Npas4 is involved in synaptic and cognitive function by regulating the transcription of Neuroligin-1 and N-cadherin Authors: Yuran Gui, Yusi Chen, Qian Guo, Yong Luo, Fumin Yang, Yacoubou Abdoul Razak Mahaman, Jian-Zhi Wang, Rong Liu, Hongbin Luo, Xiji Shu, Xiaochuan Wang, Fang Huang Published: 2026 ABSTRACT Ischemic injury has been reported to induce mild to severe permanent deficits. Nevertheless, its recovery is often dynamic, depending on the plasticity of the injured neurons. The current study found that during the early stage of cerebral ischemia (1 h), Npas4 expression was increased in the hippocampus of mice, thereby transcriptionally regulated the expression of Neuroligin-1 and N-cadherin to enhance long-term potentiation. Knocking down of Npas4 reduced the level of Neuroligin-1 and N-cadherin, sequentially prevented the enhanced neural plasticity in the early cerebral ischemia. Furthermore, the downregulation of Neuroligin-1 and N-cadherin also impeded the ischemia-induced enhancement in neural plasticity without affecting the expression of Npas4. Overexpression of Npas4 in the primary neurons increased the number of VAMP2-labeled synaptic vesicles, while interfering the expression of Neuroligin-1 and N-cadherin in the primary neurons with siRNA reduced the increase of VAMP2-labeled synaptic vesicles without affecting the Npas4 expression. Interestingly, the transcriptional level of Npas4, Neuroligin-1, and N-cadherin declined in multiple brain regions of Alzheimer’s disease (AD) patients and 3 × TG AD mice. Moreover, overexpression of Npas4 increased the expression of Neuroligin-1 and N-cadherin, ameliorated the synaptic plasticity impairment of 3 × TG mice, and enhanced the cognitive function. These findings suggest a previously undiscovered biological mechanism of neural plasticity, driven by Npas4 upregulation of Neuroligin-1 and N-cadherin, which partially explain why the compensatory increase in Npas4 in the early stages of ischemic injury and the lack of Npas4 in AD progression, highlighting potential therapeutic strategies for cognitive deficits. Introduction Stroke is the second leading cause of death and the third most common cause of disability worldwide [1]. Only a minority of stroke patients fully recover, while in most cases, ischemic damage causes mild to severe permanent damage [2]. Although the functional damage caused by stroke is often severe, the recovery process is dynamic and depends on the plasticity of the damaged neurons. Plasticity changes in cortical and subcortical structures following ischemic injury have been repeatedly documented in studies of animal models and human stroke survivors, and these changes may increase susceptibility to adaptive changes and recovery in the damaged brain. Studies have shown that neurons around focal infarcts in rat cortex tend to induce long-term potentiation (LTP) more easily in vitro [3]. We refer to the enhancement of excitatory synaptic plasticity resulting from this ischemic injury as ischemic long-term potentiation(iLTP) [4, 5], which is a pathological form of synaptic response in the early stage of ischemic stroke. However, the nature of iLTP has not been clarified. Since LTP is usually associated with synaptic transmission or synaptic plasticity that contributes to learning and memory, it is speculated that increased LTP may contribute to enhanced plasticity and functional recovery of neural networks after stroke, but the specific molecular mechanism remains unclear. The optimal behavioral response to different environmental information is an important basic function of the nervous system, in which the interaction between genetic material and sensory cues is a key step in driving brain development and regulation [6–9]. Neuronal activity and subsequent calcium influx start specific signaling cascades that lead to nuclear translocation of various transcription factors, thereby rapidly inducing the immediate-early gene (IEG) expression program, which then induces transcription of subsequent response genes, and locally modulating synaptic development and plasticity [10]. NPAS4 (neuronal PAS-domain containing protein 4, a neural activity-dependent immediate-early gene) is recognized as an IEG which functions as a transcription factor with exclusive expression in the brain tissue [11–16]. Npas4 is rapidly and robustly activated in response to cerebral ischemia in the whole brain, especially in the ischemic side of the hippocampus, with the level increasing [17–19]. Previous studies have shown that chemically induced neuronal activation can also lead to increased expression of Npas4 [20, 21]. It has been demonstrated that Npas4-dependent regulation of neuronal synapse and network remodeling occurs in the neocortex, hippocampal CA1, and CA3 regions of animals [12, 21, 22]. Although increasing evidence suggests a close relationship between Npas4 and iLTP, key evidence is still lacking. As a transcription factor dependent on neuronal activity, Npas4 can recruit RNA polymerase II to the promoter and enhancer of target genes involved in neuronal activity regulation in the hippocampal CA3 subregion, thereby promoting short-term and long-term contextual memory [23, 24]. Npas4 also transcriptionally regulates the mRNA levels of other IEG induced by depolarization, including Arc/Arg3.1, c-Fos, Zif268/Egr1, and brain-derived neurotrophic factor (BDNF) [23, 25]. To investigate the pathway by which Npas4 participates in iLTP, we selected two potential target genes with synaptic adhesion function, Neuroligin-1 and N-cadherin, based on bioinformatics predictions of Npas4 binding sites and their regulatory effects on synaptic function [26], to study their specific functions in the iLTP process. Neuroligin-1 plays a role in promoting hippocampal synaptic differentiation by connecting with axonal proteins at both the pre- and postsynaptic terminals [27, 28]. Numerous studies have indicated that the function of Neuroligin-1 is related to the induction and maintenance of LTP [29–31]. However, there are few and conflicting reports on the expression changes of Neuroligin-1 during brain ischemia. One study claimed an increase in the binding of neurexin-neuroligin1-PSD-95 after transient cerebral ischemia, but changes in the protein level of Neuroligin-1 were not mentioned [32]. Other studies have shown a decrease in the protein levels of Neuroligin and N-cadherin after local transient brain ischemia reperfusion [33], and a reduction in the protein levels of Neuroligin and Neurexin after 8 h of neuronal glucose-oxygen deprivation in vitro [34]. Therefore, further research is needed to confirm the changes and roles of Neuroligin-1 during iLTP, and the transcriptional regulation of Neuroligin-1 by Npas4 and its function in this process need to be clarified. As another synaptic adhesion protein of interest in this study, N-cadherin is regulated in a calcium-dependent manner in hippocampal CA3 neuron synapses during brain ischemia, similar to Neuroligin-1 [33, 35]. Previous studies have revealed the relationship between N-cadherin and synaptic plasticity [36–39]. Studies have shown that N-cadherin expression decreases in the surrounding tissue of ischemic brain [40], and the mRNA level of N-cadherin (Cdh2) in the ischemic neocortex of mice with middle cerebral artery occlusion (MCAO) decreases after 4 and 24 h [41]. However, the expression of N-cadherin during the early iLTP process of brain ischemia and its relationship with Npas4 regulation require further investigation. The current study first demonstrated that Npas4 can promote the transcriptional expression of Neuroligin-1 and N-cadherin by binding to their transcription promotor at the cellular level. Furthermore, in the early stages of brain ischemia in animal models, upregulation of Npas4, Neuroligin-1, and N-cadherin expression was detected, along with enhanced LTP in hippocampal slices. Interference of the expression of Neuroligin-1 and N-cadherin prevented the LTP enhancement observed in the early stage of brain ischemia. Additionally, overexpression of Npas4 in cultured primary neurons promoted the formation of neuronal synaptic vesicles, a process that also depends on the expression of Neuroligin-1 and N-cadherin. A bunch of researches have indicated that Npas4, Neuroligin-1, and N-cadherin are all closely related to AD pathology and learning and memory [23, 42–49]. Whether the mechanism by which Npas4 promotes the transcriptional expression of Neuroligin-1 and N-cadherin to enhance neuronal synaptic plasticity is involved in regulating learning and memory function during the progression of AD is not yet clarified. To investigate this, we first examined the changes in Npas4, Neuroligin-1, and N-cadherin in AD. Data from the AlzData database (http://www.alzdata.org/) showed that the transcription levels of Npas4, Neuroligin-1, and N-cadherin were decreased in multiple brain regions of the AD brain. Additionally, in the hippocampal tissue of 6-month and 9-month-old 3 × TG mice, the protein levels of Npas4, Neuroligin-1, and N-cadherin were decreased compared to age-matched controls. After overexpressing Npas4 via virus infection in the brain of 5-month-old 3 × TG mice for one month, the protein levels of Neuroligin-1 and N-cadherin were upregulated, and the cognitive impairment of 3 × TG mice was restored. The results of this study indicate that in the early stages of cerebral ischemia, Npas4 enhances neuronal synaptic plasticity by promoting the transcription of Neuroligin-1 and N-cadherin and participates in iLTP. Additionally, Npas4 strengthens the synergistic promotion of Neuroligin-1 and N-cadherin on synaptic vesicle formation and synaptic plasticity, which can improve cognitive impairment caused by the deficiency of Npas4, Neuroligin-1, and N-cadherin in an animal model of Alzheimer’s disease. Materials and methods Animals The 2month old wild-type male C57BL/6 mice used in this study were purchased from the Experimental Animal Center of Tongji Medical College, Huazhong University of Science and Technology. The male 3 × TG mice which carry the Swedish mutation of APP (APP KM670/671NL), MAPT P301L, and PSEN1 M146V mutations were purchased from Shulaibao (Wuhan) Biotechnology Co., Ltd (stock no. YDS0113). All animals were kept in a spacious and suitable environment with free access to water and food and given a 12 h light/dark cycle, with room temperature maintained at 25 ± 2°C and appropriate air humidity. In this study, we exclusively used male mice. This approach is consistent with established practices in the field, where male animals are often selected to minimise experimental variability associated with oestrous cycle-dependent hormonal fluctuations in females. By using male mice, we reduce potential confounding neurofunctional effects of endogenous oestrogen, enhance reproducibility, and ensure direct comparability with the majority of existing preclinical stroke literature. All the animal experiments conducted in this study were designed and performed in strict accordance with the principles of the ARRIVE guidelines. The animals were randomly assigned to different experimental groups. The investigator was blinded to the group allocation during all the following experiments. Cell culture SH-SY5Y cells (human neuroblastoma cells) were purchased from the American Type Culture Collection (ATCC, Cat No. CRL-2266, RRID: CVCL_0019), authenticated by STR profiling, tested for mycoplasma contamination and cultured in DMEM/high glucose medium (Life Technologies, Cat No. 11965) containing 10% fetal bovine serum (FBS, Life Technologies, Cat No. 10099-141) and 100 U/mL penicillin-streptomycin (Gibco, Cat No. 15140122). The cells were maintained in a humidified atmosphere at 37 °C and 5% CO2. The process of culturing primary neurons from mice is briefly described as follows: First, poly-D-lysine (PDL, ThermoFisher Scientific, Cat No. A38904-01) is coated onto culture plates or dishes and dried under ultraviolet sterilization. Pregnant C57BL/6 mice at embryonic day 18 are anesthetized and sterilized, and then the fetuses are removed by laparotomy and placed in clean crushed ice for anesthesia. The anesthetized fetuses are then immersed separately in 75% ethanol for 10 s and physiological saline for 10 s. After decapitation in a sterile environment, the brains are quickly removed and the meninges and blood vessels are carefully dissected away. The cortex and hippocampus are separated and placed in a clean bottle filled with pre-cooled dissection solution. The brain tissue is then cut into 1 mm3 pieces with scissors and digested with trypsin (Gibco, Cat No. 25300-054) for 15 min, with gentle shaking every 5 min. The trypsin reaction is then terminated by adding an equal volume of growth medium (DMEM/F12 (Gibco, Cat No. A4192002) containing 10% FBS (HyClone, Cat No. SH30256.01) and 100 U/ml penicillin-streptomycin), and the tissue is repeatedly blown and pipetted until no large tissue pieces remain. The cell suspension is filtered through a 40 μm mesh (Cytiva, Cat No. CSS010040) and seeded onto the PDL-coated culture plates or dishes. After 6 h, the medium is changed to maintenance medium (containing 2% B27, Gibco, Cat No. 17504044, 0.5 mM glutamine (Gibco, Cat No. A4000225301), and 100 U/mL penicillin-streptomycin) to promote neuronal differentiation, and the medium is changed every 3.5 days with fresh medium. The cells are cultured at 37 °C and 5% CO2 in a CO2 incubator. Plasmids, virus and siRNA The pAAV-CMV-EGFP-2A-Npas4-3FLAG for overexpressing mouse Npas4 (GenBank accession number: NM_153553) and its vector control plasmid were constructed by GenePharma Co., Ltd. (Shanghai), the adeno-associated virus was packaged by Shanghai JiKai Gene Technology Co., Ltd. The Lenti-EF1a- Npas4-P2A-EGFP plasmid and lentivirus were also constructed and packaged by Shanghai JiKaiGene Technology Co., Ltd. Plasmids and lentivirus for interfering with the expression of mouse Npas4, Nlgn1 (GenBank accession number: NM_138666), and Cdh2 (GenBank accession number: NM_007664) were constructed and packaged by Shanghai JiKaiGene Technology Co., Ltd. The applied plasmid vector was hU6-MCS-CBh-gcGFP-IRES-puromycin, and the effective interfering sequences targeted were 5’-CCTGGATCTTAAACCCTGGAA-3’, 5’-CCTGTACTACAAGAAGGATAA-3’, and 5’-GAACAGTTTCACCTGATATTC-3’ respectively. The siRNA used for interfering with Neuroligin-1 and N-cadherin expression was synthesized by Sangon Biotech (Shanghai) Co., Ltd. All plasmids used in this study were extracted, prepared, and sequence-verified using an endotoxin-free plasmid extraction kit (Beijing TianGen Biochemical Technology Co., Ltd., Cat No. DP117). Transfection and infection of cells For transfection of inherited cytoplasmic plasmids, cells were seeded onto a 6-well culture plate and incubated for one day to allow cell attachment. Following the instruction manual, fresh serum-free medium was replaced 2 h prior to transfection. A mixture containing 2 μg of plasmid and 4 μl of Lipofectamine 2000 (Thermo Fisher Scientific, Invitrogen, Carlsbad, CA, USA, Cat No. 11668019) transfection reagent was added to each well of the 6-well plate. After 6–8 h, the medium was replaced with complete culture medium and cells were incubated for 24–48 h. Transfection time was determined based on plasmid fluorescence expression, and cells were finally collected for detection. For viral infection of primary cells, on the 4th day of primary neuron culture, the medium was half exchanged with fresh medium prior to infection. Lentivirus was introduced to achieve a titer of over 1 × 105 in the culture medium. After 10 h of infection, the medium was replaced with fresh medium, and virus infection efficiency was determined by the fluorescence 72 h later. For siRNA transfection of primary cells, on the 5th day of primary neuron culture, the medium was half exchanged with fresh medium prior to transfection. For each well of a 6-well plate, a mixture containing 1 μl of 20 μM siRNA and 4 μl of RNA-specific transfection reagent, RNATransMate (Wuhan Sangon Biotechnology Co., Ltd., catalog number: E607402), was added for transfection. After 10 h of transfection, the medium was replaced with fresh medium, and cells were analyzed 24–48 h later (6-7 DIV). Stereotaxic injection of animal brain Anesthetize the animals with 2% isoflurane and administer 0.1 mg/kg buprenorphine subcutaneously for analgesia. Check the depth of anesthesia by pinching the toe. Trim the hair and secure the mouse in a stereotaxic instrument (Stoeltling, Wood Dale, IL). Apply sterile erythromycin eye ointment (0.5%) to the eyes to prevent dryness and corneal damage. After disinfecting the head surface with iodine, make a small incision to expose the coronal and sagittal sutures of the skull at the bregma and lambda landmarks, respectively, and clean the connective tissue with a sterile cotton swab. Determine the coordinates of the hippocampus with the stereotaxic instrument at ML: +1.5 mm, AP: −2.1 mm, DV: 1.8 mm. Drill two small holes at the left and right hippocampal coordinates on the skull. Slowly inject 1 μl of AAV and/or 3 μl of LV at the respective locations. Leave the needle in place for 10 min after injection, suture the incision, and disinfect the surface with the lincomycin- lidocaine gel. Monitor the mouse’s condition on days 1, 2, 7, and 10 after surgery. Antibodies and reagents The primary antibodies used in this study are listed in Table 1. The secondary antibodies for Western blotting using Odessey imaging system were purchased from Biosciences (catalog numbers: 926-32211 and 96-32210). The fluorescent secondary antibody (Alexa Fluor™ Plus 594, goat anti-rabbit, dilution 1:500, Catalog No. A32754; IgG conjugated to Alexa Fluor 488, goat anti-rabbit, dilution 1:500, Catalog No. A11034) were purchased from Thermo Fisher Scientific, IgG conjugated to Alexa Fluor 594(goat anti-mouse, dilution 1:500, Catalog No. ab150116) was purchased from Abcam, and the nuclear stain Hoechst (Catalog No. B2261) was purchased from Sigma. The cell culture media, serum, and other reagents used for cell culture were purchased from Gibco BRL (Gaithersburg, MD, USA).Table 1Antibodies applied in the current study.AntibodyProduct codeTypeDilutionSourceβ-actin4970 sMono-, Mouse1:1000 for WBCell SignalingNpas4sc-293239Mono-, Mouse1:1000 for WBSanta CruzNpas4sc-293239Mono-, Mouse1:200 for IFSanta CruzNpas4sc-293239Mono-, Mouse1:100 for ChIPSanta CruzNeuroligin-1ab153821Poly-, Rabbit1:1000 for WBAbcamNeuroligin-1ab153821Poly-, Rabbit1:200 for IFAbcamN-cadherinab76057Poly-, Rabbit1:1000 for WBAbcamN-cadherinab76057Poly-, Rabbit1:200 for IFAbcamVAMP2ab3347Poly-, Rabbit1:200 for IFAbcamPSD-95MA1-046Mono-, Mouse1:2000 for WBThermo FisherPSD-95MA1-046Mono-, Mouse1:200 for IFThermo FisherSynaptophysinMA5-14532Mono-, Rabbit1:200 for WBThermo Fishersynapsin-1MA5-31919Mono-, Mouse1:1000 for WBThermo Fisher Table 1Antibodies applied in the current study.AntibodyProduct codeTypeDilutionSourceβ-actin4970 sMono-, Mouse1:1000 for WBCell SignalingNpas4sc-293239Mono-, Mouse1:1000 for WBSanta CruzNpas4sc-293239Mono-, Mouse1:200 for IFSanta CruzNpas4sc-293239Mono-, Mouse1:100 for ChIPSanta CruzNeuroligin-1ab153821Poly-, Rabbit1:1000 for WBAbcamNeuroligin-1ab153821Poly-, Rabbit1:200 for IFAbcamN-cadherinab76057Poly-, Rabbit1:1000 for WBAbcamN-cadherinab76057Poly-, Rabbit1:200 for IFAbcamVAMP2ab3347Poly-, Rabbit1:200 for IFAbcamPSD-95MA1-046Mono-, Mouse1:2000 for WBThermo FisherPSD-95MA1-046Mono-, Mouse1:200 for IFThermo FisherSynaptophysinMA5-14532Mono-, Rabbit1:200 for WBThermo Fishersynapsin-1MA5-31919Mono-, Mouse1:1000 for WBThermo Fisher Antibodies applied in the current study. Protein immunoblotting Tissues or cell pellets were washed with ice-cold PBS and then homogenized in RIPA buffer (Beyotime Biotechnology, Cat No. P0013C) mixed with a protease inhibitor cocktail (Beyotime Biotechnology, Cat No. P1005), and placed on ice for 10 min. The mixture was then centrifuged at 14,000 rpm for 20 min at 4 °C to collect the supernatant. Protein concentration was measured using the BCA (Beyotime Biotechnology, Cat No. P0009) method, and the protein lysate was diluted to 5 mg/ml using 4 × Loading Buffer. Samples were loaded onto a 10% polyacrylamide gradient gel (10 μg sample for Npas4, Neuroligin-1, N-cadherin and PSD-95 antibodies, 5 μg sample for synapsin-1, synaptophysin and β-actin antibodies were loaded in each lane) and electrophoresed, followed by transferring onto nitrocellulose (NC) membranes (Thermo Fisher, Cat No. 88024). The membranes were blocked with 3% skim milk (Beyotime Biotechnology, Cat No. P0216) at room temperature (25 °C) for 45 min to an hour, and then incubated with primary antibodies overnight at 4 °C. The membranes were washed with TBS containing Tween-20, incubated with secondary antibodies (anti-M or anti-R IgG conjugated with fluorescence in the 800 nm channel, Odyssey: catalog numbers 926-32210, 926-32211, Biosciences) in the dark for 1 h, washed again with TBS containing Tween-20, and then visualized using the Odyssey Infrared Imaging System (LI-COR Biosciences, Lincoln, NE, USA). Image gray-scale analysis was performed using Image J software. All quantitative analyses for every protein were performed exclusively on the major band corresponding to its expected molecular weight. RNA extraction, reverse transcription, real-time fluorescence quantitative PCR RNA was extracted from cells and tissues using Trizol reagent (Invitrogen, Waltham, MA, USA, Cat No. 15596018CN) according to the manufacturer’s instructions. The RNA concentration was determined using a spectrophotometer (DG-3022A) and adjusted to 500 ng in a 10 μl reaction volume. cDNA was synthesized using a reverse transcription kit (Applied Biosystems, Foster City, CA, USA, Cat No. 4368814). The collected cDNA was used to prepare a standard 20 μl PCR reaction mixture with TB Green® Premix Ex Taq™ II FAST qPCR (TaKaRa, Dalian, Japan, Cat No. CN830A), gene-specific primers (Table 2), and double distilled water according to the instructions. GAPDH (for SH-SY5Y cells)/Gapdh (for tissues) served as the internal reference for primers. At least two replicates were used for each data point. Real-time quantitative PCR reactions were performed on a Biosystems 7900 Prism system with annealing temperatures and cycle numbers set accordingly, and the data were analyzed based on the final results. The 2-ΔΔCt method was used for relative quantification of gene expression.Table 2Primers applied to test the mRNA level.NPAS4Forward5′-GCCCTGCCTCGCTCTTCC-3′Reverse5′-TCTCTGCCTGAATATCTCCACTCTC-3′NLGN1Forward5′-ATGATGGAAGTGTCTTGGCAAGTT-3′Reverse5′-CTTTGCAGCCTGATCGCCTGTA-3′CDH2Forward5′-GACAATGCCCCTCAAGTGTT-3′Reverse5′-TGAATTGGGGTCTGGAGTTT-3′GAPDHForward5′-CTGGGCTACACTGAGCACC-3′Reverse5′-AAGTGGTCGTTGAGGGCAATG-3′Npas4Forward5′-CAGATCAACGCCGAGATTCG-3′Reverse5′-CACCCTTGCGAGTGTAGATGC-3′Nlgn1Forward5′-TCCAAGAATGATGTGATGTTGAGTG-3′Reverse5′-TCAAAGCGGTTGGGTTTGGTA-3′Cdh2Forward5′-CACACCCTGGGGATATTGGG-3′Reverse5′-GCTGCCCTCGTAGTCAAAGA-3′GapdhForward5′-ACAACTTTGGCATTGTGGAA-3′Reverse5′-GATGCAGGGATGATGTTCTG-3′ Table 2Primers applied to test the mRNA level.NPAS4Forward5′-GCCCTGCCTCGCTCTTCC-3′Reverse5′-TCTCTGCCTGAATATCTCCACTCTC-3′NLGN1Forward5′-ATGATGGAAGTGTCTTGGCAAGTT-3′Reverse5′-CTTTGCAGCCTGATCGCCTGTA-3′CDH2Forward5′-GACAATGCCCCTCAAGTGTT-3′Reverse5′-TGAATTGGGGTCTGGAGTTT-3′GAPDHForward5′-CTGGGCTACACTGAGCACC-3′Reverse5′-AAGTGGTCGTTGAGGGCAATG-3′Npas4Forward5′-CAGATCAACGCCGAGATTCG-3′Reverse5′-CACCCTTGCGAGTGTAGATGC-3′Nlgn1Forward5′-TCCAAGAATGATGTGATGTTGAGTG-3′Reverse5′-TCAAAGCGGTTGGGTTTGGTA-3′Cdh2Forward5′-CACACCCTGGGGATATTGGG-3′Reverse5′-GCTGCCCTCGTAGTCAAAGA-3′GapdhForward5′-ACAACTTTGGCATTGTGGAA-3′Reverse5′-GATGCAGGGATGATGTTCTG-3′ Primers applied to test the mRNA level. Chromatin immunoprecipitation (ChIP) All steps were performed according to the manufacturer’s protocol for the SimpleChIP® Enzymatic Chromatin IP Kit (Cell Signaling Technology, Cat. No. 9002). SH-SY5Y cells were fixed with 1% formaldehyde (mixed with 540 μl 37% formaldehyde and 20 ml culture medium, 1000 μl per well in a 6-well plate) at room temperature for 10 min, then quenched with 100 μl of 10 × glycine (Cell Signaling, Cat No. 7005) at room temperature for 5 min. The culture medium was removed, and the cells were washed twice with cold PBS. Then, 80 μl of SDS lysis buffer containing cocktail (protease inhibitor) was added to each well on ice, and the lysate was collected. The lysate was sonicated using a Covaris S220 ultrasonic Ator at 25% power, with 4.5 s on and 9 s off, for a total of 14 cycles, to obtain the desired DNA fragment size (200–1000 bp). After sonication, the sample was centrifuged at 10,000 × g at 4 °C for 10 min to remove insoluble material. 200 μl of the lysate was taken (the rest was frozen at −80 °C and could be stored for 2 months). For each sample, 100 μl was mixed with 4.2 μl of 5 M NaCl and incubated at 65 °C for 3 h to reverse the cross-linking. Gel electrophoresis was used to confirm the success of sonication by determining the size of the DNA fragments. Another 100 μl of the lysate was mixed with 900 μl of Chip Dilution buffer containing cocktail and 60 μl of Protein G Agarose to remove endogenous DNA and proteins that might bind to Protein G Agarose. The mixture was incubated at 4 °C for 1 h, and the supernatant was collected after centrifugation at 3000 × g - 5000 × g for 1 min. From this step, 50 μl of the sample was saved at 4 °C as the Input, 200 μl was used as the positive control with 1 μl of anti-RNA polymerase, 200 μl was used as the negative control with 1 μl of Mouse IgG, and 200 μl was used as the experimental test with 2 μl of the target protein antibody. Each sample was mixed overnight at 4 °C, and then 60 μl of Protein G Agarose was added to each tube and mixed at 4 °C for 1 h. The mixture was then centrifuged at 3000–5000 × g for 1 min, and the supernatant was discarded. The pellet was washed sequentially with 1 ml of Low salt buffer, High salt buffer, LiCl, and TE buffer for 5 min each. Elution and reversal of cross-linking: Elution buffer (10 μl of 20% SDS, 20 μl of 1 M NaHCO3, and 170 μl of H2O) was prepared. 200 μl of the buffer was added to the sample saved as Input, and 100 μl was added to each of the samples used as the positive, negative, and experimental controls. The mixtures were incubated at room temperature for 15 min to elute the proteins and DNA. The supernatant was collected after centrifugation at 3000 × g - 5000 × g for 1 min, and the elution process could be repeated once. Purification: After reversal of the cross-linking, each of the samples were incubated with 1 μl of RNase at 37 °C. 30 min later, the mixture was added with 4 μl of 0.5 M EDTA, 8 μl of 1 M Tris-HCl and 1 μl of Proteinase K treatment and incubated for another 2 h at 45 °C. Each sample was purified using Binding reagent A at a ratio of 1:5. 600 μl of the mixture was added to a spin column, and centrifuged at 10,000 × g for 30 s twice. The collecting tube and the liquid were removed and the column was placed in a new tube. After Adding 50 μl of EB or pure water to the column, the tube was held on for 5 min, then centrifuged at 10,000 × g for 30 s to collect purified DNA. Prime pairs (Table 3) of different promotor binding sequences were applied to analyze the collected DNA with qPCR or DNA electrophoresis.Table 3Primers applied for the ChIP test.NLGN1-AForward5’- AAGGGCCTGGTTTTCGTCTC -3’Reverse5’- GTGTGCAGCCAAACGATGG -3’NLGN1-BForward5’- CATTTGGCCATCGTTTGGCT -3’Reverse5’- GGAAGTTCTCGGCAATCCCC -3’NLGN1-CForward5’- GGCCAGTCGCTTGTTTAACTC -3’Reverse5’- AAAGAGGTTAGCGGCACAGA -3’CDH2-AForward5’-GGAAGCCAGCAGAAATCCTTT -3’Reverse5’-GGCCAACAGTTTCAGGGTACT -3’CDH2-BForward5’- TTTTATTTCCCGCAGCCCCT -3’Reverse5’- CTTCTCGGCTCCTCTTTCCG -3’CDH2-CForward5’- CGGAAAGAGGAGCCGAGAAG -3’Reverse5’- GGATGTGGAGGTGGAAGTGG -3’ Table 3Primers applied for the ChIP test.NLGN1-AForward5’- AAGGGCCTGGTTTTCGTCTC -3’Reverse5’- GTGTGCAGCCAAACGATGG -3’NLGN1-BForward5’- CATTTGGCCATCGTTTGGCT -3’Reverse5’- GGAAGTTCTCGGCAATCCCC -3’NLGN1-CForward5’- GGCCAGTCGCTTGTTTAACTC -3’Reverse5’- AAAGAGGTTAGCGGCACAGA -3’CDH2-AForward5’-GGAAGCCAGCAGAAATCCTTT -3’Reverse5’-GGCCAACAGTTTCAGGGTACT -3’CDH2-BForward5’- TTTTATTTCCCGCAGCCCCT -3’Reverse5’- CTTCTCGGCTCCTCTTTCCG -3’CDH2-CForward5’- CGGAAAGAGGAGCCGAGAAG -3’Reverse5’- GGATGTGGAGGTGGAAGTGG -3’ Primers applied for the ChIP test. Middle cerebral artery occlusion (MCAO) Preparation of MCAO model was performed according to the methods described by Connolly [50] and Hata [51], with modifications made based on the differences in animal sizes and vessel diameters. For clot preparation, a 5 cm segment of 0.23 mm diameter fishing line (Amber, Japan) was marked at 2 cm and coated evenly with silicone (Henkel, Australia) on the surface (diameter of 0.28–0.30 mm and length of 11 mm). The line was then left to dry at room temperature, and the line end was carefully smoothed with a surgical microscope to avoid any burrs or rough edges. 2-month-old male C57BL/6 mice were anesthetized with 4 ml/kg body weight of 20% ethyl urethane injected intraperitoneally. The hindlimbs were fixed to a surgical heating pad with tape and the tongue was pulled out, while silk thread was used to pull up the incisors. A midline neck incision was made under a surgical microscope to expose the left carotid artery sheath, and the carotid artery and vagus nerve were carefully separated, avoiding any traction damage to the vagus nerve. The carotid artery was then dissected up to the bifurcation, and the external carotid artery was ligated at its proximal end and the common carotid artery was ligated at its distal end. A loose knot was made at the proximal end of the common carotid artery, and the internal carotid artery was occluded with a vascular clamp. A small incision was made between the two knots on the common carotid artery, and the prepared clot was inserted into the internal carotid artery along the common carotid artery (delivered to a depth of approximately 20–21 mm from the bifurcation point of the carotid artery according to the mark on the clot). The clot was stopped when resistance was encountered to ensure blockage of the ipsilateral middle cerebral artery. The loose knot on the common carotid artery was then moved above the incision to secure the clot. Bleeding from the wound was thoroughly stanched, and the skin was sutured. The animal was then placed in a feeding cage with insulation and allowed to recover. Tissue perfusion and sampling were performed on animals in the sham group, as well as those with occlusion for 1, 2, and 6 h, respectively. Immunofluorescence detection of primary neuron cells Place appropriately sized circular glass slides that have been alcohol-disinfected and high-pressure sterilized in a 12-well plate. After coating with PDL, primary neurons are seeded on the glass slides, followed by virus transduction and siRNA transfection. Cells are washed with PBS for 3 × 5 min and fixed with 4% paraformaldehyde (PFA; Sinopharm Chemical Reagent, Cat# 80096628) for 30 min. After washing with PBS for another 3 × 5 min, cells are permeabilized with 0.5% TritonX-100 in PBS for 25–45 min at room temperature. Cells are then washed three times with 0.1% Triton-PBS for 5 min each time and blocked with 5% BSA at room temperature for 30–60 min. The primary antibody is diluted to an appropriate concentration with TBS and added to the wells. The plate is incubated overnight at 4 °C in a humid box. After washing with 0.1% Triton-PBS for 3 × 5 min, the corresponding fluorescent secondary antibody is diluted to an appropriate concentration in TBS and incubated in a humid opaque box at 37 °C for 1 h. The nuclei are stained with DAPI or Hochest after washing with 0.1% Triton-PBS for 3 × 5 min. Finally, the slides are sealed with an antifade mounting medium containing an anti-fluorescence quencher. The sections were imaged with a confocal microscope (ZEISS) at 60 × magnification. Confocal images were acquired, and VAMP2 positive puncta counts were performed using ImageJ software. Immunofluorescence of mouse brain slices Anesthetize mice with 20% ethyl urethane intraperitoneal injection, open the chest and expose the heart. Insert a catheter needle into the aortic end of the heart and fix the needle with a hemostatic forceps. Open the right atrium with scissors. First, perfuse the circulation with 0.9% saline solution, adjust the needle direction to the maximum perfusion speed. Rinse the blood until the mouse liver turns white, and then switch to 4% paraformaldehyde perfusion. Quickly remove the head and immerse it in 4% paraformaldehyde for 48 h of fixation, then dehydrate the fixed head with 30% sucrose. Use a cryostat to make 20-μm-thick coronal sections of the mouse brain. Place the brain sections in a 6-well cell culture plate, soak in 0.1 M PB and store. Select suitable brain sections of the hippocampus, permeabilize, block, incubate with primary and secondary antibodies (IgG conjugated to Alexa Fluor 594, goat anti-mouse, dilution 1:500, Abcam, Catalog No. ab150116; IgG conjugated to Alexa Fluor 488, goat anti-rabbit, dilution 1:500, Thermo Fisher, Catalog No. A11034), stain with DAPI (1 μg / 5 ml, Sigma-Aldrich, Catalog No. 28718-90-3), and proceed with the same process as cell immunofluorescence. Finally, mount the coverslips with antifade mounting medium for fluorescence. The sections were imaged with a confocal microscope (ZEISS) at 20 × magnification under the same parameters setting in different groups. Confocal images were acquired, and integrated density analysis was performed using ImageJ software. The images to be analyzed were imported into ImageJ software and converted to 16-bit format via Image → Type → 16-bit. The Auto threshold function was then applied (Image → Adjust → Threshold → Auto), and the resulting threshold values were recorded. To ensure comparability across all measurements within the study, these same threshold values were applied consistently to all images from the same channel. The “Dark background” option was selected before confirming. Subsequently, under Analyze → Set Measurements, the parameters “Area”, “Mean gray value”, “Integrated density”, and “Limit to threshold” were enabled. Measurements were obtained using Analyze → Measure, and the Integrated density values were recorded for subsequent statistical analysis. Electrophysiological test (LTP) After light anesthesia, the whole mouse brain was quickly removed and rinsed with a mixture of ice-cold artificial cerebrospinal fluid (aCSF). The brain was then placed in the aCSF frozen into a mashed potato and was sliced using a vibratome to a thickness of 300 μm. The left and right halves of the brain slice were separated and transferred to aCSF (30 °C) that was continuously bubbled with a mixture of oxygen and carbon dioxide. After 1-2 h of oxygen adaptation, brain slices were selected and placed in a 64-electrode array (each electrode measuring 50 × 50 μm and with 8 × 8 microelectrodes at the bottom), and the position was adjusted under a microscope to ensure the CA3 and CA1 regions of the hippocampus were on the electrode contacts. The brain slice was then gently anchored and connected to the MED64 system (Alpha MED Sciences, Panasonic), and perfused with the oxygen and artificial cerebrospinal fluid at a rate of 1-2 ml/min. Local excitatory postsynaptic potentials (EPSP) of CA1 neurons were recorded by stimulating the mossy fibers of CA3, and stable recording was acquired for approximately 40 min. LTP was induced using three sets of high-frequency stimulation (100 Hz, 1 s per set) and was recorded for approximately 70 min. Five animals were recorded for each group, with approximately six brain slices recorded for each animal, and statistical analysis were performed with the collected data. Open field test (OFT) Animals were placed individually in a 45 × 45 × 45 cm3 white opaque box (cleaned with a 75% alcohol cotton ball when switching animals) at room temperature (25 °C) and allowed to acclimate for 5 min. They were then allowed to move freely inside the box for 5 min, and their movements were monitored using a video system (HVS Imagen, UK). The bottom of the box was divided into 5 × 5 grids to determine the exploratory and anxiety-related behaviors by total number of times that animals crossing each area. Data were collected and analyzed. Novel object recognition (NOR) After the open field test, a novel object recognition test was conducted on the following day in the same 45 × 45 × 45 cm3 white opaque box (which was cleaned with 75% alcohol swabs when replacing each animal) with 5 × 5 grids. On the first day of the test, two objects (A and B) with different shapes and colors (and no significant attractiveness difference) were placed at opposite corners of the box. The animal was placed alone in the box and allowed to explore for 5 min. The number of times the animal touched each object and the time spent in the grid containing each object were recorded. On the second day, the test was conducted with a novel object C replacing object A, which had a different shape and color but was not significantly different in attractiveness to mice. The animal was placed alone in the box and allowed to explore for 5 min. The number of times the animal touched each object and the time spent in the grid containing each object were recorded. The novel object recognition index was calculated as follows: the time spent exploring the novel object on day two divided by the total time spent exploring both objects, or the number of touches with the novel object on day two divided by the total number of touches with both objects. The results were then analyzed to assess the animal’s short-term spatial learning and memory abilities. Barnes maze Barnes Maze is a circular platform made of organic board that can rotate. It has a diameter of 122 cm and 18 equidistant circular holes with a diameter of 10 cm around the platform edge. One of the holes (called the target hole) is connected to a dark box (called the target box), while the other circular holes are empty and not connected to any object. The target box is designed as a drawer for easy access to the animal inside. The target box is invisible from the platform surface, and the maze is elevated 140 cm from the ground. The animal can escape to the target box through the target hole. One day before the experiment, the animal is placed individually in the target box through the target hole for 4 min to adapt. The animal is then confined to a plastic circular barrel (diameter 20 cm, height 27 cm) in the center of the maze for 5 s. After removing the barrel, the timer is started, and the observer records the animal’s behavior using a video system. If all four limbs of the animal enter the target box, it is considered an escape, and the animal is left in the box for 30 s. Each animal was observed for up to 3 min, and if the animal cannot find the target box within this period, it is removed from the maze and placed in the target box for 30 s. The maze is cleaned between tests of two animals. Animals are trained twice a day for four consecutive days. From the second training, the maze is randomly rotated by one to several holes with the target box remained in the same position to prevent animals from relying on odor instead of memory to locate the target hole. The following parameters are recorded during the experiment: the latency to explore any hole, the latency to reach the target box, and the number of errors made by each animal (one error is defined as the animal stretching its head towards or exploring any non-target hole, including focusing on the same non-target hole). Statistical analysis In cell-based and animal experiments, the experiment was performed by members who were blinded to the sample information. In animal experiments, the information was opened after analyzing all samples. Sample size was determined by Power and Precision (Biostat). All data are expressed as mean ± SD and analyzed using GraphPad Prism statistical software (GraphPad Software). Non-paired t-test with Welch correction was used to evaluate the significance level between two groups. One-way ANOVA and Bonferroni’s multiple comparison test were applied when there were more than two groups. Two-way ANOVA and Bonferroni’s post-hoc test were used to compare differences between groups divided by two independent factors. A value of P < 0.05 was considered statistically significant. Results Npas4 binds to the promoter regions of Neuroligin-1 and N-cadherin and regulates their expression Research has shown that Npas4 plays an important role in regulating neuronal synaptic plasticity by coordinating the expression of multiple target genes during neuronal excitatory response [22, 23], while Neuroligin-1 and N-cadherin are essential for regulating the morphology and function of neuronal synapses. Previous genome-wide chromatin immunoprecipitation sequencing (ChIP-seq) analysis revealed strong enrichment of Npas4 binding in the upstream regions near the transcription start sites of the Nlgn1 and Cdh2 genes [26]. To clarify the transcriptional regulatory role of Npas4 on Neuroligin-1 and N-cadherin, Npas4 was overexpressed in cultured SH-SY5Y cells. Results showed that overexpression of Npas4 increased the protein levels of Neuroligin-1 and N-cadherin (Figs. 1A, B), and qPCR analysis showed that Npas4 overexpression also increased the mRNA levels of NLGN1 and CDH2 (Fig. 1C). We found that the DNA regions strongly enriched for Npas4 binding all contained cacga or acgtg gene sequences [26], and subsequently identified three such sequence locations near the transcription start sites of NLGN1 and CDH2 genes and designed primers based on the gene sequences (Fig. 1D). ChIP analysis showed that Npas4 bound to the cacga gene sequences in the upstream promoter regions of NLGN1 and CDH2 genes and that this binding was enhanced after Npas4 overexpression (Fig. 1E-G), indicating that Npas4 can promote the transcription and expression of NLGN1 and CDH2 genes by binding to their upstream promoter regions.Fig. 1Npas4 regulates the expression of NLGN1 and CDH2 in SH-SY5Y cells.A, B SH-SY5Y cells were transfected with Npas4 plasmid, and expression of the proteins in cells 48 h later were detected by western blot. (n = 3 for each group; Three times the experiments were repeated). C The mRNA levels of NPAS4, NLGN1, and CDH2 in Npas4-overexpressed SH-SY5Y cells were evaluated by qPCR. (n = 6 for each group; Three times the experiments were repeated). D Schematic representation and localization of Npas4 binding motif across the NLGN1 and CDH2 loci indicating the 3 amplicons (A, B, and C) used for ChIP-qPCR or ChIP-PCR analysis. E ChIP-qPCR analysis of Npas4 localization at the NLGN1 and CDH2 loci. Npas4-overexpressed SH-SY5Y cells were subjected to ChIP analysis with an anti-Npas4 antibody. (Three times the experiments were repeated). F, G PCR and analysis for the ChIP assay to confirm the transcriptional binding motif. Error bars denote standard deviations for 3 independent PCR reactions from a single ChIP assay that was representative of several that were performed. Mean ± SD, *P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 vs. Control. Student’s t test between two groups. Npas4 regulates the expression of NLGN1 and CDH2 in SH-SY5Y cells. A, B SH-SY5Y cells were transfected with Npas4 plasmid, and expression of the proteins in cells 48 h later were detected by western blot. (n = 3 for each group; Three times the experiments were repeated). C The mRNA levels of NPAS4, NLGN1, and CDH2 in Npas4-overexpressed SH-SY5Y cells were evaluated by qPCR. (n = 6 for each group; Three times the experiments were repeated). D Schematic representation and localization of Npas4 binding motif across the NLGN1 and CDH2 loci indicating the 3 amplicons (A, B, and C) used for ChIP-qPCR or ChIP-PCR analysis. E ChIP-qPCR analysis of Npas4 localization at the NLGN1 and CDH2 loci. Npas4-overexpressed SH-SY5Y cells were subjected to ChIP analysis with an anti-Npas4 antibody. (Three times the experiments were repeated). F, G PCR and analysis for the ChIP assay to confirm the transcriptional binding motif. Error bars denote standard deviations for 3 independent PCR reactions from a single ChIP assay that was representative of several that were performed. Mean ± SD, *P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001 vs. Control. Student’s t test between two groups. For accuracy, we further studied the effect of Npas4 on some other synaptic-associated molecules, including PSD-95, synaptophysin and synapsin-1. The result showed that Npas4 overexpression didn’t alter the protein levels of PSD-95, synaptophysin, or synapsin-1 in primary neuron (Supplementary Fig. 1). Early-stage cerebral ischemia results in upregulation of Npas4, Neuroligin, and N-cadherin expression, along with enhanced long-term potentiation (LTP) During cerebral ischemia, abnormalities appear in the activity of neurons, but the pattern and mechanism of these changes are unclear. To investigate the regulatory role of Npas4 in neuronal activity during cerebral ischemia, we used a Middle cerebral artery obstruction (MCAO) model to induce brain ischemia in C57BL/6 mice, and took brain tissue from the MCAO group and the sham group at 1, 2, and 6 h after ischemia. We measured the mRNA levels of Npas4, Nlgn1, and Cdh2 in the hippocampal tissue of the mice. The results showed that compared to the sham group, the mRNA level of Npas4 increased at 1 h after ischemia and gradually returned to the same level as the control group at 2 and 6 h after ischemia. The mRNA levels of Nlgn1 and Cdh2 showed a similar pattern of change to Npas4, increasing at 1 h after ischemia and gradually returning to the same level as the control group at 2 and 6 h after ischemia (Fig. 2A). This suggests that during cerebral ischemia, Npas4 may play a transcriptional regulatory role in the gene expression of Nlgn1 and Cdh2, showing a consistent pattern of change.Fig. 2The expression level of Npas4, Neuroligin-1, and N-cadherin after cerebral ischemia.A The mRNA levels of Npas4, Nlgn1, and Cdh2 in the hippocampus of mice with sham operation (Ctrl) and mice accepted middle cerebral artery occlusion for 1, 2 and 6 h (MCAO 1 h, MCAO 2 h, MCAO 6 h) mice were analyzed by qPCR. (n = 6 for each group, three times the experiments were repeated; Mean ± SD, *** P < 0.001 vs. MCAO 1 h group. one-way ANOVA and Bonferroni’s multiple comparison test among four groups). B The protein level and localization of Npas4 and Neuroligin-1 in the hippocampus CA1 and CA3 areas of mice with sham operation (Ctrl) and mice accepted middle cerebral artery occlusion for one hour (MCAO) were detected by immunofluorescence. C The protein level and localization of Npas4 and N-cadherin in the hippocampus CA1 and CA3 areas of mice with sham operation (Ctrl) and mice accepted middle cerebral artery occlusion for one hour (MCAO) were detected by immunofluorescence. D–F Integrated density of Npas4 (D), Neuroligin-1 (E), and N-cadherin (F) in the CA1 and CA3 areas of immunofluorescence image were analyzed by ImageJ. (n = 5 for each group, three times the experiments were repeated; Mean ± SD, ** P < 0.01, **** P < 0.0001 vs. MCAO group, unpaired Student’s t test between two groups). The expression level of Npas4, Neuroligin-1, and N-cadherin after cerebral ischemia. A The mRNA levels of Npas4, Nlgn1, and Cdh2 in the hippocampus of mice with sham operation (Ctrl) and mice accepted middle cerebral artery occlusion for 1, 2 and 6 h (MCAO 1 h, MCAO 2 h, MCAO 6 h) mice were analyzed by qPCR. (n = 6 for each group, three times the experiments were repeated; Mean ± SD, *** P < 0.001 vs. MCAO 1 h group. one-way ANOVA and Bonferroni’s multiple comparison test among four groups). B The protein level and localization of Npas4 and Neuroligin-1 in the hippocampus CA1 and CA3 areas of mice with sham operation (Ctrl) and mice accepted middle cerebral artery occlusion for one hour (MCAO) were detected by immunofluorescence. C The protein level and localization of Npas4 and N-cadherin in the hippocampus CA1 and CA3 areas of mice with sham operation (Ctrl) and mice accepted middle cerebral artery occlusion for one hour (MCAO) were detected by immunofluorescence. D–F Integrated density of Npas4 (D), Neuroligin-1 (E), and N-cadherin (F) in the CA1 and CA3 areas of immunofluorescence image were analyzed by ImageJ. (n = 5 for each group, three times the experiments were repeated; Mean ± SD, ** P < 0.01, **** P < 0.0001 vs. MCAO group, unpaired Student’s t test between two groups). However, mRNA levels often do not reflect the actual expression level of a protein, so we also measured the protein levels of Npas4, Neuroligin-1, and N-cadherin. Firstly, immunofluorescence results showed that after 1 h of MCAO, the fluorescence intensity of Npas4, Neuroligin-1, and N-cadherin proteins in the cortex and hippocampal CA1, CA3 and DG regions of the ischemic side was increased compared to the sham group, and the co-localization of Npas4 with Neuroligin-1 and N-cadherin was significantly enhanced (Fig. 2B-F, Supplementary Fig. 2), which again reflects the high correlation between Npas4, Neuroligin-1, and N-cadherin in protein expression. Western blot results also showed that after 1 h of MCAO, the protein levels of Npas4, Neuroligin-1, and N-cadherin in the hippocampal tissue of the ischemic side were upregulated (Fig. 3A, B).Fig. 3Synaptic plasticity in the early stage of cerebral ischemia.A, B The protein level of Npas4, Neuroligin-1, and N-cadherin in the hippocampus of mice with sham operation (Ctrl) and mice suffered middle cerebral artery occlusion for one hour (MCAO) were detected by western blot. (Mean ± SD; n = 9 in each group; Three times the experiments were repeated; *P < 0.05, **P < 0.01, **** P < 0.0001 MCAO compared with control, unpaired Student’s t test between two groups). C Electrophysiology analysis. MCAO enhanced the LTP measured by fEPSP in C57BL/6 mice (Mean ± SD; n = 4 in each group; Three times the experiments were repeated; **P < 0.01 MCAO compared with control, two-way ANOVA and Bonferroni’s post hoc test). Shown traces were representative fEPSPs of 4 samples recorded before and after 3 traces of HFS. D Representative recording traces: dark blue line, baseline; light blue line, LTP 90–100 min. E Quantification of fEPSP amplitude from the final 10 min of recordings (90–100 min in Fig. 3C) normalized to basal levels (Mean ± SD; n = 4 in each group; Three times the experiments were repeated; **P < 0.01 MCAO compared with control, unpaired Student’s t test between two groups). Synaptic plasticity in the early stage of cerebral ischemia. A, B The protein level of Npas4, Neuroligin-1, and N-cadherin in the hippocampus of mice with sham operation (Ctrl) and mice suffered middle cerebral artery occlusion for one hour (MCAO) were detected by western blot. (Mean ± SD; n = 9 in each group; Three times the experiments were repeated; *P < 0.05, **P < 0.01, **** P < 0.0001 MCAO compared with control, unpaired Student’s t test between two groups). C Electrophysiology analysis. MCAO enhanced the LTP measured by fEPSP in C57BL/6 mice (Mean ± SD; n = 4 in each group; Three times the experiments were repeated; **P < 0.01 MCAO compared with control, two-way ANOVA and Bonferroni’s post hoc test). Shown traces were representative fEPSPs of 4 samples recorded before and after 3 traces of HFS. D Representative recording traces: dark blue line, baseline; light blue line, LTP 90–100 min. E Quantification of fEPSP amplitude from the final 10 min of recordings (90–100 min in Fig. 3C) normalized to basal levels (Mean ± SD; n = 4 in each group; Three times the experiments were repeated; **P < 0.01 MCAO compared with control, unpaired Student’s t test between two groups). Long-term potentiation (LTP) of field excitatory postsynaptic potentials (fEPSPs) in the hippocampal CA1 region reflects synaptic plasticity of neurons, and is the molecular basis of learning and memory. The hippocampal CA1 region is also one of the brain areas most susceptible to the effects of cerebral ischemia. In order to clarify the changes in neuronal synaptic plasticity when the protein levels of Npas4, Neuroligin-1, and N-cadherin increase after 1 h of cerebral ischemia, we used a 64-channel electrode to detect the LTP of CA1 neurons in the hippocampus 1 h after MCAO. Comparing to the sham group, the average fEPSP slope of CA1 neurons in the ischemic side of hippocampus was increased, indicating an enhancement of synaptic transmission and plasticity in CA1 neurons after 1 h of cerebral ischemia in mice (Fig. 3C-E). Npas4 upregulates neuronal plasticity by increasing Neuroligin-1 and N-cadherin expression in the early stage of cerebral ischemia Numerous studies have shown that Npas4, Neuroligin-1, and N-cadherin can all regulate neuronal excitability and synaptic plasticity. However, it is unclear whether the enhancement of synaptic transmission and plasticity in the hippocampal CA1 neurons in the early stage of cerebral ischemia is related to the increased protein expression levels of Npas4, Neuroligin-1, and N-cadherin. To investigate this issue, we injected interfering viruses targeting Npas4, Neuroligin-1, and N-cadherin into the hippocampal area of C57BL/6 mice. First, protein immunoblotting results showed that downregulation of Npas4 led to a decrease in the protein levels of Neuroligin-1 and N-cadherin (Fig. 4B). In addition, downregulation of Neuroligin-1 or N-cadherin did not affect the protein expression levels of Npas4 or the other synaptic adhesion molecule (Fig. 4C and D). These findings suggest that Npas4 regulates the expression of Neuroligin-1 and N-cadherin through transcriptional control, but Neuroligin-1 and N-cadherin as downstream regulatory targets do not affect each other’s expression levels, nor do they affect the upstream transcription factor Npas4.Fig. 4Effects of Npas4 transcriptional regulation of Neuroligin-1 and N-cadherin expression on synaptic plasticity in early stage of cerebral ischemia.A Experimental design for LV treated C57BL/6 mice. B-G The protein level of Npas4, Neuroligin-1, and N-cadherin after knock down (Sh) of Npas4 (B), Neuroligin-1 (C), and N-cadherin (D) respectively in the hippocampus of C57BL/6 mice were detected by western blot with grayscale analysis (E-G). Mean ± SD; n = 6 in each group; Three times the experiments were repeated; ***P < 0.001, ****P < 0.0001 Sh compared with Vector group, Student’s t test between two groups. H-J Electrophysiology analysis. LTP of fEPSPs after sham operation (Ctrl), or accepted middle cerebral artery occlusion for one hour (MCAO) in the hippocampus of C57BL/6 mice with Npas4 (H), Neuroligin-1 (I), or N-cadherin (J) knock down (Mean ± SD; n = 4 in vector-Ctr and vector-MACO groups, n = 4 in all the other groups ; Three times the experiments were repeated; **P < 0.01 MCAO compared with control, two-way ANOVA and Bonferroni’s post hoc test). Shown traces were representative fEPSPs of 4 samples recorded before and after 3 traces of HFS (High frequency stimulation). K Quantification of fEPSP potentiation from the final 10 min of recordings (90–100 min in Fig. 4H-J) normalized to basal levels (Mean ± SD; n = 4 in each group; Three times the experiments were repeated; **P < 0.01, ***P < 0.001 compared with Sh-Vector+MCAO group, one-way ANOVA and Bonferroni’s post hoc test in 8 groups). L Representative recording traces: dark blue line, baseline; light blue line, LTP 90–100 min in 8 groups. Effects of Npas4 transcriptional regulation of Neuroligin-1 and N-cadherin expression on synaptic plasticity in early stage of cerebral ischemia. A Experimental design for LV treated C57BL/6 mice. B-G The protein level of Npas4, Neuroligin-1, and N-cadherin after knock down (Sh) of Npas4 (B), Neuroligin-1 (C), and N-cadherin (D) respectively in the hippocampus of C57BL/6 mice were detected by western blot with grayscale analysis (E-G). Mean ± SD; n = 6 in each group; Three times the experiments were repeated; ***P < 0.001, ****P < 0.0001 Sh compared with Vector group, Student’s t test between two groups. H-J Electrophysiology analysis. LTP of fEPSPs after sham operation (Ctrl), or accepted middle cerebral artery occlusion for one hour (MCAO) in the hippocampus of C57BL/6 mice with Npas4 (H), Neuroligin-1 (I), or N-cadherin (J) knock down (Mean ± SD; n = 4 in vector-Ctr and vector-MACO groups, n = 4 in all the other groups ; Three times the experiments were repeated; **P < 0.01 MCAO compared with control, two-way ANOVA and Bonferroni’s post hoc test). Shown traces were representative fEPSPs of 4 samples recorded before and after 3 traces of HFS (High frequency stimulation). K Quantification of fEPSP potentiation from the final 10 min of recordings (90–100 min in Fig. 4H-J) normalized to basal levels (Mean ± SD; n = 4 in each group; Three times the experiments were repeated; **P < 0.01, ***P < 0.001 compared with Sh-Vector+MCAO group, one-way ANOVA and Bonferroni’s post hoc test in 8 groups). L Representative recording traces: dark blue line, baseline; light blue line, LTP 90–100 min in 8 groups. Based on these results, we further investigated changes in neuronal activity after cerebral ischemia in mice with downregulated expression of Npas4, Neuroligin-1, or N-cadherin. Accordingly, the average fEPSP slope of CA1 neurons in the ischemic hippocampal tissue with downregulated Npas4, Neuroligin-1, or N-cadherin was lower than that in the Sh-Vector group. In each of the expression interference groups, the average fEPSP slope of CA1 neurons in the ischemic hippocampus tended to increase but without significant differences compared to the corresponding sham group (Fig. 4H-J). Moreover, compared with the Sh-Vector-Ctrl group, the Sh-Npas4-Ctrl, Sh-Nlgn1-Ctrl, and Sh-Cdh2-Ctrl groups showed a trend of decrease in the average fEPSP slope of CA1 neurons in the hippocampal tissue, but without significant differences (Fig. 4K). These findings suggest that iLTP in CA1 neurons of C57BL/6 mice is dependent on the increased expression of Npas4, and it is likely that the increased expression of Npas4 promotes the transcriptional expression of Neuroligin-1 and N-cadherin, thereby enhancing neuronal synaptic transmission and plasticity. Npas4 promotes synaptic vesicle formation in neurons by upregulating the expression of Neuroligin-1 and N-cadherin It is noteworthy that research has found that both the formation of neuronal synapses and the formation of synaptic vesicles during synaptic transmission depend on the collaborative interaction between N-cadherin and Neuroligin-1 [52]. Moreover, Neuroligin-1 promotes the frequency and release probability of mEPSCs (miniature excitatory postsynaptic potentials) at mature synapses, which requires the coordination of N-cadherin [53]. To further confirm the above findings on neuronal morphology and structure, we investigated the effects of Npas4 transcriptional regulation of Neuroligin-1 and N-cadherin on neuronal synaptic vesicle formation in cultured mouse primary neurons (Fig. 5A). Results showed that compared to the control group, overexpression of Npas4 significantly increased the number of presynaptic vesicles labeled with VAMP2 (Vesicle-associated membrane protein 2 or Synaptic vesicle protein synaptobrevin 2) on the neurites of the transfected primary neurons. However, knocking down Neuroligin-1 or N-cadherin in Npas4 overexpressed neurons significantly reduced the number of synaptic vesicles labeled with VAMP2 (Fig. 5B, C). We also measured the dendritic spine number by colocalization of EGFP and PSD-95 staining in neuron dendritic, the result showed that compared to the control group, overexpression of Npas4 significantly increased the dendritic spine number, knocking down Neuroligin-1 or N-cadherin in Npas4 overexpressed neurons significantly reduced the number of dendritic spines (Supplementary Fig. 3). Protein immunoblotting results once again confirmed that Npas4 overexpression in cultured primary neurons promoted the upregulation of Neuroligin-1 and N-cadherin expression, while knocking down Neuroligin-1 expression with interference RNA did not affect Npas4 and N-cadherin expression, and knocking down N-cadherin expression with interference RNA did not affect Npas4 and Neuroligin-1 expression (Fig. 5D, E). These results indicate that Npas4 promotes synaptic vesicle formation in neurons by upregulating the expression of Neuroligin-1 and N-cadherin, and that Npas4-mediated synaptic vesicle formation depends on the co-expression of Neuroligin-1 and N-cadherin.Fig. 5Effects of Npas4 transcriptional regulation of Neuroligin-1 and N-cadherin expression on synaptic vesicle clustering in cultured primary neurons.A Experimental design for primary neurons treated with LV and siRNA. BMouse primary neurons at DIV7(transfected with Ctrl-EGFP or Npas4-EGFP LV at 4 DIV, and transfected with Ctrl-siRNA or Nlgn1-siRNA or Cdh2-siRNA at 5 DIV) were stained with VAMP2 to show synaptic vesicle clustering on the axon. (Scale bars: 10μm). C Quantification of the presynaptic vesicle clusters. (Mean ± SD; n = 5 in each group; Three times the experiments were repeated; **P < 0.01 compared with Npas4-EGFP+Ctrl siRNA group, one-way ANOVA and Bonferroni’s post hoc test among 4 groups). D-E The protein level of Npas4, Neuroligin-1, and N-cadherin after NPAS4 overexpression and knock down of Neuroligin-1 or N-cadherin respectively in primary neurons were detected by western blot with grayscale analysis (D). Mean ± SD; n = 4 in each group; Three times the experiments were repeated; *P < 0.05, **P < 0.01 compared with Ctrl-EGFP+Ctrl-siRNA group, ###P < 0.001, ####P < 0.0001 compared with Npas4-EGFP+Ctrl siRNA group, one-way ANOVA and Bonferroni’s post hoc test among 4 groups. Effects of Npas4 transcriptional regulation of Neuroligin-1 and N-cadherin expression on synaptic vesicle clustering in cultured primary neurons. A Experimental design for primary neurons treated with LV and siRNA. BMouse primary neurons at DIV7(transfected with Ctrl-EGFP or Npas4-EGFP LV at 4 DIV, and transfected with Ctrl-siRNA or Nlgn1-siRNA or Cdh2-siRNA at 5 DIV) were stained with VAMP2 to show synaptic vesicle clustering on the axon. (Scale bars: 10μm). C Quantification of the presynaptic vesicle clusters. (Mean ± SD; n = 5 in each group; Three times the experiments were repeated; **P < 0.01 compared with Npas4-EGFP+Ctrl siRNA group, one-way ANOVA and Bonferroni’s post hoc test among 4 groups). D-E The protein level of Npas4, Neuroligin-1, and N-cadherin after NPAS4 overexpression and knock down of Neuroligin-1 or N-cadherin respectively in primary neurons were detected by western blot with grayscale analysis (D). Mean ± SD; n = 4 in each group; Three times the experiments were repeated; *P < 0.05, **P < 0.01 compared with Ctrl-EGFP+Ctrl-siRNA group, ###P < 0.001, ####P < 0.0001 compared with Npas4-EGFP+Ctrl siRNA group, one-way ANOVA and Bonferroni’s post hoc test among 4 groups. Npas4, Neuroligin-1 and N-cadherin are downregulated in the brains of individuals with Alzheimer’s disease and mouse models Synaptic plasticity of neurons is the molecular basis of learning and memory, and synaptic vesicle formation is the structural basis of learning and memory. There are also studies showing that Npas4, Neuroligin-1, and N-cadherin are related to cognitive function. However, it is currently unclear how these three molecules change during the course of AD pathology and whether their changes are involved in the progression of AD. To investigate the changes of Npas4, Neuroligin-1, and N-cadherin in AD, we first referred to the transcriptome sequencing data of multiple brain regions of AD patients and normal individuals in the AlzData database (http://www.alzdata.org/index.html) [54]. The results showed that compared to normal brains, the mRNA level of NPAS4 was significantly downregulated in the temporal cortex of AD patients (Fig. 6A), the mRNA level of NLGN1 was significantly downregulated in the temporal cortex and hippocampus of AD patients (Fig. 6B, C), and the mRNA level of CDH2 was significantly downregulated in the frontal cortex (Fig. 6D), suggesting a decreased expression of Npas4, Neuroligin-1, and N-cadherin in multiple brain regions of AD patients. We also wonder these changes in human AD brain are age- or sex-dependent manner. In order to throw light on this, we conducted further analyses by using the transcriptome data of human AD brain, results showed that expression level of NPAS4 in temporal cortex is reduced in AD-male compared to Control-male cohort, but not in female; expression level of NLGN1 both in temporal cortex and hippocampus are reduced in AD-female compared to Control-female cohort, but not in male; change of CDH2 level in frontal cortex is in age-independent manner; changes of NPAS4, NLGN1, and CDH2 levels are in sex-independent manner (Supplementary Fig. 4).Fig. 6Npas4, Neuroligin-1, and N-cadherin levels in AD human and 3 × TG mice brain.A-D Cross-platform normalized expression levels of NPAS4, NLGN1, and CDH2 in control and AD brains, all differential expression results were adjusted for age and sex of samples. Data represent mean ± SD (two-tailed student’s t test). A Dataset of NPAS4 in Temporal Cortex: GSE37263, GSE36980, GSE5281. B Dataset of NLGN1 in Temporal Cortex: GSE15222, GSE37263, GSE36980, GSE5281. C Dataset of NLGN1 in Hippocampus: GSE28146, GSE29378, GSE36980, GSE48350, GSE5281. D Dataset of CDH2 in Frontal Cortex: GSE33000. E-J Protein level of Npas4, Neuroligin-1, and N-cadherin in the hippocampus of 3-month (E and F), 6-month (G and H), and 9-month (I and J) old wild-type control and 3 × TG mice were detected by western blot with grayscale analysis. Mean ± SD; n = 6 in each group; Three times the experiments were repeated; **P < 0.01, ***P < 0.001 3 × TG compared with wild-type control, Student’s t test between 2 groups. Npas4, Neuroligin-1, and N-cadherin levels in AD human and 3 × TG mice brain. A-D Cross-platform normalized expression levels of NPAS4, NLGN1, and CDH2 in control and AD brains, all differential expression results were adjusted for age and sex of samples. Data represent mean ± SD (two-tailed student’s t test). A Dataset of NPAS4 in Temporal Cortex: GSE37263, GSE36980, GSE5281. B Dataset of NLGN1 in Temporal Cortex: GSE15222, GSE37263, GSE36980, GSE5281. C Dataset of NLGN1 in Hippocampus: GSE28146, GSE29378, GSE36980, GSE48350, GSE5281. D Dataset of CDH2 in Frontal Cortex: GSE33000. E-J Protein level of Npas4, Neuroligin-1, and N-cadherin in the hippocampus of 3-month (E and F), 6-month (G and H), and 9-month (I and J) old wild-type control and 3 × TG mice were detected by western blot with grayscale analysis. Mean ± SD; n = 6 in each group; Three times the experiments were repeated; **P < 0.01, ***P < 0.001 3 × TG compared with wild-type control, Student’s t test between 2 groups. We then further detected the protein levels of Npas4, Neuroligin-1, and N-cadherin in the hippocampus of 3, 6, and 9-month-old 3 × Tg AD mice. The results showed that compared to the wild-type control mice of the same age, there were no significant changes in the protein levels of Npas4, Neuroligin-1, and N-cadherin in the hippocampus of 3-month-old 3 × Tg AD mice (Fig. 6E, F). However, in the hippocampus of 6-month and 9-month-old 3 × Tg AD mice, the protein levels of Npas4, Neuroligin-1, and N-cadherin were significantly decreased (Fig. 6G-J), indicating that the protein expression levels of Npas4, Neuroligin-1, and N-cadherin in the hippocampus of 3 × Tg AD mice decreased with age. Studies have revealed that compared to wild-type control mice of the same age, 3 × Tg AD mice showed reduced LTP and impaired basal synaptic transmission at the age of 6 months [55], which is consistent with the decrease in protein expression levels of Npas4, Neuroligin-1, and N-cadherin that we detected. This suggests that the decrease in protein expression levels of Npas4, Neuroligin-1, and N-cadherin in the hippocampus of 3 × Tg AD mice may be involved in mediating synaptic plasticity damage in AD. Overexpression of Npas4 upregulates Neuroligin-1 and N-cadherin and alleviates AD-like pathology in 3 × Tg mice To further clarify the regulatory role of Npas4 in the transcriptional control of Neuroligin-1 and N-cadherin in the AD process and their modulation of synaptic function and animal cognition, we performed brain stereotaxic injection of virus to overexpress Npas4 in the hippocampus of 5-month-old 3 × Tg AD mice (Fig. 7A). At 6 months of age, Western blot showed that compared with the empty vector control group, the protein levels of Neuroligin-1 and N-cadherin increased with the overexpression of Npas4 in the hippocampus of 3 × Tg AD mice (Fig. 7B, C). qPCR results also showed that the mRNA levels of Nlgn1 and Cdh2 increased with the overexpression of Npas4 (Fig. 7D), indicating that Npas4 overexpression promoted the transcriptional regulation of Nlgn1 and Cdh2 in 3 × Tg AD mice.Fig. 7Regulation of Npas4 on synaptic plasticity and cognitive function in 3 × Tg AD mice.A Experimental design for AAV treated 3 × TG mice. B-C Protein level of Npas4, Neuroligin-1, and N-cadherin after Npas4 over expressed in the hippocampus of 3 × Tg AD mice were detected by western blot (B) with grayscale analysis (C). Mean ± SD; n = 6 in each group; Three times the experiments were repeated; **P < 0.01, ****P < 0.0001 compared with Vector group, Student’s t test between 2 groups. D The mRNA levels of Npas4, Nlgn1, and Cdh2 in Npas4 over expressed hippocampus of 3 × Tg AD mice were analyzed by qPCR. Mean ± SD; n = 6 in each group; Three times the experiments were repeated; ****P < 0.0001 compared with Vector group, Student’s t test between 2 groups. E Zone crossing times were recorded in the Open field test to analyze exploration and anxiety behavior after Npase4 overexpression in 3 × Tg AD mice. Mean ± SD; n = 7 in each group; compared with Vector group, Student’s t test between 2 groups. F Recognition index was calculated (Times to explore novel object divided by times to explore both objects) during training (Day1) and testing (Day2) in Novel object recognition test to analyze recognition memory after Npase4 over expressed in 3 × Tg AD mice. Mean ± SD; n = 7 in each group; **P < 0.01 compared with Vector group, Student’s t test between 2 groups. G-H Latency(G) and Error times (H) to enter the box in dark hole during the training (Day1-Day4) were recorded in the Barnes maze test to analyze spatial learning ability after Npase4 over expression in 3 × Tg AD mice. Mean ± SD; n = 7 in each group; *P < 0.05, ***P < 0.001, ****P < 0.0001 compared with Vector group, two-way ANOVA and Bonferroni’s post hoc test among 4 groups. I Correct ratio was calculated (Times to approach the target dark hole divided by the total times to approach all the holes) during the testing (Day5) of Barnes maze test to analyze spatial memory after Npase4 over expression in 3 × Tg AD mice. Mean ± SD; n = 7 in each group; **P < 0.01 compared with Vector group, Student’s t test between 2 groups. J The total times to approach all the holes during the testing (Day5) of Barnes maze test was recorded to make sure the mice in two groups with undifferentiated motor and exploration ability. Mean ± SD; n = 7 in each group; compared with Vector group, Student’s t test between 2 groups. K Electrophysiology analysis. Npas4 overexpression enhanced the LTP in 3 × Tg AD mice. LTP of fEPSPs. (Mean ± SD; n = 3 in each group; Three times the experiments were repeated; **P < 0.01 compared with Vector group, two-way ANOVA and Bonferroni’s post hoc test). Shown traces were representative fEPSPs of 3 samples recorded before and after 3 traces of HFS. L Representative recording traces, dark blue line, baseline; light blue line, LTP 90–100 min between 2 groups. M Quantification of fEPSP potentiation from the final 10 min of recordings (90–100 min in Fig. 7J) normalized to basal levels. Mean ± SD; n = 3 in each group; Three times the experiments were repeated; **P < 0.01 compared with Vector group, Student’s t test between 2 groups. Regulation of Npas4 on synaptic plasticity and cognitive function in 3 × Tg AD mice. A Experimental design for AAV treated 3 × TG mice. B-C Protein level of Npas4, Neuroligin-1, and N-cadherin after Npas4 over expressed in the hippocampus of 3 × Tg AD mice were detected by western blot (B) with grayscale analysis (C). Mean ± SD; n = 6 in each group; Three times the experiments were repeated; **P < 0.01, ****P < 0.0001 compared with Vector group, Student’s t test between 2 groups. D The mRNA levels of Npas4, Nlgn1, and Cdh2 in Npas4 over expressed hippocampus of 3 × Tg AD mice were analyzed by qPCR. Mean ± SD; n = 6 in each group; Three times the experiments were repeated; ****P < 0.0001 compared with Vector group, Student’s t test between 2 groups. E Zone crossing times were recorded in the Open field test to analyze exploration and anxiety behavior after Npase4 overexpression in 3 × Tg AD mice. Mean ± SD; n = 7 in each group; compared with Vector group, Student’s t test between 2 groups. F Recognition index was calculated (Times to explore novel object divided by times to explore both objects) during training (Day1) and testing (Day2) in Novel object recognition test to analyze recognition memory after Npase4 over expressed in 3 × Tg AD mice. Mean ± SD; n = 7 in each group; **P < 0.01 compared with Vector group, Student’s t test between 2 groups. G-H Latency(G) and Error times (H) to enter the box in dark hole during the training (Day1-Day4) were recorded in the Barnes maze test to analyze spatial learning ability after Npase4 over expression in 3 × Tg AD mice. Mean ± SD; n = 7 in each group; *P < 0.05, ***P < 0.001, ****P < 0.0001 compared with Vector group, two-way ANOVA and Bonferroni’s post hoc test among 4 groups. I Correct ratio was calculated (Times to approach the target dark hole divided by the total times to approach all the holes) during the testing (Day5) of Barnes maze test to analyze spatial memory after Npase4 over expression in 3 × Tg AD mice. Mean ± SD; n = 7 in each group; **P < 0.01 compared with Vector group, Student’s t test between 2 groups. J The total times to approach all the holes during the testing (Day5) of Barnes maze test was recorded to make sure the mice in two groups with undifferentiated motor and exploration ability. Mean ± SD; n = 7 in each group; compared with Vector group, Student’s t test between 2 groups. K Electrophysiology analysis. Npas4 overexpression enhanced the LTP in 3 × Tg AD mice. LTP of fEPSPs. (Mean ± SD; n = 3 in each group; Three times the experiments were repeated; **P < 0.01 compared with Vector group, two-way ANOVA and Bonferroni’s post hoc test). Shown traces were representative fEPSPs of 3 samples recorded before and after 3 traces of HFS. L Representative recording traces, dark blue line, baseline; light blue line, LTP 90–100 min between 2 groups. M Quantification of fEPSP potentiation from the final 10 min of recordings (90–100 min in Fig. 7J) normalized to basal levels. Mean ± SD; n = 3 in each group; Three times the experiments were repeated; **P < 0.01 compared with Vector group, Student’s t test between 2 groups. We also assessed changes in animal cognition and other behaviors through behavioral testing. Results from the open field test showed no significant differences in the number of times the animals moved between areas (Fig. 7E), indicating that overexpression of Npas4 did not affect the exploratory and anxiety-related behaviors of 3 × Tg AD mice. Results from the novel object recognition test showed that on the second day of the test, the overexpression of Npas4 increased the animals’ preference for exploring new objects compared to the empty vector control group (Fig. 7F), indicating that Npas4 overexpression improved the recognition memory of 3 × Tg AD mice. Results from the Barnes maze test showed that during the training period, the latency to find the escape hole in the overexpression group decreased from the third day of training compared to the empty vector control group (Fig. 7G), and the number of exploratory errors decreased relatively from the second day of training (Fig. 7H), indicating that after Npas4 overexpression, the spatial learning ability of 3 × Tg AD mice improved. During the memory test on the fifth day, the ratio of correct explorations of the escape hole significantly increased in the overexpression group (Fig. 7I), while the total number of hole explorations in both groups showed no statistical differences (Fig. 7J), indicating that after Npas4 overexpression, the spatial memory ability of 3 × Tg AD mice improved without affecting their exploration and motor functions. To clarify whether the improvement of cognitive function in 3 × Tg AD mice by Npas4 overexpression was achieved through the relief of synaptic functional damage in the brain, we conducted electrophysiological measurements of synaptic plasticity and transmission in the hippocampal neurons of animals. The results showed that compared with the empty vector control group, the average fEPSP slope and LTP of CA1 neurons in the hippocampal tissue of Npas4 overexpression group mice increased (Fig. 7K, M), indicating that Npas4 overexpression improved the spatial learning and memory ability of 3 × Tg AD animals by relieving synaptic functional damage in their brains. These results reveal that the deficiency of Npas4 in the brains of 3 × Tg AD animals leads to a decrease in the expression of key proteins regulating synaptic function, Neuroligin-1, and N-cadherin, causing impaired synaptic transmission, plasticity, and animal cognition. Therefore, overexpressing Npas4 in the hippocampus of AD animal brains can upregulate the expression of Neuroligin-1 and N-cadherin, leading to improved synaptic transmission ability and animal cognition. Lowering Npas4, Neuroligin-1, or N-cadherin separately can all lead to cognitive impairment in C57BL/6 mice In order to further verify the effects of Npas4, Neuroligin-1, and N-cadherin on cognitive function, the current study performed behavioral tests on C57BL/6 mice in which Npas4, Neuroligin-1, or N-cadherin was separately knocked down. The results of the open field test showed no significant differences in the number of crossings between the different areas of the animals (Fig. 8A), indicating that the reduced expression of Npas4, Neuroligin-1, or N-cadherin did not affect the exploration and anxiety-related behavior of C57BL/6 mice. The results of the novel object recognition test showed that on the second day of the test, compared with the empty vector control group, the reduced expression of Npas4, Neuroligin-1, or N-cadherin all led to a reduction in the animals’ preference for exploring new objects to varying degrees (Fig. 8B), indicating that the reduced expression of Npas4, Neuroligin-1, or N-cadherin all impaired the recognition memory of C57BL/6 mice. The results of the Barnes maze test showed that during the training period, the reduced expression of Npas4, Neuroligin-1, or N-cadherin all led to a decrease in the latency to find the escape hole compared to the empty vector control group (Fig. 8C), and the number of exploratory errors on the fourth day of training was relatively reduced (Fig. 8D), indicating that the spatial learning ability of C57BL/6 mice was impaired after the reduced expression of Npas4, Neuroligin-1, or N-cadherin. On the fifth day of the memory test, the ratio of correct explorations of the escape hole by the mice with reduced expression of Npas4, Neuroligin-1, or N-cadherin was significantly decreased (Fig. 8E, F), while the total number of holes explored in the two groups showed no statistical differences (Fig. 8E, G), indicating that after the reduced expression of Npas4, Neuroligin-1, or N-cadherin, the spatial memory of C57BL/6 mice was impaired without affecting their exploration and motor function. Among all the above testing indices, the knockdown of Neuroligin-1 or N-cadherin separately could affect animal cognitive function, while the knockdown of Npas4 had the most significant impact on cognitive function, indicating that the knockdown of Npas4 not only affected the learning and memory ability of animals by simultaneously reducing the expression of Neuroligin-1 and N-cadherin, but may also exacerbate animal cognitive impairment through other mechanisms, which requires further investigation.Fig. 8Regulation of Npas4, Neuroligin-1 or N-cadherin on cognitive function in C57BL/6 mice.A Tracking and Zone crossing times were recorded in the Open field test to analyze exploration and anxiety-related behavior after Npas4, Neuroligin-1 or N-cadherin knock down in C57BL/6 mice. Mean ± SD; n = 10 in Vector, n = 12 in si-Npas4, n = 9 in si-Nlgn1, n = 11 in si-Cdh2 groups; compared with Vector group, one-way ANOVA and Bonferroni’s post hoc test among 4 groups. B Tracking during testing (Day2) was recorded, Recognition index was calculated (Times to explore novel object divided by times to explore both objects) during training (Day1) and testing (Day2) in Novel object recognition test to analyze recognition memory after Npas4, Neuroligin-1 or N-cadherin knock down in C57BL/6 mice. Mean ± SD; n = 10 in Vector, n = 12 in si-Npas4, n = 9 in si-Nlgn1, n = 11 in si-Cdh2 groups; *P < 0.05, ***P < 0.001, ****P < 0.0001 compared with Vector group, one-way ANOVA and Bonferroni’s post hoc test among 4 groups. C Latency(C) and Error times (D) to enter the box in dark hole during the training (Day1-Day4) were recorded in the Barnes maze test to analyze spatial learning ability after Npas4, Neuroligin-1 or N-cadherin knock down in C57BL/6 mice. Mean ± SD; n = 10 in Vector, n = 12 in si-Npas4, n = 9 in si-Nlgn1, n = 11 in si-Cdh2 groups; *P < 0.05, ***P < 0.001, ****P < 0.0001 compared with Vector group, one-way ANOVA and Bonferroni’s post hoc test among 4 groups. E Tracking during the testing (Day5) of Barnes maze test was recorded. F Correct ratio was calculated (Times to approach the target dark hole divided by the total times to approach all the holes) during the testing (Day5) of Barnes maze test to analyze spatial memory after Npas4, Neuroligin-1 or N-cadherin knock down in C57BL/6 mice. Mean ± SD; n = 10 in Vector, n = 12 in si-Npas4, n = 9 in si-Nlgn1, n = 11 in si-Cdh2 groups; *P < 0.05, **P < 0.01, ****P < 0.0001 compared with Vector group, one-way ANOVA and Bonferroni’s post hoc test among 4 groups. G The total times to approach all the holes during the testing (Day5) of Barnes maze test was recorded to make sure the mice in four groups with undifferentiated motor and exploration ability. Mean ± SD; n = 10 in Vector, n = 12 in si-Npas4, n = 9 in si-Nlgn1, n = 11 in si-Cdh2 groups; compared with Vector group, one-way ANOVA and Bonferroni’s post hoc test among 4 groups. Regulation of Npas4, Neuroligin-1 or N-cadherin on cognitive function in C57BL/6 mice. A Tracking and Zone crossing times were recorded in the Open field test to analyze exploration and anxiety-related behavior after Npas4, Neuroligin-1 or N-cadherin knock down in C57BL/6 mice. Mean ± SD; n = 10 in Vector, n = 12 in si-Npas4, n = 9 in si-Nlgn1, n = 11 in si-Cdh2 groups; compared with Vector group, one-way ANOVA and Bonferroni’s post hoc test among 4 groups. B Tracking during testing (Day2) was recorded, Recognition index was calculated (Times to explore novel object divided by times to explore both objects) during training (Day1) and testing (Day2) in Novel object recognition test to analyze recognition memory after Npas4, Neuroligin-1 or N-cadherin knock down in C57BL/6 mice. Mean ± SD; n = 10 in Vector, n = 12 in si-Npas4, n = 9 in si-Nlgn1, n = 11 in si-Cdh2 groups; *P < 0.05, ***P < 0.001, ****P < 0.0001 compared with Vector group, one-way ANOVA and Bonferroni’s post hoc test among 4 groups. C Latency(C) and Error times (D) to enter the box in dark hole during the training (Day1-Day4) were recorded in the Barnes maze test to analyze spatial learning ability after Npas4, Neuroligin-1 or N-cadherin knock down in C57BL/6 mice. Mean ± SD; n = 10 in Vector, n = 12 in si-Npas4, n = 9 in si-Nlgn1, n = 11 in si-Cdh2 groups; *P < 0.05, ***P < 0.001, ****P < 0.0001 compared with Vector group, one-way ANOVA and Bonferroni’s post hoc test among 4 groups. E Tracking during the testing (Day5) of Barnes maze test was recorded. F Correct ratio was calculated (Times to approach the target dark hole divided by the total times to approach all the holes) during the testing (Day5) of Barnes maze test to analyze spatial memory after Npas4, Neuroligin-1 or N-cadherin knock down in C57BL/6 mice. Mean ± SD; n = 10 in Vector, n = 12 in si-Npas4, n = 9 in si-Nlgn1, n = 11 in si-Cdh2 groups; *P < 0.05, **P < 0.01, ****P < 0.0001 compared with Vector group, one-way ANOVA and Bonferroni’s post hoc test among 4 groups. G The total times to approach all the holes during the testing (Day5) of Barnes maze test was recorded to make sure the mice in four groups with undifferentiated motor and exploration ability. Mean ± SD; n = 10 in Vector, n = 12 in si-Npas4, n = 9 in si-Nlgn1, n = 11 in si-Cdh2 groups; compared with Vector group, one-way ANOVA and Bonferroni’s post hoc test among 4 groups. Discussion Ischemic injuries, such as stroke, usually result in severe functional impairments that are permanent. Currently, there are no effective prevention or treatment strategies for these injuries, mainly because the specific process and mechanism of neuronal damage caused by ischemic injury are not clear [56]. Studies have shown that synaptic plasticity of neurons is enhanced during brain ischemia [11, 17], which may be a self-repair response of the brain after ischemic injury, but the specific mechanism is not yet clear. After brain ischemic injury, an early neuron activity-dependent gene, Npas4, is rapidly and strongly activated with evidently increased expression in the brain [17–19], but it is not clear whether the activation of Npas4 is related to the regulation of neuronal activity. We speculate that, during brain ischemia, as neurons are subjected to various complex stimuli, such as calcium influx and specific signal cascades activation, leading to nuclear translocation of various transcription factors and rapid induction of IEGs expression programs, which in turn induce the transcription of subsequent responsive genes and locally regulate synaptic transmission and plasticity. Npas4, as an IEG activated in this process, has also been shown in numerous studies to be a neuron activity-dependent transcription factor that regulates the expression of proteins critical to neuronal activity, including c-Fos, BDNF, Synaptotagmin-10, IQSEC3 [10, 22, 23, 57, 58]. Our research has clarified that synaptic plasticity of neurons is enhanced and Npas4 expression increases in the early stage of brain ischemia. Through in vitro studies, we have also demonstrated for the first time that Npas4 can promote the expression of Neuroligin-1 and N-cadherin by binding to the promoter regions of NLGN1 and CDH2. Moreover, in the early stage of brain ischemia, upregulation of Npas4 expression also promotes the expression of Neuroligin-1 and N-cadherin proteins. Our interference experiments have also confirmed that the enhancement of synaptic plasticity in the early stage of brain ischemia depends on the expression of Npas4, Neuroligin-1, and N-cadherin, which are all essential. Studies have shown that the cooperative action of Neuroligin-1 and N-cadherin promotes synaptic formation and transmission in neurons, and this cooperation is critical to neuronal discharge activity [52, 53]. Our research has also confirmed that overexpression of Npas4 can promote the formation of synaptic vesicles in primary cultured neurons, and Npas4 promotes the increase in synaptic vesicles by upregulating the expression of both Neuroligin-1 and N-cadherin. Therefore, based on our research, it can be concluded that in the early stage of brain ischemia, the upregulation of early gene Npas4 expression promotes the transcription of Neuroligin-1 and N-cadherin, and the cooperation of Neuroligin-1 and N-cadherin promotes the formation and transmission of neuronal synaptic vesicles, leading to the observed increase in synaptic plasticity. Whether the promotion of Npas4 expression is caused by neuronal activity promoting calcium influx and activation of specific signal cascades remains to be studied. Importantly, we know that increased synaptic plasticity of neurons is critical for regulating neural network function and individual behavior. Therefore, our research provides a theoretical basis and possible targets for the treatment of ischemic injuries in the brain. Neuronal synaptic function is closely related to learning and memory, and numerous studies have suggested that Npas4, Neuroligin-1, and N-cadherin are highly correlated with cognitive function regulation [23, 44–47, 49]. In order to explore the regulatory role of Npas4 on the transcriptional regulation of Neuroligin-1 and N-cadherin in cognitive function and whether it is involved in AD pathogenesis, we studied and found out the decrease in the transcription levels of NPAS4, NLGN1, and CDH2 in multiple brain regions of AD patients, and the decrease in the protein levels of Npas4, Neuroligin-1, and N-cadherin in the hippocampus of 3 × TG AD animal brains. These results are consistent with the results of multiple previous studies. One study found that Npas4 expression decreased in the brains of APP-knockout animals. This is because the intracellular domain of AICD, a cleavage product of APP, is a transcription factor for Npas4, and after APP knockout, AICD is also reduced, leading to a decrease in Npas4 expression [43]. Based on this study and our results, we speculate that in 3 × TG AD animal brains which carry the Swedish mutation of APP (APP KM670/671NL), MAPT P301L, and PSEN1 M146V mutations, although the Swedish mutation of APP is more susceptible to being cleaved by BACE1 to generate intracellular β-CTF (C99), β-CTF (C99) is extremely unstable and is rapidly cleaved into Aβ and AICD whose level changes in the 3 × TG AD animal brain are not yet clear [59–61], leading to a decrease in NPAS4 level. However, whether the reason for the decrease in Npas4 expression level is mediated by AICD or by other more important mechanisms also requires further study. In addition, reports show that the level of Neuroligin-1 in the cerebrospinal fluid of AD patients is decreased [49], and the neuroinflammation induced by Aβ deposition inhibits histone acetylation, enhances histone methylation levels, and decreases the expression of NLGN1 through epigenetic regulation [48]. In 3 × TG AD animal brains, Aβ plaques begin to deposit at six months of age [55], and our research results show that the protein level of Neuroligin-1 is also decreased at six and nine months of age. In the AD process, epigenetic regulation induced by neuroinflammation and the transcriptional regulation of response genes caused by the decrease of Npas4 may jointly lead to the decrease of Neuroligin-1 level, but the specific and major mechanism still needs to be explored. Numerous studies have shown that Npas4, Neuroligin-1, and N-cadherin play important regulatory roles in learning and memory, and our study also reveals that the loss of Npas4, Neuroligin-1, and N-cadherin may be an important link to synaptic plasticity damage in hippocampal neurons and cognitive dysfunction of 3 × TG AD mice. Before irreversible damage occurs, the re-expression of Npas4 can promote the transcriptional expression of Neuroligin-1 and N-cadherin, improve neuronal synaptic transmission and plasticity, and improve animal learning and memory. Our findings may provide new insights into the treatment of AD. Conclusion Our research for the first time confirms that Npas4, as a neuron activity-dependent transcription factor, promotes the transcriptional expression of two synaptic adhesion proteins, Neuroligin-1 and N-cadherin (Graphical abstract 1). During the early stage of ischemic brain injury, as the regulator of neuron activity, the expression of Npas4 increases, which promotes the expression of Neuroligin-1 and N-cadherin, which leads to an increase in the formation of neuron synaptic vesicles, and thereby an increase in synaptic transmission and synaptic plasticity, resulting in the enhancement of LTP, namely iLTP. This may be a self-repair process of neurons during cerebral ischemia. The current study also found that the transcription levels of NPAS4, NLGN1, and CDH2 were decreased in multiple brain regions of AD patients, and the protein levels of Npas4, Neuroligin-1, and N-cadherin were decreased in the hippocampus of 3 × TG AD mice. Overexpression of Npas4 can improve the synaptic plasticity of hippocampal neurons and subsequently improve the cognitive function of 3 × TG AD mice by upregulating the protein expression of Neuroligin-1 and N-cadherin. Therefore, in 3 × TG AD animals, the reduction of Npas4 and the subsequent decrease in the expression of Neuroligin-1 and N-cadherin is a key mechanism for synaptic damage leading to cognitive dysfunction in animals. Supplementary information Supplementary Figures Supplementary Figures