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

(A) Schematic flowchart of mating strategy to obtain CCK Cre/Ai14 mice. (B) Fluorescence image of the dorsal hippocampus slice of CCK Cre/Ai14 stained with neuron marker (NeuN, rabbit anti-488). Scale bar: 1000 μm. (C) CCK positive neurons were double labelled by tdTomato (Red) and rabbit anti-488 (GFP) in the area CA1 and CA3. Scale bar: 50 μm. (D) Quantification of CCK+ neurons in the CA1 area and CA3 area. (E) Example of co-immunofluorescent staining of CCK positive neurons with excitatory neuronal marker (CamKIIα) in CA3 area. Scale bar: 100 μm. (F) Fluorescence image of the colocalization between CCK protein (Pro-CCK) and excitatory marker in CA3 area. Scale bar: 1000 μm. (G) Schematic diagram of virus injection. AAV-EF1α-DIO-EYFP (6.5 E+12 vg/ml, 50 nl) was injected into CA1 area of CCK Cre mice. (H) Representative image of AAV injection in the hippocampal CA1 (left, scale bar: 1000 µM). Detailed view of retrogradely labeled neurons in area CA3 (right, scale bar: 100 μm). (I) Representative image showing the colocalization between the Cre-dependent retro AAV (gfp) and CamKIIα (rabbit anti-589: red). Scale bar: 100 μm. (J) Co-immunofluorescent staining of GFP with the excitatory neuronal marker CamKIIα in area CA3 of CCK Cre mice.

Excitatory CA3 neurons secret the neuropeptide CCK.

(A) Schematic diagram of virus injection. AAV9-DIO-CamKIIα-mCherry (5.0 E+12 vg/ml, 250 nl) was injected into CA3 area of CCK Cre mice. (B) Viral expression in the hippocampal CA3 area and its projections in the CA1 area, scale bar: 1000 µm. (C) A magnified view of the square area (1; scale bar, 50 µm). (D) Schematic drawing of the brain slice recording in the hippocampus. (E) Representative L-fEPSP of CCK+ SC projections by light stimulation before and after application of CNQX+ APV. Scale, 0.2 mV by 10 msec. (F) Quantitative analysis of the glutamatergic transmission in CA3CCK-CA1 projections. (G) Schematic showing the neurological principle of the CCK sensor. CCK ligand binds the sensor and induces a change in fluorescence of CCK sensor. (H) Schematic diagram showing injection of AAV9-hSyn-CCK sensor 2.3 (5.75E + 12 vg/ ml, 200 nl) and AAV9-CamKIIα-DIO-ChrimsonR-mCherry (5.0 E + 12 vg/ ml, 250 nl) into the CA3 and CA1 area. followed by laser stimulation and photometry recording. (I) CCK sensor expression in the CA1 area around fiber tip, scale bar: 1000 µm, and ChrimsonR expressing CA3-CA1 projectionsof CCK-Cre mice. (J) Magnified images showing the stimulation site and recording site. (K) Hypothetical model depicting the neuronal activity-dependent CCK release from CA3 pre-synapse in the CA1 area. (L) The heatmap results showing the Ca2+ response in each mice after the L-TBS in GFP group (upper; N= 3, n = 6 trials) and CCK-sensor group (lower; N= 3, n = 6 trials). (M) Averaged fluorescence increases in response to optogenetic stimulation (635 nm L-TBS; N = 3 animals, n = 6 trials for each group). (N) Quantification of averaged fluorescence in the CCK-Cre mice (averaged the Δ F/F after L-TBS within 3 s). *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) Schema of the novel location task and Morris water maze task. (B) Viral injection location in the CA3 area, and labeled CCK positive neurons in the CA3 of the CCK Cre mice. AAV9-CaMKIIa-DIO-GCaMP6s (5.00E+12 vg/ ml, 300 nl); Scale bar: 1000 μm and 100 μm. (C) The mice displayed significantly more interactions with the novel placed object than with the familiar object. (D) Heatmap shows the Δ F/F average traces from a single subject in the training and testing phase. N= 6 mice. (E) The Δ F/F average traces from all subject animals (N = 6), aligned to the time of object interaction. (F) Mean GCaMP6s signals of the CA3CCK neurons during bouts of object exploration in training and testing trials (average 1.5 seconds data after the indicated events). (G) Schematic drawing depicts the experimental setup for Ca2+ recording during the MWM task. (H) Plot of the Δ F/F (black) and their corresponding escape latencies (red) during the MWM task. (I) Heatmap and mean GCaMP6s signals of the original-trained state (trial 1). (J) Heatmap and mean GCaMP6s signals of the well-trained state (trial 9). (K) Summary of the Δ F/F between the trial 1 and trial 9 (average 10 seconds data after the indicated events). ∗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 of Cre-dependent color-switch labeling in the CA3-CA1 projections of CCK-Cre mice (AAV9-CaMKIIa-DIO-hM4D(Gi)-mCherry, 6.50 + 12 vg/mL, 300 nl; AAV9-hSyn-DIO-mCherry, 6.50 + 12 vg/mL; 300 nl). (B) Viral expression with the cannula track in the area CA1, scale bar, 1,000 μm. Magnified image is shown on the right (scale bar, 100 μm). (C) Schematic of the Morris water maze (MWM) task. The hidden platform is placed in the southwestern (SW) site (quadrant 3). (D) Quantitative data analysis for swimming speed of two groups of mice in locating the platform during the visible platform task. (E) Quantitative data analysis for swimming time of two groups of mice in locating the platform during the visible platform task. (F) Escape latency of two groups of mice during training phases of the MWM. (G) Swimming traces of two groups of mice in the spatial probe trial of memory retention test. (H) The proportion of total time of two groups of mice spent in each quadrant. (I) Protocol of the novel object location task. Details is depicted in the method section. (J) The total exploration time for either object shows no significant statistical difference. (K) Controls (Only expressing mCherry) exhibited obviously more interactions with the novel placed object than with the familiar object compared to the HM4D(Gi) group. (L) HM4D(Gi) group scored significantly less on the discrimination ratio compared to 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 diagram of virus injection. AAV9-CaMKIIa-DIO-hM4D(Gi)-mCherry (5.0 E+12 vg/ml, 300 nl) or AAV9-Syn-DIO-mCherry (5.0 E+12 vg/ml, 300 nl) was injected into CA3 area of CCK Cre mice. (B) Left: viral expression in the hippocampal CA3 area and its projections in the CA1 area, scale bar: 1000 µM; Right: a magnified view of the square area (1; scale bar, 100 μm). (C) Schematic drawing of the brain slice recording in the hippocampus. (D) CNO infusion caused a remarkable decline in E-fEPSP from the HM4D(Gi) expressing slices compared to controls (mCherry). Scale, 0.2 mV by 10 msec. (E) Quantitative data analysis for the E-fEPSP between the experimental group and controls. (F) The protocol of electrical theta burst stimulation (E-TBS) for eliciting LTP. (G) LTP was attenuated in the HM4D(Gi) expressing slices compared to controls (mCherry). Scale, 0.2 mV by 10 ms. (H) Quantitative data analysis for the E-fEPSP between the experimental group and 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.

(A) Schematic diagram of virus injection. AAV9-CaMKIIa-DIO-(mCherry-bGH pA-U6)-shRNA-CCK (5.0 E+12 vg/ml, 300 nl) or AAV9-CaMKIIa-DIO-(mCherry-bGH pA-U6)-shRNA-Scramble (5.0 E+12 vg/ml, 300 nl) was injected into area CA3 of CCK Cre mice. (B) Left: viral expression in the hippocampal CA3 and its projections in the CA1 area, scale bar: 1000 µM; Right: a magnified view of the square area (1; scale bar, 100 μm). (C) Quantification of the CCK mRNA expression in the CA3 of CCK Cre mice that infected with anti-CCK or anti-scramble shRNA. (D) Schematic drawing of the brain slice recording in the hippocampus. (E) Quantitative data analysis for swimming speed of two groups of mice in locating the platform during the visible platform task. (F) Quantitative data analysis for swimming time of two groups of mice in locating the platform during the visible platform task. (G) Escape latency of two groups of mice during training phases of the MWM. (H) Swimming traces of two groups of mice in the spatial probe trial of memory retention test. (I) The proportion of total time of two groups of mice spent in each quadrant. (J) Protocol of the novel object location task. Details is depicted in the method section. (K) The total exploration time for either object shows no significant statistical difference. (L) Controls (anti-scramble) exhibited obviously more interactions with the novel placed object than with the familiar object compared to the anti-CCK group. (M) Anti-CCK group scored significantly less on the discrimination ratio compared to the anti-scramble group. (N) LTP was attenuated in the anti-CCK expressing slices compared to controls (anti-scramble). Scale, 0.2 mV by 10 ms. (O) Quantitative data analysis for the E-fEPSP between the experimental group and controls. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; ns, not significant. Data are reported as mean ± SEM.

Summary of the neural effects