Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2013 May 1;33(18):7961-74.
doi: 10.1523/JNEUROSCI.1186-12.2013.

Differential GABAB-receptor-mediated Effects in Perisomatic- And Dendrite-Targeting Parvalbumin Interneurons

Affiliations
Free PMC article

Differential GABAB-receptor-mediated Effects in Perisomatic- And Dendrite-Targeting Parvalbumin Interneurons

Sam A Booker et al. J Neurosci. .
Free PMC article

Abstract

Inhibitory parvalbumin-containing interneurons (PVIs) control neuronal discharge and support the generation of theta- and gamma-frequency oscillations in cortical networks. Fast GABAergic input onto PVIs is crucial for their synchronization and oscillatory entrainment, but the role of metabotropic GABA(B) receptors (GABA(B)Rs) in mediating slow presynaptic and postsynaptic inhibition remains unknown. In this study, we have combined high-resolution immunoelectron microscopy, whole-cell patch-clamp recording, and computational modeling to investigate the subcellular distribution and effects of GABA(B)Rs and their postsynaptic effector Kir3 channels in rat hippocampal PVIs. Pre-embedding immunogold labeling revealed that the receptors and channels localize at high levels to the extrasynaptic membrane of parvalbumin-immunoreactive dendrites. Immunoreactivity for GABA(B)Rs was also present at lower levels on PVI axon terminals. Whole-cell recordings further showed that synaptically released GABA in response to extracellular stimulation evokes large GABA(B)R-mediated slow IPSCs in perisomatic-targeting (PT) PVIs, but only small or no currents in dendrite-targeting (DT) PVIs. In contrast, paired recordings demonstrated that GABA(B)R activation results in presynaptic inhibition at the output synapses of both PT and DT PVIs, but more strongly in the latter. Finally, computational analysis indicated that GABA(B) IPSCs can phasically modulate the discharge of PT interneurons at theta frequencies. In summary, our results show that GABA(B)Rs differentially mediate slow presynaptic and postsynaptic inhibition in PVIs and can contribute to the dynamic modulation of their activity during oscillations. Furthermore, these data provide evidence for a compartment-specific molecular divergence of hippocampal PVI subtypes, suggesting that activation of GABA(B)Rs may shift the balance between perisomatic and dendritic inhibition.

Figures

Figure 1.
Figure 1.
Immunoreactivity for GABAB1 and Kir3 subunits in hippocampal PVIs. A, Confocal images of the cell body layer of the hippocampal CA1 area showing immunofluorescent labeling for PV (left, green) and GABAB1 (middle, red) and their colocalization in the superimposed image (right). Note the strong immunolabeling for the GABAB1 subunit in the soma of a PV-negative interneuron (arrow) and the lower but discernible labeling in PV-positive somata (arrowheads). B, Confocal images illustrating the colocalization of PV (left, green) with Kir3.1 (top row, middle, red), Kir3.2 (middle row, middle), and Kir3.3 subunits (bottom row, middle). Immunoreactivity for all the three Kir3 subunits is present in PV-positive interneurons in and near the CA1 str. pyramidale (arrowheads).
Figure 2.
Figure 2.
GABABR subunits are present on the plasma membrane of PV-immunopositive interneurons. A, B, Electron micrographs showing immunoreactivity for GABAB1 (A, immunoparticles) and GABAB2 (B, immunoparticles) in PV-positive dendrites (Den, peroxidase reaction end-product) contacted by several presynaptic boutons (b) in the CA1 str. radiatum. The majority of the particles for the receptor subunits were found on the extrasynaptic plasma membrane of the dendrites (arrows). C, Summary bar graphs of the surface density of immunogold particles in immunoperoxidase-negative spiny dendrites of putative PCs and PV-positive dendrites, somata, and boutons. D, Consecutive serial electron micrographs of a PVI bouton in the CA1 str. pyramidale showing immunoreactivity for GABAB1 (arrows). Note that the bouton makes a symmetrical putative GABAergic synapse (double arrows) with a pyramidal cell soma (S).
Figure 3.
Figure 3.
Kir3 subunits are present on the plasma membrane of PV-immunoreactive interneurons. A–C, Electron micrographs of consecutive sections illustrating the localization of immunogold particles (arrows) for the Kir3.1 (A), Kir3.2 (B), and Kir3.3 (C) subunits on dendritic shafts of PV-expressing neurons (Den, peroxidase reaction end-product), which were contacted by many presynaptic boutons (b). D–F, Bar graphs of the mean surface density of immunogold particles for the three Kir3 subunits on CA1 PC and PV-positive dendrites.
Figure 4.
Figure 4.
Slow GABABR-mediated synaptic responses in PVIs of the hippocampal CA1 area. A, Slow IPSCs in a PC evoked by a single stimulus (a) or trains of three (b) and five (c) stimuli elicited via an extracellular electrode placed at the border of str. radiatum and str. lacunosum-moleculare. IPSCs were blocked by application of the selective GABABR antagonist CGP (5 μm, d). Inset illustrates the recording arrangement for the experiments. B, The same stimulation protocol elicited slow IPSCs in PVIs. Inset shows confocal images of the PV immunolabeling (left, green pseudocolor) in the biocytin-filled cell body (right, blue pseudocolor). Scale bar, 20 μm.
Figure 5.
Figure 5.
Single interneurons can evoke GABABR-mediated postsynaptic responses in PCs and PVIs. A, B, In simultaneous recordings (schematic in A), trains of action potentials in various types of interneurons (upper traces) produced small, but discernible GABABR-mediated slow IPSCs (lower black traces) in both CA1 PCs (A) and PVIs (B). Families of traces on the left illustrate the discharge pattern and voltage responses of the recorded neurons to depolarizing and hyperpolarizing current pulses (−250–250 pA, 500 ms). Inset in B shows confocal images of the PV-immunoreactivity (arrow, left, green pseudocolor) in a biocytin-filled dendrite of the cell (arrow, right, blue pseudocolor; scale bar, 5 μm). In all pairs, this IPSC was blocked by application of CGP (5 μm, lower gray traces). C, Summary plot of GABABR-mediated IPSC amplitudes in CA1 PCs (four cells) and PVIs (five cells). Identity of the presynaptic partner is indicated by abbreviation near the points. NGFC indicates neurogliaform cell.
Figure 6.
Figure 6.
Differential expression of GABABR-mediated synaptic and whole-cell currents in PT and DT PVIs. A, Left: Reconstruction of a PT PVI, a putative basket cell, with dense axonal arbor (red) in the str. pyramidale (str. pyr.). Right: Reconstruction of a DT PVI, a putative bistratified cell, with axonal arbor (red) in the str. radiatum (str. rad.) and oriens (str. ori.). Insets show the immunoreactivity for PV (top, in green) in the visualized biocytin-filled cell bodies (bottom, blue pseudocolor; scale bar, 20 μm); traces show the responses of the two interneurons to a family of hyperpolarizing and depolarizing current pulses (−250 to 250 pA, 50 pA steps, 500 ms duration); note the FS discharge pattern of both cells. B, Pharmacologically isolated slow IPSCs evoked in the PT cell (left, average of 10 traces) and the DT interneuron (right) by trains of five extracellular stimuli delivered to the border of str. radiatum and str. lacunosum-moleculare (str. l-m). IPSCs had large amplitude in PT cells but were absent or had very small amplitudes in DT neurons. C, Time course plots of whole-cell current (IWC, top) and the IPSC amplitude (bottom) before and during sequential bath application of baclofen (10 μm) and CGP (5 μm) from a subset of PT cells (A, four cells) and DT cells (B, six cells) in which both drugs were applied. D, Summary bar graph of IPSC amplitudes in PCs and PT and DT PVIs elicited by single stimuli and trains of three and five stimuli. E, Summary bar graph of the baclofen-induced IWC measured in PCs and PT and DT interneurons. F, Representative plot of the voltage dependence of the baclofen-induced IWC in a PT interneuron. Voltage dependence was calculated by subtracting current responses to ramp commands after CGP and during application of baclofen. Inset, summary of reversal potentials (ER) of the baclofen-induced currents obtained from the calculated voltage dependence in PCs and PT neurons.
Figure 7.
Figure 7.
Presynaptic GABABRs inhibit synaptic transmission at PVI interneuron to PC synapses. A, Left: Reconstruction of a PT PVI, a putative basket cell, with dense axonal arbor (in red) in str. pyramidale (str. pyr.) and a postsynaptic PC (in blue). Right: Reconstruction of a DT PVI with axonal arbor (red) mainly found in the str. radiatum (str. rad.) and oriens (str. ori.) and a postsynaptic PC (in blue). Insets illustrate the immunoreactivity for PV (left, green pseudocolor) in the biocytin-filled cell bodies (right, blue; scale bar, 20 μm); traces show the responses of the presynaptic PVIs (top) and the postsynaptic PCs (bottom) to a family of hyperpolarizing and depolarizing current pulses (−250 to 250 pA, 50 pA steps, 500 ms duration). B, Action potentials (red, top traces) in the presynaptic PT (left panel) and DT (right panel) PVIs were followed by fast IPSCs (blue, bottom traces, averages of 30 responses) in the synaptically coupled PCs. Note that IPSC amplitudes in both pairs were reduced during baclofen application but recovered in CGP. C, D, Amplitudes of unitary IPSCs plotted as a function of time for the two PVI-PC pairs illustrated in B. The time of bath application of baclofen and CGP is indicated by the boxes above. E, Summary bar charts of the normalized ISPC amplitudes during baclofen and CGP application from PT-PC (light gray bars) and DT-PC (dark gray bars) pairs (number of pairs shown underneath bars); mean data are overlain by individual pair data (open circles). ns, p > 0.05; *p < 0.05; **p < 0.01; ***p < 0.001.
Figure 8.
Figure 8.
Slow GABABR-mediated conductances modulate the discharge at theta frequencies in a FS neuron network model. A, Raster plots of the activity in a mixed network model of 200 FS cells (bottom) and 20 TCs (top). Discharge frequency of TCs was 4 Hz. The compound GABAB conductance (GGABA-B) elicited in a single FS by TCs is illustrated between the raster plots. Peak amplitude of the unitary GABAB conductance was set to 25 μS cm−2. B, FS network frequency (top) and coherence (bottom) plotted against the unitary conductance amplitude (gGABA-B). Upper boundary of the shaded area in the frequency plot indicates the instantaneous frequencies measured at the trough of the compound conductance; lower values correspond to frequencies at the peak of the conductance. C, Raster plots of the activity in the mixed network with the TCs discharging at 10 Hz. GGABA-B was adjusted to match the steady-state peak amplitude of the compound conductance to that at 4 Hz TC discharge (A). Note the switch from gamma (66 Hz) to theta frequency activity (10 Hz) as the GABAB conductance summates progressively. D, Steady-state peak amplitude of the conductance normalized to the peak conductance elicited by first action potentials in TCs (top) and the proportion of the phasic component (bottom) during steady-state summation plotted against the discharge frequency of TCs. E, Phase of FS discharge relative to the TC cycle plotted against the discharge frequency of TCs.

Similar articles

See all similar articles

Cited by 19 articles

See all "Cited by" articles

Publication types

MeSH terms

LinkOut - more resources

Feedback