Ventral hippocampal diacylglycerol lipase-alpha deletion decreases avoidance behaviors and alters excitation-inhibition balance

Abstract

The endogenous cannabinoid, 2-arachidonoylglycerol (2-AG), plays a key role in the regulation of anxiety- and stress-related behavioral phenotypes and may represent a novel target for the treatment of anxiety disorders. However, recent studies have suggested a more complex role for 2-AG signaling in the regulation of stress responsivity, including increases in acute fear responses after 2-AG augmentation under some conditions. Thus, 2-AG signaling within distinct brain regions and circuits could regulate anxiety-like behavior and stress responsivity in opposing manners. The ventral hippocampus (vHPC) is a critical region for emotional processing, anxiety-like behaviors, and stress responding. Here, we use a conditional knock-out of the 2-AG synthesis enzyme, diacylglycerol lipase α (DAGLα), to study the role of vHPC 2-AG signaling in the regulation of affective behavior. We show that vHPC DAGLα deletion decreases avoidance behaviors both basally and following an acute stress exposure. Genetic deletion of vHPC DAGLα also promotes stress resiliency, with no effect on fear acquisition, expression, or contextual fear generalization. Using slice electrophysiology, we demonstrate that vHPC DAGLα deletion shifts vHPC activity towards enhanced inhibition. Together, these data indicate endogenous 2-AG signaling in the vHPC promotes avoidance and increases stress reactivity, confirming the notion that 2-AG signaling within distinct brain regions may exert divergent effects on anxiety states and stress adaptability.

Keywords: Anxiety; Endocannabinoid; Fear conditioning; Stress; Ventral hippocampus.

Fig. 1

vHPC DAGLα deletion decreases avoidance behaviors basally and following acute restraint stress A) Schematic of experimental design and timeline. DAGLα^fl/fl male and female mice were bilaterally injected with virus expressing AAV-GFP (25) or AAV-Cre (28) into vHPC. Unpaired, two-tailed t-test used for all analysis. *, p < 0.05; **, p < 0.01; ***, p < 0.001 B) Representative image of AAV-GFP injection (left) and AAV-Cre injection (right); scale bar represents 2000μm C) Schematic of elevated plus maze (EPM). Darker shaded regions represent the closed arm; lighter shader regions correspond with the open arm D) Representative heat maps of AAV-GFP (left) or AAV-Cre (right) injected mice during the EPM test E) Quantification of total distance travelled during EPM in mice injected with AAV-GFP (n = 25) or AAV-Cre (n = 28) F) Quantification of the number of open arm entries during EPM test (GFP: n = 25; Cre: n = 28) G) Assessment of total time spent in the open arm during EPM test (GFP: n = 25; Cre: n = 28) H) Quantification of the %distance travelled in the open arm (GFP: n = 25; Cre: n = 28) I) Schematic of restraint stress and subsequent elevated zero maze (EZM) testing of mice injected with AAV-GFP (10) and AAV-Cre (14); darker shaded regions represent the closed arm; lighter shader regions correspond with the open arm J) Representative heat maps during EZM of AAV-GFP (left) or AAV-Cre (right) injected mice K) Quantification of total distance travelled during EZM in mice injected with AAV-GFP (n = 10) or AAV-Cre (n = 14) in the vHPC L) Assessment of the total number of entries into open arm during the EZM test (GFP: n = 10; Cre: n = 14) M) Quantification of the total time spent in open arm during EZM test (GFP: n = 10; Cre: n = 14) N) Assessment of the % distance travelled in open arm during EZM (GFP: n = 10; Cre: n = 14).

Fig. 2

vHPC DAGLα deletion decreases stress-induced avoidance in the NIH assay and promotes stress resiliency A) Timeline of repeated novelty-induced hypophagia (rNIH) experiment in a separate cohort of mice. DAGLα^fl/fl male mice were bilaterally injected with virus expressing AAV-GFP (15) or AAV-Cre (12) into the vHPC. 2-Way ANOVA (C, F), unpaired, two-tailed, t-test (D-E, G-H, J), or Chi square (K–L) used for analysis. *, p < 0.05; ***, p < 0.001; ****, p < 0.0001 B) Schematic of foot shock stress. Mice were exposed to 5 days of foot shock stress, which consisted of 6 tone (30s)-shock (2s, 0.7 mA) pairings C) Quantification of the latency to consume Ensure over training days and NIH test days D) Quantification of feeding latency during basal NIH test, without shock (GFP: n = 15; Cre: n = 12) E) Assessment of feeding latency after 5 consecutive days of foot shock (GFP: n = 15; Cre: n = 12) F) Quantification of total Ensure consumption across training and NIH test days G) Assessment of consumption of Ensure on basal NIH test (GFP: n = 15; Cre: n = 12) H) Quantification of total consumption on NIH test, after 5 days of foot shock stress (GFP: n = 15; Cre: n = 12) I) Criteria used to separate mice into susceptible or resilient following NIH. Mice with a stress-induced latency of <120s were classified as resilient J) Quantification of the average stress-induced change in latency after one day (left) or 5 days (right) of foot shock stress (GFP: n = 15; Cre: n = 12) K) Assessment of the proportion of stress resilient mice that were injected with AAV-GFP compared to AAV-Cre, after one day of foot shock L) Assessment of the proportion of stress resilient mice that were injected with AAV-GFP compared to AAV-Cre, after 5 days of foot shock.

Fig. 3

vHPC DAGLα deletion has no effect on conditioned freezing or generalization A) Schematic of fear conditioning paradigm. Following rNIH testing, the same cohort of mice were fear conditioned in Context A, during which mice were exposed to 6 tone (CS+, 30s)-shock (US, 2s, 0.7 mA) pairings. 24 h after conditioning, mice were tested for tone (CS+)-induced fear expression in Context B, where mice were exposed to 6 tones in the absence of shock. On day 3, mice were placed back in Context A to assess contextually conditioned fear behavior, and exposed to 6 tones in the absence of shock. Unpaired, two-tailed, t-test (B, D-F, H-J, L-M), or 2-Way ANOVA (C, G, K) used for analysis. B) Quantification of baseline %freezing time during fear conditioning (GFP: n = 11; Cre: n = 10) C) Assessment of the %time spent freezing over 6 tone (CS+) presentations during conditioning day D) Quantification of the average %time spent freezing during CS+ (GFP: n = 11; Cre: n = 10) E) Quantification of the average %freezing time in-between tone/shock presentations (IEI) (GFP: n = 11; Cre: n = 10) F) Assessment of generalized freezing in Context B (GFP: n = 11; Cre: n = 10) G) Quantification of conditioned freezing across tones (CS+) in Context B. H) Quantification of the average conditioned freezing to the CS+ (GFP: n = 11; Cre: n = 10) I) Assessment of %time spent freezing during IEI in Context B (GFP: n = 11; Cre: n = 10) J) Quantification of contextually conditioned freezing during re-exposure to Context A (GFP: n = 11; Cre: n = 10) K) Quantification of %time spent freezing across tones (CS+) in Context A L) Assessment of the average %freezing time during CS+ (GFP: n = 11; Cre: n = 10) M) Quantification of the %time spent freezing in-between tones in Context A (GFP: n = 11; Cre: n = 10).

Fig. 4

vHPC DAGLα deletion shifts vHPC activity towards greater synaptic inhibition. A) Schematic of experimental design. DAGLα^fl/fl mice were injected with AAV-GFP or AAV-Cre. At least 5 weeks later, brain slices containing the vHPC from male (circle) and female (triangle) mice were obtained for whole-cell slice electrophysiology. Glutamatergic and GABAergic synaptic transmission was assessed from the same cell. N,n reflect N = number of cells, from n = number of mice. Unpaired, two-tailed t-test used for analysis *, p < 0.05; **, p < 0.01; ***p < 0.001; ****p < 0.0001. B) Representative trace of spontaneous inhibitory postsynaptic currents (sIPSCs) (top, black) that can be blocked with GABA_A receptor antagonist, picrotoxin (50 μM) (top, gray) and representative trace of spontaneous excitatory postsynaptic currents (sEPSCs) (bottom, black) that can be blocked with AMPA/NMDA receptor antagonists, CNQX (20 μM)/D-AP5 (50 μM) (bottom, gray). C) Quantification of sEPSC frequency from slices obtained from AAV-GFP (n = 12,3) or AAV-Cre (n = 16,4) injected mice (left). Cumulative probability plot of sEPSC interevent interval (IEI) with inset of representative traces from AAV-GFP (green) and AAV-Cre (black) injected mice (right). D) Quantification of sIPSC frequency (AAV-GFP: n = 12,3; AAV-Cre: n = 15,4) (left). Cumulative probability plot of sIPSC IEI with inset of representative traces from AAV-GFP (green) and AAV-Cre (black) injected mice (right). E) Assessment of the excitatory (sEPSC)/inhibitory (sIPSC) (E/I) frequency ratio (AAV-GFP: n = 12,3; AAV-Cre: n = 10,4). F) Quantification of the amplitude of sEPSCs in the vHPC from AAV-GFP (n = 12,3) or AAV-Cre (n = 16,4) injected mice. G) Quantification of sIPSC amplitude in slices obtained from DAGLα^fl/fl mice injected with AAV-GFP (n = 12,3) or AAV-Cre (n = 16,4). H) Assessment of the E/I amplitude ratio comparing AAV-GFP (n = 12,3) injected mice to AAV-Cre (n = 12,4) injected mice. I) Quantification of sEPSC ½ width (AAV-GFP: n = 12, 3; AAV-Cre: n = 14, 4). J) Quantification of sIPSC ½ width (AAV-GFP: n = 12, 3; AAV-Cre: n = 15, 4). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 5

Acute pharmacological DAGL inhibition biases vHPC output towards greater inhibition. A) Schematic of experimental design. Slices from male (circle) and female (triangle) wild-type mice were incubated in Veh (Vehicle; ACSF) or DO34 (ACSF + DO34; 2.5 μM). sEPSCs and sIPSCs were recorded from the same cell. N,n reflect N = number of cells, from n = number of mice. Unpaired, two-tailed t-test used for analysis *, p < 0.05; **, p < 0.01. B) Quantification of sEPSC frequency in slices incubated in Veh (n = 14,5) or DO34 (n = 16,5) (left). Cumulative probability plot of sEPSC interevent interval (IEI) (right) and representative traces of sEPSCs from Veh (black) and DO34 (blue) incubated slices (inset). C) Quantification of sIPSC frequency (Veh: n = 13,5; DO34: n = 18,5) (left). Cumulative probability plot of sIPSC IEI with inset of representative traces of sIPSCs from Veh (black) and DO34 (blue) incubated slices (right). D) Analysis of E/I ratio frequency (Veh: n = 12,5; DO34: n = 16,5). E) Quantification of sEPSC amplitude in slices incubated in Veh (n = 14,5) or DO34 (n = 15,5). F) Quantification of sIPSC amplitude (Veh: n = 13,5; DO34: n = 18,5) G) Analysis of E/I amplitude ratio following DO34 incubation (n = 17,5) compared to Veh (n = 12,5). H) Assessment of sEPSC ½ width from slices incubated in Veh (n = 14,5) or DO34 (n = 17,5). I) Quantification of the ½ width of sIPSCs (Veh: n = 12,5) (DO34: n = 18,5). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)