Fig 2: Schematic representation of VIP+ IN connectivity in the mouse BLA. Putative VIP+ IS-INs (magenta) primarily contact other IN subtypes: CCK+/CB1+ basket cells (cyan; comprising CCK+/CB1+/VIP+, light magenta), neurogliaform cells (NGF/NPY; green), and other IS-INs (VIP; magenta), in addition to SOM and PV (orange; Krabbe et al., 2016), which in turn innervate principal neurons (PN) at different plasma membrane domains. Therefore, VIP+ INs are in an ideal position to provide feedforward inhibition to these different INs and facilitate disinhibition of the entire somatodendritic domain of principal cells. An exception is represented by the neurogliaform cells (NGF/NPY) that do not form synapses (axon ending in a circle distant from the PN) onto principal neuron dendrites, but contribute to their inhibition via volume transmission (Manko et al., 2012). CCK+/CB1+/VIP+ have reciprocal connections with VIP+ IS-INs. VIP+ IS-INs receive putative glutamatergic cortical (VGluT1) and subcortical (thalamic and hypothalamic) inputs (VGluT2). Black lines indicate GABAergic inputs and thicker black lines represent higher strength of innervation. Dashed black lines indicate connectivity reported in studies other than the present one. Red lines indicate glutamatergic inputs.

Fig 3: Generation and characterization of Cnr1GFP– flox/GFP– flox: Aldh1l1-CreERT2 mice. (A) Schematic diagram showing the generation of Cnr1GFP– flox/GFP– flox mice by insertion of loxp-SA-GFP-pA-loxp (GFP-flox) sequence into the intron of Cnr1 gene via the CRISPR/Cas9 system. (B) Western blotting showing the expression of CB1R in the nucleus accumbens of Cnr1+/+, Cnr1GFP– flox/GFP– flox, and Cnr1GFP– flox/GFP– flox: Aldh1l1-CreERT2 mouse brain. (C) Representative montage of the hippocampus showing the expression of CB1R in Cnr1GFP– flox/GFP– flox and Cnr1GFP– flox/+ mice. Scale bar: above, 500 μm; below, 50 μm. (D) Representative montage of the nucleus accumbens showing the expression of GFP and S100β in Cnr1GFP– flox/GFP– flox and Cnr1GFP– flox/GFP– flox: Aldh1l1-CreERT2 double-transgenic mice injected with tamoxifen. Arrows indicate the co-expression of GFP and S100β. Scale bar: 50 μm.

Fig 4: Astrocytes express CB1-Rs and DGLα in close proximity to each other in the rat superficial spinal dorsal horn. (a–d) Micrographs of a single 1-µm-thick laser scanning confocal optical section illustrating the co-localization between immunolabeling for GFAP (a marker for astrocytes, magenta; (a) CB1-R (yellow; b) and DGLα (cyan; c). Two puncta immunoreactive for CB1-R or DGLα that are located close together within the confines of a profile stained for GFAP are marked with arrows and appear in mixed colors in the merged image (d). Scale bar: 5 µm. (e) Box plot showing the distribution of distances between CB1-R immunoreactive spots and the closest DGLα immunoreactive spots that were recovered within the confines of GFAP immunoreactive profiles.

Fig 5: Activation of CB1-Rs evokes Ca2+ transients in cell bodies of cultured spinal astrocytes. (a–c) Graphical representation of changes in [Ca2+]i recorded in rat cultured astrocytes following the application of 10 µM 2-AG (a), 10 µM WIN (b), and10 µM anandamide (AEA) (c). (d) Box plot illustrating the amplitudes of calcium transients of rat cultured astrocytes in response to the application of 10 µM 2-AG, 10 µM WIN, and 10 µM AEA. Asterisks indicate significant differences between the amplitudes of calcium transients evoked by 2-AG, WIN and AEA (p = 0.0007, 2-AG vs AEA; p = 0.0354 WIN vs AEA). e: Micrographs showing the fluorescent intensity of a Fluo-8-AM-loaded cultured rat spinal astrocyte at different time points after the application of 10 µM WIN. Scale bar: 50 µm. (f–i) Graphical representations of the time courses of calcium transients evoked by the application of 10 µM WIN (f), 10 µM anandamide and 180 µM ATP (g) in cultured rat spinal astrocytes, and 10 µM WIN and 180 µM ATP in astrocytes cultured from wild type (h) and CB1-R knock out (i) mice. (j) Box plot showing the amplitudes of the calcium transients normalized to ATP responses in the responding cells cultured from rats, wild type mice, and CB1–R knock out mice. Dashed and continuous lines within the boxes represent the mean and median values of the data sets, respectively. Asterisks indicate significant difference between the amplitudes of calcium transients evoked by WIN in wild type and CB1-R knock out mice (p = 0.0118).