Effect of Oral Amino Acids on Counterregulatory Responses and Cognitive Function During Insulin-Induced Hypoglycemia in Nondiabetic and Type 1 Diabetic People

Institutional review board approval was obtained for these studies. Ten healthy nondiabetic volunteers (5 men, age 32 ± 7 years, BMI 23 ± 2 kg/m², C-peptide 1.2 ± 0.2 nmol/l, A1C 4.6 ± 0.3% [means ± SD]) were studied along with 10 subjects with type 1 diabetes on long-term intensive insulin treatment (6 men, age 30 ± 8 years, diabetes duration 17 ± 7.8 years, BMI 22.7 ± 1.8 kg/m², A1C 7.4 ± 1.0% [means ± SD]). At the time of the study, all type 1 diabetic subjects were free of any detectable microangiopathic complication and were negative at the screening for autonomic neuropathy, as judged on the basis of a standard battery of cardiovascular tests (21).

The study was carried out according to the Declaration of Helsinki after obtaining written informed consent from all subjects. All nondiabetic and diabetic volunteers were studied on three different occasions in a random, computer-generated sequence at 2- to 3-week intervals with the modified hyperinsulinemic glucose clamp technique used to either maintain euglycemia or induce hypoglycemia. In diabetic subjects, care was taken to avoid preprandial, postprandial, and nocturnal blood glucose <72 mg/dl over the week before studies as previously reported (22). Briefly, the glycemic targets were blood glucose 110–130 mg/dl in the fasting state, before meals, and at bedtime, and blood glucose 140–180 mg/dl 2 h after meals. Blood glucose was measured in the fasting state, before meals, and 2 h after meals and three times a week at 0300 h. In addition, blood glucose was measured whenever subjects believed their sugar was low. Patients were advised to decrease or increase the dose of basal insulin if their fasting blood glucose was repeatedly <110 mg/dl or >130 mg/dl, and to decrease or increase the dose of rapid-acting insulin at meals if their 2-h postprandial blood glucose was repeatedly <140 mg/dl or >180 mg/dl. Adjustments of rapid-acting insulin dose were made according to the carbohydrate content of meals. Four subjects had six values of blood glucose between 55 and 72 mg/dl in the week before the study, which caused the postponing of the study to the next week. Finally, to avoid hypoglycemia the day before the experiment, the total daily insulin dosage was cut by 20%, and patients were asked to contact one investigator by phone to receive advice on insulin doses.

On the morning of the studies, all nondiabetic and diabetic subjects were admitted to the General Clinical Research Center of the Department of Internal Medicine at ∼0700 h. A hand vein of the nondominant arm was cannulated retrogradely and maintained in a hot box (∼60°C) for sampling of arterialized-venous blood (23). A superficial vein of the ipsilateral arm was also cannulated for infusion of insulin and glucose (see below). The two veins were maintained patent by means of 0.9% NaCl infusion (0.5 ml/min). In the diabetic subjects, an intravenous infusion of human regular insulin (diluted to 1 unit/ml in 2 ml of the subject's blood and 0.9% NaCl to a final volume of 100 ml) was begun at ∼0730 h in a feedback fashion, using a syringe pump (Harvard Apparatus, Ealing, South Natick, MA), according to an algorithm described previously (24), to reach the target plasma glucose of 100 mg/dl by 0830 h.

In all studies at time 0900 h (time 0 min), in both nondiabetic and diabetic subjects, intravenous insulin at the rate of 2 mU · kg⁻¹ · min⁻¹ was started and continued until the end of the study (i.e., 160 min). Intravenous glucose at variable rate was also infused to maintain euglycemia (plasma glucose at 90 mg/dl) throughout the study on one occasion (clamped euglycemic study), whereas on the other two occasions, the rate of glucose infusion was decreased after 30 min to allow plasma glucose to fall and reach the target plasma glucose of 47 mg/dl at 120 min. This hypoglycemic plateau was maintained for the next 40 min, that is, until the end of the study (clamped hypoglycemic studies).

On each occasion, at 30 min, subjects ingested over a 5-min period a 200-ml drink containing either a mixture of amino acids (3.1 g lysine, 1.7 g histidine, 3.1 g arginine, 2.8 g aspartate, 2.3 g threonine, 2 g serine, 1 g glutamate, 3.3 g proline, 2.7 g glycine, 1.7 g alanine, 1.1 g cysteine, 3 g valine, 0.7 g methionine, 2.7 g isoleucine, 4.6 g leucine, 4.1 g tyrosine, 0.5 g phenylalanine, 0.9 g tryptophan, and 2.1 g glutamine) or a seemingly identical placebo rendered palatable by flavoring it with artificial fruit flavor and adding 2 g sucrose and 1 g aspartame. All drinks were prepared and administered by a research nurse not involved in the further execution of the study. The mixture of amino acids was given both in the clamped euglycemia study (euglycemia–amino acid study) and in one of the two clamped hypoglycemia studies (hypoglycemia–amino acid study), whereas in the other clamped hypoglycemia study subjects ingested placebo (hypoglycemia-placebo study).

In all studies, blood samples were drawn at 5- to 10-min intervals for bedside plasma glucose measurement and at 30-min intervals for measurement of plasma insulin, C-peptide, pancreatic polypeptide, amino acid, counterregulatory hormone, and nonglucose substrate concentration (see below).

A semiquantitative symptom questionnaire (25) was administered every 30 min. Subjects were asked to score from 0 (none) to 5 (severe) on each of the following symptoms: seven autonomic/neurogenic (adrenergic: heart pounding, tremor, anxiety, and irritability; cholinergic: sweating, hunger, and tingling); five neuroglycopenic (difficulty in thinking, weakness, dizziness, blurred vision, and drowsiness); and three nonspecific (thirst, nausea, and headache) (26). The sum of each of these constituted the total symptom score.

In addition, at baseline, before inducing hypoglycemia, and at the hypoglycemic plateau (indicated as “time −30,” “time 0,” and “time 120,” respectively) and at similar times during the euglycemic studies, cognitive function was assessed by applying a battery of hypoglycemia-sensitive tests: Trail-Making parts A and B (27); verbal fluency (28); verbal memory (28); digit vigilance (28); forward and backward digit span (29); Stroop word, color, and color-word (interference) subtests (30); and the Paced Auditory Serial Addition Test (PASAT; 2 and 3 s) (31). Tests were always performed in this order.

Plasma glucose was measured by means of a Beckman glucose analyzer (Glucose Analyzer II; Beckman Instruments, Fullerton, CA). Plasma insulin, C-peptide, pancreatic polypeptide, counterregulatory hormone, glucagon, adrenaline, norepinephrine, glycerol, β-OH-butyrate, lactate, and alanine were measured by previously described assays (32). Plasma concentrations of amino acids were measured by high-performance liquid chromatography with postcolumn o-phalaldehyde derivatization (33). To remove antibody-bound insulin, plasma was mixed with an equal volume of 30% polyethylene glycol immediately after blood collection in both type 1 diabetic and nondiabetic subjects (34). A1C was determined by high-performance liquid chromatography using a Hi-AUTO A1C, TM HA 8121 apparatus (DIC; Kyoto Daiichi, Kogaku, Japan). Plasma free fatty acid (FFA) concentrations were measured using a commercial kit (Wako NEFA C test kit; Wako Chemicals, Neuss, Germany).

All data were subjected to repeated-measures ANOVA with Huynh-Feldt adjustment for nonsphericity (35). The ANOVA model included the sequence of studies as the between-subjects factor, whereas test condition (hypoglycemia-placebo/hypoglycemia–amino acid/euglycemia–amino acid) and time were the within-subjects factors. Subjects were entered into the model as random factors. The factor group (nondiabetic/diabetic subjects) was entered into the model as a between-subjects factor. If there were significant differences between baseline values, these were used as covariates. In this way, the data over the serial time points could be adjusted for any differences in baseline values (35). Post hoc comparisons (Newman-Keuls test) were carried out to pinpoint specific differences on significant interaction terms.

The area under the curves (AUCs) for counterregulatory hormones and substrates at the clamped hypoglycemia period (120–160 min) were calculated according to the trapezoidal rule. The incremental AUC (iAUC) for glucagon response was calculated by subtracting plasma glucagon values before amino acid ingestion from the total AUC.

A modified Bonferroni procedure (36) for multiple cognitive test adjustments was used to maintain an overall type 1 error rate of 5% (α = 0.05).

Data are given as means ± SE, except where SD is specified. We considered differences to be statistically significant if the P value was 0.05 or less. We conducted the statistical analyses by using NCSS 2007 software (Kaysville, UT) and Statistica software, version 6.0 (StatSoft, Tulsa, OK).

In all study conditions, plasma glucose was maintained at the preselected plateaus, with no differences between groups and study conditions (P > 0.2).

Plasma insulin concentrations achieved during the study did not differ between nondiabetic and diabetic subjects or in relation to treatment conditions. On average, baseline plasma insulin concentrations were lower in nondiabetic compared with diabetic subjects (P < 0.001) in all study conditions.

The rates of glucose infusion were lower during the hypoglycemia plateau in the hypoglycemia–amino acid study compared with hypoglycemia-placebo and euglycemia–amino acid studies in nondiabetic subjects (2.6 ± 0.4 vs. 5.5 ± 0.6 and 8.8 ± 0.9 mg · kg⁻¹ · min⁻¹, respectively, P < 0.001) and in diabetic subjects (3.1 ± 0.7 vs. 5.4 ± 0.5 and 9.4 ± 0.9 mg · kg⁻¹ · min⁻¹, respectively, P < 0.001). Glucose infusion rates were not different between nondiabetic and diabetic subjects (P = 0.661).

Plasma glucagon concentrations were similar at baseline in nondiabetic and diabetic subjects in all study conditions. In the hypoglycemia-placebo study, plasma glucagon concentrations after an initial decrease at 60 min increased (P = 0.001) in nondiabetic subjects, whereas it did not increase in diabetic subjects (P = 0.867). The ingestion of amino acids in the hypoglycemia–amino acid study increased significantly the response of glucagon to hypoglycemia in nondiabetic and, to a lesser extent, in diabetic subjects compared with the hypoglycemia-placebo study (Table 1). Amino acids stimulated glucagon response also in the euglycemia–amino acid study, which was maximal between 75–120 min and then tended to decrease to baseline values at 120–160 min. In diabetic subjects, the response of glucagon to hypoglycemia stimulated by amino acids was lower than that of nondiabetic subjects. However, it was superimposable on that of nondiabetic subjects in the hypoglycemia-placebo study (P = 0.581; Table 1). Interestingly, although the iAUC of glucagon response in the hypoglycemia–amino acid study was 4.7 times greater in nondiabetic compared with diabetic subjects (7,468 ± 835 and 1,599 ± 142 ng/l, respectively, P < 0.001), the estimated contribution of amino acids per se in directly stimulating glucagon secretion, as derived from the euglycemia–amino acid study, was not statistically different in both nondiabetic and diabetic subjects (19 ± 4 and 35 ± 7%, respectively, P = 0.061).

Plasma C-peptide concentrations in nondiabetic subjects decreased at 30 min in all studies and then remained suppressed in the hypoglycemia-placebo study, whereas they increased after the ingestion of amino acids more in the euglycemia–amino acid than in the hypoglycemia–amino acid study (P < 0.05, Fig. 2). By the end of the study, however, C-peptide concentrations returned to basal values and were suppressed in the euglycemia–amino acid and hypoglycemia–amino acid studies, respectively. Plasma C-peptide concentrations were undetectable in diabetic subjects in all studies.

Plasma pancreatic polypeptide concentration increased during hypoglycemia-placebo and hypoglycemia–amino acid studies in both nondiabetic and diabetic subjects, although the peak response was higher in the hypoglycemia–amino acid study in nondiabetic subjects but not in diabetic subjects (Table 1). Pancreatic polypeptide levels did not change in euglycemia–amino acid studies.

Plasma adrenaline levels increased in the hypoglycemia-placebo and hypoglycemia–amino acid studies in both nondiabetic and diabetic subjects. However, responses were lower in diabetic compared with nondiabetic subjects (Table 1; P < 0.05). In the euglycemia–amino acid study, adrenaline levels did not change compared with basal values in both groups (Table 1).

Plasma norepinephrine concentrations were not different in all study conditions in both nondiabetic and diabetic subjects (Table 1; P > 0.2).

Responses of plasma cortisol and growth hormone increased similarly in response to hypoglycemia-placebo and hypoglycemia–amino acid compared with euglycemia–amino acid in both nondiabetic and diabetic subjects (Fig. 3; Table 1).

Plasma FFA levels decreased similarly in all studies in both nondiabetic and diabetic subjects. However, FFA concentrations at 120–160 min were higher in the hypoglycemia-placebo and hypoglycemia–amino acid compared with euglycemia–amino acid studies (P = 0.018) in nondiabetic subjects. They were not different in diabetic subjects (P = 0.712).

Plasma glycerol concentrations decreased from basal values in all studies in both nondiabetic and diabetic subjects. However, similar to FFA, glycerol levels were higher in the hypoglycemia-placebo and hypoglycemia–amino acid compared with euglycemia–amino acid studies (P = 0.018) only in nondiabetic subjects.

Plasma β-OH-butyrate concentrations, after an initial decrease to nadir values at 90 and 60 min in nondiabetic and diabetic subjects, respectively, increased in the hypoglycemia–amino acid and euglycemia–amino acid but not in the hypoglycemia-placebo studies in both nondiabetic (P < 0.001) and diabetic subjects (P < 001).

Plasma lactate concentrations increased in all studies in both nondiabetic and diabetic subjects. In nondiabetic subjects, plasma lactate increased more in the hypoglycemia-placebo compared with both the hypoglycemia–amino acid (P = 0.025) and the euglycemia–amino acid (P = 0.008) studies. In diabetic subjects, lactate was similar in the hypoglycemia-placebo, hypoglycemia–amino acid, and euglycemia–amino acid studies (P > 0.2). Overall, plasma lactate response in nondiabetic was greater than that of diabetic subjects (nondiabetic vs. diabetic subjects: hypoglycemia-placebo, P < 0.001; hypoglycemia–amino acid, P = 0.005; and euglycemia–amino acid, P = 0.046).

Branched and nonbranched chain amino acid (BCAA and N-BCAA, respectively) concentrations were similar at baseline in nondiabetic and diabetic subjects in all studies (Fig. 5). After the ingestion of amino acids, both BCAA and N-BCAA levels increased similarly in the hypoglycemia–amino acid and euglycemia–amino acid studies. In the hypoglycemia-placebo studies, both BCCA and N-BCCA concentrations tended to decrease from basal values by the end of study with no difference between the two groups (Fig. 5).

The score of autonomic and neuroglycopenic symptoms increased in the hypoglycemia-placebo and hypoglycemia–amino acid studies with no difference (P > 0.2) between studies in both nondiabetic and diabetic subjects, whereas it did not change in the euglycemia–amino acid studies in nondiabetic and diabetic subjects (Fig. 6).

With the exception of Trail-Making part A, digit vigilance, and Stroop words and colors tests in nondiabetic subjects and of Trail-Making part A and digit vigilance tests in diabetic subjects, all cognitive tests deteriorated significantly during hypoglycemia compared with euglycemia–amino acid, in both hypoglycemia-placebo and hypoglycemia–amino acid (Table 2). However, the degree of deterioration was lower in hypoglycemia–amino acid than in hypoglycemia-placebo in the following tests: Trail-Making part B, PASAT (2 s), digit span forward, and Stroop colored words and verbal memory tests in nondiabetic subjects and Trail-Making part B, digit span backwards, and Stroop color tests in diabetic subjects.

The present study was undertaken to examine the effects of the ingestion of a mixture of amino acids on the counterregulatory, symptomatic, and cognitive responses to hypoglycemia. Both nondiabetic and type 1 diabetic subjects were studied. First, the results indicate that oral amino acids enhance glucagon response to hypoglycemia in both nondiabetic and diabetic subjects, although to a lesser extent in the latter. Second, oral amino acids affect the responses of β-cells of pancreatic islets, as shown by the lower suppression of C-peptide. Third, amino acids preserve several aspects of cognitive responses to hypoglycemia, which, to the best of our knowledge, is a novel finding.

In type 1 diabetic subjects, the glucagon response to hypoglycemia with amino acid administration was nearly similar to the glucagon response to hypoglycemia alone (without amino acid administration) in the nondiabetic subjects. Such a result is in line with recent evidence indicating that the sensitivity of pancreatic α-cells to amino acids is markedly reduced in diabetic subjects compared with nondiabetic subjects and that a near-normal glucagon response is seen when the amino acid alanine is infused in diabetic subjects in hypoglycemia (13).

Interestingly, the near-normal response of glucagon to hypoglycemia after oral amino acids observed in the present study in diabetic subjects is also similar to that described after ingestion of a mixed meal (15). It is currently believed that the protein component of the meal is responsible for such an effect. As reported previously, the mechanisms of amino acid–induced amplification of the glucagon response to hypoglycemia likely involve both direct and indirect effects (13). In this study, a direct stimulatory effect of oral amino acids on glucagon response was evident at 60–90 min in both the euglycemic clamp and the hypoglycemic clamp before hypoglycemia was established. However, the stimulatory effect of oral amino acids on glucagon release was modulated by plasma glucose concentration. It was amplified by the induction of hypoglycemia in both nondiabetic and diabetic subjects but suppressed by euglycemia in both groups of subjects. This indicates that amino acids directly stimulate glucagon secretion from the α-cell (37) and that hypoglycemia amplifies the direct stimulatory effect of amino acids on glucagon release. Furthermore, oral amino acids may also act through different indirect mechanisms that involve gastroenteric peptides such as glucagon-like peptide 1 (38) and a decrease of tonic intra-islet α-cell inhibition by insulin and hypoglycemia (39). With regard to the latter, it is possible that the initially greater stimulation of β-cell function by amino acids was followed by a greater decrement in intra-islet insulin concentration during hypoglycemia, resulting in a greater decrease in tonic intra-islet α-cell inhibition by insulin and signaling for increased glucagon secretion during hypoglycemia in normal nondiabetic subjects (39). Most likely, the decrement in intra-islet insulin may be an important factor for the increase in glucagon secretion during hypoglycemia (40).

The ingestion of amino acids induced a greater peak response of pancreatic polypeptide compared with placebo in the hypoglycemic clamp studies in nondiabetic but not in diabetic subjects. It is well known that pancreatic polypeptide secretion after ingestion of a mixed meal is mostly mediated by the vagus nerve or by the extravagal cholinergic system (41). In addition, hypoglycemia per se represents a strong stimulus to pancreatic polypeptide secretion (41). Furthermore, autonomic vagal neuropathy blunts pancreatic polypeptide response to hypoglycemia (41). Although the diabetic subjects we studied had no evident signs of autonomic neuropathy, we cannot exclude the possibility that the lower response of pancreatic polypeptide concentrations to hypoglycemia in these subjects compared with nondiabetic subjects might be related to subclinical autonomic neuropathy or to other diabetes-related causes, including diabetes duration (41).

Elevation of plasma amino acid concentration after ingestion of amino acids did not affect counterregulatory hormones, with the exception of glucagon, and symptomatic responses to hypoglycemia, whereas the elevation clearly resulted in preservation of some aspects of cognitive function during hypoglycemia. This is a new finding that points toward a net effect of amino acids in supporting cognition while having no relevant effects on hormonal and symptomatic responses to hypoglycemia. It should be noted that previous studies indicate only a partial or no role of amino acids in sustaining cognition during hypoglycemia. In fact, Evans et al. (18) have shown that alanine infusion can sustain performance of Stroop word and color tests during hypoglycemia, although a contribution of elevated lactate levels to performance in those tests cannot be completely excluded (18). In contrast, M'Bemba et al. (12) found no effect of infusion of a dipepetide made of alanine and glutamine on a four-choice reaction time test. One possible explanation for a specific effect of amino acids on cognitive function is that elevated plasma amino acids do not affect metabolism and neurotransmission of brain centers that physiologically direct counterregulation and symptoms, whereas they may sustain brain areas capable of controlling hypoglycemia-induced deterioration of cognitive function. Therefore, in contrast to the role of other nonglucose substrates such as lactate (19) and ketones (20), amino acids limit their effect only to cognitive function. That would seem advantageous because the reduced counterregulatory hormone release and lower symptoms associated with lactate (19) or ketone (20) use by the brain might weaken defensive responses to hypoglycemia and generate hypoglycemia unawareness, a well-known risk factor for severe hypoglycemia (42), even in the face of preserved cognitive function.

Earlier evidence indicates that amino acids can serve as energy sources in the brain during insulin-induced hypoglycemia (43). Therefore, under such circumstances, they can be used by neurons and glia cells as an alternative substrate to glucose to limit the detrimental effect of glucopenia on cerebral metabolism and function. However, it has been recently shown that brain amino acid (alanine and leucine) uptake is not sufficient to offset energy deficit due to reduced glucose uptake during hypoglycemia (44) and that increased availability of amino acids does not result in net brain uptake of amino acids during hypoglycemia (45). Additionally, several amino acids such as tryptophan, tyrosine, phenylalanine, histidine, and arginine are used by the brain for the synthesis of various neurotransmitters and neuromodulators (46) that play a variety of functions in the brain, including a role in the regulation of mood state, fatigue, attention, and memory (47). Therefore, on theoretical grounds, our study result of preservation of some aspects of cognitive function during hypoglycemia supplemented with oral amino acids might be due either to a direct use of amino acids by neuronal cells to derive energy and maintain cognitive function or to their conversion to neurotransmitters and/or neuromodulators able to affect cognition. However, whichever the mechanism, levels of different classes of amino acids in the brain depend largely on their availability in blood, and when individuals are exposed to stressful conditions, such as hypoglycemia, brain requirements for specific amino acids may be particularly critical. In the present study, supplementation of oral amino acids not only prevented the insulin-induced fall in plasma amino acids observed in the hypoglycemia and placebo studies, it actually increased the overall concentrations of amino acids. Likely, these increased concentrations favored amino acid use by the brain.

The battery of tests that we used assessed a broad range of cognitve aspects. Overall, amino acid ingestion sustained cognitive domains pertaining to memory, attention, psychomotor efficiency, information processing, and immediate memory in nondiabetic subjects and all of these except immediate memory in diabetic subjects. This inter-group difference can be most likely attributed either to diabetes or to a general verbal ability and/or general intelligence condition. Notably, these results have been obtained with plasma amino acid concentrations in the physiological range of the postprandial condition in humans (∼2–3 mmol/l) (15).

Plasma β-OH-butyrate concentrations were suppressed in all studies. However, they were less suppressed after oral amino acid ingestion in both euglycemia and hypoglycemia. Their use by the brain as an alternative substrate to sustain brain metabolism and function during hypoglycemia cannot be completely excluded. However, this is unlikely because, first, the levels of β-OH-butyrate in our study were much lower than those that have been shown to be effective in maintaining cognitive function during hypoglycemia in previous studies (20), and second, β-OH-butyrate diminished counterregulatory and symptomatic responses to hypoglycemia in those studies (20) but not in our present study.

In conclusion, the present study indicates that oral amino acids improve some aspects of cognitive function in response to hypoglycemia and potentiate (in nondiabetic subjects) and recover (in subjects with type 1 diabetes) responses of glucagon to hypoglycemia. Additional studies are required to prove that these beneficial effects of oral amino acids during hypoglycemia translate into clinical advantages for people with type 1 diabetes and to possibly demonstrate that they contribute to reduction of the risk of severe hypoglycemia in type 1 diabetes.

This work received Juvenile Diabetes Research Foundation International Grant 1-2005-176 and National Institutes of Health Grant RO1-DK-41973.