PubMed1993
The autoreceptor-mediated control of GABA release was simulated on a personal computer using commercially available software (STELLA/ITHINK). The experimental data to be matched were taken from previous publications. A basic model was able to fairly accurately reproduce frequency dependencies of GABA release in the presence and absence of uptake inhibition as well as concentration-response curves for changes in release produced by the agonist, (-)-baclofen, or by relatively low concentrations of the antagonists, phaclofen and CGP 35348. Obvious mismatch was observed at high concentrations of a potent antagonist, at a stimulation frequency of 2 Hz. Whereas the experimental data indicate a 3-fold increase in release as compared to controls, simulation predicts a 7-fold increase. By adaptation of the model, simulation data were obtained indicating that this mismatch was not due to (a) the autoreceptor occurring as two subtypes with different affinities for antagonists, (b) the occurrence of an agonist and antagonist state of the autoreceptor, with the latter prevailing at low synaptic concentrations of endogenous GABA, and (c) overruling of uptake inhibition by markedly elevated synaptic GABA concentrations. On the other hand, a simple restriction of the amount of transmitter able to be released per time unit produced much better matching data. A refined model assuming a restricted replacement capacity for exocytotically emptied synaptic vesicles at their docking sites gave similar results. As a consequence, we shall attempt to address this possibility experimentally. Simulation can never prove a case in the positive sense. It can, however, help to exclude ill-matching solutions of a problem and to prioritize among possible ones, which then must be experimentally addressed. We found simulation with this user-friendly software extraordinarily useful, also and not least because it necessitates and stimulates very intense dealing with a subject.