An in vitro method for assessment of amino acid bidirectional transport and intracellular metabolic fluxes in mammary epithelial cells.

Understanding uptake of AA by mammary tissue as supply varies is critical for predicting milk component production. Our objective was to develop an in vitro method to quantify cellular uptake, efflux, and intracellular metabolism of individual AA that could be implemented for evaluating these factors when AA supply and profile are varied. Bovine primary mammary epithelial cells were grown to confluency and exposed to medium with an AA profile and concentration similar to lactating dairy cow plasma for 24 h. Cells were then preloaded in medium enriched with 15N-labeled AA for 24 h followed by removal of the 15N-labeled medium and incubation with medium enriched with 13C-labeled AA for 0, 15, 60, 300, 900, 1,800, and 3,600 s. Extracellular free AA and intracellular free and protein-bound AA were analyzed for concentrations and isotopic enrichment by gas chromatography-mass spectrometry. A dynamic, 12-pool model was constructed representing extracellular and intracellular free and protein-bound pools of an AA, and their respective 15N and 13C isotopes. Markov chain Monte Carlo simulation (n = 5,000) was conducted to evaluate prediction errors by deriving standard errors and posterior distributions for rate constants, fluxes, and pools. Cellular Ala influx and efflux were higher than Leu, reflecting Ala role in driving system L transport and the high capacity of sodium-dependent transport. The Ala and Leu turnover rates were 181 and 95, 580 and 857, and 74 and 157% per hour for extracellular, intracellular, and fast protein-bound pools, respectively. The intracellular and extracellular Ala to Leu ratios were quite different, meaning the blood AA profile is not the AA profile provided for protein translation. The high level of exchange and rapid turnover of pools provide a mechanism for matching the AA supplies to the precision necessary for translation. This also understates the importance of using experimental medium similar to what is observed in vivo given that some AA depend on other AA for influx (exchange driven). The average root mean squared prediction error across the isotope enrichments, pools, and concentrations was 9.7 and 14.1% for Ala and Leu, respectively, and collinearity among parameters was low, indicating adequate fit and identifiability. The described model provides insight on individual AA transport kinetics and a method for future evaluation of AA transport and intracellular metabolism when subjected to varying AA supplies.