INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

THE PRODUCTION OF GLUCOSE by liver and kidney provides requisite fuel to the brain and is therefore under complex regulation (4, 14). Thus, under a wide variety of circumstances such as fasting (14) or exercise (4), the blood glucose level is valiantly defended despite extreme variability in metabolic need. This fine regulation is aberrant in type 2 diabetes, which is characterized by peripheral insulin resistance and chronic hyperglycemia. The observed hyperglycemia after overnight fasting has been ascribed to a failure to suppress endogenous glucose production (EGP) in the face of peripheral insulin resistance (12, 13), implying defect(s) in the regulation of hepatic gluconeogenesis (GNG) and/or glycogenolysis (GLY) (11, 26, 27). Additionally, impairment of insulin's ability to suppress EGP after a meal is at least partially responsible for the impaired glucose tolerance observed in type 2 diabetes (13).

The role of hormonal (28, 36) and metabolite control on GNG and glycogen degradation in liver has been well studied. However, the relative roles of GNG and GLY in liver remain controversial, because, for technical reasons, GNG has proven difficult to quantify (28, 36). Further confounding the understanding of GNG and GLY is the apparent compensatory interrelationship ("hepatic interregulation") between these two pathways (19, 21, 22, 24, 31, 39). Thus basal EGP remained constant when GNG was acutely increased by infusion of gluconeogenic precursors (19, 21, 22) or was inhibited with ethanol (24, 31). In rats, this interregulatory effect was demonstrated despite markedly decreased concentrations of liver glycogen. Collectively, these data suggest that an initial modification of the gluconeogenic rate is followed by compensatory changes in the glycogenolytic rate, thus maintaining a constant EGP and satisfying the energetic needs of the central nervous system.

Despite the potentially crucial role that intracellular hepatic interregulation may play in controlling glycemia in health and diabetes, little is known regarding the mechanism of this phenomenon (19, 21, 22, 24, 31, 39). Possible sites of interregulation include coordinated modification of glycogen and/or gluconeogenic enzyme activities and/or pathway regulation by glucose 6-phosphate or other key intermediates. Furthermore, it is currently not clear whether the role of hepatic interregulation is associated with modulation of GNG only or represents a general mechanism that maintains a glucose output proportional to need regardless of which pathway(s) is affected.

One approach to determine whether the intraregulation phenomenon is independent of which pathway is affected is to study whether specific suppression of GLY leads to a compensatory increase in GNG. This was done in the present study by using 1,4-dideoxy-1,4-imino-D-arabinitol (DAB), a novel, potent, and specific inhibitor of hepatic glycogen phosphorylase (GP), whose effects have been shown both in vitro and in vivo (1, 17). The drug was infused in overnight-fasted (11 h), conscious mongrel dogs under basal and glucagon-stimulated conditions. GNG was measured using the dideuterated water method of Kalhan et al. (23) and Landau et al. (25). We asked whether or not a specific reduction in glycogen degradation in liver would result in an equal and opposite increase in GNG to maintain constant EGP. In a similar study, Shiota et al. (34) reported that inhibition of GLY enhanced glucagon-stimulated gluconeogenic precursor uptake by the liver of conscious dog. Their findings are discussed.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Animals. Eight conscious male mongrel dogs (25.1 ± 3.8 kg) were studied according to Principles of Laboratory Animal Care (National Institutes of Health Publication no. 85-23, revised 1985) and California law. Animals were housed under controlled kennel conditions (12:12-h light-dark cycle) in the Keck Medical School (University of Southern California) Vivarium. Animals had free access to standard chow (25% protein, 9% fat, 49% carbohydrate, and 17% fiber; Wayne Dog Chow, Alfred Mills, Chicago, IL) and tap water. Animals were used for an experiment only if they had a hematocrit >38%, a good appetite, and normal body temperature and stools. Food was withdrawn 11 h before experiments. Chronic catheters were implanted 7 days before the experiments, as previously described (32). One catheter was placed in the portal vein 4 cm upstream from the porta hepatis for portal infusions of insulin and glucagon. Slow infusions at this site are equally distributed among the lobes of the liver (6). A second catheter was placed in the femoral vein and advanced to the inferior vena cava for the infusion of tracer, somatostatin, and DAB. A third catheter was placed in the jugular vein for the sampling of mixed venous blood. On the morning of each experiment, a catheter was acutely inserted into the saphenous vein for the variable infusion of glucose.

Experimental protocol. Each dog underwent four protocols under euglycemic clamps performed in random order. The protocols were separated by 3 wk to clear ²H₂O from the circulation. At t = 240 min, a bolus of ²H₂O (10 g/kg) was injected into the femoral vein to achieve a 1% enrichment of body water. D-[3-³H]glucose was given as a primed, continuous infusion beginning at t = 150 min (25 µCi prime + 0.25 µCi/min infusion; NEN Research Products, Du Pont, Boston, MA). Also at this time, a femoral infusion of somatostatin (1 µg · kg¹ · min¹; Bachem, Torrance, CA) was started to suppress endogenous insulin and glucagon release, and basal portal replacement infusions of insulin (porcine insulin: 0.3 mU · kg¹ · min¹; Novo Nordisk, Copenhagen, Denmark) and glucagon (porcine glucagon: 1.3 ng · kg¹ · min¹; Sigma Chemical, St. Louis, MO) were begun. Basal samples were taken at t = 240 (denoting the fasting concentration) and 150 min (denoting the basal concentration). Blood samples were drawn every 15 min from t = 150 to 30 min, and glucose was measured on-line at these and all later time points. Glucose was clamped at basal (t = 150 min) values by a variable infusion of glucose (Glc_inf) labeled with D-[3-³H]glucose (2.7 µCi/g) to minimize large changes in the specific activity (16). After tracer equilibration, blood samples were drawn at t = 30, 20, 10, and 0 min. At t = 0 min, DAB dissolved in saline or saline vehicle was given as a primed, continuous infusion for 210 min (1.943 mg/kg prime + 0.0187 mg · kg¹ · min¹ infusion, Novo Nordisk). The selected dose of DAB (5.87 mg/kg) was based on preliminary experiments in dogs (K. Lundgren and L. Ynddal, unpublished observations) to obtain a constant plasma concentration of DAB. The full pharmacokinetic profile of DAB is currently not known and is a subject for further investigation. At t = 60 min, the dogs were infused portally with high glucagon (GN; 5 ng · kg¹ · min¹) or 0.9% saline (SAL) in addition to the basal replacement concentrations of insulin and glucagon. Thus each animal (n = 8) was subjected to four protocols: DAB + GN, SAL + GN, DAB + SAL, and SAL only. Mixed venous blood samples for assays were collected every 5 min from t = 0 to t = 30 min, followed by collections every 10 min to t = 60 min, again every 5 min from t = 60 to t = 90 min, again every 10 min to t = 120 min, and every 15 min from t = 120 to t = 180 min. Finally, blood samples were obtained at t = 180, 190, 200, and 210 min. All solutions were infused at a flow rate of 0.25 ml/min, except for DAB, which was infused at 0.2175 ml/min. Samples for determination of plasma glucose, tritiated glucose, glycerol, lactate, insulin, and glucagon concentration were collected in 50 µl of EDTA (2 g/100 ml for 1.5 ml blood) in tubes coated with lithium fluoride and heparin. Trasylol (aprotinin: 75 µl/ml blood; Miles, Kankakee, IL) was added to samples for glucagon measurement to inhibit proteolysis of the hormone. Samples for the measurement of free fatty acids (FFA) were collected in EDTA with paraoxon to suppress lipoprotein lipase (40). Samples were centrifuged immediately in a vacuum centrifuge, and the plasma was separated into microcentrifuge tubes. Plasma samples were either kept on ice and processed the same day or stored at 80°C until assayed. For comparison of rates of GNG, GLY, and EGP, we defined four periods as follows: period 1 (P1), 30 to 0 min; period 2 (P2), 40-60 min; period 3 (P3), 90-110 min, and period 4 (P4), 180-210 min. The outline of the infusion protocol is shown in Fig. 1.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Time schedule and protocol for injections and infusions. Dogs were fasted overnight for 11 h before onset of the protocol. At t = 210 min, fasting blood samples were taken, a bolus of 2H2O (10 g/kg) was given, and the experiment was started. At t = 150 min, basal blood samples were taken, and a primed infusion of [3-3H]glucose was started. In addition, infusion of somatostatin (SRIF) and replacement concentrations of glucagon and insulin were initiated at t = 150. A primed 1,4-dideoxy-1,4-imino-D-arabinitol (DAB) or saline (SAL) infusion was initiated at t = 0 min, and infusion of high concentrations of glucagon (GN) or SAL was initiated at t = 60 min. Period 1 (P1) was defined from t = 30 to 0 min; period 2 (P2) from t = 40 to 60 min; period 3 (P3) from t = 90 to 110 min, and period 4 (P4) was defined from t = 180 to 210 min. Experiments were concluded at t = 210 min. See METHODS for further details.

Assays. Glucose and lactate were measured with a YSI 2700 autoanalyzer (Yellow Springs Instrument, Yellow Springs, OH) immediately after sampling. Glycerol was measured with a 912 Hitachi analyzer (Boehringer Mannheim, Mannheim, Germany). Samples for D-[3-³H]glucose were deproteinized with Ba(OH)₂ and ZnSO₄, and supernatants were evaporated in a vacuum, reconstituted in water, and counted in Ready Safe scintillation fluid (Beckman liquid scintillation counter; Beckman Instruments, Fullerton, CA). Tracer infusates were processed in the same manner. FFA were measured with a kit from Wako (NEFA C; Wako Pure Chemical Industries, Richmond, VA). Glucagon was assayed using a kit obtained from Linco Research (St. Louis, MO). Insulin was measured by ELISA, on the basis of two murine monoclonal antibodies that bind to different epitopes on the insulin molecule (2). Novo Nordisk provided materials for the insulin assay, including the dog standard. Deuterium incorporation at carbon-6 (C6) of glucose was determined using the hexamethylenetetramine derivative, as described by Kalhan et al. (23) and Landau et al. (25) and modified by A. Sunehag (unpublished observation). This method measures GNG from pyruvate and underestimates GNG by the contribution from glycerol and to some extent by a potential, incomplete equilibration between the deuterium enrichment in body water and at glucose C6 (23, 25). The conclusions from the present studies are based on comparisons among experiments. Thus a potential, incomplete equilibration of the deuterium enrichment is unlikely to differ among the experiments. Furthermore, to evaluate potential differences in the contribution from glycerol, glycerol concentrations were measured in all experiments. Total GNG can be determined from the deuterium enrichment at glucose C5; however, this is an extremely tedious and time-consuming method. The deuterium enrichment in hexamethylenetetramine was analyzed by gas chromatography-mass spectrometry (HP6890/5973; GC column: HP 5: 25 m × 0.25 mm × 1.0 µm, Hewlett-Packard, Palo Alto, CA) in the electron impact mode with selected monitoring of mass-to-charge ratios 140 and 141. Hexamethylenetetramine enrichments were converted to the corresponding glucose enrichments by use of a standard curve prepared from [1-²H]glucose (99 atom % ²H, Cambridge Isotope Laboratories, Andover, MA) after conversion to sorbitol (30). Deuterium enrichment in body water (represented by plasma water) was measured by isotope ratio mass spectrometry (Finnigan Delta-E, Finnigan MAT, San Jose, CA) after reduction to hydrogen gas according to accepted methods (37, 38).

Calculations. EGP was calculated using Steele's model (32). During periods 1 and 4, the specific activity of glucose (SA_Glc) and the concentration of measured hormones and metabolites were stable. Periods 2 and 3 were included for a comparison of glucose kinetics just before GN was started (P2) and when the new level of plasma glucagon was reached (P3) for a closer examination of the proposed interregulatory mechanism. Estimates of EGP in the latter two periods were based on the assumption of stable concentrations of metabolites and SA_Glc. GNG-to-EGP ratios can be calculated as the deuterium bound to C6 of glucose divided by plasma ²H₂O content when steady state is achieved (23, 25). A recent study in preterm human babies indicates that the deuterium-C6 method provides an accurate method of quantifying GNG, so long as glycerol GNG is not included (35). In the present study, a correction factor was necessary, because the glucose pool was diluted by a variable infusion of exogenous glucose, which did not affect the deuterium enrichment of plasma water. The correction factor was calculated on the basis of a single compartment model (see APPENDIX). The rate of GLY was calculated as (EGP  GNG).

Statistical analysis. All results are given as means ± SE. Balanced ANOVA was used to test for the effects of DAB and glucagon on all dog outcome variables. One dog did not complete the DAB + GN and SAL + GN protocols; this dog was otherwise similar to the rest of the group. To allow performance of a balanced ANOVA, the average values from the remaining seven dogs were substituted for the missing values. When significance was reached by ANOVA, a paired Student's t-test was used to compare P1 and P4. This paired Student's t-test was based on n = 7, or n = 8 when possible, and did not include estimated data.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Hormone replacement. The outline of the protocol is shown in Fig. 1. During P1, the average concentration of plasma glucagon of 45 ± 6 ng/l (Fig. 2) matched the fasting concentration (Table 1). Although glucagon concentration remained unchanged in the absence of GN [Fig. 2: 41 ± 6 vs. 45 ± 6 ng/l, P4 vs. P1, P = nonsignificant (NS)], infusion of glucagon at 5 ng · kg¹ · min¹ tripled plasma glucagon to 134 ± 8 ng/l during P4 (Fig. 2, P < 0.001). DAB infusion did not affect glucagon concentration in any of the protocols (P = NS). Infusion of insulin (0.3 mU · kg¹ · min¹) underreplaced the fasting and basal concentrations (Tables 1 and 2, P < 0.001). Plasma concentrations of FFA and glycerol during the experiments were similar to basal concentrations (Tables 1 and 2, P = NS) but were modestly reduced compared with fasting concentrations (Tables 1 and 2, P < 0.001).

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Summary of plasma concentrations of glucagon. Dogs were infused with basal replacement concentrations of glucagon (1.3 ng · kg1 · min1) throughout the experiment or basal replacement followed (at t = 60 min) by GN (5 ng · kg1 · min1); see METHODS for details. No effect of DAB on plasma glucagon concentrations was observed. Vertical bars indicate SE values.

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Effect of DAB and GN on the deuterium enrichments of body water (left) and carbon 6 (C6) of plasma glucose (right). Four protocols were performed on each dog: DAB + GN (n = 7), SAL + GN (n = 7), DAB + SAL (n = 8), and SAL only (n = 8), and the deuterium enrichment of body water and C6 of glucose was measured during the 4 study periods for the determination of gluconeogenesis (GNG), as described in METHODS. Deuterium enrichment of plasma water and C6 of glucose was stable in all periods (P1-P4) in all protocols. Vertical bars indicate SE values.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Plasma levels of glucose and rates of glucose infusion. Plasma glucose levels (left) and glucose infusion rates (Glcinf; right) are shown. Plasma glucose levels were clamped at basal (t = 150 min) values by variable infusions of glucose (Glcinf), as described in METHODS. Four protocols were performed on each dog: DAB + GN (n = 7), SAL + GN (n = 7), DAB + SAL (n = 8), and SAL only (Control; n = 8). GN caused a transient increase above the clamped level; however, no differences in plasma glucose levels were observed between any groups when P1 and P4 were compared. Vertical bars indicate SE values.

Glucose turnover. With basal hormone replacement (control), EGP demonstrated a modest decline during the experiment (Fig. 5), declining from 1.79 ± 0.18 to 1.66 ± 0.17 mg · kg¹ · min¹, from P1 to P4. This decline was due to falling GLY, which was reduced by 30% over the experiment (Table 3, P = NS).

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Effect of DAB and GN on the rate of endogenous glucose production (EGP). Four protocols were performed on each dog: DAB + GN (n = 7), SAL + GN (n = 7), DAB + SAL (n = 8), and SAL only (n = 8), and dynamic changes in EGP were measured as described in METHODS. DAB impaired the effect of glucagon on the rate of EGP (P < 0.05, n = 7). Vertical bars indicate SE values.

1

1

1

1

1

1

View larger version (33K): [in this window] [in a new window]   Fig. 6.   Effect of DAB and GN on the rate of gluconeogenesis (GNG) and glycogenolysis (GLY) during P4. Four protocols were performed on each dog: DAB + GN (n = 7), SAL + GN (n = 7), DAB + SAL (n = 8), and SAL only (n = 8), and dynamic changes in the rate of GLY (top) and GNG (bottom) were measured as described in METHODS. Total EGP = GNG + GLY. GN stimulated GLY (*P < 0.05, n = 7) and GNG (***P < 0.001, n = 7). Additional infusion of DAB suppressed this effect of glucagon on GLY without a compensatory increase of GNG. DAB alone had no effect on either GNG or GLY during P4 (P = NS, n = 8). Vertical bars indicate SE values.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION APPENDIX REFERENCES

Despite the central role of the liver in the regulation of blood glucose in health and diabetes, questions continue to exist regarding the regulation and relative roles of the two pathways contributing to EGP, GNG, and GLY (19, 24, 31, 39). When GNG was inhibited with ethanol in patients with type 2 diabetes (31) or with rats (24) or when GNG was acutely increased by infusion of gluconeogenic precursors (19, 21, 22), hepatic glucose output remained constant, suggesting that an initial modification of the gluconeogenic rate is followed by compensatory changes in the glycogenolytic rate. This mechanism has been termed hepatic autoregulation (19, 21, 22, 24, 31, 39). Here, however, we prefer the term interregulation, because the term autoregulation has traditionally been used to describe the mechanism by which the blood glucose controls the overall hepatic glucose output (29).

Currently, it is not clear whether the suggested interregulation is a phenomenon solely associated with modulation of GNG or a general mechanism that maintains a constant EGP regardless of which pathway is affected. Shiota et al. (34) reported that inhibition of GLY by BAY R 3401 enhanced glucagon-stimulated gluconeogenic precursor uptake by the liver of conscious dog. Also, maximal estimate of GNG was higher in the drug group than in placebo, whereas the authors observed no differences when comparing minimal estimates of GNG.

Here, using the compound DAB, a novel inhibitor of GP and glycogen breakdown (1, 17), we report no changes in GNG upon inhibition of GLY with or without glucagon stimulation. The difference in findings might be explained as a deposition of gluconeogenic carbon as glycogen, which would not be detected by the dideuterated water method that we used in our study. Thus DAB had no influence on glycogen synthesis in primary hepatocytes (1) or on lactate deposition into glycogen in the perfused rat liver (K. Fosgerau, N. Westergaard, and J. Breinholt, unpublished observations). In contrast, the amount of glycogen synthesized from gluconeogenic carbon with infusion of BAY R 3401 in dogs was higher in the drug group than in the placebo group (34). Alternatively, the difference in findings might be explained as a difference in the mechanism of action between compounds DAB and BAY R 3401. Thus, in hepatocytes, DAB had no effect on either protein phosphorylase 1, the enzyme responsible for dephosphorylation of GPa to GPb, or on phosphorylase kinase (1), whereas BAY R 3401 promoted dephosphorylation of GP (3).

Notably, maximal estimate of GNG in the study of Shiota et al. (34) was obtained by assuming that all of the gluconeogenic precursors taken up by the liver were completely converted to glucose. However, a significant amount of the gluconeogenic precursors was actually deposited as glycogen. Because the amount of glycogen synthesized from gluconeogenic carbon was higher in the drug group than in the placebo group (34), the maximal estimate of GNG is relatively overestimated in the former group.

The basal rate of EGP in dogs treated with DAB was compared with that in dogs not treated with DAB (Fig. 5). At t = 10, 15, 20, and 50 min, the rate of EGP was lower in dogs treated with DAB (P < 0.05). This is believed to be the result of an inhibition of basal GLY, because the average rates of GNG were not affected by DAB during P2. However, there was a tendency toward an increase in EGP immediately before the infusion of GN at 1 h. Also, in the absence of GN infusion, GLY during P4 was equal in dogs treated with DAB compared with controls (Fig. 6), suggesting that the infusion protocol used for DAB did not lead to full inhibition of GP. Alternatively, increased lysosomal hydrolysis of glycogen may diminish the effect of DAB on glycogen breakdown, because DAB is a weak inhibitor of mammalian glucosidases (1) and 10% of the hepatic glycogen is located within the lysosomes (18).

Glucagon is recognized as a critical physiological regulator of EGP and basal glucose homeostasis (7-10, 15, 33). Elevated concentrations of glucagon have stimulating effects on both the gluconeogenic and glycogenolytic pathways, albeit following different time courses (9, 33). Thus a sustained elevated glucagon concentration causes a rapid and marked activation of GLY followed by a declining rate of this pathway despite a sustained period of GN ["evanescent" effect (9, 33)], just as we observed (Table 3). In contrast, the rate of GNG is slowly upregulated with elevated concentrations of glucagon, and this effect is sustained (9, 33). It is possible that the initial marked increase in GLY observed with elevated concentrations of glucagon limits glucagon's stimulatory effect on GNG at first and that the rate of GNG subsequently increases only as GLY declines (33). The existence of such a mechanism would further support the existence of the proposed hepatic interregulatory mechanism.

As expected, elevated glucagon had a transient stimulatory effect on GLY followed by a slow increase in GNG in the present study (Table 3). However, when DAB was given before glucagon, the glucagon stimulation of EGP was sharply reduced. This inhibitory effect of DAB was sustained throughout the experiment, as demonstrated by the fact that EGP above basal in dogs infused with DAB + GN was only 36% of the increment in EGP observed in animals treated with glucagon alone. We observed no differences in plasma glucose levels between any groups when P1 and P4 were compared (Fig. 3, P = NS). However, there was a tendency toward an increased plasma glucose level in response to SAL + GN infusion during P3. This response in plasma glucose can be ascribed to the effect of glucagon on GLY (7-10, 15, 33). The glucose level was restored (P4) at the clamp level by adjustment of Glc_inf (Fig. 3) and as a result of an increased rate of disposal (R_d) (data not shown). With infusion of DAB + GN, we observed increased plasma glucose level during P4, which can be ascribed to the effect of glucagon on GNG (7-10, 15, 33). Thus, judging from the enrichment data of C6 of glucose (Fig. 4), there was a tendency toward increasing deuterium incorporation during P4 after a GN infusion, indicating that GNG was, in fact, increasing at this stage and that the maximum rate of GNG was not reached in the length of the study period. The compensatory adjustment of Glc_inf resulted in the observed glucose level, which was above, but not statistically different from, the P1 level. However, it cannot be excluded that the metabolism of DAB may cause a relief of the observed GLY inhibition during P4. Previous studies in enzyme preparations of GP (17) and in the perfused liver (K. Fosgerau, H. Westergaard, and J. Breinholt, unpublished results) revealed no effect of glucose on the inhibitory action of DAB.

GNG was measured during four periods (P1-P4) by means of deuterium labeling of glucose secondary to infusion of deuterated water (23, 25). This method is a straightforward approach to measuring GNG; however, it underestimates GNG by the contribution from glycerol and potentially by an incomplete equilibration between the deuterium enrichment in body water and at C6 of glucose. In the present study, there were no differences with regard to plasma concentrations of glycerol and FFA between the experiments, indicating that DAB did not affect lipolysis or glycerol delivery from adipose tissue. Thus it is very unlikely that DAB had any effect on the gluconeogenic contribution from glycerol. Deuterium labeling of glycogen would lead to an overestimation of GNG; however, in the present study, deuterated water was given during fasting conditions, where labeling of glycogen is believed to have no significance because glycogen synthase is not activated (20, 29). Furthermore, it was previously shown that DAB has no effect on glycogen synthesis in primary hepatocytes (1) or in the perfused rat liver (K. Fosgerau, H. Westergaard, and J. Breinholt, unpublished observations). Consequently, it is not believed that the choice of method for estimating GNG had any impact on our conclusions regarding GNG.

Because GNG was equal with or without DAB, the lower glucose production observed with DAB can be ascribed to an inhibition of glucagon-stimulated GLY. Thus, as shown in Fig. 6, DAB inhibited glucagon-stimulated GLY by 46% during P4, confirming the mechanism of action of DAB and again demonstrating that it is effective in vivo (1, 17). Also, because GNG in our hands was not affected by infusion of DAB regardless of the concentration of glucagon, it is apparent that, in the fasting normal dog, hepatic interregulation may not exist for GNG being regulated by GLY. Thus this suggests that hepatic interregulation is a phenomenon for which GNG may alter the rate of GLY, but not the converse (19, 21, 22, 24, 31, 39). Of course, we did not alter the rate of GNG in the present study, so that the existence of the hepatic interregulation phenomenon overall is not disproved by the present results, in that changes in GNG may indeed alter GLY. Also, at the concentration of DAB used, glucagon-stimulated GLY was not completely inhibited during P4. It remains possible that inhibition of GLY must be more pronounced for an effect to be seen on GNG. Finally, the possibility continues to exist that the apparent constancy of EGP might be an extrahepatic phenomenon. Our studies were carried out under clamped conditions. It remains a possibility that the plasma glucose itself (20, 29), other blood-borne compounds such as FFA (5, 23, 28), or the central nervous system may be important extrahepatic regulators of liver function that guarantee an appropriate delivery rate of glucose to the peripheral tissues regardless of the balance between GNG and GLY. Also, we measured total EGP and did not separate out the relative roles of liver vs. kidney in glucose production. Possibly, a decrease in hepatic GNG was followed by an increase in kidney GNG; however, this latter possibility appears unlikely, because glucose production by the kidney provides only a minor fraction of total GNG for blood glucose homeostasis after an overnight fast (14).

In conclusion, DAB inhibited glucagon-stimulated EGP via inhibition of GLY but did not affect the absolute contribution of GNG to EGP with or without glucagon stimulation. These data do not support an effect of changing GLY on GNG. However, it is still possible that the changes in GNG can affect the glycogenolytic rate, suggesting that the interrelationship between glycogen degradation and GNG in controlling EGP is complex. Because the glucagon challenge is regarded as a model of type 2 diabetes, these data suggest that inhibition of glycogen phosphorylase might prove beneficial in the treatment of this disease.