Cholecystokinin B receptor antagonists for the treatment of depression via blocking long-term potentiation in the basolateral amygdala

Depression is a common and severe mental disorder. Evidence suggested a substantial causal relationship between stressful life events and the onset of episodes of major depression. However, the stress-induced pathogenesis of depression and the related neural circuitry is poorly understood. Here, we investigated how cholecystokinin (CCK) and CCKBR in the basolateral amygdala (BLA) are implicated in stress-mediated depressive-like behavior. The BLA mediates emotional memories, and long-term potentiation (LTP) is widely considered a trace of memory. We identified that the cholecystokinin knockout (CCK-KO) mice impaired LTP in the BLA, while the application of CCK4 induced LTP after low-frequency stimulation (LFS). The entorhinal cortex (EC) CCK neurons project to the BLA and optogenetic activation of EC CCK afferents to BLA-promoted stress susceptibility through the release of CCK. We demonstrated that EC CCK neurons innervate CCKBR cells in the BLA and CCK-B receptor knockout (CCKBR-KO) mice impaired LTP in the BLA. Moreover, the CCKBR antagonists also blocked high-frequency stimulation (HFS) induced LTP formation in the BLA. Notably, CCKBR antagonists infusion into the BLA displayed an antidepressant-like effect in the chronic social defeat stress model. Together, these results indicate that CCKBR could be a potential target to treat depression.


INTRODUCTION
Depression is a leading cause of disability and a major contributor to the overall global burden of disease (WHO 2021), affecting more than 280 million people worldwide (GHDx).According to WHO, there is a 25% increase in depression cases during the Covid-19 pandemic [1].Considering the increasing incidence of depression and delayed onset of action of current antidepressants [2][3][4], developing new antidepressants with a specific novel target and with no or minimal side effects becomes increasingly important for society.
Considerable evidence suggests that stressful events or experiences may increase the risk of depression [5,6].Kendler indicated the substantial causal relationship between stressful life events and the onset of episodes of major depression [7].Stress leads to neural plasticity and synaptic markers changes in the brain's valence-coding systems that are strongly associated with depression, post-traumatic stress disorders (PTSD), and anxiety disorders [8][9][10][11][12].Amygdala deep nuclei in the brain perform a pivotal role in memory encoding and modulation [13], and the basolateral amygdala (BLA) is regarded as prominent among several valence-coding brain regions [14][15][16].Several studies revealed that depression is correlated with the hyperactivated amygdala [17][18][19].However, amygdala neurocircuitry mediating depression is poorly understood.
The entorhinal cortex (EC) is one of the major input structures to BLA [20,21].EC has been implicated in spatial and contextual fear memory regulation in animals [22][23][24].Recent studies show that EC projection to the visual cortex regulates depressive-like behavior [25], and EC projections enhance the neurogenesis in the dentate gyrus, which modulates depressive-like behavior [26].EC is enriched with cholecystokinin (CCK)-positive neurons [27], and EC CCK neurons send dense projections to BLA [20].
Cholecystokinin (CCK) has long been recognized as a neuropeptide involved in depression [28][29][30][31].A clinical study of 105 patients suffering from major depression showed that patients who had made one or more suicide attempts tended to have higher cerebrospinal fluid (CSF) CCK levels [29].Currently, there are two known CCK receptors in the central nervous system: CCK A receptor (CCKAR) and CCK B receptor (CCKBR).CCKAR is expressed in the brain region e.g., the hippocampus, and certain midbrain regions while CCKBR is abundantly expressed in the neocortex, amygdala, and sparely the hippocampus [20,[32][33][34].Previous studies from our group unveiled that CCK and CCKBR mediate neuroplasticity as well as visual and sound associative memory formation in the auditory cortex [35][36][37], and trace fear memory formation in the amygdala [20].Other studies indicated that CCKBR antagonist L-365,260 exhibited antidepressant effects in forced swim tests of mice [28], and CCKBR antagonist CI-988 prevented chronic social defeat-induced depressive behaviors and increased cortical CCK release in the rat model [30].Although these studies showed that CCK and CCKBR are implicated in depression, the neural circuit mechanism of CCK-dependent pathogenesis of depression and the CCKBR antagonist-mediated antidepressant effect remain unclear.We hypothesized that CCK might facilitate aversive memory formation by enabling LTP in the BLA, enhancing the development of depression.CCKBR antagonists would suppress the formation and consolidation of aversive memories, leading to the treatment of depression by blocking LTP in the amygdala.
In the present study, we tested the above hypotheses by using in vitro recording, optogenetic manipulation, histology, viral gene knockout, drug manipulation, fiber photometry recording, and behavioral analysis.First, we examined whether CCK could switch LTP formation in the BLA via in vitro slice recording.Then it was examined whether optogenetic activation of CCK projections from EC-BLA could promote stress susceptibility.Further, we confirmed this effect is mediated by CCK by detecting CCK release in the BLA, and knockdown of CCK in the EC-BLA circuit.Next, we investigated the effects of CCKBR antagonists on neural plasticity in the BLA.Finally, the antidepressant effects of CCKBR antagonists were tested on the mouse model via intraperitoneal and local infusion into BLA.

MATERIAL AND METHODS Animals
All experimental procedures were reviewed and approved by the Animal Subjects Ethics Sub-Committees of the City University of Hong Kong.The following transgenic mice were used: CCK-ires-Cre (Cck tm1.1(Cre)Zjh /J, C57BL/6 J background, stock # 012706, Jackson Laboratory), CCKBRKO (Cckbr tm1Kpn /J, 129S1 background, stock # 006369, Jackson Laboratory), and CCK-CreER (Cck tm2.1(Cre/ERT2)Zjh /J, C57BL/6 J background, stock # 012710, Jackson Laboratory), the corresponding wild-type littermates background mice were also used.Male retired breeder ICR mice were used for physical defeat during CSDS.C57BL/6 J male adult mice (7 weeks of age) were used to establish the CSDS model of depression from the same vendor, the Chinese University of Hong Kong.We used mice in our current study and all mice were randomly allocated into the experimental or control group.All behavioral tests were carried out during the dark phase.Mice were housed in a 12-hour light/ 12-hour dark cycle (dark from 9:00 to 21:00) and given food and water ad libitum.

In vitro electrophysiology
ACSF was prepared with the following protocol: 124 mM NaCl, 2.5 mM KCl, 1 mM NaH2PO4, 10 mM D glucose, 25 mM NaHCO3, 2 mM CaCl2, and 1 mM MgCl2, at pH 7. 35-7.45.Before use, ACSF was well-oxygenated (95% O2/5% CO2, v/v).For brain slices preparation, animals were deeply anesthetized with isoflurane.Coronal sections (thickness, 300 μm) across the objective brain region, the amygdala, were cut in well-oxygenated icecold ACSF using a vibratome.The brain slices were allowed to recover for 1.5 h at 30 ± 1 °C in oxygenated ACSF.After recovery, one slice was positioned on the microelectrode array system probe (MED-PG515A) on the stage of the microscope.Once the slice was settled, a fine mesh and anchor were carefully overlaid on the slices to ensure stabilization during recording.Then, the slices were continuously perfused with oxygenated 30 ± 1 °C ACSF with the aid of a peristaltic pump during electrophysiological recording.The 4*4 channel microelectrode array was located on the targeted brain region under a light microscope.One microelectrode was selected for stimulation, referring to the microscope photo.For the selection of the best stimulation site, monopolar, biphasic constant-current pulses (0.2 ms in duration) were generated by Mobius software with a 2 s interval.The fEPSPs were evoked at several of the remaining sites.The fEPSPs were amplified by an amplifier and displayed on the monitor screen.Microelectrodes were screened, and the best stimulation site was selected.After 30 min of recovery of the slices with stabilization of the baseline fEPSP responses, an input-output curve was determined by using the measurements of the fEPSPs amplitude (output) in response to a series of ascending stimulation intensities in 10 mA steps (input).For LTP induction, 30-50% saturated intensity was selected as the stimulation intensity for the baseline synaptic response recording.fEPSPs responded stably for at least 15 min set as a baseline.Then, a TBS protocol (4 sets with 10 s intervals, each set consisting of 5 trains at 5 Hz, 4 pulses at 100 Hz as one train) was given at the stimulation site.The stimulation intensity of TBS was adjusted to elicit 75% of the saturated intensity to introduce LTP.For the drug effects test, after baseline recording, perfused ACSF diluted with CCKBR antagonist (YM022 or YF476) for 10 min before TBS; or perfused ACSF diluted with CCKBR agonist (CCK4) for 5 min with LFS (75% of the saturated intensity,1HZ, 100 pulses), then fresh ACSF washed out drug after TBS/LFS.fEPSP responses were recorded for at least 60 min after TBS/ LFS.Changes in the fEPSP amplitudes were analyzed as a percentage change from the baseline.The average normalized amplitudes of the first 10 min before and the last 10 min after TBS/LFS were compared.For comparison of the LTP magnitude between groups, the averaged values of the last 10 min were compared statistically.The slices with unstable baselines were excluded from the study.

Immunochemistry
Mice were deeply anesthetized with pentobarbital (100 mg/kg, i.p. injection) and perfused transcardially with 20 ml of phosphate-buffered saline (PBS) followed by 20 ml of 4% paraformaldehyde (PFA) in 0.1 MPBS, pH 7.4.Brains were removed and postfixed overnight in 4% PFA at 4 °C in 0.1 M PBS.Coronal sections (thickness; 50 μm) were cut in 0.1 M PBS using a vibratome and stored at −20 °C in an antifreeze solution (50% PBS, 20% ethylene glycol, 30% glycerin).Targeted slices across the amygdala (−1.70 mm relative to the bregma) were selected under a light microscope and washed three times with PBS.The slices were incubated in a blocking solution for 1.5 hours at room temperature.Blocking solution prepared with the following: 5% normal goat serum, 0.2% Triton X-100, with PBS.Then, the slices were shaken with primary antibodies in a blocking solution overnight at 4 °C.The primary antibodies used were c-Fos (1:1000, ab208942) and CCKBR (1:1000, PA3-201).After three washes with PBS, sections were incubated with Alexa Fluor 594 (1:500), and Alexa Fluor 488 (1:1000) secondary antibodies at room temperature for three hours.After another three 7 min washes in PBS, immunofluorescence was assessed using a Ni-E upright fluorescence microscope (Nikon).The percentage of c-Fos/CCKBR colocalized neurons in each group was calculated as the percentage of the total number of c-Fos/CCKBR colocalized neurons counted within the total number of c-Fos expressing neurons counted.

Western blotting
The mice were decapitated, and their brains were rapidly isolated on the ice.The BLA tissue was collected in a tube containing a mixture of RIPA and protein inhibitors.Tissue was homogenized and after centrifuge supernatant was collected for protein quantification.The total protein in the samples was quantified using the BCA method.An equal amount of protein was loaded on 8% SDS-polyacrylamide gel and transferred to the PVDF membrane.Membranes were incubated with primary antibodies AMPA (Cat # ab31232; 1:1000), NMDA (cat # A13863; 1:1000), CAMKII (cat # ab52476; 1:1000), and CCK (cat # ab27441; 1:1000) for respective samples.After overnight incubation at 4 C, the membrane was washed with TBST 3 times 5 min each wash and then incubated with secondary antibody (cat # ab6721; 1:2000) for 2 hours at room temperature.The blot was visualized using an ECL substrate with a BIO-RAD Chemidoc machine.

Fiber photometry
To detect invivo CCK release we adopt the fiber photometry system.Mice were injected with AAV-CAMKII-DIO-ChrimsonR-cherry into the EC (AP = −4.25 mm from bregma, ML = ± 3.85 mm, and DV = −2.55mm from dura matter) and AAC-hSyn-CCK2.3 into the BLA (Ap = −1.6;ML = ± 3.40; and DV = −4.00from dura matter), 4-week later the fibers were implanted over the BLA (Ap = −1.6;ML = ± 3.40; and DV = −3.85from dura matter) and EC(AP = −4.25 mm from bregma, ML = ± 3.85 mm, and DV = −2.25 mm from dura matter), and after 1-week fiber photometry recording was performed.The GPCR activation-based sensor was developed by inserting a circular-permuted green fluorescent protein into the intracellular domain of CCK-BR [38].When there is CCK release it binds to the CCKBR and emits a fluorescence signal which can be detectable by the fiber photometry system.Mice were connected to the fibers and after baseline recording, EC CCK neurons were stimulated with red light, and CCK release was detected in the BLA via the fiber.To detect the real-time neural activity of BLA in response to aversive stimuli (Footshock), We injected AAV-CAMKII-GCAMP6s into the BLA, and 3 weeks after virus expression fiber photometry was performed.Mice were connected to fiber and after habituating the chamber for two minutes footshock was delivered and calcium activity was recorded.The data were analyzed by pMAT an independent source software for fiber photometry data analysis [39].

Chronic social defeat stress (CSDS) model of depression
It was performed as described previously [40].Retired breeder male CD1 mice were screened on three consecutive days to validate their aggressive characteristics.During ten days of the chronic social defeat stress procedure for the experimental group, male intruder C57 mice were introduced into the resident CD1 mouse home cage.The novel aggressive CD1 mice physically attacked the C57 mice for 10 min.After that, using a perforated acrylic board partition, the resident cage was divided into two, wherein the residents and intruders could be maintained in sensory contact for 24 h.For 10 successive days, a C57 mouse was introduced to a novel aggressive CD1 mouse cage every day.After ten days of CSDS, the animals were housed singly, and social interaction behavior was tested 24 h later.For the sham group, two C57 mice were housed in one cage separated by a perforated acrylic partition.A novel partner was rotated every day, and physical contact with each other was avoided.Mice with severe physical injury are excluded from the study.

Two-trial subthreshold social defeat stress (SSDS) model
It was performed as described previously [41,42].Before the trial, retired resident breeder CD1 mice were screened for aggressive behavior.In a two-trial social defeat experiment, an intruder CCK-Cre mouse was placed into the home cage of an aggressive novel resident CD1 mouse for 10 min physical defeat.Then, with a perforated acrylic board partition dividing the resident home cage into two, residents and intruders maintained sensory contact for 10 min.After 10 min of the sensory stress phase, the CCK-Cre mouse was returned to its home cage for 5 min.Then, the second round of 10 min of physical defeat and 10 min sensory stress in the home cage of another novel CD1 mouse was carried out.After two defeat tails, the intruder CCK-Cre was returned to its home cage and accepted the social interaction test the following day.During the sensory stress phase, bilateral intra-BLA laser stimulation stimulated the EC-BLA circuit.The laser power was adjusted to 10 mW before each experiment.For ChETAmediated stimulation, the optical fibers were connected to a 473 nm blue laser diode.

Social interaction test (SI)
The social interaction behavior of mice that underwent social defeat was tested in an open-field arena (44 × 44 × 44 cm 3 ) containing an empty acrylic cage (9.5 × 9.5 × 8 cm 3 ).The social interaction test included no target and target phases.In the no-target stage, a mouse was placed in the center of the open field arena facing the empty acrylic cage for 2.5 min.The animals' track was analyzed with Smart.3 video-tracking software and the time spent in the social interaction zone surrounding the acrylic cage (a 6 cm region around) was measured.The animal was then returned to the home cage for 1 min.In the target stage, a novel, aggressive CD1 mouse was placed in the acrylic cage, and the same metric was measured.From these two stages, a social interaction ratio was calculated (the time spent in the interaction zone in the target phase/the time spent in the interaction zone in no target phase).

Sucrose preference test (SPT)
Animals that underwent CSDS or SSDS were singly housed and habituated with two 50 ml tubes of water for 1 week before social defeat stress and 1 day before SPT.Then, the experimental animals were given access to one tube of water and one tube of 1% sucrose solution for 24 hours and were weighed at 10:00, 15:00, and 18:00.The bottle position (left or right) was switched after each weight measurement to avoid a side preference.The sucrose preference ratio was calculated by dividing the total consumption of sucrose solution by the total consumption of both water and sucrose solution.

Tail suspension test (TST)
Mice were suspended by their tails with adhesive tape (1 cm from the tail tip) approximately 50 cm above the surface.Acrylic board partitioning was applied to ensure no contact.Plastic tubes were placed over the tail to ensure that the mice could neither climb nor hang on to their tails.The animals were video recorded from the side for 6 min, and the total immobility time was quantitated by Smart.3 video tracking software over the full 6 min post hoc.

Forced swimming test (FST)
Mice were individually placed into a cylinder of water (23 °C ± 1 °C, 16 cm diameter × 24 cm high) for 6 min.The water depth was set to prevent animals from touching the bottom with their tails or hind limbs.Animal behavior was video recorded from the side.The total immobility time of the last 4 min was quantitated by Smart.3 video-tracking.The immobility of the mice was defined as remaining floating or motionless, with only movements necessary to maintain balance in the water.

Open field test (OFT)
Mice were individually introduced to the central zone of an open field (50 cm length, 50 cm width, 40 cm height) with dim light for 5/10 min.Their exploration in the arena was recorded by a video camera and analyzed by Smart 3.0 software.Locomotor activity was evaluated as the total distance traveled.The anxiety-like behavior level was defined by the time spent in the center region (central 30*30 cm zone).The number of fecal boli deposited accounted for anxiety after the test.

Drug delivery
For the acute behavior test, the drug was administered by i.p. injection 30 min before the test.For drug delivery in the CSDS model, C57 mice were treated with 10 mg/kg fluoxetine or 3.0ug/kg YM022 once a day via i.p. injection for ten days.Control C57 mice received the vehicle of the YM022 injection.The injection site changed to left or right daily.

Pharmacokinetics
Materials and instrument: MS2 Turnover type oscillator was purchased from IKA Work's Guangzhou (China), and a 5415Rhigh speed tabletop centrifuge was purchased from Eppendorf AG (Germany).Chromatographic analyses were performed with an HPLC system consisting of an LC-10ADvp Pump (SHIMADZU, Japan), an MPS3C automatic sample (Gerstel Autosampler, Germany), and API3000 triple quadrupole tandem mass spectrometer (AB co., U.S.).Propranolol (≥90% in purity, internal standard, IS) was purchased from Sigma Chemical Co.(China).Formic acid (HPLC grade) and methanol (HPLC grade) were purchased from DIKMA Co. (China).All chemicals and solvents were analytical grades.And the water was purified using a Millipore (AK, USA) laboratory ultra-pure water system (0.2 μm filter).Animal: Male KM mice, weighing 32-36 g (Southern China Medical University, China) were utilized for the studies.The protocol was approved by the Animal Care and Use Committee, GIBH.Animals were maintained on standard animal chow and water ad libitum, in a climatecontrolled room (23 ± 1°C, 30-70% relative humidity, a minimum of 10 exchanges of room air per hour, and a 12-h light/dark cycle) for one week before experiments.Pharmacokinetic studies: The compound was dissolved in a solution containing 2% DMSO, 4% ethanol, 4% castor oil, and 90% DD H2O; Pharmacokinetic properties of mice (male) were determined following PO administration.Animals were randomly distributed into 8 experimental groups (n = 4).The mice were given 5 mg/kg

RESULTS
CCK signaling is critical for LTP induction in the BLA Stress-induced synaptic plastic changes including long-term potentiation (LTP) and long-term depression (LTD) are strongly associated with depressive disorders [43,44].These emotional events regulation is associated with neuronal activity or neuroplasticity in the amygdala [45,46], and the aversive memory can be inactivated and reactivated by LTD and LTP [47].Therefore, we first examined the dependency of CCK in LTP induction in the BLA.We adopted a brain-slice electrophysiological system with multiple slices and microelectrode arrays (Fig. 1A) to compare the LTP induction in the BLA between the CCK-KO (CCK-/-) mice and its wildtype control.CCK-KO mice showed an impaired TBS-induced LTP in the BLA, as compared to their wild-type controls (Fig. 1B-D ).Behavioral analysis showed that the CCKKO mice showed normal locomotion and decreased anxiety-like behavior as indicated by increased center time (%) (Fig. S1 A-D).However, there was no difference in immobile time between CCKKO mice and wild-type control during TST and FST (Fig. S1 E, F).We next adopted the LFS protocol with CCK4 and without CCK4 in wild-type C57 mice, to rule out the global CCK gene knockout effect and more reliable control conditions.It is well-known that LFS does not induce LTP in the field-potential recording setting.LFS induces long-term depression (LTD) when both the number of stimulation pulses (typically 900 pulses) and the current of stimulation are large [48].As previously reported, we concluded that HFS induced the release of CCK from the presynaptic terminals, leading to LTP in the neocortex and amygdala [20,35].We then hypothesize that the application of the CCKBR agonist (CCK4) could induce LTP in the BLA even under the LFS paradigm.Indeed, application of CCK4 with LFS (100 pulses) induced LTP in the BLA, whereas application of the vehicle did not induce LTP (Fig. 1H-J; first 10 min before vs. last 10 min after LFS in C57 mice with the vehicle, 99.4 ± 0.5% vs. 108.2± 4.6%; first 10 min before vs. last 10 min after LFS in C57 with CCK4, 99.4 ± 0.97% vs. 196.9.2 ± 34.6%).Both the loss-of-function and gain-of-function invitro electrophysiological studies supported the critical role of CCK signaling in LTP induction or neuroplasticity in the BLA.
We confirmed the injection site of the AAV virus in the EC and their projections in the BLA (Fig. 2A).The duration spent in the social interaction zone when a novel social target was present in the SI test showed no difference in the two groups (Fig. 2B, C; ChETA vs. EYFP with a target, 52.6 ± 13.0 s vs. 52.9± 8.0 s).However, the preference for sucrose solution of the ChETA group between cheta and EYFP group (Fig. S2 E-H).These results indicated that the activation of CCK projections from the EC to BLA facilitated stress susceptibility and induced anhedonia phenotype, a loss of capability to perceive pleasure, one of the most stringent parameters in depression.
It is previously known that the stimulation of EC CCK terminals in the auditory cortex mediates CCK release that facilitates associative memory formation via induction of LTP in the auditory cortex [35].Therefore, we hypothesized that the depressive effect may be mediated by CCK release from the EC CCK neurons into BLA.To monitor the dynamics of CCK, we adopted the G protein-coupled receptor (GPCR) activation-based CCK sensor [20,49], which is a circular-permutated green fluorescent protein inserted into the intracellular domain of CCKBR.Once the CCKBR binds with CCK, the increased fluorescence intensity can be measured by fiber photometry.The AAV-Syn-CCK-Sensor was injected into the BLA and AAV-CAMKII-DIO-ChrimsonR-mCherry was injected into the EC of CCK-CRE mice (Fig. 2E).The virus was well expressed six weeks after injection.We found that the laser activation of EC CCK neurons increased the average intensity of the fluorescent signal when compared with the baseline signal before stimulation (Fig. 2F-G; Before Laser vs After Laser, −0.2 ± 0.2 % vs 0.8 ± 0.3 %).Together, these results indicate the activation of the EC CCK-BLA pathway releases CCK into the BLA and mediates the depressive-like phenotype.
EC CCK -BLA circuit-specific knockdown of CCK impaired the optogenetic activation-induced stress susceptibility To further confirm that the CCK mediates the stress susceptibility in EC to BLA pathways.We adopted a candidate approach [50] to the circuit specifically the knockout of CCK.We selectively knocked out the CCK in the EC-BLA projection by injecting a retrogradely transported virus encoding the guide RNA into the BLA and another virus encoding Cas9 into the EC and assessed the effect of EC-BLA CCK-cKO on stress susceptibility (Fig. 3A).We verified that our retrograde CRISPR-Cas9 approach had high efficiency in targeting CCK expression in EC-BLA projection neurons but not glutamate marker level (Fig. S3 C-E).Indeed, EC-BLA CCK-cKO impaired stress susceptibility, which is indicated by increased time in the interaction zone (Fig. 3B-D; Control vs. CCK-cKO with a target, 19.4 ± 5.04 s vs. 52.1 ± 4.1 s), increased sucrose preference (Fig. 3E; Control vs CCK-cKO, 73.7 ± 2.02 vs. 83.3± 1.86), increased center time in OFT (Fig. 3F-H; Control vs. CCK-cKO, 9.8 ± 1.62 vs. 15.7 ± 1.92), and decreased immobile time in TST (Fig. S3B) in CCK-cKO mice compared with control group mice.Furthermore, we found that the EC CCK terminals innervate with BLA CCKBR cells (Fig. 3I).

CSDS activated CCKBR neurons at BLA
It is previously known that the CCKBRs are expressed intensively in the amygdala [20,34], and the amygdala showed hyperactivity when a person is sad or clinically depressed [18,19,51].We found that the footshock aversive stimuli increase the BLA neuronal activities as indicated by an increase in calcium response from 0.08 ± 0.28 to 7.6 ± 1.42 when the footshock was delivered (Fig. S4  A, B).Furthermore, we found that the CCKBR knockout mice (CCKBR-/-) showed reduced C-fos-CCKBR colocalized cells compared with wild-type control mice (Fig. S4 C, D).In the next experiment, we examined if a standard 10-day CSDS protocol [40] could induce the hyperactivated amygdala and whether the activated neurons correlate with CCKBRs.Mice underwent CSDS and after that SIT was performed to evaluate the social avoidance behavior as an indicator of depressive-like behavior in mice (Fig. 4A).The time spent in the social interaction zone in the presence of a social target of the mice who underwent the CSDS is significantly less than that of sham group mice (Fig. 4B, C; Sham vs CSDS with Target, 106.7 ± 7.7 vs 64.1 ± 11.1).We measured c-Fos and CCKBR expression in the BLA when the mice were re-exposed to novel CD1 mice.Notably, c-Fos expression in the BLA of CSDS mice was significantly higher than in the sham group mice (Fig. 4D, E; Sham vs CSDS C-fos per mm, 7.8 ± 2.6 vs 48.3 ± 5.2).Furthermore, c-Fos-positive neurons activated by CSDS were colocalized with CCKBR neurons, and the number of colocalized cells was higher in CSDS group mice compared with sham control group mice (Fig. 4F; Sham vs CSDS C-fos & CCKBR Cells (%), 31.8 ± 2.0 vs 91.5 ± s2.1).These results indicated that CSDS mediates social avoidance and increased the expression of CCKBR-c-fos positive neurons in the BLA.

CCKBR antagonist suppressed LTP induction in the BLA
It has been shown that the EC CCK terminals innervate with the CCKBR in the BLA.Next, we investigated the role of CCKBR in LTP induction in the BLA using in-vitro slice recording.First, we used transgenic mice lacking the CCKBR gene (CCKBR-/-) to explore whether knockout mice impaired the LTP (Fig. 5A).Results exhibited a deficit of TBS-induced LTP in the BLA of CCKBR-/mice, as compared to their wild-type (WT) control (Fig. 5B-D; first 10 min before vs. last 10 min after TBS CCKBR-/-mice, 100.0 ± 0.3% vs. 102.7 ± 2.4%; first 10 min before vs. last 10 min after TBS WT mice, 99.8 ± 0.5% vs. 135.4± 6.0%).CCKBR-/-mice behavior analysis exhibited decreased depressive and anxiety-like behavior as indicated by decreased immobile time during TST and FST tests (Fig. S5 I-K) and increased center time in the OFT when compared with wild-type control mice (Fig. S5 A, B).Besides, we didn't find any significant difference in time in the interaction zone (Fig. S5L) and sucrose preference (Fig. S5M) between control and CCKBR-/mice.However, CCKBR-/-mice showed increased distance traveled in the OFT test (Fig. S5C).To confirm that increased distance traveled is a result of decreased anxiety but not because of impaired motor behavior, we analyze the distance traveled in the home cage.In the home cage, both the control wild-type and CCKBR-/-mice showed the same total distance traveled (Fig. S5E, F).Furthermore, we changed the environment partially similar to the home cage by changing the open field floor with home cage bedding.CCKBR-/-and control mice showed similar total distance in the modified open field (Fig. S5G, H).Hence, CCKBR-/-mice showed less anxiety and depressive-like behavior without impaired motor behavior.It should be noted here that some researchers have interpreted high activity or increased exploratory behavior as an index of low emotionality [52][53][54][55] while others conceive exploratory behavior as being independent of emotionality [56].One has to acknowledge that differences in locomotor activity can confound emotional behavior [57].These results showed that the CCKBR-/-mice showed less anxiety and depressive-like behavior without impaired motor behavior.We next used CCKBR antagonists in wild-type C57 mice, to rule out the global CCKBR gene knockout effect and a reliable control group using the same C57 mice with the vehicle.A natural question was whether a CCKBR antagonist could impair the LTP induction in the BLA.YM022 is a selective CCKBR antagonist, with an IC50 for CCKBR 68 pM [58].Perfusion of YM022 (1 nM/10 nM) suppressed the LTP induction in the BLA of C57 mice, whereas perfusion of the vehicle did not (Fig. 5E-G; first 10 min before vs. last 10 min after TBS with the vehicle, 99.20 ± 0.3% vs. 172.3± 18.0%; first 10 min before vs. last 10 min after TBS with 1 nM of YM022, 99.6 ± 0.9% vs. 122.1 ± 2.8%; first 10 min before vs. last 10 min after TBS with 10 nM YM022, 99.6 ± 0.4% vs. 110.2± 5.6%).We further confirm the results using another selective CCKBR antagonist YF476 with an IC50 for CCKBR 0.1 nM [59], 10 nM of YF476 fully blocked TBS induced LTP in the BLA of C57 mice (Fig. S5N-Q; first 10 min before vs. last 10 min after TBS with 10 nM YF476, 101.0 ± 0.7% vs. 106.5 ± 9.0%; first 10 min before vs. last 10 min after TBS with the vehicle, 101.1 ± 0.8% vs. 187.5 ± 17.4%).Together, these results revealed that CCKBR is critical for the induction of LTP in the BLA.
CCKBR antagonists showed antidepressant effects in the acute TST and OFT According to Pardridge, around 98% of small molecules cannot penetrate the blood-brain barrier (BBB) and failed to show their response to central nervous system disorders [60].To investigate the antidepressant-like effect of CCKBR antagonist in the animal model, first, we examined whether the CCKBR antagonists YM022 and YF476 can penetrate the blood-brain barrier (BBB) because without penetrating the BBB a drug could not reach the target to show the antidepressant-like effects.After single oral administration of YM022 (5 mg/kg) and YF476 (5 mg/kg), the plasma and brain samples were obtained at several time points after dosing: 15, 30 min, and 1, 2, 4, 6, 8, 24 hours, and analyzed by LC-MS/MS analysis (Fig. S6A).The peak concentration of YM022 in brain tissue was 6.0 ± 1.8 ug/L at 30 min postdose; pharmacokinetic profiles of YM022 in the plasma showed that maximum concentration(Cmax) was 215.2 ug/L, time for Cmax to occur (Tmax) was 0.38 hour, halflife (T1/2) was 1.3 hours, and area under the plasma concentrationtime curve AUC (0-∞) was 321.8 ug/L*h (Fig. S6B, C).The peak concentration of YF476 in brain tissue was 32.7 ± 6.8 ug/L at 15 min post-dose; pharmacokinetic parameters of YF476 in plasma Cmax was 341.5 ug/L, Tmax was 0.38 hour, T1/2 was 0.5 hours, and AUC (0-∞) was 228.1 ug/L*h (Fig. S6 B, D).The results indicated that YM022 and YF476 can penetrate the BBB.Determination of the total immobility time (despair behavior) in the TST allows a quick assessment of potential antidepressant effects, various classes of antidepressant drugs decrease immobility time in the TST [61,62].Hence, we examined the effects of acute administration of CCKBR antagonists on the duration of immobility in the TST.We compared the effects of the CCKBR antagonists, YM022 and YF476, with the saline control and current antidepressant, fluoxetine (selective serotonin receptor inhibitors, SSRIs) as the positive control.The fluoxetine at a dose of 10 mg/kg significantly decreased the total immobility time in the TST; of four different concentrations of YM022 (0.3 μg/kg, 0.6 μg/kg, 3.0 μg/kg, and 30 μg/kg) that were tested, mice showed a considerably less immobility time with doses of 0.6 μg/kg and 3.0 μg/kg (Fig. S6 E,  F).Moreover, mice treated with YF476 also showed a significantly decreased immobility time with a dose of 1.3 μg/kg and 3.0 μg/kg (Fig. S6G).As the period of immobility is an indicator of hopelessness during the TST test, CCKBR antagonists, both YM022 and YF476, demonstrated a similar beneficial effect as compared to fluoxetine a positive control.Next, we adopted the OFT to evaluate the anxiolytic effect of YM022.We administrated intraperitoneally YM022 (3.0 ug/kg) 30 min before the OFT.We observed no difference in the total distance traveled between the vehicle and YM022 treated mice during the 10-min test time (Fig. S6K), but YM022 treated mice spent significantly more time in the center (Fig. S6H-J).These results revealed that CCKBR antagonist YM022 has anxiolytic effects with a dose of 3.0 ug/kg.

Antidepressant effects of CCKBR antagonists in the CSDS model
Next, we adopted a standard CSDS model to examine the CCKBR antagonist antidepressant effect.We administrated YM022 30 min before each physical defeat to prevent the development of depressive-like behavior in mice.We conducted SI and SPT respectively, 24 hours after the final social defeat (Fig. 6A).YM022 (3.0 μg/kg) and positive control fluoxetine (10 mg/kg) treated mice spent significantly more time at the social interaction zone with the social target than vehicle-treated mice (Fig. 6B, C; time in SI zone with the target, sham 79.8 ± 3.4 vs Veh 58.5 ± 5.6 vs YM022 79.3 ± 5.3 vs Fluoxetine 88.8 ± 6.2), and the social interaction ratio in the drug-treated group is also increased significantly (Fig. 6D; SI ratio, sham 1.6 ± 0.1 vs Veh 0.9 ± 0.1 vs YM022 1.5 ± 0.1 vs Fluoxetine 1.5 ± 0.2).Besides, it has been found that only the vehicle group mice spent significantly more time in corners than the sham, YM022, and fluoxetine-treated mice, to actively avoid the novel social target (Fig. 6E; Time in the corner with the target, sham 16.0 ± 1.6 vs Veh 35.0 ± 5.2 vs YM022 15.9 ± 2.9 vs Fluoxetine 16.9 ± 3.2).The corner ratio showed a significant difference accordingly (Fig. 6F; Corner time ratio, sham 0.7 ± 0.1 vs Veh 2.1 ± 0.2 vs YM022 0.6 ± 0.1 vs Fluoxetine 0.9 ± 0.2).Consistent with social interaction behavior results, sham, YM022, and fluoxetine groups consumed significantly more sucrose liquid than vehicle group mice (Fig. 5G; SPT, sham 78.9 ± 1.8 vs Veh 66.9 ± 1.7 vs YM022 74.5 ± 1.7 vs Fluoxetine 74.6 ± 1.6).These results demonstrated that CCKBR antagonist YM022 effectively prevented CSDS-induced social avoidance and anhedonia in mice.

Antagonizing BLA CCKBR mediates antidepressant-like effects
To confirm that CCKBR antagonist mediates antidepressant-like effect through BLA, we implanted a cannula over the BLA for local drug infusion, and 1-week after recovery mice underwent CSDS, CCKBR antagonist YM022 was locally infused before every physical defect (Fig. 7A).The YM022 treatment group showed significantly increased SI time in the presence of the target when compared with the control ACSF group (Fig. 7B, C; time in SI zone with the target, ACSF 50.4 ± 9.2 vs YM022 87.3 ± 5.5), and the social interaction ratio was also increased (Fig. 7C; SI ratio, ACSF 0.8 ± 0.2 vs YM022 1.7 ± 0.1).YM022 infusion into BLA increased the sucrose solution intake compared with the ACSF group (Fig. 7E; SPT, ACSF 61.8 ± 4.1 vs YM022 75.1 ± 2.0).Furthermore, we compared the immobile time in the TST, center zone time, and total distance traveled in the OFT.Mice with local BLA infusion of YM022 showed significantly reduced immobile time (Fig. 7F; ACSF 118.5 ± 12.5 vs YM022 78.0 ± 10.9), increased center time in OFT (Fig. 7G-H; ACSF 7.5 ± 1.0 vs YM022 12.3 ± 1.3), without any change in total distance traveled (Fig. 7I; ACSF 33.5 ± 1.9 vs YM022 38.7 ± 2.6).Lastly, we found that the CCKBR antagonist YM022 treatment decreased the C-fos activity in the BLA compared with ACSF-treated mice (Fig. S7A, B).

DISCUSSION
Despite the clinical [29,63], and preclinical studies [28,30,64] reporting the critical role of cholecystokinin and its receptor in a depression state, less attention has been paid to understanding how CCK and CCKBR are involved in depression.In this study, we combined the in vitro electrophysiological recording, optogenetic manipulation, c-fos & CCKBR staining, viral gene knockout, anatomical, and behavioral analysis to examine the critical role of CCK and CCKBR in depression.By comparing the degree of EPSPs from CCK-KO, CCKBR-KO, and C57 with CCKBR agonist CCK4, with CCKBR antagonist YM022, with Vehicle, and control mice, we provided a critical role of CCK and CCKBR in the induction of LTP.We identified that the EC CCK -BLA pathway is critical for stress susceptibility, and CCKBR antagonist YM022 prevents the development of depressive-like behavior.Overall, our data suggest that the depressive-like phenotypes and neuroplasticity in the BLA are mediated by CCK and CCKBR signaling.Synaptic plasticity including LTD and LTP has been conventionally accepted as the cellular substrate of learning and memory [65].Although the extent to which synaptic plasticity constitutes the cellular substrate of a particular behavioral phenotype is still under debate [66,67], it may serve as the substrate of depression.Some recent studies supported that stress-induced synaptic plasticity changes mediate depression [68,69].We observed that the CCK and CCKBR knockout mice showed impaired LTP formation.Notably, CCKBR antagonists YM022 and YF476 block the TBS-induced LTP in the BLA.Given that the BLA is the extremely important nucleus among numerous valence-coding brain regions [14,15], which shows structural and functional changes in depressed patients [14,70,71], current findings suggest that the CCK-CCKBR signaling within the BLA could modulate the depression through LTP.Indeed, this is supported by a previous study that revealed that the fear memories can be activated and deactivated by LTP and LTD within the amygdala [47].Future investigations are required to elucidate if LTP changes in the BLA could act as a potential therapeutic target of antidepressants.
Previous studies reported that the BLA received strong inputs from the EC [21] and most of the EC to BLA projections are CCKpositive [20], and these pathways modulate the fear memory [20].We demonstrated that the activation of EC CCK to BLA afferents enabled stress susceptibility.Interestingly, this effect was facilitated by CCK release in the BLA, and EC-BLA circuit-specific knockdown of CCK abolished the stress susceptibility.These results indicate the endogenous activity of the EC CCK -BLA circuit in conveying information related to stress susceptibility.Although EC CCK -BLA pathways facilitate stress susceptibility.However, the role of CCK neurons generally in stress processing in other brain   regions and other neural circuits can not be ruled out because CCK is highly expressed in the brain [72,73].For example, a previous study showed that the BLA CCK inputs to the nucleus accumbens (NAc) regulate depressive-like behavior via CB1 receptors [41].We found that the CCK level is the same in the EC and BLA.It would be interesting and worthy to investigate the CCK level in other brain regions and how CCK local neurons and input projections from other brain regions at the circuit level are implicated in stress processing in other brain regions.
There is a substantial causal relationship between stressful life events and the onset of episodes of major depression [7].Here, we identified that CSDS induced depressive-like behavior and enhanced the c-fos (a neuronal activity marker) [74,75] activity in the BLA.These findings are consistent with previous studies showing depression leads to hyperactivity of the amygdala [76,77].Of note, we found that the CSDS induced an elevated number of CCKBR-activated cells compared with the control group, and CCKBRKO or CCKBR antagonist local infusion into the BLA mediates the antidepressant-like effects, suggesting that these cells modulate the depression.Previous studies from our group showed that the activation of EC CCK terminals in the auditory cortex mediates CCK release which binds to CCKBR and facilitates the LTP induction and associative memory formation in the auditory cortex [27,35].In the present study, we showed that the EC CCK terminal innervates with the CCKBR cells in the BLA.Thus, current data suggest that the EC CCK terminal activation released CCK in the BLA which binds to CCKBR and facilitates the stress susceptibility while antagonizing CCKBR alleviates the stress susceptibility.
Regardless of the considerable effort of pharmaceutical companies for decades, no CCKBR antagonist has been developed into a drug, solubility, and oral bioavailability are two such reasons.Because of poor systemic availability, CI-988 (CCKBR antagonist) demonstrated a relatively modest effect on CCK4-induced panic symptoms in healthy volunteers [78].In the present study, we examined PK parameters and BBB penetration of the CCKBR antagonist YM022 and found the parameters were acceptable.Remarkably, our results demonstrated that compared to the current antidepressant fluoxetine, CCKBR antagonist YM022 exhibited an antidepressant effect with a 3333.3-foldlower dosage.
Previously, we have shown that the NMDA receptor triggers the CCK release and facilitates the formation of LTP in the auditory cortex [35].Considerable evidence also showed the key role of AMPA receptors in the formation and regulation of synaptic plasticity [79,80], and BLA-hippocampal innervation modulates stress-induced depressive-like behaviors through AMPA receptors (Ma, Li et al. 2021).Some studies have shown an increased AMPA subunit level in the amygdala of chronically stressed mice [81,82], and patients with depressive disorders also showed elevated AMPA receptor levels in the anterior cingulate cortex [83].Ketamine mediates antidepressant effects through BDNF-AMPA receptor signaling [84][85][86], and BDNF receptor TrKB knockout mice showed an impaired level of the AMPA receptor subunit [87].Here, we performed western blot analysis and found that the CCKBRKO mice exhibited decreased AMPA levels compared with the control (Fig. S7D-F), indicating that CCK-CCKBR binding may increase the AMPA receptors trafficking accompanied by LTP formation, and CCKBR downregulation with reduced AMPA trafficking impaired LTP formation, which subsequently results in reduced depressive-like behavior.However, further research is necessary to directly correlate the CCKBR association with AMPA alteration in stress processing and LTP.In conclusion, CCKBR could be a new potential target for the treatment of depressive-like phenotypes in mice.
compared with that of the control mice (Fig.2D; ChETA vs. EYFP, 65.4 ± 3.9% vs. 77.8± 3.0%).Moreover, we found the trend of increased immobility in the tail suspension test (TST), and the percentage time in the center of the open field (OFT) is decreased without any change in the total distance

Fig. 2
Fig.2Optogenetic activation of the EC cck -BLA circuit with SSDS induced depressive-like behaviors.A Top: The workflow of the SSDS model with optogenetic manipulation.Bottom left: Schematic diagram of virus injection into EC and optical fiber implantation into BLA, and Virus expression CHETA-EYFP or EYFP was specifically expressed in CCK neurons at EC (up) and projections to BLA (bottom), scale bar, 500 μm.Bottom right: Paradigms of 20 Hz optogenetic activation of CCK fibers in the EC-BLA circuit during two sensory phases of SSDS.B Schematic diagram of SI (left).C The time spent in the social interaction zone without a target and with a target among EYFP and Cheta groups (Two-way ANOVA: Interaction: F (1, 48) = 0.046, P < 0.8; EYFP with No-Target vs. Target N = 13, t = 0.75, P < 0.90; Cheta with No-Target vs. Target N = 13, t = 1.1, P < 0.5; Bonferroni's adjustment).D Sucrose preference percentage among EYFP and Cheta groups (EYFP vs. Cheta N = 13, df = 24, t = 2.7, P < 0.013; two-tail unpaired t-test).E Top left: Schematic representation of the experimental design.Top right: virus expression in EC (ChrimsonR-mCherry and BLA (CCK-GFP).Bottom left: Fiber photometry setup.F Average trace of before and after laser stimulation of EC CCK neurons.G Change in average response of CCK in individual mice (Before Laser vs.After Laser N = 5, df = 4, t = 2.8, P < 0.04; two-tail paired ttest).All data are the mean ± SEM.

Fig. 4
Fig. 4 CSDS activated CCKBR neurons in the BLA.A Experimental design.B Representative track traces of the no-target phase (top) and the target phase (bottom) among sham and CSDS mice.C Time spent in the social interaction zone in the absence or presence of a social target (Two-way ANOVA, Interaction: F (1, 26) = 2.0, P < 0.1, CSDS (N = 10) vs sham (N = 5) with a target, t = 3.2, P < 0.008; Bonferroni's adjustment).D Representative images of C-fos and CCKBR cells in the BLA.Left, c-Fos (red) expression and colocalization with CCKBR (green) in the BLA; scale bar, 200 μm.Right, an enlarged view of the white rectangle in the image on the left, c-Fos immunopositive neurons (left), and colocalization with CCKBR (right), 20 μm.E Quantification of c-Fos density (sham (N = 3 mice, 10 slices) vs CSDS (5 mice, 20 slices) per mm, t = 5.3, P < 0.0001; two-tail unpaired t-test).F Percentage of the c-Fos positive cells that colocalized with CCKBR (sham vs CSDS per mm, t = 20.2,P < 0.0001; two-tail unpaired t-test).All data are the mean ± SEM.