucagon study visit, implying that elevated glucagon concentrations induced hepatic insulin resistance in the ND individuals. Raising overnight glucagon concentrations in ND subjects to those frequently observed in T2D subjects increased EGP throughout the night via increased GGL and GNG. The majority of T2D subjects did not require exogenous dextrose overnight as their plasma glucose concentrations were already elevated. Furthermore, glucose concentrations in T2D subjects did not return to euglycemic range despite lowering overnight glucagon concentrations during the basal glucagon visit. Of note, the peripheral glucagon concentrations achieved during the overnight period in protocol 2 in the presence of endogenous blockade were higher than the peripheral concentrations observed in our prior study (4) conducted without somatostatin-induced pancreatic hormone blockade. However, as have been elegantly demonstrated in the canine model by Cherrington et al (22-24), the liver is normally exposed to ~2-fold higher insulin and glucagon concentrations than the peripheral tissues, referred to as the portal-peripheral gradient, under conditions of daily living. However, unlike in the canine model, it is not possible to cannulate the portal venous system in humans. Hence, extrapolating from the canine data and our prior study (4), we believe that both the insulin and glucagon concentrations achieved in protocol 2 are representative of anticipated portal concentrations during the normal nocturnal period. Taken together, the results indicate that inhibiting the normal nocturnal rise in cortisol concentrations had a modest impact on the differences in glucose concentrations in T2D when compared to the effect observed with glucagon manipulation.

Acute hyperglucagonemia increases EGP primarily through stimulation of GGL. However, this effect is transient. Chronic hyperglucagonemia lasting ~13 h results in sustained increase in EGP in lean ND humans, as long as insulin secretion is inhibited (25). Hyperglucagonemia stimulates GNG when complemented by factor(s) that concomitantly increase(s) gluconeogenic substrate delivery to the liver (eg, relative or absolute lack of insulin leading to increased lipolysis and proteolysis). However, these observations have to be considered in perspective since, to the best of our knowledge, unlike our current study, direct measurement of rates of GNG were not undertaken in any of these prior reports. In contrast to ND subjects where both GNG and GGL were higher with elevated rather than basal glucagon, in T2D subjects, GNG but not GGL were higher through the night. This observation is intriguing given the more prominent and stated role of glucagon on GGL rather than on GNG (26). A possible explanation could be related to lower postprandial hepatic glycogen reserves in T2D than in ND subjects as has been confirmed using nuclear magnetic resonance methods (27). Mechanistically, this has been, at least in part, explained by reduced rates of glycogen synthesis through the direct uridine 5′-diphosphate–glucose pathway because of a functional defect in hepatic glucokinase activity in T2D subjects (28,29). Hence, it may be reasonable to assume that in T2D subjects lower hepatic glycogen content would lead to an earlier depletion of hepatic glycogen reserves as the night progressed. Consequently, lack of substrate (viz, hepatic glycogen) translated to lack of additional contribution of GGL to EGP even when glucagon concentrations were elevated. Interestingly, rates of GNG in ND subjects were also higher with elevated than basal glucagon throughout the night implying a robust effect of glucagon generally on stimulation of GNG in humans with and without T2D, especially in the postabsorptive phase than has been previously recognized.

Like all studies, these studies have limitations. While we directly measured EGP and GNG by established methods, rates of GGL were calculated. However, to the best of our knowledge, these studies are the first in humans to directly estimate the effects of physiological changes in cortisol and glucagon concentrations on EGP and its components throughout the night. We did not measure hepatic glycogen content that would have provided additional information on the hormonal effects on net nocturnal glycogen balance. Future studies are needed to investigate these further. We did not conduct additional experiments in protocol 1 without cortisol replacement nor in protocol 2 without glucagon replacement. Besides adding a third study visit for both protocols thus increasing the experimental burden on our study volunteers, we reasoned that in protocol 1, conducting such an experiment in the presence of absolute cortisol deficiency due to metyrapone would pose significant risk of hypotension, nausea, and vomiting, especially in those with T2D. We elected not to conduct additional visit without glucagon replacement in protocol 2 since absolute nocturnal hypoglucagonemia is not germane to conditions of daily living. Furthermore, careful and elegant studies by Cherrington et al in the canine model (23) have clearly demonstrated direct and indirect effects of insulin per se on hepatic glucose metabolism.

Lowering postprandial hyperglucagonemia with consequent improvement in postprandial glucose control has been achieved clinically in T2D and type 1 diabetes with use of glucagon-like peptide-1 receptor agonists. Recent proof of concept studies in humans have also demonstrated beneficial effects of glucagon receptor blockade on fasting glucose concentrations and HbA1c (30-32). Observations from the current study provide the scientific rationale that modulation of nocturnal glucagon effects should favorably impact glucose control in individuals with T2D.

Additional Information

Disclosure Summary: The authors have nothing to disclose.

Data Availability: The datasets generated during and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.