Location dependent differences between somatic and dendritic IPSPs

Inhibitory synapses are thought to be distributed throughout the dendrites and the soma of pyramidal cells in several cortical regions. Recent experimental evidence suggests that inhibitory synapses of different locations are associated with specific actions based on their position. It has been observed that IPSP responses to GABA have different time-courses corresponding to the different locations of receptors. This can be explained by either the electrotonic filtering by the dendritic tree or different subspecies of GABA(A) receptors giving rise to distinct responses. Pearce (1993) has found that these differences are explained by different ratios of two pharmacologically distinct species of GABA(A) receptors in the dendrites versus the soma. In paired recordings conducted by Miles and coworkers (1996), where response properties could be directly related to specific synapses, IPSPs could be fitted with single exponentials which raises the possibility that the differences are due to filtering effects. This is interesting because if different IPSPs are due to post-synaptic receptors with identical kinetics then it is curious how the different actions attributed to these synapses arise.

First, we tried to reproduce the somatic IPSP measured by Miles et al. 1996 by finding the optimal values for the parameters of our synaptic conductance kinetics. In this case we modeled applying inhibition to the soma and, as throughout our whole work, recording the changes of membrane potential in the somatic compartment. We made a detailed exploration of the parameter space and in the following simulations we used the parameters and parameter intervals found to be best for reproducing the measured averages and standard deviations resp. of the three characteristic properties (time to peak, amplitude and duration at half apmlitude) of somatic IPSPs.

We then started to change the location of inhibition, but not the parameters of the synaptic kinetics, by applying it to different dendritic compartments, each time further from the soma. We recorded the membrane potential at the soma and examined the change in parameters of the IPSP and the total error of the fit with respect to the dendritic IPSP measured by Miles et al. 1996. There was an optimal electrotonic distance needed for reproducing the average dendritic IPSP that translated into an anatomical distance of 250-500 micron. Deviation of dendritic IPSPs was also successfully reproduced by changing the parameters only in the intervals found previously.

Our simulations showed that it is possible to reproduce the differences seen between somatic and dendritic IPSP as measured at the soma by Miles and coworkers, thus raising the possibility that although there are a number of known GABA(A) receptors, interneuron to pyramidal cell synapses use the same one regardless of location. This is supported by that we simulated the electrotonic attenuation with morphological distances comparable to that seen in reconstructed cells. Future simulations can describe the functional effects of this electrontonic attenuation of IPSPs on the bursting behavior of pyramidal cells.

While trying to make variations in IPSPs by changing the kinetic parameters of the synapse we were faced with the (not contraintuitive) fact that at least two properties of PSPs (time to peak and duration at half amplitude) did not vary independently, indeed there was a pronounced linear relationship between them. This gives a hint that the properties most commonly used to characterize PSPs by electrophysiologists are at least redundant and probably not appropriately chosen. Further investigations could enlighten the possible consequences of this linearity and also the parameters it depends from.

This work was done by Máté Lengyel and Ádám Kepecs