Branched chain amino acids (BCAA) are known to exert insulinogenic effect. Whether this effect is mediated by incretins (GLP-1, GIP) is not known. The aim of the study was to show incretin effect of BCAA, i.e. whether the oral administration of BCAA elicits higher insulin and incretin response when compared with IV route of the same dose of BCAA. Eighteen healthy, male subjects participated in three tests: IV application of BCAA solution (30.7±1.1 g of BCAA, IV BCAA) second was oral ingestion of BCAA capsules in the same dose (ORAL BCAA) and third experiment was oral placebo (PLACEBO). Glucose, insulin, GLP-1, GIP, valine, leucine and isoleucine levels were measured at time interval for up to 4 h.

Study protocols: Three tests were conducted on a different occasion (at least 7 days in between). First test (IV BCAA test) comprised of IV application of BCAA solution (Nutramin VLI 3%, Fresenius Kabi, KGaA, Germany; Leucine 43%, Isoleucine 24%, Valine 33%) in total dose of 0,4 g/kg in a 2 h infusion (mean solution volume 1023±34.6mL, mean total dose of BCAA 30.7±1.1g) to respect maximum recommended infusion rate. Second test (ORAL BCAA test) comprised of at once oral ingestion of BCAA capsules (BCAA capsules, Reflex Nutrition, UK, Brighton; Leucine 50%, Isoleucine 25%, Valine 25%) in a single dose of 0.4g/kg (mean total dose of BCAA 30.7±1.1 g) administered over 30s, washed down with 500mL of tap water. Third test (ORAL PLACEBO test) comprised of at once oral ingestion of placebo capsules (methylcellulose, University hospital institution pharmacy prepared) in a single dose of 0.4 g/kg administered over 30 s, washed down with 500 mL of tap water. The content of placebo capsules was weighted and packed identically with BCAA capsules. Nor participants nor the staff administering capsules and carrying out the protocol did not know about capsules content so that randomized double-blinded settings could have been maintained.

Blood analysis: Basal peripheral venous blood was drawn from each subject after 12 h of fasting (-15 min) and then at time 0, 15, 30, 45, 60, 90, 120, 150, 180, 210, 240 min throughout the intervention. Plasma was immediately separated and all samples were frozen at -80 °C until analysis were performed. For GLP-1 and GIP analysis BD P800 Blood Collection System was used with coat developed to preserve metabolic peptides in blood (cocktail of proteases, esterase and DPP-IV inhibitors, anticoagulant 3,6 mg di-potassium ethylenediaminetetraacetic acid (K2EDTA), Becton, Dickinson and Co., New Jersey, USA).

Parameters of glucose homeostasis were assessed: plasma glucose using hexokinase reaction (Konelab Glucose analyzer, Thermo Fisher Scientific, Oy., Finland) and serum insulin using solid phase competitive chemiluminescent enzyme immunoassay (Immulite 2000, Siemens A.G., Germany).

Commercially available ELISA kits were used for GLP-1 and GIP analysis: for GLP-1 Immuno-Biological Laboratories (Immuno-Biological Laboratories, Gunma, Japan) and for GIP Millipore (EMD Millipore Corporation, Bilerica, MA, USA).

Serum levels of BCAAs were determined by method of capillary electrophoresis (CE) with contactless conductivity detection, which has been already described in details [18]. Shortly: CE measurements were carried out using HP3DCE system (Agilent Technologies, Waldbronn, Germany) equipped with a built-in contactless conductivity detector. Separation took place in a fused-silica capillary (31.4 cm in total length, 14.7 cm to detector, 25 μm inside diameter, 363 μm outside diameter, Composite Metal Services, UK) at the controlled temperature of 25 °C. The inner surface of the capillary is covered using INST coating solution (Biotaq, U.S.A.) to prevent electro-osmotic flow before its first use [19]. The CE separation is performed in an optimized background electrolyte with composition 3.2 mol/l acetic acid in 20% v/v methanol, pH 2.0. The achieved separation time was 125 s at electric field intensity of 0.96 kV/cm and simultaneous application of a hydrodynamic pressure of 50 mbar. The separation efficiency in blood serum equaled 461,000 plates/m for valine and isoleucine, and 455,000 plates/m for leucine; the detection limits are equal to 0.4 µM for all three amino acids. The relative standard deviation values for repeatability of the migration time equaled 0.1% for measurements during a single day and 0.3% for measurements on different days; the relative standard deviation values for repeatability of the peak areas equaled 2.3 - 2.6% for measurements during a single day and 2.7 - 4.6% for measurements on different days. Blood samples were collected in test tubes containing ethylenediaminetetraacetic acid (EDTA). The obtained serum samples were stored in a freezer at -20 °C until the analysis. Prior to the analysis, the unfrozen serum samples were deproteinized by mixing 250 µl of serum with 750 µL of acetonitrile. Deproteinization was performed in an Eppendorf tube after shaking for 30s. Then the serum samples were centrifuged at an acceleration of 4 g for 45 s; 800 µL of the obtained supernatant were taken for CE analysis.

Statistical analysis: Data are presented in text, tables and figures as means ± standard error of the mean (SEM) and values of p<0.05 were considered statistically significant. Secretion responses for insulin, GIP, GLP-1, valine, leucine and isoleucine were calculated for each subject as incremental areas under the curve (iAUC). iAUCs calculation allow for different individual baseline values. iAUCs were calculated using the trapezoid model, from 0-240 min for glucose, insulin and BCAA and 0-120 for GLP-1 and GIP. All individual values below the baseline were excluded and each subject in respective study was their own reference. Secretion responses for glucose were calculated as decremental AUC (dAUC) as the response is predominantly negative. All individual values above the baseline were excluded in the case. Data for statistical comparison for iAUC and individual values were tested for normality and for normally distributed data, samples were compared using general linear model ANOVA with Bonferroni's multiple comparison to test the statistical significance of differences between groups. Repeated measures ANOVA test was used to assess time effect for every respective treatment. Mixed model ANOVA was used to assess treatment and time vs treatment effect. Differences between groups were identified using Bonferroni's multiple comparisons tests. GraphPad Prism, release 5.03 (GraphPad software, San Diego) was used to perform all statistical procedures.

Inclusion Criteria:

• healthy volunteers

Exclusion Criteria:

• age under 18 years, obesity, any chronic disease related to energy metabolism (particularly diabetes, thyreopathy, hypertension, dyslipidemia, atherosclerosis etc.), any chronic medication, smoking and regular alcohol consumption