. 2023 Oct 4;111(19):3053-3067.e10.

doi: 10.1016/j.neuron.2023.06.023. Epub 2023 Jul 21.

A cortico-amygdala neural substrate for endocannabinoid modulation of fear extinction

Ozge Gunduz-Cinar¹, Laura I Castillo², Maya Xia², Elise Van Leer², Emma T Brockway², Gabrielle A Pollack², Farhana Yasmin³, Olena Bukalo², Aaron Limoges², Sarvar Oreizi-Esfahani², Veronika Kondev⁴, Rita Báldi³, Ao Dong⁵, Judy Harvey-White⁶, Resat Cinar⁷, George Kunos⁶, Yulong Li⁵, Larry S Zweifel⁸, Sachin Patel³, Andrew Holmes⁹

Affiliations

¹ Laboratory of Behavioral and Genomic Neuroscience, National Institute on Alcohol Abuse and Alcoholism, NIH, Bethesda, MD 20892, USA. Electronic address: ozge.gunduzcinar@nih.gov.
² Laboratory of Behavioral and Genomic Neuroscience, National Institute on Alcohol Abuse and Alcoholism, NIH, Bethesda, MD 20892, USA.
³ Northwestern Center for Psychiatric Neuroscience, Department of Psychiatry and Behavioral Sciences, Northwestern University Feinberg School of Medicine, Chicago, IL 60611, USA.
⁴ Vanderbilt Brain Institute, Vanderbilt University, Nashville, TN 37232, USA.
⁵ Peking University School of Life Sciences, PKU-IDG/McGovern Institute for Brain Research, Peking-Tsinghua Center for Life Sciences, Academy for Advanced Interdisciplinary Studies, Peking University, Beijing 100871, China.
⁶ Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, NIH, Bethesda, MD 20892, USA.
⁷ Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, NIH, Bethesda, MD 20892, USA; Section on Fibrotic Disorders, National Institute on Alcohol Abuse and Alcoholism, NIH, Bethesda, MD 20892, USA.
⁸ Department of Psychiatry and Behavioral Sciences, University of Washington, Seattle, WA 98195, USA; Department of Pharmacology, University of Washington, Seattle, WA 98195, USA.
⁹ Laboratory of Behavioral and Genomic Neuroscience, National Institute on Alcohol Abuse and Alcoholism, NIH, Bethesda, MD 20892, USA. Electronic address: andrew.holmes@nih.gov.

PMID: 37480845
PMCID: PMC10592324
DOI: 10.1016/j.neuron.2023.06.023

A cortico-amygdala neural substrate for endocannabinoid modulation of fear extinction

Ozge Gunduz-Cinar et al. Neuron. 2023.

. 2023 Oct 4;111(19):3053-3067.e10.

doi: 10.1016/j.neuron.2023.06.023. Epub 2023 Jul 21.

Authors

Affiliations

¹ Laboratory of Behavioral and Genomic Neuroscience, National Institute on Alcohol Abuse and Alcoholism, NIH, Bethesda, MD 20892, USA. Electronic address: ozge.gunduzcinar@nih.gov.
² Laboratory of Behavioral and Genomic Neuroscience, National Institute on Alcohol Abuse and Alcoholism, NIH, Bethesda, MD 20892, USA.
³ Northwestern Center for Psychiatric Neuroscience, Department of Psychiatry and Behavioral Sciences, Northwestern University Feinberg School of Medicine, Chicago, IL 60611, USA.
⁴ Vanderbilt Brain Institute, Vanderbilt University, Nashville, TN 37232, USA.
⁵ Peking University School of Life Sciences, PKU-IDG/McGovern Institute for Brain Research, Peking-Tsinghua Center for Life Sciences, Academy for Advanced Interdisciplinary Studies, Peking University, Beijing 100871, China.
⁶ Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, NIH, Bethesda, MD 20892, USA.
⁷ Laboratory of Physiologic Studies, National Institute on Alcohol Abuse and Alcoholism, NIH, Bethesda, MD 20892, USA; Section on Fibrotic Disorders, National Institute on Alcohol Abuse and Alcoholism, NIH, Bethesda, MD 20892, USA.
⁸ Department of Psychiatry and Behavioral Sciences, University of Washington, Seattle, WA 98195, USA; Department of Pharmacology, University of Washington, Seattle, WA 98195, USA.
⁹ Laboratory of Behavioral and Genomic Neuroscience, National Institute on Alcohol Abuse and Alcoholism, NIH, Bethesda, MD 20892, USA. Electronic address: andrew.holmes@nih.gov.

PMID: 37480845
PMCID: PMC10592324
DOI: 10.1016/j.neuron.2023.06.023

Abstract

Preclinical and clinical studies implicate endocannabinoids (eCBs) in fear extinction, but the underlying neural circuit basis of these actions is unclear. Here, we employed in vivo optogenetics, eCB biosensor imaging, ex vivo electrophysiology, and CRISPR-Cas9 gene editing in mice to examine whether basolateral amygdala (BLA)-projecting medial prefrontal cortex (mPFC) neurons represent a neural substrate for the effects of eCBs on extinction. We found that photoexcitation of mPFC axons in BLA during extinction mobilizes BLA eCBs. eCB biosensor imaging showed that eCBs exhibit a dynamic stimulus-specific pattern of activity at mPFC^→BLA neurons that tracks extinction learning. Furthermore, using CRISPR-Cas9-mediated gene editing, we demonstrated that extinction memory formation involves eCB activity at cannabinoid CB1 receptors expressed at vmPFC^→BLA synapses. Our findings reveal the temporal characteristics and a neural circuit basis of eCBs' effects on fear extinction and inform efforts to target the eCB system as a therapeutic approach in extinction-deficient neuropsychiatric disorders.

Keywords: CRISPR-Cas9; GRAB(eCB2.0) biosensor; PTSD; amygdala; anxiety; endocannabinoid; extinction; fear; optogenetics; prefrontal cortex.

Published by Elsevier Inc.

PubMed Disclaimer

Conflict of interest statement

Declaration of interests S.P. is or has been a scientific consultant for H. Lundbeck A/S and Psy Therapeutics, unrelated to the current work.

Figures

**Figure 1:. Photoexcitation of mPFC^→BLA neurons facilitates fear extinction through BLA eCBs.**
(**A-C**) Behavioral (A) and *in vivo* optogenetic (B) procedure for mPFC^→BLA photoexcitation during partial (10-trial) extinction training (YFP n=11/ChR2 n=8). Virus expression in mPFC and BLA (C). (D) Freezing increased over conditioning trials (RM-ANOVA: F(1,13)=63.960, P<0.0001, η_p²=0.73) and decreased over extinction trial-blocks (RM-ANOVA: F(1,17)=56.610, P<0.0001, η_p²=0.34). Lower freezing in opsin groups on early extinction (first 5-trial block) vs extinction retrieval 1 (RM-ANOVA: F(1,17)=5.683, P=0.0291, η_p²=0.17) and on extinction retrieval 2 (RM-ANOVA: F(1,17)=11.170, P=0.0039, η_p²=0.28). (**E,F**) Measurement of photoexcitation-related BLA eCB levels (AEA-YFP n=15/AEA-ChR2 n=15, AEA-YFP n=15/AEA-ChR2 n=14) **(E)**. Higher AEA, not 2-AG (P>0.05), levels in ChR2 versus YFP (unpaired t-test: t(28)=2.887, P=0.0074, η_p²=0.23 (F). (**G,H**) Systemic SR141716A (SR, CB1R antagonist) prior to combined extinction training/mPFC^→BLA photoexcitation (YFP-Vehicle (Veh) n=11/YFP-SR n=9/ChR2-Veh n=9/ChR2-SR n=12) (G). Freezing increased over conditioning trials (RM-ANOVA: F(1,36)=209.800, P<0.0001, η_p²=0.75). Freezing decreased over extinction trial-blocks (RM-ANOVA: F(1,37)=116.700, P<0.0001, η_p²=0.37). Lower freezing in ChR2 versus YFP on extinction retrieval 1 (ANOVA opsin x drug interaction: F(1,37)=4.742, P=0.0359, η_p²=0.10, Holm-Šídák's tests: YFP-Veh versus ChR2-Veh: P=0.0488, all others P>.05) and retrieval 2 (ANOVA opsin x drug interaction: F(1,37)=2.054, P=0.1602, η_p²=0.04, Holm-Šídák's tests: YFP-Veh versus ChR2-Veh: P=0.0023, all others P>.05) (H). (**I,J**) Intra-BLA SR141716A prior to combined extinction training/mPFC^→BLA photoexcitation (YFP-Veh n=10/YFP-SR n=8/ChR2=Veh n=15/ChR2-SR n=11) (I). Freezing increased over conditioning trials (RM-ANOVA: F(1,40)=306.400, P<0.0001, η_p²=0.77). Freezing decreased over extinction trial-blocks (RM-ANOVA: F(1,40)=76.000, P<0.0001, η_p²=0.28). Lower freezing in ChR2 versus YFP on extinction retrieval 1 (ANOVA opsin x drug interaction: F(1,40)=4.511, P=0.0399, η_p²=0.09, Holm-Šídák's tests: YFP-Veh versus ChR2-Veh: P=0.0029, all others P>.05) and retrieval 2 (ANOVA opsin x drug interaction: F(1,39)=2.323, P=0.1355, η_p²=0.05, Holm-Šídák's tests: YFP-Veh versus ChR2-Veh: P=0.0417, all others P>.05) (J). Data are represented as mean±SEM. *P<0.05. Table S1 reports full statistical results for all figures.

**Figure 2:. Stimulus-related temporal patterning of eCBs at mPFC^→BLA neurons during extinction.**
(**A-C**) Behavior (A) and *in vivo* fiber photometry (B) procedure for mPFC^→BLA GRAB_eCB2.0 biosensor recordings (n=16 mice). Virus expression in mPFC and BLA (C). (D) Increased GRAB_eCB2.0 signal during conditioning US (paired t-test for US versus baseline: t(15)=6.451, P<0.0001, η_p²=0.74) (not CS P>0.05) presentation. (E) Example photometry traces showing increased GRAB_eCB2.0 signal during extinction training CS off periods. (F,G) Increased GRAB_eCB2.0 signal at CS off (not CS on, P>0.05) during extinction training (paired t-test for first 5-trial block (early extinction) versus baseline: t(15)=5.100, P=0.0001, η_p²=0.63, paired t-test for last 5-trial block (late extinction) versus baseline: t(15)=3.754, P=0.0019, η_p²=0.48) (F). Reduced CS off-related GRAB_eCB2.0 signal on late versus early extinction trials (RM-ANOVA event-effect: F(1,30)=8.524, P=0.0066, η_p²=0.04, test-phase effect: F(1,30)=27.110, P<0.0001, η_p²=0.37, interaction: F(1,30)=1.700, P=0.2023, η_p²=0.02, Holm-Šídák's tests: early-CS off versus early-CS on: P=0.0001, late-CS off versus late-CS on: P=0.0098, late-CS off versus early-CS off: P=0.0212) (G). (H,I) Increased GRAB_eCB2.0 signal at CS off (paired t-test for CS off versus baseline: t(12)=5.085, P=0.0003, η_p²=0.68), decreased at CS on (paired t-test for CS on versus baseline: t(12)=3.234, P=0.0072, η_p²=0.47) periods during fear renewal, but not during extinction retrieval (P>0.05) (H). Higher CS off-related GRAB_eCB2.0 signal on renewal versus retrieval 2 (RM-ANOVA event effect: F(1,26)=4.232, P=0.0498, η_p²=0.03, test-phase effect: F(1,26)=19.390, P=0.0002, η_p²=0.31, interaction: F(1,26)=4.204, P=0.0505, η_p²=0.07, Holm-Šídák's tests: renewal-CS off versus renewal-CS on: P=0.0003, renewal-CS off versus retrieval-CS off: P=0.0126) (I). Data are represented as mean±SEM population average Z-scores normalized to 5-second pre-event baseline. Time-normalized AUC values correspond to 28 post-CS and 2 seconds post-US during conditioning (D), and 30 seconds post-CS on and CS off for extinction training, retrieval and renewal (F-I). ****P<0.0001, ***P<0.001, **P<0.01, #P<0.05.

**Figure 3:. Extinction-related eCBs at mPFC^→BLA neurons in females and non-conditioned males.**
(**A-C**) In females (n=11 mice) freezing increased over conditioning trials (paired t-test for trial 1 versus trial 3: t(10)=5.369, P=0.0003, η_p²=0.74) and decreased from early (first trial-block) to late (last trial-block) extinction training (paired t-test: t(10)=4.006, P=0.0025, η_p²=0.62). Lower freezing on retrieval versus early extinction (paired t-test: t(10)=4.092, P=0.0022, η_p²=0.63) (A). Increased GRAB_eCB2.0 signal during conditioning US presentation (paired t-test for US versus baseline: t(10)=3.659, P=0.0044, η ²=0.57) (not CS P>0.05) presentation (B) and during CS off periods on early (paired t-test for CS off versus baseline: t(10)=2.927, P=0.0151, η_p²=0.46) (not late P>0.05) extinction training (C). (**D-F**) In unpaired conditioned males (n=7 mice), freezing did not increase over conditioning trials (paired t-test for trial 1 versus trial 3: t(6)=2.443, P=0.0503, η_p²=0.50) and was low during extinction and retrieval (D). Increased GRAB_eCB2.0 signal during conditioning US presentation (paired t-test for US versus baseline: t(6)=3.469, P=0.0133, η_p²=0.67) (not CS P>0.05) presentation (E) and no change on extinction trials (P<0.05) (F). (**G-H**) In CS-only conditioned males (n=7 mice), freezing did not increase over conditioning trials (P<0.05) and was low during extinction training and retrieval (G). Decreased GRAB_eCB2.0 signal during conditioning CS presentation (paired t-test for CS versus baseline: t(6)=0.311, P=0.7663, η_p²=0.02) (not the US-equivalent period, P>0.05) (H) and showed no change during extinction training CS on (P>0.05) or CS off periods (P>0.05) (I). (**J-L**) In US-only conditioned males (n=6 mice), freezing increased over conditioning trials (paired t-test for trial 1 versus trial 3: t(5)=3.503, P=0.0172, η_p²=0.71) and was low on extinction trials. (J). Decreased GRAB_eCB2.0 signal during conditioning US presentation (paired t-test for US versus baseline: t(5)=3.181, P=0.0245, η_p²=0.67) (not CS-equivalent period, P>0.05) (K) and CS off period during early (paired t-test for CS off versus baseline t(5)=3.183, P=0.0245, η_p²=0.67) (not late, P>0.05) extinction training (L). Data are represented as mean±SEM population average Z-scores normalized to 5-second pre-event baseline. Time-normalized AUC values correspond to 28 post-CS and 2 seconds post-US during conditioning, and 30 seconds post-CS on and CS off for extinction training, retrieval and renewal. ***P<0.001, **P<0.01, *P<0.05 with multiple-comparison Bonferroni correction for extinction GRAB_eCB2.0 data.

**Figure 4:. eCB-CB1R signaling modulates mPFC^→BLA neurotransmission.**
(**A-D**) *In vitro* recordings of CP55–940 (CP, CB1R agonist, Veh n=4 mice/22 cells, CP n=4 mice/15 cells), JZL184 (JZL, MAGL inhibitor, Veh n=3 mice/19 cells, JZL n=3 mice/14 cells) and PF-3845 (PF, FAAH inhibitor, Veh n=5 mice/22 cells, PF n=5 mice/18 cells) effects on ChR2-mediated optically-evoked EPSCs at vmPFC^→BLA synapses (A). Higher oEPSC paired-pulse ratio (PPR) in slices incubated with CP (unpaired t-test for CP versus Veh: t(35)=3.344, P=0.0020, η_p²=0.24) (B), JZL (unpaired t-test for JZL versus veh: t(31)=3.587, P=0.0011, η_p²=0.29) (C) or PF (unpaired t-test versus veh: t(38)=2.389, P=0.0220, η_p²=0.13) (D), with example traces. (**E-G**) *In vitro* recordings of SR141716A (SR, CB1R antagonist) effects on ChR2-mediated oEPSCs at mPFC^→BLA synapses following vmPFC neuron high-frequency stimulation (HFS) (n=3 mice/6 cells Veh/n=3 mice/6 cells SR) (E). Versus Veh, SR attenuated an HFS-induced decrease in oEPSCs, with example traces (ANOVA drug x time interaction: F(35,385)=1.764, P=0.0058, η_p²=0.03) (F). Lower oEPSC PPR 25–30 minutes post-HFS in slices incubated with SR (unpaired t-test for SR versus Veh: t(10.97)=2.214, P=0.0489, η_p²=0.31) (G). (**H-I**) *In vitro* recordings of effects of CP on ChR2-mediated oEPSCs at mPFC^→BLA synapses following virus-mediated *Cnr1* deletion from vmPFC^→BLA neurons (n=3 mice/6 cells Veh/n=3 mice/6 cells SR) (H). CP decreased oEPSCs in *Cnr1* intact not deleted mice (ANOVA drug x time interaction: F(1,17)=15.030, P=0.0012, η_p²=0.20), with example traces (I). Data are represented as mean±SEM .***P<0.001, **P<0.01, *P<0.05.

**Figure 5:. CRISPR-Cas9 *Cnr1* deletion from vmPFC^→BLA not dmPFC^→BLA neurons impairs extinction.**
(**A-C**) Viral strategy for vmPFC^→BLA CRISPR-Cas9 *Cnr1* mutation in Rosa26^fsTRAP mice (A). Less GFP-enriched purified RNA (not GFP-unbound input RNA) expression in mutated versus intact (unpaired t-test: t(11)=2.543, P=0.0273, η_p²=0.37, intact n=6/mutated n=7) (B). Example images from combined BaseScope *in situ* hybridization and immunohistochemistry showing *Cnr1* mRNA loss in Cre-expressing BLA-projecting vmPFC neurons in mutated (not intact) (C). (**D,E**) *In vitro* recordings of CP55–940 (CP, synthetic cannabinoid, CB1R agonist) on ChR2-mediated optically-evoked EPSCs at vmPFC^→BLA synapses (D) showing lesser effect in mutated versus intact (ANOVA time x group: F(34,408=4.208, P<0.0001, η_p²=0.07, n=3 mice/6 cells intact/n=3 mice/6 cells mutated), with example traces (E). (**F,G**) vmPFC^→BLA *Cnr1* mutation (F). Freezing increased over conditioning trials (RM-ANOVA: F(1,18)=279.100, P<0.0001, η_p²=0.88) and decreased over extinction trial-blocks (RM-ANOVA: F(1,18)=124.400, P<0.0001, η_p²=0.74). Less freezing on extinction retrieval 1 in mutated versus intact (RM-ANOVA F(1,18)=6.338, P=0.0215, η_p²=0.11) (intact n=9, mutated n=11) (G). (**H,I**) dmPFC^→BLA *Cnr1* mutation (H). Freezing increased over conditioning trials (RM-ANOVA: F(1,13)=87.360, P<0.0001, η_p²=0.69), decreased over extinction trial-blocks (RM-ANOVA: F(1,13)=45.370, P<0.0001, η_p²=0.58) and was similar in mutated and intact across test-phases (intact n=7, mutated n=8) (I). (**J,K**) Less light/dark test compartment time in vmPFC^→BLA *Cnr1* mutated (J) (unpaired t-test: t(18)=2.307, P=0.0331, η_p²=0.23, intact n=9/mutated n=11), not dmPFC^→BLA *Cnr1* mutated (intact n=8/mutated n=8) (K), versus intact. (**L,M**) Less novel open field distance moved (unpaired t-test: t(14)=2.467, P=0.0272, η_p²=0.30, intact n=8/mutated n=8) in dmPFC^→BLA *Cnr1* mutated (M), not vmPFC^→BLA *Cnr1* mutated (L), versus intact. Data are represented as mean±SEM. ***P<0.001, *P<0.05.

See this image and copyright information in PMC

Comment in

Endocannabinoids modulate fear extinction controlled by a cortical-amygdala projection.
Adhikari A. Adhikari A. Neuron. 2023 Oct 4;111(19):2948-2950. doi: 10.1016/j.neuron.2023.09.011. Neuron. 2023. PMID: 37797579

References

1. Craske MG, Stein MB, Eley TC, Milad MR, Holmes A, Rapee RM, and Wittchen HU (2017). Anxiety disorders. Nature reviews. Disease primers 3, 17024. 10.1038/nrdp.2017.24. - DOI - PMC - PubMed
1. Pavlov IP (1927). Conditioned reflexes (Oxford University Press).
1. Hariri AR, and Holmes A (2015). Finding translation in stress research. Nat Neurosci 18, 1347–1352. 10.1038/nn.4111. - DOI - PMC - PubMed
1. Gunduz-Cinar O, Hill MN, McEwen BS, and Holmes A (2013). Amygdala FAAH and anandamide: mediating protection and recovery from stress. Trends Pharmacol Sci 34, 637–644. 10.1016/j.tips.2013.08.008. - DOI - PMC - PubMed
1. Mayo LM, Rabinak CA, Hill MN, and Heilig M (2021). Targeting the Endocannabinoid System in the Treatment of Posttraumatic Stress Disorder: A Promising Case of Preclinical-Clinical Translation? Biol Psychiatry 10.1016/j.biopsych.2021.07.019. - DOI - PMC - PubMed

Publication types

Actions
Actions

MeSH terms

Actions
Actions
Actions
Actions
Actions
Actions
Actions
Actions

Save citation to file

Email citation

Add to Collections

Add to My Bibliography

Your saved search

Create a file for external citation management software

Your RSS Feed

A cortico-amygdala neural substrate for endocannabinoid modulation of fear extinction

Affiliations

A cortico-amygdala neural substrate for endocannabinoid modulation of fear extinction

Authors

Affiliations

Abstract

Conflict of interest statement

Figures

Comment in

References

Publication types

MeSH terms

Substances

Grants and funding

LinkOut - more resources

Full Text Sources

Research Materials