Glucose effectiveness, the ability of glucose per se to suppress endogenous glucose production (EGP), is lost in type 2 diabetes mellitus (T2DM). Free fatty acids (FFA) may contribute to this loss of glucose effectiveness in T2DM by increasing gluconeogenesis (GNG) and impairing the response to hyperglycemia. Thus, we first examined the effects of increasing plasma FFA levels for 3, 6, or 16 h on glucose effectiveness in nondiabetic subjects. Under fixed hormonal conditions, hyperglycemia suppressed EGP by 61% in nondiabetic subjects. Raising FFA levels with Liposyn infusion for ≥3 h reduced the normal suppressive effect of glucose by one-half. Second, we hypothesized that inhibiting GNG would prevent the negative impact of FFA on glucose effectiveness. Raising plasma FFA levels increased gluconeogenesis by ∼52% during euglycemia and blunted the suppression of EGP by hyperglycemia. Infusion of ethanol rapidly inhibited GNG and doubled the suppression of EGP by hyperglycemia, thereby restoring glucose effectiveness. In conclusion, elevated FFA levels rapidly increased GNG and impaired hepatic glucose effectiveness in nondiabetic subjects. Inhibiting GNG could have therapeutic potential in restoring the regulation of glucose production in type 2 diabetes mellitus.
- free fatty acids
- diabetes mellitus
fasting hyperglycemia in type 2 diabetes mellitus (T2DM) is characterized by increased endogenous glucose production (EGP) (11, 16) predominantly due to increased gluconeogenesis (GNG) (2, 47). Considerable evidence has shown potent direct effects of glucose to suppress EGP in healthy individuals (6, 17, 31), with marked suppression even in the absence of an insulin response (48). Loss of this “glucose effectiveness” contributes significantly to increased EGP (19, 31, 33). Elevated circulating free fatty acid (FFA) levels appear to play a pivotal role in the loss of glucose effectiveness in T2DM (2, 47), since FFAs increase EGP by stimulating GNG and by affecting key regulatory enzymes (1, 24, 25, 30).
Therefore, we wished to define 1) the time-dependent effects of raising FFA on glucose effectiveness in nondiabetic individuals, 2) the extent to which an increase in GNG is responsible for these defects, and 3) whether EGP can be inhibited in the presence of elevated plasma FFA by inhibiting GNG with ethanol. Our results confirm the striking effects of elevated plasma FFA to impair glucose effectiveness and suggest that increased GNG contributes importantly to this loss of regulation.
RESEARCH DESIGN AND METHODS
Eighteen healthy subjects (13 men and 5 women) were recruited. Informed, written consent was obtained in accordance with the policies of the Institutional Committee on Clinical Investigations at the Albert Einstein College of Medicine, which approved the study. The volunteers were taking no medications, had no family history of T2DM, and were not involved in any other research study. The average age and BMI were 47.4 ± 3.3 yr and 28.5 ± 1.3 kg/m2, respectively. A 2-h oral glucose tolerance test was performed to ensure normal glucose tolerance. On the basis of National Institute on Alcohol Abuse and Alcoholism guidelines for the study of ethanol, individuals who were pregnant or had a history of liver disease, depression, or alcohol problems were excluded from the study (34).
The study was divided into two main parts. The first part was designed to determine whether there are time-dependent effects of increased FFA on glucose effectiveness. The second part determined the rates of GNG and the impact of inhibiting GNG in the presence of elevated FFA on glucose effectiveness. All participants were studied after an overnight fast. Most subjects were admitted to the Clinical Research Center on the morning of the study. If deuterated water was ingested or Liposyn was infused for the duration of 16 h (Lip 16 h), the subjects were admitted to the Clinical Research Center the night before the study.
Time-Dependent Effects of FFA on Glucose Effectiveness
Euglycemic hyperglycemic pancreatic clamp studies.
The experimental pancreatic clamp protocols lasted 6 h and consisted of an initial 4-h euglycemic period (∼90 mg/dl) followed by a 2-h hyperglycemic period (180 mg/dl), as described previously (2, 21, 23). All experiments consisted of 360-min insulin/somatostatin (250 μg/h) infusions, with replacement of glucoregulatory hormones (glucagon 1.0 ng·kg−1·min−1, growth hormone 3.0 ng·kg−1·min−1) to maintain fixed levels of these hormones throughout. Glucose fluxes were measured with HPLC-purified [3-3H]glucose (prime infusion of 22 μCi, continued at 0.15 μCi/min) (2, 47).
To determine glucose effectiveness, glucose fluxes were compared between euglycemia and hyperglycemia in the study subjects under the different conditions outlined below. The subjects were studied under paired conditions with a control (saline infusion) and a Liposyn infusion study ≥1 mo apart, as described below (Fig. 1).
All subjects underwent a saline control study, with the duration of saline infusion paired with one of the following Liposyn infusion studies.
Lip 3 h (n = 8).
Infusion of lipid emulsion was conducted (20% Liposyn, 0.42 ml/min) for 3 h (t = 180–360) to reproduce the moderately elevated FFA levels observed in poorly controlled T2DM.
Lip 6 h (n = 12).
Infusion of lipid emulsion was conducted throughout the 6-h clamp studies.
Lip 16 h (n = 7).
Infusion of lipid emulsion started at 10 PM the night prior to the clamp study and continued throughout the duration of clamp study.
Inhibiting GNG in the Presence of Elevated FFA
Euglycemic hyperglycemic pancreatic clamp studies.
The pancreatic clamp studies were performed as described above (23). However, the experimental protocols lasted 7 h and consisted of an initial 3.5-h euglycemic period followed by a 3.5-h hyperglycemic period (2, 21). The length of these studies was extended to 7 h to ensure a sufficient duration of ethanol infusion for optimal GNG inhibition during the hyperglycemic period.
To determine the impact of inhibiting GNG on glucose effectiveness in the presence of elevated FFA, rates of EGP were compared between euglycemia and hyperglycemia under the conditions outlined below. All subjects were studied under the following three conditions, with each study ≥1 mo apart (Fig. 2).
Baseline 7-h saline control studies (normoglycemic hyperglycemic pancreatic clamp studies; n = 7) were conducted.
Lip+ (n = 7).
Seven-hour euglycemic hyperglycemic pancreatic clamp studies with infusion of lipid emulsion (20% Liposyn; Abbott Laboratories, North Chicago, IL) were conducted at the rate of 0.42 ml/min for the final 6 h to reproduce the moderately elevated FFA levels observed in poorly controlled T2DM. The 6-h infusions of Liposyn matched the intermediate duration (Lip 6 h) of the previous studies.
Lip+/Et+ (n = 7).
Infusion of both lipid emulsion (Liposyn) for the final 6 h of the studies and ethanol (to inhibit GNG) during the hyperglycemic phase (t = 210–420) of the 7-h euglycemic hyperglycemic pancreatic clamp studies was started.
Ethanol infusions titrated to reach plasma levels of 0.08 g/dl were started at t = 210 min. To avoid venous irritation, 98% ethanol was diluted with 0.9% saline for a final concentration of 6%. The physiologically based pharmacokinetic model, a three-compartment model of alcohol mass flow rate comprised of the liver, vasculature, and peripheral body water, was used to calculate the ethanol infusion rates (39). Target ethanol levels were established within 20 min and monitored by measuring breath alcohol levels using a hand-held breath alcohol analyzer and confirmed with blood ethanol measurements.
All GNG measurements were made during the euglycemic study phase. Accurate measurements of GNG could not be performed in the presence of changing glucose levels and exogenous glucose infusion. Maintaining glucose levels at ∼180 mg/dl would require substantially greater rates of exogenous, unlabeled glucose infusion relative to rates of hepatic glucose production. The infused unlabeled glucose would consequently dilute the deuterated glucose generated by GNG, artificially lowering the measured rates of GNG (data not shown). During the Lip+/Et+ studies described above, ethanol infusions were initiated only during the hyperglycemic phase of the studies to avoid prolonged ethanol exposure. Therefore, to accurately determine rates of GNG in the Lip+/Et+ group, we performed additional short euglycemic pancreatic clamp studies in four subjects, infusing both Liposyn and ethanol (Lip+/Et+) for the duration of 3.5 h.
Subjects drank deuterated water (total of 5 g D2O/kg total body water) at 8 PM, 11 PM, and 3 AM the night before each study. Body water was estimated to be 50% of body weight in women and 60% in men. Deuterated water was ingested slowly over ∼30 min/dose to avoid dizziness. Any other ingested water was enriched to 0.5% with D2O to maintain isotopic steady state. Blood was drawn to determine the ratio of deuterium enrichment on carbons 2 and 5 of glucose during the final 15 min of the euglycemic study phase (t = 210).
Plasma glucose, [3-3H]glucose-specific activity, insulin, C-peptide, glucagon, FFA, glycerol, and lactate were measured as described previously (2, 40). Blood ethanol measurements were performed using an enzymatic in vitro assay (ethyl alcohol kit; Roche Diagnostics, Mannheim, Germany). Measurements of gluconeogenesis were performed at Case Western Reserve University School of Medicine using an established method that measures the deuterium enrichment at carbons 2 and 5 on plasma glucose (26, 44).
Rates of glucose appearance and glucose uptake (GU) were calculated using Steele's steady-state equation (26). Rates of EGP were calculated as described previously (42). Data for glucose turnover, plasma hormones, and substrate concentrations represent the mean values during the final 60 min of the euglycemic and the final 60 min of the hyperglycemic periods.
Analysis of the data was performed using SPSS Version 11.5. For averaged data, Student's t-test was employed, using paired t-tests for comparisons of euglycemic and hyperglycemic intervals. Analysis of variance (ANOVA) was used to compare Lip 3 h, Lip 6 h, and Lip 16 h and to compare Lip−/Et−, Lip+/Et−, and Lip+/Et+ studies, and significant differences identified by ANOVA were further analyzed by paired t-tests. All data are presented as means ± SE unless specified otherwise. A P value of <0.05 was considered significant.
General Study Conditions
Following an overnight fast, plasma insulin concentrations averaged 11.5 ± 2.5 μU/ml (all studies), plasma glucose levels 99.9 ± 2.6 mg/dl (P < 0.001), and basal (t = 0) plasma FFA levels 446.6 ± 59.7 μM (Tables 1 and 3).
During the studies, plasma glucose levels averaged 100.2 ± 2.9 mg/dl during the euglycemic and 183.3 ± 1.4 mg/dl during the hyperglycemic study periods. Steady-state conditions were achieved because glucose-specific activity was constant following tracer equilibration during both euglycemia and hyperglycemia in each individual subject and study.
The average insulin infusion rates required to maintain euglycemia were 0.15 ± 0.02 μU·kg−1·min−1 for all study types. Plasma glucagon levels remained stable in all study types. C-peptide levels were suppressed by somatostatin infusion in all studies and did not differ among types of studies in either basal, euglycemic, or hyperglycemic periods.
Time-Dependent Effects of FFA on Glucose Effectiveness
Euglycemic hyperglycemic pancreatic clamp studies.
There was a trend toward higher plasma insulin levels in the 6- and 16-h FFA studies compared with the control studies in the same subjects, which was significant during the euglycemic phase [plasma insulin levels ∼23 (P = 0.01) and 48% (P = 0.03) higher in the Lip 6 h and Lip 16 h studies, respectively]. This is consistent with past studies demonstrating FFA-induced decreases in hepatic insulin clearance (4, 49). Of note, the FFA-induced impairments of EGP suppression are all the more striking in light of the elevated insulin levels.
Plasma lactate levels trended downward during the hyperglycemic phases of the Lip 16 h studies and were significantly lower in the Lip 6 h studies compared with the control studies in the same subjects (Table 1). These findings are consistent with FFA-induced stimulation of GNG, with increased consumption of lactate evident by 6 h.
Below, the results of all saline control studies are combined. In the tables, all study results are compared with the paired saline control studies in the same subjects, e.g., Lip 3 h vs. Lip−.
The average rate of glucose infusion required to maintain the target hyperglycemic plateau during the last 60 min of the hyperglycemic period was 4.1 ± 0.5 mg·kg−1·min−1. EGP was suppressed by 60.9 ± 6.3% with hyperglycemia. The percent increase in GU between the euglycemic and hyperglycemic study periods was 142.1 ± 28.8% (Table 2). Plasma FFA concentrations were 227.1 ± 30.4 μM during euglycemia and 189.9 ± 23.9 μM during hyperglycemia (P < 0.001; Table 1).
The infusion of Liposyn raised plasma FFA to levels comparable with those previously seen in T2DM (23). Significant elevations in FFA levels were attained after ∼1 h of Liposyn infusion. Glycerol levels were elevated during Liposyn infusion in Lip+ studies (Lip 3 h = 220.8 ± 9.0 μU/ml vs. Lip− = 76.3 ± 5.8 μU/ml, Lip 6 h = 213.1 ± 7.6 μU/ml vs. Lip− = 75.0 ± 5.0 μU/ml, Lip 16 h = 211.2 ± 10.8 μU/ml vs. Lip− = 77.9 ± 7.1 μU/ml; Table 1). To control for this finding, we previously infused glycerol alone in six subjects to achieve plasma glycerol levels comparable with those seen during Liposyn infusion (21). Glycerol infusion alone did not affect glucose effectiveness because the percent decrease in EGP and increase in GU in response to hyperglycemia was not significantly different from matched control studies.
The average rate of glucose infusion required to maintain the target hyperglycemic plateau during the last 60 min of the hyperglycemic period was decreased significantly during the Lip+ studies compared with the saline control studies in the same subjects (Lip 3 h = 2.3 ± 0.3 mg·kg−1·min−1 vs. Lip− = 4.3 ± 0.6 mg·kg−1·min−1, P = 0.006; Lip 6 h = 1.7 ± 0.3 mg·kg−1·min−1 vs. Lip− = 3.4 ± 0.4 mg·kg−1·min−1, P = 0.006; Lip 16 h = 2.6 ± 0.5 mg·kg−1·min−1 vs. Lip− = 4.7 ± 0.5 mg·kg−1·min−1, P = 0.005), suggesting a decrease in glucose effectiveness with increased FFA levels in all three groups (Table 2). The elevated FFA levels resulted in a significant blunting of the percent suppression of EGP with hyperglycemia in Lip studies of all three durations (Lip 3 h = 33.3 ± 5.5% vs. Lip− = 65.8 ± 5.9%; Lip 6 h = 32.3 ± 4.9% vs. Lip− = 56.2 ± 4.5%, Lip 16 h = 32.6 ± 5.9% vs. Lip− = 60.7 ± 8.4%; Fig. 3A and Table 2). There was no difference in suppression of EGP between Lip 3 h, Lip 6 h, and Lip 16 h [P = not significant (NS) by ANOVA]. The percent increase in GU was ∼50% reduced in each Lip study type relative to saline control studies in the same subjects. Stimulation of GU by hyperglycemia was reduced to 63.9 ± 18.0% in the Lip 3 h vs. 148.8 ± 31.5% in the Lip− (P = 0.003), 45.1 ± 12.6% in the Lip 6 h vs. 116.8 ± 20.9% in the Lip− (P = 0.009), and 82.3 ± 21.7% in the Lip 16 h vs. 160.6 ± 34.1% in the Lip− (P = 0.027) (Fig. 3B and Table 2).
Inhibiting GNG in the Presence of Elevated FFA
Euglycemic hyperglycemic pancreatic clamp studies.
Plasma insulin and C-peptide levels did not differ between the euglycemic and hyperglycemic study periods (Table 3). Plasma lactate levels were stable in the Lip−/Et− and Lip+ studies. During the hyperglycemic phase of the Lip+/Et+ studies, there was a significant increase in lactate. These findings are consistent with previous studies reporting ethanol-induced lactate elevations (18, 41) presumably due to a decrease in consumption of lactate via GNG.
The average rate of glucose infusion required to maintain the target hyperglycemic plateau during the last 60 min of the hyperglycemic period was 3.9 ± 0.5 mg·kg−1·min−1 (Table 4). EGP was suppressed by 61.4 ± 4.3% with hyperglycemia (Fig. 4). Furthermore, the percent increase in GU between the euglycemic and hyperglycemic study periods was 110.4 ± 18.5% (Fig. 5). Plasma FFA concentrations for the study subjects during euglycemia were 229.0 ± 32.0 and 198.9 ± 28.3 μM during hyperglycemia (Table 3).
The infusion of Liposyn raised FFA to levels comparable with those seen in poorly controlled T2DM (Table 3). Significant elevations in FFA levels were attained after ∼1 h of Liposyn infusion and maintained throughout the studies. Glycerol levels were also substantially elevated during Liposyn infusion in Lip+ studies (Lip+ = 216.2 ± 30.6 μU/ml vs. Lip−/Et− = 58.9.0 ± 13.1 μU/ml, P = 0.004). The average rate of glucose infusion required to maintain the target hyperglycemic plateau during the last 60 min of the hyperglycemic period was significantly lower during the Lip+ studies compared with the Lip−/Et− studies in the same subjects (Lip+ = 2.0 ± 0.5 mg·kg−1·min−1 vs. Lip−/Et− = 3.9 ± 0.5 mg·kg−1·min−1, P = 0.048) (Table 4). Plasma insulin levels in the Lip+ studies did not differ from those in the Lip−/Et− studies.
The elevated FFA levels resulted in a significant blunting of the percent suppression of EGP with hyperglycemia in Lip+ studies (Lip+ = 34.2 ± 3.7% vs. Lip−/Et− = 61.4 ± 4.4%, P = 0.0097; Fig. 4). The percent increase in GU during hyperglycemia in the Lip+ study type trended lower relative to the Lip−/Et− studies in the same subjects, although the difference was not statistically significant (Lip+ = 53.3 ± 21.0% vs. Lip−/Et− = 110.4 ± 18.5%, P = 0.092; Fig. 5 and Table 4).
Plasma ethanol levels were 90.4 ± 5.2 mg/dl at t = 300 min, 85.6 ± 4.9 mg/dl at t = 360 min, and 86.3 ± 3.3 mg/dl at t = 420 and remained stable throughout the studies (P = NS by ANOVA). The FFA values attained during the final hour of the Liposyn and ethanol coinfusion studies were comparable with the Lip+ studies (Lip+/Et+ = 510.9 ± 97.1 μM/l vs. Lip+ = 602.4 ± 59.1 μM/l, P = 0.17). Similarly, there were no statistical differences in the glycerol levels between the Lip+ and Lip+/Et+ studies. An upward trend in glycerol levels with onset of the ethanol infusion was likely caused by decreased GNG.
Importantly, there was substantially greater suppression of EGP in the Lip+/Et+ studies compared with the Lip+ studies (Lip+/Et+ = 65.8 ± 5.1% vs. Lip+ = 34.2 ± 3.7%, P = 0.004; Fig. 4). The percent decrease in EGP in the coinfusion studies was comparable with that in the saline control studies (Lip+/Et+ = 65.8 ± 5.1% vs. Lip−/Et− = 61.4 ± 4.4%, P = 0.6; Table 4). The stimulation of glucose uptake during the hyperglycemic phase of the Lip+/Et+ studies was markedly lower than that of the Lip−/Et− studies (Lip+/Et+ = 10.2 ± 8.8% vs. Lip−/Et− = 110.4 ± 18.5%, P = 0.001; Fig. 5). GU trended downward in the Lip+/Et+ studies compared with the Lip+ studies, although it did not reach statistical significance (Lip+/Et+ = 10.2 ± 8.8% vs. Lip+ = 53.3 ± 21.0%, P = 0.1). However, glucose infusion rate was significantly decreased with infusion of both Liposyn and ethanol (Lip+/Et+ = 2.2 ± 0.4 mg·kg−1·min−1 vs. Lip−/Et− = 3.9 ± 0.5 mg·kg−1·min−1, P = 0.02; Table 4). These findings are consistent with previous studies confirming that insulin-mediated GU is significantly reduced in the presence of systemic ethanol, particularly in nondiabetic individuals (3).
Under euglycemic pancreatic clamp (Lip−/Et−) conditions, rates of GNG averaged 0.9 ± 0.1 mg·kg−1·min−1 (accounting for ∼39% of EGP) (Fig. 6). When FFA levels were elevated by Liposyn infusion throughout the 3.5-h studies, rates of GNG increased to 1.4 ± 0.1 mg·kg−1·min−1 (P = 0.006). However, ethanol completely prevented the Liposyn-induced rise in GNG, with GNG decreasing significantly to baseline rates of 0.9 ± 0.1 mg·kg−1·min−1 (P = 0.008 between Lip+ and Lip+/Et+ studies, P = 0.999 for Lip−/Et− vs. Lip+/Et+).
Loss of glucose effectiveness contributes significantly to the worsening of hyperglycemia in T2DM. Since increased FFA levels stimulate EGP, we proposed that they might be a key player in this loss of regulation in T2DM. Thus, it was our intent to examine 1) the time-dependent effects of raising FFA on EGP and glucose effectiveness in nondiabetic individuals, 2) the extent to which an increase in GNG is responsible for these defects, and 3) whether glucose effectiveness can be restored in the presence of elevated plasma FFA by inhibiting GNG with ethanol. Indeed, there was rapid blunting of hepatic glucose effectiveness following only 3 h of Liposyn infusion. The magnitude of this effect was equal and remained constant at 3, 6, and 16 h of fatty acid infusion. Significant increases in GNG were noted after only 2 h of Liposyn infusion. Furthermore, FFA-induced loss of glucose effectiveness was completely restored by ethanol infusion, likely due to inhibition of GNG, in the face of increased plasma FFA levels.
Effects of Elevated FFA on Glucose Effectiveness
The major cause of fasting hyperglycemia in T2DM is believed to be increased EGP (11, 16). In nondiabetic individuals, the majority of suppression of EGP by glucose has been attributed to glucose-induced reductions in FFA levels (7). Indeed, hyperglycemia is known to inhibit lipolysis and decrease FFA (27). Plasma FFA levels seem to play a critical role in glucose effectiveness and hepatic EGP and rise in proportion to worsening glycemic control (19). We previously reported that individuals with optimal glycemic control retain normal suppression of both FFA and EGP in response to hyperglycemia, and 72 h of intensive insulinization in poorly controlled T2DM normalized glucose effectiveness and plasma FFA levels (19). In the current studies, hepatic glucose effectiveness was rapidly impaired after only 3 h of Liposyn infusion. A key feature of the study was to reproduce the moderate elevations in FFA levels typical of T2DM, since insulin signaling and other parameters can be affected by supraphysiological rises in plasma FFA levels. Indeed, all durations of Liposyn infusion attained FFA levels comparable with what we have previously reported during clamp studies in T2DM subjects (21, 23).
Increased plasma FFA levels also inhibited peripheral glucose effectiveness, defined as the ability of hyperglycemia per se to stimulate whole body glucose uptake. This finding was in accordance with previous studies reporting impaired peripheral glucose effectiveness with elevated plasma FFA levels under conditions of physiological insulin levels (28). In T2DM, the ability of glucose to stimulate its own uptake may be decreased due to decreased numbers of glucose transporters at the plasma membrane (35). Since FFA seem to directly influence the expression (28) and translocation (50) of skeletal muscle glucose transporters, this could be another means by which elevated FFA levels impact glucose metabolism in T2DM.
FFA-Induced Increases in GNG
FFAs have been shown to increase GNG both in vivo and in vitro (8, 24). GNG is the pathway inappropriately increased in T2DM, accounting for the high overall rates of EGP (9, 29). Increased levels of plasma FFA have been shown to potently stimulate hepatic GNG through various mechanisms, including the enhanced gene expression of gluconeogenic enzymes phosphoenolpyruvate carboxykinase and fructose-1,6-bisphosphatase (2, 43). The prompt effects of raising FFAs described in these studies would be consistent with rapid changes in GNG.
Of note, all GNG measurements in the current studies were taken during the euglycemic study phase. Maintaining glucose levels at ∼180 mg/dl would require substantially greater rates of exogenous, unlabeled glucose infusion relative to rates of hepatic glucose production. The infused, unlabeled glucose would consequently dilute the deuterated glucose generated by GNG, artificially lowering the measured rates of GNG (data not shown). Therefore, to accurately quantify GNG, it was necessary to measure GNG during euglycemia. To minimize ethanol exposure, ethanol was infused only during the hyperglycemic phase of the studies. Therefore, we performed a small number of additional studies in which the combined effects of FFA and ethanol on GNG were quantified under stable, euglycemic conditions. Indeed, the elevation of plasma FFA for only 2 h stimulated GNG significantly.
Importantly, in most physiological settings, marked increases or decreases in GNG alone are not enough to change EGP in either nondiabetic or T2DM individuals (13, 22, 37, 38, 45). A hepatic “autoregulatory” mechanism maintains constant EGP and plasma glucose in the presence of stable hormone and plasma glucose levels. This is due to compensatory changes in glycogenolysis. During euglycemic conditions this mechanism remains intact, even in the face of FFA-induced increases in GNG (9). Our studies confirm this finding, given that Liposyn infusion increased rates of GNG but failed to alter rates of EGP. Because autoregulation would be intact under euglycemic conditions, ethanol should likewise not affect basal EGP during euglycemia, since its inhibitory effects on GNG should be compensated by increases in glycogenolysis (32). This observation was likewise confirmed in our studies. However, increases in plasma glucose levels rapidly inhibit glycogenolysis in nondiabetic individuals and would impair this hepatic autoregulation. Although we did not measure glycogenolysis directly in these studies, we would predict on the basis of the findings of Chu et al. (14) that increased FFA in the presence of hyperglycemia would suppress glycogenolysis. Indeed, the current studies demonstrate the loss of hepatic autoregulation, evidenced by a rise in EGP in response to Liposyn infusion during hyperglycemia.
Besides stimulating GNG, increased plasma FFAs also affect EGP through changes in flux through the two key hepatic enzymes glucokinase (GK) and glucose-6-phosphatase (G-6-Pase) (24, 25, 30). Increased FFA availability decreases the expression of GK, simultaneously promoting the expression of and allosterically activating G-6-Pase (12, 15). Both of these effects impair glucose effectiveness, since defective GK activity would impair the liver's ability to “sense” changes in plasma glucose levels (5, 20) and increased G-6-Pase flux would result in increased rates of EGP.
Infusion of Ethanol in the Presence of Increased FFAs Restores Glucose Effectiveness
As discussed above, glucose effectiveness was dramatically and rapidly blunted in the face of elevated FFAs. Here we show that ethanol infusion in the presence of increased FFAs completely restores glucose effectiveness. Under euglycemic conditions, we have demonstrated the inhibitory effects of ethanol on GNG. Although we were not able to directly measure GNG during hyperglycemia, it is likely that the inhibition of GNG also accounted for the restoration of glucose effectiveness in the presence of elevated FFA. However, it cannot be excluded that ethanol independently suppresses GNG and EGP under hyperglycemic conditions. Of note, moderate alcohol consumption has been shown to lower the risk of developing T2DM (36). Since ethanol is associated with other concerns, new approaches of inhibiting GNG are currently being developed. Recently, a fructose-1,6-bisphosphatase inhibitor has been shown to lower blood glucose in monkeys and diabetic rodents (46).
Notably, there was a nearly complete loss of peripheral glucose effectiveness in the presence of both Liposyn and ethanol infusions. This is likely due to the combined inhibitory effects of increased plasma FFA and ethanol levels on peripheral glucose uptake. In support of this, Avogaro et al. (3) reported significantly decreased whole body insulin-mediated glucose uptake with systemic ethanol infusion in both nondiabetic and T2DM individuals but found it to be particularly evident in the nondiabetic subjects.
To conclude, these studies confirm the significant impact of increased plasma FFA levels on glucose effectiveness. In particular, the substantial effects of increased FFA levels on the ability of glucose to suppress EGP appear to be due in great part to FFA-induced stimulation of GNG. Remarkably, ethanol infusion in the presence of increased FFAs completely restored hepatic glucose effectiveness. Since the loss of glucose effectiveness in T2DM may contribute significantly to hyperglycemia, inhibiting GNG shows promise as a potential target for intervention.
This work was supported by funding from the American Diabetes Association, National Institutes of Health Grants 5-R01-DK-069861 and 1-P01-AG-021654 (to M. Hawkins), and National Center for Research Resources Grant 1K23-RR-02335-01 (to P. Kishore).
We thank Laura Clintoc, Angela Stangarone, and Dr. Corina Fratila for assisting with the studies, Robin Sgueglia for determinations performed in the Diabetes Research and Training Center Hormone Assay Core (P60-DK-20541), Dr. Hillel Cohen for biostatistical assistance, and the staff of the Albert Einstein College of Medicine's Clinical Research Center (M01-RR-12248) for outstanding patient care. We acknowledge Dr. Harry Shamoon for many helpful discussions and the late Dr. Bernard Landau for the gluconeogenesis measurements. We also thank Dr. Sean O'Connor for providing us with the physiologically based pharmacokinetic program that enabled us to perform ethanol infusions.
This work was presented in preliminary form at the 64th Scientific Sessions of the American Diabetes Association, June 4–8, 2004, in Orlando, FL.
- Copyright © 2009 the American Physiological Society