Excitation of Diverse Classes of Cholecystokinin Interneurons in the Basal Amygdala Facilitates Fear Extinction

Photostimulation of GABAergic cells in the BA. A, Schematic drawing showing the intersectional viral strategy used to target CCK⁺ INs in CCK-Cre;Dlx5/6-FLP double-transgenic mice. B, Example of INTRSECT AAVdj-hSyn-Con/Fon-hChR2(H134R)-EYFP-WPRE (INTRSECT-ChR2) expression in the BA (CeA = central amygdala). C, Schematic drawing represents a horizontal slice with viral expression shown in green (Hip = hippocampus). Whole-cell patch clamp recordings were performed in non-green cells, likely in PNs in slices prepared from double-transgenic mice injected with AAV containing INTRSECT-ChR2. D, E, Averaged traces of five consecutive PSCs obtained in three different neurons evoked by light illumination (blue arrow). High variability both in peak amplitude (D) and decaying phase (D, E) is typical for events evoked in different neurons. The traces in E are peak scaled. Dashed lines show where the peak amplitude for fast and slow components of evoked currents was measured. F, Traces from an experiment measuring the antagonist-sensitivity of light-evoked responses. Gabazine (5 µM) wash-in eliminated the fast GABA_A-mediated component, while CGP 5699A (1 µM) blocked the remaining slow GABA_B-mediated component. Importantly, no inward, i.e., EPSC, could be observed in the presence of the GABA receptor antagonists, indicating that the applied intersectional strategy allowed us to excite selective GABAergic cells. G, Peak amplitude of the fast components in evoked responses measured in the same neurons was significantly larger than the peak amplitude of the slow components. H, The fast components were blocked by bath application of gabazine (*paired t test). I, The slow components were eliminated by CGP 5699A wash-in. J, The area, i.e., the charge of the fast and slow components evoked in the same neurons, was not different. GABA_A receptor-mediated fast responses were isolated by subtracting the responses recorded in the presence of gabazine from the control traces and their area was measured. Example traces (subtracted) are shown in F. The area of the GABA_B receptor-mediated slow components were determined on the traces recorded in the presence of gabazine. K, Averaged traces taken from an example experiment indicate that the light-evoked PSCs are smaller on wash-in of a CB1R agonist, CP 55,940 (2 µM). L, In all experiments tested, bath application of CP 55,940 significantly reduced the peak amplitude of the fast component. M, Averaged traces taken from an experiment showing that, in the presence of the CB1R antagonist, AM251 (2 µM), bath application of CP 55,940 (2 µM) did not cause a reduction in the peak amplitude. N, Preincubation of the slices in AM251 prevented the CP 55,940-induced reduction of the peak amplitude of light-evoked postsynaptic responses. O–Q, A portion of EYFP-expressing axon terminals is immunoreactive for CB1 (arrows). R, Approximately 40% of EYFP-expressing axonal varicosities were immunopositive for CB1R (156 EYFP+ varicosities were tested in two mice); *p < 0.05 fast versus slow, +gabazine versus control (Ctr), +CGP versus in gabazine (Gab), +CP versus Ctr. n.s., non-significant.

GABAergic cells labeled with intersectional viral strategy represent different populations of INs. A, Example of EYFP-expressing cells (panel A1) that either are (panel A2, arrow) or are not (panel A3, asterisk) also immunopositive for NPY (scale bar = 10 µm). B, Example of two EYFP-expressing cells (panel B1, arrow and arrowhead) that both contain PV (panel B2), but only one of which is also immunopositive for Calb (panel B3, arrow; scale bar = 10 µm). C, Pie chart showing the ratio of EYFP-expressing neurons that contain PV or NPY. Note a large proportion of cells does not express either PV or NPY. D, Example of EYFP-expressing cell (panel D1, asterisk) and non-overlapping cells immunopositive for CaMKII (panels D2, D3; scale bar = 10 µm).

Action potential features distinguish GABAergic cell types labeled with intersectional viral strategy. A, Schematic drawing depicts a horizontal slice with viral expression shown in green (Hip = hippocampus). Whole-cell patch clamp recordings were performed in INTRSECT-ChR2-transfected GABAergic cells (green circles) visualized by blue light illumination. B, Traces exemplifying differences in the full width at action potential half maximum (FWHM), 50% decay of AHP and maximum firing rate for the three electrophysiologically distinct IN groups: fast-spiking INs (FS INs) in orange, CCK⁺ basket cells (CCKBCs) in blue, and NGFCs in cyan. C, 3D plot showing the separation of 33 intracellularly labeled EYFP+ INs based on the three action potential parameters. D–G, Examples of four distinct types of EYFP-expressing INs intracellularly filled by whole-cell recording in vitro. In each case, a maximal intensity projection of a 3D confocal image of the labeled INs is shown together with its firing pattern and the EYFP expression at the soma level. D, An example for a PV+ basket cell (PVBC) identified based on its firing pattern, Calb and PV positivity in its axonal boutons (white arrows in insets) and forming no close appositions (red arrows) with ankyrin G (AnkG)-labeled axon initial segments (delimited by green arrowheads). E, An example for a PV+ axo-axonic cell (AAC) identified based on its firing pattern, PV positivity and Calb negativity in its axonal boutons (white arrows in insets) and forming close appositions by its axonal boutons (red arrows) with an AnkG-labeled axon initial segment (delimited by green arrowheads). F, An example for a CCKBC identified based on its firing pattern and on the CB1 content in its axonal boutons (white arrows in insets). G, An example of a NGFC based on its dendritic and axonal morphology and characteristic firing pattern. H, Pie chart showing the ratio of identified IN types in a group of EYFP-expressing neurons in the BA that were randomly sampled in slice preparations. For D–G depictions of maximal intensity projections of intracellularly filled cells, scale bar = 40 µm, insets = 5 µm; firing pattern scale bar x-axis = 100 ms, y-axis = 10 mV.

In vivo photostimulation of transfected INs during extinction training. A, Schematic of the intersectional strategy used to target BA INs in CCK-Cre;Dlx5/6-Flp double-transgenic mice with INTRSECT pAAV-nEF1a-Con/Fon-hChR2(H134R)-EYFP-WPRE or AAVdj-hSyn-Con/Fon-Arch3.3-EYFP-WPRE virus and shine blue or green light during each CS presentation of extinction-training. For fiber placement maps see Extended Data Figure 5-1. B, Photoexcitation, but not photoinhibition, during extinction training reduced freezing on light-free extinction retrieval the following day; n = 9–13 per group; *p < 0.05 INTRSECT-ChR2 versus INTRSECT-EYFP. #p < 0.05 T0 versus T1 extinction training for INTRSECT-ChR2.