The Tissue Concentration of UDP-N-acetylglucosamine Modulates the Stimulatory Effect of Insulin on Skeletal Muscle Glucose Uptake*

(Received for publication, October 7, 1996, and in revised form, November 22, 1996)

Meredith Hawkins Dagger , Ivo Angelov , Rong Liu , Nir Barzilai § and Luciano Rossetti

From the Division of Endocrinology and the Diabetes Research and Training Center, Albert Einstein College of Medicine, Bronx, New York 10461

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
Acknowledgments
REFERENCES


ABSTRACT

To delineate the biochemical mechanism by which increased availability of GlcN impairs insulin action on skeletal muscle glucose uptake, we replenished the uridine pool during GlcN administration. Co-infusion of uridine with GlcN prevented the GlcN-induced fall in skeletal muscle UDP-glucose levels (24.9 ± 5.3 versus 10.1 ± 2.9 nmol/g; p < 0.01) and further increased the skeletal muscle UDP-GlcNAc levels (198.4 ± 26.3 versus 96.0 ± 8.4 nmol/g; p < 0.01). Greater reductions in the rates of glucose infusion (~53%), glucose uptake (~43%), and glycogen synthesis (~60%) were observed with the addition of uridine. Similarly, the infusion of uridine alone markedly increased the skeletal muscle levels of both UDP-glucose (55.2 ± 14.2 versus 17.8 ± 6.1 nmol/g; p < 0.01) and UDP-GlcNAc (86.8 ± 8.8 versus 35.9 ± 8.4 nmol/g; p < 0.05) and induced marked insulin resistance. The decrease in insulin action on peripheral glucose uptake was highly correlated with the increase in skeletal muscle UDP-GlcNAc levels. Finally, immunoisolation of GLUT4-containing vesicles revealed that the rate of labeled GlcN incorporation was ~100-fold greater following GlcN compared with saline infusions (p < 0.01). We suggest that the marked reduction in insulin action induced by GlcN and uridine is mediated by increased accumulation of muscle UDP-N-acetylhexosamines, perhaps via altered glycosylation of protein(s) in GLUT4-containing vesicles.


INTRODUCTION

A decrease in the ability of insulin to stimulate the uptake of glucose in skeletal muscle is a major feature of non-insulin-dependent diabetes mellitus (NIDDM)1 (1). Recent evidence obtained by 31P NMR spectroscopy of skeletal muscle glucose 6-phosphate (2) and by modeling of forearm glucose transport and uptake (3) has revealed that this defect in insulin action in subjects with NIDDM is due to decreased muscle glucose transport/phosphorylation. Glucose transport in skeletal muscle occurs by facilitated diffusion using the insulin-regulated type of glucose transporter, GLUT4 (4-7). Since skeletal muscle GLUT4 gene expression is similar in subjects with NIDDM and nondiabetic controls (8, 9), defective translocation or activation of GLUT4 in response to insulin has been hypothesized (10). However, the mechanism(s) responsible for this impairment in the glucose transport system remains to be delineated.

In primary culture of rat adipocytes, Marshall et al. (10, 11) demonstrated that increased glucose flux into the hexosamine biosynthetic pathway is the mechanism by which prolonged exposure to high levels of glucose and insulin resulted in impaired glucose transport. Additional in vitro and in vivo studies support the hypothesis that recruitment of this pathway is a major mechanism by which hyperglycemia causes insulin resistance (12-15). Recently, we infused GlcN in conscious rats at a rate designed to reproduce the flux of glucose into the hexosamine pathway during hyperglycemia (16). This was associated with increased accumulation of skeletal muscle UDP-N-acetylhexosamines, marked time-dependent impairments in the ability of insulin to promote peripheral glucose uptake and glycogen synthesis, and a pronounced decline in skeletal muscle UDP-Glc concentrations. More important, Baron et al. (17) demonstrated that the marked impairment in insulin-mediated glucose disposal induced by increased GlcN availability was associated with decreased muscle GLUT4 translocation to the sarcolemmal fraction in response to insulin. Similarly, hyperglycemia resulted in the accumulation of skeletal muscle hexosamine metabolites, decreased skeletal muscle UDP-hexose (UDP-Glc and UDP-Gal) levels, and increased UDP-hexosamine/UDP-hexose ratios (18). Thus, it has been hypothesized that this quantitatively small pathway of glucose utilization may link the rate of glucose flux to the activity of the glucose transport system by functioning as a "glucose sensor" capable of decreasing the rate of glucose uptake when detecting an increased availability of hexose phosphates (11, 19, 20).

The rate-limiting step for the entry of glucose carbons into this pathway is the conversion of fructose 6-phosphate to glucosamine 6-phosphate by the enzyme glutamine:fructose-6-phosphate aminotransferase. Downstream metabolic products are formed by subsequent acetylation and uridylation. The end product of this pathway (UDP-GlcNAc) serves as substrate in the biosynthesis of glycoproteins, and its accumulation provides an index of the amount of carbon flux through this pathway (Fig. 1A). Consumption of UTP in the synthesis of hexosamine end products causes depletion of the intracellular uridine pool (21-23). More important, the decreased cellular availability of UTP is expected to limit the synthesis of both UDP-glucose and UDP-GlcNAc while favoring the accumulation of intermediate products of the hexosamine biosynthetic pathway (22). The above effects may modulate the action of insulin on glucose metabolism. As an example, the decrease in UDP-Glc available as substrate for glycogen formation may contribute to the GlcN-induced impairment in insulin action on glucose uptake and glycogen synthesis (16). Alternatively, uridine availability may alter the balance between intermediate and end products within the glucosamine pathway (22); hence, depletion of intracellular uridine may either limit or potentiate the metabolic effects of GlcN administration on insulin action.


Fig. 1. A, role of uridine in the glycogenic and hexosamine biosynthetic pathways. Both glucose and GlcN enter skeletal muscle and adipose cells via the glucose transport system. Glucose is rapidly phosphorylated to Glc-6-P, thus becoming available for glycogen synthesis and glycolysis. The hexosamine pathway receives ~1-3% of incoming glucose via the conversion of Fru-6-P (F-6P) to glucosamine 6-phosphate by the rate-limiting enzyme glutamine:fructose-6-phosphate aminotransferase. Alternatively, infused GlcN is directly phosphorylated to GlcN-6-P. Intermediate metabolic products such as GlcNAc result from subsequent acetylation. The uridine (UDP/UTP) pool contributes to the uridylation of both glucose (to UDP-glucose) and the GlcN metabolites. The end product of the glucosamine pathway (UDP-GlcNAc) serves as substrate in glycoprotein biosynthesis, and its accumulation provides an index of the amount of carbon flux through this pathway. B, schematic representation of the experimental protocol used to assess the effects of uridine repletion on the ability of GlcN to induce insulin resistance in normal rats. Paired euglycemic insulin (18 milliunits/kg/min, ~500 microunits/ml) clamps were performed at 0-2 and 5-7 h during 7-h continuous infusions of GlcN (30 µmol/kg/min) or saline (1.2 ml/h) in conscious nondiabetic rats. Infusions of equimolar uridine (-GlcN-U, n = 4; +GlcN-U, n = 6) or saline (-GlcN-S, n = 5; +GlcN-S, n = 10) were maintained throughout the 7 h. Variable infusion rates of a 25% glucose solution were used to maintain base-line blood glucose levels throughout the studies. Tritiated glucose was infused during both study intervals to measure peripheral glucose uptake. [14C]Glucosamine infusions were added at t = 5 h in the +GlcN-S and -GlcN-S groups.
[View Larger Version of this Image (50K GIF file)]


In this study, we attempt to identify those metabolic/cellular effects of a prolonged increase in GlcN availability that are required for the induction of skeletal muscle insulin resistance. This study provides evidence that the inhibitory effects of GlcN on insulin-mediated glucose uptake are due to increased accumulation of UDP-GlcNAc and UDP-GalNAc, the end products of the GlcN pathway, rather than depletion of UDP-glucose.


EXPERIMENTAL PROCEDURES

Animals

Twenty-five normal male Sprague-Dawley rats (Charles River Laboratories, Wilmington, MA) were housed in individual cages and subjected to a standard light (6 a.m. to 6 p.m.)-dark (6 p.m. to 6 a.m.) cycle. The rats were anesthetized with intraperitoneal injection of pentobarbital (50 mg/kg of body weight), and indwelling catheters were inserted into the right internal jugular vein and in the left carotid artery as described previously (24-27). The venous catheter was extended to the level of the right atrium, and the arterial catheter was advanced to the level of the aortic arch. The in vivo studies were performed 5-7 days following catheter placement.

Euglycemic Clamp Study

Studies were performed in awake, unstressed, normoglycemic 6-h fasted rats using the euglycemic clamp technique in combination with [3-3H]glucose infusion as described previously (24-27). In the current protocol (Fig. 1B), primed continuous infusions of glucosamine (30 µmol/kg/min; n = 16) or saline (1.2 ml/h; n = 9) were administered for 7 h (16). Paired euglycemic insulin (18 milliunits/kg/min) clamp studies in combination with [3-3H]glucose were performed in all animals during the first 2 h (t = 0-2 h) and the final 2 h (t = 5-7 h) of either glucosamine (+GlcN) or saline (-GlcN) infusions. Beginning 1 h prior to the onset of the study (t= -1 h), the animals received 8-h infusions either of equimolar uridine (+GlcN-U, n = 6; -GlcN-U, n = 4) or of saline (+GlcN-S, n = 10; -GlcN-S, n = 5). A primed continuous infusion of HPLC purified [3-3H]glucose (8-µCi bolus, 0.4 µCi/min; DuPont NEN) was initiated at t = 0 and 5 h and maintained during both 2-h clamps. To estimate the rate of incorporation of the infused GlcN into GLUT4-containing vesicles, [14C]glucosamine (8-µCi bolus, 0.4 µCi/min of infusion; DuPont NEN) was also infused from 5 to 7 h in a subgroup of +GlcN-S and -GlcN-S rats. A variable infusion of 25% glucose solution was started at time 0 and adjusted every 10 min during the clamp intervals, maintaining basal plasma glucose concentrations.

Plasma samples for determination of [3H]glucose specific activity were obtained at 10-min intervals throughout the insulin infusions. Samples for measurement of plasma insulin concentration were obtained at t = 0, 60, 120, 300, 360, and 420 min. The total volume of blood sampled was ~5.0 ml/study; to prevent volume depletion and anemia, a solution (1:1, v/v) of ~6.0 ml of fresh blood (obtained by heart puncture from a littermate of the test animal) and heparinized saline (10 units/ml) was infused. At the end of the in vivo studies, rats were anesthetized (60 mg of pentobarbital/kg of body weight, intravenously); the abdomen was quickly opened; and rectus abdominal muscle was freeze-clamped in situ with aluminum tongs precooled in liquid nitrogen (16, 26). The time from injection of the anesthetic to freeze-clamping of the muscle was <30 s. All tissue samples were stored at -80 °C for subsequent analysis. The study protocol was reviewed and approved by the Institutional Animal Care and Use Committees of the Albert Einstein College of Medicine.

Whole Body Glycolysis

The rates of glycolysis were estimated as described previously (26). Briefly, plasma-tritiated water specific activity was determined by liquid scintillation counting of the protein-free supernatant (Somogyi filtrate) before and after evaporation to dryness. Since tritium at C-3 of glucose is lost to water during glycolysis, it can be assumed that plasma tritium is present either in the form of tritiated water or [3-3H]glucose (28, 29).

Glycogen Formation in Vivo and Glycogen Synthase Activity

Muscle glycogen synthesis was estimated by subtracting the glycolytic rate from the glucose uptake (26). Muscle glycogen concentration was determined following digestion with amyloglucosidase as described previously (25-27). Muscle glycogen synthase activity was measured by a modification (25, 27) of the method of Thomas et al. (30) and is based on the measurement of the incorporation of radioactivity into glycogen from UDP-[U-14C]glucose.

Immunoisolation of GLUT4-containing Vesicles

Frozen muscles were weighed and pulverized and then homogenized in ice-cold HES buffer (20 mM HEPES, 1 mM EDTA, and 250 mM sucrose, pH 7.4) using a two-step process as described by Rodnick et al. (31). A membrane fraction was prepared by three sequential centrifugations at 3000 × g with extraction of the pellet, and then all supernatants were combined and centrifuged at 180,000 × g for 90 min. The resulting high speed pellet was resuspended in HES buffer by vertical pipetting 100 times and then was stored at 4 °C. This membrane fraction was used for subsequent immunoisolation and gel filtration of GLUT4-containing vesicles. Magnetic beads (6.8 × 108/ml) with sheep anti-mouse IgG bound to the surface were pelleted by magnet, and the supernatant was discarded. Nonspecific binding was blocked by incubating the beads for 20 min at 22 °C with 10 mg/ml BSA (in PBS with 1 mM EDTA), followed by two washes with PBS at 4 °C for 20 min each. Affinity-purified monoclonal anti-GLUT4 antibody (1F8; Charles River Pharmservices, Southbridge, MA) was added to the beads (31, 32), which were incubated at 4 °C for 14 h in PBS with 1 mg/ml BSA, 1 mM EDTA, and 0.02% sodium azide and then washed three times with PBS/BSA (1 mg/ml) at 4 °C. 250 µl of 180,000 × g muscle pellet suspension (100 µg of protein) was added to the beads in the presence of PBS/BSA (1 mg/ml) and incubated for 4.5 h at 4 °C, with the total volume being 1 ml. Beads were washed twice with PBS/BSA (1 mg/ml), and the pellet (containing magnetic beads and pelleted membranes) was resuspended in 100 µl of PBS. Samples of the suspension were pipetted onto Whatman No. 3M paper. The levels of radioactivity from [14C]GlcN in the membrane fraction were determined by liquid scintillation counting of the paper after evaporation to dryness. The nonspecific radioactivity (i.e. dpm not due to oligosaccharide side chains) was estimated by treatment of the immunoisolated vesicles with endoglycosidase F. Specifically, the isolated membranes were incubated with or without 1 unit of endoglycosidase F for 17 h at 37 °C. After incubation, vesicles were immunoprecipitated as described above, and radioactivity was measured. The radioactivity due to [3H]glucose entrapment within the vesicles was not significantly diminished by the endoglycosidase treatment, while >90% of the radioactivity derived from [14C]GlcN was removed by the procedure. Finally, to verify specific recovery of GLUT4 protein in the GLUT4-containing vesicles, 30-µl samples of isolated membranes from the -GlcN-S (n = 3) and +GlcN-S (n = 4) groups were separated by nonreducing 10% polyacrylamide gel electrophoresis and immunoblotted with saturating concentrations of anti-GLUT4 antibody before and after immunoisolation.

The rate of incorporation of [14C]glucosamine from the UDP-GlcNAc pool into oligosaccharide chains of GLUT4-containing vesicles was calculated as the 14C content in immunoabsorbed vesicles (in dpm/100 mg of tissue) divided by the skeletal muscle specific activity (in dpm/nmol) of the precursor pool, i.e. UDP-GlcNAc, upon separation by HPLC.

Analytical Methods

Plasma glucose was measured by the glucose oxidase method (Glucose Analyzer II, Beckman Instruments), and plasma insulin was measured by radioimmunoassay using rat and porcine insulin standards. Regression analysis of the slopes of the rate of appearance of 3H2O (used in the calculation of the rates of glycolysis) was performed at 60-min intervals throughout the study. Glucose 6-phosphate concentrations were measured spectrophotometrically (33).

Muscle UDP-Glc, UDP-Gal, UDP-GlcNAc, and UDP-GalNAc concentrations were obtained through two sequential chromatographic separations and UV detection (27, 34, 35). Plasma GlcN concentrations were determined by HPLC following quantitative derivatization with phenyl isothiocyanate as described by Anumula and Taylor (36).

All values are presented as the mean ± S.E. Comparisons between groups were made using repeated measures analysis of variance where appropriate. Where F-ratios were significant, further comparisons were made using Student's t tests (paired difference test and small-sample test for independent samples). Regression analysis of linear correlation was performed using the Statview statistical program for Macintosh.


RESULTS

General Characteristics of the Animals

At the time of the study, the mean body weights of the animals in the +GlcN groups were similar to those in the saline-infused (-GlcN) groups (326 ± 12 versus 294 ± 20 g; p > 0.05). The mean plasma glucose concentrations at base line were 7.0 ± 0.2 mmol/liter in the +GlcN groups and 7.1 ± 0.3 mmol/liter in the -GlcN groups (p > 0.1).

Plasma Glucose, Glucosamine, and Insulin Concentrations

The infusion of glucosamine (30 µmol/kg/min) over the 7-h studies increased the plasma glucosamine concentration by ~100-fold in both +GlcN groups (to 1.9 ± 0.2 mM). The plasma glucose concentration was maintained at the basal level during all the insulin clamp studies. The plasma insulin levels were maintained constant at ~570 nM throughout the 7 h in both study groups. The coefficients of variation in plasma glucose and insulin levels were <5 and 10%, respectively, in all studies.

Skeletal Muscle UDP-GlcNAc, UDP-GalNAc, UDP-Glc, and UDP-Gal Concentrations

The co-infusion of uridine during prolonged GlcN administration prevented the depletion in the skeletal muscle concentrations of UDP-Glc and UDP-Gal that was observed in +GlcN-S. Uridine infusion in the absence of GlcN elevated UDP-Glc levels by ~2.5-fold. While 7-h GlcN infusion resulted in ~2.5-fold elevation in skeletal muscle UDP-GlcNAc concentrations in the +GlcN groups (+GlcN-U and +GlcN-S) relative to their saline controls, there were significant and substantial increases in the skeletal muscle concentrations of both UDP-GlcNAc and UDP-GalNAc in the presence of uridine repletion (+GlcN-U and -GlcN-U) (Table I).

Table I.

Skeletal muscle UDP-Glc, UDP-Gal, UDP-GlcNAc, and UDP-GalNAc concentrations at the end of paired euglycemic hyperinsulinemic clamp studies in 6-h fasted rats in the presence of 7-h glucosamine (+GlcN) or saline (-GlcN) infusions with (-GlcN-U, n = 4; +GlcN-U, n = 6) and without (-GlcN-S, n = 5; +GlcN-S, n = 5) repletion of uridine

Values are means ± S.E.
Uridine Saline
GlcN
 - +  - +

UDP-Glc (nmol/g) 17.8  ± 6.1 55.2  ± 14.2a 10.1  ± 2.9a 24.9  ± 5.3b
UDP-Gal (nmol/g) 5.2  ± 0.5 16.3  ± 3.2c 4.1  ± 0.8 8.7  ± 1.2d
UDP-GlcNAc (nmol/g) 35.9  ± 9.7 84.8  ± 8.8a 94.7  ± 9.7a 198.4  ± 26.3b
UDP-GalNAc (nmol/g) 9.4  ± 2.3 21.3  ± 1.6a 13.8  ± 3.0 27.0  ± 4.1d
UDP-GlcNAc/UDP-Glc 2.1  ± 0.5 1.8  ± 0.3 12.3  ± 3.6c 9.4  ± 1.6

a  p < 0.05 versus -GlcN-S.
b  p < 0.01 versus +GlcN-S.
c  p < 0.025 versus -GlcN-S.
d  p < 0.025 versus +GlcN-S.

In Vivo Metabolic Parameters

Infusions of GlcN and uridine in normal rats resulted in significant decreases in insulin-mediated glucose disposal relative to each respective saline-infused control group. Infusion of GlcN and uridine together appeared to have additive effects on in vivo metabolic parameters, just as they had on skeletal muscle UDP-GlcNAc concentrations. The rates of glucose infusion (5-7 h = 16.3 ± 2.6 versus 0-2 h = 34.7 ± 2.6 mg/kg/min; p < 0.005) and the rates of tissue glucose uptake thus decreased to the greatest degree in the +GlcN-U group (Fig. 2A and Table II).


Fig. 2. A, percent decrease in glucose infusion rate (GIR) for paired euglycemic hyperinsulinemic clamp studies (18 milliunits/kg/min) during the first 2 h (0-2 h) and the last 2 h (5-7 h) of 7-h glucosamine (+GlcN) or saline (-GlcN) infusions with (-GlcN-U and +GlcN-U) and without (-GlcN-S and +GlcN-S) equimolar uridine infusions. B, skeletal muscle Glc-6-P concentrations at the end of the four experimental protocols described above (-GlcN-S, -GlcN-U, +GlcN-S, and +GlcN-U). *, p < 0.05 versus -GlcN-S; §, p < 0.005 versus -GlcN-S; dagger , p < 0.005 versus +GlcN-S.
[View Larger Version of this Image (14K GIF file)]


Table II.

Rates of glucose uptake, glycolysis, and glycogen synthesis during the paired insulin clamp studies performed during the first and last 2 h of 7-h saline and GlcN infusions with or without concomitant uridine infusion (±uridine) and the percent decrease in these parameters from the initial to the final 2 h


Uridine Saline
GlcN
 - +  - +

mg/kg/min
Glucose uptake
  0-2 h 43.2  ± 2.6 45.4  ± 1.8 45.8  ± 2.3 44.4  ± 1.5
  5-7 h 38.2  ± 2.3 27.9  ± 0.9a,b 32.4  ± 1.7c 25.5  ± 2.2b,c
  % decrease 12  ± 1 39  ± 5d 29  ± 3b 43  ± 5e
Glycolysis
  0-2 h 19.2  ± 1.9 17.6  ± 1.8 16.8  ± 1.1 13.8  ± 1.3
  5-7 h 17.2  ± 1.3 15.6  ± 0.6 16.9  ± 1.3 12.6  ± 1.1
  % decrease 10  ± 3 11  ± 4 0  ± 5 9  ± 4
Glycogen synthesis
  0-2 h 23.9  ± 0.8 27.8  ± 0.4d 29.0  ± 1.6d 30.5  ± 1.0
  5-7 h 21.0  ± 1.6 12.3  ± 0.6a,b 15.4  ± 2.1c 12.2  ± 1.5a
  % decrease 12  ± 5 56  ± 2d 47  ± 5b 60  ± 5e

a  p < 0.005 versus 0-2 h.
b  p < 0.05 versus -GlcN-S.
c  p < 0.05 versus 0-2 h.
d  p < 0.005 versus -GlcN-S.
e  p < 0.025 versus +GlcN-S.

Fig. 2A depicts the percent decrease in the rate of glucose infusion from the initial 2 h to the final 2 h of 7-h GlcN (+GlcN) or saline (-GlcN) infusion with (+GlcN-U and -GlcN-U) or without (+GlcN-S and -GlcN-S) uridine repletion. The glucose infusion rate exhibited a 53 ± 6% decrease in the +GlcN-U group as compared with a 25 ± 4% decrease with GlcN alone (+GlcN-S: 5-7 h = 28.4 ± 1.8 versus 0-2 h = 38.2 ± 2.6 mg/kg/min; p < 0.005 versus +GlcN-U). Similarly, glucose uptake decreased by 43 ± 5% with both GlcN and uridine infusion (+GlcN-U), but only by 29 ± 3% in the +GlcN-S group (p < 0.025). Infusion of uridine alone (-GlcN-U) also resulted in significant reductions in glucose infusion rate (5-7 h = 24.8 ± 1.8 versus 0-2 h = 37.4 ± 2.4 mg/kg/min; p < 0.025) (Fig. 2A) and skeletal muscle glucose uptake (Table II).

The observed decreases in glucose infusion rate and in peripheral glucose uptake were largely due to marked reductions in the rate of glycogen synthesis, which also decreased to the greatest extent in the +GlcN-U group (Table II). It is noteworthy that the initial rates (0-2 h) of glycogen synthesis were higher with the addition of uridine. There was also an increase in the initial rate of glycogen synthesis with increased GlcN availability. The rates of glycolysis were not significantly altered between the two study intervals.

Fig. 3 depicts the significant correlation between skeletal muscle UDP-GlcNAc concentrations and the percent decrease in glucose infusion rate observed from the initial 2 h to the final 2 h of 7-h GlcN or saline infusions, with and without uridine repletion (r2 = 0.758; p < 0.005).


Fig. 3. Percent decrease in glucose infusion rate plotted as a function of skeletal muscle UDP-GlcNAc concentrations following 7-h GlcN or saline infusions with paired euglycemic insulin clamp studies with (+GlcN-U, n = 6; -GlcN-U, n = 4) and without (+GlcN-S, n = 10; -GlcN-S, n = 5) uridine repletion. GIR, glucose infusion rate.
[View Larger Version of this Image (11K GIF file)]


Skeletal Muscle Substrate Concentrations

Muscle Glycogen Concentrations

There was a greater accumulation of skeletal muscle glycogen at the end of the 7-h GlcN infusions with the addition of uridine (+GlcN-U = 19.9 ± 0.4 mg/g) compared with the infusion of saline (+GlcN-S = 16.2 ± 0.6 mg/g; p < 0.05). Uridine infusion over the 7-h time interval modestly increased total glycogen concentration in the saline groups as well (-GlcN-U = 21.3 ± 0.4 versus 18.5 ± 1.7 mg/g, wet weight; p < 0.05).

Muscle Glucose 6-Phosphate Concentration

Skeletal muscle glucose 6-phosphate concentrations were similarly and significantly decreased following 7-h GlcN and uridine infusions. There was a further reduction in Glc-6-P levels in the +GlcN-U group, apparently reflecting the proportionally greater decrease in glucose uptake (Fig. 2B).

Kinetic Analysis of Muscle Glycogen Synthase

Most of the short-term effects of insulin on skeletal muscle glycogen synthase activity are due to dephosphorylation of the enzyme. This in turn results in heightened affinity of the enzyme for its substrate, i.e. decreased Km for UDP-glucose. Following GlcN infusion alone (+GlcN-S), insulin caused normal activation of the enzyme with decreased Km. However, no significant decrease in the Km for UDP-glucose occurred in the uridine-treated rats; thus, insulin was unable to normally activate glycogen synthase in either the -GlcN-U or +GlcN-U group (+GlcN-U Km = 0.27 ± 0.05 versus +GlcN-S Km = 0.17 ± 0.02 mM, p < 0.05; -GlcN-U Km = 0.30 ± 0.05 versus -GlcN-S Km = 0.15 ± 0.02 mM, p < 0.01).

Immunoisolation of GLUT4-containing Vesicles

Fig. 4A contrasts two representative GLUT4 immunoblots before (lane B) and after (lane A) immunoprecipitation of skeletal muscle membranes. The intensity of the lane A bands suggests the almost complete recovery of GLUT4 protein in the GLUT4-containing vesicles. Fig. 4B demonstrates that ~100-fold greater amounts of [14C]glucosamine were incorporated from the UDP-GlcNAc pool into the GLUT4-containing vesicles following GlcN relative to saline infusion (+GlcN-S versus -GlcN-S; p < 0.01).


Fig. 4. Immunoisolation of GLUT4-containing vesicles from skeletal muscle. A, 30 µl of nonreducing 10% polyacrylamide gels were immunoblotted with anti-GLUT4 antibody before (lane B) and after (lane A) immunoprecipitation with saturating concentrations of anti-GLUT4 monoclonal antibody. B, shown is the amount of incorporation of [14C]glucosamine into GlcNAc in the membranes of the GLUT4-containing vesicles following 7-h infusions of GlcN (+GlcN) or saline (-GlcN) during paired insulin clamp studies (expressed in nM [14C]glucosamine/100 mg of tissue). The nonspecific radioactivity (i.e. dpm not due to oligosaccharide side chains) was estimated by treatment of the immunoisolated vesicles with endoglycosidase F (+EG-F). Specifically, the isolated membranes were incubated with or without 1 unit of endoglycosidase-F for 17 h at 37 °C. After incubation, vesicles were immunoprecipitated as described above, and radioactivity was measured. The radioactivity due to [3H]glucose entrapment within the vesicles was not significantly diminished by the endoglycosidase treatment, while >90% of the radioactivity derived from [14C]GlcN was removed by the procedure. *, p < 0.01 versus -GlcN-S.
[View Larger Version of this Image (14K GIF file)]



DISCUSSION

This study demonstrates that uridine repletion enhanced the formation of uridylated end products of the hexosamine biosynthetic pathway and resulted in substantial resistance to the metabolic effects of insulin. While infusion of GlcN resulted in significant depletion of skeletal muscle UDP-Glc, simultaneous and equimolar uridine infusions completely normalized the intracellular UDP-Glc pool. Substantial and comparable decreases in skeletal muscle glucose uptake and glycogen synthesis occurred following 7-h continuous infusions of both GlcN (+GlcN-S) and uridine (-GlcN-U) relative to saline (-GlcN-S), while GlcN and uridine infusions together (+GlcN-U) had considerably more effect on these parameters. In fact, the accumulation of skeletal muscle hexosamines (UDP-GlcNAc and UDP-GalNAc) was significantly and similarly increased by ~3-fold in the presence of either GlcN or uridine infusions alone, while the infusion of GlcN together with uridine repletion resulted in ~6-fold increased accumulation. Thus, GlcN and uridine administration appeared to have additive effects, both on the accumulation of end products of the glucosamine pathway and on the induction of insulin resistance. Indeed, there was a highly significant correlation between skeletal muscle UDP-GlcNAc levels and the percent reduction in glucose infusion rate and glucose uptake. More important, increasing the levels of UDP-GlcNAc by infusion of uridine in the absence of increased GlcN availability and UDP-Glc depletion was equally effective in inducing skeletal muscle insulin resistance.

Formation of glucose 6-phosphate in skeletal muscle decreased with glucosamine infusion, both in the uridine-treated and the saline-infused rats. The greatest effect (lowest Glc-6-P levels) was observed in the +GlcN-U group, in which the decline in peripheral glucose uptake was also greatest. As noted in previous studies in which continuous glucosamine infusion resulted in reduced insulin-mediated glucose uptake with decreased Glc-6-P levels (16, 37), this indicates that the rate-limiting defect in glucose uptake with increased GlcN availability must be either glucose transport or phosphorylation. Conversely, while glycogen synthesis was significantly decreased in both groups, the lack of accumulation of Glc-6-P revealed that this was not the key regulatory site for the decrease in glucose uptake. This finding is consistent with the previous observations that increased GlcN availability was associated with decreased GLUT4 translocation (13, 17).

Intracellular uridine availability may regulate both glycogen storage, through the formation of UDP-Glc, and the synthesis of such vital cellular products as glycoproteins via its incorporation into the UDP-hexosamines. Consequently, there is an important interrelationship between the intracellular uridine (UDP/UTP) pool and the hexosamine biosynthetic pathway. This connection has been extensively explored in oncology literature, examining the effects of uridine depletion and repletion on the efficacy of GlcN as a chemotherapeutic agent. A number of in vitro and in vivo studies showed that GlcN administration either to cultured carcinoma cells or to conscious rats resulted in depletion of the intracellular UTP pool (21, 23) and actually limited the accumulation of UDP-GlcNAc and other end products of the hexosamine pathway. Additionally, in the presence of uridine depletion, there was increased formation of such intermediate products of the hexosamine pathway as N-acetylglucosamine 1-phosphate and N-acetylglucosamine 6-phosphate (22). However, the subsequent addition of uridine resulted both in depletion (i.e. normalization) of the concentrations of these hexosamine intermediates (22) and in increased formation of the UDP-N-acetylhexosamine end products (21-23, 38, 39). Finally, it was noted that the amount of UDP-GlcNAc accumulation was associated with the functional efficacy of GlcN administration, i.e. with its antitumor activity (38).

In this study, we generated marked alterations in the muscle concentration of key metabolites in the GlcN and glycogenic pathways to gain insight into the cellular mechanism(s) by which increased flux through the GlcN pathway causes insulin resistance. Consistent with the above observations, substantially higher skeletal muscle UDP-N-acetylhexosamine levels (UDP-GlcNAc and UDP-GalNAc) were measured in both experimental groups receiving uridine repletion relative to their controls. The increased accumulation of UDP-GalNAc with uridine repletion was particularly noteworthy in the absence of GlcN, indicating that the increased availability of UDP/UTP had a stimulatory role in the formation of end products of the hexosamine pathway. These findings also suggest that the depletion of intracellular uridine following prolonged GlcN infusion had in fact limited the amount of flux through the hexosamine pathway down to the level of UDP-GlcNAc and hence ultimately the amount or composition of oligosaccharide chains in newly formed glycoproteins. This could also explain the highly significant correlation between skeletal muscle UDP-GlcNAc levels and the percent decrease in glucose uptake in all groups. On the contrary, the GlcN-induced depletion of the muscle UDP-Glc pool did not appear to play any significant role in the induction of insulin resistance. Thus, repletion of uridine apparently unmasked a greater potential for GlcN to affect insulin action by enhancing the formation of downstream products of the hexosamine pathway.

The observation that increased GlcN availability led to a marked increase in the rate of incorporation of 14C-labeled GlcN into skeletal muscle GLUT4-containing vesicles suggests that one or more proteins associated with these vesicles is rapidly glycosylated. Thus, it is likely that GlcN administration resulted in either altered composition of the oligosaccharide chains in newly formed glycoproteins or an increase in the absolute rate of protein glycosylation.

Indeed, UDP-GlcNAc is a critical component of both O-linked and N-linked protein glycosylation. There is a relative enrichment of proteins bearing O-linked GlcNAc in cell nuclei (40), and transcription factors can be rapidly modified by O-linked glycosylation (41). The latter effects may play an important role in the long-term consequences of increased GlcN availability on the gene expression of key intracellular proteins. However, our current findings are most consistent with the hypothesis that altered post-translational glycosylation of a vesicular protein with an important role in GLUT4 vesicular trafficking or membrane docking could result in decreased GLUT4 translocation or activation and thus in defective activation of glucose transport by insulin.

Activation of the GlcN pathway by either GlcN or uridine infusion was associated with a profound suppression of glycogen synthesis, accounting for nearly all the impairment in glucose uptake, and the degree of suppression was greater in the presence of increased GlcN availability with concomitant uridine repletion. As noted above, we previously reported similar GlcN-induced impairments in glucose uptake and glycogen synthesis, together with accumulation of UDP-hexosamines and depletion of UDP-Glc (16, 37). This GlcN-induced suppression of glycogen synthesis had occurred despite normal activation of skeletal muscle glycogen synthase by insulin. Thus, the profound decrease in substrate availability appeared to be responsible for the reduction in glycogen synthesis. It is thus particularly noteworthy that the percent decrease in glycogen synthesis in the current studies was greater with uridine repletion than with GlcN alone. Thus, depletion of UDP-Glc was not solely responsible for GlcN-induced resistance to the peripheral effects of insulin either on glucose uptake or on glycogen synthesis.

There was indeed a significant impairment in the activation of muscle glycogen synthase by insulin in the two groups receiving uridine infusions. The observation of a diminished ability of insulin to activate skeletal muscle glycogen synthase in the presence of an increased availability of both GlcN and uridine may be attributed to the higher levels of muscle UDP-GlcNAc achieved in this group. In fact, higher concentrations of this metabolite may be required to affect insulin action on glycogen synthase than on glucose transport. However, the concentration of skeletal muscle UDP-GlcNAc increased comparably with the addition of uridine and GlcN alone. A potential explanation for the lack of effect of GlcN infusion on glycogen synthase activation may be found in the known inhibitory action that elevated levels of GlcN per se exert on the rate of glycosylation (13, 42). In fact, the concomitant increase in the concentration of GlcN may restrain the effects of increased availability of UDP-GlcNAc on its metabolic targets. The increased efficiency of the skeletal muscle UDP-GlcNAc generated during uridine versus GlcN infusion is also supported by the heightened effects of the former on the in vivo rates of glucose uptake and glycogen synthesis.

Despite a greater percent decrease in the rate of peripheral glycogen synthesis over 7 h of GlcN infusion, we noted an overall increase in total skeletal muscle glycogen content with the addition of uridine infusion. There was a significant time-dependent effect of uridine on the rate of glycogen synthesis: a transient initial increase, likely reflecting the increased concentration of UDP-Glc, was followed by a marked decrease (Table II).

In conclusion, the marked decrease in insulin-mediated glucose uptake and glycogen synthesis observed with increased GlcN availability appears to be proportional to the amount of flux into the hexosamine pathway and is not the consequence of skeletal muscle UDP-glucose depletion. Furthermore, the addition of uridine during both GlcN and saline infusions resulted in enhanced accumulation of skeletal muscle UDP-hexosamines that was highly correlated with the degree of peripheral insulin resistance. Thus, increased formation of UDP-GlcNAc generated either by increased GlcN availability or by enhanced uridylation causes a severe impairment in insulin-stimulated glucose uptake in skeletal muscle. The substantial decrease in Glc-6-P levels indicated that a defect in glucose transport and/or phosphorylation was responsible for the decrease in glucose uptake. Abnormal glycosylation of proteins that play a key role in the translocation of GLUT4-containing vesicles could explain such a reduction in glucose transport. We consequently propose that GlcN-induced resistance to the effects of insulin on glucose uptake and glycogen synthesis is mediated by increased accumulation of downstream hexosamine metabolites, perhaps via altered protein glycosylation of GLUT4-containing vesicles.


FOOTNOTES

*   This work was supported in part by Grants DK 45024 and DK48321 from the National Institutes of Health, by grants from the Juvenile Diabetes Foundation, and by Albert Einstein Diabetes Research and Training Center Grant DK 20541. This work was presented in preliminary form at the 56th Annual Meeting of the American Diabetes Association, San Francisco, June 8-11, 1996. The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Dagger    Recipient of post-doctoral research fellowships from the Juvenile Diabetes Foundation and the Medical Research Council of Canada.
§   Supported by Clinical Investigator Award KO8-AG00639 from the National Institutes of Health.
   To whom correspondence should be addressed: Div. of Endocrinology, Dept. of Medicine, Albert Einstein College of Medicine, 1300 Morris Park Ave., Bronx, NY 10461. Tel.: 718-430-4118/4215; Fax: 718-430-8557; E-mail: Rossetti@AECOM.YU.EDU.
1    The abbreviations used are: NIDDM, non-insulin-dependent diabetes mellitus; HPLC, high performance liquid chromatography; BSA, bovine serum albumin; PBS, phosphate-buffered saline.

Acknowledgments

We thank Meizhu Hu, Wei Chen, and Robin Squeglia for excellent technical assistance.


REFERENCES

  1. DeFronzo, R. A., Bonadonna, R. C., and Ferrannini, E. (1992) Diabetes Care 15, 318-368 [Abstract]
  2. Rothman, D. L., Shulman, R. G., and Shulman, G. I. (1992) J. Clin. Invest. 89, 1069-1075 [Medline] [Order article via Infotrieve]
  3. Bonadonna, R. C., Del Prato, S., Saccomani, M. P., Bonora, E., Gulli, G., Ferrannini, E., Bier, D., Cobelli, C., and DeFronzo, R. A. (1993) J. Clin. Invest. 92, 486-494 [Medline] [Order article via Infotrieve]
  4. James, D. E., Strube, M., and Mueckler, M. (1989) Nature 338, 83-87 [CrossRef][Medline] [Order article via Infotrieve]
  5. Fukumoto, H., Kayano, T., Buse, J. B., Edwards, Y., Pilch, P. F., Bell, G. I., and Seino, S. (1989) J. Biol. Chem. 264, 7776-7779 [Abstract/Free Full Text]
  6. Charron, M. J., Brosius, F. C., Alper, S. L., and Lodish, H. F. (1989) Proc. Natl. Acad Sci. U. S. A. 86, 2535-2539 [Abstract]
  7. Klip, A., Ramlal, T., Bilan, P. J., Cartee, G. D., Gulve, E. A., and Holloszy, J. O. (1990) Biochem. Biophys. Res. Commun. 172, 728-736 [Medline] [Order article via Infotrieve]
  8. Garvey, W. T., Maianu, L., Zhu, J.-H., Hancock, J., and Golichowski, A. (1993) Diabetes 42, 1773-1785 [Abstract]
  9. Yki-Jarvinen, H., Vuorinen-Markkola, H., Koranyi, L., Bourey, R., Tordjman, K., Mueckler, M., Permutt, A., and Koivisto, V. (1992) J. Clin. Endocrinol. & Metab. 75, 795-799 [Abstract]
  10. Garvey, W. T., Olefsky, J. M., Matthaei, S., and Marshall, S. (1987) J. Biol. Chem. 262, 189-197 [Abstract/Free Full Text]
  11. Marshall, S., Bacote, V., and Traxinger, R. R. (1991) J. Biol. Chem. 266, 4706-4712 [Abstract/Free Full Text]
  12. Traxinger, R. R., and Marshall, S. (1991) J. Biol. Chem. 266, 10148-10154 [Abstract/Free Full Text]
  13. Robinson, K. A., Sens, D. A., and Buse, M. G. (1993) Diabetes 42, 1333-1346 [Abstract]
  14. Giaccari, A., Morviducci, L., Zorretta, D., Sbraccia, P., Leonetti, F., Caiola, S., Buongiorno, A., and Bonadonna, R. C. (1995) Diabetologia 38, 518-524 [CrossRef][Medline] [Order article via Infotrieve]
  15. Crook, E. D., Daniels, M. C., Smith, T. M., and McClain, D. A. (1993) Diabetes 42, 1289-1296 [Abstract]
  16. Rossetti, L., Hawkins, M., Chen, W., Gindi, J., and Barzilai, N. (1995) J. Clin. Invest. 96, 132-140 [Medline] [Order article via Infotrieve]
  17. Baron, A., Zhu, J.-S., Zhu, J.-H., Weldon, H., Maianu, L., and Garvey, W. T. (1995) J. Clin. Invest. 96, 2792-2801 [Medline] [Order article via Infotrieve]
  18. Robinson, K. A., Weinstein, M. L., Lindenmeyer, G. E., and Buse, M. G. (1995) Diabetes 44, 1438-1446 [Abstract]
  19. Buse, M. G., Robinson, K. A., Marshall, B. A., and Mueckler, M. (1996) J. Biol. Chem. 271, 23197-23202 [Abstract/Free Full Text]
  20. Rossetti, L. (1995) Clin. Invest. Med. 18, 255-260 [Medline] [Order article via Infotrieve]
  21. Holstege, A., and Keppler, D. (1986) J. Natl. Cancer Inst. 76, 485-492 [Medline] [Order article via Infotrieve]
  22. Pederson, N., Knop, R., and Miller, W. (1992) Cancer Res. 52, 3782-3786 [Abstract]
  23. Erichsen, C., Christensson, P., Jakobsson, B., Eriksson, G., Yngner, T., Jonsson, P., and Stenram, U. (1987) Anticancer Res. 7, 77-80 [Medline] [Order article via Infotrieve]
  24. Rossetti, L., Smith, D., Shulman, G. I., Papachristou, D., and DeFronzo, R. A. (1987) J. Clin. Invest. 79, 1510-1515 [Medline] [Order article via Infotrieve]
  25. Rossetti, L., and Laughlin, M. R. (1989) J. Clin. Invest. 84, 892-899 [Medline] [Order article via Infotrieve]
  26. Rossetti, L., and Giaccari, A. (1990) J. Clin. Invest. 85, 1785-1792 [Medline] [Order article via Infotrieve]
  27. Rossetti, L., and Hu, M. (1993) J. Clin. Invest. 92, 2963-2974 [Medline] [Order article via Infotrieve]
  28. Karlander, S., Roovete, A., Vranic, M., and Efendic, S. (1986) Am. J. Physiol. 251, E530-E536 [Abstract/Free Full Text]
  29. Hostler, K. Y., and Landau, B. R. (1967) Biochemistry 6, 2961-2964 [Medline] [Order article via Infotrieve]
  30. Thomas, J. A., Schlender, K. K., and Larner, J. (1968) Anal. Biochem. 25, 486-499 [Medline] [Order article via Infotrieve]
  31. Rodnick, K. J., Slot, J. W., Studelska, D. R., Hanpeter, D. E., Robinson, L. J., Geuze, H. J., and James, D. E. (1992) J. Biol. Chem. 267, 6278-6285 [Abstract/Free Full Text]
  32. Holman, G. D., Kozka, I. J., Clark, A. E., Flower, C. J., Saltis, J., Habberfield, A. D., Simpson, I. A., and Cushman, S. W. (1990) J. Biol. Chem. 265, 18172-18179 [Abstract/Free Full Text]
  33. Michal, G. (1985) in Methods of Enzymatic Analysis (Bergmeyer, N. U., ed), Vol. VI, pp. 191-198, VCH Verlagsgesellschaft mbH, Weinheim, Germany
  34. Giaccari, A., and Rossetti, L. (1989) J. Chromatogr. 497, 69-78 [Medline] [Order article via Infotrieve]
  35. Giaccari, A., and Rossetti, L. (1992) J. Clin. Invest. 89, 36-45 [Medline] [Order article via Infotrieve]
  36. Anumula, K. R., and Taylor, P. B. (1991) Anal. Biochem. 197, 113-120 [Medline] [Order article via Infotrieve]
  37. Hawkins, M., Barzilai, N., Chen, W., Angelov, I., Hu, M., Cohen, P., and Rossetti, L. (1996) Diabetes 45, 1734-1743 [Abstract]
  38. Calvaruso, G., Taibi, G., Torregrossa, M. V., Romano, N., and Tesoriere, G. (1986) Life Sci. 39, 2221-2227 [Medline] [Order article via Infotrieve]
  39. Pels Rijcken, W. R., Overdijk, B., Van den Eijnden, D. H., and Ferwerda, W. (1995) Biochem. J. 305, 865-870 [Medline] [Order article via Infotrieve]
  40. Holt, G. D., and Hart, G. W. (1986) J. Biol. Chem. 261, 8049-8057 [Abstract/Free Full Text]
  41. Jackson, S., and Tijan, R. (1988) Cell 55, 125-133 [Medline] [Order article via Infotrieve]
  42. Elbein, A. D. (1987) Annu. Rev. Biochem. 56, 497-534 [CrossRef][Medline] [Order article via Infotrieve]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.