Quantitative Prediction of Vasopressin Secretion Using a Computational Population Model of Rat Magnocellular Neurons

The goal of this study was to create a realistic and quantitative simulation of vasopressin (AVP) secretion under iso-osmotic and short-term challenged plasma osmolality. The relationship between AVP concentration ([AVP]) and plasma osmolality was computed using a sophisticated and integrated model that chronologically simulates (1) the overall firing rate of the hypothalamus' magnocellular neuronal (MCN) population, (2) the propagation of the spike activity down the axons, (3) the fatigue and facilitation mechanisms of AVP release at the axon terminals and (4) the [AVP] pharmacodynamics based on the trains of AVP release. This global simulation predicted that the differential MCN sensitivity to dynorphin would be the most critical mechanism underlying the individual variability of MCN firing behaviors (silence, irregular, phasic and continuous firing patterns). However, at the level of the MCN population, the simulation predicted that the dynorphin factor must be combined with the distribution of the resting membrane potentials among the MCNs to obtain a realistic overall firing rate in response to a change in osmolality. Moreover, taking advantage of the integrated model, the simulation predicted that the selective removal of the frequency-dependent facilitation of AVP secretion has a major impact on the overall [AVP]-to-osmolality relationship (mean absolute change of 2.59 pg/ml); the action potential propagation failure, while critical, has a smaller quantitative impact on the overall [AVP] (0.58 pg/ml). The present integrated model (from a single MCN to a quantitative plasma [AVP]) improves our knowledge of the mechanisms underlying overall MCN firing and AVP excitation-secretion coupling.