Rapid quantification of extracellular neurotransmitters in mouse brain by PESI/MS/MS and longitudinal data analysis using the R and Stan-based Bayesian state-space model

Abstract

We developed a methodology for rapid quantification of extracellular neurotransmitters in mouse brain by PESI/MS/MS and longitudinal data analysis using the R and Stan-based Bayesian state-space model. We performed a rapid analysis for quantifying extracellular l-glutamic acid (L-Glu) and gamma-aminobutyric acid (GABA) in the mouse striatum by combined use of probe electrospray ionization/tandem mass spectrometry (PESI/MS/MS) and in vivo brain microdialysis. We optimized the PESI/MS/MS parameters with the authentic L-Glu, GABA, L-Glu-13C5,15N1, and GABA-D6 standards. We constructed calibration curves of L-Glu and GABA with the stable isotope internal standard correction method (L-Glu-13C5,15N1, and GABA-D6), demonstrating sufficient linearity (R > 0.999). Additionally, the quantitative method for L-Glu and GABA was validated with low-, middle-, and high-quality control samples. The intra- and inter-day accuracy and precision were 0.4%–7.5% and 1.7%–5.4% for L-Glu, respectively, and 0.1%–4.8% and 2.1%–5.7% for GABA, respectively, demonstrating high reproducibility of the method. To evaluate the feasibility of this method, microdialyses were performed on free-moving mice that were stimulated by high-K+-induced depolarization under different sampling conditions: 1) every 5 min for 150 min (n = 2) and 2) every 1 min for 30 min (n = 3). We applied the R and Stan-based Bayesian state-space model to each mouse's time-series data considering autocorrelation, and the model successfully detected abnormal changes in the L-Glu and GABA levels in each mouse. Thus, the L-Glu and GABA levels in all microdialysates approximately increased up to two- and seven-fold levels through high-K+-induced depolarization. Additionally, a 1-min temporal resolution was achieved using this method, thereby successfully monitoring microenvironmental changes in the extracellular L-Glu and GABA of the mouse striatum. In conclusion, this methodology using PESI/MS/MS and Bayesian state-space model allowed easy monitoring of neurotransmitters at high temporal resolutions and appropriate data interpretation considering autocorrelation of time-series data, which will reveal hidden pathological mechanisms of brain diseases, such as Parkinson's disease and Huntington's disease in the future.