Mice can form the trace fear memory with trace fear conditioning protocol, which places a trace interval between offset of the conditioned stimulus (CS) and onset of the unconditioned stimulus (US). Despite this tiny change in the temporal relationship between CS and US, neural circuit mediating trace fear memory formation differs from delay fear memory formation, whose protocol does not contain this trace interval. It has been universally accepted that amygdala, an almond shape structure lying deeply in the brain, is the emotional hub. Moreover, neural plasticity happens in the lateral nucleus of amygdala (LA) is widely thought as the synaptic basis underlying fear memory formation.

Studies already found that trace fear conditioning can recruit more brain regions compared to delay paradigm. Among them, entorhinal cortex (EC) was discovered to be specifically involved in trace fear memory formation. However, the modulation mechanism of trace fear memory formation by EC is still unclear.

Serving as the most abundant neuropeptide in the central nervous system, cholecystokinin (CCK) has been unveiled to participate in many processes and functions of the brain, including depression, anxiety, learning, and memory. Here we found lacking CCK can cause the deficit in trace fear memory formation of mice. Subjected to an auditory-cued trace fear conditioning, transgenic CCK knockout (CCK-/-) mice showed less freezing behavior towards CS, compared to their wildtype (WT) control. Moreover, this deficit was independent of the training paradigm with long or short trace interval. Also, the CCK-/- mutant showed bad discrimination between US-paired CS and US-unpaired control auditory stimulus, which also indicated a deficit in establishing the association between CS and US. Further confirmations excluded other auditory or fear-related abnormalities may cause the deficit of conditioned fear formation.

Secondly, we failed to induce long term potentiation (LTP) of auditory evoked potential (AEP) in LA of CCK-/- mice, while LTP was readily induced in their WT control. Moreover, this LTP induction deficit could be rescued by exogenous sulfated CCK octapeptide (CCK-8s) application, after immunohistochemistry (IHC) confirmed the existence of CCK receptor in LA of CCK-/- mice. In accordance with the physiological rescue, intravenous and intraperitoneal administration of CCK tetrapeptide (CCK-4) was also able to rescue the deficit of CCK-/- in trace fear conditioning.

Thirdly, EC-LA connection was investigated by injecting retrograde neural tracer in LA. Then we used both pharmacological and chemogenetic tools to interfere with EC during trace fear conditioning and found that this suppressed intervention caused the decline of freezing percentage. Moreover, Cre-dependent retrograde virus labeling revealed that EC CCK neurons could send neural projection to LA. So we used Cre-dependent inhibitory chemogenetic virus to target CCK neurons located in EC, and inhibited these neurons during trace fear conditioning. Similarly, results showed this cell type-specific inhibition also cause the decline of freezing percentage.

Furthermore, we applied the Cre-dependent virus to transfect entorhinal cortical CCK-expressing neurons with excitatory optogenetic opsin. Then we inserted optrode into LA, for recording AEP and stimulating EC-derived CCK axons distributed in LA. We found that high-frequency activation of CCK axons can induce LTP of AEP. High-frequency neural activation was believed to trigger the release of neuropeptide from axon terminals. Therefore neural plasticity in LA was modulated by EC-derived CCK.

Finally, to confirm the role of EC-derived CCK in fear memory formation, we transfected EC CCK neurons with inhibitory optogenetic opsin and fluorescent reporter. The fluorescence signal of EC axons was found in LA. Then optic cannulae were implanted towards LA to illuminate these CCK positive axons during trace fear conditioning. And we found this intervention could impair trace fear memory formation, which suggested the critical role EC-derived CCK plays in trace fear memory formation.

In conclusion, our current study showed that CCK from EC regulate trace fear memory formation via enable neural plasticity happens in LA. During fear conditioning, CS and US input converged at LA. After training, the CS-US association was established so that CS alone can trigger downstream fear expression. Our study demonstrated EC-derived CCK is significantly involved in this CS-US association. Despite this, detailed mechanism of CCK release, binding and downstream signaling pathway for modulating CS-US association, still needs further investigations.