COMPUTER-SIMULATIONS OF MORPHOLOGICALLY RECONSTRUCTED CA3 HIPPOCAMPAL-NEURONS

1. We tested several hypotheses with respect to the mechanisms and pro cesses that control the firing characteristics and determine the spati al acid temporal dynamics of intracellular Ca2+ in CA3 hippocampal neu rons. In particular, we were interested to know 1) whether bursting an d nonbursting behavior of CA3 neurons could be accounted for in a morp hologically realistic model using a number of the known ionic conducta nces; 2) whether such a model is robust across different cell morpholo gies; 3) whether some particular nonuniform distribution of Ca2+ chann els is required for bursting; and 4) whether such a model can reproduc e the magnitude and spatial distribution of intracellular Ca2+ transie nts determined from fluorescence imaging studies and can predict reaso nable intracellular Ca2+ concentration ([Ca2+](i)) distribution for CA 3 neurons. 2. For this purpose we have developed a highly detailed mod el of the distribution and densities of membrane ion channels in hippo campal CA3 bursting and nonbursting pyramidal neurons. This model repr oduces both the experimentally observed firing modes and the dynamics of intracellular Ca2+. 3. The kinetics of the membrane ionic conductan ces are based on available experimental data. This model incorporates a single Na+ channel, three Ca2+ channels (Ca-N, Ca-L, and Ca-T), thre e Ca2+-independent K+ channels (K-DR, K-A, and K-M), two Ca2+-dependen t K+ channels (K-C and K-AHP), and intracellular Ca2+-related processe s such as bufffering, pumping, and radial diffusion. 4. To test the ro bustness of the model, we applied it to six different morphologically accurate reconstructions of CA3 hippocampal pyramidal neurons. In ever y neuron, Ca2+ channels, Ca2+-related processes, and Ca2+-dependent K channels were uniformly distributed over the entire cell. Ca2+-indepe ndent K+ channels were placed on the soma and the proximal apical dend rites. For each reconstructed cell we were able to reproduce bursting and nonbursting firing characteristics as well as Ca2+ transients o an d distributions for both somatic and synaptic stimulations. 5. Our sim ulation results suggest that CA3 pyramidal cell bursting behavior does not require any special distribution of Ca2+-dependent channels and m echanisms. Furthermore, a simple increase in the Ca2+-independent K+ c onductances is sufficient to change the firing mode of our CA3 neurons from bursting to nonbursting. 6. The model also displays [Ca2+](i) tr ansients and distributions that are consistent with fluorescent imagin g data. Peak [Ca2+](i) distribution for synaptic stimulation of the no nbursting model is broader when compared with somatic stimulation. Som atic stimulation of the bursting model shows a broader distribution in [Ca2+](i) when compared with the nonbursting model. Synaptic stimulat ion in both models produces a [Ca2+](i) distribution that has a peak a round the site of stimulation. 7. In conclusion, this model is able to reproduce realistic bursting, spike Frequency adaptation, and [Ca2+]( i) dynamics of hippocampal CA3 neurons using several reconstructed mor phologies. In almost all casts changes only in the Ca2+-independent K channel densities and distributions on and near the soma were necessa ry to reproduce the same electrophysiological behavior in different mo rphologies. Different modes of firing were not dependent on varying Ca 2+ and Ca2+-dependent K+ channel distribution, or on geometric constra ints of the cell, but on Ca2+-independent K+ channel densities and dis tributions on and near the soma. Morphological factors such as cell ge ometry and dendritic surface-to-volume ratios, however. did influence [Ca2+](i) transients and distributions.