Cell type-specific separation of subicular principal neurons during network activities

Abstract

The hippocampal output structure, the subiculum, expresses two major memory relevant network rhythms, sharp wave ripple and gamma frequency oscillations. To this date, it remains unclear how the two distinct types of subicular principal cells, intrinsically bursting and regular spiking neurons, participate in these two network rhythms. Using concomitant local field potential and intracellular recordings in an in vitro mouse model that allows the investigation of both network rhythms, we found a cell type-specific segregation of principal neurons into participating intrinsically bursting and non-participating regular spiking cells. However, if regular spiking cells were kept at a more depolarized level, they did participate in a specific manner, suggesting a potential bimodal working model dependent on the level of excitation. Furthermore, intrinsically bursting and regular spiking cells exhibited divergent intrinsic membrane and synaptic properties in the active network. Thus, our results suggest a cell-type-specific segregation of principal cells into two separate groups during network activities, supporting the idea of two parallel streams of information processing within the subiculum.

Fig 1. Sharp wave and gamma network oscillations within the subiculum.

(A) Spectrogram (top) with color-coded power spectral density (PSD) exemplifies the transition from spontaneously occurring sharp wave-ripples (SWR) to gamma frequency oscillations within the subiculum. The corresponding LFP recordings are displayed below. The application of kainic acid (KA, onset is marked by black line) abolishes the SWR rhythm and induces, after a brief transitory state, a stable oscillatory gamma rhythm. The recording interruptions of the top spectrograms and the underlying LFP traces are 12 s (middle) and 25 min (right). Red lines mark three examples that are illustrated below with higher temporal resolution (SWR, transition, gamma). (A, bottom, left) The SWR (filtered 2–300 Hz), the corresponding SPW (2–50 Hz) and the ripple components (100–300 Hz) supplemented by the color-coded power spectral density wavelet transform. (A, bottom, right) The boxplot depicts the distribution of the wavelet peak power spectral frequencies of 100 analyzed consecutive ripple events of the upper example trace. (B) Sharp waves of both polarities are exemplified on the left with each SWR trace (2–300 Hz, top), the ripple trace (100–300 Hz, middle) and the corresponding wavelet transform as color-coded power spectral density plot (bottom). The boxplot (right) illustrates the distribution of the mean SWP rates of all slices investigated (n = 42).

Fig 2. Gamma frequency network oscillations within the subiculum.

(A) Gamma frequency oscillations were found in all of the examined subicular regions (middle, ventral, and dorsal). Sketches illustrate horizontal (middle, ventral and dorsal) slice preparation. The position of the scissors indicate the cuts made around the perimeters of the subicular region with the resulting subicular minislices marked by an asterisk. (A, right next to the sketch) Two example LFP recordings obtained from intact (grey, I, top trace) and isolated (black, II, bottom trace) middle (top), ventral (middle) and dorsal (bottom) slices are displayed together with the corresponding power spectra (A, middle column, color code according to the example traces). (A, right) The population data of the oscillatory frequency (top histogram) and spectral power (bottom histogram) exhibits no significant difference of the intact compared to the isolated subicular slices in network oscillatory gamma frequency as well as in spectral power (values and numbers in Table 1) except for the ventral subicular slices (p = 0.034, significance level indicated by asterisk). (B) Gamma frequency oscillations recorded from the medial (grey, I) and lateral (black, II) subiculum within the sagittal slice preparations. Same type of illustration as in (A), the subicular region is marked by an asterisk. The population histogram for frequency and power (values and numbers in Table 2) did not reveal a significant difference.

Fig 3. Behavior of subicular IB cells during spontaneous subicular SPWs.

(A) Current-voltage relationship of a IB cell (left) with current injection steps of −300 pA, −100 pA, and +140 pA, respectively, displayed with the attached microphotograph of a biocytin-stained IB cell (right). These cells exhibit the typical pyramidal shaped cell body, prominent apical dendrites that travel through the molecular layer reaching the hippocampal fissure (hf), and basal dendrites that spread within the pyramidal cell layer. The axon leaves the subiculum (Sub) via the alveus. IB cells respond to a hyperpolarizing current injection with a sag in membrane potential whereas a positive current pulse leads to burst firing. (B) Example of simultaneous extracellular LFP (top trace) and intracellular (bottom trace) recordings at RMP is shown on the left. The intracellular recording reveals phase-locked synaptic responses as well as a full-blown AP (truncated for clarity) with respect to the LFP SPWs. The spike time histogram (n = 16 IB cells) on the right illustrates a clear peak of AP generation in close vicinity to the SPW peak. The vertical line marks the maximum mean SPW deflection as time point 0. (C) EPSPs and IPSPs are displayed in correlation to the LFP (left). The EPSPs and IPSPs were recorded at −80 mV and at 0 mV, respectively. (right) Accumulated mean EPSP/IPSP with respect to the maximum SPW peak deflection.

Fig 4. Behavior of subicular RS cells during spontaneous subicular SPW activity.

(A) Neuronal discharge pattern and microphotograph of an RS cell. RS cells do not fire bursts and show little or no sag potential. They respond to a depolarizing current injection with a train of single APs. RS cells show a typical pyramidal cell morphology. (A, right) The accommodation behavior (n = 19 IB cells; n = 16 RS cells) reveals a significant difference between subicular IB and RS cells (level of significance indicated by the asterisks, p < 0.0001). (B) Example of a RS cell at RMP (top left) and under the condition of depolarizing current injection (bottom left) during spontaneous SPW is given with the same type of illustration as in Fig 3. Intracellular example recording of a silent RS cell at RMP (top left) depicting a SPW associated postsynaptic depolarization without AP generation. The population data (n = 7, right next) do not contain any APs, but aggregate a phase-locked mixed postsynaptic current (PSP) instead. When depolarized by current injection previously silent RS cells display a tonic AP firing mode (bottom left). In stark contrast to the IB cells, data of spontaneous active and depolarized RS cells reveal a prominent SPW peak correlated pause of AP generation (n = 11, bottom, middle left). EPSPs and IPSPs examples are displayed in correlation to the LFP (middle right). The EPSPs and IPSPs were recorded at −80 mV and at 0 mV, respectively. (right) Accumulated mean EPSP/IPSP with respect to the maximum SPW peak deflection. (C) EPSP (left) and IPSP (right) amplitudes for both cell classes. IB cells receive significantly higher synaptic excitation than RS neurons (indicated by the asterisk, p = 0.033; IB: n = 12; RS: n = 13), while there was no significant difference in IPSP (p = 0.52; IB: n = 3; RS: n = 4).

Fig 5. Temporal correlation of subicular IB cell activity to subicular field gamma frequency oscillations.

(A) Example of simultaneous extracellular LFP (top trace) and intracellular (bottom trace) recordings demonstrates regular AP firing (truncated for clarity) during gamma network oscillations. The spike time histogram (n = 8) on the right reveals a bimodal phase-locked behavior with a prominent pause of AP generation around the peak of the gamma cycle. The vertical line marks the maximum mean gamma peak deflection as time point 0. (B) Single intracellular recording cutouts (40 ms each) of consecutive triggered LFP gamma cycles illustrating the typical IB cell AP pattern for oscillatory network gamma activity together with the mean LFP trace of these recordings on top. (C) Example traces of synaptic potentials (left). EPSP and IPSP were recorded at −80 mV and 0 mV, respectively. Maximal cumulative postsynaptic potential peak deflections occur before the maximum of LFP gamma cycle (right).

Fig 6. Temporal correlation of subicular RS cell activity to gamma frequency network oscillations.

(A) Examples of simultaneous LFP (top) and intracellular (bottom) recordings during gamma frequency oscillations. In stark contrast to IB neurons, without current injection RS cells usually do not generate APs during gamma frequency oscillations and the data (n = 7) solely reveals a phase-locked mixed postsynaptic current (PSP, middle left). A depolarizing current injection initiates AP generation (middle right). Depolarized and spontaneous active RS cells (n = 9) show a distribution of AP generation (right) similar to the one observed in IB cells. (B) Example traces of synaptic potentials on the left. EPSP and IPSP were recorded at −80 mV and 0 mV respectively. Maximal cumulative postsynaptic potential peak deflections occur before the maximum of LFP gamma cycle. (C) There was no significant difference between the two classes of subicular PCs concerning the EPSP (IB: n = 8, RS: n = 11; p = 0.86) and IPSP amplitude (IB: n = 5, RS: n = 5; p = 0.56).

Fig 7. Phase distribution of APs and synaptic inhibition during SPW and gamma frequency oscillations.

APs (black) and synaptic inhibition (red) of IB (left column) and RS (right column) cells projected on a standard phase polar diagram for SPW (upper row; 80 ms projected time window) and gamma frequency oscillations (lower row; 24 ms projected time window). The peak of network activity is located at 180°, the ascending slope is encoded with lower, the descending with higher values. APs are illustrated based on the cumulative AP time histograms (Figs 3B, 4B, 5A and 6A) with respect to the projected time window, synaptic inhibition is represented by the distribution of time points for maximal deflection in the cellular cumulative IPSPs (Figs 3C, 4B, 5C and 6B). All values are normalized to the respective maximum activity level. During SPW oscillations RS cells, spontaneously active or depolarized, exhibit an earlier AP activity (n = 11) as well as an earlier peak of synaptic inhibition (n = 4) with respect to the IB cell phase distribution (n = 16, n = 3, respectively). Hence, according to the different phase distributions the activity of both cell types is separated in the time domain during SPW network activity. In contrast, during gamma frequency oscillations both subicular cell classes reveal comparable activity pattern with two distinct AP peaks (IB: n = 8; RS: n = 9) and an intermediate prominent inhibition (IB: n = 5; RS: n = 5).