Cholecystokinin (CCK) was first described as a substance within intestinal extract that can stimulate gallbladder contraction and evacuation (Ivy and Oldberg 1928). Later it was found to be one of the most abundant neuropeptides in the brain (Beinfeld et al. 1981). The bioactive form of CCK, sulfated CCK-8 or CCK8s, is thought to regulate various neuronal processes, including neuronal excitability (Ma et al. 2006; Deng et al. 2010; S. Wang et al. 2011), anxiety (Daugé and Léna 1998), food intake (Gibbs, Young, and Smith 1973), as well as several types of learning and memory (Huston et al. 1998).

Lesions in the EC or the hippocampus lead to memory deficits including temporal, spatial, episodic memory, as well as associative fear conditioning (Amaral, Zola-Morgan, and Squire 1986; Nakazawa et al. 2010; Squire, Stark, and Clark 2004). Our earlier studies showed that activation of CCK positive neurons that originated from the entorhinal cortex (EC) induces long-term potentiation (LTP) and neuroplasticity in the auditory cortex (Li et al. 2013).

In the present study, we found that Cck-CreER mice, which show dramatic low expression of Cck mRNA in the whole brain, exhibit severe deficits in spatial memory and theta burst stimulation (TBS) induced hippocampal LTP formation. In the Morris Water Maze Test, compared to normal behaving control C57 mice, Cck-CreER mice could not shorten the latency used in looking for the hidden platform during the five days of learning and spent nearly equivalent time in each quadrant in the probe test on Day 6. TBS-induced LTP in the classical hippocampal CA3 to CA1 pathway was also impaired in Cck-CreER mice. In contrast, the improved learning curve of platform finding and acquired spatial memory of the hidden platform in CCK Type B Receptor (CCKBR)-KO mouse group were still observed, along with an intact CA3 to CA1 LTP in the acute study. This means CCKBR is not responsible for CCK’s function in hippocampal spatial function and TBS-induced LTP formation. I thus injected Devazepide, antagonist of another CCK receptor, Type A or CCKAR, into the hippocampus and found it sufficient to block TBS-LTP in the CA3 to CA1 pathway. This result is consistent with the dense mRNA expression pattern, and predominant immunostaining pattern of CCKAR in the hippocampus (Datson et al. 2001, Allen Brain Atlas, and our own results).

Neuropeptides, such as CCK, are usually released from the terminals after high-frequency stimulation of the neurons. We hypothesized that CCK required for LTP formation is released after TBS stimulation. Thus, I utilized Cck-ires-Cre and optogenetic tools to precisely study CCK release from CCK positive neurons.

CCK positive interneurons are abundant and comprise one of the major GABAergic neuron groups in the hippocampus. Whether these interneurons contribute to TBS-induced hippocampal LTP formation remains unknown. In a recent study, local inhibitory CCK neurons were found to receive input from long-range inhibitory neurons originating in the lateral EC and to modulate tri-synaptic LTP in the hippocampus (J. Basu et al. 2016). With AAV-EF1a-DIO-ChR2-EYFP virus injected in the hippocampus of Cck-ires-Cre mice, hippocampal CCK positive neurons were manipulated with either low-frequency (LF: 1Hz) or high-frequency (HF: 80Hz) of 473nm blue light; yet neither blue light stimulation induced LTP in the CA3 to CA1 pathway. Both manipulations led to robust and mild Long-term Depression (LTD) respectively. We thus hypothesize that the CCK causing hippocampal TBS-LTP is not from local CCK neurons but projecting neurons of EC, the hub between neocortex and the hippocampus. The same virus was then injected into medial and lateral EC, and the CCK positive projections to the hippocampus were activated by LF and HF blue light. While LF blue light did not induce LTP, HF blue light which was paired with LF ES stimulation of CA3, initiated LTP in the CA3 to CA1 pathway. The light-induced LTP was accompanied by an increase of CCK concentration in the hippocampus. Using novel AAVs, which encode Cre dependent ChR2 and shRNA against Cck mRNA, I confirmed that presynaptic Cck knocking down in EC CCKergic neurons prevented the formation of LTP in the CA3 to CA1 pathway. The nature of these projecting CCK neurons was further studied by immunohistochemical staining of excitatory neuron markers (CaMKIIa and VGLUT1) and inhibitory markers (GAD1, GAD2, and VGAT). I found both excitatory and inhibitory CCK neurons synapse on hippocampal neurons. A further study showed that HF light activation of excitatory CCKergic projections to the hippocampus, which expressed AAV-CamkIIa-DIO-ChR2 virus, did initiate CA3 to CA1 LTP of similar increment to the result of AAV-EF1a-DIO-ChR2 manipulation. In contrast, ontogenetically activating the EC to hippocampus inhibitory terminals, including CCK positive ones, failed to induce LTP. The results support the hypothesis that high frequency stimulation induce CCK released from EC CCKergic terminals in hippocampus, and initiate LTP formation in CA3 to CA1 pathway.

At the behavioral level, the EC to hippocampus CCK projection strengthens the connectivity between place cells after HF blue light stimulation and LF ES pairing. Enhancement of intra-place cell connections was first confirmed in an acute study showing that both theta burst electrical stimulation (ES-TBS) and CCK perfusion, successfully initiated LTP in the longitudinal CA1-CA1 pathway. Then in a spatial fear memory test, a device containing one optical fiber and three electrodes was implanted in the hippocampus of C57 and Cck-ires-Cre mice, the latter of which received Cre-dependent virus injection in EC. Three groups of artificial place cells (APCs) were induced to represent three different regions in the chamber by low-frequency electrical stimulation via each electrode. After four days of training, footshock (FS) was given in one of the regions to induce a spatial fear memory of that region, and electrical stimulation was given to the corresponding electrode. In this way, the footshock region was represented by the corresponding APCs and electrode. Under anesthesia, another two electrodes were used to pair the corresponding APCs to the FS APCs with HF blue light stimulation or control treatment (LF blue light pairing, LF ES or null pairing). After at least three days of pairing, the mice showed decreased occupancy in the HF-light-paired regions, while the control regions showed no decreased occupancy. This indicated that the fear against the FS region extended to HF-light-paired region, and it reflected that HF activation of the EC to hippocampus CCK projection strengthens place cells connectivity at the behavioral level.

Conclusion
In this study, I found that the excitatory CCKergic neurons in the entorhinal cortex densely project to the hippocampus. High-frequency stimulation (ES-TBS or HF light stimulation) of these CCK terminals/neurons could release CCK and switch on the LTP in the hippocampal CA3 to CA1 pathway. This long-term enhanced plasticity was reflected in a spatial fear memory behavioral test. I also demonstrated that the activation of CCK-positive neurons in the hippocampus does not induce LTP in the CA3 to CA1 pathway. In summary, the study demonstrated a crucial function of the EC to hippocampus CCK-positive projection in the LTP formation in the hippocampus.