Overview: To compare the impact of amino acid consumption in conjunction with exercise (Study 1) young women (n=7, 25±2 yrs. Height: 161±2 cm; Weight: 60±2 kg; BMI: 24±1) completed two identical resistance training sessions that included 8 sets of 15 repetitions of single-leg calf press exercise on a seated leg press machine at 70% of 15RM (Astill et al. 2017). The eight sets were preceded by two warm-up sets at ~40% of maximum effort. Each exercise session was separated by at least one week. Immediately following the exercise bout, subjects consumed an oral bolus of essential amino acids (EAA) or a placebo. Each exercise session was conducted after an overnight fast. At least one week before the exercise studies, the calf muscle strength of the dominant leg was assessed using a 15-repetition maximum (RM), i.e., the maximum weight that can be completed 15 times (Astill et al. 2017). All exercise sessions were supervised by a research team member. For Study 2, a single EAA bolus identical to Study 1 was given to young (n=7; 4 men, 3 women; 27±1 yr.; Height: 169±4 cm; Weight: 69±6 kg; BMI: 24±1) and older adults (n=6; 1 man, 5 women; 68±2 yrs.; Height: 172±5 cm; Weight: 71±4 kg; BMI: 24±2). Participants in Study 2 did not exercise. Both studies were approved by the Institutional Review Board of Purdue University, West Lafayette, IN (IRB#1704019133 and IRB#1904022075) and registered on ClinicalTrials.gov (NCT04067479 and NCT04064528).
Amino Acid Bolus: Subjects in both studies consumed an identical bolus of EAA containing 3.5 grams of leucine (Dickinson et al. 2014; Dickinson et al. 2017; Glynn et al. 2010), 3 g of proline, 2 g glycine, 1.1 g histidine, 1.0 g isoleucine, 1.55 g lysine, 0.30 g methionine, 1.55 g phenylalanine, 1.45 g threonine, and 1.2 g valine. Amino acids (Ajinomoto Health & Nutrition North America, Inc) were mixed in a noncaloric, non-caffeinated carbonated beverage (Crystal Light). The placebo beverage consisted of Crystal Light only. The leucine dose was chosen based on work demonstrating its effectiveness at stimulating skeletal muscle protein synthesis (Dickinson et al. 2014), and for comparison to previous studies (Dickinson et al. 2017). We decided to include greater glycine because of the recent preclinical evidence implying that glycine can improve tendon properties (Vieira et al. 2015).
Microdialysis Study 1: Immediately after the exercise session, an ethylene oxide sterilized microdialysis fiber was inserted in the peritendinous space anterior to the Achilles tendon after preparation of the skin with an antiseptic (povidone-iodine) and local anesthetic [lidocaine 1%; (Astill et al. 2017; Gump et al. 2013)]. Subjects consumed the EAA bolus 90-minutes after fiber insertion or two-hours post-exercise. The two hours allowed for one hour of fiber equilibration, a thirty-minute baseline collection, and 30 minutes to insert the fiber. Microdialysis samples were collected every 30-minutes for five hours after amino acid consumption. For each experimental day, subjects fasted for 12-hours before arrival at Purdue University. Two subjects also completed a pilot microdialysis trial with no exercise no earlier than one week from the exercise experiments. This pilot experiment was included to determine if fiber insertion resulted in substantial changes in peritendinous amino acid concentrations. The results of these pilot experiments were not included in statistical analysis.
Microdialysis Study 2: A microdialysis fiber was inserted in the peritendinous space of the Achilles tendon after a 12-hour fast. One hour after fiber insertion, subjects consumed a bolus of EAA identical to Study 1. Microdialysis samples were collected every 30-minutes for four hours after amino acid consumption.
Sample Analysis. Amino acid concentrations were determined with high-performance liquid chromatography (Agilent Technologies 1100 HPLC System, Santa Clara, CA). Microdialysis samples deproteinized with 10%TCA (1:1 dilution) and further diluted with 1:1 with 0.1N HCl. Diluted samples were immediately centrifugated for 10 minutes at 4°C (10,000 g). The supernatant was removed and transferred to an HPLC vial. Amino acids were eluted using gradient elution with mobile phase A (10 mM Na2HPO4, 10 mM Na2B4O7, pH 8.2, and 5 mM NaN2, pH 8.2) and mobile phase B (45:45:10 of HPLC-grade acetonitrile, methanol, and water (Long 2017). Separation of amino acids was achieved using an Eclipse Plus C18 4.6x100 mm, 3.5mm column (Agilent) with a Restek Ultra C18 Guard Column (Restek Corporation, Bellefonte, PA). Peaks were monitored at 230 nm excitation/450 nm emission (G1321A, Agilent). Individual amino acid concentrations were determined by comparison with a standard curve (AAS18, MilliporeSigma, St. Louis, MO). The concentration of pro-collagen 1a1 concentration in the peritendinous space was determined at select time points after amino acid consumption using a DuoSetÒ ELISA from R&D Systems (DY6220-05, Minneapolis, MN
Statistics. Study 1: The concentration pro-collagen Ia1 was analyzed using a linear mixed-effects regression model. The fixed effects in this model were birth control usage, the group indicator, BMI, and time. In addition, the mixed model for pro-collagen Iα1 had random effects for a subject and a spatial power covariance structure to account for the unequally spaced time points in which the measurements were collected. We performed residual diagnostics to evaluate the assumptions of normal error terms, constant variance for the random error terms, and independent errors for the linear mixed-effects regression models. The residual diagnostics for the analyses of the raw outcome variables indicated that these assumptions were violated. In contrast, the diagnostics for the analyses of the logarithmically transformed outcomes indicated that the assumptions were satisfied. As such, we analyzed the logarithmically transformed outcome variables in our mixed-effects model.
A small number of amino acid time points were randomly lost because participants required a restroom break during the microdialysis experiment. The individual amino acid concentrations were analyzed separately using linear mixed-effects regression models that accounted for the correlation in a subject’s repeated measurements. In these models, the fixed effects were birth control usage, the group indicator, BMI, time, and the initial value of the respective amino acids. An autoregressive covariance structure of order 1 was used. As before, residual diagnostics were performed to assess the validity of the model assumptions and based on these diagnostics. The logarithmically transformed amino acid concentrations were ultimately analyzed. Multiple comparison tests were performed based on the Dunnett-Hsu adjustment approach for the fixed effects for strong control of Type 1 errors. The Kenward-Rogers adjustment was used to perform the mixed-effects models' statistical tests. Interactions between group and time were evaluated using tests and model fit statistics such as AIC, BIC, and AICc.
Study 2: Individual amino acid concentrations were evaluated with a mixed-effects regression model with random intercepts for subjects and continuous-time first-order autoregressive correlation structure [AR(1)] (Simpson et al. 2010). A Kenward-Roger adjustment was used for hypothesis testing of fixed effects (Time and Age). For amino acids and pro-collagen, we noted that the residuals were not normally distributed, and that the constant variance assumption was not valid. Thus, the raw outcome data were log-transformed before the analysis. Multiple comparisons were performed using a Dunnett’s test. Individual time points were compared to the 30-minute sample collection for any significant time effects. All statistical analyses (Study 1 and 2) were completed in SAS (SAS Institute, Inc), and figures were generated using GraphPad Prism 9.0.1 (GraphPad Software, San Diego, CA).