The distribution profile of CCK-positive neurons in the dorsal hippocampus.

(A) Schematic illustrating the breeding strategy used to generate CCK-Cre/Ai14 reporter mice, in which CCK-expressing neurons are labeled with tdTomato. (B) Representative fluorescence image of a dorsal hippocampal slice from a CCK-Cre/Ai14 mouse immunostained with the pan-neuronal marker NeuN (rabbit anti-NeuN, Alexa Fluor 488), showing overall hippocampal anatomy. Scale bar, 1000 µm. (C) Higher-magnification images showing CCK-positive neurons identified by tdTomato expression (red) colocalized with NeuN immunofluorescence (green) in hippocampal CA1 and CA3 regions. Scale bar, 50 µm. (D) Quantification of the number of CCK-positive neurons in the CA1 and CA3 regions of the dorsal hippocampus. (E) Representative image of co-immunofluorescent labeling of CCK-positive neurons with the excitatory neuronal marker CaMKIIα in the CA3 region. Scale bar, 100 µm. (F) Fluorescence image showing colocalization of endogenous CCK protein (Pro-CCK) with an excitatory neuronal marker in the CA3 region. Scale bar, 1000 µm. (G) Schematic diagram illustrating the viral injection strategy. A Cre-dependent AAV (AAV-EF1α-DIO-EYFP; 6.5 × 10¹² vg/mL, 50 nL) was injected into the CA1 region of CCK-Cre mice to label CCK-expressing neurons and their projections. (H) Representative fluorescence images showing AAV expression in the hippocampal CA1 region (left) and retrogradely labeled neurons in the CA3 region (right). Scale bars: 1000 µm (left), 100 µm (right). (I) Representative image showing colocalization of retrogradely labeled GFP-positive neurons with the excitatory neuronal marker CaMKIIα (rabbit anti-CaMKIIα, red) in the CA3 region. Scale bar, 100 µm. (J) Co-immunofluorescent staining demonstrating colocalization of GFP with CaMKIIα in CA3 neurons of CCK-Cre mice, confirming the excitatory identity of retrogradely labeled CCK-expressing neurons. *p < 0.05, **p < 0.01, ***p < 0.001; ns not significant. Data are reported as mean ± SEM.

Excitatory CA3 neurons secret the neuropeptide CCK.

(A) Schematic diagram illustrating the viral injection strategy. AAV9-DIO-CaMKIIα-mCherry (5.0 × 10¹² vg/mL, 250 nL) was injected into the CA3 region of CCK-Cre mice to selectively label excitatory CCK-expressing neurons. (B) Representative fluorescence image showing viral expression in the hippocampal CA3 region and labeled projections to the CA1 area. Scale bar, 1000 µm. (C) Higher-magnification image of the boxed region in (B), highlighting labeled CA3 axons projecting to CA1. Scale bar, 50 µm. (D) Schematic illustrating the experimental configuration for hippocampal slice electrophysiological recordings. (E) Representative light-evoked field excitatory postsynaptic potentials (L-fEPSPs) recorded from CA1 following optical stimulation of CA3 CCK-positive Schaffer collateral (SC) projections, shown before and after application of the glutamatergic receptor antagonists CNQX and APV. Scale bars, 0.2 mV and 10 ms. (F) Quantitative analysis of glutamatergic synaptic transmission mediated by CA3 CCK-positive projections to CA1, summarized across recordings. (G) Schematic illustration of the CCK sensor principle. Binding of the CCK ligand to the genetically encoded sensor induces a conformational change that results in altered fluorescence intensity. (H) Experimental schematic showing co-injection of AAV9-hSyn-CCK sensor 2.3 (5.75 × 10¹² vg/mL, 200 nL) and AAV9-CaMKIIα-DIO-ChrimsonR-mCherry (5.0 × 10¹² vg/mL, 250 nL) into the CA1 and CA3 regions, respectively, followed by optogenetic stimulation and fiber photometry recording. (I) Representative fluorescence image showing expression of the CCK sensor in the CA1 region surrounding the optical fiber tip (left; scale bar, 1000 µm) and ChrimsonR-expressing CA3-CA1 projections in CCK-Cre mice (right). (J) Higher-magnification images illustrating the optical stimulation site in CA3 and the photometry recording site in CA1. (K) Schematic model depicting activity-dependent release of CCK from CA3 presynaptic terminals onto CA1 neurons. (L) Heatmap representation of calcium-dependent fluorescence responses following optogenetic long theta-burst stimulation (L-TBS) in control GFP-expressing mice (upper; N = 3 animals, n = 6 trials) and CCK sensor-expressing mice (lower; N = 3 animals, n = 6 trials). (M) Averaged ΔF/F fluorescence responses evoked by optogenetic stimulation (635 nm L-TBS) in GFP and CCK sensor groups (N = 3 animals per group, n = 6 trials each). (N) Quantification of the mean fluorescence response in CCK-Cre mice, calculated as the averaged ΔF/F within 3 s following L-TBS. *p < 0.05, **p < 0.01, ***p < 0.001; ns not significant. Data are reported as mean ± SEM.

CA3CCK neurons fire actively during hippocampal-dependent tasks.

(A) Schematic overview of the behavioral paradigms used in this study, including the novel object location (NOL) task and the Morris water maze (MWM) task. (B) Representative fluorescence images showing viral injection into the CA3 region and expression of GCaMP6s in CCK-positive neurons of CCK-Cre mice following injection of AAV9-CaMKIIα-DIO-GCaMP6s (5.0 × 10¹² vg/mL, 300 nL). Scale bars, 1000 µm (left) and 100 µm (right). (C) Behavioral performance during the NOL task showing that mice spent significantly more time interacting with the object placed in a novel location compared with the familiar location. (D) Heatmap representation of ΔF/F fluorescence traces from a representative mouse during the training and testing phases of the NOL task (N = 6 mice). (E) Averaged ΔF/F fluorescence traces from all mice (N = 6), aligned to the onset of object exploration. (F) Mean GCaMP6s fluorescence signals of CA3 CCK-expressing neurons during object exploration bouts in training and testing trials, quantified as the average ΔF/F within 1.5 s following exploration onset. (G) Schematic illustration of the experimental setup for calcium imaging during the Morris water maze task. (H) Representative plot showing the relationship between normalized ΔF/F signals (black) and escape latency (red) across training trials in the MWM task. (I) Heatmap and mean GCaMP6s fluorescence signals during the initial learning phase (trial 1) of the MWM task. (J) Heatmap and mean GCaMP6s fluorescence signals during the well-trained phase (trial 9) of the MWM task. (K) Summary quantification of ΔF/F signals comparing trial 1 and trial 9, calculated as the average fluorescence within 10 s following trial onset. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; ns, not significant. Data are reported as mean ± SEM.

Chemogenetic inhibition of the excitatory CA3CCK-CA1 pathway impairs behavioral tasks.

(A) Schematic illustrating Cre-dependent viral labeling and chemogenetic manipulation of CA3–CA1 projections in CCK-Cre mice. AAV9-CaMKIIα-DIO-hM4D(Gi)-mCherry (6.5 × 10¹² vg/mL, 300 nL) or control AAV9-hSyn-DIO-mCherry (6.5 × 10¹² vg/mL, 300 nL) was injected to selectively target excitatory CCK-expressing neurons. (B) Representative fluorescence image showing viral expression in the CA1 region with the corresponding cannula track. A higher-magnification image of the boxed area is shown on the right. Scale bars, 1000 µm (left) and 100 µm (right). (C) Schematic of the Morris water maze (MWM) task. The hidden platform was located in the southwestern (SW) quadrant (quadrant 3). (D) Quantification of swimming speed during the visible platform task, showing no significant differences between control and hM4D(Gi)-expressing mice. (E) Quantification of latency to locate the visible platform, indicating comparable sensorimotor performance between groups. (F) Escape latency across training days during the hidden platform phase of the MWM, comparing control and hM4D(Gi) groups. (G) Representative swimming trajectories of control and hM4D(Gi)-expressing mice during the spatial probe trial. (H) Percentage of total time spent in each quadrant during the probe test, showing reduced preference for the target quadrant in hM4D(Gi)-expressing mice. (I) Schematic of the novel object location (NOL) task (see Methods for details). (J) Total object exploration time during the NOL task, showing no significant difference between groups. (K) Control mice expressing mCherry exhibited a stronger preference for the object in the novel location compared with hM4D(Gi)-expressing mice. (L) Quantification of discrimination index demonstrating a significant reduction in spatial discrimination performance in hM4D(Gi)-expressing mice compared with controls. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; ns, not significant. Data are reported as mean ± SEM.

Chemogenetic inhibition of the excitatory CA3CCK-CA1 pathway impairs LTP formation.

(A) Schematic illustrating the viral injection strategy. AAV9-CaMKIIα-DIO-hM4D(Gi)-mCherry (5.0 × 10¹² vg/mL, 300 nL) or control AAV9-hSyn-DIO-mCherry (5.0 × 10¹² vg/mL, 300 nL) was injected into the CA3 region of CCK-Cre mice to selectively target excitatory CCK-expressing neurons. (B) Representative fluorescence images showing viral expression in the hippocampal CA3 region and labeled projections to the CA1 area. A higher-magnification image of the boxed region is shown on the right. Scale bars, 1000 µm (left) and 100 µm (right). (C) Schematic illustration of the hippocampal slice electrophysiological recording configuration. (D) Representative electrically evoked field excitatory postsynaptic potentials (E-fEPSPs) recorded in CA1 following stimulation of CA3 Schaffer collateral inputs before and after application of clozapine-N-oxide (CNO). Chemogenetic inhibition significantly reduced E-fEPSP amplitude in slices expressing hM4D(Gi) compared with mCherry-expressing controls. Scale bars, 0.2 mV and 10 ms. (E) Quantification of E-fEPSP responses showing reduced synaptic transmission in hM4D(Gi)-expressing slices relative to control slices following CNO application. (F) Schematic of the electrical theta-burst stimulation (E-TBS) protocol used to induce long-term potentiation (LTP). (G) Representative traces showing that LTP induction was attenuated in hM4D(Gi)-expressing slices compared with mCherry controls. Scale bars, 0.2 mV and 10 ms. (H) Summary quantification of E-fEPSP responses following E-TBS, demonstrating significantly reduced LTP in the hM4D(Gi) group relative to controls. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; ns, not significant. Data are reported as mean ± SEM.

RNA interference of excitatory CA3CCK expression attenuates hippocampal functions.

Schematic illustrating the viral strategy for CCK knockdown in CA3 CCK-expressing neurons. CCK-Cre mice received injections of AAV9-CaMKIIα-DIO-(mCherry-bGH pA-U6)-shRNA-CCK or control AAV9-CaMKIIα-DIO-(mCherry-bGH pA-U6)-shRNA-Scramble (5.0 × 10¹² vg/mL, 300 nL) into the CA3 region. (B) Representative fluorescence images showing viral expression in the hippocampal CA3 region and labeled projections to the CA1 area. A higher-magnification image of the boxed region is shown on the right. Scale bars, 1000 µm (left) and 100 µm (right). (C) Quantification of CCK mRNA expression in the CA3 region of CCK-Cre mice infected with anti-CCK shRNA or scramble control shRNA, confirming efficient knockdown of CCK expression. (D) Schematic illustration of the hippocampal slice electrophysiological recording configuration. (E) Quantification of swimming speed during the visible platform task of the Morris water maze (MWM), showing no significant difference between anti-CCK and anti-scramble groups. (F) Quantification of latency to locate the visible platform, indicating intact sensorimotor performance in both groups. (G) Escape latency during the hidden platform training phase of the MWM for anti-CCK and anti-scramble groups. (H) Representative swimming trajectories of anti-CCK and anti-scramble mice during the spatial probe trial of the MWM. (I) Percentage of total time spent in each quadrant during the probe test, showing reduced preference for the target quadrant in anti-CCK mice compared with controls. (J) Schematic of the novel object location (NOL) task (see Methods for details). (K) Total object exploration time during the NOL task, showing no significant difference between anti-CCK and anti-scramble groups. (L) Control mice expressing scramble shRNA exhibited a greater preference for the object in the novel location compared with anti-CCK mice. (M) Quantification of discrimination index demonstrating significantly reduced spatial discrimination performance in anti-CCK mice relative to scramble controls. (N) Representative traces showing that long-term potentiation (LTP) induced by electrical stimulation was attenuated in hippocampal slices expressing anti-CCK shRNA compared with scramble controls. Scale bars, 0.2 mV and 10 ms. (O) Summary quantification of electrically evoked field excitatory postsynaptic potentials (E-fEPSPs) following LTP induction, demonstrating reduced synaptic potentiation in the anti-CCK group relative to controls. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; ns, not significant. Data are reported as mean ± SEM.

Summary model illustrating the role of the perforant pathway in regulating spatial memory and hippocampal synaptic plasticity.

Excitatory CA3 neurons secret the neuropeptide CCK.

(A) Representative fluorescent images show that colocalization of the CaMKIIα and CCK+ neurons. Scale bar = 50 µm. (B) Quantitative analysis shows the ratio between CCK+ neurons and CaMKIIα in the CA3 area.

CA3CCK neurons fire actively during hippocampal-dependent tasks.

(A) Representative fluorescence images showing colocalization of CCK-positive neurons with the excitatory neuronal marker CaMKIIα in the CA3 region. Scale bar = 50 µm. (B) Quantification of the proportion of CCK-positive neurons that co-express CaMKIIα in the CA3 area.

Chemogenetic inhibition of the excitatory CA3CCK-CA1 pathway impairs behavioral tasks.

(A) Representative fluorescent images show that colocalization of the CaMKIIα and CCK neurons. Scale bar = 50 µm. (B) Quantitative analysis shows the ratio between CCK neurons and CaMKIIα in the CA3 area.

RNA interference of excitatory CA3CCKexpression attenuates hippocampal functions.

(A) Representative fluorescent images show that CCK-shRNA (left panel) significantly reduced CCK expression in CA3 -positive neurons compared with the CCK-Scramble group (right panel). Scale bar = 50 µm. (B) Quantitative analysis shows the number of CCK neurons, defined by colocalization of CCK and Pro-CCK, in the CCK-shRNA and CCK-Scramble groups (n = 9 slices for each group). ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; ns, not significant. Data are reported as mean ± SEM.