Complementary roles of cholecystokinin- and parvalbumin-expressing GABAergic neurons in hippocampal network oscillations

Abstract

In the hippocampal CA1 area, a relatively homogenous population of pyramidal cells is accompanied by a diversity of GABAergic interneurons. Previously, we found that parvalbumin-expressing basket, axo-axonic, bistratified, and oriens-lacunosum moleculare cells, innervating different domains of pyramidal cells, have distinct firing patterns during network oscillations in vivo. A second family of interneurons, expressing cholecystokinin but not parvalbumin, is known to target the same domains of pyramidal cells as do the parvalbumin cells. To test the temporal activity of these independent and parallel GABAergic inputs, we recorded the precise spike timing of identified cholecystokinin interneurons during hippocampal network oscillations in anesthetized rats and determined their molecular expression profiles and synaptic targets. The cells were cannabinoid receptor type 1 immunopositive. Contrary to the stereotyped firing of parvalbumin interneurons, cholecystokinin-expressing basket and dendrite-innervating cells discharge, on average, with 1.7 +/- 2.0 Hz during high-frequency ripple oscillations in an episode-dependent manner. During theta oscillations, cholecystokinin-expressing interneurons fire with 8.8 +/- 3.3 Hz at a characteristic time on the ascending phase of theta waves (155 +/- 81 degrees), when place cells start firing in freely moving animals. The firing patterns of some interneurons recorded in drug-free behaving rats were similar to cholecystokinin cells in anesthetized animals. Our results demonstrate that cholecystokinin- and parvalbumin-expressing interneurons make different contributions to network oscillations and play distinct roles in different brain states. We suggest that the specific spike timing of cholecystokinin interneurons and their sensitivity to endocannabinoids might contribute to differentiate subgroups of pyramidal cells forming neuronal assemblies, whereas parvalbumin interneurons contribute to synchronizing the entire network.

Figure 1.

In vivo firing patterns and visualization of a CCK-expressing basket cell (T64c). A, Reconstruction of the soma and dendrites (orange) is shown complete; the axon (yellow) is shown only from three sections of 65 μm thickness for clarity. st. rad., Stratum radiatum; st. pyr., stratum pyramidale; st. or., stratum oriens. Scale bar, 100 μm. B, Light (DAB reaction) and immunofluorescence microscopic visualization of the labeled cell (asterisk) demonstrates expression of CCK but not PV. Note an adjacent PV-expressing cell (arrow). Scale bar, 20 μm. C, Serial sections of a filled axonal bouton (b) of the cell making a synaptic junction (black arrows) with a pyramidal soma (s). Note the presence of a large granulated vesicle (white arrow), likely to contain neuropeptides. Scale bar, 20 nm. D, Action potential of the recorded cell filtered between 0.8 and 5 kHz. Calibration: 1 ms, 0.5 mV. E-H, In vivo firing patterns of the labeled cell. During theta oscillations (E), the cell fired preferentially at the ascending phase of the theta waves (filtered between 3 and 6 Hz) recorded extracellularly in the pyramidal cell layer by a second electrode. Note the difference in interspike interval between first and second versus second and third spike within a theta cycle (E). During ripple episodes (filtered between 90 and 140 Hz), the cell was sometimes activated (F) and sometimes silenced (G). During the labeling, positive current was applied in 200 ms on/off mode (H), modulating the firing of the cell. Note that a ripple episode during the current-on phase silenced the cell. Asterisks mark true ripple episodes, whereas other fast oscillations are artifacts of current steps. lfp, 0.3-200 Hz. Calibration: lfp, 0.5 mV; theta, 0.2 mV and 0.2 s; ripples, 0.1 mV and 0.2 s; spikes, 0.5 mV.

Figure 2.

In vivo firing patterns and visualization of a CCK-expressing, apical dendrite-innervating cell (T46d). A, Reconstruction of the soma and dendrites (orange) is shown complete; the axon (yellow) is shown only from three sections of 65 μm thickness for clarity. st. rad., Stratum radiatum; st. pyr., stratum pyramidale; st. or., stratum oriens; st. l.m., stratum lacunosum moleculare. Scale bar, 100 μm. B, Superimposed immunofluorescence micrograph of the neurobiotin-labeled cell (blue) and CCK (red) immunoreactivity in the Golgi apparatus. Scale bar, 20 μm. C, In vivo firing patterns of the cell. Note that during the initial slow oscillation, the cell fired very rarely, but after a light foot pinch (arrow), the local field potential changed to theta oscillation and the cell strongly increased its firing rate. Calibration: lfp and spikes, 1 mV, 2 s.

Figure 3.

In vivo firing patterns and visualization of a CCK-expressing perforant-path-associated cell (T82e). A, In vivo firing patterns of the cell. Note that during the initial ripple episodes, the cell did not fire but became active during the subsequent theta oscillations, when it fired at the ascending phase of the theta waves recorded in the stratum pyramidale with a second electrode. Calibration: 0.2 s; lfp and spikes, 0.5 mV; ripples, 0.1 mV. B, Reconstruction of the soma and dendrites (orange) is shown complete; the axon (yellow) is shown from selected series of sections as indicated (bottom section number marks more caudal position). Scale bar (same for all 3 projections), 100 μm. C, Immunofluorescence micrograph of the neurobiotin-labeled cell (blue), CCK (green), and calbindin (red) immunoreactivity. Scale bar, 20 μm. st. rad., Stratum radiatum; st. pyr., stratum pyramidale; st. l.m., stratum lacunosum moleculare; DG, dentate gyrus; st. m., stratum moleculare; st. g., stratum granulosum; sec., section.

Figure 4.

In vivo firing patterns and visualization of a CCK-expressing perforant-path-associated cell (T123b). A, B, Reconstruction of the soma and dendrites (orange) is shown complete; the axon (yellow) is shown from selected series of sections as indicated (bottom section number marks more caudal position). Asterisks mark the continuation of the main axon in A and B. Scale bar, 100 μm. Laminar boundaries are adjusted to reflect the position of processes as far as possible, but because of the tangential cutting plane, boundaries strongly shifted from section to section. A schematic sagittal view representing the cell shows rostrocaudal dimensions. C, In vivo firing pattern of the cell. Note the absence of firing during ripple episodes. Calibration: lfp, 0.5 mV, 0.2 s; ripples, 0.1 mV; spikes, 0.5 mV. D, Immunofluorescence micrographs of the cell labeled by neurobiotin (green) showing CCK immunoreactivity (red) in the soma and immunoreactivity for the cannabinoid receptor CB1 (orange) in the axonal membrane. Immunopositive axons from unlabeled cells are also evident. Scale bars: top, 20 μm; bottom, 10 μm. st. rad., Stratum radiatum; st. pyr., stratum pyramidale; st. l.m., stratum lacunosum moleculare; st. m., stratum moleculare; st. g., stratum granulosum; st. or., stratum oriens; sec., section; no., number.

Figure 5.

Postsynaptic targets and cannabinoid receptor decoration of axonal boutons from CCK-expressing cells. A, Axonal boutons of an apical dendrite innervating cell (T159a) labeled by neurobiotin (green) are immunopositive for CCK, CB1 receptor, and VGLUT3 as shown by immunofluorescence (false color images). Scale bar, 10 μm. B, Electron micrograph showing the labeled bouton (b) of a CCK-expressing basket cell (T152b) making a type II synapse (arrow) onto the soma (s) of a pyramidal cell. The bouton makes another type II synapse (arrow) with an invaginating protrusion [dendritic protrusion (dp)] from a dendrite. C, The filled bouton (b) of a perforant-path-associated cell (T82e) makes a type II synapse (arrow) onto a small dendritic shaft (d) of a pyramidal cell. A consecutive section reveals that the bouton is penetrated by a protrusion from this dendrite. dp, Dendritic protrusion. D, A filled bouton of the apical dendrite innervating cell (T46d) makes a type II synapse (arrow) onto a dendrite of an interneuron (IN) also receiving a type I synapse (double arrow). Note the bouton wrapping around the dendrite. Scale bars:B-D, 0.2 μm. E, A labeled bouton of the apical dendrite innervating cell (T46d) makes a type II synapse (arrow) onto the shaft of an apical dendrite (ad) of a pyramidal cell. The inset shows the synaptic junction in higher magnification. Scale bar, 0.5 μm; inset, 0.2 μm.

Figure 6.

CCK-expressing interneurons exhibit distinct firing patterns during theta and ripple oscillations. A, During theta oscillations, CCK-expressing interneurons (color traces of individual cells) fire with highest probability at the ascending phase of the theta wave recorded extracellularly in stratum pyramidale. For each cell, 1352 ± 1121 (range, 42-3693) theta cycles were recorded. For clarity, two theta cycles are shown. Top, Schematic theta wave; 0, 360, and 720° mark the troughs. The averaged firing probability of the CCK-expressing interneurons and, for comparison, of eight PV-expressing basket cells [5 published by Klausberger et al. (2003)] is shown in gray columns. Note that CCK-expressing cells fire out of phase with PV-expressing basket cells. Because the somata of pyramidal cells receive input from PV- and CCK-expressing basket cells at a probable ratio of 2:1 (Nyiri et al., 2001), the average firing probability of the PV-expressing basket cells was multiplied by 0.66 and added to the average firing probability of the three identified CCK-expressing basket cells multiplied by 0.33. The weighted sum of the firing probabilities of CCK- and PV-expressing basket cells still provides a theta-modulated GABAergic input to the soma of pyramidal cells. B, During ripple episodes, the average firing probability of the CCK-expressing interneurons showed no change (gray columns) relative to the preripple and postripple time. In contrast, the PV-expressing basket cells strongly increased their firing. Individual CCK-expressing interneurons (color traces) exhibited variable firing patterns ranging from slightly excited at the beginning and end of the ripple episodes to weakly inhibited. For each cell, 55 ± 36 (range, 18-113) ripple episodes were recorded. The start, maximum amplitude, and end (bottom) of the normalized ripple episodes are marked as -1, 0, and 1, respectively. C, Comparison of ripple episodes with and without the firing of basket cell T64c. The peri-ripple periods preceding or after ripples with (bottom) or without (middle) basket-cell firing show similar firing rates, indicating that the instantaneous firing frequency of the cell did not influence whether it would fire during a certain ripple episode.

Figure 7.

A subset of interneurons in drug-free and behaving animals show comparable firing patterns to CCK-expressing interneurons in anesthetized rats. A, The firing probability of six cells (color traces) is shown during theta oscillation and ripple oscillations; the mean is shown as gray columns. Theta oscillations were recorded while the animal was exploring an environment; ripples were recorded in slow-wave sleep. Note that none of these cells, selected by theta firing phase alone, exhibited a strong increase or inhibition of firing probability during ripple episodes like PV-expressing cells. One cell (red, top) showed a weak increase in firing at the beginning and end of the ripple episode, another (light blue) increased its firing slightly after the ripples, and the other four cells showed no change of firing during and around the ripple events. B, C, Firing sequences of an interneuron (red trace in A) during theta (B) and ripple (C) oscillations recorded by a tetrode in the pyramidal layer. The cell fired at the ascending phase of the theta cycle and at the beginning and end of the ripple episode. Calibration: 0.1 s, 0.2 mV.

Figure 8.

Hypothetical contribution of CCK-expressing interneurons to sparse coding by CA1 pyramidal cells during theta oscillations. The majority of pyramidal cells (dark blue) are silent as the animal explores a given place, whereas place cell (light blue) firing is theta modulated (spikes, vertical lines over a schematic theta wave). As the rat enters the place field of a pyramidal cell, the neuron starts to fire before the peak of the theta wave. The cell fires at gradually earlier phases during consecutive theta cycles. It is assumed that all pyramidal cells receive GABAergic input from PV-expressing basket cells (green), axo-axonic cells (purple), and several types of CCK-expressing cell (red). The PV basket cells fire, on average, before the place cells fire. The CCK-expressing interneurons fire, on average, at the ascending theta phase, at the same time when place cells start to fire. Thus, the glutamatergic excitation of place cells (yellow) must overcome the peak of inhibition by CCK-expressing and axo-axonic cells for the cell to start firing as the animal enters the place field. Only CCK-expressing cells have CB1 receptors on their presynaptic boutons (black squares), and when activated, these suppress GABA release. Place cells often fire complex spikes early in the place field leading to strong calcium influx, which in cooperation with muscarinic acetylcholine receptor activation would lead to the formation of endocannabinoids and a reduction of GABA release from only those CCK terminals that innervate the place cell. The CCK-expressing interneurons will fire and continue to release GABA to the majority of pyramidal cells, which are silent, keeping a high threshold for activation. In contrast, the place cell continues to fire with high frequency as a result of decreased inhibition and increasing excitation. Thus, the specific spike timing of CCK-expressing interneurons during theta oscillations and the expression of cannabinoid receptors on their terminals contribute to increasing the difference in firing between activated and nonactivated pyramidal cells.