Kinetics and Connectivity Properties of Parvalbumin- and Somatostatin-Positive Inhibition in Layer 2/3 Medial Entorhinal Cortex

Abstract

Parvalbumin (Pvalb + )-and somatostatin (Sst + )-positive cells are the two largest subgroups of inhibitory interneurons. Studies in visual cortex indicate that synaptic connections between Pvalb + cells are common while connections between Sst + interneurons have not been observed. The inhibitory connectivity and kinetics of these two interneuron subpopulations, however, have not been characterized in medial entorhinal cortex (mEC). Using fluorescence-guided paired recordings in mouse brain slices from interneurons and excitatory cells in layer 2/3 mEC, we found that, unlike neocortical measures, Sst + cells inhibit each other, albeit with a lower probability than Pvalb + cells (18% versus 36% for unidirectional connections). Gap junction connections were also more frequent between Pvalb + cells than between Sst + cells. Pvalb + cells inhibited each other with larger conductances, smaller decay time constants and shorter delays. Similarly, synaptic connections between Pvalb + and excitatory cells were more likely and expressed faster decay times and shorter delays than those between Sst + and excitatory cells. Inhibitory cells exhibited smaller synaptic decay time constants between interneurons than on their excitatory targets. Inhibition between interneurons also depressed faster, and to a greater extent. Finally, inhibition onto layer 2 pyramidal and stellate cells originating from Pvalb + interneurons were very similar, with no significant differences in connection likelihood, inhibitory amplitude, and decay time. A model of short-term depression fitted to the data indicates that recovery time constants for refilling the available pool are in the range of 50-150 ms and that the fraction of the available pool released on each spike is in the range 0.2-0.5. SignificanceTwo large and distinct classes of interneurons in medial entorhinal cortex (mEC) include parvalbumin (Pvalb + )-and somatostatin (Sst + )-positive cells. Previous work has demonstrated unique functions with regards to spatial tuning and network oscillations for these two interneuron populations. Potential differences in kinetics of inhibition and likelihood of connection from these two interneuron groups, however, have not been quantified. Here, using fluorescence to guide intracellular recordings, we quantified the synaptic connections from both types of interneurons. We indicate that Sst + and Pvalb + express different synaptic kineticsare target-cell specific. In contrast to neocortical measures, we find substantial connections between Sst + interneurons.

Interneuronal network model of theta-nested fast oscillations predicts differential effects of heterogeneity, gap junctions and short term depression for hyperpolarizing versus shunting inhibition

“…This phase preference is consistent with that observed in PV+ basket cells in CA1 [36]. Based on STD parameters in PV+ mEC interneurons [32], this preference will be expressed even at the lower range of fast gamma in mEC (~65 Hz). Our results strongly suggest that any ING in mEC is mediated by hyperpolarizing rather than shunting inhibition, and that artificially manipulating the synaptic reversal potential to make it more depolarizing should decrease optogenetically-evoked ING that persists after blocking excitatory synapses.…”

“…For this reason this number is not well-constrained in our model. In the mEC, the probability of both electrical and chemical synapses drops off dramatically at distances between somata that are 125-150 μm apart [32]. In order to apply the connection probabilities, the average number of PV+ cells that are within that distance of a typical soma of a PV+ cell must be estimated.…”

“…The connection probabilities and conductances for both electrical and chemical synapses were taken from [32]. Chemical synapses were modeled by an inhibitory postsynaptic conductance with a biexponential waveform…”

“…. We used a 27% probability that any pair of neurons was connected by gap junctions [32]. Peak conductances were obtained from data originally collected for [32].…”

“…Our study is specific to layer 2/3 medial entorhinal cortex because both intrinsic and synaptic parameters of the network were constrained by data from this region [32]. The only other study that has previously attempted to capture the true extent of heterogeneity in the intrinsic properties of the PV+ interneuronal network, in the external segment of the globus pallidus [58], did not address network synchronization.…”