Effect of adrenal medullectomy on metabolic responses to chronic intermittent hypoxia in the frequently sampled intravenous glucose tolerance test

To our knowledge, this is the first study examining the role of adrenal catecholamines in glucose metabolism during intermittent hypoxia (IH) in unanesthetized unrestrained C57BL/6J mice. We report that IH did not affect fasting glucose and insulin levels nor insulin sensitivity and insulin secretion during, whereas glucose effectiveness was decreased. Adrenal medullectomy decreased fasting blood glucose and insulin levels in mice exposed to IH but had no effect on glucose metabolism, insulin secretion, and insulin sensitivity.

Abstract

Obstructive sleep apnea is associated with type 2 diabetes. We have previously developed a mouse model of intermittent hypoxia (IH) mimicking oxyhemoglobin desaturations in patients with sleep apnea and have shown that IH increases fasting glucose, hepatic glucose output, and plasma catecholamines. We hypothesize that adrenal medulla modulates glucose responses to IH and that such responses can be prevented by adrenal medullectomy. We performed adrenal medullectomy or sham surgery in lean C57BL/6J mice, which were exposed to IH or intermittent air (control) for 4 wk followed by the frequently sampled intravenous glucose tolerance test (FSIVGTT) in unanesthetized unrestrained animals. IH was administered during the 12-h light phase (9 AM to 9 PM) by decreasing inspired oxygen from 21 to 6.5% 60 cycles/h. Insulin sensitivity (S_I), insulin independent glucose disposal [glucose effectiveness (S_G)], and the insulin response to glucose (AIR_G) were determined using the minimal model method. In contrast to our previous data obtained in restrained mice, IH did not affect fasting blood glucose and plasma insulin levels in sham-operated mice. IH significantly decreased S_G but did not affect S_I and AIR_G. Adrenal medullectomy decreased fasting blood glucose and plasma insulin levels and increased glycogen synthesis in the liver in hypoxic mice but did not have a significant effect on the FSIVGTT metrics. We conclude that, in the absence of restraints, IH has no effect on glucose metabolism in lean mice with exception of decreased S_G, whereas adrenal medullectomy decreases fasting glucose and insulin levels in the IH environment.

NEW & NOTEWORTHY To our knowledge, this is the first study examining the role of adrenal catecholamines in glucose metabolism during intermittent hypoxia (IH) in unanesthetized unrestrained C57BL/6J mice. We report that IH did not affect fasting glucose and insulin levels nor insulin sensitivity and insulin secretion during, whereas glucose effectiveness was decreased. Adrenal medullectomy decreased fasting blood glucose and insulin levels in mice exposed to IH but had no effect on glucose metabolism, insulin secretion, and insulin sensitivity.

intermittent hypoxia (IH) is a cardinal manifestation of obstructive sleep apnea (OSA) (12). OSA leads to high cardiovascular morbidity and mortality (22, 23, 33, 49), which has been attributed to metabolic dysfunction (9, 10, 23, 25, 33, 49). OSA is independently associated with insulin resistance and glucose intolerance (26, 31, 32, 34, 35), which are abated by treatment with continuous positive airway pressure (27). Short-term exposure of healthy volunteers to IH also leads to insulin resistance and glucose intolerance suggesting IH as a major contributing factor (21).

A mouse model of IH mimicking the oxygen profile of patients with OSA has been developed (13, 30, 36). In this model, IH has been shown to induce fasting hyperglycemia (13, 30, 43, 44), glucose intolerance, and insulin resistance and to impair non-insulin-mediated glucose uptake and insulin secretion (17). Chronic IH exposure significantly affected glucose metabolism in the liver increasing hepatic glucose output (44). The hepatic glucose output is regulated by both insulin-dependent and insulin-independent mechanisms and depends on hepatic gluconeogenesis and metabolism of glycogen. The IH-induced increase in hepatic output has been attributed to upregulation of gluconeogenesis (29, 44).

The effects of IH on glucose metabolism have been linked to activation of the sympathetic nervous system (SNS), which augments the hepatic glucose output, and increased production of adrenal catecholamines suppressing insulin secretion by pancreatic β-cells (43, 44). However, the effects of IH on metabolism of glycogen, which is strongly influenced by the sympathetic nervous system and catecholamines (4, 5, 11, 37), are not known. The effects of the sympathetic nervous system and adrenal catecholamines were previously assessed by blood glucose testing from the tail of unanesthetized animals, while plasma insulin levels were obtained under anesthesia at the time of death (43). Both techniques could stress animals influencing the results.

In the current study we examined effects of adrenal medullectomy on glucose metabolism during chronic IH by performing the frequent sampling intravenous glucose tolerance test (FSIVGTT) in conscious unrestrained mice. The FSIVGTT with minimal model analyses allows an assessment of insulin-dependent [insulin sensitivity (S_I)] and insulin-independent [glucose-effectiveness (S_G)] glucose disposal, as well as the acute insulin response to glucose (AIR_G) and the disposition index (DI), the product of AIR_G and S_I and a strong predictor of progression of glucose intolerance in humans (4). The effects of non-IH-related stress were prevented by obtaining blood samples from indwelling catheters and by minimizing blood loss by reinfusing red blood cells. In addition, we measured glycogen levels in the liver and assessed the effect of IH and adrenal medullectomy on key enzymes of hepatic glycogen metabolism, glucokinase (GK) and glycogen synthase kinase (GSK) (1, 15, 16, 24). We hypothesized that chronic IH will induce fasting hyperglycemia, insulin resistance, and glucose intolerance and suppress insulin secretion and that the effects of IH will be prevented by adrenal medullectomy.

MATERIALS AND METHODS

Animals.

Thirty-eight adult male C57BL/6J mice, 6–8 wk of age were procured from the Jackson Laboratory (Bar Harbor, ME) and housed in a 22°C laboratory with a 12-h light-dark cycle (light phase 9 AM to 9 PM). Mice underwent adrenal medullectomy (n = 19) or sham surgery (n = 19). After a 2-wk recovery, animals were exposed to chronic intermittent hypoxia (IH) or intermittent air (IA) for 4 wk while fed a regular chow diet. The left femoral arterial and venous lines were implanted on week 4 of exposure under 1–2% isoflurane anesthesia and the FSIVGTT was performed after 72-h recovery in conscious unrestrained mice followed by death. The study was approved by the Johns Hopkins University Animal Use and Care Committee (Institutional Animal Care and Use Committee Protocol MO12M309) and complied with the American Physiological Society Guidelines for Animal Studies.

Adrenal medullectomy.

Adrenal medullectomy was performed as previously described (43). Briefly, mice were anesthetized with 1–2% isoflurane, shaved, and prepped with chlorhexidine. A 1-cm dorsal midline incision was performed between the first and third lumbar vertebra. The left adrenal gland was located lateral and cranial to the spleen and the right adrenal gland was located cranial to the right kidney. The adrenal glands were exteriorized. Small incisions were made on the adrenal capsule bilaterally and the medulla was gently squeezed out. The adrenal capsule and attached fat pads were returned to the abdominal cavity and the skin incision was closed. Burpenorphine (0.01 mg/kg) was administered subcutaneously at the end of surgery to minimize discomfort. Sham surgery was performed in an identical fashion, except that adrenal medulla was not removed.

Intermittent hypoxia.

IH was performed as previously described (43, 44). Briefly, a gas control delivery system was designed employing programmable solenoids and flow regulators, which controlled the flow of air, nitrogen, and oxygen into cages. During each cycle of intermittent hypoxia, the percentage of O₂ decreased from ∼21 to ∼6–7% over a 30 s period, followed by a rapid return to ∼21% over the subsequent 30-s period. We have previously shown that this regimen of IH induces oxyhemoglobin desaturations from 99 to ∼65%, 60 times/h (14, 36). A control group was exposed to an identical regimen of IA delivered at the same flow rate as IH. All animals had free access to water. The IH animals had free access to food. The IA group was weight matched to the IH group by varying food intake as previously described (14). IH and IA were administered during the light phase (9 AM to 9 PM) for 4 wk.

Frequently sampled intravenous glucose tolerance test.

The FSIVGTT was performed in conscious mice during week 4 of exposure to IH or IA as previously described (4, 17). Briefly, under 1–2% isoflurane anesthesia catheters (MRE025; Braintree Scientific) were chronically implanted in the left femoral artery and vein. The catheters were perfused throughout the recovery period by an infusion pump with a sterile saline solution containing heparin (20 U/ml). Animals were allowed 72 h to recover from the surgery. IH or IA exposures were continued during recovery and the FSIVGTT. On the day of the FSIVGTT animals were fasted from 8 AM and the test was performed between 1 and 3 PM A glucose bolus of 1 g/kg D-50 was given over a 15-s period through the venous catheter. Glucose levels were measured in whole blood sampled from the arterial line using a handheld glucometer (Accu-Check Aviva; Roche, Indianapolis, IN) 10 min and immediately before the glucose bolus (times: -10 and 0 min) and then at 1, 2, 3, 4, 5, 6, 8, 10, 12, 14, 16, 18, 20, 25, 30, and 60 min. Arterial blood was collected for plasma insulin levels at −10, 0, 1, 2, 4, 8, 12, 16, 20, 30, and 60 min. Red blood cells were reinfused to prevent blood loss. Plasma was separated and immediately frozen at -80°C and later tested for insulin levels with an ELISA kit from Millipore (Billerica).

Minimal model.

Blood glucose and plasma insulin curves from FSIVGTT were analyzed in the minimal model according to Bergman et al. (6) with modifications for the faster glucose disposal in mice compared with humans as previously described (4, 17). Specifically, the mixing phase was defined at 0–2 min; the AIR_G was calculated based on insulin values between 0 and 4 min; S_G was modeled between 2 and 5 min (4).

Enzyme activities.

Hexokinase and glucokinase activity measurements were performed as previously described (47). Briefly, 30~50 mg freeze-clamped livers were homogenized in 1 ml buffer containing 50 mmol/l HEPES, 100 mmol/l KCl, 1 mmol/l EDTA, 5 mmol/l MgCl₂, and 2.5 mmol/l dithioerythritol. Homogenates were centrifuged at 100,000 g for 45 min to sediment the microsomal fraction. The enzymatic activities were assayed in the postmicrosomal fraction with no added glucose (0 mM glucose), 0.5 mM glucose, 8 mM glucose, and 100 mM glucose in the presence of phosphatase inhibitors. K_m values for glucose are ~0.03 mM for hexokinase I (HKI), 0.2–0.3 mM for HKII, 0.003 mM for HKIII, and 7–8 mM for GK. Therefore, total HK, HKI, HKII, and HKIII activity was measured as glucose phosphorylation rate in the presence of 0.5 mM glucose. V_0.5 and V_max of GK activity were determined as the glucose phosphorylation rate in the presence of 8 and 100 mM glucose minus the glucose phosphorylation rate in the presence of 0.5 mM glucose, respectively.

Real-time PCR, immunoblot, and glycogen.

Total RNA was extracted from liver using Trizol (Life Technologies, Rockville, MD) and cDNA was synthesized using Advantage RT for PCR kit from Clontech (Palo Alto, CA). Real-time reverse-transcriptase PCR (RT-PCR) was performed with primers from Invitrogen (Carlsbad, CA) and Taqman probes from Applied Biosystems (Foster City, CA). The sequences of primers and probes for mouse 18S were previously described (19). Mouse phosphoenolpyruvate carboxykinase (PEPCK), glucose 6-phosphatase (G-6-Pase), HKI-III, and glucokinase mRNA were been measured with the Applied Biosystems premade primers and probes. The mRNA expression levels were referenced to 18S rRNA and the values were derived according to the 2^−ΔΔCt method (40).

GSK3β phosphorylation was measured in liver tissue total lysate by Western blot using GSK3β and p-GSK3β rabbit monoclonal antibodies from Cell Signaling Technology (Danvers, MA), and goat anti-rabbit-HRP conjugate from Bio-Rad (Hercules, CA). Glyceraldehyde 3-phosphate dehydrogenase (GAPDH; loading control) was detected with mouse monoclonal antibodies from Sigma (St. Louis, MO) and goat anti-mouse HRP conjugate from Bio-Rad. SDS-PAGE and Western blot were performed using Bio-Rad precast gel system. Twenty micrograms protein were applied per lane. Liver glycogen was measured using a kit from Biovision (Milpitas, CA).

Statistical analysis.

All values are reported as means ± SE unless otherwise specified. All the data in the study were checked for normality with a χ²-goodness of fit test or ladder of powers (48) if the indexes were estimated from the FSIGTT data. Statistical significance for all comparisons between normally distributed values was determined by two-way ANOVA test with Bonferroni post hoc correction for multiple comparisons. Statistical significance for all comparisons between nonnormally distributed values was determined by Kruskal-Wallis and Mann-Whitney tests. P < 0.05 was considered significant. For indexes estimated from the FSIVGTT data, we conducted robust regression where each of the indexes were considered to be associated with the interaction between groups (medullectomy and sham groups) and treatment (air or intermittent hypoxia). Post hoc, we conducted pairwise comparisons of differences across all levels of factor variables with and without adjustments for changes in body weight. The FSIVGTT data were analyzed with and without adjustments for fasting blood glucose levels. Statistical analysis was performed using Minitab 16 (State College, PA) and STATA 14 (College Station, TX).

RESULTS

All mice were of the same age and similar body weight before exposure (Table 1). IH resulted in weight loss in both sham-operated and medullectomized mice, which was significantly greater than in the air exposure groups. Despite our attempts to weight match controls by food restriction, the IA mice consumed slightly more food than the IH animals (3.3% more in sham-operated mice and 6.7% more in medullectomized mice, Table 1). We have previously shown in this model that adrenal medullectomy in our hands effectively depletes adrenal catecholamines without any effect on corticosterone (43).

Table 1.

Basic characteristics of C56BL/6J mice

	Sham Surgery		Adrenal Medullectomy
	IA	IH	IA	IH
Number of mice	8	11	9	10
Age, wk
Day 0	8.3 ± 0.5	8.4 ± 0.4	8.5 ± 0.4	8.7 ± 0.4
Day 28	12.3 ± 0.4	12.4 ± 0.4	12.5 ± 0.4	12.7 ± 0.4
Weight, g
Day 0	22.8 ± 0.45	23.4 ± 0.8	23.4 ± 0.6	24.1 ± 0.7
Day 28	21.1 ± 0.7	21.2 ± 0.6	22.4 ± 0.7	21.6 ± 0.5
Weight loss, g	1.7 ± 0.3	2.2 ± 0.4*	1.1 ± 0.3	2.5 ± 0.3*
Food intake, g/day	3.0 ± 0.1	2.9 ± 0.1*	3.0 ± 0.1	2.8 ± 0.03*

Values are means ± SE. IA, intermittent air; IH, intermittent hypoxia.

^*P < 0.05 compared with IA.

IH significantly decreased fasting blood glucose in both sham-operated and medullectomized mice compared with IA mice (P = 0.03, Fig. 1A), but this effect was no longer present after adjustment for weight loss during the exposure. IH decreased fasting serum insulin levels in medullectomized mice and this effect also was no longer significant after weight loss adjustment (Fig. 1B). Mice subjected to adrenal medullectomy showed lower fasting blood glucose and serum insulin levels during IH compared with sham-operated mice (Fig. 1), independent of changes in body weight.

Fig. 1.

Fasting blood glucose (A) and fasting serum insulin levels (B). *P < 0.05 for the effect of hypoxia.

IH induced glucose intolerance in FSIVGTT (Fig. 2, A and B). Blood glucose levels throughout FSIVGTT were significantly higher in IH-exposed mice, compared with normoxia, regardless of the presence of adrenal medulla. After adjustment for fasting blood glucose levels, there was a clear separation of the glucose profiles with higher glucose levels in IH mice throughout the FSIVGTT time course (Fig. 2B). Later in the time course of FSIVGTT, medullectomized mice exposed to IH demonstrated significantly lower absolute blood glucose levels compared with sham-hypoxia mice, but this difference disappeared after adjustment for the lower baseline glucose levels (Fig. 2, A and B). The glucose profiles in normoxic mice were identical between sham-operated and medullectomized groups. The insulin profiles were similar between all four groups of mice (Fig. 2C). The minimal model analysis showed that overall IH decreased glucose effectiveness (ΔS_G, P = 0.024) and trended to increase the area under the curve for glucose (AUC_G) levels (P = 0.058) in both sham-operated and medullectomized mice (Fig. 3). Medullectomized mice had significantly lower glucose effectiveness during IH compared with IA conditions (ΔS_G, P = 0.001). IH did not affect insulin sensitivity (S_I), acute insulin response to glucose (AIR_G), the disposition index (DI), and area under the curve for insulin (AUC_I) levels (Fig. 3). Adjustments for body weight differences and weight loss during exposure did not modify the results of the minimal model analysis. Adjustments for fasting blood glucose showed that IH significantly increased AUC_G (P = 0.02), whereas all other comparisons were not affected.

Fig. 2.

Absolute blood glucose levels (A), change in blood glucose levels from baseline (time 0, B), and plasma insulin levels (C) during the frequently sampled intravenous glucose tolerance test (FSIVGTT). At time 0 an intravenous injection of 1 g/kg of D50 was administered over a 15-s interval and blood glucose and plasma insulin were sampled over a 1-h period. P < 0.05 for the effect of adrenal medullectomy and intermittent hypoxia on glucose levels; P < 0.001 for the effect of intermittent hypoxia on change in glucose levels from baseline.

Fig. 3.

The minimal model of FSIVGTT. The effect of intermittent hypoxia and adrenal medullectomy on insulin sensitivity (S_I), glucose effectiveness (S_G), the acute insulin response to glucose (AIR_G), the disposition index (DI), the areas under curve for blood glucose level adjusted for the baseline (AUC_G) and plasma insulin levels adjusted for the baseline (AUC_I). *P < 0.05 for the overall effect of intermittent hypoxia; †P < 0.001 for the effect of intermittent hypoxia in medullectomized mice.

To explore mechanisms by which adrenal catecholamines affect glucose metabolism during IH we measured hepatic gene expression of key enzymes of gluconeogenesis G-6-Pase and PEPCK as well as gene expression and enzymatic activity of HKI-III and GK. IH did not affect GK mRNA levels, whereas adrenal medullectomy increased GK gene expression in the liver (Fig. 4). Neither IH nor adrenal medullectomy affected HKI-III, G-6-Pase, and PEPCK gene expression levels (not shown). Hepatic GK enzymatic activity was increased at both Km concentration of glucose of 8 mmol (V_0.5) and at a high glucose concentration of 100 mmol (V_max) in medullectomized mice exposed to IH but not to IA (Fig. 5). Neither IH nor medullectomy affected HKI-III activity (Fig. 5).

Fig. 4.

Glucokinase gene expression in the liver.

Fig. 5.

Liver hexokinase and glucokinase activities. A: V_max activity of glucokinase was measured at 100 mM of glucose. B: V_0.5 activity of glucokinase was measured at Km concentration of glucose (8 mM). C: a sum of hexokinases I, II, and III activities was measured at glucose concentration of 0.5 mM.

IH did not affect liver glycogen levels in sham-operated mice (Fig. 6A). In contrast, IH significantly increased liver glycogen content in medullectomized mice (P < 0.05 for the interaction between medullectomy and IH). IH did not affect GSK phosphorylation (Fig. 6B), which suggests that glycogen synthase activity was unchanged. Adrenal medullectomy significantly increased GSK phosphorylation, which indicates that GSK activity decreased while glycogen synthase activity increased (42).

Fig. 6.

A: liver glycogen levels and phosphorylation of the β-subunit of glycogen synthase kinase 3 (GSK3β). B: representative immunoblots of phosphorylated GSK3β (p-GSK3β), total GSK3β and a housekeeping protein glyceraldehyde 3-phosphate dehydrogenase (GAPDH) run on the same membrane. C: optical density ratio of p-GSK3β to total GSK3β.

DISCUSSION

To our knowledge, this is the first study examining the role of adrenal catecholamines in glucose metabolism during IH in unanesthetized unrestrained C57BL/6J mice. We report that IH did not have an independent effect on fasting glucose and insulin levels nor insulin sensitivity and insulin secretion during the FSIVGTT, whereas glucose effectiveness was decreased. The main finding of the study was that adrenal medullectomy decreased fasting blood glucose and insulin levels in mice exposed to IH but had no effect on glucose metabolism, insulin secretion, and insulin sensitivity during the FSIVGTT. In addition, adrenal medullectomy significantly increased liver glycogen levels and hepatic GK activity during IH.

The effect of IH and adrenal medullectomy on fasting glucose and insulin levels.

In contrast to our previous reports (39, 43), IH did not raise fasting blood glucose levels in sham-operated animals (Table 2). The main difference between our current study and previous experiments was that mice were unrestrained and unhandled during glucose testing with blood drawn from the arterial line. Furthermore, blood loss and related stress was prevented by reinfusion of red blood cells. It is likely that IH and handling interact to increase catecholamine levels leading to hyperglycemia. Notably, IH did not affect fasting blood glucose and insulin levels in the previous report in catheterized unrestrained mice (17).

Table 2.

Comparison of the effects of IH and adrenal medullectomy on glucose and insulin according to Shin et al. (43) and the current study

	Shin et al. (43)		Current Study
	IH	MED	IH	MED
Mouse restraints	Yes	Yes	No	No
Mouse handling	Yes	Yes	No	No
Fasting blood glucose	↑	=	=	↓
Fasting serum insulin	=	↑	=	↓
Glucose tolerance test	Intraperitoneal	Intraperitoneal	Intravenous	Intravenous
Insulin sensitivity (SI)	↓	=	=	=
Glucose effectiveness (SG)	NA	NA	↓	=
AIRG	↓	↑	=	=
AUCG	=	↓	↑ =	=

AIR_G, acute insulin response to glucose; AUC_G, glucose area under the curve; IH, intermittent hypoxia; MED, adrenal medullectomy; NA, nonapplicable.

We have previously reported that adrenal medullectomy dramatically decreases adrenal catecholamine levels abolishing the IH-induced increases in plasma epinephrine, whereas corticosterone production remains intact (43). Our current data demonstrate that depletion of adrenal catecholamines significantly decreases fasting blood glucose and plasma insulin levels in hypoxic mice. These data differ from our previous report which failed to find an effect of adrenal medullectomy on fasting glucose levels (Table 2). However, experiments were performed in restrained animals (43) and it is conceivable that hepatic sympathetic innervation compensated for the lack of plasma epinephrine at the stressful conditions. Given that the blood glucose level in the fasting state is largely determined by hepatic glucose output (38, 46), we speculate that removal of adrenal medulla decreases hepatic glucose output at hypoxic conditions.

Improvement in control of fasting glucose in medullectomized mice exposed to IH could be either due to decreased gluconeogenesis or decreased glycogen degradation. We have previously reported upregulation of PEPCK, a key enzyme of gluconeogenesis, under IH condition (29, 44), which was not observed in the present study. We speculate that these differences are primarily related to significant differences in body weight and age of experimental animals. Mice were ~20% heavier in the previous reports (29, 44) and obesity significantly augments metabolic responses to IH (7). Our current data demonstrate that IH increased liver glycogen level in medullectomized mice in association with increased activity of an important enzyme of glycogen synthesis, hepatic GK. Adrenal medullectomy increased GK gene expression in both hypoxic and normoxic mice, but increased activity was observed only in hypoxic mice. Thus adrenal medullectomy increases liver GK activity during IH.

GK catalyzes the first reaction of glucose metabolism converting glucose to G-6-phosphate, which can ether enter glycolysis or the glycogen synthesis pathway (1). Hepatic GK is chronically regulated by insulin at the transcriptional level (1). It is conceivable that improved insulin sensitivity in hypoxic medullectomized mice at fasting conditions resulted in upregulation of GK gene expression (45). Acute regulation of GK activity occurs posttranscriptionally by glucokinase regulatory protein (GKRP). Increased intracellular glucose causes the release of glucokinase from the nuclear GCK-GKRP complex allowing glucokinase to be transported into the cytoplasm where it regulates glucose phosphorylation (1, 24, 45). Given relatively low fasting concentrations of glucose, GK is usually inactive at fasting conditions. However, in hypoxic medullectomized mice GK activity was increased even at K_m concentration of 8 mM of glucose, which is equivalent to 144 mg/dl and not very different from fasting blood glucose levels. It is conceivable that improved hepatic insulin sensitivity in hypoxic medullectomized mice results in accelerated glucose intracellular transport in hepatocytes, GK activation, and glycogen synthesis.

Adrenal medullectomy also increased phosphorylation of GSK suppressing its activity and consequently increasing activity of glycogen synthase. Adrenal medullectomy increased GSK phosphorylation (i.e., glycogen synthesis) in both IH-exposed and normoxic mice. Thus combined upregulation of GK and glycogen synthase could accelerate glycogen synthesis and decrease fasting blood glucose levels in medullectomized mice during IH.

The effect of IH and adrenal medullectomy on glucose and insulin levels during FSIVGTT.

IH for 4 wk resulted in glucose intolerance and decreased glucose effectiveness in FSIVGTT in mice, which is consistent with previous observations (17). However, postprandial (i.e., postglucose bolus) insulin secretion and insulin resistance were not affected. This apparent discrepancy with previous FSIVGTT and hypersulinemic euglycemic clamp data may be related to the fact that the mice were significantly leaner in the current study, which can markedly influence metabolic effects of IH (8).

We have previously evaluated the effect of adrenal medullectomy on glucose and insulin metabolism using the intraperitoneal glucose tolerance and insulin tolerance tests (43). In contrast to our previous observations (43), we did not observe suppression of the insulin response to glucose during IH and adrenal medullectomy did not augment basal and glucose-stimulated insulin secretion in IH-exposed mice in the present study (Table 2). Although glucose and insulin responses to the intraperitoneal and intravenous glucose loads are different (43), stress caused by physical restraint and handling in the previous study could be an important factor explaining differences in metabolic outcomes. This high level of stress might have suppressed insulin secretion (2, 18, 20). We hypothesize that a relatively low level of stress in unrestrained unhandled mice in the current study negated any effect of adrenal medullectomy (Table 2). Thus our data suggest that, in unrestrained lean mice, IH leads to impaired utilization of the glucose load via mechanisms independent of insulin and adrenal catecholamines.

Two major insulin independent mechanisms of glucose utilizations are 1) glucose regulated liver GK activity, and 2) glucose uptake in the brain. If low glucose effectiveness during IH was caused by decreased glucose utilization in the liver, there would be a decrease in hepatic GK activity. However, IH increased GK activity in medullectomized mice. Our findings imply that impairment in glucose effectiveness during IH was extrahepatic. Severe hypoxia decreases glucose uptake by the brain (41). IH reduces the bioavailability of nitric oxide in the cerebral circulation and blunts vasodilatory responses to hypoxia, which may suppress cerebral glucose uptake (28). Given that the IH exposure paradigm in our study results in severe tissue hypoxia (3, 36), it is conceivable that impaired cerebral glucose uptake is implicated in disturbances of glucose metabolism during FSIVGTT.

Conclusions and implications.

In conclusion, our data demonstrated that, in the absence of physical restraints and handling, chronic IH does not induce significant changes in glucose metabolism in lean mice with exception of a decrease in glucose effectiveness. Decreased glucose effectiveness appears to be related to extrahepatic insulin independent mechanisms of glucose utilization, possibly in the brain. Adrenal medullectomy lowers fasting blood glucose levels in hypoxic mice by upregulating glycogen synthesis in the liver but does not have a significant impact on glucose utilization. Our data emphasize limitations of commonly done glucose testing in restrained animals for IH research and confirm previous observations that leanness may protect against detrimental effects of IH on glucose metabolism.

GRANTS

This work was supported by the National Heart, Lung, and Blood Institute Grants R01-HL-080105, R01-HL-128970 and R01-HL-133100, and the American Sleep Medicine Foundation Grant 133-BS-15 (to V. Y. Polotsky) and by the American Heart Association Grant 12POST11820001 (to M.-K. Shin).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

M.-K.S., M.S., D.S., and V.Y.P. conceived and designed research; M.-K.S., W.H., H.J., S.B.-F., and M.S. performed experiments; M.-K.S., W.H., H.J., S.B.-F., M.S., D.S., and V.Y.P. analyzed data; M.-K.S., M.S., D.S., and V.Y.P. interpreted results of experiments; M.-K.S. and V.Y.P. prepared figures; M.-K.S., M.S., D.S., and V.Y.P. edited and revised manuscript; V.Y.P. drafted manuscript; V.Y.P. approved final version of manuscript.