Amino Acid Ingestion Strongly Enhances Insulin Secretion in Patients With Long-Term Type 2 Diabetes

Ten long-term diagnosed male type 2 diabetic patients were selected to participate in this study. Exclusion criteria were impaired renal or liver function, obesity (BMI >30 kg/m²), cardiac disease, hypertension, diabetes complications, and exogenous insulin therapy. Most subjects (n = 8) were using oral antidiabetic agents (mainly sulfonylureas). Ten healthy, normoglycemic, male adults, matched for age and BMI, participated as control subjects. In the diabetic subjects, blood glucose-lowering medication was witheld for 3 days before participation in the trials and throughout the entire experimental period. Subjects were screened for glucose intolerance/type 2 diabetes by an oral glucose tolerance test (OGTT) according to the World Health Organization criteria of 1999 (29). Subjects’ characteristics are provided in Table 1. All subjects were informed about the nature and risks of the experimental procedures before their informed consent was obtained. All clinical trials were approved by the local Medical Ethical Committee.

Before selection in the study, all subjects performed an OGTT. After an overnight fast, subjects reported to the lab and a blood sample was collected. Thereafter, a bolus of 75 g glucose (dissolved in 250 ml water) was ingested, and 2 h later a second blood sample was obtained (Table 1).

Each subject participated in two trials, separated by 1 week, in which the insulin response to the ingestion of two drink compositions (carbohydrate only [CHO] or carbohydrate plus the free amino acid/protein mixture [CHO+PRO]) was determined. Subjects were seated and remained inactive during both 3-h trials. Drinks were provided in randomized order and double-blind fashion. Beverages were vanilla flavored to make the taste comparable in all trials. Subjects were instructed to refrain from heavy physical labor and to keep their diet as constant as possible for 2 days before the trials.

After an overnight fast, subjects reported to the laboratory at 0830. A teflon catheter (Baxter BV, Utrecht, the Netherlands) was inserted into an antecubital vein, and a resting blood sample was drawn (0 min). Immediately thereafter, subjects drank an initial bolus (3 ml/kg) of a given test drink (CHO or CHO+PRO trial). Repeated boluses (3 ml/kg) were ingested every 30 min until 90 min. Blood samples were drawn at 15-min intervals for measurement of plasma glucose and insulin concentrations. In addition, growth hormone (GH) and cortisol concentrations were determined at 0, 60, and 120 min. IGF-I and IGF binding protein 3 (IGF-BP3) concentrations were determined at 0 and 120 min. In four subjects within each group, additional blood samples were collected at 2, 4, 6, 8, and 10 min for insulin and glucose analysis.

At 0, 30, 60, and 90 min, subjects received a beverage volume of 3 ml/kg to ensure a given dose of 0.7 g · kg⁻¹ · h⁻¹ carbohydrate (50% glucose and 50% maltodextrin) with or without an additional 0.35 g · kg⁻¹ · h⁻¹ of an amino acid/protein hydrolysate mixture (CHO+PRO or CHO trial, respectively). This hydrolysate/amino acid mixure consisted of a wheat protein hydrolysate (50%, 0.17 g · kg⁻¹ · h⁻¹), free leucine (25%, 0.09 g · kg⁻¹ · h⁻¹), and free phenylalanine (25%, 0.09 g · kg⁻¹ · h⁻¹). Glucose and maltodextrin were obtained from AVEBE (Veendam, the Netherlands), crystalline amino acids from BUFA (Uitgeest, the Netherlands), and the wheat protein hydrolysate was prepared by Quest International (Naarden, the Netherlands). Both drinks were uniformly flavored by adding 0.2 g sodiumsaccharinate, 1.8 g citric acid, and 5 g cream vanilla flavor (Quest International) per liter of beverage.

Blood (10 ml) was collected in EDTA-containing tubes and centrifuged at 1,000g and 4°C for 10 min. Aliquots of plasma were immediately frozen in liquid nitrogen and stored at −80°C. Glucose (Uni Kit III, 07367204; Roche, Basel) was analyzed with the COBAS FARA semiautomatic analyzer (Roche). Plasma insulin (Insulin RIA 100 kit; Pharmacia, Uppsala, Sweden) and glucagon (DPS, Los Angeles, CA) concentrations were analyzed by radioimmunoassay. Sequential immunometric assays were used to determine plasma GH (Nichols Institute Diagnostics, San Juan Capistrano, CA) and cortisol (Chiron Diagnostics, East Walpole, MA) concentrations. Plasma IGF-I was performed using an immunoradiometric assay (DSL-5600 ACTIVE; DSL Deutschland, Sinsheim, Germany) with an intra-assay variance of 3.0% and an interassay variance of 1.5%. Plasma IGF-BP3 concentrations were determined by another immunoradiometric assay (DSL-6600 ACTIVE; DSL Deutschland) with an intra-assay variance of 3.2% and an interassay variance of 0.5%.

Data are expressed as means ± SE. The plasma responses were calculated as areas under the curve above baseline values. Statistical analysis of the data were performed using multiway ANOVA. Changes in time within each group were checked for statistical significance using repeated-measures ANOVA. Differences between trials were determined for statistical significance using the Scheffé’s post hoc test (P < 0.05).

In both the CHO and CHO+PRO trials, fasting plasma insulin concentrations (0 min) were similar between groups. In all trials, plasma insulin concentrations increased significantly during the first 105 min, after which levels tended to decrease (Fig. 1A). After expressing the insulin response as area under the curve (above baseline values), we observed significant differences between both trials and groups (Fig. 1B). Coingestion of the protein hydrolysate/amino acid mixture (CHO+PRO) resulted in a substantial increase in insulin response that was an additional 189 and 114% higher than the ingestion of carbohydrates only (CHO) in the type 2 diabetic and healthy control subjects, respectively (P < 0.01). The additional increase in insulin response in the CHO+PRO trial was not significantly different between groups (3.77 ± 0.53 and 5.43 ± 1.03 mU · ml⁻¹ · 2 h⁻¹, respectively). The insulin responses in both the CHO and CHO+PRO trials were significantly lower in the type 2 diabetic subjects than in the normoglycemic control subjects (P < 0.01).

Fasting plasma glucose concentrations were substantially higher in the type 2 diabetic subjects than in the normoglycemic control subjects (11.6 ± 0.7 and 5.0 ± 0.1 mmol/l, respectively, P < 0.01). Plasma glucose concentrations increased from 0 min onward in the type 2 diabetic subjects in both the CHO and CHO+PRO trials and tended to plateau over the last 30 min of the 2-h period (Fig. 2A). No significant differences were observed within this 2-h time frame between the different trials in the diabetic group. In the normoglycemic control subjects, plasma glucose concentrations increased over the first 30–45 min, after which concentrations returned to baseline values (Fig. 2A). Though plasma glucose concentrations tended to be lower between 45 and 90 min in the CHO+PRO trial, no significant differences were observed between the CHO and CHO+PRO trials within this 2-h time frame (P = 0.06).

Plasma glucose concentrations remained substantially higher in the type 2 diabetic patients than in the control group in both the CHO and CHO+PRO trials (P < 0.01). After expressing the plasma glucose response as area under the curve (above baseline values), we observed no significant differences between trials in each group (Fig. 2B). However, both the glucose responses in the CHO and CHO+PRO trial were significantly higher in the type 2 diabetic subjects than in the normoglycemic control subjects (P < 0.01). In four subjects in each group, additional blood samples were taken at 2, 4, 6, 8, and 10 min to check for the potential prevalence of a primary secretory insulin response. However, on a whole-body level, we were unable to demonstrate such a response; both glucose and insulin concentrations increased during the first 10 min after ingestion of the first beverage but remained below the values recorded at 15 min (data not shown).

In both groups, plasma glucagon concentrations increased over time in all trials (P < 0.05) (Table 2). No significant differences were observed between groups. Within groups, a stronger increase in plasma glucagon in the CHO+PRO than the CHO trial was observed in the control group only (P < 0.05). Fasting plasma GH and cortisol concentrations tended to be higher in the type 2 diabetic subjects than in the control subjects (P = 0.10 and P = 0.14, respectively). In both the type 2 diabetic and control subjects, plasma GH concentrations decreased over time in both the CHO and CHO+PRO trials (Table 2). However, due to large variances in plasma GH concentrations, this time effect only reached statistical significance in the CHO+PRO trial in the diabetic group (P < 0.05). Plasma cortisol concentrations significantly decreased over time in both groups within both trials (P < 0.05). Further analysis of plasma GH, cortisol, IGF-I, and IGF-BP3 concentrations did not reveal any significant differences between groups or between trials. In four diabetic patients and four control subjects, IGF-BP3 concentrations were determined at several points in time, which did not show any change (data not shown).

The diabetic patients selected to participate in this study had been diagnosed with type 2 diabetes for >8 years and were about to switch from taking oral antidiabetic agents only (mainly sulfonylureas) to starting exogenous insulin therapy. In the selected patients, hyperinsulinemia, as a compensatory effect to reduce prevailing hyperglycemia, was not present anymore. Fasting plasma insulin concentrations were not different from values observed in the control group (Fig. 1), but glucose concentrations were substantially higher (Table 1 and Fig. 2). The OGTT revealed substantial glucose intolerance in the type 2 diabetic group (Table 1).

As hyperglycemia is no longer accompanied by compensatory hyperinsulinemia in long-term type 2 diabetic patients, it is often assumed that the insulin secretory capacity of the β-cells is substantially reduced. However, our data argue against this assumption. The insulin response following carbohydrate intake only (CHO) was less than half in the type 2 diabetic patients compared with the normoglycemic control subjects (P < 0.01) (Fig. 1) and clearly demonstrates the reduced sensitivity of the β-cell to glucose. However, the insulin response in these patients could be substantially increased (by an additional 189%) by coingestion of the amino acid/protein mixture (P < 0.01). In the control group, insulin responses increased by 114% in the CHO+PRO trial (P < 0.01). The additional increase in insulin response following coingestion of the free amino acid/protein mixture was not significantly different between groups, which seems to suggest that even in these long-term type 2 diabetic patients, insulin secretion in response to other stimuli is not impaired.

The increase in insulin concentration following the combined ingestion of carbohydrate with leucine, phenylalanine, and a protein hydrolysate has been shown to be highly correlated with the increase in plasma leucine and phenylalanine concentration (21). Substantial progress has recently been made in the understanding of the mechanisms and signaling pathways by which leucine (12,15–17), arginine (13–16,30), and phenylalanine and its derivatives (18) stimulate β-cell function and insulin secretion in in vitro studies using isolated β-cells and insulin-secreting pancreatic β-cell lines. Clearly, the large responsiveness to the coingestion of the insulinotropic amino acid/protein mixture in the present study strongly suggests that amino acids can provide a glucose-independent stimulus to the pancreatic β-cells that is preserved in patients with long-term type 2 diabetes.

Although insulin responses were strongly increased in the CHO+PRO trials, the mean glucose responses were not significantly reduced within this time frame, neither in the type 2 diabetic patients nor in the healthy control subjects (Fig. 2). This can be explained by the fact that in the present study, repeated boluses of a carbohydrate-containing solution were provided. The latter resulted in a substantial amount of carbohydrate being ingested over a relatively short period. Therefore, it is more than likely that a potential increase in plasma glucose disposal is being obscured by accelerated intestinal glucose absorption. Other studies with longer time frames have reported a plasma glucose-lowering effect after coingestion of protein and/or amino acids with carbohydrates, which is in line with our findings. For example, Gannon et al. (5) demonstrated an increase in insulin secretion following the combined ingestion of a single bolus of glucose with intact proteins (both egg protein and cottage cheese) in type 2 diabetic subjects and reported an 11 and 20% decrease in glucose area response over a 5-h postprandial period. The plasma glucose-lowering effect in that study became apparent mainly during the last 3 h. In accordance, we have reported a significant decrease in glucose response (−48%) following the coingestion of the applied amino acid/protein mixture in trained cyclists over a 5-h postexercise period (19).

Theoretically, the insulinotropic mixture used in the present study might stimulate glucagon secretion, as purified pancreatic α-cells have been shown to secrete glucagon in response to an amino acid mixture (31). To determine whether increased glucagon secretion in the CHO+PRO trials could have accelerated hepatic glycogen breakdown and thus contributed to the lack of a plasma glucose-lowering effect within the 2-h time frame in both groups, we also measured plasma glucagon concentrations (Table 2). However, a significant difference in the increase in plasma glucagon concentration between trials was observed within the control group only. The absence of such an effect in the type 2 diabetic group may partly be explained by the prevailing hyperglycemia, because high glucose concentrations have been shown to inhibit amino acid-induced glucagon release through a direct insulin-independent action on the α-cells (31). Though no significant change in the 2-h glucose response was observed in either the diabetic or control group, this does not necessarily imply that plasma glucose disposal was not improved, as explained above. Clearly, further research is needed to establish the magnitude to which free amino acid/protein supplementation can improve glucose disposal and reduce plasma glucose concentrations over time. Subsequently, it should also be investigated whether such dietary supplements can increase the efficiency of traditional glucose-lowering medication and/or provide an effective and practical means to reduce, delay, or even prevent the diabetic patients’ dependency on conventional blood glucose-lowering medication and/or exogenous insulin administration.

Because of the proposed physiological role of GH and IGF-I in insulin sensitivity (32), we also investigated direct or indirect effects of oral free amino acid/protein supplementation on plasma GH, cortisol, IGF-I, and IGF-BP3 concentrations in both the type 2 diabetic patients and normoglycemic control subjects (Table 2). However, no significant differences in plasma GH, cortisol, IGF-I, or IGF-BP3 were observed between trials or groups.

In conclusion, the insulin secretory reponse following carbohydrate ingestion is substantially impaired in long-term type 2 diabetic patients. However, the insulin response can be substantially increased by the coingestion of a free amino acid/protein mixture. Because the applied amino acid/protein mixture, containing free leucine, phenylalanine, and an ample source of essential amino acids, has a high insulinotropic potential in type 2 diabetic patients, future research is warranted to investigate the use of such amino acid-containing mixtures as tools to increase postprandial glucose disposal and improve glucose homeostasis.