The medial prefrontal cortex (mPFC) has been implicated in the extinction of emotional memories, including conditioned fear. We found that ventral hippocampal (vHPC) projections to the infralimbic (IL) cortex recruited parvalbumin-expressing interneurons to counter the expression of extinguished fear and promote fear relapse. Whole-cell recordings ex vivo revealed that optogenetic activation of vHPC input to amygdala-projecting pyramidal neurons in the IL was dominated by feed-forward inhibition. Selectively silencing parvalbumin-expressing, but not somatostatin-expressing, interneurons in the IL eliminated vHPC-mediated inhibition. In behaving rats, pharmacogenetic activation of vHPC→IL projections impaired extinction recall, whereas silencing IL projectors diminished fear renewal. Intra-IL infusion of GABA receptor agonists or antagonists, respectively, reproduced these effects. Together, our findings describe a previously unknown circuit mechanism for the contextual control of fear, and indicate that vHPC-mediated inhibition of IL is an essential neural substrate for fear relapse.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
We thank A. Woodruff for comments on the manuscript. We thank L. Xu, University of North Carolina Vector Core, University of Pennsylvania Vector Core, and the Institute of Molecular Genetics of Montpellier for producing viruses. This work was supported by the US National Institutes of Health (R01MH065961 to S.M.; F31MH107113 to T.D.G.; F31MH112208 to T.F.G.), a McKnight Memory and Cognitive Disorders Award to S.M., and Australian Research Council (CE140100007) and National Health and Medical Research Grants to P.S.
Integrated supplementary information
Comparison of hippocampal inputs onto principal neurons in the IL and PL reveal significantly larger EPSCs in IL neurons compared to caudal or rostral PL responses (rostral PL: n = 12; caudal PL: n = 5; IL: n = 21 from 7 animals; one-way ANOVA with Dunnett’s multiple comparison test, F2,35 = 6.62, *P = 0.037). Dots represent data from individual cells. Error bars indicate means ± s.e.m.
a, Example of a HPC-driven response in a pyramidal neuron, containing the initial excitatory response, followed by an inhibitory component. The 5-fold enlarged inset shows the time-locked onset of individual responses (grey) to the light stimulation (blue) without any response-failures. Averaged response shown in black. Analysis of EPSCs show very low response jitter (b) (individual jitter of neurons shown as grey dots; n = 19) and response latency (c) (n = 14), both typical parameters for monosynaptic responses. Error bars indicate means ± s.e.m.
a, Schematic showing experimental setup. Optical terminal stimulation of ventral hippocampal inputs and whole-cell recording of pyramidal neurons in the IL. For the spiking suppression experiments (d), local electrical stimulation (battery symbol) was used. b, Current-clamp responses to hippocampal terminal release before (black trace) and after (red trace) the application of the GABAB-receptor antagonist CGP55845 (n = 3), revealing the fast (yellow) and slow IPSCs (green). c, Bar graphs with response patterns of IL principal neurons to optical hippocampal stimulation containing fast (top, yellow) and slow (bottom, green) inhibitory conductances (percentage of the total). d, Investigation of spiking suppression of IL principal neurons by the hippocampus. Electrical suprathreshold stimulation was used to evoke spiking (left) while presenting optical hippocampal stimulation 150 ms before the spiking event (n = 8). Blue bars represent optical stimulation.
Supplementary Figure 4 Feed-forward inhibition onto pyramidal neurons is mediated by local IL interneurons.
a, Schematic for electrical stimulation (battery symbol) of IL tissue ex vivo in the presence of AMPA- and NMDA-receptor antagonists NBQX and APV, respectively, to isolate GABAergic transmission. b, Voltage clamp (left) and current-clamp recordings (right) revealed inhibitory responses that contain both fast (yellow) and slow (green) inhibitory components. c, Application of the GABAB-receptor antagonist CGP55845 (red trace) isolated the fast, inhibitory conductance, which was blocked by the GABAA/C-receptor antagonist picrotoxin (green trace). Holding voltage: −60 mV.
a, Schematic design of experimental approach alongside a representative image depicting DREADD-expressing neurons of animals implanted with multichannel recording arrays into the IL (40 μm coronal section; white bar inset = 250 μm). b, Vehicle (VEH) injections did not cause a significant change in the spontaneous activity (20 s bins) of IL neurons (n = 15 for hM4Di; n = 27 for mCherry control) (left). When the hM4D(Gi)-expressing animal was injected with 1 mg/kg (middle) or 3 mg/kg (right) of CNO, IL neurons (n = 18 for 1 mg/kg; n = 16 for 3 mg/kg) exhibited a significant reduction in spontaneous firing relative to neurons of control virus-infected animals (n = 25 for 1 mg/kg; n = 23 for 3 mg/kg; repeated measures ANOVA, main effects of virus: F1,41 = 10.912, **P = 0.0020 for 1 mg/kg; F1,37 = 24.375, ***P < 0.0001 for 3 mg/kg).
Supplementary Figure 6 Pharmacological inactivation of the IL impedes retrieval of extinguished fear.
Test data show mean baseline freezing (3 min), mean freezing during nine 5-ITI blocks (30-sec ITIs) and a during a post-trial period (150 sec) following infusions of muscimol or vehicle into the IL (MUSC, n = 6; VEH, n = 10; repeated measures ANOVA, main effect of drug: F1,14 = 35.78, ***P < 0.0001). Corresponding conditioning and extinction data are shown in Fig. 4. Error bars indicate means ± s.e.m.