Role of Excess Glycogenolysis in Fasting Hyperglycemia Among Pre-Diabetic and Diabetic Zucker (fa/fa) Rats

Male obese Zucker (fa/fa) and lean Zucker (+/+) rats were a generous gift from Dr. Roger H. Unger at the Veterans Affairs North Texas Health Care System (Dallas, Texas). [3,4-¹³C₂]glucose (99%) was purchased from Omicron Biochemicals (South Bend, IN). [U-¹³C₃]propionate (99%), ²H₂O (99.9%), and deuterated acetonitrile (99.8%) were obtained from Cambridge Isotopes (Andover, MA). Deuterated N,N-dimethylformamide-d₇ (99.5%) and other common chemicals were purchased from Aldrich (Milwaukee, WI).

This study was approved by the institutional animal care and use committee at the University of Texas Southwestern Medical Center. Four different groups of male Zucker rats were studied: 8- and 13-week-old fa/fa rats and age-matched +/+ rats. Immediately before surgery to place a chronic indwelling venous catheter, a urine sample was collected from 12-week-old fa/fa rats; development of glucosuria was checked by urine glucose level (DiastixR; Bayer, Munich, Germany). After the right jugular vein was cannulated under ketamine-xylazine anesthesia, animals were allowed to recover for 7 days with free access to water and standard chow containing 6% fat. Then, 1 day before the study, rats were fasted for 24 h with free access to water. At t = −30 min, rats received an intraperitoneal injection of ²H₂O (20 μl/g rat), and at t = 0 they received a bolus infusion of [3,4-¹³C₂]glucose (35 μmol/kg) followed by continuous infusion of [3,4-¹³C₂]glucose and [U-¹³C₃]propionate (1 and 13 μmol · kg⁻¹ · min⁻¹, respectively) via the jugular vein catheter for 90 min. At the end of the 90-min infusion period, whole blood (∼8–10 ml, the entire available blood volume) was drawn from the inferior vena cava under anesthesia with pentobarbital sodium (50 mg/kg body wt), and the liver and skeletal muscle tissue from the hind limbs were freeze-clamped and kept at −80°C for subsequent processing.

Whole blood was immediately deproteinized by adding cold perchloric acid to a final concentration of 7% by volume. After neutralization with KOH and centrifugation, the supernatant was lyophilized. Dried glucose was converted into monoacetone glucose (MAG) (Fig. 1) as described previously (10,15). A 5- to 8-g portion of liver and skeletal muscle tissue was used for glycogen extraction and purification (16). Isolated glycogen was dissolved in ∼5 ml of 10 mmol/l sodium acetate solution (pH 4.8) and incubated with amyloglucosidase (50 mg glycogen per 20 units) for 4 h at 50°C. After freeze-drying, the hydrolyzed glycogen was converted to MAG.

All NMR spectra were collected using a Varian INOVA 14.1 T spectrometer (Varian Instruments, Palo Alto, CA) equipped with a 3-mm broadband probe as described previously (10). NMR spectra were analyzed using the curve-fitting routine of NUTS, a PC-based NMR spectral analysis program (Acorn, Fremont, CA). Glucose turnover was estimated from the dilution of infused [3,4-¹³C₂]glucose using ¹³C NMR to analyze MAG derived from blood glucose at the end of the infusion protocol (12). The contributions of glycogen, glycerol, and phosphoenol pyruvate (PEP) to blood glucose was obtained directly from the deuterium enrichment at positions 2, 5, and 6_S as determined from the ²H NMR of MAG (13,17). A ¹³C NMR isotopomer analysis—based on the ¹³C-¹³C spin-coupled multiplets of carbon 2 (C2) of MAG—that yields relative fluxes in the citric acid cycle has been reported previously (10,14).

²H enrichment of the H2 position of glucose was measured using N,N-dimethylformamide-d₇ as an internal reference, as described previously (18). Also the natural abundance of methyl groups of MAG (Fig. 1) was used as internal references for the estimation of ²H excess enrichment for low–enriched glycogen samples.

Plasma glucose was assayed by the glucose oxidase method (Sigma). Plasma free fatty acids (FFAs) were measured enzymatically using a commercially available kit (Wako Chemical, Richmond, VA). Plasma insulin was determined using a sandwich enzyme-linked immunosorbent assay kit (Alpco Diagnostics, Windham, NH). Freeze-clamped liver and muscle tissues were pulverized under liquid nitrogen, and a small portion (∼0.5 g) of each tissue was used for glycogen assay (19).

The data are the means ± SE. Comparisons between groups were performed using one-way ANOVA. Differences in mean values were considered statistically significant at P < 0.05.

Both 8- and 13-week-old fa/fa rats were heavier and had higher liver-to-body ratios than their age-matched controls (Table 1). Both age- groups of fa/fa rats were hyperglycemic after fasting and had higher levels of plasma FFAs and plasma insulin than their age-matched controls. Insulin levels in 13-week-old fa/fa rats tended to be lower than that of 8-week-old fa/fa rats, although this difference did not reach statistical significance. Liver glycogen was essentially depleted in controls, but it remained high in fa/fa rats, particularly the 13-week-old animals (24.0 ± 8.6 and 86.9 ± 15.4 μmol glucosyl unit/g wet weight in 8- and 13-week-old animals, respectively). Animals in all groups had similar low levels of skeletal muscle glycogen (∼2 μmol glucosyl unit/g wet weight).

Typical ¹³C NMR spectra of MAG derived from blood glucose of an fa/fa rat and a +/+ rat are shown in Fig. 1. All six carbons originating from glucose are easily resolved, and consequently the tracer used for glucose turnover, [3,4-¹³C₂]glucose, is detected at 75.3 ppm (signal from C3) and 80.5 ppm (signal from C4). The signal from each carbon is further split by spin-spin coupling, so the NMR signature of [3,4-¹³C₂]glucose is unique. At a fixed rate of infusion of [3,4-¹³C₂]glucose, the area of the characteristic doublets will decrease proportionally to the rate of hepatic glucose production (Fig. 1C). At steady state, [3,4-¹³C₂]glucose was 1.79 ± 0.16% (8 weeks of age +/+), 1.52 ± 0.21% (8 weeks of age fa/fa), 2.00 ± 0.44% (13 weeks of age +/+), and 1.72 ± 0.17% (13 weeks of age fa/fa) of total plasma glucose. Glucose production (v1) (Table 2) was significantly higher (∼18–27%) in fa/fa rats of both age-groups compared with their age-matched controls, despite higher levels of plasma insulin.

The ratio of H6_S to H2 in the ²H NMR spectrum is a direct measure of the fraction of blood glucose produced by gluconeogenesis from the TCA cycle (17). By simple inspection of the ²H NMR spectra in Fig. 2, it is easily seen that a significant fraction of blood glucose is derived from PEP in all groups because the H6_S resonance is high compared with H2 resonance in all spectra. This is to be expected because the animals were fasting. Total gluconeogenesis from both glycerol and the TCA cycle is measured by H5/H2 (13). By inspection of Fig. 2, it is easily seen that the H5-to-H2 ratio is somewhat less than 1 in the 8-week-old fa/fa animals (Fig. 2B) and significantly less than 1 in the 13-week-old fa/fa animals (Fig. 2D). Because glycogenolysis is the difference between gluconeogenesis and glucose production, these spectra demonstrate higher glycogenolysis in both fa/fa groups. The higher hepatic glycogen content in fa/fa animals (Table 1) correlates nicely with a higher contribution from glycogenolysis to glucose production in these animals, as determined by ²H NMR analysis (Table 2). The difference between H5 and H6_S was modest in all four groups, indicating that the fractional contribution of glycerol to endogenous glucose production ranged from 16 to 19%.

The absolute rates of glucose production from glycogen, PEP, and glycerol were determined by multiplying the individual fractional contributions and the glucose production rates. These values are summarized in Table 2. Glycogenolysis was 6.3 ± 3.6 and 9.4 ± 4.0 μmol · kg⁻¹ · min⁻¹ in 8- and 13-week-old fa/fa rats, respectively, but it was negligible in both +/+ rats. Endogenous glucose production from either glycerol or PEP was not significantly higher in fa/fa rats compared with age-matched controls.

Entry of [U-¹³C₃]propionate into the TCA cycle and subsequent turnover produces a mixture of ¹³C isotopomers in all intermediates and eventually in glucose derived from the TCA cycle. Examples of the C2 multiplet pattern of the ¹³C NMR spectrum of MAG from each group are shown in Fig. 3. In each case, these spectra are expansions of the C2 region shown in Fig. 1. Analysis of the C2 multiplets provides an estimate of flux through PEP carboxykinase (PEPCK), pyruvate cycling, and gluconeogenesis, all relative to citrate synthase flux. Because gluconeogenesis is commonly thought to be increased in type 2 diabetes, the finding that gluconeogenesis from the TCA cycle, relative to flux through citrate synthase, was ∼30% lower among fa/fa rats compared with age-matched controls was unanticipated (Table 2).

The overall pattern of glucose production becomes much clearer once the individual measures of glucose production, fractional sources of glucose production by ²H NMR, and TCA cycle fluxes by ¹³C NMR are combined to measure fluxes through multiple pathways (Table 2 and Fig. 4). An interesting picture emerges. TCA cycle flux is higher by ∼40% in fa/fa animals in both age-groups compared with controls, and the flux through PEPCK (v6) was also higher by ∼24–50%. However, despite these differences, gluconeogenesis through PEP (v4) was not higher in fa/fa animals compared with age-matched controls because of pyruvate cycling (v5) in fa/fa rats. Thus, the hepatic TCA cycle is significantly more active in fa/fa rats, but this does not translate into greater glucose production from the level of the TCA cycle because three carbon units leaving the cycle in the form of PEP are shunted back into cycle intermediates by pyruvate cycling mechanisms. This suggests that pyruvate cycling acts as a rheostat in modulating gluconeogenesis from the level of the TCA cycle, perhaps whenever glucose is available via glycogenolysis.

In Fig. 5, ²H NMR spectra were scaled so that the natural abundances of methyl signals are approximately equal. Therefore, each spectrum in Fig. 5 represents an equal number of glucose molecules. With this in mind, it is easy to appreciate that the fraction of hepatic glycogen enriched in ²H is dramatically lower in fa/fa animals compared with controls, regardless of age (Fig. 5). Among 8-week-old animals, the ²H enrichment in position 2 was nearly 10-fold lower in fa/fa animals (Fig. 5B). A similar pattern was evident in the 13-week-old animals. This marked difference between fa/fa and control animals is attributable largely to high glycogen content in fa/fa animals. For example, the mass of hepatic glycogen in 8-week-old controls was ∼205 g body wt per rat × 19.6 mg liver/g body wt × 1.7 μmol glycogen/g liver, or 6.8 μmol hepatic glucosyl units per +/+ rat. Because the ²H enrichment in H2 was 1.37% in glycogen compared with 2.3% in plasma glucose, ∼60% of this hepatic glycogen mass was synthesized over the course of the experiment (2 h), or ∼0.16 μmol · kg body wt⁻¹ · min⁻¹. Similarly, the mass of hepatic glucosyl units was ∼188 μmol per 8-week-old fa/fa rat, ∼25 times more than the age-matched controls, whereas the rate of glycogen synthesis was ∼0.31 μmol · kg body wt⁻¹ · min⁻¹, roughly double the value seen in controls. In summary, higher levels of “old” glycogen are responsible for the obvious differences seen in Fig. 5 between fa/fa and control animals in ²H spectra of hepatic glucosyl units.

The ²H enrichment patterns in MAG from hepatic glycogen were strikingly different between fa/fa and +/+ animals. In contrast, the ²H enrichment patterns in MAG from skeletal muscle glycogen were indistinguishable between groups, and overall absolute ²H excess enrichment of the H2 position of the glucosyl unit was low (∼0.2%) (Table 1) in all groups. However, the ²H distribution pattern was unusual compared with either blood glucose or hepatic glycogen: the highest ²H enrichments were in H1 and H2, whereas the remaining H3–H6 positions were enriched at a much lower level (Fig. 6).

These observations demonstrate a profound age-dependent disruption in glycogen metabolism among fa/fa rats. Compared with age-matched controls, the fa/fa animals were heavier, had greater liver mass per gram body weight, and had far more glycogen per gram liver. After a 24-h fast, the hepatic glycogen content per fa/fa rat was >25 times that of control among the 8-week-old animals and >100 times that of control among the 13-week-old animals. Although the fraction of plasma glucose derived from glycogen was also increased, gluconeogenesis from glycerol and the citric acid cycle remained quite active, so the overall fraction of glucose derived from glycogen was <20%. In this widely used animal model of type 2 diabetes, fasting hyperglycemia is caused by a combination of continual gluconeogenesis and excess glycogenolysis.

Many features of type 2 diabetes are observed in the fa/fa rat, and for this reason it is a popular model for study (3). The current observations among these fa/fa animals, such as fasting hyperglycemia, hyperinsulinemia, elevated plasma FFAs, and elevated endogenous glucose production, are typical of human patients, particularly those with “severe” type 2 diabetes, as defined by Basu et al. (20) as fasting glucose >9 mmol/l. However, it should be emphasized that the data presented here highlight similarities but also substantial discrepancies between glucose metabolism measured in this rat model compared with measurements in patients with type 2 diabetes. For example, Clore et al. (4) measured the glycemic response to glucagon after 64 h of fasting, a time when gluconeogenesis should be maximally stimulated. With the assumption that the glycemic response represents glycogen, patients with type 2 diabetes had higher liver glycogen after a 3-day fast compared with an overnight fast, and glycogen after a 3-day fast was higher among patients with type 2 diabetes compared with nondiabetic individuals. Muller et al. (21) found that hepatic glycogen measured in liver biopsies was increased among obese human subjects. These observations are consistent with the elevated glycogenolysis and increased hepatic glycogen content reported here. In contrast, hepatic glycogen measured directly by ¹³C NMR spectroscopy, considered highly reliable, was reduced among patients with diabetes in studies by Magnusson et al. (22), Hundal et al. (23), and Kunert et al. (24). The rate of glycogenolysis among patients with type 2 diabetes is also controversial. Reduced (22,25), unchanged (5), and increased (20) rates of glycogenolysis all have been reported. The associated metabolic features of diabetes may also play a role in glycogen metabolism because elevated fatty acids may inhibit glycogen mobilization (26). These discrepancies may be attributable to differences in methods for assessing glycogenolysis and gluconeogenesis, or to the study of patients at different stages of disease. It is conceivable that glycogen metabolism among humans with diabetes for several years may be quite different from metabolism in a rodent with insulin resistance for several weeks, at most.

Despite the relatively early reports of abnormal hepatic glycogen metabolism among subjects with diabetes, the cause of preserved glycogen stores in humans or rats during a modest fast is unclear. Clore et al. (4) proposed that increased gluconeogenesis could be responsible for glycogen sparing, but this possibility was discounted because suppressed gluconeogenesis did not deplete liver glycogen in type 2 diabetic patients after a 3-day fast (27). Previous studies of hepatocytes from fa/fa rats suggested increased glycogen synthesis (7,28). In the current study, the ²H enrichment patterns of the glucosyl unit of glycogen provided valuable information about the sources of hepatic glycogen in two respects. First, ²H enrichment in the H2 position was ∼20-fold less than that of blood glucose in fa/fa rats (Table 1). This level of enrichment of liver glycogen in fa/fa animals is far too low to attribute preservation of liver glycogen to increased glycogen synthesis (glyconeogenesis) during the period of this study. Second, the ²H NMR spectrum also demonstrated active glycogen synthesis through gluconeogenic pathways in fasted fa/fa rats and even among controls, though at a lower rate (Fig. 5). The latter finding is somewhat unexpected because glycogen synthesis is not thought to be active in fasted normal animals. The ²H NMR data indicate that both glyconeogenesis and reduced glycogen turnover relative to the mass of glycogen contribute to the preserved glycogen stores in fa/fa animals.

Citrate synthase flux (v7) and efflux (v6) of carbon skeletons from the TCA cycle of fa/fa rats were higher than controls. Nevertheless, glucose production from the TCA cycle was not increased because of increased reentry into the TCA cycle. Efflux of a four-carbon skeleton into a three-carbon intermediate followed by recarboxylation has been termed “pyruvate cycling.” This phrase refers to combined flux through two distinct futile cycles: oxaloacetate → PEP → pyruvate → oxaloacetate, and oxaloacetate → malate → pyruvate → oxaloacetate. Thus, in pyruvate cycling, pyruvate kinase and malic enzyme are two key enzymes. NADPH is produced during the conversion of malate to pyruvate through malic enzyme, which is then consumed in fatty acid chain elongation as a part of lipogenesis. Thus, increased pyruvate cycling of fa/fa rats may be needed because of increased requirement for NADPH if energy reserves are to be stored as fat.

The unique ²H distribution pattern of skeletal muscle glycogen suggests an unemphasized aspect of glyconeogenesis in skeletal muscle. Lactate, amino acids, muscle glycolytic intermediates, and glycerol are known or potential substrates for glyconeogenesis (29). However, those substrates and known pathways for glycogen synthesis would not result in high ²H enrichment of both the H1 and H2 position of the glucosyl unit of muscle glycogen. Chandramouli et al. (30) reported ²H labeling into the H1 and H2 positions of glucose from ²H₂O via phosphoglucose isomerase and phosphomannose isomerase (PMI) activities, which interchanges mannose 6-phosphate, glucose 6-phosphate, and fructose 6-phosphate. The involvement of PMI in skeletal muscle glyconeogenesis has not been described, but mRNA levels of PMI were high in human skeletal muscle (31). Another possibility for the high enrichment of H1 and H2 is via glycogen cycling and intramolecular ¹H (or ²H) exchange at the phosphoglucose isomerase level (32). In an earlier study, however, ²H NMR analysis of liver glycogen with active glycogen cycling did not demonstrate exceptionally high enrichment in both H1 and H2 (10). Thus, the possibility of high enrichment through intramolecular ²H exchange seems low in muscle.

In this study, steady-state metabolic conditions were not confirmed in each rat over the infusion time course because it was not practical due to the inherent low sensitivity of NMR. Previously, we reported that a primed 90-min infusion was sufficient to accurately measure endogenous glucose production in rats by comparing the 90-min results with those after a primed 150-min infusion (12). In the current study, to eliminate the possibility of overestimation of endogenous glucose production in hyperglycemic fa/fa rats because of a low priming dose or short infusion duration, steady state was confirmed by performing a 120-min infusion with 13-week-old fa/fa rats (n = 3). Endogenous glucose production of the rats was 59.6 ± 4.7 μmol · kg⁻¹ · min⁻¹, essentially the same value as a 90-min infusion for the same group of rats (56.1± 6.2) (Table 2), satisfying the steady-state assumption. Isotopic steady state was also assumed for metabolism of [U-¹³C₃]propionate in this study. The analysis of glucose C2 (or MAG C2) depends only on the multiplet patterns detected in the ¹³C NMR spectrum, not on the absolute ¹³C enrichment. The 13-week-old fa/fa rats with a 120-min infusion showed a similar multiplet pattern of glucose C2 (Fig. 3D) as the same group of rats with a 90-min infusion. This observation is consistent with an early study that indicated that a 45-min infusion is sufficient to reach a steady-state condition during infusion of [U-¹³C₃]propionate in rats (33). Together, these data indicate that a primed 90-min infusion was sufficient to reach steady state.

The current results illustrate the need to develop practical methods that would allow clinical investigators to easily quantify glycogen content and turnover in patients. Our results indicate that hepatic glycogenolysis plays a significant role in glucose overproduction in fa/fa rats, especially in older animals. Both 8- and 13-week-old fa/fa rats had higher fluxes of hepatic pyruvate cycling, PEPCK flux, and TCA cycle compared with age-matched controls. These abnormal metabolic fluxes involve multiple pathways in hepatic metabolism, even in relatively young animals, suggesting that enzyme fluxes may be potential markers for predicting onset of overt type 2 diabetes as well as potential targets to prevent the transition from a pre-diabetic state into overt diabetes.

This study was supported by grants from the National Institutes of Health (RR-02584, DK16194, and HL-34557).

We thank Charles Storey and Angela Milde for their excellent work in animal experiments and Erin Smith for technical support.