Statistical analysis of synaptic transmission at the calyx of Held synapse

Abstract

In order to distinguish pre- and postsynaptic mechanisms contributing to short-term synaptic depression at the calyx of Held synapse in the rat auditory pathway, non-stationary EPSC fluctuation analysis was theoretically developed and experimentally applied.Classical stationary fluctuation analysis estimates synaptic parameters (quantal size, number of release sites) from the variance-mean relationship of EPSCs, recorded in several stationary epochs of different mean EPSC amplitude varied e.g. by [Ca2+] ext (Clements and Silver, Trends Neurosci. 23:105-13, 2000). Non-stationary EPSC fluctuation analysis is an extension of this, where by application of short trains of stimuli the synapse is repetitively driven through states of different EPSC amplitude due to short-term synaptic plasticity. In addition to the classical variance-mean relationship non-stationary EPSC fluctuation analysis considers the covariance in the amplitude of successive responses due to vesicle depletion. This has important advantages compared to the classical approach. It yields an estimate of the quantal size for each EPSC in a stimulus train, allowing to detect use-dependent changes in this parameter. Moreover it yields an estimate of the number of release sites, which is affected less by heterogeneity in the release probability than the classical variance-mean approach. Furthermore correlation in the size of successive responses due to postsynaptic effects (Quastel, Biophys J. 72:728-53, 1997) was integrated in the analytical approach. The newly developed analytical algorithms were tested successfully on data generated with Monte Carlo simulations of synaptic transmission.Postsynaptic whole cell recordings with presynaptic fiber stimulation were performed in brainstem slices from 8 to 10 d old rats. Trains of 5 stimuli (100 Hz) at 4 mM [Ca2+]ext caused synaptic depression down to about 20 %. Successive EPSC during the train displayed significant, negative correlation. Non-stationary EPSC fluctuation analysis revealed, that depression was accompanied by a reduction in the effective quantal size to about 30 %. The quantal size reduction was abolished by combined application of 100 microM cyclothiazide and 1 mM kynurenic acid, which suggest a postsynaptic effects (desensitization, saturation) as underlying mechanisms. The quantal size estimates from the new covariance approach agreed well with mEPSC data and the observed reduction in quantal size was consistent with results from experiments where non-stationary and stationary variance-mean analysis were combined. Estimates for the number of release sites were higher from the covariance approach (approx 1900) than from the classical variance-mean approach (approx 800), suggesting the presence of heterogeneity in the release probability. Estimates of the number of releasable vesicles from depleting stimulus trains (approx 1250) lay in between these numbers.