Fructose Amplifies Counterregulatory Responses to Hypoglycemia in Humans

We studied seven lean nondiabetic volunteers (five men and two women, aged 33 ± 3 years, BMI 24.2 ± 1.42 kg/m², HbA_1c 5.9 ± 0.2%). All participants were in good health, were not taking any medications, and had no family history of diabetes. Each subject participated in two stepped hypoglycemic clamp studies separated by an interval of 6 weeks. During each hypoglycemic clamp, subjects received either fructose (1.2 m · kg⁻¹ · min⁻¹) or saline intravenously in random order. Clamp studies were performed after an overnight fast.

Informed written consent was obtained in accordance with policy of the Committee on Clinical Investigations of the Albert Einstein College of Medicine.

Subjects were admitted to the General Clinical Research Center (GCRC) for each experiment. At 0700 on the day of study, two indwelling cannulae were inserted, one in an antecubital vein for infusions and the second placed in a retrograde fashion in a distal hand vein of the contralateral forearm for blood sampling. To obtain arterialized venous blood samples, this hand was maintained at 55°C in a thermoregulated Plexiglas box. At t = −120 min, a primed-continuous infusion of HPLC-purified [3-³H]glucose was initiated with a bolus of 21.6 μCi followed by continuous infusion of 0.15 μCi/min for the entire period of study.

The specific activity of infused dextrose was kept equivalent to plasma glucose specific activity by addition of [3-³H]glucose to the infusate, using the method of Finegood et al. (25) to prevent negative R_a artifacts. At t = 0 min, a primed-continuos infusion of insulin (Humulin Regular; Eli Lilly, Indianapolis, IN) at a rate of 0.4 mU · kg⁻¹ · min⁻¹ was initiated, and a variable infusion of 20% dextrose was begun to maintain the plasma glucose concentration at 5.0 mmol/l for 50 min (step one of the clamp). At t = +50 min and every 50 min thereafter, the plasma glucose concentration was decreased by 0.5 mmol/l decrements for 50 min, each by reducing the dextrose infusion rate accordingly. The plasma glucose was clamped at the desired range by varying the dextrose infusion according to plasma glucose measured at 5-min intervals with targets of 4.4, 3.9, and 3.3 mmol/l. At t = +50 min (plasma glucose ∼5.0 mmol/l), an infusion of fructose 1.2 mg · kg⁻¹ · min⁻¹or saline was initiated and maintained constant until the end of the study. This infusion rate of fructose was selected in order to achieve plasma fructose levels of ∼200 μmol/l (26,27). Such low-circulating levels of fructose would be regulatory for GK translocation, without providing increased substrate for gluconeogenesis (28). However, to examine whether fructose infusion could directly influence the rates of glucose production or peripheral glucose uptake in the absence of changes in the counterregulatory hormones, we performed another set of studies in five healthy lean volunteers matched for age, BMI, and HbA_1c with the participants in the stepped hypoglycemia studies (three men and two women, BMI 23.1 ± 2.81 kg/m², HbA_1c 5.4 ± 0.5%). Using the pancreatic clamp technique to suppress glucagon secretion and lowering the plasma glucose to 4.2 mmol/l (above the expected glycemic threshold for other counterregulatory hormones secretion), we examined the effects of fructose per se on EGP. Glucose tracer was administered as previously described. At t = 0 min, an infusion containing somatostatin (250 μg/h) (Bachem, King of Prussia, PA), growth hormone (3.0 ng · kg⁻¹ · min⁻¹) (Genentech, San Fransisco, CA), glucagon (1.0 ng · kg⁻¹ · min⁻¹ (Eli Lilly), and 1 mg/ml albumin diluted in saline was initiated and maintained throughout the study. Insulin (Humulin Regular) infusion was also initiated at t = 0 and maintained at a fixed rate of 0.4 mU · kg⁻¹ · min⁻¹. Plasma glucose was maintained at basal concentrations during the first 120 min by a variable infusion of dextrose 20% (i.e., the 5.0 mmol/l nominal glucose step). At t = 120 min, plasma glucose was lowered to 4.2 mmol/l by decreasing the dextrose infusion rate, and simultaneously an infusion of either fructose (1.2 mg · kg⁻¹ · min⁻¹) or saline was initiated and maintained for the final 120 min of the study (the 4.2-mmol/l nominal glucose step). Thus, by suppressing glucagon and growth hormone secretion (with somatostatin) and maintaining the plasma glucose above the threshold for counterregulatory hormone secretion, the direct effect of fructose (versus saline) on glucose counterregulation was assessed. At the end of the clamp, all of the infusions were discontinued, and the subject was given a meal and discharged from the GCRC.

Fasting blood for HbA_1c was collected on the morning of each clamp study. During the clamps, blood samples were obtained for the determinations of plasma insulin, C-peptide, glucagon, epinephrine, norepinephrine, cortisol, growth hormone, and lactate, as well as for glucose kinetics.

Plasma glucose was measured with a Beckman glucose analyzer (Fullerton, CA), using the glucose oxidase method. Plasma [3-³H]glucose radioactivity was measured in duplicate on the supernatants of barium hydroxide–zinc sulfate precipitates of plasma samples after evaporation to dryness to eliminate tritiated water.

The methods for measurement of plasma insulin, C-peptide, glucagon, epinephrine, norepinephrine, cortisol, growth hormone, lactate, and their intra- and interassay variations have been previously reported (29).

The data in the text, figures, and tables are presented as means ± SE. Steele’s equation was used for calculation of glucose turnover as described elsewhere (30). Values for EGP and glucose uptake, obtained at 10-min intervals, were averaged over the final 30 min of each glucose step for each individual subject. The glycemic thresholds for activation of counterregulatory hormone secretion was calculated as the glycemic level at which there was an increase of >2 SDs above the basal plasma hormone concentration (31,32). The area under curve was calculated using the trapezoidal method for the whole duration of the clamp (31). Statistical analyses were performed using repeated measures ANOVA for multiple comparisons and paired Student’s t test for comparing means before and after the intervention (fructose infusion). A value of P < 0.05 was considered significant.

Plasma glucose concentrations during both clamps (fructose and control) are shown in Fig. 1A. Plasma glucose concentrations at t = 0 min were 5.1 ± 0.1 mmol/l in the fructose studies and 4.9 ± 0.1 mmol/l in the control studies (P = NS). No significant difference was noted between the fructose and control studies during the 5.0-, 4.4-, or 3.9- mmol/l plasma glucose steps (4.9 ± 0.1, 4.6 ± 0.1, and 4.0 ± 0.1 mmol/l, respectively, in the fructose studies and 4.9 ± 0.1, 4.5 ± 0.1, and 3.9 ± 0.1 mmol/l, respectively, in the control studies). However, during the 3.3 mmol/l plasma glucose step, there was a significantly higher plasma glucose concentration in the fructose studies compared with control studies (3.8 ± 0.1 vs. 3.4 ± 0.1 mmol/l, respectively, P = 0.009). The higher plasma glucose levels in the fructose studies during this glucose step occurred despite the fact that the glucose infusion was discontinued and coincided with elevated plasma concentration of epinephrine and glucagon. Thus, using the same insulin infusion rate (0.4 mU · kg⁻¹ · min⁻¹) in the fructose and control studies, plasma glucose could not be lowered to the desired glucose step in the fructose studies, confirming that the glucose counterregulatory systems were activated.

Glucose infusion rates in the experimental protocol are depicted in Fig. 1B. During the first and second glucose steps, average glucose infusion rates were comparable in both studies (7.2 ± 1.1 and 7.7 ± 1.1 μmol · kg⁻¹ · min⁻¹ in the fructose and control studies, respectively, P = NS). However, during the 3.9-mmol/l glucose step, the mean rate of glucose infusion was lower in the fructose studies (4.6 ± 0.9 vs. 7.4 ± 1.1 μmol · kg⁻¹ · min⁻¹, P = 0.03), and during the 3.3-mmol/l glucose step, the average glucose infusion rate was 0.5 ± 0.1 in the fructose studies versus 5.2 ± 1.2 μmol · kg⁻¹ · min⁻¹in the control studies (P < 0.001). It is important to emphasize that during the final 20 min of the clamp (t = 180–200 min), glucose infusion was stopped in all fructose studies because plasma glucose concentration was above the desired (3.3 mmol/l) glucose step.

Basal plasma insulin concentrations were nearly identical in both studies, averaging 24.6 ± 7.2 pmol/l in the fructose studies and 30.6 ± 13.2 pmol/l in the control studies (P = NS). Similarly, there was no significant difference in plasma insulin concentration during both clamps, averaging 186.6 ± 8.4 pmol/l in the fructose studies and 180 ± 6.8 pmol/l in the control studies (Fig. 2A). Plasma C-peptide concentrations were comparable in both sets of studies at baseline (0.49 ± 0.06 nmol/l in the fructose studies and 0.52 ± 0.06 pmol/l in the control studies, P = NS) and remained very similar during the clamp, suppressing in both studies to 0.03 ± 0.01 pmol/l at the hypoglycemic nadir, at t = 200 min (Fig. 2B).

During the 5.0- and 4.4- mmol/l glucose steps, plasma epinephrine concentrations remained near or at basal values and were similar in both fructose and control studies (193 ± 6 pmol/l with fructose and 219 ± 14 pmol/l without fructose, P = NS [Table 1]). Further reduction in plasma glucose to 3.9 mmol/l was associated with an increment in plasma epinephrine in both sets of studies, though the threshold for epinephrine release occurred at a higher plasma glucose concentration in the fructose than in the control studies (4.0 ± 0.1 vs. 3.8 ± 0.1 mmol/l, respectively, P = 0.006) (Fig. 3A). Moreover, the average maximal increase in plasma epinephrine was also significantly higher during the 3.9- and 3.3-mmol/l glucose steps in the fructose than in the control studies, equaling 1,152 ± 131 and 797 ± 93 pmol/l, respectively, during the 3.9 mmol/l glucose step, and 3,411 ± 448 pmol/l and 2,784 ± 333 pmol/l, respectively, during the 3.3 mmol/l glucose step (both P < 0.05). The area under the epinephrine curve during the whole duration of the clamps was also significantly higher in the fructose than in the control studies (2,855 ± 202 vs. 2,336 ± 142 pmol · l⁻¹ · h⁻¹, P < 0.05) (Fig. 3B).

Plasma glucagon concentrations were equivalent during the 5.0- and 4.4-mmol/l glucose steps in both studies (Table 1). However, the threshold for glucagon release occurred at a significantly higher plasma glucose concentration in the fructose than the control studies (4.1 ± 0.1 and 3.9 ± 0.2 mmol/l, respectively, P = 0.03) (Fig. 4A). Similarly, the magnitude of glucagon release during the 3.9- and 3.3- mmol/l glucose steps was significantly higher in the fructose than in the control studies (117 ± 12 and 185 ± 14 ng/l in the fructose studies, respectively, and 97 ± 14 and 133 ± 17 ng/l in the control studies, respectively (P < 0.05 for both). The area under the glucagon response curve was higher in the fructose than in the control studies (292 ± 28 vs. 215 ± 22 ng · l⁻¹ · h⁻¹, P < 0.05) (Fig. 4B).

In contrast to the amplified epinephrine and glucagon responses in the fructose studies as compared with the control studies, plasma norepinephrine, growth hormone, and cortisol levels were similar in both sets of studies (Table 1).

Plasma lactate concentrations were similar in both studies and averaged 0.8 ± 0.1 and 0.8 ± 0.2 μmol/l (P = NS) in the fructose and control studies, respectively, at baseline and 0.9 ± 0.1 and 1.1 ± 0.2 μmol/l (P = NS) in the fructose and control studies, respectively, during the 3.3-mmol/l nominal glucose step.

[3-³H]glucose specific activity was effectively maintained in both sets of studies during the clamps (Table 2).

Mean fasting EGP was similar in the two studies (10.1 ± 1.1 and 10.5 ± 1.1 μmol · kg⁻¹ · min⁻¹ in the fructose and control studies, respectively, P = NS). During the 5.0- and 4.4-mmol/l glucose steps (after the initiation of insulin infusion), EGP was equally suppressed by ∼45% in both studies. During the next two clamp steps (3.9 and 3.3 mmol/l), EGP increased to 78 and 83%, respectively, of basal in the fructose studies and to 47 and 62%, respectively, in the control studies (P < 0.05 for both) (Fig. 5A), demonstrating a significant amplification of EGP in the fructose studies.

Glucose disposal rates were similar before insulin infusion in all studies. With the initiation of the clamp, glucose disposal rates increased equally in the fructose and control studies by ∼85% (to 20.1 ± 1.8 and 21.2 ± 1.2 μmol · kg⁻¹ · min⁻¹, respectively, P = NS). Consequently, during the 3.9- and 3.3-mmol/l glucose steps, glucose disposal rates decreased by ∼30% in both studies. However, the fructose studies demonstrated a more significant decrease in glucose uptake (R_d) than the control studies (11.8 ± 1.4 and 14.1 ± 0.9 μmol · kg⁻¹ · min⁻¹, respectively, P < 0.05, during the 3.9-mmol/l glucose step and 12.8 ± 1.8 and 15.7 ± 1.6 μmol · kg⁻¹ · min⁻¹, respectively, P < 0.05, during the 3.3-mmol/l glucose step) (Fig. 5B). It should be noted that the reduction in R_d during the fructose studies at the 3.3-mmol/l glucose step may be underestimated because plasma glucose was not lowered to the same level as in the control studies.

To exclude an effect of fructose on nonhypoglycemic hepatic glucose fluxes, we performed these studies at two plasma glucose steps: 5.0 and 4.2 mmol/l. Thus, we examined the effects of fructose on EGP excluding the action of the counterregulatory hormones. The study protocol and plasma glucose levels during both clamps (with and without co-infusion of fructose) were well maintained and similar during both clamps (Fig. 6). Mean plasma glucose concentrations at the nominal 5.0- and 4.2-mmol/l glucose steps were 5.2 ± 0.5 and 4.2 ± 0.2 mmol/l in the fructose studies and 5.1 ± 0.4 and 4.2 ± 0.1 mmol/l in the control studies, respectively (P = NS).

Plasma insulin, C-peptide, glucagon, epinephrine, norepinephrine, growth hormone, and cortisol concentrations were comparable in both clamps and are shown in Table 3.

Plasma lactate concentrations were similar in both studies and averaged 0.8 ± 0.2 and 0.7 ± 0.1 μmol/l in the fructose and control studies, respectively, at baseline and 0.9 ± 0.2 and 0.8 ± 0.3 μmol/l in the fructose and control studies, respectively, during the 4.2 mmol/l nominal glucose step.

The mean rates of EGP (Fig. 7) were similar at baseline (10.5 ± 0.9 and 10.6 ± 0.5 μmol · kg⁻¹ · min⁻¹ for fructose and control studies, respectively, P = NS), during the nominal 5.0-mmol/l glucose step (5.2 ± 0.9 and 4.7 ± 0.6 μmol · kg⁻¹ · min⁻¹ for fructose and control studies, respectively, P = NS), and during the nominal 4.2 mmol/l glucose step (3.9 ± 0.9 and 3.7 ± 0.6 μmol · kg⁻¹ · min⁻¹ for fructose and control studies, respectively, P = NS). Glucose uptake was also comparable (Fig. 7) in both clamps at baseline (9.8 ± 1.2 and 10.1 ± 0.8 μmol · kg⁻¹ · min⁻¹ for fructose and control studies, respectively, P = NS), during the nominal 5.0 mmol/l glucose step (15.7 ± 1.7 and 14.8 ± 1.3 μmol · kg⁻¹ · min⁻¹ for fructose and control studies, respectively, P = NS), and during the nominal 4.2-mmol/l glucose step (14.9 ± 1.4 and 15.8 ± 1.6 μmol · kg⁻¹ · min⁻¹ for fructose and control studies, respectively, P = NS).

Our results demonstrate that intravenous infusion of catalytic amounts of fructose induce an amplification of the counterregulatory response to mild hypoglycemia in nondiabetic subjects. The augmentation of hormonal counterregulation was characterized by both a shift in the threshold for epinephrine and glucagon secretion (a significant release of these hormones occurred at a higher plasma glucose concentration) and a higher magnitude of epinephrine and glucagon release (as reflected by both peak and AUC concentrations) compared with matched control studies. These modestly augmented hormonal responses to hypoglycemia were accompanied by a markedly amplified EGP response and a relative greater reduction of glucose uptake. Indeed, because of this enhanced counterregulatory response in the fructose infusion studies, we were unable to lower the plasma glucose levels to <3.8 mmol/l. Thus, despite the same plasma insulin concentrations (∼180 pmol/l in both studies), we could not attain the final glucose step (3.3 mmol/l target plasma glucose) in the fructose infusion studies, a factor that contributes to underestimation of the magnitude of the counterregulatory response in these studies but emphasizes the potency of the enhanced glucose counterregulation in preventing further hypoglycemia. The fact that the activation of EGP with fructose was disproportionately higher than the difference in the thresholds and magnitudes of epinephrine and glucagon release suggests that fructose may have an additional and independent effect on EGP during mild hypoglycemia. Thus, the fructose effect on glucose counterregulation in the stepped hypoglycemia studies could be attributed to the augmentation of epinephrine and glucagon responses and also to direct effect of hypoglycemia per se. Interestingly, the rates of EGP and glucose uptake did not change with fructose infusion (compared with saline) during the pancreatic clamp when the counterregulatory hormones were at baseline, although the plasma glucose levels in this protocol were clamped above the threshold for triggering hypoglycemia counterregulation.

Accumulating evidence suggests that the brain—in particular the VMH—plays an important role in glucose sensing. In dogs, cerebral perfusion of glucose abolishes the counterregulatory response in the presence of peripheral hypoglycemia (33). In rats, selective destruction of the VMH by injection of ibotenic acid or localized perfusion of concentrated glucose solutions into the ventromedial nuclei abolishes the counterregulatory response (34,35). In contrast, selective glycopenia in the cells within the VMH by perfusion of 2-deoxyglucose activates counterregulation, even in the presence of normal circulating blood glucose concentration (36). Unfortunately, however, there are no data regarding these potential sites for hypoglycemia sensing and integration of the counterregulatory response in humans.

The mechanism by which fructose infusion might augment the counterregulatory response to hypoglycemia is not known. In vitro studies have shown that fructose-1-phosphate can activate glucokinase translocation by binding to the GKRP. GKRP inhibits GK in an allosteric manner that is competitive with respect to glucose. Conversely, fructose-6-phosphate binds to GKRP and greatly enhances its binding to and inhibition of GK, whereas fructose-1-phosphate promotes the dissociation of this complex and activates GK. Under basal glucose conditions, GKRP inhibits GK activity by its sequestration in the perinuclear area, thus rendering it inaccessible to incoming glucose. Fructose-1-phosphate, generated from fructose through the action of hepatic FK, binds to the regulatory protein and thereby decreases its affinity for GK. This alters the configuration of GK and permits its immediate translocation to the cytosol, both of which are responsible for increased enzyme activity. The operation of this rapid regulatory mechanism on EGP during hyperglycemia has been recently demonstrated in conscious dogs by administering small amounts of fructose either intraportally or intraduodenally (26,27). However, in cells that lack FK, GK phosphorylates fructose (via hexokinase) into fructose-6-phosphate, which may severely inhibit GK activity.

With respect to the role of GK in hypoglycemia, there is a paucity of published data the normal physiology in humans. Interestingly, patients with maturity-onset diabetes of the young (MODY)-II, whose mutation results in a GK protein with decreased activity in the β-cell (and presumably in the liver), appear to have an amplified counterregulatory response to low blood glucose, underscoring the potential role of GK in the defense against hypoglycemia (37). Specifically, in MODY II patients, Guenat et al. (37) reported higher glycemic thresholds for glucagon release and perhaps for EGP activation compared with nondiabetic subjects and patients with type 2 diabetes. These findings suggest that a partial decrease in GK activity in these patients may influence hypoglycemia counterregulation by acting at the level of hypoglycemia sensing because only the thresholds for (and not the magnitude of) glucagon and EGP were altered. The same authors performed similar studies in subjects with MODY III (hepatocyte nuclear factor-1α [HNF-1α] mutations) and demonstrated no significant difference in the counterregulatory response to hypoglycemia compared with nondiabetic subjects and patients with type 2 diabetes, suggesting that mutations of HNF-1α in MODY III do not modulate counterregulatory responses to hypoglycemia (38).

We have used fructose as a pharmacological modulator of GK during hypoglycemia in nondiabetic subjects. An acute infusion of low-dose fructose induced a shift in the thresholds for epinephrine and glucagon release to a higher plasma glucose concentration and an increase in the magnitude of hormone secretion. Most importantly, glucose counterregulation (characterized by an increase in EGP) increased significantly and dramatically with fructose infusion. This remarkable effect of fructose infusion on hypoglycemia counterregulation could be explained, in part, by an apparent inhibitory action of fructose on GK activity in cells that lack FK (e.g., hypoglycemia-sensing cells in the hypothalamus). In support of this hypothesis, it has been demonstrated that certain areas in the brain considered to be responsible for glucose sensing, such as the VMH, possess glucose-sensitive neurons that increase their firing rate when ambient glucose decreases (39). These neurons also possess the key biochemical elements required for glucose sensing (GK, GKRP, and K_ATP channels) (16,40). Although fructose does not pass readily through the blood brain barrier, these hypothalamic glucose-sensitive neurons may be exposed to the plasma glucose concentrations prevailing in the systemic circulation and thus respond to physiological variations in plasma glucose concentration (39).

Another mechanism that may be responsible for the increase in EGP with fructose infusion during hypoglycemia is by transient activation of hepatic GK to increase the glycogen pool (G-6-P→G-1-P) for further utilization. In support of this mechanism, Petersen et al. (24) have recently demonstrated that infusion of a low dose of fructose to nondiabetic subjects during a hyperinsulinemic-euglycemic clamp induces a threefold increase in net hepatic glycogen synthesis. Because net hepatic glycogen synthesis has been shown to be diminished in poorly controlled type 1 diabetic patients (41), previous activation of glycogen synthase and increasing net hepatic glycogen content with fructose infusion could positively contribute to EGP during subsequent hypoglycemia. Taken together, our data suggest that fructose has a potent modulatory effect on GK, inducing a significant amplification of the counterregulatory response to hypoglycemia in nondiabetic subjects. This effect may be attributed to fructose’s action on the counterregulatory hormonal response and also to a potentially direct effect on EGP. Indeed, our results clearly demonstrate that fructose administration in the face of impending hypoglycemia induces an earlier and more powerful counterregulatory response compared with control. Furthermore, the significant increase in glucose production during the 3.9- and 3.3-mmol/l glucose steps was present in all subjects, and our inability to lower the plasma glucose to the target glycemic level (3.3 mmol/l) in the fructose infusion studies was also noted in all subjects, factors that further strengthen our findings.

The fact that in the fructose infusion studies both epinephrine and glucagon responses were amplified suggests that fructose action occurred at the level of glucose sensing or at a more proximal integration center. Amplification of the glucagon response only, the epinephrine secretion only, or the EGP response alone could have suggested a local effect on the α-cell, autonomic activation, or CNS (or hepatic) regulation of EGP, respectively.

In conclusion, this study describes a novel finding that the counterregulatory hormonal response to hypoglycemia can be enhanced or reset in nondiabetic subjects with fructose. The augmentation of the counterregulatory response was characterized by a shift in the glycemic threshold for epinephrine and glucagon secretion (which occurred at a higher plasma glucose level) and an increased magnitude of these hormones. Augmenting the counterregulatory response to hypoglycemia could have major implications in the management of patients with type 1 diabetes in whom intensive insulin treatment, although advantageous, is often avoided because of the risk of developing defective counterregulation and severe hypoglycemia. However, whether fructose has the same effects on hypoglycemia counterregulation in patients with type 1 diabetes remains to be determined.

This study was supported by grants from the National Institutes of Health (DK 20541 to L.R. and H.S. and DK 4502411 to L.R.) and the Endocrine Fellows Foundation (to I.G.).

The authors would like to acknowledge the nursing staff of the Albert Einstein General Clinical Research Center (RR12248) for their superb care of the subjects.